Determinants of reverse knowledge transfer for emerging market multinationals: the role of complexity, autonomy and embeddedness

Determinants of reverse knowledge transfer for emerging market multinationals the role of complexity, autonomy and embeddedness R F Q1 A S m s a o o e © P K R S s a F – 0 b 1 2 3 4 5 6 7 8 9 10 11 12[.]

Revista de Administração

http://rausp.usp.br/ Revista de Administração xxx (2016) xxx–xxx

emergentes: el papel de la complejidad, la autonomía y la integración

Franciane Freitas Silveiraa,∗, Roberto Sbragiab, Henry Lopez-Vegac, Fredrik Telld

Q1

aUniversidade Federal do ABC, São Bernardo do Campo, SP, Brazil

bUniversidade de São Paulo, São Paulo, SP, Brazil

cJönköping University, Jönköping, Sweden

dUppsala University, Uppsala, Sweden

Received 29 June 2016; accepted 19 August 2016 Scientific Editor: Maria Sylvia Macchione Saes

Abstract

Subsidiariesconductinnovationactivitiesinforeignmarketseithertocapturevaluableknowledgethatisnecessarytoadapttheirproductstolocal marketsortocreatevaluableknowledgeforheadquarters.Foremergingmarketmultinationals,moststudieshaveoverlookedthedeterminantsof successfulreverseknowledgetransferfromsubsidiarieslocatedinemerginganddevelopedmarkets.Thispaperanalyzedtheresponsesofasurvey administeredto78Brazilianmultinationalsthatownsubsidiariesindevelopedandemergingmarkets.Wefoundthatknowledgecomplexity devel-opedatthesubsidiary,itsautonomyandembeddednessintheforeignmarketdeterminethesuccessfulreverseknowledgetransfertoheadquarters

ofemergingmarketmultinationals.Thispapercontributestopreviousstudiesofreverseknowledgetransferbyunderlyingthemaindriversfor emergingmarketmultinationals

©2016DepartamentodeAdministrac¸˜ao,FaculdadedeEconomia,Administrac¸˜aoeContabilidadedaUniversidadedeS˜aoPaulo–FEA/USP PublishedbyElsevierEditoraLtda.ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/)

Keywords:Reverse knowledge transfer; Emerging multinationals; Brazilian multinationals

Resumo

Subsidiáriasrealizamatividadesdeinovac¸ãoemmercadosestrangeiros,querparacapturaroconhecimentovaliosoqueénecessárioparaadaptar seusprodutosaosmercadoslocaisouparacriarconhecimentodealtovalorparaasede.Nocontextodemultinacionaisdemercadosemergentes,

amaioriadosestudostêmnegligenciadoosdeterminantesdatransferênciadeconhecimetnoprovenientesdesubsidiárias(transferênciareversa) Foramanalisadasasrespostasdeumapesquisarealizadacom78multinacionaisbrasileirasquepossuemsubsidiáriasemmercadosdesenvolvidos

∗Correspondingauthorat:AlamedadaUniversidade,s/n◦–CEP09606-045,SãoBernardodoCampo,SP,Brazil.

E-mail:franciane.silveira@ufabc.edu.br (F.F Silveira).

Peer Review under the responsibility of Departamento de Administrac¸ão, Faculdade de Economia, Administrac¸ão e Contabilidade da Universidade de São Paulo – FEA/USP.

http://dx.doi.org/10.1016/j.rausp.2016.12.007

0080-2107/© 2016 Departamento de Administrac¸˜ao, Faculdade de Economia, Administrac¸˜ao e Contabilidade da Universidade de S˜ao Paulo – FEA/USP Published

by Elsevier Editora Ltda This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ).

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

eemergentes.Verificou-sequeacomplexidadedoconhecimentodesenvolvidonasubsidiária,bemcomoasuaautonomiaeinserc¸ãonomercado externodeterminam ofluxodetransferência reversadeconhecimento naempresamultinacional emergente.Estetrabalhoenriquece estudos anterioressobretransferênciareversadeconhecimentodestacandoosprincipaisdriversparaasmultinacionaisdosmercadosemergentes

©2016DepartamentodeAdministrac¸˜ao,FaculdadedeEconomia,Administrac¸˜aoeContabilidadedaUniversidadedeS˜aoPaulo–FEA/USP

PublicadoporElsevierEditoraLtda.Este ´eumartigoOpenAccesssobumalicenc¸aCCBY(http://creativecommons.org/licenses/by/4.0/)

Palavras-chave: Trasnferência reversa de conhecimento; Multinacionais emergentes; Multinacionais brasileiras

Resumen

Filialesrealizanactividadesdeinnovaciónenlosmercadosextranjeros,yaseaparacapturarelconocimientovaliosoqueesnecesarioparaadaptar susproductosalosmercadoslocales,oconelfindecrearconocimientodealtovalorparasusede.Respectoalasmultinacionalesdemercados emergentes,enlamayorpartedelosestudiosnosehadadoladebidaatenciónalosfactoresdeterminantesdelatransferenciadeconocimientoa partirdefiliales(transferenciainversa).Enesteestudioseanalizanlasrespuestasdeunaencuestarealizadaa78multinacionalesbrasile˜nasque poseenfilialesenmercadosdesarrolladosyemergentes.Losresultadosindicanquelacomplejidaddelconocimientodesarrolladoenlafilial,así comosuautonomíaeinserciónenelmercadoexternodeterminanelflujodetransferenciainversadeconocimientoenlaempresamultinacional emergente.Con estetrabajo,secolabora aldesarrollodelosestudios acercadelatransferencia inversadeconocimiento,conénfasis enlos principalesdriversparalasmultinacionalesdemercadosemergentes

©2016DepartamentodeAdministrac¸˜ao,FaculdadedeEconomia,Administrac¸˜aoeContabilidadedaUniversidadedeS˜aoPaulo–FEA/USP

PublicadoporElsevierEditoraLtda.Esteesunart´ıculoOpenAccessbajolalicenciaCCBY(http://creativecommons.org/licenses/by/4.0/)

Palabras clave: Transferencia inversa de conocimiento; Multinacionales emergentes; Multinacionales brasile˜nas

Introduction

Themultinationalenterprise(MNE)isadifferentiated

net-workinwhichitscontrolledsubsidiariesvarywidelyinterms

of duties andresponsibilities(Nohria &Ghoshal, 1994).For

example,whilesomesubsidiariesevolvethroughthe

headquar-ters’mandatesothersfocusontheirowninitiatives(Mudambi,

Piscitello,&Rabbiosi,2014).Sincethelate1990s,the

recog-nitionthat headquarters operate as knowledge receivers from

their internationally dispersedsubsidiaries hasgained

signifi-cance ininternational business research (Ambos, 2015).The

strategicimportanceoftheMNE’subsidiarieshascontinuedto

grow,inthatitisan accesspathwaytoknowledgeandtothe

technology situatedatthesubsidiaries’localmarkets (Borini,

Oliveira,Silveira,&Concer,2012;Criscuolo&Narula,2007;

Frost &Zhou, 2005), whichcanactively contribute to value

creationandsubsequentgainofcompetitiveadvantageforthe

entireMNE(Bartlett&Ghoshal,1989;Cantwell&Mudambi,

2005;Yang,Mudambi,&Meyer,2008)

An underlying idea is that MNE make use of knowledge

generatedbyforeignsubsidiaries Fromthisperspective,

sub-sidiaries upgrade their competence enhancing role such as

market expansion,costreduction andsupplieradaptation and

begintoplayamoreactiverole throughknowledge

develop-ment For example, foreign subsidiaries might develop new

products, newtechnologies, create new practices, new skills

thatwilllatershapetheirowncompetencecreatingpathwaysas

wellasaccumulatedifferentdegreesoftechnologicalcapability

(Birkinshaw,1997;Borinietal.,2012;Borini,Costa,Bezerra,&

Oliveira,2014;Cantwell&Mudambi,2005;Figueiredo&Brito,

2011;Frost,Birkinshaw,&Ensign,2002;Ghoshal&Bartlett,

1988; Govindarajan & Trimble, 2012; Mudambi, Mudambi,

&Navarra,2007; Nohria&Ghoshal,1997).Moreover,

com-petence creating subsidiaries could enhance their innovation

outcomes which enables them to compete domestically and internationally (Bell & Pavitt, 1995; Cantwell & Mudambi, 2005;Figueiredo&Brito,2011).Fromasubsidiaryperspective, reverseknowledgetransfer(RKT)givesvisibilitytosubsidiaries thatcouldleveragetheirstrategicpositioninthemultinational network(Borinietal.,2012;Holm&Pedersen,2000)

Thesefactorshavehighlightedthatreverseknowledge trans-ferisakeyvariableinthestudyofcross-borderknowledgeflows

inMNEs(Ambos,2015).Asaresult,theknowledgetransferin thereversedirection,thatis,fromsubsidiariestoMNE headquar-ters,hasemergedasaprominentthemeininternationalbusiness studies(Ambos,2015;Ambos,Ambos,&Schlegelmilch,2006; Q2

Criscuolo, 2005;Frost&Zou, 2005;Gupta&Govindarajan,

2000; Hakanson & Nobel, 2001; Rabiosi, 2008; Rabiosi & Santangelo, 2011;Rabiosi, 2011;Yangetal., 2008).Whilea numberofarticlesexploretheantecedents,successamountand successkey-factorsindifferentfunctionalconFigurationsatthe multinationalcorporation(Ambos,2015),additionalresearchis needed (Michailova&Mustaffa,2012).First,sincethe trans-fer ofknowledge inMNEs hasgrown considerablyin recent years,becomingthereforemorepronetovariousdefinitionsand measurementsofthesameconstructsresultinginconclusions, oftencontradictoryandambiguous.Second,whilerecognizing theimportanceofinvestigatingtherelationshipofthesubsidiary withexternalcompanieslocatedinthehostcountries,the liter-ature often focusesonly onthe research of knowledge flows within theMNE.Thisnarrowattentionconsiderssubsidiaries areprimarilyrecipientsofknowledge(Michailova&Mustaffa,

2012)

Fromemergingmultinationalsenterprises(EMNEs)’s view-point, the ability to transfer knowledge in reverse direction seemstobeevenmorecrucial Forexample, authorssaythat the EMNEsstrategicmodelsareguidedbythepursuitof for-eigncapabilities,suchastechnologicalknowledge,whichcan

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94

becombinedwiththe existingresources(Bartlett&Ghoshal,

2000).Thatissobecause,insteadofinternationalizingtoutilize

existingadvantages,emergingmarketmultinationalswill

inter-nationalizeaimingatacquiringnewadvantagesandcapabilities

(Guillén&García-Canal,2009;Mathews,2006;Ramamurti&

Singh,2009)andshouldtodoitquickerthantraditional

multi-nationalsdidintheirexpansionpaths(Mathews,2006)

InthecontextofBrazilianmultinationals,recentstudieshave

soughttounderstandtheprimaryfactorsthatinfluencetheRKT

ThestudyofBorinietal.(2012)arguesthatthereverse

knowl-edge transfer is a function of the strategic guidance of the:

(1)subsidiaries’R&Dlaboratories,(2) integration

(communi-cation) betweenheadquarters and subsidiaries, (3) subsidiary

entrepreneurialorientation,(4)subsidiarylifetimeand(5)entry

viagreenfieldinvestments.Moreover,thestudyofBezerraand

Borini(2015)teststheimpactthatanationdevelopmentexerts

onthereverseinnovationtransferinproductsandprocesses.In

thisstudy,theysoughttounderstandwhichdeterminantsofRKT

arepresentinBrazilianmultinationals.Inourstudy,weshowthat

RKTisrelatedtothedegreeof:(1)knowledgecomplexitythatis

being transferred, (2) subsidiary autonomy and (3) external

embeddednes.Ashighlightedbynumerousauthors(Mimbaeva

etal.,2007;VanWijk,Jansen,&Lyles,2008),suchaspectsare

Q3

identifiedaskeyonesforunderstandingtheRKTphenomenon

Althoughtherearemanykindsofknowledgetobetransferred

throughconventionaland/orreversedirection,thisstudyfocuses

specificallyonthetechnologicaltypeofknowledge(ofproduct

andprocess).Ourfindingsarebasedinananalysisofthesurvey

responsesadministeredto78Brazilianmultinationalsthatown

subsidiariesindevelopedandemergingmarkets

Asacontribution,itisexpectedthatourstudyadds

knowl-edgetotheinternationalbusinessestheory,sincetheknowledge

transferhasbeentreatedasakeyfactorofcompetitiveadvantage

ofMNEs(Borinietal.,2012,2014;Govindarajan&Ramamurti,

2011)and,specifically,oftheemergingmultinationals

compa-nies(Cuervo-Cazurra,2012;Immelt,Govindarajan,&Trimble,

2009;Ramamurti,2008).Sincemostresearchthatexplainsthis

phenomenonisbasedonMNEswithsubsidiariesand

headquar-tersindevelopedcountries,lessaffectedbyinstitutionaldistance

(Rabbiosi,2011;Yangetal.,2008),onecannotassumethatthe

Q4

factorsthatinfluenceRKTfromforMNEsarethesameasthose

forEMNEs(Borinietal.,2016).Apractitionercontributionof

thisstudyseekstoinformEMNEmanagersaboutthestrategic

driversofRKT

Thispaperisstructuredasfollows:thenextsection,

“Con-Q5

ceptualframework”sectionpresentstheproposeddeterminants

ofreversetechnologytransfer.“Methodology”sectionoutlines

our research strategy and field procedures “Findings”

sec-tionpresentsourresultsanddiscussestheimplicationsofour

findingsforfirmsinemergingmarkets.Finally,“Conclusion”

sectionpresentsourmainconclusions,somelimitationsofthe

study,andavenuesforfurtherresearch

Theliteraturearguesthatknowledgetransfer,whether

aris-ingfrominternalorexternalsources,hasanimportantimpact

onorganizationalperformanceandinnovationcapacity(Lyles

&Salk,1996;Powell,Koput,&Smith-Doerr,1996;Tsai,2001; VanWijketal.,2008).Theunderlyingideaisthatthetransferred knowledge contributes to the development of organizational capabilities that are difficult to imitateand can later lead to betterperformance(Szulanski,1996).Knowledgetransfer sti-mulates the combination of the existing knowledge with the newly acquired oneandincreasesthecapability of aunitfor carrying out new combinations (Jansen, Van Den Bisch, & Volberda,2005)

However, transferringknowledgebetween unitsof asame organizationisnoteasierthanconductingexternalknowledge transfers(Kogut&Zander,1992).Thisisparticularlythecase whenitcomestoRKT.Thisprocesscanbeevenmore challeng-ing,sincewhile“[ ]theconventionaltransferisaprocessof teaching,thereversetransferisaprocessofpersuading(Yang

etal., 2008)” Inthiscase, the effortismuch higherbecause its effectiveness depends on convincing headquarters There-fore,thetransferdependsonheadquarter’sassessmentthatthe featuresandrelevanceofthesubsidiary’sknowledgeiscrucial

sothatthereversetransferdoesoccur.TheRKTisdefinedas

“an intra-organizationalexchange of information, technology

or know-how frominternational subsidiaries(located in host countries)tocorporateheadquarters(homecountries).Theterm

‘reverse’ is usedto distinguishthesetransfers fromthe more conventional formof ‘forward’transfers – fromheadquarters

tosubsidiaries–,and‘lateral’transfers betweensubsidiaries” Q6

(Ambos,2015)

Somestudieshavehighlightedthatsubsidiariescreate com-petitive advantages for MNEs when valuable knowledge is transferred to the headquarters (e.g Gupta & Govindarajan,

2000;Hakanson&Nobel,2001;Rabbiosi,2011; Yangetal.,

2008).ForMNEs,someof the determinants ofRKT include the:(1)knowledgefeaturesbeingtransferred(Mimbaevaetal., 2007),(2)organizationalcharacteristics(size,age,autonomy) (Frostetal., 2002;Gupta&Govindarajan,2000),(3) roleof organizational mechanisms (Hakanson & Nobel, 2001; Rab-biosi, 2011), (4) the subsidiaries’ roles (Ambos et al., 2006; Rabbiosi, 2011;Yang etal., 2008),(5) thehost country eco-nomicdevelopment(Cantwell&Mudambi,2005;Frostetal., 2002;Gupta&Govindarajan,2000),(6)theabsorptivecapacity (Ambosetal.,2006),(7)theknowledgerelevance(Yangetal.,

2008),(8)theinternalembeddedness(subsidiary/headquarters) and (9) the external embeddedness (subsidiary/partners) (Figueiredo,2011;Meyer,Mudambi,&Narula,2011)

Following, it is explained how knowledge characteristics, suchascomplexity,autonomy,andexternalembeddedness influ-enceRKTfromsubsidiariestoheadqueartersofEMNEs

Subsidiary’s autonomy

Subsidiary’s autonomy could be defined as the extent to which a subsidiary is allowed to make decisions on its key strategic issues(Mudambi &Navarra,2004),withouta head-quartersdirectintervention(Roth&Morrison,1992).Ahigher level of autonomyisoftenrelatedtoknowledge creationand development at the MNE (Ghoshal & Nohria, 1989; Gupta

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197

198

199 200 201 202 203 204

&Govindarajan,1991;Nohria &Ghoshal,1994),since

inde-pendentsubsidiaries,(1)havestrategicmandates(Birkinshaw,

Hood,&Jonsson,1998),(2)make quickdecisions(Cantwell

&Piscitello, 1999),(3)recognizeandtakeadvantageof local

opportunities(Frostetal.,2002),(4)developnewknowledgeas

oflocalknowledgebases(Andersson,Forsgren,&Holm,2002),

(5)generateintrinsicmotivationonindividuals(Mudambietal.,

2007),(6)haveinitiativeandwillingnesstosharetheknowledge

acquired(Gupta&Govindarajan,2000;Tsai,2002).Onthe

con-trary,alowlevelofautonomy,maylimitthesubsidiaryfreedom,

hindering its knowledge creation anddevelopment capability

(Ghoshal & Bartlett, 1988) Foss and Pedersen (2002) also

explainthathighlevelsofsubsidiaryautonomy–associatedwith

lossofcontrol–couldbeovercomedbytheincreasein

knowl-edge exchange amongst subsidiaries While opposite results

havealsobeenreported(Frostetal.,2002;Gammelgaard,Holm,

&Pedersen,2004),mostresearches havesuggestedmainlya

positiverelationshipbetweenknowledge decentralization and

transfer(Cantwell&Mudambi,2005;Foss&Pedersen,2002;

VanWijketal.,2008)

Recently,Rabiosi(2008)arguedthat RKTiscoupledwith

subsidiary autonomy, i.e mechanisms of personal

communi-cation between subsidiary and headquarters Yet, regarding

subsidiaries of EMNEs, it is argued that, dueto their recent

progressintheinternationalmarketand,therefore,duetotheir

earlyage, theyarestronglydependent onheadquarters’

deci-sionmakingpower(Dunning,1993).Thismightnotbedifferent

inBrazilianMNEs,thattendtobemorecentralizing,limiting

therefore their subsidiaries’ knowledge creation possibilities

(Chu&Wood,2008).Thisisanunfavorablesituationforthe

developmentofexistingandnewknowledgeattheheadquarter

However,inthesamewayastraditionalMNEs,the

international-izationprocessofEMNEsrequiresthecapabilitytoacquireand

developknowledge(Mathews,2006).Hence,subsidiariesplay

acentralroleinthepursuitofnewknowledge(Borini&Fleury,

2011) Different authors state that EMNEs survival depends

evenmoreheavilyonresourcesthathavebeendevelopedabroad

whencomparedtothemultinationalsfromdevelopedcountries

(Guillén&García-Canal,2009;Mathews,2006).Therefore,this

studyadvocatesthatsubsidiaryautonomyiscriticalforRKTin

EMNEs,whichallowsustohypothesizethat:

H1. The greater the subsidiary autonomy, the greater the

reverseknowledgetransfer

Knowledge complexity

The increasing specialization and sophistication in R&D

requires companies to integrate distinct knowledge areas to

developnewproducts.Asaresultknwoledgeturnstobehighly

complexanddifficulttoconductintra-knowledgetransfers.A

paradox emerges: the greater the number of functional areas

andscientific disciplines necessary todevelop newproducts,

themorecomplexitistotransfertheknowledge(Ciabuschi&

Martín,2012).Knowledgecomplexityisassociatedtothe

ampli-tudewhichistheextentofspecializationfields(Grant,1996)and

theambiguityofthereferencedknowledge(Reed&DeFillippi,

1990) The greater the number of techniques, organizational routines,peopleandresourcesinvolvedthatareconnectedtoa particularknowledge,themorecomplexitbecomes.These con-ditionsmoderatetheinformationamountthatmustbeprocessed fortheunderstandingofcomponentsinvolved(Simonin,1999) Thus, management scholars tend to agree on the idea that complexity hinders knowledge transfersince it decreases the receiver’sabilitytoidentify,understandandintegratethe knowl-edgetobeacquired(Simonin,1999).Yet,oppositeresultshave been found in the literature (Mimbaeva et al., 2007) Since, complexknowledgeisthemostvaluabletothecompany’s com-petitiveness.Studieshaveshown,forexample,thatglobalteams areabletosharecomplexknowledgethroughrulesandcodes commontotheexchangingarea(Reddy,2008) Q7

IntheEMNEperspective,additionaleffortstosharethiskind

of knowledgecanbe advantageoussince,as itsimitationand substitution is hampered,it may be usefultothe building of strategic capabilities (Nair,Demirbag, &Mellahi, 2015)due

to the prevailing need to use the available foreign resources (Mathews, 2006) A study conducted in Indian multination-als, for example, found that RKT happens regardless of the knowledgecomplexity.InBrazilianMNEs,itissuspectedthat onlylesscomplexknowledgefromsubsidiariesistransferredin reversedirection,consideringthattheforeignsubsidiariesrole

isdeterminedbytheBrazilianheadquarters(Galina&Moura,

2013) whichstill holds greatercentralization inthedecisions andinnovations.Accordinglyit wasformulatedthefollowing hypothesis:

H2. Thelowerthecomplexityofthesubsidiary’sR&D knowl-edge,thegreaterthedegreeofreverseknowledgetransfer

Local embeddedness

EmbeddednesisrelatedtothenotionthatMNE’scompetitive performancecanbefacilitatedthroughthesocialrelationships they create with several business players such as customers, universities and local research institutions (Grabher, 1993; Granovetter, 1985; Uzzi, 1996) More specifically, embed-dedness refers to the mutualadaptation of activitiesbetween two companies as much as a common understanding of the collective targets and appropriate ways to work in a social system (Tsai & Ghoshal, 1998) Therefore, it is considered

as astrategic resourcefor MNEs.It provides easyaccess to the resources and capabilities that are outside the company (Anderssonetal.,2002;Uzzi&Gillespie,2002)thatareable

to generate a large knowledge transfer among the partners (Figueiredo,2011;Uzzi&Gillespie,2002)

Thedegreeofembeddednessbyforeignsubsidiaries, mea-sured by the proximityto localpartners, reflects subsidiary’s abilitytoabsorbknowledgefromitslocalnetwork,which some-timesmightresultinnewknowledgecreation(Andersonetal., Q8

2002).Thisscenariotendstodirectlyfostersubsidiary’s innova-tivecapacity,i.e.improvementofexistingproductsandservices

or newproduct,service,technologydevelopment(Andersson,

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286

287 288

289

290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309

etal.,2002;Håkanson&Nobel,2001;Yamin&Otto,2004)

Indirectly,localsubsidiaryembeddednesscanfosterknowledge

transfertootherMNE’sunits(Powelletal.,1996;Yamin&Otto,

2004),constructing,inturn,thesubsidiary’spowerrelationships

withintheMNE(Andersson,Forsgren,&Holm,2007)

Higher levels of subsidiary embeddedness are related to

an understanding of the context in which the local

knowl-edgeresides Frequently,subsidiariesinteract withits closest

networkof local companiesandinstitutions inordertolearn

aboutcustomersandtechnologiesand,therefore‘capture’the

localknowledge(Figueiredo,2011).Subsequently,itmustuse

theconnectivityalreadyestablishedwithin theMNEnetwork

for transferring the knowledge in reverse direction (Meyer

et al., 2011; Najafi-Tavani, Giroud, & Andersson, 2013)

RegardingMNEs of emerging markets,Child andRodriguez

(2005),Mathews(2006)andLuoandTung(2007)emphasize

the importance of relationships and knowledge opportunities

availableatsubsidiarieshostingmarkets.Forexample,

provid-ing easy access to technologies found in developed markets

(Figueiredo,2005).RamamurtiandSingh(2009,pp.126–127)

showthatEMNEscanpursueseveraldifferentstrategies,such

as“low-costpartners”,“globalconsolidators”and“globalfirst

movers”.Basedonthesearguments,thefollowinghypothesisis

suggested:

H3. Thegreatertheembeddednessofaforeignsubsidiary,the

higherthereverseknowledgetransfer

Methodology

Sample and data collection

Thesample ofthisstudyconsists of BrazilianMNEswith

manufacturing,salesorR&Dsubsidiariesabroad.Weexpected

thatsubsidiarieswithmorestrategicactivitieswouldhavemore

opportunitiesoftransferringknowledgetotheheadquartersin

reversedirection.Thedatawascollectedusinganeletronic

sur-veywithBrazilianMNEssubsidiariesestablishedabroad(see

Appendix1).Duetothenon-existenceofanofficialnumberof

Brazilianmultinationalsowningsubsidiarieswitheither

man-ufacturingorR&Dcentersinstalledabroad.Thefirststepwas

toidentifyBrazilianmultinationalspresentingthese

character-isticsfromsecondarydatasources,suchasGINEBRAProject

(Management System for the Internationalization of Brazilian

Enterprises) that resulted in the publication ‘Business

Man-agementfor the Internationalization of BrazilianCompanies’

(coordinatedbyFleury,2010),anannualsurveyoftheFundac¸ão

DomCabral(Dom Cabral Foundation),ValorEconômico

(Eco-nomic Value),andSOBEET(Brazilian Society of Transnational

Corporations)surveysaswellasdatafromtheBrazilian

Multi-nationalsObservatory(Center of Brazilian Multinationals)of

theESPM(School of Higher Education in Advertising and

Mar-keting).

Inthissecondarysources,63multinationalcompanieswere

listed, being possible to identify 240 subsidiaries with

for-eign manufacturing operations and/or R&D centers Of this

population,39Brazilianmultinationalsparticipatedinthe sur-vey (61.9%), with 78 responses, corresponding to 32.5% of allsubsidiaries.Thismeansthatinsomecasesresponseswere receivedfrommorethanonesubsidiaryperheadquarter

Intheattempttoidentifypossibleshortcomingsor misunder-standingsinthesurvey,apre-testwasconductedtogetherwith specialists from academiaandindustry (Cooper&Schindler,

2003)whichhelptogeneratenewinsightsandadjustmentsin thequestionnaire.Following,theelectronicsurveywassentto participants,withafollow-upphone-calltoclarifyanyquestions fromrespondents.Thetotalperiodofdatacollectionwasfive months,from October 2013up toFebruary2014 Responses werecollectedfrom R&Doffices andthe respondentsranged fromsubsidiarydirector,internationalbusinessandR&D direc-tor,andengineeringmanagers

Measures Dependent variable

Thedependentvariable(reverseknowledgetransfer–RKT) represents, overthelastthreeyears, therate ofRKTof tech-nologyand market knowledge that the subsidiary transferred back tothe headquarters.In ordertodetailthetypesof tech-nologicalcontent,itwasappliedtheIammarino,Padilla-Pérez, andVonTunzelmann(2008)scale,whichwasvalidated previ-ouslybyotherauthors(Lall,1992;Bell&Pavitt,1995;Ariffin andFigueiredo(2003)),whorankthetechnologicalknowledge Q9

transferintermsofproductandprocess.Onafive-pointscale (rangingfrom1“notatall”to5“toaverygreatextent”).For ensuring therobustnessresults,itwas alsoinsertedadummy variablewhichallowedtherespondenttoindicatethecasesin whichthe subsidiaryhadneverdone orhaddone thereverse transferofaspecificproductorprocessknowledge(0or1)

Independent variables

The knowledge complexity construct measures the num-ber of interdependenttechnologies, routines, individuals,and resources linkedtoaparticularknowledge orasset(Simonin,

1999).Moreover,thecomplexityconstructwasmeasuredusing

a six-item Likert scale based on responses (1=strongly dis-agree;5=stronglyagree)(adaptedfromSimonin,2004;Zander

&Kogut,1995).Thesubsidiaryautonomymeasureindicatesthe extenttowhichasubsidiaryisallowedtomakedecisionsabout itskeystrategicissues(Rabbiosi,2011).Themeasureof sub-sidiaryautonomywasbasedonascaleoriginallydevelopedby

GhoshalandNohria(1989)andlaterusedbyBirkinshawetal (1998)andRabbiosi(2011).Afive-itemLikertscaleassessedit Thesubsidiaryembeddednessindicatesthecollaborationdegree withthelocalnetworks.Inparticular,thisstudyfocusesonthe subsidiary embeddedness withlocalcustomersandsuppliers ThisconstructwasdevelopedbasedonAnderssonetal.(2002, 2005).Afive-itemLikertscaleassessedit

Controls variables

The MNE literaturesuggestsseveralfactorsthat might be correlatedtoRKT.Inparticular,itisexpectedthatsubsidiaries

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363 364 365 366 367 368 369 370 371 372 373 374 375 376 377

378

379 380 381 382 383 384 385 386 387 388 389 390 391 392 393

394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411

412 413 414

morelikelytotransferreverseknowledge

Subsidiary location. Thehostcountryhasbeenrelatedto

fac-tors that impact the subsidiary development and positioning

(Birkinshaw & Hood, 1998; Gupta & Govindarajan, 2000;

Mudambi&Cantwell,2005;Rabiosi,2011)aswellasthenature

Q10

ofRKT(Yangetal.,2008).Particularly,thishappensbecausethe

subsidiary’scapabilitiesandskillscouldreflectthecountry

tech-nologicalandinstitutionalforces,suchaslegalandinstitutional

factors(forexamplepatentprotectionandindustrialincentives)

that ensure the proliferation of innovation The assumption,

therefore,isthatcompaniesinemergingmarketsgetinvolvedin

lessinnovationthancompaniesindevelopedmarkets,duetothe

lackofhightechnologyinemergingmarkets(Vernon-Wortzel&

Wortzel,1998).Thus,thehighertheeconomicdevelopmentof

thesubsidiary’shostcountry,thegreaterthebenefitsearnedby

theheadquartersarisingfromthetransferredknowledge(Frost

etal.,2002)

Foremerged market MNEs, subsidiarieslocated in

devel-opedor high-incomecountriescanimpacttherate andspeed

ofRKT,sincetheresourcesavailableinthesemarketscanhelp

increasetheheadquartersbreadthandnovelty(Mathews,2006)

Onthecontrary,Aulakh(2007)andCuervo-CazurraandGenc

Q11

(2008)arguethatemergingmarketMNEshavethesame

knowl-edgeresourcesthanthoseoperatingindevelopedcountries.In

thespecificcaseofBraziliancompanies,Bezerraetal.(2015)

Q12

have concluded that MNEs’ Brazilian subsidiarieslocated in

developedcountriestransfermoreknowledgeinreverse

direc-tion thansubsidiarieslocatedin emergingcountries.In order

tocapturethesubsidiarylocationeffectsonthelevelsofRKT,

thedummyvariablelow-incomecountries(0)andhigh-income

countries(1)wereaddedtothemodel

Subsidiary’s age. Moreancient subsidiariescouldhavesome

advantagesoverneweronesdueto(1)theincreased

informa-tionandresources,(2)thehigherdevelopmentofR&Dskills,

(3)acquiredexperienceandexpertise,and(4)increasedlearning

curveeffects.Thereforetheymightbelessdependenton

knowl-edge from headquarters (Foss & Pedersen,2002; Yamin and

Otto,2004).Previousstudiesshowbothpositiveandnegative

effectsoforganizations’sageregardingthelearningand

innova-tionoutcomes(Sørensen&Stuart,2000).Whilepositiveeffects

are justified by the knowledge increase, accumulated

experi-ence and possession of stronger relationships with suppliers,

andcustomersthatenabletheinnovationprocessimprovement

(Cohen&Levinthal,1990).Negativeeffectsareassociatedwith

upgrade difficulties of more mature companies with external

technologicaladvances,attheriskofbecominginertandlimited

forlearningandadaptingtonewcircumstances.Inotherwords,

thereisalossofinnovativecapacity(Sørensen&Stuart,2000;

Tushman&Anderson,1986)

Inthisregard,intheBrazilianmultinationalscontext,Bezerra

etal (2015)found that theyounger asubsidiary, the greater

its extent of RKT.Thus, despite the inconclusivefindings of

subsidiaries’age,itisexpectedthatoldersubsidiariesaremore

likelytodevelopandtransferbackknowledgetoheadquarters

than recentlyestablished subsidiaries Particularly, duetothe periodofexistenceofBraziliansubsidiariesismuchlowerwhen comparedtoemergedmarketsubsidiaries.Inordertocapturethe subsidiaryageeffectsonthelevelsofRKT,thedummyvariable young(0)andoldsubsidiary(1)wereaddedtothemodel.The detailsofeachvariable,includingindicatorsandauthors,used

asbackgroundispresentedinAppendix1

Data analysis

Adescriptiveanalysiswascarriedouttoidentifythe frequen-ciesofrespondents’answersforallconstructscomprisedinthe survey.ThePartialLeastSquare–StructuralEquation Model-ing(PLS-SEM)wasusedtoassessthedeterminants’influence

of RKT(Hair,Hult,Ringle,&Sarstedt,2014).The structural modelwasestimatedonSmartPLS3.0(Ringleetal.,2014)using Q13

the ‘path’weightingscheme.Thedecisiontousethismethod took intoaccount anumber of criteria,including (1) the fact that theindicatorsdonothaveanormaldistribution,whichis oneoftheassumptionsfortheuseofthemaximumlikelihood method(ML);(2)theuseofintervalscales(Joreskog&Wold,

1982);(3)itsabilitytodealwithmorecomplexmodelsas com-pared toLISREL(Henseler,Ringle,&Sinkovics,2009);and (4)thesmallsamplesize

Since the PLSalgorithm formulation(Hui &Wold,1982; Lohmöller, 1989) is recognized that it is biased and is only

“consistent at large”,which meansthat the bias decreases as the numberof indicatorsby latentvariable isincreased.This issueoccursbecausetherelationshipsamongstlatentvariables (correlations andpathcoefficients) are estimated as from the factorialscores,whichareobtainedasasumoraweighted aver-ageoftheirindicators,includingthemeasurementerrors.This factistreatedascorrelationattenuationinthemethodological references relatedtopsychometrics,forexample(Nunnally& Bernstein,1994,p.212).However,despitethisbias,Hairetal (2014,p 79)mentionsomesimulationswhereit isidentified thatthebiasissmallforpracticalpurposes.Forfourandeight indicatorsbylatentvariable,ChinandNewsted(1999,p.333)

foundabiasequalto0.05.Tominimizethisbias(attenuation)the latentvariablesweremeasuredwithfivetosixindicatorseach, reachingreliabilityvalues(compositereliabilityandCronbach’s alpha)higherthan0.8(Table2)

Additionally, toassessthisbiassize,thedisattenuated cor-relations were calculated (or “correction” for attenuation as explained by Nunnally and Bernstein (1994, p 241) of the dependentvariable(RKT)withtheotherindependentvariables (Table1)

Itisobservedthatthehighestbiaswasequalto0.053.Asthis

isasmallbiasforpracticalpurposesandisintheconservative direction(underestimatingthepopulationparameter),theresults wereconsideredadequateforpurposesofresultsinterpretation fromthe pointof viewofstatistical significanceandpractical importance

Anotherwaytocheckthesamplesizeadequacyisthrough analyzing the statistical power sensitivity, performed with G*Power3software(Faul,Erdfelder,Buchner,&Lang,2009) Forasampleof78respondents,withasignificancelevelof5%

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470 471 472 473 474 475 476

477

478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524

Table 1

Disattenuated correlations of the dependent variable RKT.

Disattenuated correlation with RKT (using CR) 0.161 0.375 0.307 0.343 0.463

andstatisticalpowerof0.80(Cohen,1998),thetest‘sensitivity

analysis’foundthat themodel isabletodetectaneffectsize

of0.1574,whichisconsideredamediumeffect(Cohen,1998)

Using the population effect formula f2=R2/(1−R2) (Cohen,

1998),itwasconcludedthattheresearchwilldetectaminimum

R2of0.1507

Findings

TherespondentsincludedalargevarietyofBrazilian

multi-nationals ranging from natural resources (12%), consumer

goods(21%),basicinputs(32%),manufacturing(19%),system

assembly(10%)andrawmaterialsforconstruction(6%).The

respondingsubsidiarieslocationswere:LatinAmerica(42%),

NorthAmerica(24%),Asia (14%),Europe(14%)andAfrica

(5%).Atthe countrylevel, thelargestnumberof subsidiaries

areintheU.S.(15%),Argentina(15%),Colombia(10%)and

Mexico (9%) Moreover, China (6%) already appears as an

importantdestinationforBraziliansubsidiaries.Astothesize

andnumberofemployeesatthesubsidiary,56%ofresponding

subsidiariesareintherange100–1000employees,followedby

14%ofsubsidiariesemployingmorethan1000workers.This

descriptive statistics shows that a relative percentageof

sub-sidiariesconsistofconsolidatedcompaniesabroad.Asregard

tothesubsidiaries’age,themajority(69%)isundertenyearsof

age,22%arebetweentenandnineteenyearsandonly9%are

morethan20yearsof activities.Theentrymodeof Brazilian

subsidiariesabroadrepresents77%acquisitionsand23%direct

investmentorgreenfieldinvestment

Evaluation of the measurement model

In measuring the constructs, the model was conducted by

evaluatingtheconvergent,discriminantandreliabilityvalidity

AspresentedinTable1,theconstructs(alsocalledlatent

vari-ables)weremeasured using reflectiveindicatorstoverify the

adequatereliabilityoftheCronbach’salphavalues.Inaddition,

all latent variablesachieved convergent validity, that is, they

haveanaveragevarianceextracted(AVE)higherthan0.5,and

compositereliabilityhigherthan0.7(Hairetal.,2014;Henseler

etal.,2009;Tenenhaus,Vinzi,Chatelin,&Lauro,2005)

How-ever,threeitemsofthescaleshadtoberemovedfromthemodel

sothat the AVEreached the referencevalue (3.7;4.2;4.4 in

Appendix1).Thediscriminantvalidityisverifiedbythe

For-nellLarckercriterionandwasevaluatedthroughthecross-loads

analysis Thisfacilitated todetermine whether a constructis

trulydistinctfromotherconstructsthroughempiricalpatterns

Basedonthisresult,itwasnotedthatallcorrelationsamongstthe

latentvariablesweresmallerthanthesquarerootoftheaverage varianceextractedoftheirlatentvariables(Fornell&Larcker,

1981).Thus,itcanbesaidthatthemodelpresentedconvergent, discriminantandreliabilityvalidity.Themeans,standard devi-ations,reliabilityestimatesandfactorcorrelationsarereported

inTable2

Assessment of the structural model

Thestructuralmodelisabletospecifytherelationship pat-ternsamongsttheconstructs.Themodelwasassessedusingfive criteria:(i)pathcoefficients(β);(ii)pathsignificant(p-value);

(iii) varianceexplain(R2); (iv)effectsize (f2)and(v) predic-tiverelevance(Q2).AccordingtoHairetal.(2014),themain criteria for the structural model evaluationarethe coefficient

ofdetermination(R2)andthelevelandsignificanceofthepath coefficients(β).Tocalculatethem,thepathweightingscheme andabootstrappingtechniquewereusedwith78observations and 500 randomsamples toestimate the t-values inorderto assessthe significance.Forsocialscience researches,R2 val-uesof0.26,0.13and0.02areconsideredstrong,moderateand weak,respectively(Cohen,1998)

Continuing Fichman and Kemerer (1997), in addition to the full model, we have evaluated two nested models (con-trolmodelandtheoreticalmodel).Intotal,thesethreemodels were accessed to evaluate the true impactand the additional explanatorypowerofthetheoreticalvariablesafterthevariance explainedbythecontrol.Thefullmodelincludesallthisstudy variables,thecontrolmodelincludesonlythecontrolvariables, andthe theoretical model includesthe hypothesized relation-ships.Comparisonsamongstthethreemodelsaresummarized

inTable3 TheR2valueresultsforthefullmodel(includingcontrol vari-ables)indicatethatthevarianceof36%inRKTwasexplained

bythemodel.Thisresultisconsideredsubstantialandprovides evidencethat the model iscapableof explainingthe depend-entvariable(Cohen,1998).Whencomparingtheresultsofthe adjusted R2 (33%) with the sensitivity analysis on statistical power,itisfoundaR2valuewellabovetheminimumdetectable

bythemodel,whichis15%

A comparison between the full model and control model (location and age) shows that the control model explains an incrementalvarianceonR2 of19%onthedependentvariable (RKT).Thedeltabetweenthecontrolmodelandthefullmodel was(R2=0.17).Thisresultsuggeststhat,despitehaving pre-sentedamoderateresult,controlvariablesalonedonotprovide

asolidbasisthroughwhichonecanunderstandandpredictRKT patterns

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569 570 571 572 573 574

575

576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614

Table 2

Evaluation of the measurement model.

Note 1: In bold on the diagonal, there are values of the square root of the average variance extracted.

Note 2: AVE, average variance extracted; C.R., composite reliability; C.A., Cronbachs alpha.

Note 3: AVE benchmarks: 0.5; composite reliability: 0.7; Cronbach’s alpha: 0.6.

Table 3

Significance test results of the structural model path coefficients.

Values oftwere calculated through bootstrapping with 500 resamples and 78 cases per sample.

* p≤ 0.05.

**p≤ 0.01.

***p≤ 0.001

Comparing the full model and the theoretical model, the

incrementalvariance derivedbythemodelisaround 30%for

RKT.Resultsindicatethatthetheoreticalmodelinthisstudyis

substantiveenoughtoexplainthevarianceintheresearchmodel

However,controlvariableswereresponsibleforaconsiderable

proportionofthevarianceintheR2valueofRKT.Asthe

pre-dictedpathsforthestructuralmodel,allthehypothesizedwere

statisticallysignificant.The confidencelevel intheprediction

modelwasmeasuredbytheindicatorQ2whichmustbehigher

thanzero TheQ2 valuetoconstruct‘RKT’is0.171ensuring

themodelpredictiverelevance(Hairetal.,2014;Henseleretal.,

2009)

Theeffectsize(f2)measuresthemagnitudeofan

indepen-dent variable on adependent variable (Tabachnick& Fidell,

2007) The exogenousconstructs omission of the model can

beusedtoassessinwhichcasetheseomittedconstructshave

substantialimpactontheendogenousconstructs.Cohen(1998)

providedvaluesof0.02,0.15and0.35consideredweak,

mod-erateandstrong,respectively Thef2 isalsocalculatedbyR2

included=f2−R2excluded/1−R2included(Hairetal.,2014)

Following,Table3showsthesignificanceresultsofeachpath

amongstthelatentvariablesandtheeffectsize

Theresultssupporttwoofthethreehypothesesstatements

Hypothesis H1 shows that autonomy has a positive and

sig-nificant effect on reverse transfer (β=0.19, p≤0.05) The

effectsize (f2)of 0.05 indicates that the constructsubsidiary

autonomyhasaweakeffectontheendogenouslatentvariable

RKT(Cohen,1998) HypothesisH2 states thatthe lowerthe

complexityofsubsidiary’sR&Dknowledge,thelargertherate

of reverse technology transfer to headquarters Surprisingly, this study’s results showed that knowledge complexity hasa significant, but positive effect on reverse transfer (β=0.25,

p≤0.01) Thisrelationshipischaracterized byaweakeffect (0.08)ontheendogenouslatentvariable‘RKT’(Cohen,1998) Finally,theresultsshowedthatsubsidiaryembeddednesshasa significant andpositiveeffect(0.15)onRKT,whichconfirms

H3hypothesis(β=0.35,p≤0.001).Thisrelationshipis charac-terizedbyamoderatetostrongeffectontheendogenouslatent variable‘RKT’(Cohen,1998)

With regardtothe controlvariables,the localizationeffect waspositiveandsignificant(β=0.19,p=0.05)forRKT, indi-catingthatsubsidiarieslocatedindevelopingcountriesaremore likelytotransferknowledgeinreversedirection.Alsofor the subsidiaryagevariablethecoefficientissignificant(β=−0.18,

p=0.01)butthenegativesignindicatesthatRKTismorelikely

to occur from young subsidiaries, confirming the findings of Bezerraetal.(2015)

Discussion

Despite ambiguousevidenceabout RKT inBrazil(Fleury

& Fleury, 2011),this study found that Braziliansubsidiaries withahighautonomydegreearemorecapableoftransferring knowledgebacktoheadquarters,confirmingourhypothesisH1

An argument on the positiveeffect of autonomy for RKT is basedonthe ideathatthesubsidiariesindependenceprovides

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661

662

663 664 665 666 667 668

greateraccess tolocalknowledge databases,knowledgefrom

localpartners and possibilities toinnovate (Anderssonet al.,

2002;Ciabuschi&Martín,2012;Gupta&Govindarajan,1991;

Mudambi&Cantwell,2005).Hence,subsidiary autonomyis

recognizedasanimportantpredictorofreverseknowledge

trans-ferinthecontextofEMNEs.Autonomyempowerssubsidiaries

toexploretheirownbusinessandmarketopportunitiessothat

they can make use of external sources to their competitive

advantage Taking into account that Brazilian multinationals

are still at an early stage of internationalization, it is a new

phenomenon the fact that their subsidiaries have been

seek-ingfor autonomy andindependencefrom their headquarters’

decisions

Thispaperidentifiedthatknowledgecharacteristicsand

sub-sidiarycharacteristicsdeterminetherateofreverseknowledge

transfer from subsidiaries to emerging market MNEs First,

fromtheknowledgecharacteristicsviewpoint,itwaspossibleto

showthattheknowledgecomplexitylevelhasapositiveimpact

ontheextent of RKT.Thisfinding iscontrarytothisstudy’s

hypothesis(H2),whichsuggestedthatthelowerthesubsidiary

knowledgecomplexity,thegreatertheRKT.Itissuspectedthat

one of the reasons for this intriguing, but interesting result,

mayberelatedtotheknowledgecomplexityparadox,because,

whileknowledgetransferencountershighercostsproblems,itis

themostcompensatorytypeofknowledgetotheheadquarters

Thus,itissuspectedthattheBrazilianmultinationalstryto

trans-ferthemostcomplexknowledgedevelopedintheirsubsidiaries,

regardlessofthecomplexitylevelsassociated,whichincludes

theinvolvementtoagreaterextent,oftheheadquarterssothat

thistype of transferactually materializes (Nairet al., 2015)

Sucharesultisalsoinlinewiththeframeworkoflearningand

effectiveleverage (Mathews,2006) ofthe EMNE’s resources

andnetworksabroad(LLLchart).Otherpossibleexplanationis

theeffectofsubsidiary’srole.Forexample,moreinnovative

sub-sidiariesmighttransfermorecomplexR&Dknowledge,which

suggeststhatimplementerandcontributorsubsidiariesmaynot

transfer(ortransfertoalesserextent)complextypeknowledge

Insummary,althoughtheinitialH2wasnotsupported,thisresult

providesanopportunitytosuggestthatinnovativesubsidiaries

mayengageincomplexknowledgetransferandthusbecomea

competitiveplayer

OurresultsalsosupportthehypothesisH3whichproposes

thatlocalembeddednessimpactstherateofRKT.Itwasfound

thatembeddednesswithsuppliersandcustomers,inotherwords,

localbusinessnetworksincreasethepossibilityofgainingaccess

tonewknowledge, whichcan subsequentlybe transferred to

EMNEs This paper confirms that subsidiaries from

emerg-ing market multinationals become internationalized in order

toexploreknowledgeandexistingcapabilitiesinforeign

mar-kets as well as to develop new knowledge and capabilities

throughknowledgeavailableinthesubsidiaries’host

environ-ment(Narula,2012).Forsubsidiariesisessentialtobeembedded

inlocalbusinessnetworkstoobtaindistinctiveknowledge

devel-opment.Newconnectionswithlocalnetworksallowsubsidiaries

toperforminnovativetasks forheadquarters, instead oftasks

limited to adaptation of products and processes to the local

market(Borini&Fleury,2011)

Withregardtothefirstcontrolvariable(location),theresults indicatedthatsubsidiarieslocatedindevelopedmarkets,such

asNorthAmericaandEurope,areprobablytheonesthatmost transferknowledgetotheirheadquarters.Thisresultisinline withseveralcontributionsintheliteraturewhichstatethatthe innovationcapacityofsubsidiarieslargelydependsonthehost countriesadvantages(Gupta&Govindarajan,2000;Mudambi

&Cantwell,2008;Yangetal.,2008)

BasedonpreviousfindingsaboutEMNEs,twoperspectives canbepresented.Thefirstperspective,ledbyCuervo-Cazurra andGenc(2008),Ramamurti(2008),KhannaandPalepu(2011),

Cuervo-Cazurra (2012) and Ramamurti (2012), argues that EMNEs have anew typeof capability,unlike the traditional MNEscapabilities,whichisrelatedtotheabilityofcopingwith the institutional deficienciesto whichtheyare exposed This current advocates that emerging MNEs, for having operated

inenvironmentspresentingdifficultconditions,suchas under-developed premises, corrupt bureaucracies, poor educational institutions and unstable governments, have the “advantages

of adversity.”The secondperspective, ledbyauthors suchas

Mathews (2006) andChild andRodriguez (2005),arguethat MNEsplacetheirsubsidiariesindevelopedcountriesasaway

toleveragetheir productive,technologicalandmarketing effi-ciency, following an asset-seeking strategy, looking for their competitiveadvantagesincrease.Therefore,thepreferencesof emerging MNEs for developing markets exemplifytheir ten-dency to explore the “institutional voids”. However when it comes to subsidiaries that transfer knowledge in the reverse direction,theyare morelikelytobeincountrieswherethere arebetterinfrastructureconditions,businesssupportinstitutions andfavorablelegalenvironment

Regarding the second control variable (age), the results surprisingly indicatedthat therewas asignificantcorrelation, thoughnegative,betweenageandRKT.Thus,theyoungerthe subsidiary, the morelikely the existence of RKT.A possible explanationforthisunexpectedresultisthefactthatexperience leads to efficiency gains, but on the other hand, in environ-mentswherechangesoccurveryrapidly,theadjustmentbetween organizational capabilities and market demands declines, as the subsidiariesgrowolder, havinginviewthat moremature companiestakelongertoincorporatethemostcurrent techno-logical developments (Sørensen& Stuart, 2000).It is inthis perspectivethatageandaccumulatedskillscanbecome disad-vantageswhencomparedtoyoungersubsidiaries.Particularly, this occurs with regard to the company’sability to adapt or developmajortechnologicalchanges(Sørensen&Stuart,2000; Tushman & Anderson, 1986) With respect to the group of emerging MNEs, younger subsidiariesmay be more influen-tial in the headquarters’ knowledge exactly because they are abletobemoreagileanddynamicinrelationtotechnological developments

Conclusions, limitations and further research

Thispaperexplainedreverseknowledgeflowsinsubsidiaries

of emerging market multinationals and tested the impact of threedeterminantsinBrazilianmultinationals(Govindarajan&

669

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776

777

778 779 780

knowl-edge flow of RKT in Brazilian MNEs are suggested First,

in comparison withtraditional MNEs, Brazilian MNEs have

a higher interest in reverse technology transfer, due to the

higherimportanceofsubsidiariesforheadquarters.Second,

sub-sidiariesofBrazilianMNEswilltransferproducts’knowledge

justwithabasicandintermediatelevelof technological

com-plexity (Ariffin&Figueiredo, 2003;Iammarinoetal., 2008)

Third,ontheprocessofRKTinforeignsubsidiariesofBrazilian

MNEs,thisworkexploredtheimpactofknowledgecomplexity

characteristicsaswellsubsidiarycharacteristics,i.e.autonomy

andembeddedness.TheresultsshowedthatRKTispositively

affected by knowledge complexity, subsidiary autonomy and

embeddednessofforeignsubsidiarieswithcustomersand

sup-pliers Fourth, it was assessed the effect of the subsidiary’s

location and age on the RKT The results indicate that

sub-sidiaries located in developed countries are more likely to

transfer knowledge in reverse direction as well as younger

subsidiaries This paper’sempirical implicationssuggest that

subsidiarieswithhigheraccesstolocalknowledgewillbebetter

positionedtoacquirenewknowledgeandconsequentlytransfer

it backtoheadquarters.The externalembeddednesshasbeen

indicatedasanimportantdeterminantofRKT.Fromthe

view-pointofpracticalimplications,itisnecessarythatsubsidiaries

investinmechanismsofrelationshipandknowledgeexchange

to establish strong collaborations with local partners These

findings mayalso beusefulfor policy makersinas much as

understandingtheinnovationtransferpatternisakeycomponent

ofacountry’sinnovationsystem

Animportantlimitationofthisstudyisthatthisresearchis limited tothenarrowcontextofBraziliansubsidiaries, which thereforeimposeslimitstotheresultsgeneralization.Second, thesamplesizeandsamplecompositionturnitdifficulttomake far-reaching generalizations of its results Third, the survey methodprovidesasnapshotthatreducestheinformationsource credibility,theaccesstotherightpeople,theresponsescontrol, andtheutilizationofonlyonerespondentbycompany.Fourth, itschoiceofcontrolvariables,whichcouldhavecoveredother aspects, possibly stakeholdersin the achievedresult Finally,

it is assumed somerestrictionsrelated totheunit of analysis andtheinformationfromheadquarters.Furtherresearchescould explore the autonomy and integration degree of subsidiaries fromemergingmarketsmultinationals

Conflicts of interest

Theauthorsdeclarenoconflictsofinterest Q14

BreschiandLissoni (2001),DimaggioandPowell(1983),

Lall (1983), Minbaeva (2007), Reddy (2011), and Ringle, Wende,andWill(2005)

Dependent variables

Reverse knowledge

transfer (RKT)

1.1 Development of new production process; 1.2 Development of new equipment and/or tools; 1.3 Development of new products; 1.4 Know-how and expertise in the form of plans, models, instructions, guides, formulas, specifications, designs, plans, technical drawings, and/or prototypes to design new products; 1.5 Results of research into new materials and specifications; 1.6 Results of research and development (R&D) into new product generations.

Ariffin and Figueiredo (2003); Bell and Pavitt (1995); Iammarino et al.

(2008); Lall (1992); Yang

et al (2008)

Independent variables

Complexity 2.1 Its understanding requires prior learning from other related technological knowledge; 2.2 Its

understanding requires a large amount of information; 2.3 It is the product of many interdependent routines, individuals and resources; 2.4 It includes many different skills or competencies; 2.5 It is technologically sophisticated and difficult to deploy; 2.6 It is complex (vs simple)

Simonin (2004); Zander and Kogut (1995)

Autonomy 3.1 Implementation of changes in products and services; 3.2 Development of new products and

services; 3.3 Implementation of changes in production processes; 3.4 Entry into new markets in the country; 3.5 Procurement and supply chain management; 3.6 Management of Purchasing and Supply Chain; 3.7 Hiring and firing of the subsidiary workforce.

Ghoshal and Nohria (1989);

Birkinshaw et al (1998);

Rabiosi (2011) External

embeddedness

(with customers,

suppliers)

4.1 Customers/suppliers has fully participated in the development of technological knowledge in the subsidiary; 4.2 Customers/suppliers showed important initiatives for the development of technological knowledge in the subsidiary; 4.3 Customers/suppliers satisfied the requirements in developing technological knowledge in the subsidiary; 4.4 The technological subsidiary knowledge was partially developed within this Customers/suppliers’ premises; 4.5 The cooperation with customers/suppliers has been characterized by frequent interactions.

Lane and Lubatkin (1998) , Andersson et al (2005) and Najafi-Tavani et al (2013)

Moderating variables

Subsidiary’s location 5.1 Low-income countries (0); 5.2 High-income countries (1) Mudambi and Cantwell

(2005) Subsidiary’s age 6.1 subsidiaries under 10 years old (0); 6.2 Subsidiaries with over 10 years old (1) Ambos and Schlegelmilch

(2007); Rabiosi (2011)

781

782

783

784

785

786

787

788

789

790

791

792

793

794

795

796

797

798

799

800

801

802

803

804

805

806

807

808

809

810 811 812 813 814 815 816 817 818 819 820 821 822 823

824

825

826

827 828 829

830

831