Does the explanatory power of the OLI approach differ among sectors and business functions? Evidence from firm-level data

A ani is, Spy os; Hollens ein, Heinz; S ucki, Tobias A icle Does he explana o y powe o he OLI app oach di e among sec o s and business unc ions? E idence om i m-le el da a Economics: The Open-Access, Open-Assessmen E-Jou nal P o ided in Coope a ion wi h: Kiel Ins i u e o he Wo ld Economy – Leibniz Cen e o Resea ch on Global Economic Challenges Sugges ed Ci a ion: A ani is, Spy os; Hollens ein, Heinz; S ucki, Tobias (2016) : Does he explana o y powe o he OLI app oach di e among sec o s and business unc ions? E idence om i m-le el da a, Economics: The Open-Access, Open-Assessmen E-Jou nal, ISSN 1864-6042, Kiel Ins i u e o he Wo ld Economy (I W), Kiel, Vol. 10, Iss. 2016-12, pp. 1-46, h ps://doi.o g/10.5018/economics-ejou nal.ja.2016-12 This Ve sion is a ailable a : h ps://hdl.handle.ne /10419/141667 S anda d-Nu zungsbedingungen: Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen Zwecken und zum P i a geb auch gespeiche und kopie we den. Sie dü en die Dokumen e nich ü ö en liche ode komme zielle Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich machen, e eiben ode ande wei ig nu zen. So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen (insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en, gel en abweichend on diesen Nu zungsbedingungen die in de do genann en Lizenz gewäh en Nu zungs ech e. Te ms o use: Documen s in EconS o may be sa ed and copied o you pe sonal and schola ly pu poses. You a e no o copy documen s o public o comme cial pu poses, o exhibi he documen s publicly, o make hem publicly a ailable on he in e ne , o o dis ibu e o o he wise use he documen s in public. I he documen s ha e been made a ailable unde an Open Con en Licence (especially C ea i e Commons Licences), you may exe cise u he usage igh s as speci ied in he indica ed licence. h p://c ea i ecommons.o g/licenses/by/4.0/ Recei ed Decembe 4, 2015 Published as Economics Discussion Pape Decembe 8, 2015 Re ised May 11, 2016 Accep ed May 13, 2016 Published June 1, 2016 © Au ho (s) 2016. Licensed unde he C ea i e Commons License - A ibu ion 4.0 In e na ional (CC BY 4.0) Vol. 10, 2016-12 | June 01, 2016 | h p://dx.doi.o g/10.5018/economics-ejou nal.ja.2016-12 Does he Explana o y Powe o he OLI App oach Di e Among Sec o s and Business Func ions? E idence om Fi m-le el Da a Spy os A ani is, Heinz Hollens ein, and Tobias S ucki Abs ac As se ices and goods di e wi h espec o impo an cha ac e is ics, one may expec ha he de e minan s o in e na ionalisa ion a e no iden ical o se ices and manu ac u ing. Su p isingly, he e is p ac ically no i m-le el esea ch con as ing he wo sec o s in his espec . To ill his gap, he au ho s aim a iden i ying o he wo sec o s he de e minan s o a i m’s p opensi y o engage in expo s o FDI as well as he ac o s de e mining a i m’s FDI s a egy in e ms o business unc ions. They ind ha an OLI-based model is well sui ed o explaining he p opensi y o go in e na ional as well as he di e ences be ween speci ic o ms o FDI in e ms o business unc ions bo h o manu ac u ing and se ices. As expec ed, he explana o y powe o he OLI app oach is s onge o manu ac u ing han se ice ac i i ies. Da a a ailable upon eques (see he ollowing link): h p://www.economics-ejou nal.o g/da a-a ailabil y-dp-2015-67/ iew JEL F23 Keywo ds Manu ac u ing s. se ices in e na ionalisa ion; o sho ing s. expo s; in e na ionalisa ion o business unc ions; mul ina ional companies; in e na ional business s a egy Au ho s Spy os A ani is, KOF Swiss Economic Ins i u e, ETH Zu ich, CH-8092 Zu ich, [email p o ec ed] Heinz Hollens ein, KOF Swiss Economic Ins i u e, ETH Zu ich, CH-8092 Zu ich, [email p o ec ed] Tobias S ucki, KOF Swiss Economic Ins i u e, ETH Zu ich, CH-8092 Zu ich, [email p o ec ed] Ci a ion Spy os A ani is, Heinz Hollens ein, and Tobias S ucki (2016). Does he Explana o y Powe o he OLI App oach Di e Among Sec o s and Business Func ions? E idence om Fi m-le el Da a. Economics: The Open-Access, Open-Assessmen E-Jou nal, 10 (2016-12): 1—46. h p://dx.doi.o g/10.5018/economics- ejou nal.ja.2016-12 www.economics-ejou nal.o g 2 1 In oduc ion The long- e m end o an inc easing sha e o se ice sec o FDI has accele a ed o e he las wen y yea s (Dunning and Lundan 2008; Kundu and Me chan 2008). Ne e heless, empi ical esea ch on FDI is s ill concen a ed on manu ac u ing. As he basic cha ac e is ics o se ices and goods di e – he main speci ics o se ices a e in angibili y, insepa abili y o p oduc ion and consump ion, he e ogenei y, pe ishabili y and es ic ed owne ship (Buckley e al. 1992) – one would expec ha he de e minan s o in e na ionalisa ion a e no he same (o no ele an o same ex en ) in he wo sec o s (see, e.g., Boddewyn e al. 1986; Dunning 1989; Con ac o e al. 2003). This may hold ue al hough he sepa a ion line be ween goods and se ices has become qui e blu ed (see, e.g., G ön oos 1999; Jack e al. 2008), e lec ing he g owing se ice con en o manu ac u ing and a ce ain endency owa ds an indus ialisa ion o pa s o he se ice sec o (Miozzo and Miles 2002; Gue ie i and Meliciani 2005). Agains his backg ound, i is su p ising ha i m-le el econome ic esea ch dealing wi h he in e na ionalisa ion o se ice i ms by way o FDI is sca ce and s ill domina ed by he analysis o speci ic indus ies.1 The ew s udies ha co e he en i e se ice sec o , wi h only some excep ions (Li and Guisinge 1992, Kundu e al. 2008), a e de o ed o aspec s o in e na ionalisa ion ha we do no analyse in he p esen con ibu ion.2 Fu he mo e, and mos impo an ly, he e is ha dly any empi ical wo k con as ing se ices and manu ac u ing wi h espec o he ac o s de e mining FDI. In ac , we a e awa e o only h ee s udies which sys ema ically explo e _________________________ 1 Econome ic indus y s udies became a ailable, wi h some excep ions, only in ecen yea s. To men ion a e Lin (2010) and Na ayanan and Bha (2010) o ICT/so wa e; Con ac o and Kundu (1998), B own e al. (2003), León-Da de e al. (2010) and Villa e al. (2012) o he ho el indus y; U sacki and Ve insky (1992), Nachum and Wymbs (2005) and Ma iscal e al. (2012) o inancial se ices; Te ps a and Yu (1988); Rod iguez and Nie o (2012) o some knowledge-in ensi e se ice indus ies. 2 Fo example: o eign en y mode choice (E amilli and Rao 1993); ou sou cing (Mu ay and Ko abe 1999); impac o in e na ionalisa ion on i m pe o mance (Con ac o e al. 2003); causal ela ionship be ween manu ac u ing FDI and FDI in business se ices (Ne ussi and Schwellnus 2010). www.economics-ejou nal.o g 3 di e ences be ween hese wo sec o s in his espec .3 Fi s ly, Lejp as (2009) analyses o Ge many, using i m-le el su ey da a, a i m’s choice be ween domes ic ac i i y and expo ing (accompanied by FDI and, al e na i ely, wi hou FDI) and be ween domes ic ac i i y and FDI (wi hou expo s). The explana o y a iables mainly e lec a i m’s capabili ies (inno a i eness, size, e c.) and i s “en i onmen ” (compe i ion; a ac i eness o domes ic loca ions). Secondly, Wagne (2014) in es iga es, based on Ge man da a, whe he he well documen ed pecking o de o manu ac u ing i ms wi h espec o he way o going in e na ional (only he mos p oduc i e i ms a e ac i e in FDI, less p oduc i e ones a e expo e s, he leas p oduc i e companies exclusi ely se e domes ic ma ke s) is also cha ac e is ic o he se ice sec o . This s udy is based on he “he e ogeneous i m app oach” (Meli z 2003), which is qui e o en applied in ecen yea s, bu only o manu ac u ing. Finally, Py and Ha em (2009) analyse he loca ion o o eign in es men p ojec s o F ench companies, wi h hos coun y a ibu es se ing as explana o y a iables. This s udy is, o ou knowledge, he only one ha di e en ia es be ween in es men s in ol ing speci ic business unc ions. In he p esen con ex , howe e , his pape is no e ec i ely ele an as i ocuses on he loca ion choice, which is no he aim o ou analysis ha deals wi h he de e minan s o he choice o he ype o o eign ac i i y. All in all, we may conclude ha he empi ical knowledge wi h espec o he simila i ies and di e ences be ween manu ac u ing and se ices i ms wi h espec o he de e minan s o he in e na ionalisa ion o economic ac i i ies o i ms is limi ed. The p esen s udy aims a illing his esea ch gap by iden i ying and compa ing o he wo sec o s (a) he d i e s o he in e na ionalisa ion o i m ac i i y in e ms o expo s and FDI, and (b) he de e minan s o he choice be ween speci ic o ms o FDI in e ms o business unc ions. In o de o analyse hese p oblems we o mula e wo empi ical models. In Model I (“INT_p o- pensi y”), we es ima e he p obabili y o a i m o belonging o one o he ollowing (mu ually exclusi e) h ee ca ego ies: “selling on domes ic ma ke s only” (DOMESTIC), “se ing, addi ionally, expo ma ke s (wi hou any o - _________________________ 3 Ano he wo s udies ha deal wi h di e ences be ween he wo sec o s wi h espec o he di ec o eign p esence a e B ou he s and B ou he s (2003) and Jaklic e al. (2012). The o me e e s o he en y mode choice (wholly-owned subsidia ies s. join en u es), he la e o he impac o in e na ional ou sou cing/o sho ing on i m pe o mance. www.economics-ejou nal.o g 4 sho ing)” (EXPORT), and “o sho ing” (FDI). In Model II (“INT_ unc ion”), we de e mine he likelihood o a i m o loca ing ab oad speci ic (combina ions o ) business unc ions, ha is (a) “o sho ing dis ibu ion and/o p oduc ion” (FDI_DP) and (b) “o sho ing R&D ac i i ies, in addi ion o dis ibu ion and/o p oduc ion” (FDI_RDP). The wo g oups o i ms, which a e subca ego ies o FDI, again a e mu ually exclusi e. Theo e ical amewo k o he empi ical analysis is he well-known OLI pa adigm (Dunning 2000). We expec ha he explana o y powe o he OLI a iables is lowe in he case o he se ice sec o han o manu ac u ing, p ima ily because o he la ge he e ogenei y o i s ac i i ies, he pa icula impo ance o so ac o s (as opposed o echnology- ela ed a iables) as well as he signi icance o idiosync a ic and si ua ion-speci ic ac o s. Mo eo e , we a gue ha he explana o y powe o OLI a iables is s onge o a complex s a egy o FDI in e ms o business unc ions han o a less complex one, because he cos s incu ed by he i s s a egy a e highe and may hus be o e come only in he p esence o la ge O-ad an ages. Finally, we men ion al eady a his s age ha he empi ical es s o hese hypo heses a e p ima ily based on he OL-pa o he OLI pa adigm, as he da a a hand allow only a e y ough measu emen o I- ad an ages (see Subsec ion 4.2). The analysis is an ex ension o Hollens ein (2005), who es ima ed, using Swiss da a o 1998, OLI-based models ha a e s uc u ally simila o Model I and II bu did so only o he en i e business sec o . The esea ch also goes beyond he al eady men ioned analysis o Lejp as (2009), who disagg ega ed he business sec o in manu ac u ing and se ices and es ima ed a model compa able o ou Model I ha dis inguishes be ween “expo ing only” and “di ec o eign p esence”. This esea che , howe e , did no u he di e en ia e wi hin he ca ego y o i ms wi h FDI as we do in Model II (di e ences be ween wo combina ions o business unc ions). Simila ly, ou con ibu ion goes beyond Wagne (2014) who also es ima ed Model I only. By es ima ing Model I and II we hus signi ican ly add o p e ious e idence on he di e ences be ween manu ac u ing and se ice companies wi h espec o he de e minan s o in e na ional ac i i ies. The pape is based on a unique i m-le el da ase con aining in o ma ion om 1921 companies o he Swiss business sec o ha esponded o a comp ehensi e su ey we conduc ed in 2010 among a andom sample d awn om he o icial en e p ise census o 2008 ( esponse a e: 42%). The a ailable da a allow a ich www.economics-ejou nal.o g 5 speci ica ion o he explana o y a iables o he wo models, in pa icula ly wi h espec o he OL-pa o he OLI model.4 Using Swiss da a may be o gene al in e es as he p ocess o in e na ionalisa ion o he Swiss economy is pa icula ly ad anced (UNCTAD 2015). In line wi h ou hypo heses, we ind, o bo h sec o s, ha an OLI-based model (in pa icula he OL-pa ) is well sui ed no only o explaining he p opensi y o i ms o go in e na ional by means o expo s and/o FDI (Model I) bu also o iden i y he di e ences be ween speci ic o ms o FDI in e ms o business unc ions (Model II). In all models, he explana o y powe o he OLI app oach is s onge o manu ac u ing han o se ices. The esul s o manu ac u ing a e in line wi h he s ages iew o in e na ionalisa ion, his is only pa ly he case o he se ice sec o . The s uc u e o he pape is as ollows: In he nex sec ion, we p esen he concep ual amewo k and he ela ed hypo heses. In Sec ion 3, we desc ibe he da abase and gi e some in o ma ion on he incidence o in e na ional ac i i ies o Swiss companies. Sec ion 4 is de o ed o model speci ica ion, and in Sec ion 5 we p esen he empi ical esul s and compa e hem wi h p e ious wo k. In he inal sec ion, we summa ise and d aw some conclusions. 2 Concep ual amewo k and ela ed hypo heses 2.1 Theo e ical backg ound Since Hyme (1976), he heo y o in e na ional in es men is based on he assump ion o impe ec ma ke s. Unde hese condi ions, i m-speci ic capabili ies yield a compe i i e edge independen o he economic a ac i eness o di e en loca ions (see Ca es 1982) and, mo e o malised, he “new ade heo y” (see, e.g., Helpman 1984). Mo eo e , he “ ansac ion cos heo y” s a es ha a i m engages in FDI whene e he cos s o se ing up and unning a ansna ional o ganisa ion o ac i i ies a e lowe han hose o ex e nal ma ke ansac ions (Rugman 1981; Henna 1982; Buckley and Casson 1985; Williamson 1985). In _________________________ 4 The same da a ha e been success ully used o explaining he loca ion o he FDI o Swiss i ms (A ani is e al. 2015) as well as hei choice be ween equi y-based and. non-equi y o eign en y modes (Hollens ein and Be ge 2015). www.economics-ejou nal.o g 6 addi ion, he e a e many pa ial hypo heses explaining speci ic aspec s o in e na ionalisa ion ha a e oo ed in di e en sub-disciplines o economics such as indus ial o ganisa ion, managemen sciences, e olu iona y economics (Dunning 2000). As ea ly as in he 1970s, Dunning a gued ha no single app oach is able o explain a i m’s in e na ional ac i i ies (Dunning 1977, 1979). He p oposed an eclec ic heo y o in e na ional p oduc ion, he well-known OLI pa adigm, which he u he de eloped o e he yea s o accoun o changing ea u es o he in e na ional economy and new heo e ical app oaches. In he mos ecen e sion (Dunning and Lundan 2008) he OLI model applies no only o in e na ional p oduc ion bu also o o he business unc ions. In addi ion, i emphasises he s a egic aspec s o in e na ionalisa ion mo e explici ly by d awing on he “ esou ce-based” (We ne el , 1984) o “dynamic capabili y” (Teece e al. 1997) iew o he i m, o he concep o he “knowledge-based company” (Kogu and Zande 1993). We pos ula e ha he OLI pa adigm is an app op ia e heo e ical amewo k o speci ying he explana o y pa o he wo models we es ima e in his pape ha deals wi h he p opensi y o i ms o in e na ionalise hei ac i i ies (Model I: INT_p opensi y) and wi h he likelihood o i ms o loca ing ab oad speci ic (combina ions o ) business unc ions (Model II: INT_ unc ion). As men ioned in Sec ion 1, we expec ha he explana o y powe o he OLI a iables is highe in he case o he manu ac u ing han he se ices sec o (Model I) and highe o complex s a egies o FDI in e ms o business unc ions han o less complex ones (Model II). The OLI model basically accoun s o h ee g oups o explana o y a iables: 1) “Owne ship-speci ic (O) ad an ages”, which a ise mainly om he a ailabili y o i m-speci ic knowledge, human capi al, manage ial skills, p ope y igh s, ma ke ing ou le s, access o inance o in e na ional expe ience. Fi ms cha ac e ised by la ge O-ad an ages a e mo e likely o go in e na ional by way o expo ing and/o FDI. 2) “Loca ion-speci ic (L) ad an ages“, which oo in di e ences be ween o eign and domes ic loca ions wi h espec o ac o cos s, poli ical s abili y, he egula o y amewo k o dis ance. This ype o ad an ages de e mines why a company chooses an FDI-based s a egy o in e na ionalisa ion a he han one ha www.economics-ejou nal.o g 7 is (pu ely) expo -o ien ed (L-ad an ages o o eign loca ions as a p econdi ion o FDI). L-ad an ages, ob iously, would also be ele an o explaining whe e a i m loca es i s FDI; howe e , his aspec is no ele an o ou analysis as we do no aim a explaining loca ion choices. 3)“In e nalisa ion (I) ad an ages”, which allow i ms o educe isks and cos s o ansac ions in impe ec ma ke o echnology o key in e media e p oduc s. I- ad an ages a e ele an o choosing among se e al modes o o eign engagemen s (e.g., wholly-owned subsidia y s. join en u es s. non-equi y based co- ope a ions s. o he ela i ely loose ela ionships including expo s). In his pape , we only dis inguish be ween wo modes o o eign ac i i y, ha is a) FDI (i.e. equi y-based o eign p esence) and b) expo s (which may (o may no ) be complemen ed by non-equi y based o eign co-ope a ion). In he p esence o signi ican I-ad an ages, FDI is he p e e ed mode o in e na ionalisa ion. We no e again (see Sec ion 1) ha , due o insu icien da a, he empi ical analysis is p ima ily based on he OL-pa o he OLI model. We use some o he explana o y a iables ha complemen he h ee key elemen s o he OLI model. Fi s , depending on he ma ke en i onmen (deg ee o compe i ion, ma ke g ow h), i may be bene icial (o necessa y) o a i m o ex end i s ac i i ies o o eign loca ions (e.g., “ i s mo e ” o “ ollow he leade ” s a egy). Second, we p esume ha indus y a ilia ion is an impo an con ol a iable as he ele ance o O-, L- and I-ad an ages a e likely o di e among indus ies, in pa icula in he (he e ogeneous) se ice sec o (Dunning 1989). The “model o he he e ogeneous i ms” (see he seminal pape s o Meli z 2003 and Helpman e al. 2004) would be ano he app oach one could use o analysing in e na ionalisa ion s a egies o i ms. Recen empi ical esea ch dealing wi h manu ac u ing la gely suppo s he main p oposi ion o his model, s a ing ha only he mos p oduc i e i ms en e o eign ma ke s by FDI, he less p oduc i e ones a e expo e s and he leas p oduc i e companies exclusi ely se e domes ic ma ke s ( o a e iew see, e.g., G eenaway and Knelle 2007 o Wagne 2011). In he se ice sec o , howe e , he “pecking o de ” wi h espec o p oduc i i y seems o be di e en . Acco ding o Wagne (2014) expo e s o se ices a e mo e p oduc i e han se ice companies ha en e o eign ma ke s by means o FDI. No wi hs anding he e idence o his widely used app oach, we p e e he OLI model as amewo k o analysis as i p o ides a mo e de ailed iew www.economics-ejou nal.o g 8 o he d i e s o in e na ionalisa ion han he “p oduc i i y app oach”. In ac , he a iable “p oduc i i y” is a kind o a “black box” ep esen ing a whole bundle o speci ic s eng hs (and weaknesses) o a i m which in luence he choice be ween di e en in e na ionalisa ion s a egies. By using he OLI model we a e able o iden i y he ele ance o he indi idual a iables included in he p oduc i i y black box. In doing so, we ollow o he esea che s such as Cas ellani and Zan ei (2007) o Cas ellani and Gio anne i (2010) who emphasised he need o looking inside his black box. They ind o he I alian manu ac u ing sec o ha using inno a ion and o he knowledge- ela ed a iables signi ican ly educes he explana o y powe o p oduc i i y. Falk and Hags en (2015), using Swedish da a, go he same esul o compu e and business se ices. Mo eo e , as will be shown in Subsec ion 4.2, we do no use p oduc i i y as addi ional a iable o ou OLI model, as i is highly co ela ed wi h he O- and I- a iables; including p oduc i i y would hus bias he es ima es (mul icollinea i y).5 2.2 Hypo heses In a i s s ep, we seek o iden i y he de e minan s o he p opensi y o a i m o in e na ionalise i s ac i i ies (Model I: INT_p opensi y) based on he en i e business sec o , al hough we p ima ily a e in e es ed in di e ences in his espec be ween manu ac u ing and se ices i ms. In doing so, we a e able o gene ally “ alida e” ou OLI-based app oach, which hen may se e as e e ence o he subsequen analysis o di e ences/simila i ies be ween he manu ac u ing and he se ice sec o . As se ou in he p e ious subsec ion, O-ad an ages o i ms su ice o explain why hey go in e na ional a all (by means o expo s and/o FDI). Mo eo e , FDI is he p e e ed way o in e na ionalisa ion i a company, addi ionally, bene i s om L-ad an ages o o eign loca ions and is able o in e nalise (in e na ional) ma ke ela ions (I-ad an ages). We hus pos ula e (Model I: INT_p opensi y): _________________________ 5 Mo eo e , eg ession esul s, no p esen ed he e, show ha he a iable “labou p oduc i i y” is s a is ically no signi ican in an equa ion whe e i is added o he OI-pa o he OLI model. www.economics-ejou nal.o g 15 a iable o Model I and Model II espec i ely, and (b) he sample unde lying he es ima es (en i e business sec o , manu ac u ing, se ices). 8 The es ima ion o model I p o ides wo pa ame e ec o s, he i s one e e ing o he ca ego y DOMESTIC, and he second one o FDI wi h EXPORT used as e e ence ca ego y (see Table 4). In he case o Model II, we abula e only he pa ame e ec o o FDI_RDP as FDI_DP is used as e e ence ca ego y. Table 2: Speci ica ion o he dependen a iables Dependen a iable De ini ion Model I INT_ p opensi y The i m belongs o one (and only one) o he ollowing ypes o i ms: DOMESTIC: Fi ms wi h domes ic sales EXPORT: Fi ms wi h domes ic and expo sales (bu no o sho ing) FDI: Fi ms wi h any kind o di ec o eign p esence (in addi ion o domes ic and expo sales) EXPORT is used as e e ence ca ego y Model II INT_ unc ion The i m belongs o one (and only one) o he ollowing ypes o i ms: DOMESTIC: Fi ms wi h domes ic sales EXPORT: Fi ms wi h domes ic and expo sales (bu no o sho ing) FDI_DP: Fi ms ha ing o -sho ed dis ibu ion and/o p oduc ion FDI_RDP: Fi ms ha ing o -sho ed esea ch and de elopmen (R&D) in addi ion o dis ibu ion and/o p oduc ion FDI_DP is used as e e ence ca ego y _________________________ 8 Al hough we p esume ha he deg ee o in e na ionalisa ion is posi i ely ela ed o he numbe o business unc ions loca ed ab oad (see Subsec ion 2.2), he mul inomial logi p ocedu e is mo e adequa e han he o de ed p obi model used, o example, by Basile e al. (2003). Only in he o me case, we ge o each explana o y a iable speci ic coe icien s o he di e en s a egies o in e na ionalisa ion. www.economics-ejou nal.o g 16 The econome ic analysis is based on c oss-sec ion da a. Consequen ly, all a iables, in p inciple, could be endogenous, wha would imply biased pa ame e es ima es. The endogenei y p oblem, o some ex en , is a enua ed as a subs an ial numbe o he explana o y a iables migh be s uc u al in na u e, hus a e only sligh ly changing o e ime (e.g., he sha e o employees wi h e ia y-le el educa ion o he exis ence o R&D ac i i ies). Ne e heless, we canno e ade he gene al endogenei y p oblem inhe en in c oss-sec ion analyses. The e o e, a he han making causal claims we in e p e he es ima ed coe icien s as condi ional co ela ions ha , howe e , do no p eclude an e alua ion o ou hypo heses. 4.2 Explana o y a iables The explana o y a iables cap u e he mos impo an aspec s o O-, L- and I- ad an ages, he i ms’ ma ke en i onmen and con ol a iables e e ing o some s uc u al i m cha ac e is ics (indus y a ilia ion, o eign owne ship). In Table 3 we p esen he exac de ini ion, measu emen and sign expec a ion o he a iables, and Table A.3 in he Appendix shows he ela ed desc ip i e s a is ics. O-ad an ages We expec ha his ca ego y o a iables ep esen ing i m-speci ic capabili ies is posi i ely ela ed o in e na ional ac i i y by means o expo s and/o FDI. We conside , i s ly, wo indica o s o he inno a ion capaci y o i ms, ha is in-house R&D ( &d) and he sales sha e o new o signi ican ly imp o ed p oduc s (inno_sales). Mo eo e , we ake in o accoun he use o high-le el human esou ces ( e ia y_academic). We also inse a measu e o he e ec i eness o he p o ec ion o knowledge (app op iabili y), which co e s pa en ing and o he legal igh s (b ands, copy igh s) as well as in o mal app op iabili y mechanisms (e.g., ime lead o sec ecy). Finally, we include wo indica o s o a i m’s in e na ional expe ience, which is a co e a iable o he s ages iew o in e na ionalisa ion (Johanson and Vahlne 1977), i.e. he age o a company ( i m_age) and he ele ance o expe ience- ela ed obs acles o o eign ac i i ies (obs _expe ience). Fi m age should be posi i ely ela ed o in e na ionalisa ion, whe eas we expec a nega i e sign o he obs acle a iable as i is in e sely ela ed o o eign expe ience. www.economics-ejou nal.o g 17 Table 3: Speci ica ion o he explana o y a iables o Model I o II Explana o y a iable Desc ip ion Expec ed signs O-ad an ages &d R&D ac i i ies, 2006-2008 (yes/no, dummy a iable) + inno_sales Sales sha e o inno a i e p oduc s, loga i hm, 2008 + e ia y_academic Sha e o employees wi h academic educa ion, loga i hm, 2008 + app op iabili y E ec i eness o knowledge p o ec ion (dummy a iable (high/low e ec i eness) based on he a e age o i m assessmen s on a 5-poin Like scale o ou means o p o ec ion: pa en s, adema ks, copy igh s, in o mal mechanisms) + i m_age Age o he i m (yea s), loga i hm, 2008 + O- ela ed obs acles o in e na ionalisa ion: (dummy a iable based on a i m’s assessmen o he ele ance o his speci ic obs acle on a 5-poin Like scale: alue 1 (high) o sco es 4 o 5, o he wise 0 (low)) obs _expe ience Lack o o eign expe ience - L-disad an ages o hos loca ions: L- ela ed obs acles o in e na ionalisa ion: (dummy a iables based on i m assessmen s o he ele ance o speci ic obs acles on a 5-poin Like scale: alue 1 (high) o sco es 4 o 5, o he wise 0 (low)) obs _dis ance La ge geog aphical dis ance - obs _cul u e La ge cul u al di e ence - obs _cos High cos s o in e na ionalisa ion - obs _ egula ion Res ic i e egula ions in o eign loca ions - obs _ins abili y Poli ical ins abili y in a ge coun ies - obs _local_ alue En o ced gene a ion o high local alue added - obs _ ech_ ans e Obliga ion o ans e echnology o local ac o s - obs _IPR_p o ec ion Insu icien p o ec ion o IPRs in hos coun ies - I-ad an ages, i m size size, size2 Numbe o employees (in 1000) and i s squa e, 2008 + / ? coope a ion Co-ope a ion wi h domes ic i ms (yes/no, dummy a iable) + Table 3 con inued www.economics-ejou nal.o g 18 Table 3 con inued Explana o y a iable Desc ip ion Expec ed signs Ma ke en i onmen compe i ion Sha e o i ms a he 3-digi indus y le el which a e con on ed wi h s ong p ice compe i ion, 2008: Fi ms wi h an assessmen o 4 o 5 on a i e poin Like scale + demand_ end Sha e o i ms a he 3-digi indus y le el which bene i om s ongly g owing ma ke s in he pe iod 2006-2011: Fi ms wi h an assessmen o 4 o 5 on a i e poin Like scale) + Con ol a iables o eign The i m is a o eign-owned subsidia y p ima ily ocusing on he domes ic ma ke (yes/no; dummy a iable) - egional_ma ke The local/ egional ma ke a home is su icien ly la ge: (dummy a iable based on a i m’s assessmen on a i e poin Like scale: alue 1 (high) o sco es 4 o 5, o he wise 0 (low)) - indus y Indus y dummies (2-digi ) in manu ac u ing (16) and se ices (9) L-ad an ages In many s udies L-ad an ages se e o explain whe e a i m loca es i s FDI p ojec s. Acco dingly, cha ac e is ics o po en ial a ge egions/coun ies (wage cos s, ac o endowmen , e c.) a e used as measu es o L-ad an ages. In con as , we aim a explaining why a i m uses he one o he o he way o going in e na ional (i.e. o sho ing s. expo ing). To his end, we inse as explana o y a iables a se o i m-le el indica o s ep esen ing disad an ages o o eign coun ies as a whole (i.e. wi hou di e en ia ion among o eign loca ions) as compa ed o Swiss loca ions. These a iables e lec a i m’s assessmen o he ele ance (high/low) o a numbe o loca ion- ela ed obs acles o in es ing ab oad (e.g., “ he egula o y amewo k in hos coun ies is un a ou able o FDI”). The e o e, we expec a nega i e ela ionship be ween he obs acle a iables and FDI. Mo e speci ically, he L-pa o he empi ical model is made up o eigh measu es o obs acles, which acco ding o an explo a o y analysis a e (s a is i- www.economics-ejou nal.o g 19 cally) he mos pe inen ones. In acco dance wi h he g a i y model o in e na ional ade and in es men he model con ains a iables ep esen ing he geog aphical and cul u al dis ance be ween Swi ze land and o eign loca ions (obs _dis ance, obs _cul u e). Fu he mo e, we conside high cos s o in es ing ab oad (obs _cos ). We also inse some measu es depic ing a se o egula ion- ela ed L-disad an ages o hos coun ies: es ic i e egula o y amewo k in gene al (obs _ egula ion); obliga ion o gene a e locally a la ge sha e o alue added (obs _ alue); o ced echnology ans e o local i ms/ins i u ions (obs _ ech_ ans e ); insu icien en o cemen o IPRs in hos coun ies (obs _IPR_p o ec ion). Fu he mo e, we include a measu e o he ele ance o poli ical ins abili y in o eign loca ions as an impedimen o in e na ional ac i i y (obs _ins abili y).9 As men ioned be o e, he “obs acle a iables” a e expec ed o be nega i ely ela ed o FDI. Howe e , in ce ain cases we would no be su p ised o ind a posi i e sign indica ing ha a i m pe cei es (o becomes awa e o ) a speci ic obs acle only when i is eally con on ed wi h such an impedimen (e.g., “insu icien en o cemen o IPRs in hos coun ies”; see he discussion o he es ima ion esul s in Sec ion 5). This in e p e a ion o a posi i e sign in case o “obs acles a iables” is widely-used in inno a ion esea ch (see, e.g., Galia and Leg os 2004; Tou igny and Le 2004), and mo e ecen ly also in he in e na ionalisa ion li e a u e (see, e.g., Ramme and Schmiele 2008; A ani is and Hollens ein 2011). I-ad an ages, i m size I-ad an ages, which a e ele an o choosing be ween o eign en y modes (in ou case be ween FDI and expo s), should be posi i ely ela ed o FDI. Howe e , i is di icul o ind sa is ac o y indica o s o his ype o ad an ages. Fi s ly, om a heo e ical poin o iew, I-ad an ages should be measu ed a he le el o _________________________ 9 Based on an explo a o y analysis o he obs acles o be p esen in o eign loca ions, we do no include some o he po en ial impedimen s o which he “Swiss In e na ionalisa ion Su ey 2010” p o ides da a. To men ion a e high coo dina ion cos s, high inancial isks, insu icien inance, es ic ions o he ans e o p o i s, legal unce ain ies, lack o quali ied manpowe . Fo some o hese obs acles i is qui e su p ising ha hey appa en ly a e no ele an . This is pa ly due o co ela ions wi h obs acles we include in he model (e.g., co ela ion o obs _ins abili y wi h “legal unce ain y”, o obs _cos wi h “high inancial isks” and “insu icien inance”). www.economics-ejou nal.o g 20 indi idual ansac ions (i.e. sepa a ely o each in es men p ojec loca ed ab oad) a he han a i m le el, which is no easible wi h he da a a hand. Secondly, e en a he i m le el, we ha e o ely on p oxy a iables ha a e qui e gene al. We use wo indica o s o I-ad an ages, i.e. i m size and co-ope a ion expe ience ha we in e p e as measu es o a i m’s “capaci y o in e nalize ma ke ela ions”. We assume ha la ge i ms a e in a be e posi ion han small ones o educe ansac ion cos s by in e nalising ma ke ela ionships: economies o scale in he go e nance o o eign a ilia es o join en u es, highe powe o ba gain wi h o eign uni s, ad an ages in moni o ing quali y-s anda ds o be me by a ilia es. Fi m size may hus e ec i ely be used as an o e all indica o o I-ad an ages (al hough i also cap u es size- ela ed O-ad an ages ha a e no explici ly speci ied in he empi ical model). We measu e i m size by he numbe o employees (size) and, o accoun o a possible nonlinea ela ionship, by i s squa e (size2). Fo he quad a ic e m we expec an insigni ican o a nega i e coe icien , he la e meaning ha i m size only ma e s up o a ce ain h eshold. Mo eo e , we main ain ha expe ience gained om co-ope a ing wi h o he companies (coope a ion) can be used as an indica o o I-ad an ages as i may enhance a i m’s capabili y o in e nalise ma ke ela ionships. All in all, as ou speci ica ion o I-ad an ages is no e y sa is ac o y, we in e p e he empi ical esul s o Model I and II, in he i s ins ance, by e e ing o he OL-pa o he OLI model. Ma ke en i onmen We expec ha in ensi e compe i ion on a i m’s ma ke s (compe i ion) en o ces a company o become ac i e in o eign loca ions o is an incen i e o do so (“ ollow he leade ” o “ i s mo e ” s a egy). Fu he mo e, s ongly g owing wo ld ma ke s in he ield o a i m’s ac i i ies (demand_ end) is ano he incen i e o in e na ional engagemen s. We hus expec o bo h a iables a posi i e sign. Con ol a iables We ake accoun o he ac ha some i ms may epo ha hey do no see any obs acle o FDI jus because hey ne e hough o doing business ab oad. To his end, we inse a con ol a iable indica ing ha a i m’s “local/ egional ma ke a home is la ge enough” ( egional_ma ke ), implying ha i does no see any need o go in e na ional (nega i e ela ionship o in e na ionalisa ion). By including his www.economics-ejou nal.o g 21 a iable we make su e ha he es ima es o he “obs acle a iables” (L- disad an ages) a e no biased. Fu he mo e, we con ol o he ac ha o eign-owned companies ( o eign) o en a e p ima ily o ien ed owa ds he Swiss ma ke ; his a iable should hus be nega i ely ela ed o expo ing as well as o FDI. Finally, we inse a la ge se o indus y dummies (indus y) ep esen ing, o example, mac oeconomic condi ions p e ailing in he speci ic indus ies in he e e ence yea 2010. In addi ion, indus y dummies con ol o a (possible) “omi ed a iable bias”. They also should make su e ha he explici ly speci ied a iables e ec i ely cap u e beha iou al di e ences among i ms a he han indus y-speci ic cha ac e is ics. 5 Resul s 5.1 Model I: INT p opensi y 5.1.1 Business sec o (H1) Remembe ha we e alua e he coe icien s o he a iables explaining DOMESTIC and FDI agains hose o he i ms ha a e in e na ionally ac i e solely h ough domes ic and expo sales (EXPORT). The e o e, a s a is ically signi ican nega i e sign o he coe icien o DOMESTIC in combina ion wi h a signi ican posi i e sign o he coe icien o FDI (deno ed as DOMESTIC<EXPORT<FDI) o O- and I- a iables indica es a mono onically inc easing posi i e e ec o a ce ain a iable on a i m’s in e na ionalisa ion, hus ex ending ac i i ies om only domes ic ac i i ies o expo s and u he o FDI (see, e.g., a iable &d). This can be he case also pa ially, namely i he coe icien only inc eases ei he om DOMESTIC o EXPORT (DOMESTIC<EXPORT≅FDI), e.g., a iable app op iabili y, o om EXPORT o FDI (DOMESTIC≅EXPORT<FDI), e.g., inno_sales. In case o he O- ela ed obs acle “lack o o eign expe ience” (“obs _expe ience”) i is he o he way ound as his a iable ep esen s a disad an age o a i m; in his case, we expec an inc easing posi i e e ec on in e na ional ac i i y i DOMESTIC>EXPORT>FDI. Mo eo e , as he L- ela ed obs acles e lec disad an ages o hos coun ies, o sho ing is p e e ed o expo ing i EXPORT>FDI. No ice ha L-disad an ages a e i ele an o explaining a shi om DOMESTIC o EXPORT; we hus expec www.economics-ejou nal.o g 22 ha he coe icien s o he L- a iables a e s a is ically insigni ican in case o DOMESTIC. The Columns 1 and 2 o Table 4 show ha O-ad an ages a e e y impo an d i e s o in e na ionalisa ion as all coe icien s o he O- a iables a e signi ican ly posi i e (and, as expec ed, nega i e in case o obs _expe ience). Some coe icien s mono onically inc ease ( &d, e ia y_academic), o he s inc ease pa ly (inno_sales, app op iabili y, i m_age), and he coe icien s o he O- ela ed obs acle obs _expe ience mono onically dec ease. The es ima es hus s ongly suppo he O-pa o he model. The esul s o he a iables ep esen ing L-disad an ages o o eign loca ions con i m he hypo hesis ha hey a e ele an only o he shi om EXPORT o FDI; none o he coe icien s o DOMESTIC is signi ican . In con as , we ind s a is ically signi ican e ec s o he ansi ion om EXPORT o FDI o p ac ically all L- ela ed obs acles. Fou o hem show he pos ula ed nega i e sign, meaning ha hey de e a i m om being di ec ly ac i e a o eign loca ions: (la ge) geog aphic dis ance (obs _dis ance), high cos s o in e na ionalisa ion (obs _cos ), obliga ion o p oduce locally a subs an ial pa o he alue added (obs _local_ alue), and, inally, o ced echnology ans e (obs _ ech_ ans e ). We do no ind a signi ican e ec o poli ical ins abili y in po en ial hos coun ies (obs _ins abili y). Con a y o ou expec a ions, we ob ain o h ee L- ela ed obs acle a iables a posi i e sign: cul u al di e ence (obs _cul u e), es ic i e egula o y en i onmen (obs _ egula ion) and insu icien p o ec ion o IPRs in hos coun ies (obs _IPR_p o ec ion). In he case o hese a iables, we a gue (see Subsec ion 4.2) ha a posi i e sign e lec s he ac ha i ms pe cei e (o ge awa e o ) he speci ic hind ance no un il hey ha e al eady engaged in di ec o eign ac i i ies (FDI). Cul u al di e ences ha make in e na ional ansac ions mo e di icul may mani es hemsel es only as he company has become di ec ly p esen ab oad; o example, FDI may acili a e he adap a ion o p oduc s o local as es. Mo eo e , i o en is di icul o en o ce IPRs wi hou knowledge o local condi ions and wi hou (pe sonal) ela ionships o local ac o s; hence, i IPRs a e impo an o pene a ing o eign ma ke s, a di ec p esence ab oad is supe io o expo ing. Conside a ions along simila lines apply in he case o egula o y obs acles, as knowledge o local condi ions usually is necessa y o op imally adap o egula o y www.economics-ejou nal.o g 23 Table 4: Resul s o Model I: INT_p opensi ya Business sec o Manu ac u ing Se ices Explana o y a iable DOMESTIC FDI DOMESTIC FDI DOMESTIC FDI O-ad an ages &d –.708*** .823*** –.707*** .883*** –.508 .868** (.19) (.18) (.26) (.22) (.35) (.39) inno_sales .050 .177*** .022 .182** .057 .167 (.06) (.06) (.08) (.08) (.08) (.11) e ia y_academic –.257*** .187*** –.286** .125 –.274*** .195* (.06) (.07) (.11) (.09) (.08) (.12) app op iabili y –.276*** .015 –.197* –.004 –.248* .072 (.08) (.09) (.12) (.11) (.13) (.17) i m_age .074 .350*** .139 .445*** –.043 .201 (.09) (.10) (.14) (.13) (.13) (.17) O- ela ed obs acles obs _expe ience .449** –.473** .446 –.643** .740** .044 (.20) (.22) (.28) (.29) (.31) (.43) L-(dis)ad an ages L- ela ed obs acles obs _dis ance .044 –.422** –.252 –.351 .525* –.501 (.19) (.22) (.28) (.27) (.31) (.46) obs _cul u e .138 .842*** .309 1.26*** –.176 .175 (.24) (.24) (.37) (.32) (.34) (.43) obs _cos –.279 –.565*** –.077 –.445* –.858*** –.846** (.18) (.19) (.24) (.24) (.28) (.38) obs _ egula ion .204 .684*** .087 .806*** .221 .563 (.22) (.22) (.37) (.29) (.32) (.41) obs _ins abili y –.160 .345 –.393 .163 .009 .554 (.25) (.23) (.38) (.29) (.39) (.45) obs _local_ alue .024 –.506** .070 –.476 .043 –.473 (.27) (.26) (.38) (.33) (.41) (.56) obs _ ech_ ans e –.180 –.575** –.120 –.756** .021 –.447 (.37) (.29) (.50) (.35) (.61) (.75) obs _IPR_p o ec ion .248 .510* .176 .925*** –.247 –1.04 (.34) (.28) (.46) (.33) (.51) (.73) Table 4 con inued www.economics-ejou nal.o g 24 Table 4 con inued Business sec o Manu ac u ing Se ices Explana o y a iable DOMESTIC FDI DOMESTIC FDI DOMESTIC FDI I-ad an ages, i m size size 2E-05 6E-04*** -6E-04 1E-03** 6E-05 3E-04 (2E-04) (2E-04) (9E-04) (5E-04) (3E-04)) (3E-04) size2 –1E-07 –9E-06** –7E-05 –1E-04* 2E-06 –1E-06 (4E-06) (4E-06) (2E-04)) (5E-05)) (3E-05) (3E-05)) coope a ion –.029 .262* .162 .448** –.315 –.061 (.15) (.15) (.23) (.19) (.21) (.28) Ma ke en i onmen compe i ion –.294 1.25* –.584 .989 –.020 2.15** (.61) (.70) (1.1) (1.2) (.74) (.92) demand_ end –2.48*** 1.20 –3.81*** 1.43 –.862 3.08** (.65) (.76) (1.0) (1.1) (.91) (1.3) Con ol a iables o eign –.117 –2.08*** –.192 –2.18*** –.118 –2.06*** (.20) (.28) (.32) (.37) (.29) (.48) egional_ma ke 1.26*** –.979*** 1.53*** –.704*** 1.05*** –1.18*** (.14) (.18) (.21) (.25) (.23) (.30) indus y Yes Yes Yes S a is ics N 1’921 1’000 757 Wald χ2 3851.6*** 1072.8*** 1786*** Pseudo R2 .338 .332 .246 a Fi ms wi h “domes ic sales only” (DOMESTIC) and hose wi h “di ec o eign ac i i y” (FDI), e alua ed agains he e e ence g oup o i ms wi h “expo and domes ic sales only” (EXPORT); mul inomial logi es ima es. The mul inomial logi model es ima es o each explana o y a iable wo slope pa ame e s wha allows o e alua e whe he he esponses DOMESTIC and FDI espec i ely signi ican ly di e om he e e ence le el EXPORT. The es ima es o he in e cep s and he indus y dummies ha e been h oughou omi ed. The signi icance o he pa ame e s is indica ed wi h ***, ** and * esp. ep esen ing he 1%, 5% and 10%-le el wi h obus s anda d e o s in b acke s. The cons uc ion/ene gy sec o (N=164) is included in he es ima es o he o al business sec o (N=1921) bu excluded in case o manu ac u ing (N=1000) and se ices (N=757). www.economics-ejou nal.o g 31 5.2.1 Manu ac u ing s. se ices (H4) In manu ac u ing (Table 5, Column 2), we ind di e ences be ween he wo s a egies o in e na ionalisa ion FDI_RDP and FDI_DP o eigh ou o eigh een a iables. The pa e n o hese di e gences is p ac ically he same as in he en i e business sec o . In manu ac u ing, he ele ance o O-ad an ages seems o be somewha s onge as we ind signi ican e ec s o some u he O- ela ed ac o s, i.e. he “sha e o employees wi h e ia y-le el educa ion” (“ e ia y_academic”) and “expe ience wi h o eign ac i i ies” (“obs _expe ience”). These addi ional e ec s a e compa ible wi h he speci ic cha ac e and he equi emen s o o eign R&D in es men s. These indings, in combina ion wi h he esul s o Model I (see Table 4), clea ly indica e (simila o he esul s o he en i e business sec o ) a g adual and con inuous p ocess o in e na ionalisa ion o manu ac u ing i ms as i is pos ula ed by he s ages iew o in e na ionalisa ion (DOMESTIC<EXPORT< FDI_DP<FDI_RDP). The indings a e la gely consis en wi h hose o a s udy dealing wi h he manu ac u ing sec o o he I alian economy (Cas ellani and Zan ei 2007). The es ima es o he se ice sec o (Table 5, Columns 3 and 4) su e om mul icollinea i y p oblems. Due o he collinea i y be ween &d and e ia y_academic, he coe icien o he la e a iable becomes signi ican ly nega i e in he p esence o &d (Column 3). In es ima es wi hou &d, he a iable e ia y_academic is insigni ican (Column 4). Also because o mul icollinea i y, he coe icien o he a iable obs _expe ience is signi ican ly posi i e in Column 3 and becomes insigni ican in Column 4. We conside he es ima es in Column 4 as mo e alid in econome ic e ms. The e o e, ou commen s e e o hese esul s. A compa ison be ween Column 4 and Column 2 o he able shows ha he explana o y powe o he OLI model wi h espec o he choice be ween he wo ypes o FDI s a egy is lowe o se ices han o manu ac u ing (wha is also indica ed by he adjus ed R2). Mo eo e , he pa e n o explana ion subs an ially di e ges be ween he wo sec o s. In he case o se ices, we ind s a is ically signi ican coe icien s o FDI_RDP ( ep esen ing de ia ions om he e e ence s a egy FDI_DP) o six o he se en een a iables co e ing all pa s o he OLI model (wi h he excep ion o he ma ke en i onmen ), as agains eigh in he case www.economics-ejou nal.o g 32 o manu ac u ing. Howe e , he pa e n o explana ion in e ms o he signi ican a iables qui e s ongly di e s be ween he wo sec o s. In he se ice sec o , he coe icien s o h ee O- a iables (sha e o employees wi h e ia y-le el educa ion, app op iabili y, o eign expe ience) a e no longe signi ican ; a leas he esul s o he inno a ion a iable “sales sha e o inno a i e p oduc s” emain he same as o manu ac u ing. These di e gences p ima ily e lec some peculia i ies o inno a ion- ela ed cha ac e is ics o se ices such as he in ensi e use o non- echnological know-how which does no need o be s ongly p o ec ed om compe i ion ( aci knowledge). Among he L- (dis)ad an ages, geog aphical dis ance (obs _dis ance) and he obliga ion o p oduce locally (obs _local_ alue) exe a signi ican in luence on he choice o he wo s a egies only in se ices, whe eas – wha is plausible – he a iable “insu icien en o cemen o IPRs” (obs _IPR_p o ec ion) is ele an only in manu ac u ing (whe e R&D ac i i ies a e mo e p e alen han in se ices). The di e ences be ween he wo sec o s wi h espec o I-ad an ages a e p obably no so ele an as hey pe ain only o “coope a ion“, whe eas he i m size e ec is p ac ically iden ical. Finally, in acco dance wi h manu ac u ing, he wo s a egies o in e na ionalisa ion do no di e wi h espec o he e ec o he ma ke en i onmen (compe i ion, demand_ end). The “pe o mance” o he OLI model is weake o se ices no only wi h espec o Model II (choice be ween he s a egies FDI_RDP and FDI_DP) bu , as shown in Subsec ion 5.1, also o Model I (in pa icula wi h espec o he shi om EXPORT o FDI; see Table 4). In se ices, he mos ele an s ep o in e na ionalisa ion ha can be explained by he OLI app oach is he shi om “exclusi ely domes ic ac i i ies” o “ o eign ac i i ies in gene al” (expo s and/o FDI). The speci ic di e ences be ween expo ing and FDI (wi h o wi hou R&D), al hough hey exis , a e clea ly less p onounced han in manu ac u ing. The pa e n o in e na ionalisa ion o se ices i ms may hus be cha ac e ised by DOMESTIC<EXPORT≅FDI_DP<FDI_RDP. The p ocess o in e na ionalisa ion, hough g adual in bo h sec o s, seems o be less con inuous in se ices han in manu ac u ing whe e i is DOMESTIC<EXPORT<FDI_DP<FDI_RDP). To sum up, he indings (Model II) o he wo sec o s a e in line wi h hypo hesis H4. The explana o y powe o he OLI model we applied o explain he choice be ween wo FDI s a egies in e ms o speci ic combina ion o business unc ions ( he mo e a - eaching one including R&D, he o he one pe aining only www.economics-ejou nal.o g 33 o dis ibu ion and/o p oduc ion) is clea ly lowe in se ices han in manu ac u ing. Mo eo e , he di e ences be ween he wo sec o s wi h espec o he ela i e signi icance o he explana o y a iables o Model II a e qui e plausible. Mo eo e , i u ns ou ha he p ocess o in e na ionalisa ion, hough g adual in manu ac u ing as well as in se ices, is less con inuous in he se ice sec o . This inding, which is based on es ima es o Model I and II, is no su p ising in iew o p e ious (mos ly heo e ical o in e iew-based) esea ch, which shows ha o some se ice companies expo ing may be he bes op ion o s a ing in e na ionalisa ion whe eas o o he s a di ec p esence ab oad is op imal (see, e.g., Ca man and Langea d 1980; Boddewyn e al. 1986; Bell 1995; G ön oos, 1999). Hence, he e migh be no s a egy o in e na ionalisa ion o se ices ha is gene ally supe io , wha is consis en wi h ou esul s which imply ha X≅FDI. Econome ic e idence p o ided by Wagne (2014) also suppo s he iew o a less g adual and con inuous p ocess o in e na ionalisa ion in se ices han in manu ac u ing. Applying he model o i m he e ogenei y o he se ice sec o , he e en inds ha , in se ices, expo e s a e mo e p oduc i e han i ms wi h FDI (implying ha in e na ionalisa ion ollows he pa h D<FDI<X). 6 Summa y and conclusions Econome ic esea ch dealing wi h he in e na ionalisa ion o i ms s ill ocuses on manu ac u ing. As he cha ac e is ics o se ices and goods di e in se e al espec s, one may expec ha he de e minan s o he i ms’ in e na ional ac i i ies a e no he same o he wo sec o s. Howe e , he e is ha dly any empi ical i m- le el e idence on whe he his p oposi ion is co ec . The e o e, we i s aimed a iden i ying econome ically he ac o s de e mining why i ms in e na ionalise hei ac i i ies and why hey choose a speci ic s a egy o in e na ionalisa ion (expo ing s. FDI). We hen asked (wha is a he co e o ou in e es ) whe he he de e minan s o hese wo ways o going in e na ional di e be ween manu ac u ing and se ices. Second, we analysed (possible) di e ences be ween speci ic o ms o FDI in e ms o combina ions o business unc ions, again con as ing he wo sec o s. We used he well-known OLI model as heo e ical www.economics-ejou nal.o g 34 backg ound o he empi ical analysis, which is based on a la ge i m-le el da ase co e ing he Swiss business sec o . We speci ied wo models which we es ima ed o he en i e business sec o and, sepa a ely, o manu ac u ing and se ices. The i s one (“INT_p opensi y”) de e mines he p obabili y o a company o belonging o one o he ollowing ca ego ies o i ms: a) “selling on domes ic ma ke s only”, b) “se ing, addi ionally, expo ma ke s” and, c) “o sho ing”. The second model (“INT_ unc ion”) ocuses on in e na ionalised companies and de e mines he p obabili y o a i m o go in e na ional by means o a) “se ing expo ma ke s”, b) “o sho ing o dis ibu ion and/o p oduc ion”, and c) “o sho ing o R&D, in addi ion o dis ibu ion and/o p oduc ion”. The es ima es o he wo models a e la gely in line wi h ou hypo heses. Fi s ly, we ind ha he OLI model is well sui ed o explaining why i ms in e na ionalise hei ac i i ies as expo e s o , al e na i ely, by means o o sho ing (p opensi y o in e na ionalise). This holds ue in pa icula o he business sec o as a whole and o manu ac u ing, whe eas, as hypo hesised, he explana o y powe o he OLI model is lowe in he case o se ices. Secondly, he OLI model also allows o explaining di e ences be ween wo speci ic o ms o o sho ing in e ms o (combina ions o ) business unc ions, i.e. “o sho ing o dis ibu ion and/o p oduc ion” s. “o sho ing o R&D, in addi ion o dis ibu ion and/o p oduc ion”. This p ima ily holds ue o manu ac u ing bu ends o be he case also o he se ice sec o . Thi d, he es ima es con as ing he manu ac u ing and he se ice sec o yield di e ences wi h espec o he ela i e impo ance o he indi idual explana o y a iables ha (mos ly) a e economically plausible; his is ue o he model explaining he choice be ween expo ing and FDI as well as he choice o a speci ic FDI s a egy in e ms o business unc ions. Fou h, he es ima es o bo h models seem o be consis en wi h he s ages iew o in e na ionalisa ion, in pa icula in he manu ac u ing sec o bu , o a lesse ex en , also in he se ice sec o whe e he p ocess o in e na ionalisa ion, howe e , is less con inuous han in manu ac u ing. Conside ing he scope and indings o he analysis and gi en he lack o s udies dealing wi h he opic o his pape , we subs an ially add o exis ing knowledge ega ding he de e minan s o he in e na ionalisa ion o i m ac i i y. In pa icula , he pape yields new insigh s by sys ema ically con as ing es ima es o manu ac u ing and se ices i ms, as well as by in es iga ing he d i e s o www.economics-ejou nal.o g 35 di e en o ms o in e na ional ac i i ies in e ms o (combina ions o ) business unc ions. The analysis shows ha manu ac u ing and se ices companies subs an ially di e wi h espec o he ac o s de e mining he p opensi y o in e na ionalisa ion as a whole as well as di e en ia ed by business unc ion. Al hough he explana o y powe o he OLI model is lowe o se ices han o manu ac u ing i emains an adequa e amewo k o analysis o bo h sec o s. Fu he mo e, i u ns ou ha he d i e s o in e na ionalisa ion p ima ily e lec i m-speci ic beha iou whe eas indus y-speci ic ac o s (which a e emphasised in p e ious wo k) a e o mino impo ance. The s udy has a numbe o limi a ions. A i s one e e s o he speci ica ion o he OLI model. The measu emen o I-ad an ages (which is di icul anyway) is no e y sa is ac o y, p ima ily due o de iciencies o he da abase. As a consequence, he empi ical indings, in he i s place, con i m he OL-pa o he OLI model (al hough, in gene al, he I- a iables also yielded he expec ed esul s). Fu he esea ch could also help o imp o e o he elemen s o he OLI model. Fo example, i would be bene icial o ex end he O-pa , o example by including inancial a iables. Secondly, due o he c oss-sec ion na u e o he da a, he indings ha e o be in e p e ed as condi ional co ela ions a he han as causal ela ionships; ne e heless, i is s ill possible o assess whe he he empi ical esul s a e consis en wi h he pos ula ed hypo heses. As a consequence, econome ic s udies making use o longi udinal da a would be highly welcome. Such wo k would be mo e adequa e o analyse he dynamics o in e na ionalisa ion (e.g., he s ages iew o in e na ionalisa ion). www.economics-ejou nal.o g 36 Re e ences A ani is, S., and H. Hollens ein (2011). How do Di e en D i e s o R&D In es men in Fo eign Loca ions A ec Domes ic Fi m Pe o mance? An Analysis Based on Swiss Panel Mic o Da a. Indus ial and Co po a e Change 20(2): 605–640. h p://icc.ox o djou nals.o g/con en /20/2/605. ull.pd A ani is, S., H. Hollens ein, and T. S ucki (2015). The Cha ac e is ics o MNEs Loca ed in Di e en Hos Regions. Swiss Jou nal o Economics and S a is ics 151(4): 261– 298. h p://www.sjes.ch/published.php?Pape N =2015-IV-1&Yea =2015 Basile, R., A. Giun a, and J.B. Nugen (2003). Fo eign Expansion by I alian Manu ac u ing Fi ms in he Nine ies: An O de ed P obi Analysis. Re iew o Indus ial O ganiza ion 23(1): 1–24. h p://link.sp inge .com/a icle/10.1023/B%3AREIO.0000005561.07176.30 Bell, J. (1995). The In e na ionalisa ion o Small Compu e So wa e Fi ms: A Fu he Challenge o “S age” Theo ies. Eu opean Jou nal o Ma ke ing 29(8): 60–75. h p://www.eme aldinsigh .com/doi/pd plus/10.1108/03090569510097556 Boddewyn, J.J., M. Halb ich, and A.C. Pe y (1986). Se ice Mul ina ionals: Concep ualiza ion, Measu emen and Theo y. Jou nal o In e na ional Business S udies 17(3): 41–57. h p://www.js o .o g/s able/pd /154932.pd ?accep TC= ue B ou he s, K.D., and L.E. B ou he s (2003). Why Se ice and Manu ac u ing En y Mode Choices Di e : The In luence o T ansac ion Cos Fac o s, Risk and T us . Jou nal o Managemen S udies 40(5): 1179–1204. h p://onlinelib a y.wiley.com/doi/10.1111/1467-6486.00376/epd B own, J.R., C.S. De , and Z. Zhou (2003). B oadening he Fo eign Ma ke En y Mode Decision: Sepa a ing Owne ship and Con ol. Jou nal o In e na ional Business S udies 34(4): 393–413. h p://www.js o .o g/s able/3557163 Buckley, P.J., and M.C. Casson (1985). The Economic Theo y o he Mul ina ional En e p ise. McMillan: London. Buckley, P.J., Pass, C.L., and K. P esco (1992): The In e na ionalisa ion o Se ices Fi ms: A Compa ison wi h he Manu ac u ing Sec o . Scandina ian In e na ional Business Re iew 1(1): 39–56. h p://www.sciencedi ec .com/science/a icle/pii/0962926292900355 Ca man, J.M., and E. Langea d (1980). G ow h S a egies o Se ice Fi ms. S a egic Managemen Jou nal 1(1): 7–22. h p://onlinelib a y.wiley.com/doi/10.1002/smj.4250010103/epd www.economics-ejou nal.o g 37 Cas ellani, D., and G. Gio anne i (2010). P oduc i i y and he In e na ional Fi m: Dissec ing He e ogenei y. Jou nal o In e na ional Policy Re o m 13(1): 25–42. h p://www. and online.com/doi/pd /10.1080/17487870903546226 Cas ellani, D., and A. Zan ei (2007). In e na ionalisa ion, Inno a ion and P oduc i i y: How Do Fi ms Di e in I aly. The Wo ld Economy 30(1): 156–176. h p://onlinelib a y.wiley.com/doi/10.1111/j.1467-9701.2007.00875.x/epd Ca es, R.E. (1982). Mul ina ional En e p ise and Economic Analysis. Camb idge Uni e si y P ess: Camb idge (Mass.). Cla k, T., and G. Mallo y (1997). The Impac o he S a egic Choice on he In e na ionalisa ion o he Fi m. In G. Ch yssochoidis, C. Milla , and J. Clegg (Eds.). In e na ionalisa ion S a egies. Palg a e Macmillan: London and New Yo k. Con ac o , F.J., and S.K. Kundu (1998). Modal Choice in a Wo ld o Alliances: Analyzing O ganiza ional Fo ms in he In e na ional Ho el Sec o . Jou nal o In e na ional Business S udies 29(2): 325–357. h p://www.js o .o g/s able/pd /155606.pd ?accep TC= ue Con ac o , F.J., Kundu, S.K., and C.-C. Hsu (2003). A Th ee S age Theo y o In e na ional Expansion: The Link be ween Mul ina ionali y and Pe o mance in he Se ice Sec o . Jou nal o In e na ional Business S udies 34(1): 5–18. h p://www.js o .o g/s able/3557136 Dunning, J.H. (1977). T ade, Loca ion o Economic Ac i i y and he Mul ina ional En e p ise: A Sea ch o an Eclec ic App oach. In B. Ohlin, P.O. Hesselbo n, and P.J. Wiskman (Eds.). The Alloca ion o In e na ional Ac i i y. London: McMillan. Dunning, J.H. (1979). Explaining Changing Pa e ns o In e na ional P oduc ion: In De ence o he Eclec ic Theo y. Ox o d Bulle in o Economics and S a is ics 41(4): 269–295. h p://onlinelib a y.wiley.com/doi/10.1111/j.1468-0084.1979.mp41004003.x/pd Dunning, J.H. (1989). Mul ina ional En e p ises and he G ow h o Se ices: Concep ual and Theo e ical Issues. Se ice Indus ies Jou nal 9(1): 5–39. h p://www. and online.com/doi/pd /10.1080/02642068900000002 Dunning, J.H. (2000). The Eclec ic Pa adigm as an En elope o Economic and Business Theo ies o MNE Ac i i y. In e na ional Business Re iew 9(2): 163–190. h p://www.sciencedi ec .com/science/a icle/pii/S0969593199000359 Dunning, J.H., and S.M. Lundan (2008). Mul ina ional En e p ises and he Global Economy, Second Edi ion. Edwa d Elga : Chel enham. www.economics-ejou nal.o g 38 Ekeledo, I., and K. Si akuma (1998). Fo eign Ma ke En y Mode Choice o Se ice Fi ms: a Con ingency Pe spec i e. Jou nal o he Academy o Ma ke ing Science 26(4): 274–292. h p://link.sp inge .com/a icle/10.1177%2F0092070398264002 E amilli, M.K. (1990). En y Mode Choice in Se ice Indus ies. In e na ional Ma ke ing Re iew 7(5): 50–62. h p://www.eme aldinsigh .com/doi/pd plus/10.1108/EUM0000000001535 E amilli, M.K., and C.P. Rao (1993). Se ice Fi ms’ In e na ional En y Mode Choice: A Modi ied T ansac ion-Cos Analysis App oach. Jou nal o Ma ke ing 57: 19–38. h p://web.b.ebscohos .com/ehos /pd iewe /pd iewe ?sid=297b3b85-ad67-45db- 8 8-2706d2739503% 40sessionmg 114& id=12&hid=123 Falk, M., and E. Hags en (2015). Expo Beha iou o Mic o Fi ms in he Swedish Compu e and Business Se ices Indus ies. Economics: The Open Access, Open Assessmen E-Jou nal 9(2015-32): 1–24. h p://www.economics-ejou nal.o g/economics/jou nala icles/2015-32 Galia, F., and D. Leg os (2004). Complemen a i ies be ween Obs acles o Inno a ion: E idence om F ance. Resea ch Policy 33: 1185–1199. h p://www.sciencedi ec .com/science/a icle/pii/S0048733304000952 Gassmann, O., and M. on Zed wi z (1999). New Concep s and T ends in In e na ional R&D O ganiza ion. Resea ch Policy 28(2–3): 231–250. h p://www.sciencedi ec .com/science/a icle/pii/S0048733398001140 G eenaway, D., and R. Knelle (2007). Fi m He e ogenei y, Expo ing and Fo eign Di ec In es men . Economic Jou nal 117(Feb .): F134–161. h p://onlinelib a y.wiley.com/doi/10.1111/j.1468-0297.2007.02018.x/epd G ön oos, C. (1999). In e na ionalisa ion S a egies o Se ices. Jou nal o Se ices Ma ke ing 13(4/5): 290–297. h p://www.eme aldinsigh .com/doi/pd plus/10.1108/08876049910282547 Gue ie i, P., and V. Meliciani (2005). Technology and In e na ional Compe i i eness: The In e dependence Be ween Manu ac u ing and P oduce Se ices. S uc u al Change and Economic Dynamics 16: 489–502. h p://www.sciencedi ec .com/science/a icle/pii/S0954349X05000202 Helpman, E. (1984). A Simple Theo y o In e na ional T ade wi h Mul ina ional Co po a ions. Jou nal o Poli ical Economy 92(3): 451–471. h p://www.js o .o g/s able/1837227 www.economics-ejou nal.o g 39 Helpman, E., M. Meli z, and S. Yeaple (2004). Expo e sus FDI wi h He e ogenous Fi ms. Ame ican Economic Re iew 94(1): 300–316. h p://pubs.aeaweb.o g/doi/pd plus/10.1257/000282804322970814 Henna , J.-F. (1982). A Theo y o Mul ina ional En ep ise. Uni e si y o Michigan P ess: Ann A bo . Hollens ein, H. (2005). De e minan s o In e na ional Ac i i ies: A e SME’s Di e en ? Small Business Economics 24(5): 431–450. h p://link.sp inge .com/a icle/10.1007/s11187-005-6455-x Hollens ein, H., and M. Be ge (2015). Choice o Fo eign R&D En y Mode and I s Rela ion o Fi m Pe o mance: A Fi m-le el Analysis o Swi ze land and Aus ia. Economics: The Open-Access, Open-Assessmen E-Jou nal 9 (2015-34): 1-46. h p://dx.doi.o g/10.5018/economics-ejou nal.ja.2015-34 Hyme , S. (1976). The In e na ional Ope a ions o Na ional Fi ms: A S udy o Di ec Fo eign In es men . MIT P ess: Camb idge (MA). Jack, R., As-Sabe , S., Edwa ds, R., and P. Buckley (2008). The Role o Se ice Embeddedness in he In e na ionalisa ion P ocess o Manu ac u ing Fi ms. In e na ional Business Re iew 17: 442–451. h p://www.sciencedi ec .com/science/a icle/pii/S0969593108000140 Jaklic, A., Ci jako ic, J., and A. Chidlow (2012). Explo ing he E ec s o In e na ional Sou cing on Manu ac u ing e sus Se ice Fi ms. Se ice Indus ies Jou nal 32(7): 1193–1207. h p://www. and online.com/doi/pd /10.1080/02642069.2012.662496 Johanson, J., and J.-E. Vahlne (1977). The In e na ionalisa ion P ocess o he Fi m. A Model o Knowledge De elopmen and Inc easing Fo eign Ma ke Commi men s. Jou nal o In e na ional Business S udies 8(1): 22–32. h p://www.js o .o g/s able/pd /254397 Knigh , G. (1999). In e na ional Se ices Ma ke ing: Re iew o Resea ch, 1980-1998. Jou nal o Se ices Ma ke ing 13(4/5): 347–360. h p://www.eme aldinsigh .com/doi/pd plus/10.1108/08876049910282619 Kogu , B., and U. Zande (1993). Knowledge o he Fi m and he E olu iona y Theo y o he Mul ina ional Co po a ion. Jou nal o In e na ional Business S udies 24(4): 625– 645. h p://www.js o .o g/s able/pd /155168.pd Kundu, S.K., Kuma , V., and S.D. Pe e s (2008). Impac o Owne ship and Loca ion Fac o s on Se ice Mul ina ionals` In e nalisa ion. Se ice Indus ies Jou nal 28(5): 567–580. h p://www. and online.com/doi/pd /10.1080/02642060801988068 www.economics-ejou nal.o g 40 Kundu, S.K., und H. Me chan (2008). Se ice Mul ina ionals: Thei Pas , P esen , and Fu u e. Managemen In e na ional Re iew 48(4): 371–377. h p://link.sp inge .com/a icle/10.1007/s11575-008-0021-0#/page-1 Lejp as, A. (2009). De e minan s o In e na ionalisa ion: Di e ences be ween Se ice and Manu ac u ing SMEs. DIW Discussion Pape s No. 886. Ge man Ins i u e o Economic Resea ch (DIW): Be lin. h p://www.diw.de/documen s/publika ionen/73/diw_01.c.97798.de/dp886.pd León-Da de , F., Villa -Ga cía, C., and J. Pla-Ba be (2010). En y Mode Choice in he In e na ionalisa ion o he Ho el Indus y: A Holis ic App oach. Se ice Indus ies Jou nal 31(1): 107–122. h p://www. and online.com/doi/pd /10.1080/02642069.2010.485198 Li, J., and S. Guisinge (1992). The Globaliza ion o Se ice Mul ina ionals in he “T iad” Regions: Japan, Wes e n Eu ope and No h Ame ica. Jou nal o In e na ional Business S udies 23(4): 675–696. h p://www.js o .o g/s able/pd /155154.pd ?accep TC= ue Lin, F-J. (2010). The De e minan s o Fo eign Di ec In es men in China: The Case o Taiwanese Fi ms in he IT indus y. Jou nal o Business Resea ch 63(5): 479–485. h p://www.sciencedi ec .com/science/a icle/pii/S0148296309001088 Ma iscal, A.S., Zhang, Y., and J.L. Pasqual (2012). In e na ionalisa ion o Mul ina ional Banks: A S udy o Fo eign Di ec In es men in Se en La in Ame ican Coun ies. Se ice Indus ies Jou nal 32(7): 1149–1170. h p://www. and online.com/doi/abs/10.1080/02642069.2012.662494 Meli z, M. (2003). The Impac o T ade on In a-indus y Realloca ion and Agg ega e Indus y P oduc i i y. Econome ica 71(6): 1695–1725. h p://onlinelib a y.wiley.com/doi/10.1111/1468-0262.00467/epd Miozzo, M., and I. Miles (2002). The Rela ion Be ween he In e na ionalisa ion o Se ices and he P ocess o Inno a ion: a Resea ch Agenda, in: M. Miozzo and I. Miles (Eds.), In e na ionalisa ion, Technology and Se ices. Edwa d Elga : Chel enham (UK). Mu ay, Y.K., and M. Ko abe (1999). Sou cing S a egies o U.S. Se ice Companies: A Modi ied T ansac ion –Cos Analysis. S a egic Managemen Jou nal 20: 791–809. h p://onlinelib a y.wiley.com/doi/10.1002/%28SICI%291097- 0266%28199909%2920:9%3C791:: AID-SMJ49%3E3.0.CO;2-U/epd Nachum, L., and C. Wymbs (2005). P oduc Di e en ia ion, Ex e nal Economies and MNE Loca ion Choices: M&As in Global Ci ies. Jou nal o In e na ional Business S udies 36(4): 415–434. h p://www.js o .o g/s able/3875301 Please no e: You a e mos since ely encou aged o pa icipa e in he open assessmen o his a icle. You can do so by ei he ecommending he a icle o by pos ing you commen s. Please go o: h p://dx.doi.o g/10.5018/economics-ejou nal.ja.2016-12 The Edi o © Au ho (s) 2016. Licensed unde he C ea i e Commons License - A ibu ion 4.0 In e na ional (CC BY 4.0).