Towards a multi-perspective model of reverse knowledge transfer in multinational enterprises: A case study of Coats plc

European Management Journal (2013) 31, 179–195

journal homepage: www.elsevier .com/ locate /emj

Towards a multi-perspective model of reverseknowledge transfer in multinational enterprises:A case study of Coats plc

Martina McGuinness a, Mehmet Demirbag a,*, Sasanka Bandara b

a Management School, The University of Sheffield, 9 Mappin Street, Sheffield S1 4DT, UKb Stretchline (Private) Limited, Lot 89, BEPZ, Walgama, Malwana, Sri Lanka

0263-2373/$ - see front mattehttp://dx.doi.org/10.1016/j.e

* Corresponding author. Tel.2223348.

E-mail address: m.demirba

r ª 201mj.2012

: +44 (0)

g@sheffi

KEYWORDSReverse knowledgetransfer;MNE;Knowledge management

Summary The conventional role of corporate headquarters as the sole engine of knowl-edge creation within the multinational enterprise (MNE) is changing. Increasingly, geo-graphically dispersed subsidiaries need to function as neural networks, reversetransferring knowledge to headquarters. This study explores the attributes that stimulatereverse knowledge transfer within MNEs. Firstly, a multi-perspective model of reverseknowledge transfer (RKT) is created conceptualising the process as predicated uponknowledge creation and knowledge transfer. Constructs of this model are then opera-tionalised within Coats plc, a UK based multinational and global market leader in the tex-tile industry. A key benefit of the proposed new model is that it allows for maps ofcapabilities in RKT to be created. This has important implications for management prac-tice by providing managers with a potentially significant tool for strategic analysis basedupon an integrative perspective on factors underpinning RKT.ª 2012 Elsevier Ltd. All rights reserved.

Introduction

MNEs dominate today�s business landscape and it is esti-mated that they collectively account for over 90% of currentworld trade (Petras & Veltmeyer, 2007). The world�s largestMNEs are equivalent in their economic importance to somemedium size economies (Bartlett, Ghoshal, & Beamish,2007), also pioneering most of the world�s research anddevelopment (R&D) and owning most of the patents world-wide (Buckley, 2006). The process of globalisation continues

2 Elsevier Ltd. All rights reserved.03.013

114 2223441; fax: +44 (0)114

eld.ac.uk (M. Demirbag).

to impact upon the MNE by, for example, influencing anincreasingly precise use of location and ownership as asource of competitive advantage (Buckley, 2009; Buckley& Ghauri, 2004). However, at the same time, this increasesthe pressure to balance the needs of the �global� (andaccompanying centralisation) and the �local� (with itsemphasis on decentralisation). One illustration of a re-sponse to these conflicting demands is the emergence ofthe transnational structure whereby the organisation nei-ther completely centralises nor decentralises key activities.Instead, these activities are dispersed, specialised andinterdependent (Bartlett & Ghoshal, 2002), whilst at thesame time an integrated network of assets and capabilitiesis maintained so that the MNE can remain efficient and

.

http://dx.doi.org/10.1016/j.emj.2012.03.013

mailto:[email protected]




180 M. McGuinness et al.

flexible. Globalisation means that factors of paramountimportance to the MNE are technology, knowledge and cap-ital rather than traditional assets such as land (Buckley,2009; Demirbag & Glaister, 2010). Capturing and leveragingthese factors becomes increasing complex and challengingin the �global factory� creating further difficulties in address-ing the imperative to transfer knowledge effectively acrossorganisational boundaries (Kogut & Zander, 1992). At thesame time, effective knowledge transfer is vital for theMNE. Indeed, some commentators argue, the MNE owes itsvery existence to its superior ability to internally transferknowledge, more effectively and efficiently than throughmarket mechanisms (Foss & Pedersen, 2002; Kogut &Zander, 1993). Clearly then, a key challenge for the MNEin the 21st century is to respond, proactively, to the inher-ent difficulties of effective knowledge transfer within thecontext of an ever shifting paradigm of globalisation.

Traditional models of knowledge transfer as exemplifiedby the Hymer–Kindleberger approach (Hymer, 1976) focusupon the conventional forward transfer of knowledge fromheadquarters to foreign affiliates. However, one impact ofglobalisation is that knowledge transfer takes place acrossmultiple dimensions (space, time, language, culture etc.)as well as in multiple directions (forward, backward and lat-eral). As an exemplar, empirical evidence shows that for-eign direct investments decisions (FDI) can be driven (atleast partially) by the desire to gain knowledge (Foss &Pedersen, 2002; Frost, 2001). This presents the spectre ofreverse knowledge transfer (RKT) as an important consider-ation in the MNE�s search for competitive advantage. A com-prehensive review of recent publications indicates that RKTis emerging as an area of increasing interest to researchers(Michailova & Mustaffa, in press). This study adds, concep-tually and empirically, to this burgeoning topic of researchby developing and empirically analysing a conceptual frame-work for RKT. Further, whilst empirical research on knowl-edge transfer has been critiqued for being largely situatedin developed economies (Buckley, Clegg, & Tan, 2003; Yang,Mudambi, & Meyer, 2008), or predominantly Eastern Europe(Cui, Griffith, Cavusgil, & Dabic, 2006; Yang et al., 2008),this study presents data garnered from a case study of aBritish global MNE which spans RKT process from developedto developed economies (US–UK) as well as less developedto developed economies (Sri Lanka–UK; Turkey–UK). A caseapproach is particularly appropriate where data is gatheredfrom cross-border and cross-cultural settings (Ghauri, 2004)and novel perspectives are offered with respect to the geo-graphical (Michailova & Mustaffa, in press) and economicdimensions of the data utilised. From a methodological per-spective, there is also some novelty in terms of the study�sempirical approach, given the under-utilisation of the casemethod in international business research (Marschan-Piekkari & Welch, 2004; Welch, Piekkari, Plakoyiannaki, &Paavilainen-Mantymaki, 2011). Finally, the use of the tex-tile industry as the focus of the research offers significantbenefits. This is a mature industry which has been subjectto significant change resulting from the forces of globalisa-tion and technological development and is highly competi-tive with manufacturing globally distributed (Buxey,2005). This case study presents insights based upon datagathered from two of subsidiaries based in emerging econo-mies for which the textile industry is of significant economic

importance. Furthermore, although work has been carriedout from a cross-industry perspective which includes matureindustries (Piscitello & Rabbiosi, 2006; Rabbiosi, Mudambi,& Piscitello, 2007), there is a lack of empirical work onRKT focused solely in the textile manufacturing industry.Liberalisation and low production costs in less-developedand emerging economies have been influential in the locat-ing of subsidiaries by the MNE, and consequently the effec-tiveness of knowledge transfer and, increasingly RKT, is ofvital importance within the globally distributed textileindustry (Nordas, 2004). This study actively integrates thisindustrial dimension by studying subsidiary RKT and therebyhelps to explain mixed findings in the extant literature.Finally, Michailova and Mustaffa (in press) have highlightedthe narrow representation of empirical work on subsidiariesacross more than two countries. Our approach follows oneof their suggested guidelines for future research by examin-ing subsidiary knowledge flows in a diversified geographicalcontext through an exploration of knowledge flows with re-spect to three geographically dispersed subsidiaries (US,Turkey and Sri Lanka) and its U.K headquarters.

Literature review and model development

Historically, the search for competitive advantage has beenoriented around endogenous factors as embodied by work byPenrose (1959) and the resource-based view (RBV) or exoge-nous factors as proposed by Porter (1980) and the IndustrialOrganisation (IO) School. As the core of its paradigm, the re-source-based view (RBV) views the application of strategi-cally significant resources available at the firm�s disposal askey to competitive advantage (Barney, 1991; Penrose,1959; Teece, Pisano, & Shuen, 1997). Evolving from this per-spective, scholars have explored the process by which orga-nizations develop capabilities of strategic significance(Eisenhardt & Martin, 2000; Grant, 1996a, 1996b; Kogut &Zander, 1993; Nonaka, Byosiere, Borucki, & Konno, 1994).Kogut and Zander (1992) recognize the firm�s combinativecapability; its ability to generate new applications fromexisting knowledge, as an asset of strategic importance.There is recognition that research into organisational knowl-edge transfer is in its relatively early stages (van Wijk,Jansen, & Lyles, 2008). Nonetheless, there is consensus thatknowledge transfer is complex and inherently problematic,whether across firms (Easterby-Smith, Lyles, & Tsang,2008); within regional clusters (Tallman, Jenkins, Henry, &Pinch, 2004); across strategic alliances (Mowery, Oxley, &Silverman, 1996; Simonin, 1999); or between units withinthe firm (Szulanski, 1996). Some (such as Inkpen & Pien,2006; Inkpen & Tsang, 2005) argue that knowledge transferacross firms poses greater difficulty that between units ofthe same firm, whilst others highlight the performanceimplications of lack of knowledge flows which lead to sub-sidiary isolation (Monteiro, Arvidson and Birkinshaw, 2008).Overall then, it is safe to say that research in this area isas yet fragmented reflecting its inchoate state (seeMichailova & Mustaffa, in press for an overview of theliterature).

Knowledge flows between parent and subsidiary havebeen explored with some acknowledgement of the needfor more than unidirectional knowledge flow (Bjorkman,

Sign

ifica

ntfindings

Both

leve

lofintegrationan

dleve

lofloca

lembeddedness

increase

propensity

ofreve

rsetech

nology

tran

sfer

rProposedco

nce

ptual

modelincludesstrategicman

dateofthe

subsidiary,va

lueofkn

owledge

stock

,cu

lturaldistance

,kn

owledge

tran

sferca

pab

ilities,

organ

izational

distance

JointR&Dincreasesab

sorptive

capab

ility&

social

capital

among

partners,increasinglike

lihoodoffuture

knowledge

integration

tBenefits

positive

lyaffectedbyco

mpetitive

strengthofthehost

country,

subsidiary

strategicman

datean

dab

sorptive

capac

ityofthe

head

quarters

s,Im

pac

tofRKTonparent�sinnova

tive

ness

isgreaterwhen,person-based

mech

anismsareemploye

d,subsidiariesareco

mpetence

-creatingan

dkn

owledge

creationbenefits

from

loca

lexternal

linka

ges

Operational

structure,kn

owledge

sharingince

ntive

san

dsubsidiary

socialization,liaisonmech

anismsan

dtemporary

team

shav

eapositive

influence

onoutboundkn

owledge

tran

sfer

Knowledge

releva

nce

ismore

importan

tforreve

rseflowsthan

for

conve

ntional

flows.

Effectsaremoderatedbyloca

tionco

ntext

Towards a multi-perspective model of reverse knowledge transfer in multinational 181

Barner-Rasmussen, & Li, 2004; Buckley & Carter, 2002;Schulz, 2003) to flows in multiple directions (Buckleyet al., 2003; Fey & Furu, 2008; Persson, 2006). So, for exam-ple Buckley and his colleagues found that RKT in an interna-tional joint venture subsidiary relationship was conditionalupon the success of primary knowledge transfer as well asfactors such as global strategy and the goals, aptitudesand abilities of the joint venture partners. Nonetheless,to date work on RKT is limited in range and volume withan emphasis on subjects such as technology and innovation,or the mechanisms of knowledge flow. This study moves be-yond a simple review of existing work in the area of RKT. Itdraws together critical perspectives on RKT literature todate in order to present a holistic framework. This frame-work facilitates insights into the nature of the organisa-tional context which is conducive to effective RKT.

This study utilises Gupta and Govindarajan�s (1991,1994) methodical analysis of knowledge flow based strate-gic mandates (also called strategic contexts) of a subsidi-ary as a foundation for our analysis. The mandates ofglobal innovator and integrated player both entail high lev-els of knowledge outflows from subsidiaries to the rest ofthe corporation. Whilst the authors make no specific refer-ence to RKT it is evident that subsidiaries with global inno-vator and integrated player mandates engage in RKT andconsequently the choice of this framework is appropriateto our research. This framework is integrated with keyfindings from existing studies in order to develop a modelwhich encompasses multiple variables thereby providingmulti-perspectives on RKT within the MNE.

Table

1Keyreleva

ntstudiesonRKT.

Study

Focu

s

Haka

nsonandNobel(2001

)Theeffect

ofloca

lembeddedness

and

integrationonreve

rsetech

nology

tran

sfer

Schlege

lmilch

&Chini(200

3);

Schlege

lmilch

etal.(200

3)Attemptto

deve

lopaunifyingfram

ework

toenab

lefuture

research

onkn

owledge

tran

sfe

Frost

andZhou(200

5)

Impac

tofjointR&Dprojectsonreve

rse

knowledge

integration

Ambosetal.(200

6)Keyva

riablesim

pac

tingH.Q

.ab

ilityto

benefi

from

RKT

Piscitelloan

dRab

biosi

(200

6)Effectiveness

ofRKTan

dtran

sfermech

anism

subsidiary

�srole,au

tonomy,

loca

lco

ntext

Persson(200

6)Knowledge

outflowsfrom

subsidiaries

Yan

getal.(2008

)Aco

mparisonofco

nve

ntional

andreve

rse

knowledge

flowsin

MNEs

Development of the multi-perspectiveframework

Based upon a detailed review of the relevant literature, amodel of RKT was formulated. Table 1 below identifiesseveral key studies from this literature review, which wereutilised in determining the constructs underpinning themodel.

In research, conceptual models are most commonlycreated using a deductive approach which warrantshypothesis-testing. However, the context and constraintsof this study prohibited the use of such methods. The mod-el developed is multi-perspective in the sense that it incor-porates elements from a host of different major researchworks on knowledge transfer, grouped under four con-structs of;

(1) Potential to create knowledge.(2) Relevance of knowledge created.(3) Ability to reverse transfer new knowledge.(4) Motivation to reverse transfer new knowledge.

The phenomenon of RKT is therefore conceptualised asnecessitating both knowledge creation and knowledgetransfer capabilities. With respect to knowledge creation,two preconditions (1 and 2 above) are deemed necessaryto stimulate RKT. Firstly, a subsidiary must have potentialto create knowledge. Secondly, the new knowledge cre-ated must be of relevance to the rest of the corporation.Similarly, considering knowledge transfer, the subsidiary�s


ability and motivation (3 and 4 above) to reverse transfernew knowledge is essential to promote RKT.

Conceptual framework

The four main constructs of the framework were operationa-lised using quantifiable variables identified through a synthe-sis of literature (see Table 2 and Figure 1). The methodologyemployed in operationalisation can be exemplified using thefirst construct which is the potential to create knowledge. Itis argued that reverse technology transfer from subsidiariesis positively correlated to the level of embeddedness of thesubsidiary within its local environment and that this relation-ship is mediated through an improvement in the subsidiary�sinnovativeness (Schmid & Schurig, 2003; Hakanson & Nobel,2001; Noorderhaven & Harzing, 2009). This argument impliesthat (i) the subsidiary�s potential to create knowledge (inno-vativeness) has a positive impact on RKT and (ii) that poten-tial to create knowledge can be operationalised in terms oflocal embeddedness (Hakanson & Nobel, 2001). Followingwork by Nonaka and Takeuchi (1995) and Cohen and Levin-thal (1990), it is also argued that knowledge is always cre-ated by individuals and that an individual�s ability to createnew knowledge is largely a function of his/her level of priorrelated knowledge (Fang, Wade, Delios, & Beamish, 2007;Hong, Easterby-Smith, & Snell, 2006; Wang & Suh, 2009).This argument leads to the addition of the variable prior re-lated knowledge of the senior management team (SMT), un-der this construct. Other such variables (from significantstudies) were subsequently found and incorporated intothe model. Table 2 below summarizes the variables used tooperationalise each of the four constructs.

Underpinning variables

Each of the proposed model�s dimensions is underpinned bya number of variables as well as those necessary to calcu-

Table 2 Summary of the variables used in operationalisation of

Constructs Varia

Potential to create knowledge PriorTalenHeadqInnovLocalPerce

Relevance of knowledge created DegreOrganCultu

Ability to reverse transfer new knowledge SMT�sFunctDegreIntenLevel

Motivation to reverse transfer new knowledge SMT�sFunctpromDegreIntenMD�s

late subsidiary strategic mandates. The manner in whichthe variables were measured is discussed in the methodssection and details are presented in Appendix 1.

Potential to create knowledgeAbsorptive capacity represents the firm�s ability to recog-nise the value of new, external information, assimilate it,and apply/exploit it to commercial ends (Cohen & Levin-thal, 1990; Gupta & Govindarajan, 2000; Lane, Koka, &Pathak, 2006). Cohen and Levinthal found absorptive capac-ity to be mainly a function of the level of a firm�s prior re-lated knowledge and that it has a positive impact on theorganisation�s ability to learn and innovate. Developing thistheme, other researchers (Minbaeva, 2007; Minbaeva,Pedersen, Bjorkman, Fey, & Park, 2003; Zhao & Anand,2009) posit that the concept encompasses both employees�ability and motivation and that these are needed to facili-tate knowledge transfer within the MNE. For Nonaka andTakeuchi (1995), knowledge is created by individuals (thebasic ontological level within which knowledge can exist)but organisations can foster the conditions necessary to en-hance knowledge creation by individuals through mecha-nisms such as talent management practices (Bjorkmanet al., 2004; Reiche, Harzing, & Kraimer, 2009).

The MNE�s organisational structure is also related to po-tential to create knowledge (Foss & Pedersen, 2002;Noorderhaven & Harzing, 2009; Schulz, 2003). Empirical evi-dence has shown that a high level of centralisation dampensdown an organizations ability to innovate (Ghoshal & Bart-lett, 1988).This observation is attributed to the fact thatfreedom to experiment is required for creating innovations(Mohr, 1969) and it highlights the relevance of the degreeof centralisation between headquarters and subsidiary. Thenature of the relationship between headquarters and subsidi-ary was also found to be significant by Buckley et al. (2003)who identified themovement of local Chinese personnel fromthe Chinese subsidiary to the headquarters of the American

constructs.

bles used in operationalisation

knowledge of senior management team (SMT)t management practices within the subsidiaryuarters-subsidiary decentralisationative capacity of host countryembeddedness of the subsidiaryntage of local nationals in the SMT

e of centralisationizational similarity to headquartersral similarity to headquarterscurrent ability to reverse transfer new knowledgeionality of performance appraisal and training routinese of formal integration of the subsidiarysity of subsidiary-parent communicationof slack resourcescurrent motivation to reverse transfer new knowledgeionality of performance based compensation and merit basedotion schemese of informal integration of the subsidiarysity of subsidiary-parent communicationbonus structure

Figure 1 Cause and effect diagram representing the multi-perspective model.


subsidiary as a key element of informal RKT. Broadernetworks of relationships are also potentially important, asillustrated by Frost�s (2001) findings that foreign subsidiariesof MNEs are able to gain from regional knowledge networks.Blomstrom and Kokko (1998) find that MNEs establishingsubsidiaries in centres of excellence benefit from spillover ef-fects. Such spillovers increase a subsidiary�s potential forknowledge creation. The relative economic level of the hostcountry is used in this study, as a proxy for the country�s rel-ative innovative capacity (following Gupta & Govindarajan,2000). Reverse technology transfer from subsidiaries can bepositively correlated to the level of embeddedness of thesubsidiary within its local environment and this relationshipis mediated through an improvement in the subsidiary�s inno-vativeness (Hakanson & Nobel, 2001). A subsidiary that main-tains frequent and significant interaction with its localenvironment has an increased ability to tap into local knowl-edge networks (Frost & Zhou, 2005; Phene, Madhok, & Liu,2005; Williams, 2009). Gupta and Govindarajan (2000) usethe proportion of local nationals within a subsidiary�s SMTas one of the variables that affects its absorptive capacity.Absorptive capacity, in turn, increases the subsidiary�s abilityto learn and innovate (Ambos, Ambos, & Sclegelmilch, 2006;Cohen & Levinthal, 1990). The proportion of local nationalswithin the SMT of a subsidiary can also be used as a measurethat contributes towards inter-unit heterophily (Gupta &Govindarajan, 2000). This is important because high heter-ophily ensures that the subsidiary concerned is endowed witha degree of knowledge fromdifferent domains, in comparisonto other subsidiaries/corporate headquarters. Greater heter-ophily also implies diversity in language systems as well as inthe meanings assigned to the artefacts of communication(Gupta & Govindarajan, 2000: 481; Ambos & Ambos, 2009).Thus, our first dimension brings together factors influencingthe firm�s potential to create knowledge for RKT.

Relevance of knowledge createdIt has been found that coercion can be used in centralisedMNEs to achieve inter-organizational isomorphism (Ghoshal&Bartlett, 1988). The degree of centralisation therefore con-tributes towards ensuring that new knowledge created is ofrelevance to the rest of the corporation. In addition to coer-cion, centralisation can also ensure the relevance of knowl-edge created through an intimate relationship betweenparent and subsidiary. As the converse of decentralisation,the variable of centralisation reflects the perceived trade-off between the degree of control (relevance of knowledgecreated) and freedom to experiment (potential to createknowledge). Simonin (1999) finds that both organisationaldissimilarity and cultural dissimilarity cause ambiguity whichin turn dampens knowledge transfer. Cultural similarity/dis-tance measurements have been used extensively in interna-tional business related research (Ambos & Ambos, 2009;Ambos et al., 2006; Engelhard & Nagele, 2003). So, for exam-ple, the influential Uppsala model (Johanson &Wiedersheim,1975) posits thatMNEs gradually expand from their homebaseto countries (markets) with increasing psychic distance. Cul-tural distance is also associated with higher transaction costsdue to the difficulty of transferring skills, competencies andalso the cost of information thereby impacting upon theRKT (Ambos & Ambos, 2009; Kostova, 1999) Therefore, thesecond construct in our conceptual framework posits thatthe relevance of knowledge created is an important dimen-sion in facilitating RKT.

Ability to reverse transfer new knowledgeThis construct consists of five components. Each of thesecomponents are important determinants of subsidiaries�ability to RKT. As �knowledge engineers�, the SMT of a sub-sidiary play a vital role in both knowledge creation andtransfer processes (Nonaka & Takeuchi, 1995). Minbaeva


et al. (2003) find employees� ability and motivation to be thekey determinants of a firm�s absorptive capacity. In a similarvein, recent studies identify the ability and motivation of asubsidiary�s SMT as two of the key aspects that impact on itsability to reverse transfer new knowledge (Adenfelt &Lagerstrom, 2008; Jensen & Szulanski, 2004; Minbaeva,2007;Monteiro et al., 2008). Minbaeva et al. (2003) also findevidence to show that developing employees� ability andmotivation through the extensive use of human resourcemanagement (HRM) practices contributes to knowledgetransfer processes within MNEs.

Whilst the measure of the SMT�s current ability to trans-fer new knowledge characterises this attribute at a point intime, the measure of functionality of performance appraisaland training routines reflects the ability of the subsidiary asa whole, over a longer period in time (reflecting path-dependence) (Bjorkman et al., 2004; Bjorkman, Fey, &Hyeon Jeong, 2007; Fey & Furu, 2008).

Formal integrative (boundary spanning) mechanisms suchas liaison personnel, temporary task forces and permanentteams can be used effectively to solve problems of highequivocality (Hakanson and Nobel (2001), Bjorkman et al.,2004; Noorderhaven & Harzing, 2009; Persson, 2006; Tsai,2002) These structural mechanisms have the ability to in-crease both the quantity and richness of information ex-changed (Bjorkman et al., 2004; Daft & Lengel, 1986).

Knowledge transfer may require numerous individual ex-changes, especially when the knowledge transferred has ta-cit components (Nonaka et al., 1994). Such person based,face-to-face modes of communication are considered formsof rich-media (Daft & Lengel, 1986: 560; Gnaywali, Singal, &Mu, 2009; Hansen, Mors, & Lovas, 2005; Hakanson & Nobel,2000). Media of lower-richness include information commu-nication technology based or paper based mechanisms whichmay or may not rely on rules and procedures. Ghoshal andBartlett (1988) find a positive correlation between transmis-sion channel richness and the diffusion of innovation.

With respect to knowledge transfer, slack resources alsohave been found to be beneficial. Higher levels of slack re-sources, such as leeway in operating budgets, positively af-fect knowledge transfer since they allow subsidiaries totransfer knowledge without having to withdraw resourcescommitted to maintaining their current level of operationalactivities (Ghoshal and Bartlett, 1988:370). Consequently,we propose that these variables are influential in facilitatingthe ability to RKT in the MNE.

Motivation to RKTHRM practices can be powerful influences upon employees�motivation to engage with certain activities (Huselid,1995; Minbaeva et al., 2003; Bjorkman et al., 2007), in thiscase SMT�s current motivation to reverse transfer new knowl-edge. Functionality of performance-based compensationand merit-based promotion schemes have been identifiedas key indicators (Bjorkman et al., 2004; Fey & Furu, 2008;Minbaeva et al., 2003; Persson, 2006). It has been observedthat a subsidiary head�s motivation to engage in RKT transferwould be greater when his/her incentives are tied to the per-formance of a cluster of units as a whole (Gupta & Govind-arajan, 1991: 781).

Informal integrative mechanisms are also important(Bjorkman et al., 2004; Persson, 2006; Tsai, 2002). Ghoshal

and Bartlett (1988: 374–375) find that socialization mecha-nisms (informal processes) can be used as indicators of nor-mative integration. Integration in turn facilitates diffusionof innovations by increasing the richness of transmissionchannels (Gupta & Govindarajan, 2000). Thus, the finaldimension of our framework postulates that motivation isimportant in motivating senior managers to engage inRKT. Below is a cause and effect diagram illustrating theeighteen variables outlined above.

Methods and data collection

Case study underpins the methodology adopted for this re-search. Whilst the case approach is most widely associatedwith qualitative data, such as interviews and focus-groups,this case study is based upon quantitative data gatheredthrough an in-depth survey. Drawing upon Welch et al�s(2011) typology of methods of theorising from case studies,this case study can be best understood as a �natural experi-ment� whereby explanation in the form of cause-effect link-ages is presented. A key strength of this approach is that ofinternal validity and the ability to generalise to theory. Theconceptual context of RKT and the research context of thecompany lend themselves particularly well to a case ap-proach which seeks to build theory (Eisenhardt, 1989). AsYin (1994) notes the single, holistic case method is a partic-ularly appropriate choice of research strategy when the rel-evant theory underlying the case study is itself of an holisticnature. From the discussion above which outlined the devel-opment of the multi-perspective model, the consonance be-tween theory and method is clear. It is also worth notingthat in addition to allowing an in-depth investigation of asingle organisation and an accompanying holistic view, thecase method remains an under-represented empirical meth-od in the canon of international business research which re-flects a broader issue of the under-utilisation of qualitativemethodology (Marschan-Piekkari, 2004; Michailova & Must-affa, in press). This is despite the growing use of this re-search strategy, as well as increasing confidence in itsrigour (Hartley, 2004). Such under-representation may re-flect the particular challenges posed by the case method(Eisenhardt & Graebner, 2007) or perhaps misunderstand-ings regarding case study research (Flyvbjerg, 2006).

Research context

Coats plc (UK) is a UK based multinational which is theworld�s leading manufacturer and supplier of sewing thread.The company has a history that dates back over 250 yearsand has pioneered many developments within its industry.Increasingly, the company faces stiff multipoint competitionfrom other large MNEs as well as localised challenges fromnational firms. The three subsidiaries analysed were allwholly/majority owned by Coats plc. Both Coats American(USA) and Coats Thread Turkey (Turkey) were establishedas greenfield operations in 1991 and 1950 respectively. CoatsThread Lanka (Sri Lanka) became part of the Coats group in1991 through the acquisition of Total Group plc by Coats plc.In terms of subsidiary size, Coats American has the largestannual turnover of GBP 60 million, while Coats Thread Lankahas the lowest of GBP 17 million. Coats Thread Turkey


employs the largest number of employees (865) while CoatsThread Lanka employs the least (600). All three subsidiarieswere engaged in manufacturing, marketing/sales and logis-tics activities utilising similar technologies. In addition, theoperations of Coats American also encompassed researchand development.

Data collection

A method for primary data collection was necessary throughwhich all of the variables incorporated in the model of RKT,as well as those necessary to calculate subsidiary strategicmandates (for comparison) could be measured. The neces-sity to collect data from multiple, geographically dispersed,subsidiaries of a MNE increased the complexity of the studysignificantly. In fact, such complexities stemming fromempirical multi-unit MNE research have often been attrib-uted to the scarcity of work in this field (Chini, 2004: 68).However, the authors shared the view put forward inMichailova & Mustaffa�s (in press) guidelines for futureresearch on subsidiary knowledge flows in MNEs, namely,that for a better understanding knowledge flows withinMNEs research should focus on several geographically dis-persed subsidiaries. The choice of Coats plc as our casecompany allows us to address this guideline by examiningsubsidiaries spread across countries and indeed, yet moreinterestingly, across geographical regions. Moreover, wewere particular keen to add novelty to our data collectionby our choice of subsidiary location. This spanned not onlydeveloped and less developed economies, but also includedsubsidiaries in countries (Turkey and Sri Lanka) which havenot previously been explored from this conceptual platform(Michailova & Mustaffa, in press). Helping us to address, ona practical level, the level of complexity created by such anapproach, was the fact that one of the authors was an em-ployee of a foreign subsidiary within Coats plc. In this way,we were better able to manage the challenges of datacollection across such geographical breadth. Consequently,the choice of subsidiaries themselves was influenced by geo-graphic location, and, within a designated region, specialconsideration was given to selecting units from which goodquality data could be obtained. The managing director (MD)and heads of manufacturing, marketing/sales and logisticsfunctions were selected as key respondents within each sub-sidiary. These personnel are key players within the SMT of asubsidiary. As knowledge engineers (Nonaka & Takeuchi,1995), their job is to animate corporate level visionary ide-als in an often chaotic market reality. Mediating betweenwhat is and what should be, these individuals often needto engineer new knowledge according to the company�s vi-sion (Nonaka & Takeuchi, 1995: 154). They can also beviewed as catalysts and architects of communicationsinfrastructure and protectors of knowledge investments(Hedlund, 1994). In the specific context of Coats plc, thethree functions of manufacturing, marketing and logisticsconstitute areas of strategic importance.

Development and administration of the surveyinstrumentTwo standardized quantitative questionnaires were devel-oped, incorporating all the variables outlined previously

and one was administered to the MD of each subsidiaryand the other to the heads of manufacturing, marketing/sales and logistics. The questionnaire addressed to each ofthe three divisional heads (although standardized) had min-or variations among them to suit the context of each func-tion. Our conceptual framework for studying RKT fromCoats� subsidiaries comprises four primary constructs andeighteen secondary variables. Constructs used in the surveyinstrument and details of the scales are described in Appen-dix 1. The major characteristics of the framework warrantemphasis at this stage. Every attempt was made to opera-tionalise the constructs in terms of variables endogenousto the MNE (over which they have control). The four mainconstructs of the model act as full mediators of their ante-cedent variables and, in turn, the amalgamation of all fourconstructs contribute towards RKT intensity.

Calculating the intensity of RKTMathematically, RKT intensity from a subsidiary was con-ceptualised as the arithmetic mean across the four con-structs. The relative intensity of RKT (as a percentage)from a subsidiary can therefore be calculated as:

IRKTjðrelÞ ¼P4

z¼1CzjPnj¼1P4

z¼1Czj

� 100%

where, IRKTj(rel) is the relative RKT intensity for the jth sub-sidiary and Czj is the value of the zth construct (of the fourconstructs) for the jth subsidiary.

Similarly, calculation of composite values for each of thefour constructs (Czj�s) can be done based on the values oftheir constituent variables. For a meaningful comparisonto be made across subsidiaries the difference betweensubsidiaries for any given variable was normalised. Thesenormalized deviations can then be averaged to determinethe composite index for each construct. Algebraically, themethod can be represented as:

Czj ¼Pn

i¼1 ðVizj � V ðminÞizÞ2=r2iz

n o

n

where, Czj is the value of the zth construct for the jth sub-sidiary, Vizj is the value of the ith variable used to operation-alize the zth construct for the jth subsidiary, V(min)iz is theminimum value of the ith variable used to operationalizethe zth construct across all subsidiaries, riz

2 is the varianceof the ith variable used to operationalize the zth constructacross all subsidiaries and n is the number of variables usedto operationalize the zth construct.

Comparing and validating the model

In order to test for construct validity on the new model, analternative method by which the intensity of RKT could bemeasured was required. The method of assigning knowledgeflow based strategic mandates to subsidiaries (Gupta &Govindarajan, 1991) was found to be ideal for this purpose.In addition to providing an alternative perceptual methodfor measuring RKT intensity, the strategic mandates assignedto subsidiaries through this method also act as mechanismsfor facilitating discussion on the intensity of RKT calculated.


Strategic mandates of subsidiariesFollowing Gupta and Govindarajan (1991, 1994) subsidiarystrategic mandates were operationalised using the intensityand direction of knowledge flows. A seven-item instrumentwas adapted from Gupta and Govindarajan (1994) with theMD of each subsidiary asked to indicate the extent of knowl-edge flow (from 0 – not at all, to 6 – a very great deal) foreach of the seven business activities given the four flow con-texts of (a) knowledge outflow to headquarters (global andregional) (b) knowledge outflow to peer subsidiaries (c)knowledge inflow from headquarters and (d) knowledge in-flow from peer subsidiaries. Averages of the compositeknowledge outflow and inflow measures (across all threesubsidiaries) were calculated and used to derive the bound-aries that separate the various mandates of implementor(sic), local innovator, global innovator and integrated player(Gupta & Govindarajan, 1994: 450). The value of perceivedrelative intensity of RKT, for each subsidiary, was calcu-lated using the values obtained in context (a) above. Theformula used was;

IRKTjðrelÞ ¼OhjPnj¼1Ohj

� 100%

where, IRKTj(rel) is the perceived relative RKT intensityfor the jth subsidiary and Ohj is the value of knowledge out-flow to headquarters for the jth subsidiary.

Measures

All data used in the analysis were from the administeredquestionnaire. Variables used to create constructs and mea-surement scales and respondents of each questions are de-scribed in Appendix 1.

Findings and discussion

Composite values for each of the four main constructs of (1)potential to create knowledge, (2) relevance of knowledgecreated, (3) ability to reverse transfer new knowledge and(4) motivation to reverse transfer new knowledge, were cal-culated using the normalised values of the variables mea-sured. Table 3 below summarises these results.

The values of each of these constructs were then used tocalculate the relative intensity of RKT for each subsidiarywith Coats American (51%) recording the highest relativeintensity of RKT, followed by Coats Turkey (27%) and CoatsThread Lanka (22%).

In order to test the validity of the new model, the inten-sity of RKT was also measured using the MD�s perceptionregarding the intensity of hierarchical and lateral knowl-edge flows of a subsidiary (Gupta & Govindarajan, 1994).These results are summarised in Table 4 below.

Table 3 Calculated values for each of the four main constructs

Constructs

Potential to create knowledgeRelevance of knowledge createdAbility to reverse transfer new knowledgeMotivation to reverse transfer new knowledge

Using this alternate method, the relative intensity ofknowledge outflows to headquarters (RKT) was then calcu-lated. Here, too, Coats American (52%) showed the highestlevel of relative RKT intensity, followed by Coats Turkey(29%) and Coats Thread Lanka (19%). It is interesting to notethat the two different methods produced similar results,particularly given the fact that most measures involved per-ceptual attributes. See Table 5 for a comparison of the rel-ative RKT intensities.

It is clear that the preliminary results obtained using thenew model are comparable to the results obtained using amore direct perceptual measure of RKT intensity. Thesquare-root of the average of squared differences worksout to 2.48, indicating that on average, the new model com-putes relative RKT intensity to within ±2.48% of the actual(perceived) value computed using method 2. This observa-tion provides rudimentary evidence of the validity of themodel.

A more detailed analysis of the model reveals that not allof its variables affect RKT intensity as anticipated. As indi-cated by the perceptual measure, the three units examinedvary with respect to RKT in the order of Coats American(52%) > Coats Turkey (29%) > Coats Thread Lanka (19%).Therefore, a comparison of any constituent variable of themodel, across the three subsidiaries in the above order alsoessentially conveys if RKT intensity changes linearly withthe variable, or not. Using the first construct of potentialto create new knowledge as an example (Figure 2 below)it can be seen that RKT intensity does not increase linearlywith changes in talent management practices, localembeddedness and percentage of local nationals in theSMT (i.e the correlation is non linear). The pattern of a var-iable across the three subsidiaries can be used to infer themanner in which RKT intensity changes with that variable.The next figure (Figure 3 below) illustrates the actual vari-ation in RKT intensity, with changes in the values of eachof the three constituent variables.

Assigning strategic mandates to subsidiaries

Subsidiary strategic mandates with respect to knowledgetransfer were determined for all three subsidiaries. On aver-age, knowledge inflow from headquarters recorded thehighest value both overall (4.48) and also in the case of eachindividual subsidiary. This clearly indicates the dominanceof conventional forward knowledge transfers. Interestingly,knowledge outflow to headquarters (RKT) ranked second(2.76) and took precedence over both lateral transfer pro-cesses (inflow and outflow to peer subsidiaries).

The average values for total knowledge outflow and totalknowledge inflowwere used to distinguish between unitswithhigh (above average) and low (below average) knowledge

of the model.

Sri Lanka Turkey USA

0.66 1.78 2.121.25 0.08 1.130.85 1.68 2.190.24 1.23 2.69

42%

58%66%

89%

71%

11%

50%

83%

71%

59%

86%

25%

57%65%

73% 76% 75%

131%

Prior Knowledge Talent Management

Practices

Decentralization Local Embeddedness

Local Nationals in SMT

REL of Host Country

Sri Lanka Turkey USA

Figure 2 Comparison of potential to create new knowledge. The pattern of a variable across the three subsidiaries can be used toinfer the manner in which RKT intensity changes with that variable. (REL: relative economic level).

Table 4 Intensities of hierarchical and lateral knowledge flows.

Subsidiary Knowledgeoutflow

Total knowledge outflow Knowledge inflow Total knowledge inflow

HQ Peers HQ Peers

Sri Lanka 1.57 1.14 2.71 4.14 0.57 4.71Turkey 2.43 1.43 3.86 4.71 1.86 6.57USA 4.29 2.86 7.14 4.57 2.00 6.57Average 2.76 1.81 4.57 4.48 1.48 5.95

Table 5 Comparison of relative RKT intensities calculated using two different methods.

Subsidiaries Method 1The new model (%)

Method 2G&G (1994) (%)

Squared difference (%)

Coats Thread Lanka 22.31 18.94 0.11Coats Thread Turkey 26.79 29.31 0.06Coats American 50.90 51.75 0.01


flow values. Strategic mandates were then assigned to eachsubsidiary based on the intensity of composite knowledge in-flows and outflows, compared to the average (see Figure 4).

The results show that Coats American can be classifiedas an Integrated Player. Integrated players engage in highlevels of RKT and contribute extensively to the knowledgebase of the firm. However, they are also heavily reliant onknowledge inflows from the corporate parent. An interest-ing observation regarding the sample of subsidiaries exam-ined is worth mentioning at this stage. Coats Americanhad an R&D department, implying that it had differentstrategic mandate from the other two subsidiaries

(confirmed by the above analysis). In fact, some studieshave shown the strategic mandate of a subsidiary as acontingency to RKT (see for example Schlegelmilch & Chi-ni, 2003). The observation that RKT increases with a par-ticular indented strategic mandate, on its own, is of littlerelevance to the corporation. What is more important is todiscover the underlying factors that could potentiallyshape an indented mandate of a subsidiary, leading to adifferent emergent mandate (which is what this modelstrives to achieve).

Coats Thread Turkey falls into the category of Implemen-tor. Subsidiaries with implementor roles are heavily reliant

Figure 3 How perceived relative RKT intensity varies with talent management practices, local embeddedness and the percentageof local nationals in the SMT.

Figure 4 Subsidiary strategic mandates.


on knowledge inflows (especially from the corporate parentin this case) and are common within MNEs that opt to pursuea globalization strategy. Coats Thread Lanka (Sri Lanka) fitsthe role of a Local Innovator. Gupta and Govindarajan (1991:775), posit that the lower knowledge inflow from the rest im-plies that local innovators need to create their ownknow-how. They further add that, this new know-how is idi-osyncratic and seldom of much competitive use beyond thecontext of the host country in which the subsidiary is situ-ated, explaining the low outflow of knowledge. A closerexamination of the data for the three subsidiaries revealsthat all three subsidiaries receive very similar levels ofknowledge inflows from corporate headquarters (Sri Lanka42%, Turkey 50% and USA 57%), making this explanation unli-kely. In this regard, the strategic mandate for Coats ThreadLanka can best be described as that of an Implementor. Thenew model also provides no support in favour of Gupta andGovindarajan�s (1991) hypothesis that the new knowledgecreated by a local innovator would remain idiosyncratic.With respect to Coats Thread Lanka, it can be seen that thissubsidiary records the highest relative score for relevance ofknowledge created, indicating that the low overall level for

relative intensity of RKT recorded by this subsidiary cannotbe attributed to the lack of relevance of knowledge created.Whilst this may appear counter-intuitive at first glance, itmakes more sense when other data are considered. So,although all three subsidiaries demonstrate the same degreeof organisational similarity, the USA shows the highest levelof cultural similarity to the UK, and Sri Lanka shows the high-est level of centralisation. The degree of centralisation isparticularly important as the greater the decentralisationof the subsidiary, arguably the knowledge generated by thesubsidiary digresses from that of the corporate HQ, thus ren-dering the knowledge less valuable to the parent.

Mapping RKT capabilities within a MNE

The new multi-faceted model allows for maps of RKT capa-bilities across different subsidiaries to be created. Figure 5depicts such a map for Coats plc.

Darker boundaries were used on this map to depict areaswith low comparative scores (i.e. areas for concern), whilelighter boundaries were used to depict increasingly highscores (with white depicting the highest category).

Figure 5 A map of RKT capability within Coats plc.

Figure 6 A map of ability to transfer new knowledge within Coats plc. [Values for this map were calculated by normalising thedeviance for data on each of the variables presented.]


The map clearly shows how Coats American (as an inte-grated player with a high intensity of RKT) comparativelyoutperforms the other two subsidiaries with respect toRKT capability. Comparing all three subsidiaries, CoatsAmerican can be seen as having the highest motivation toreverse transfer new knowledge. In terms of weaknesses,Coats Turkey appears to score the lowest with respect tothe relevance of knowledge created, while Coats ThreadLanka scores low on the potential to create knowledgeand both the ability and motivation to reverse transfernew knowledge. This map can be used as guide for improv-ing RKT capabilities. If it were needed, for example, to re-verse transfer more knowledge from Coats Thread Lankathen, according to the new model, all of the three attri-

butes listed above (potential, ability and motivation) mustbe improved. The model also provides guidance in termsof how each attribute can be strengthened. Taking abilityto transfer new knowledge as an example (Figure 6), itcan be seen that, within Coats Thread Lanka, this attributecan be enhanced by (a) ensuring the SMT has higher abilityto reverse transfer new knowledge, (b) improving function-ality of performance appraisal and training routines, (c)using liaison personnel, temporary task forces and perma-nent teams to improve integration and (d) increasingintensity of subsidiary-parent communication. In CoatsThread Lanka, the only variable (in relation to ability totransfer new knowledge) which does not need to be im-proved is the level of slack resources.


Furthermore, by highlighting the relative strengths ofsubsidiaries, these maps not only indicate what attributesneed to be improved and how they can be improved, but alsoprovide targets in terms of how-much improvement is possi-ble. Using the above example, we see that Coats Americansets the internal benchmark in terms of ability of the SMTand parent-subsidiary communication, while Coats ThreadTurkey sets the internal standard for utilization of formalintegrative mechanisms and functionality of performanceappraisal and training routines. These units can thereforebe identified as best practice units in relation to each ofthe above attributes. Observation and imitation of the pro-cedures and routines within these subsidiaries, should there-fore lead to improvements within Coats Thread Lanka.

Conclusions

Grounded in received theory, the study finds that RKT phe-nomena can be characterised by 18 variables, which can begrouped under the four main constructs of the potential tocreate knowledge, relevance of the knowledge created,ability to reverse transfer new knowledge and motivationto reverse transfer new knowledge. Results showed thaton average, the value for relative RKT intensity calculatedusing the model vary by ±2.48% in comparison to the valuefor relative RKT intensity estimated based on the MD�s per-ception. The new multi-perspective model therefore, clo-sely reflects reality. Interestingly, preliminary results alsoseem to have uncovered some weaknesses in Gupta andGovindarajan�s (1991) method of assigning mandates tosubsidiaries. As in the previous section, the early work onRKT focused predominantly on innovation, R&D and technol-ogy transfer (Frost & Zhou, 2005; Hakanson & Nobel, 2001).This paper attempts to bridge some of these gaps by build-ing upon previous work in knowledge management and RKT.Further, commentators have argued that knowledge trans-fer across organisational boundaries (whether inter or in-tra-organisation) is complicated because of the differencein cultures (Easterby-Smith et al., 2008). By providingempirical evidence from a major MNE to support this argu-ment, and from the novel perspective of RKT, our work con-tributes to broader theoretical debates about knowledgetransfer and cultural distance. From the managerial per-spective, our study demonstrates the importance of ade-quate considerations of cultural difference in encouragingand facilitating knowledge flows across the MNE.

Prevailing approaches to modelling begin by testing rela-tionships between variables for statistical significance.Reflecting the limitations of the study, we bypassed thisstep by using related variables from major previous studies,for which hypotheses concerning their relationships havebeen tested and confirmed. The methodology employed,can also be described as largely systemic since no attemptis made to offer hypotheses regarding the relative intensityof RKT and individual constructs or variables of the model.Instead, intensity of RKT is viewed as being reliant on theamalgamation of the four constructs, with each of these,in turn, acting as a full mediator of its constituent variables.Given the small sample size, no statistical inferences can bemade regarding the results at this stage.

Whilst the UK was once the dominant player in this indus-try; today most of the world�s textile and apparel manufac-turing is carried out outside the UK. This fact highlights theimportance to foreign subsidiary knowledge for UK basedMNEs like Coats plc and emphasises the importance of RKTprocesses for managerial practice. In addition to its implica-tions for managerial practice, the multifaceted model cre-ated, lends itself towards further investigation within thegrowing field of RKT research. Since this study adopted a no-dal approach and focused purely on the subsidiary, it wasonly possible to measure the subsidiary�s perception regard-ing the intensity of knowledge outflow. An obvious next-stepwould be to compare the values obtained using the modelagainst the perceptions of the corporate parent. It may alsobe possible to improve the precision of the new model, byexploring more robust methods through which its constitu-ent variables can be measured.

Whilst there are certain limits imposed on the researcharising from the constraints in terms of generalisability fromone in depth case study, it is clear, nonetheless, that ournew model of RKT opens new horizons both in terms of the-ory and practice. This case study has addressed a key gap inthe literature on RKT (Michailova & Mustaffa, in press; Rab-biosi, 2011). In particular, Michailova and Mustaffa identifieda focus on knowledge inflows to subsidiaries in research car-ried out to date. This is at the expense of investigation ofsubsidiary knowledge outflows, both horizontal and vertical.This study has explored this critical gap in the literature in anovel manner through its multi-subsidiary focus across di-verse geographical and economic regions and has added toknowledge transfer research as a result. Having successfullytrialled the model during this study, larger scale research isnow called for, in support of future development.

Managerial implications

A key benefit of our new model is that it allows for maps ofRKT capabilities to be created. This has important implica-tions for management practice. Not only do these maps of-fer managers a potentially significant tool for strategicanalysis by providing an integrative perspective on factorsunderpinning RKT, they also allow an assessment of relativeRKT capabilities across different subsidiaries. One can arguethat being able to build an overarching picture of RKT acrossthe different parts of the MNE is vital given Monteiro et al.�s(2008) findings regarding the negative performance implica-tions of subsidiary isolation. Buckley et al. (2003) found thatthe global strategy of the MNE parent was influential indetermining the success of RKT. The model presented en-hances managers� abilities to explicitly and methodicallyidentify, target and review RKT within the global MNE notsimply to optimise opportunities for RKT, but also to ensureconsonance with global strategic imperatives. This is ofgreat value given data on knowledge flows within MNEswhich appear to indicate negative performance implicationsfor isolated subsidiaries isolated from knowledge flows.

For managers, the mapping process set out in this paperoffers several practical benefits. At a strategic level, thismapping process can be utilised to highlight areas whichcan (or which need) to be improved in order to stimulateRKT (in terms of the four main constructs). This can allow


the most efficient allocation of scarce resource within theMNE. Benchmarking can be carried out across organisationalunits to enhance knowledge flows in multiple directions andensure strategic fit with MNE organisational imperatives.Operationally, the maps can establish specific directions foroptimal improvements through consideration of the constit-uent variables underpinning the different constructs. In this

Table A1 Constructs, respondents and measures used to calc

Potential to create knowledgePrior knowledge of thesenior management team

Adapted from Minbaeva et al. (2003),within each subsidiary was measured iCoats units and (b) individuals hoAccordingly, the MD of each subsididifferent functional heads for each ofpoint likert scale

Talent managementpractices within thesubsidiary

Adapted from Minbaeva et al. (2003),logistics functions within each subsidextent to which they agree/disagreerespective departments as a whole;

a. A strong emphasis is placed on reand skills.

b. It is likely that vacancies in the sfrom within this subsidiary.Values acrossposite score for talent management wit

Headquarters-subsidiarydecentralization

Adapted from Vancil (1979), the MD ofhe/she had in affecting the outcomeand Govindarajan (2000), an answerranged from: (0) your opinion is not asyou, followed by cosultation with sup

Innovative capacity ofhost country

The relative economic level of the horelative innovative capacity (followinthe host country is measured by dividipower parity (GNIppp) of the host coudata for each country were obtained

Local embeddedness ofthe subsidiary

Adapted from Hakanson and Nobel (20within each subsidiary were askedsubsidiary/their respective departmegovernment) for different modes of cfrequency of communication. Thesemechanisms (which are known to be fother modes were each weighted at 1

Proportion of localnationals within the SMT

The MD of each subsidiary was asked tdifferent functional heads

Relevance of knowledgeDegree of centralisation As the converse of decentralisation, t

[100% - decentralisation]. Although tmodel may seem perplexing, this recontrol (relevance of knowledge creat

Organizational similarity Modified from Simonin (1999), the Mculture of the subsidiary with respectwithin the host country. The extent tosuch as business practices, institutiopoint scale. Responses were averagedorganizational similarity (for each of

way, managers can target and direct effective implementa-tion and utilisation of resource in order to optimise returnsto the business. Finally, they can be used to estimate targetsregarding howmuch improvement can bemade (based on thecapability of a selected benchmark unit).

Appendix A. See Table A1.

ulate RKT.

the level of prior related knowledge of each member of the SMTn comparison to (a) individuals holding similar positions in otherlding similar positions among local industry competitors.ary was asked to indicate the job-related knowledge of thethe situations (a) and (b) above, with responses graded on a five

the MD and the heads of manufacturing, marketing/sales andiary were asked to indicate, on a seven point likert scale, thewith the following statements considering the subsidiary/their

cruiting and/or retaining employees with high levels of knowledge

enior management team (including subsidiary head) would be filledboth statements and all individuals were averaged to create a com-hin each subsidiary

each subsidiary was asked to indicate the typical influence thatof a range of strategically relevant decisions. Following Guptato each questions was selected from a five point scale whichked, but decision is explained to you, through to (4) proposal byerior, with your opinion prevailingst country is used in this study, as a proxy for the country�sg Gupta & Govindarajan, 2000). The relative economic level ofng the gross national income per capita adjusted for purchasingntry, by that of the UK (and expressed as a percentage). GNIpppfrom World Bank (2007) statistics01), the MD and heads of manufacturing, marketing and logisticsto indicate the frequency of communication between thent and local institutions (customers, suppliers, universities andommunication and these were scored on a seven point scale ofresponses were reverse scored for analysis and face-to-faceorms of rich media) which were given a higher weight of 2. All

o indicate his own nationality as well as the nationalities of the

he variable of centralisation was calculated using the formulahe introduction of two contradictory variables into the sameflects reality in which a trade-off is seen between degree ofed) and freedom to experiment (potential to create knowledge)D of each subsidiary was asked to compare the organizationalto (a) the corporate headquarters and (b) other local businesseswhich the MD agreed/disagreed with statements about factors

nal heritage and corporate culture were recorded on a sevenacross all three statements to create a composite measure forthe two situations (a) and (b) above)

(continued on next page)

Table A1 (continued)

Cultural similarity Hofstede�s (2001) cultural dimensions were used as a basis for this measurement. Due to unavailabilityof cultural data on Sri Lanka, the data for Thailand (reflecting values of a predominantly Buddhistculture) and India (reflecting colonial influence) were averaged to create these statistics. The degreeof cultural similarity was calculated as:

CSj ¼P4

i¼1 jIij � Iiuj=DIðmaxÞiu� �

4where, CSj is the cultural similarity index for the jth country to the UK, Iij is the Hofstede indexfor the ith dimension and jth country, Iiu is the Hofstede�s (2001) index for the ith dimension forthe UK and DI(max)iu is the maximum possible difference the ith dimension can take, in compar-ison to its value for the UK

Ability to reverse transfer new knowledgeSMT�s current ability toreverse transfer newknowledge

Minbaeva et al. (2003) find employees� ability and motivation to be the key determinants of a firm�sabsorptive capacity. In a similar vein, this study identifies the ability and motivation of a subsidiary�sSMT as two of the key aspects that impact on its ability to reverse transfer new knowledge. The MD ofeach subsidiary was asked to evaluate the ability of the functional heads to transfer new knowledge tothe rest of the corporation. Responses had to be indicated on a five-point scale ranging from 0 (verylow) to 4 (outstanding). The average score for all members of the SMT was used as a compositemeasure of the SMT�s ability to transfer knowledge

Functionality ofperformance appraisaland training routines

Adapted from Minbaeva et al. (2003), the MD and the heads of manufacturing, marketing/sales andlogistics functions within each subsidiary were asked to indicate the extent to which they agree/disagree that there are well functioning performance appraisal and training routines in operation,designed to enhance the knowledge and skills of employees within the subsidiary/their respectivedepartments as a whole. Responses were scored on a seven point Likert-type scale ranging from 0(strongly disagree) to 6 (strongly agree) and were averaged over all individuals to create a compositescore for this variable within each subsidiary (see also talent management practices)

Formal integrativemechanisms

Following Gupta and Govindarajan (2000: 478), formal integrative mechanisms were measured byasking the MD of each subsidiary to indicate the extent to which his/her subsidiary used themechanisms of liaison personnel, temporary task forces and permanent teams to coordinate decisionsand actions with sister subsidiaries, on a scale of 0 (used rarely) to 6 (used frequently). A weightedaverage of responses to the three items was used as the final measure for each subsidiary. Permanentteams (the most complex mechanism), temporary task forces and liaison personnel (the least complexmechanism) were given weights of 3, 2 and 1 respectively

Intensity of subsidiary-parent communication

Modified from Ghoshal and Bartlett (1988), the MD and heads of manufacturing, marketing andlogistics within each subsidiary were asked to indicate the frequency of communication between thesubsidiary/respective department and Coats global headquarters for different modes ofcommunication using a 7 point scale of frequency (0 – daily, to 6 – less frequently than once ayear). As earlier, responses were reverse scored for analysis and face-to-face mechanisms (which areknown to be forms of rich media) were given a higher weight of 2 while all other modes were eachweighted at 1

Availability of slackresources

Bartlett & Ghoshal (2002) estimate slack resources in terms of leeway in operating budgets. Slackresources in terms of manning levels were also measured in this study to buttress the existing metric.Accordingly, the heads of manufacturing, marketing/sales and logistics within each subsidiary wereasked to indicate the impact of a 10% reduction in (a) operating budgets and (b) manning levels, on theday to day operations of their departments. Answers had to be selected from a scale ranging from 0(no perceptible effect) to 6 (significant disruption of activities). The responses were reverse scoredfor analysis and scores across both categories ((a) and (b) above) were averaged out to create acomposite measure for the level of slack resources.

Motivation to reverse transfer new knowledgeSMT�s current motivationto reverse transfer newknowledge

Based on Minbaeva et al. (2003), the MD of each subsidiary was asked to evaluate the motivation ofthe different functional heads to transfer new knowledge to the rest of the corporation. Responseshad to be indicated on a five-point scale ranging from 0 (very low) to 4 (outstanding). The averagescore for all members of the SMT was used as a composite measure of the SMT�s motivation to transferknowledge

(continued on next page)


Table A1 (continued)

Functionality ofperformance-basedcompensation and merit-based promotion schemes

Again, following Minbaeva et al. (2003) the MD and the heads of manufacturing, marketing/sales andlogistics functions within each subsidiary were asked to indicate the extent to which they agree/disagree with the following statement considering the subsidiary/their respective departments as awhole;‘‘There are well functioning performance-based compensation and merit-based promotionschemes in operation, designed to motivate employees’’Responses were scored on a seven- pointLikert-type scale ranging from 0 (strongly disagree) to 6 (strongly agree) and were averaged over allindividuals to create a composite score for this variable within each subsidiary

Informal integrativemechanisms

Adopting scales from Gupta and Govindarajan (2000), corporate socialization mechanisms weremeasured by asking the MD�s if they had (a) worked for one or more years in other subsidiaries, (b)participated in executive development programs involving participants from several subsidiaries (c)worked for one or more years at corporate headquarters and (d) had a mentor at corporateheadquarters. Binomial �yes� or �no� answers were provided for these questions. The total count of�yes�-responses was treated as a measure of participation in corporate socialization mechanisms

MD�s bonus structure The MD of each subsidiary was asked to indicate the percentage of his/her incentive bonus that wasbased on (a) the subsidiary�s performance and (b) on the performance of a cluster of subsidiaries, forthe most recent year


References

Adenfelt, M., & Lagerstrom, K. (2008). The development andsharing of knowledge by centres of of excellence and transna-tional teams: A conceptual framework. Management Interna-tional Review, 48, 319–338.

Ambos, T. C., & Ambos, B. (2009). The impact of distance onknowledge transfer effectiveness in multinational corporations.Journal of International Management, 15, 1–14.

Ambos, T. C., Ambos, B., & Sclegelmilch, B. B. (2006). Learningfrom foreign subsidiaries: An empirical investigation of head-quarters� benefits from reverse knowledge transfers. Interna-tional Business Review, 15, 294–312.

Barney, J. B. (1991). Firm resources and sustained competitiveadvantage. Journal of Management, 17(1), 99–120.

Bartlett, C. A., & Ghoshal, S. (2002). Managing Across Borders: TheTransnational Solution (2nd ed.). Boston, Mass: Harvard BusinessSchool Press.

Bartlett, C. A., Ghoshal, S., & Beamish, P. (2007). TransnationalManagement (5th ed.). New York: McGraw-Hill/Irwin.

Bjorkman, I., Barner-Rasmussen, W., & Li, L. (2004). Managingknowledge transfer in MNCs: the impact of headquarters controlmechanisms. Journal of International Business Studies, 35(5),443–455.

Bjorkman, I. C., Fey, F., & Hyeon Jeong, P. (2007). Institutionaltheory and MNC subsidiary HRM practices: Evidence from athree-country study. Journal of International Business Studies,38(3), 430–446.

Blomstrom, M., & Kokko, A. (1998). Multinational corporations andspillovers. Journal of Economic Surveys, 12(3), 247–277.

Buckley, P. J. (2006). The multinational enterprise and theglobalization of knowledge. Palgrave Macmillan.

Buckley, P. J. (2009). The impact of the global factory on economicdevelopment. Journal of World Business, 44(2), 131–143.

Buckley, P. J., & Carter, M. J. (2002). Process and structure inknowledge management practices of British and U.S., multi-national enterprises. Journal of International Management, 8,29–48.

Buckley, P. J., Clegg, J., & Tan, H. (2003). The art of knowledgetransfer in China�s telecommunications manufacturing industry.Management International Review, 43(Special issue 2), 67–93.

Buckley, P. J., & Ghauri, P. (2004). Globalisation, economicgeography and the strategy of multinational enterprises. Journalof International Business Studies, 35(2), 81–98.

Buxey, G. (2005). Globalisation and manufacturing strategy in theTCF industry. International Journal of Operations & ProductionManagement, 25(2), 100–113.

Chini, T. C. (2004). Effective Knowledge Transfer in MultinationalCorporations. New York: Palgrave Macmillan.

Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A newperspective on learning and innovation. Administrative ScienceQuarterly, 35(1), 128–152.

Cui, A. S., Griffith, D. A., Cavusgil, S. T., & Dabic, M. (2006). Theinfluence of market and cultural environmental factors ontechnology transfer between foreign MNCs and local subsidiar-ies: A Croatian illustration. Journal of World Business, 41(2),100–111.

Daft, R. L., & Lengel, R. H. (1986). Organizational informationrequirements, media richness and structural design. Manage-ment Science, 32(5), 554–571.

Demirbag, M., & Glaister, K. (2010). Factors determining offshorelocation choice for R&D projects: A comparative study ofdeveloped and emerging regions. Journal of ManagementStudies, 47(8), 1534–1560.

Easterby-Smith, M., Lyles, M. A., & Tsang, E. W. K. (2008). Inter-organizational knowledge transfer: Current themes and futureprospects. Journal of Management Studies, 45(4), 677–690.

Eisenhardt, K. M. (1989). Building theories from case studyresearch. The Academy of Management Review, 14(4),532–550.

Eisenhardt, K. M., & Graebner, M. E. (2007). Theory building fromcases: Opportunities and challenges. Academy of ManagementJournal, 50(1), 25–32.

Eisenhardt, K. M., & Martin, J. A. (2000). Dynamic capabilities:what are they? Strategic Management Journal, 21(10–11),1105–1121.

Engelhard, J., & Nagele, J. (2003). Organizational learning insubsidiaries of multinational companies in Russia. Journal ofWorld Business, 38, 262–277.

Fang, Y., Wade, M., Delios, A., & Beamish, P. W. (2007).International diversification, subsidiary performance, and mobil-ity of knowledge sources. Strategic Management Journal, 28,1053–1064.


Fey, C. F., & Furu, P. (2008). Top management incentive compen-sation and knowledge sharing in multinational corporations.Strategic Management Journal, 29, 1301–1323.

Flyvbjerg, B. (2006). Five misunderstandings about case-studyresearch. Qualitative Inquiry, 12(2), 219–245.

Foss, N. J., & Pedersen, T. (2002). Transferring knowledge in MNCs:The role of sources of subsidiary knowledge and organizationalcontext. Journal of International Management, 8(1), 49–67.

Frost, T. S. (2001). The geographic sources of foreign subsidiaries�innovations. Strategic Management Journal, 22(2),101–123.

Frost, T. S., & Zhou, C. (2005). R&D co-practice and ‘‘reverse’’knowledge integration in multinational firms. Journal of Inter-national Business Studies, 36, 676–687.

Ghauri, P. (2004). Designing and Conducting Case Studies inInternational Business Research. Handbook of QualitativeResearch Methods for International Business. R. Marschan-Piekkari and C. Welch, Edward Elgar: 109–124.

Ghoshal, S., & Bartlett, C. A. (1988). Creation, adoption, anddiffusion of innovations by subsidiaries of multinational corpo-rations. Journal of International Business Studies, 19(3),365–388.

Gnaywali, D. R., Singal, M., & Mu, S. C. (2009). Knowledge tiesamong subsidiaries in MNCs: A multilevel conceptual model.Journal of International Management, 15, 387–400.

Grant, M. (1996a). Toward a knowledge-based theory of the firm.Strategic Management Journal, 17, 109–122 (Special Issue).

Grant, M. (1996b). Prospering in dynamically-competitive environ-ments: Organizational capability as knowledge integration.Organization Science, 7(4), 375–387.

Gupta, A. K., & Govindarajan, V. (1991). Knowledge flows and thestructure of control within multinational corporations. Academyof Management Review, 16(4), 768–792.

Gupta, A. K., & Govindarajan, V. (1994). Organizing for knowledgeflows within MNCs. International Business Review, 3(4),443–457.

Gupta, A., & Govindarajan, V. (2000). Knowledge flows withinmultinational corporations. Strategic Management Journal, 21,473–496.

Hakanson, L., & Nobel, R. (2000). Technology characteristics andreverse technology transfer. Management International Review,40, 29–48.

Hakanson, L., & Nobel, R. (2001). Organizational characteristicsand reverse knowledge transfer. Management InternationalReview, 41(4), 396–420.

Hansen, M. T., Mors, M. L., & Lovas, B. (2005). Knowledge sharing inorganizations: Multiple networks, multiple phases. Academy ofManagement Journal, 48, 776–793.

Hartley, J. (2004). Case study research. In C. Cassell & G. Symon(Eds.), Essential Guide to Qualitative Methods in OrganizationalResearch. London: Sage.

Hedlund, G. (1994). A model of knowledge management and the N-form corporation. Strategic Management Journal, 15, 73–90(Special issue).

Hong, J. F. L., Easterby-Smith, M., & Snell, R. (2006). Transferringorganizational learning systems to Japanese subsidiaries inChina. Journal of Management Studies, 43, 1027–1058.

Huselid, M. A. (1995). The impact of human resource managementpractices on turnover, productivity, and corporate financialperformance. Academy of Management Journal, 38, 635–672.

Hymer, S. H. (1976). The International Operations of NationalFirms: A Study of Direct Foreign Investment. Cambridge, MA:MIT Press.

Inkpen, A. C., & Pien, W. (2006). An examination of collaborationand knowledge transfer: China-Singapore Suzhou Industrial Park.Journal of Management Studies, 43(4), 779–811.

Inkpen, A. C., & Tsang, E. W. K. (2005). Social capital, networks andknowledge transfer. Academy of Management Review, 30(1),146–165.

Jensen, R., & Szulanski, G. (2004). Stickiness and the adaptation oforganizational practices in cross-border knowledge transfers.Journal of International Business, 35, 508–523.

Johanson, J., & Wiedersheim, J.-P. (1975). The internationalizationof the firm - four Swedish cases. Journal of ManagementStudies, 12(3), 305–322.

Kogut, B., & Zander, U. (1992). Knowledge of the firm, combinativecapabilities, and the replication of technology. OrganizationScience, 3(3), 383–397.

Kogut, B., & Zander, U. (1993). Knowledge of the firm and theevolutionary theory of the multinational corporation. Journal ofInternational Business Studies, 24(4), 625–645.

Kostova, T. (1999). Transnational transfer of strategic organiza-tional practices: A contextual perspective. Academy of Manage-ment Review, 24, 308–324.

Lane, P. J., Koka, B. R., & Pathak, S. (2006). The reification ofabsorptive capacity: A critical review and rejuvenation of theconstruct. Academy of Management Review, 31(4), 833–863.

Marschan-Piekkari, R. & Welch, C. (2004). Qualitative researchMethods in International Business: The State of the Art.InMarschan-Piekkari, R. & C. Welch, (Eds.), Handbook of Quali-tative Research Methods for International Business: EdwardElgar.

Michailova, S., & Mustaffa, Z. (in press). Subsidiary knowledge flowsin multinational corporations: Research accomplishments, gaps,and opportunities. Journal of World Business, in press, Cor-rected Proof.

Minbaeva, D. (2007). Knowledge transfer in multinational corpora-tions. Management International Review, 47, 567–593.

Minbaeva, D., Pedersen, T., Bjorkman, I., Fey, C. F., & Park, H. J.(2003). MNC knowledge transfer, subsidiary absorptive capacity,and HRM. Journal of International Business Studies, 34(6),586–599.

Mohr, L. B. (1969). Determinants of innovation in organizations.American Political Science Review, 63(1), 111–126.

Monteiro, L. F., Arvidsson, N., & Birkinshaw, J. (2008). Knowledgeflows within multinational corporations: Explaining subsidiaryisolation and its performance implications. Organization Sci-ence, 19(1), 90–107.

Mowery, D. C., Oxley, J. E., & Silverman, B. S. (1996). Strategicalliances and interfirm knowledge transfer. Strategic Manage-ment Journal, 17, 77–91 (Special Issue).

Nonaka, I., Byosiere, P., Borucki, C. C., & Konno, N. (1994).Organizational knowledge creation theory: A first comprehen-sive test. International Business Review, 3(4), 337–351.

Nonaka, I., & Takeuchi, I. (1995). The Knowledge CreatingCompany. Oxford University Press.

Noorderhaven, N., & Harzing, A.-W. (2009). Knowledge-sharing andsocial interaction within MNEs. Journal of International BusinessStudies, 40, 719–741.

Nordas, H. K. (2004). The Global Textile and Clothing Industry postthe Agreement on Textiles and Clothing. Geneva: World TradeOrganization.

Penrose, E. T. (1959). The Theory of the growth of the firm.Oxford: Blackwell.

Persson, M. (2006). The impact of operational structure, lateralintegrative mechanisms and control mechanisms on intra-MNEknowledge transfer. International Business Review, 15,547–569.

Petras, J., & Veltmeyer, H. (2007). Multinationals on Trial: ForeignInvestment Matters. Aldershot: Ashgate.

Phene, A., Madhok, A., & Liu, K. (2005). Knowledge transfer withinthe multinational firm: What drives the speed of transfer?Management International Review, 45, 53–74 (special issue/2).


Piscitello, L., & Rabbiosi, L. (2006). How does knowledge transferfrom foreign subsidiaries affect parent companies� innovativecapacity? DRUID Summer Conference 2006 on Knowledge, Inno-vation and Competitiveness: Dynamics of Firms, Networks,Regions and Institutions, CBS, Copenhagen, Denmark, June18–20, 2006.

Porter, M. E. (1980). Competitive Strategy. New York: N.Y.,FreePress.

Rabbiosi, L. (2011). Subsidiary roles and reverse knowledge trans-fer: An investigation of the effects of coordination mechanisms.Journal of International Management, 17(2), 97–113.

Rabbiosi, L., Mudambi, R., & Piscitello, L. (2007). Mandates andMechanisms: The Impact on Reverse Knowledge Transfer inMNEs. DRUID Summer Conference 2007 on Appropriability,proximity, Routines and Innovation, CBS, Copenhagen, Denmark,June 18–20, 2007.

Reiche, B. S., Harzing, A.-W., & Kraimer, M. L. (2009). The role ofinternational assignees� social capital creating inter-unit intel-lectual capital. Journal of International Business Studies, 40,509–526.

Schlegelmilch, B. B., Ambos, B., & Chini, T. C. (2003). Are youready to learn from your offshore affiliates. European BusinessForum, 16, 50–54.

Schlegelmilch, B. B., & Chini, T. C. (2003). Knowledge transferbetween marketing functions in multinational companies: aconceptual model. International Business Review, 12(2),215–232.

Schmid, S., & Schurig, A. (2003). The development of criticalcapabilities in foreign subsidiaries: Disentangling the role of thesubsidiary�s business networks. International Business Review,12, 755–782.

Schulz, M. (2003). Pathways of relevance. Exploring inflows ofknowledge into sub-units of multinational corporations. Organi-zation Science, 14, 440–459.

Simonin, B. L. (1999). Ambiguity and the process of knowledgetransfer in strategic alliances. Strategic Management Journal,20(7), 595–623.

Szulanski, G. (1996). Exploring internal stickiness: Impediments tothe transfer of best practice within the firm. Strategic Manage-ment Journal, 17, 27–43.

Tallman, S. B., Jenkins Henry, N., & Pinch, S. (2004). Knowledgeclusters and competitive advantage. Academy of ManagementReview, 29(2), 258–271.

Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilitiesand strategic management. Strategic Management Journal,18(7), 509–533.

Tsai, W. (2002). Social structure of ‘‘coopetition’’ within amultiunit organization: Coordination, competition, and intraor-ganizational knowledge sharing. Organization Science, 13,179–190.

van Wijk, R., Jansen, J. J. P., & Lyles, M. A. (2008). Inter- and intra-organizational knowledge transfer: A meta-analytic review andassessment of its antecedents and consequences. Journal ofManagement Studies, 45(4), 830–853.

Vancil, R. F. (1979). Decentralization: Managerial Ambiguity byDesign. Homewood, Ill.: Dow Jones-Irwin.

Wang, Y., & Suh, C. S. (2009). Towards re-conceptualization of firminternationalization: Heterogeneous process, subsidiary rolesand knowledge flow. Journal of International Management, 15,447–459.

Welch, C., Piekkari, R., Plakoyiannaki, E., & Paavilainen-Manty-maki, E. (2011). Theorising from case studies: Towards apluralist future for international business research. Journal ofInternational Business Studies, 42, 740–762 (5 Special Issue:Qualitative Research in International Business).

Williams, C. (2009). Subsidiary level determinants of global initia-tives in multinational corporations. Journal of InternationalManagement, 15, 92–104.

World Bank (2007). World Bank development indicators. Accessedat <http://www.worldbank.org>.

Yang, Q., Mudambi, R., & Meyer, K. (2008). Conventional andreverse knowledge flows in multinational corporations. Journalof Management, 34(5), 882–902.

Yin, R. (1994). Case Study Research (2nd ed.). Thousand Oakes, Ca.,Sage.

Zhao, Z. J., & Anand, J. (2009). A multilevel perspective onknowledge transfer: Evidence from Chinese automotive indus-try. Strategic Management Journal, 30, 959–983.

MARTINA MCGUINNESS works in the Man-agement School at the University of Shef-field. She researches in the area ofinternational business and has a particularinterest in the role of managerial percep-tions upon organisational behaviour andstrategic decision making.

MEHMET DEMIRBAG works in the Manage-ment School at the University of Sheffield.He researches in the area of internationalbusiness and has a particular interest in therole of managerial perceptions upon organ-isational behaviour and strategic decisionmaking. Recent research interests havebeen focused around the internationalisa-tion of R&D activities, product/regionaldiversification strategies of MNEs, and MNEsfrom emerging markets.

SASANKA BANDARA works for Stretchline(Pvt) Limited as Head of Dyeing Technology.He was previously Dyehouse Manager forCoats plc in Sri Lanka. His areas of interestinclude dyeing & colouration technology,international business and knowledgemanagement.

http://www.worldbank.org

Documents

Towards a multi-perspective model of reverse knowledge transfer in multinational enterprises: A case study of Coats plc