Foreign direct investments and participation in global value chains: New evidence from advanced manufacturing industries in Central and Eastern Europe

42023, olume 26, issue 4, pp. 4–16, DOI: 10.15240/ ul/001/2023-4-001 Economics Fo eign di ec in es men s and pa icipa ion in global alue chains: New e idence om ad anced manu ac u ing indus ies in Cen al and Eas e n Eu ope Ines Ke san-Skabic1, An ea Ba isic2 1 Ju aj Dob ila Uni e si y o Pula, Facul y o Economics and Tou ism “D . Mijo Mi ko ic”, Depa men o Na ional and In e na ional Economics, C oa ia, ORCID: 0000-0001-7905-368X, [email p o ec ed]; 2 Uni e si y o Zag eb, Facul y o Economics and Business, Depa men o Economic Theo y, C oa ia, ORCID: 0000-0002-1149-6860, [email p o ec ed]. Abs ac : Fo eign di ec in es men s (FDIs) and ade wi hin global alue chains (GVC) ha e been conside ed one o he mos impo an ehicles o economic de elopmen and compe i i eness, especially o coun ies in ansi ion. The coun ies o Cen al and Eas e n Eu ope a e compe ing o a ac no only high amoun s o FDI in lows bu hey also pu emphasis on he FDIs’ sec o al s uc u e. FDI in lows in o ad anced indus ies can ep esen he basis o echnological de elopmen ha will lead o he c ea ion o a g ea e numbe o high-skill jobs and a highe le el o inno a ion, he eby in luencing he educ ion o he economic gap in ela ion o de eloped coun ies. This s udy aims o explo e he impo ance o FDI inwa d s ock o GVC pa icipa ion in he ad anced manu ac u ing indus ies in ele en EU membe s a es ha ans e ed o ma ke economies in he 1990s and a ac ed as amoun s o FDI in lows since hen. The dynamic panel da a analysis esul s indica e he impo ance o FDI inwa d s ock in he manu ac u ing indus y o he GVC pa icipa ion in ad anced manu ac u ing indus ies ac oss his se o coun ies while also s essing he impo an implica ions o manu ac u ing sha e in GDP changes o hese indus ies’ backwa d GVC pa icipa ion. The indings con i m a posi i e associa ion be ween a ac ing FDI in he manu ac u ing sec o and GVC in eg a ion in ad anced manu ac u ing indus ies. Such FDI a ge ing can be a ehicle o in e na ionaliza ion and de elopmen o high- ech and knowledge-based indus ies. Keywo ds: Fo eign di ec in es men , global alue chains, Cen al and Eas e n Eu ope, ad anced indus ies, manu ac u ing. JEL Classi ica ion: F12, F14, F21. APA S yle Ci a ion: Ke san-Skabic, I. & Ba isic, A. (2023). Fo eign di ec in es men s and pa icipa ion in global alue chains: New e idence om ad anced manu ac u ing indus ies in Cen al and Eas e n Eu ope. E&M Economics and Managemen , 26(4), 4–16. h ps://doi.o g/ 10.15240/ ul/001/2023-4-001 In oduc ion Fo eign di ec in es men s (FDIs) and global alue chains (GVCs) ha e been key ea u es o economic libe aliza ion and globaliza ion. A e he in oduc ion o he ma ke economy and hei libe aliza ion in he e y beginning o he 1990s, he Cen al and Eas Eu opean (CEE) coun ies (EU new membe s a es, EU CEE coun ies) needed esh capi al o p i a ize he majo i y o hei s a e-owned 5 2023, olume 26, issue 4, pp. 4–16, DOI: 10.15240/ ul/001/2023-4-001 Economics companies and o es ablish new ones wi h he main aim o boos hei economic g ow h. The CEE coun ies ha e a ac ed a cumula- i e amoun o EUR 842,964.4 million FDI in- wa d s ock om he beginning o he 1990s un il 2021, a ying om 5,268 EUR pe capi a in Romania o 22,321 EUR pe capi a in Es- onia (10,545 EUR pe capi a on a e age), ep esen ing om 35% o GDP in Slo enia o 94.5% o GDP in Es onia (58.4% o GDP on a e age) (WIIW, 2022). In hese coun ies, FDI in lows we e p omo - ed as a ehicle o echnological de elopmen wi h he po en ial o spu he de elopmen o knowledge-based indus ies and u he lead o hei specializa ion in highe - alue-added indus ies, enabling hei con e gence o de- eloped economies (Ju cic & Ba isic, 2018). Capi al accumula ion and echnological p og- ess spillo e e ec s de i ing om FDI in lows ha e been emphasized as impo an con ibu- o s o economic g ow h (Bo ensz ein, 1998; Iamsi a oj, 2016; Wan, 2010). Howe e , se e al pieces o esea ch poin ou di e se economic g ow h and expo ou comes o FDI in lows ac oss coun ies. They show ha posi i e e ec s ha e no ma e ialized h ough FDI channels e enly, which is o en a ibu ed o he impo - ance o abso p i e capaci y, unde lying mo i es o FDI and he indus ies ecei ing i (Alm aji & Almsa i , 2013; Damijan e al., 2013b; Ke san- Skabic & Zubin 2009). The impac o FDI hus con inues o be he ocus o academic deba es, especially ac oss hese economies (Bilas, 2020; Gaspa eniene e al., 2020). A he same ime, GVC pa icipa ion pa hs in hese coun ies ha e shown o be di e se, especially in manu ac u ing indus ies which a e conside ed an impo an sou ce o econom- ic de elopmen (Ba isic, 2020; Ke san-Skabic, 2017). Al hough in de eloped economies he impo ance o se ices has aised, i is ha d o ind a success ul de elopmen s o y o a coun y wi hou i being d i en by he indus- ializa ion (Szi mai, 2013). Economic libe aliza ion and de egula ion allowed his se o coun ies o join GVCs and ecei e FDI in lows, which a e o en ega ded o be impo an de e minan s o GVC ade as he o eign subsidia ies a e o en mo e con- nec ed o hei pa en company (MNE head- qua e ) and o en use es ablished downs eam ne wo ks. Al hough some esea ches poin ou o he e e se causa ion (GVC pa icipa ion leading o FDI) (Ca il-Caccia & Pa lo a 2020; Cipollina e al., 2021; Ma inez-Galan & Fon- ou a, 2018), FDI is mo e o en seen as one o he mos impo an de e minan s o es ab- lishing GVCs (Amado & Cab al, 2016) as i al- lows he shaping o hese p oduc ion ne wo ks o exploi he oppo uni ies in o eign ma ke s, and is expec ed o be ollowed by he in e na- ional p oduc ion agmen a ion and inc ease in GVC ade. Acco dingly, FDI in lows ha e o en been p o- claimed bene icial o os e ing GVC pa icipa ion in his se o coun ies, bu empi ical esea ch is e y limi ed. FDI inwa d s ock and high- ech p oduc sha e in expo s ha e p o en o be an impo an d i e o economy-wide GVC pa - icipa ion in EU NMS (13) (Ke san-Skabic, 2019), and a posi i e connec ion has been ound be ween bila e al FDI s ock and g oss bila e al ade and bila e al impo -con en o expo s in he case o CEE coun ies in he pe iod om 2000–2011 (Buelens & Ti pak, 2017). Specializa ion in ad anced indus ies in manu ac u ing (i.e., indus ies conside ed mo e echnology in ensi e), which o en en ail geo- g aphically a highly agmen ed p oduc ion p o- cess, is equen ly in he ocus o middle-income coun ies as echnological upg ading, and GVC in eg a ions a e o en conside ed acili a- o s o ca ch-up and a oiding he middle-income ap (Gill & Kha as, 2015). Al hough in eg a- ion in GVCs is belie ed o b ing an inc ease in p oduc i i y and os e he specializa ion in highe alue-added ac i i ies, which can eap he economies o scale, o he au ho s’ knowl- edge, no ecen esea ch in his se o coun ies has been conduc ed o explo e he e ec s o FDI in lows in he manu ac u ing indus y on ad anced manu ac u ing indus ies’ pa ici- pa ion in GVCs. Gi en he speci ic de elopmen pa h o hese coun ies since he 1990s, ac- companied by high le els o FDI in lows and in eg a ion in in e na ional p oduc ion ne wo ks, explo ing he de e minan s o GVC in eg a- ion in he indus ies conside ed impo an o hei de elopmen p o ides an in e es ing case o analysis. Thus, o ill his li e a u e gap and p o ide aluable con ibu ion o go e nmen policies ele an o a ac ing FDI, his pape aims o explo e he impo ance o FDI inwa d s ock in he manu ac u ing indus y o GVC pa ici- pa ion in ad anced indus ies in manu ac u ing in pos - ansi ion coun ies in he CEE egion 62023, olume 26, issue 4, pp. 4–16, DOI: 10.15240/ ul/001/2023-4-001 Economics (Bulga ia, C oa ia, Czechia, Es onia, Hunga y, La ia, Li huania, Poland, Romania, Slo a- kia and Slo enia). The ad anced indus ies in manu ac u ing included in he analysis we e chosen as hey a e conside ed i al o inc eas- ing compe i i eness o economies. To empi i- cally explo e hese ela ionships, dynamic panel da a models a e es ima ed using he sys em GMM es ima o . The analysis includes he da a o he pe iod om 2005 o 2018. The pape is s uc u ed as ollows. Sec ion 1 p esen s he heo e ical backg ound oge he wi h he li e a u e e iew o ela ed scien i ic pape s. Sec ion 2 explains he me hodology and da a used in he empi ical esea ch, while sec ion 3 p esen s he esul s o he empi ical esea ch and p o ides a discussion o he ind- ings. The las sec ion o e s he conclusion and policy ecommenda ions. 1. Theo e ical backg ound Mul ina ional companies’ g owing and dispe s- ing ac i i ies esul ed in inc eased ade in in- e media e p oduc s, indica ing he p oduc ion p ocess has been agmen ed in o many phases aking place in di e en places ac oss di e en coun ies. In his way, he p oduc ion p ocesses c ea e he global alue chains (GVCs). Coun- ies (i.e., p oduce s) can pa icipa e as ex- po e s o aw ma e ials and in e media e p oduc s ( o wa d linkages) and as impo e s o in e media e p oduc s om ab oad (back- wa d linkages). GVC pa icipa ion can esul om companies’ e o s and ne wo ks buil wi h business pa ne s (supplie s, cus ome s). I is also an indica o o specializa ion and inno a- ion, as i e e s o pa icipa ion in in e na ional p oduc ion chains wi h he ocus on he p oduc- ion o speci ic pa s o p oduc s (o se ices). Many ac o s can in luence he coun y’s in- ol emen in GVCs, bu o eme ging economies and/o ex-socialis economies, he easies way o inclusion in such chains is h ough he a ac- ion o FDIs. In his way, o eign companies in es as amoun s o money in b own ield o g een- ield p ojec s. E en hough he FDI in gene al can impac GVC pa icipa ion, he impac depends on he in es o s’ mo i a ion. This link is especial- ly e iden o e iciency seeking mo i a ed FDI. I he main mo i es lie in selling in he o eign ma ke s, he FDI will eplace he expo and no pa icula in luence on p oduc ion chain is likely o occu . On he o he hand, i he in es- o s seek esou ces, o cheape p oduc ion ac o s, hey will mo e a pa o he p oduc ion ab oad causing he c ea ion o e ical linkages be ween home companies and o eign ones. In ha case, in a-company ade eme ges and a big impac o FDI on GVC pa icipa ion can be expec ed. UNCTAD (2013) es ima ed ha abou 80% o global ade (in e ms o g oss expo s) is ela ed o in e na ional p oduc ion ne wo ks. In ecen empi ical esea ch, he connec- ion be ween FDI and GVC pa icipa ion has been explo ed in a ious ways. The analysis can be done by in es iga ing he connec ion be ween inwa d (o in lows o ) FDI and ade (especially ade in in e media es); ou wa d FDI and GVC (Li e al., 2021; Wang & Chen, 2020); GVC pa icipa ion and egional (p e e en ial ade) in eg a ion because in eg a ion can c ea e a a ou able en i onmen o boos ing in a- egional ade and in es men s as a con- sequence o which pa icipa ion in he global ( egional) alue chain can be expec ed as well (Pa k & Kim, 2020). Some pape s used he g a - i y app oach in explaining he links be ween ade and FDI lows as well as ade in in e media es and FDI lows (Buelens & Ti pak, 2017; Ca il- Caccia & Pa lo a, 2020). The li e a u e e iew including ele an pape s ha analyzed he impac o FDI on GVC pa icipa ion in he global con ex , ollow- ed by hose which ocused on he sample o EU coun ies, is p esen ed in addi ion. Amado and Cab al (2016) ound ha he ole o FDI is ins umen al in se ing up GVCs un by mul ina ional co po a ions. MNEs ha e a c u- cial ole in expanding in e na ional business and ade: hey accoun o one hi d o he global ou - pu and wo ld GDP, and a e esponsible o hal o he global expo s (OECD, 2018). They con ib- u ed o he g owing agmen a ion o p oduc ion seen wi hin GVCs in he pas decades (OECD, 2018). E ogo e al. (2021) applied he panel da a analysis on he sample o de eloping coun ies o he pe iod 2010–2019 and ound a posi i e impac o FDI in lows on GVC pa icipa ion ega dless o whe he he FDI is in he p ima y, seconda y o e ia y sec o . I is impo an o poin ou ha Cingolani e al. (2017) ound ha eme ging and de eloping coun ies mos ly posi ioned cen ally a ups eam and mids eam p oduc ion s ages, while high-income coun ies we e posi ioned a downs eam s ages. Speci ically, he Cen al and Eas e n Eu ope (CEE) coun ies ha e o en been sepa a ely esea ched due o hei accession 7 2023, olume 26, issue 4, pp. 4–16, DOI: 10.15240/ ul/001/2023-4-001 Economics p ocess o he EU which pushed conside able amoun s o FDI in lows om EU15 o hei Eas e n neighbou s ha s a ed joining he EU in 2004 ( ollowed by he enla gemen in 2007 which included Bulga ia and Romania and in 2013 which included C oa ia). Al hough hese coun ies ha e expe ienced p o ound changes since he beginning o 1990s, esea ch do no poin ou o clea e idence o hei ca ching-up beha iou owa ds he “old” EU membe s a es, bu s ess ha hei in eg a ion is a leng hy and non-linea p ocess (Cieslik & Wcislik, 2020). Addi ionally, i is impo an o no e hey ha e subs an ial dispa i ies in bo h he le els o FDI and deg ee o in eg a ion in GVCs ac oss sec o s and indus ies (Ba isic, 2020; Ke san Skabic, 2019). Also, hese coun ies ha e expe ienced dis inc pa hs o indus ializa ion/ deindus ializa ion. No ably, he Czech Re- public, Hunga y, Poland, Romania, Slo akia and Slo enia ha e main ained ela i ely high sha es o manu ac u ing indus ies in he o e - all GDP, while Bulga ia, C oa ia, Es onia and La ia exhibi lowe le els (WIIW, 2020) The ollowing esea ch include e idence ela ed o his opic om his egion. Damijan e al. (2013a) ound ha inwa d FDI in CEEC has changed and de ined hei expo composi- ion and posi i ely impac ed he de elopmen o p oduc ion ne wo ks in CEEC. Kaminsky and Ng (2005) ound ha FDI leads o in- c eased egional ade in eg a ion. Beha and F eund (2011) also indica ed ha in a-EU ade in in e media e goods has become mo e so- phis ica ed a e hei accession o he EU, poin ing o he inc easing ole o new membe s as supplie s o in e media e goods o he in- cumben coun ies. Amb oziak (2018) poin ed ou he g ea impo ance o Ge many o he in- eg a ion o new EU membe s a es (CEECs) in GVCs in he pe iod om 1995–2011. He ound ha Ge many became he main d i e o GVC pa icipa ion g ow h, whe e i s impo ance is highe in he backwa d linkages (as supplie s o in e media es) han in o wa d linkages. Nikolo o a (2013) esea ched he im- pac o FDI on he sec o al le el on he sample o EU27 coun ies and ound ha an inc ease in FDI is ela ed o highe p oduc ion and a highe demand o in e media e goods. This indica ed ha FDI is p ima ily e ical and ela - ed o ou sou cing ac i i ies. Hagemeje (2017) explo ed he sou ces o alue added and p o- duc i i y g ow h in he EU NMS in he pe iod 1995–2009. They ound an impo an con ibu- ion o ade (30–40%) o GDP g ow h, while connec ing he sec o al p oduc i i y g ow h wi h FDI, expo ing, and he posi ion o a sec o /coun y in GVCs, especially s essing he impo ance o impo ed in e media e goods o highe p oduc i i y ac oss sec o s. P oduc- i i y g ow h was also ound in sec o s dis an om he inal demand and hose expo ing in e media e goods. Se e al s udies explo ed he in eg a ion o his egion in GVCs. Lei ne and S eh e (2014) ound ha e ical specializa ion in ensi- ied in mos o he EU NMS and ha s onge pa icipa ion in global p oduc ion p ocesses enhanced hei pe o mance be ween 1995 and 2007. Thei esul s indica e ha expo g ow h and he deg ee o e ical specializa- ion end o ein o ce each o he . Cieslik e al. (2016) calcula ed pa icipa ion and posi ion o hese pos -socialis coun ies along GVCs in 2009 and ound mixed esul s. Coun ies wi h s onge links wi h Wes e n Eu opean coun ies we e mo e in eg a ed in GVCs, and mos o he expo e s om CEE a e posi ioned in he downs eam pa o he p oduc ion p ocess. In addi ion, hey ound he highes in e na ionaliza ion in anspo equipmen and elec onics whe e some o he coun ies we e among he global leade s in he ups eam segmen . G odzicki and Geodecki (2016) explained he co e-pe iphe y model in Eu ope based on he con ibu ion o g oups o coun ies o GVCs, while wa ning ha GVC pa icipa ion can accele a e he deindus ializa ion p ocess and ha CEE coun ies a e in a be e posi ion han he Sou he n Eu opean coun ies, due o hei con inued dependence on o eign capi- al and echnology. Besides economic de e mi- nan s o FDI in lows, some esea ch poin ed ou o impo ance o ins i u ional de e minan s including in es men policy design, e alua ion, and moni o ing ac oss Sou heas Eu ope o a ac ing FDI lows (Silajdzic & Mehic, 2022). Ada o and S eh e (2021) ocused hei esea ch on Eu opean coun ies co e ing he pe iod om 2000 o 2014 by in es iga ing impac o FDI in lows on he o ma ion o GVCs. Inwa d FDIs ha e shown o be e y in luen ial in he o ma ion o backwa d linkages and on he o he hand, ou wa d FDIs boos o wa d pa icipa ion in GVC. They also p o ided a sec- o al analysis, and hei indings a e s ongly ele an o high- ech manu ac u ing sec o s. 82023, olume 26, issue 4, pp. 4–16, DOI: 10.15240/ ul/001/2023-4-001 Economics Eu opean Commission (2020) also s essed he impo ance o syne gies be ween GVCs, FDI and na ional, egional and EU-le el de elopmen ajec o ies. They conduc ed an analysis o h ee GVC-sensi i e manu ac u ing indus ies (com- pu e and elec onics, au omo i e and ex iles and appa el) and g ouped EU membe s a es in o ou g oups acco ding o hei pa icipa ion le els: i) High GVC in eg a ion (Belgium, Czech Republic and Slo enia); ii) Low GVC in eg a ion (C oa ia, Cyp us, G eece, I aly, Po ugal, Spain and La ia); iii) Backwa d GVC in eg a ion (Li huania, Es onia, Slo ak Republic, Hunga y, Bulga ia, and I eland, Denma k, Luxembou g and Mal a); and i ) Fo wa d GVC in eg a ion (Ge many, Aus ia, Sweden, F ance, Finland, he Uni ed Kingdom, he Ne he lands, and Po- land and Romania). This epo demons a ed he impo ance o a ac i eness o o eign capi al and i s coo dina ion on he na ional and egional le els wi h o he policies. Mo eo e , economies lis ed in he Fo wa d GVC in eg a- ion g oup ecei ed and sen a o al alue o FDI signi ican ly la ge han he o he h ee g oups combined. Ca il-Caccia and Pa lo a (2020) p o ided an ex ensi e o e iew o pape s dealing wi h he connec ion be ween FDI and ade and especially be ween e ical FDI and ade. They poin ed ou he necessi y o deepen he exis ing analysis by including he s uc u e (de e minan s) o bo h. Chang e al. (2021) p e- sen ed an o e iew o 80 pape s ha connec in(ou ) lows o in(ou )wa d FDIs on GVC pa ici- pa ion, and emphasized he gaps no co e ed by exis ing esea ch. Following he indings o he li e a u e e iew, we s i e o go a s ep u he and empi ically explo e he connec ion be ween FDI in low in he manu ac u ing indus y and GVC pa icipa ion in ad anced manu ac u ing indus ies in he new EU membe s a es ex- cluding island economies – Cyp us and Mal a (EU NMS-11). Thus, his subse o coun ies in- cludes CEE coun ies which ha e expe ienced he ansi ion o ma ke economy in he 1990s and in he ollowing wo decades ha e joined he EU (CEE coun ies). 2. Resea ch me hodology The p ima y aim o his empi ical analysis is o explo e he impo ance o FDI inwa d s ock in he manu ac u ing indus y o GVC pa - icipa ion o ad anced manu ac u ing indus- ies ac oss EU CEE coun ies. The g oss expo s decomposi ion enabling he calcula ion o GVC pa icipa ion indica o s is explained in de ail in Koopman e al. (2010). Gi en i s me hodological cha ac e is ics, he mos e- quen ly used da abases o expo decomposi- ion in scien i ic esea ch ha e been he Wo ld Inpu -Ou pu da abase (WIOD) (Timme e al., 2015) and OECD T ade in Value Added da a- base (TiVA). Bo h enable he analysis o mod- e n ade pa e ns h ough GVCs, as hey p o ide da a on di ec and indi ec in e -sec- o al connec ions ac oss and wi hin coun ies (De Backe & Mi oudo , 2014). The mos common measu e o GVC pa - icipa ion is he sum o o wa d and backwa d GVC pa icipa ion, whe e backwa d GVC pa - icipa ion is he sha e o o eign alue added in g oss expo s, while o wa d GVC pa icipa ion is domes ic alue added which di ec ade pa - ne s u he eexpo , exp essed as he sha e o domes ic g oss expo s. I a coun y is down- s eam in he p oduc ion ne wo ks, i.e., ocused on he las s ages o p oduc ion, i is likely o ha e a la ge sha e o impo ed in e media e goods, which leads o high alues o backwa d pa icipa ion ela i e o o wa d, as i is he case wi h his se o coun ies (Ba isic, 2020; Ke san-Skabic, 2017). The opposi e would be ue o coun ies ocused on he ups eam GVCs ac i i ies. This empi ical analysis includes all h ee men ioned key a iables o GVC pa icipa- ion (backwa d pa icipa ion in GVCs, o wa d pa icipa ion in GVCs and o al pa icipa ion in GVCs) as dependen a iables. The da a o main dependen a iables include yea ly da a o 11 CEE coun ies (Bulga ia, Czech Repub- lic, C oa ia, Es onia, Hunga y, La ia, Li huania, Poland, Romania, Slo akia and Slo enia) in six ad anced indus ies: Manu ac u e o pha - maceu ical p oduc s and p epa a ions (C21), Manu ac u e o compu e , elec onic and op i- cal p oduc s (26), Manu ac u e o elec ical equipmen (C27), Manu ac u e o machine y and equipmen (C28), Manu ac u e o mo- o ehicles, aile s and semi- aile s (C29) and Manu ac u e o o he anspo equip- men (C30). Lis ed indus ies a e classi ied as high- echnology o medium-high- echnology acco ding o Eu os a (2022) high- ech classi i- ca ion and a e in gene al conside ed o be highly in eg a ed in GVCs, as p e iously elabo a ed. The FDI inwa d s ock in he manu ac u - ing indus y is he main independen a iable 9 2023, olume 26, issue 4, pp. 4–16, DOI: 10.15240/ ul/001/2023-4-001 Economics in he model. The s udy co e s he pe iod om 2005 o 2018, i.e., om he i s (and la ges ) EU accession wa e o hese coun ies un il he las a ailable TiVA da a. This analysis o- cuses on a pa icula ime ame d i en by he a ailabili y o da a ela ed o GVCs, selec ing he da abase ha p o ides he mos up- o-da e da a. As poin ed ou in he li e a u e e iew, ecen esea ch has used bo h s ock and lows FDI da a in explo ing his ela ionship. In his model, s ock da a a e chosen gi en hei mo e long- e m pe spec i e on he accumula ion o FDI in a coun y hus p o iding a be e unde - s anding o hei cumula i e impac on he hos economies’ GVC pa icipa ion. O he con ol a iables lis ed in Tab. 1 a e included in he model o explo e hei as- socia ion wi h he GVC pa icipa ion. Sha e o esea ch and de elopmen in GDP is used as a measu e o cen ali y o inno a ion ac oss economies, as R&D expendi u es olume has shown o be associa ed wi h highe manu ac- u ing indus y expo s, pa icula ly in he case o high- ech manu ac u ing (Hamma & Bela bi, 2021; Sandu & Ciocanel, 2014). As connec- ion be ween GVCs and indus ializa ion le els o coun ies has shown o be impo an (Baldwin & Okubo, 2019), he sha e o manu ac u ing indus y in GDP is also included as a mea- su e o coun ies’ indus ializa ion le el. Real e ec i e exchange a e is used o assess he cos compe i i eness o economies ela i e o 42 ading pa ne s. Also, ollowing Ada o and S eh e (2021) GDP pe capi a is used as a measu e o de elopmen le el and GDP as a measu e o he size o he economies. Thus, we explo e he ela ionship be ween he le el o economic de elopmen and he in e- g a ion o high- ech indus ies in GVCs, as well as ela ionship be ween he size o economies and le els o GVC pa icipa ion gi en ha smalle economies a e ypically cha ac e ized by g ea e openness. The lis and explana ion o a iables included in he analysis, oge he wi h da a sou ces used is p o ided in Tab. 1 and he co ela ion ma ix is a ailable upon eques . All lis ed a iables a e included in he model in he na u al loga i hm o m, excep he GVC_FL, gi en i s ela i ely small alues and e y small changes o e ime. Thus, GVC_FL was kep in i s o iginal o m (exp essed as he sha e o sec o ’s expo s). The numbe o obse a ions is no Va iable label Desc ip ion Sou ce GVC_BL Backwa d pa icipa ion in GVCs is he o eign alue added embodied in expo s, exp essed as he sha e o o al g oss expo s o he expo ing coun y – da a a he ad anced manu ac u ing indus y le el (%) OECD (2022) TiVA da abase GVC_FL Fo wa d pa icipa ion in GVCs is he “domes ic alue added embodied in o eign expo s, exp essed as he sha e o o al g oss expo s o he alue-added sou ce coun y” – da a a he ad anced manu ac u ing indus y le el (%) OECD (2022) TiVA da abase GVC_TOT To al pa icipa ion in GVCs (%) is calcula ed as he sum o backwa d (GVC_BL) and o wa d pa icipa ion (GVC_FL) in GVCs – da a a he ad anced manu ac u ing indus y le el (%) OECD (2022) TiVA da abase FDImanu FDI inwa d s ock in he manu ac u ing indus y (million EUR) WIIW (2022) RDinGDP G oss domes ic expendi u e on esea ch and de elopmen exp essed as he sha e o GDP (%) Eu os a (2022) Manu GDP Sha e o manu ac u ing indus y in GDP (%) WIIW REER Real e ec i e exchange a e ela i e o 42 ading pa ne s (index 2015 = 100) Eu os a (2022) GDPpc GDP pe capi a (EUR, chain linked olumes 2010) Eu os a (2022) GDP GDP (million EUR, cu en p ices) WIIW (2022) Sou ce: own Tab. 1: Desc ip ion o a iables 10 2023, olume 26, issue 4, pp. 4–16, DOI: 10.15240/ ul/001/2023-4-001 Economics he same o all a iables ac oss coun ies, making his da a se unbalanced. Desc ip i e s a is ics a e p o ided in Tab. 2. The c oss-sec ional dimension in his panel da a analysis is c ea ed by g ouping 6 ad anced manu ac u ing indus ies wi h 11 coun ies included in he model, hus c ea - ing 66 coun y-indus y g oups, while he ime dimension includes 14 yea s (2005–2018). The es ima ion based on he dynamic model, using he lagged dependen a iable, allows he dynamics o he unde lying p ocesses, which can be c ucial in ob aining consis en es ima es o he emaining pa ame e s (Bond, 2002). The endogenei y p oblem ha can appea in he s a ic analysis is o e come by applying he gen- e alized me hod o momen s (GMM) es ima- ion using ins umen al a iables. Conside ing he o he cha ac e is ics and ad an ages o he GMM es ima o s, sys em GMM es ima o is used in his esea ch (A ellano & Bo e , 1995; Blundell & Bond, 1998), using he x abond2 use w i en command (Roodman, 2009) in STATA 17. Gi en he he e oskedas ici y o he da a, he wo-s ep sys em GMM es ima o is employed in he analysis. The dynamic panel model is de ined in he ollowing o m: Yi = δYi –1 + λX ´ i + d + εi (1) whe e: Yi – he dependen a iable ec o , which is also included in he model in he lagged o m (Yi −1) as an independen a iable; X ´ i – a ma ix o main independen and con ol a iables; d – a ec o o yea dummy a iables c ea ed o explo e he ime speci ic e ec s; εi – he e o e m; δ and λ – unknown pa a- me e s which a e es ima ed h ough he model; i – he coun y-indus y g oups; – he yea . 3. Resul s and discussion 3.1 Resul s Th ee models including dependen a iables o o wa d, backwa d, and o al GVC pa icipa ion a e sepa a ely es ima ed. As a iables a e in- cluded in he model in he na u al loga i hm o m, he in e p e a ion o he esul s is exp essed in he o m o elas ici y, excep in he case o he dependen a iable in he model including o wa d GVC pa icipa ion, whe e in e p e a ion is hus di e en and e e s o a pe cen age poin inc ease o he dependen a iable. The o- bus es ima ion esul s o he wo-s ep sys em GMM model gi en by he Equa ion (1) a e p esen ed in Tab. 3. All ele an diagnos ic es s we e es ima ed o his model, as shown in Tab. 3. To check he alidi y o he se o ins umen s used in he es ima ions, we applied he A ellano and Bond au oco ela ion es (A ellano & Bond, 1991), which shows ha models do no su e om second-o de au oco ela ion. The Hansen es o o e -iden i ying es ic ions esul s show ha he ins umen s used in he models a e no Va iable Numbe o obse a ions Mean S anda d de ia ion Median Minimum Maximum ln GVCBL 924 3.66 0.34 3.70 2.06 4.37 GVC_FL 924 1.11 0.99 0.90 0.16 5.54 ln GVC_TOT 924 3.68 0.34 3.73 2.10 4.39 ln FDImanu 714 8.84 1.22 8.72 6.19 11.06 ln RDinGDP 924 −0.11 0.48 −0.15 −0.97 0.94 ln Manu GDP 924 2.80 0.23 2.84 2.26 3.18 ln REER 924 4.61 0.06 4.62 4.36 4.76 ln GDP 924 11.03 0.91 10.73 9.34 13.12 ln GDPpc 924 9.27 0.34 9.28 8.34 9.92 No e: All a iables excep he GVC_FL (gi en he cha ac e is ics o da a wi h low alues and e y low yea ly changes) a e exp essed in na u al loga i hms. Sou ce: own Tab. 2: Desc ip i e s a is ics o a iables included in he model (2005–2018) 11 2023, olume 26, issue 4, pp. 4–16, DOI: 10.15240/ ul/001/2023-4-001 Economics o e -iden i ied, i.e., he e is no misspeci ica ion in he models. So, he econome ic es s con- i m he obus ness o he p esen ed esul s. The es ima es o he model co e ing he pe iod om 2005 o 2018 o coun y (11 CEE coun ies) – indus y (6 ad anced manu ac- u ing indus ies) g oups a e consis en wi h he unde lying economic heo y. The lagged de- penden a iable has shown o be highly posi- i ely associa ed wi h dependen a iables in all h ee models, especially in he case o GVC_FL a iable. The change in he lagged dependen a iable by 1% is associa ed wi h 0.66% change o dependen a iable GVC_TOT, while i is as- socia ed wi h 0.65% change o dependen a i- able GVC_BL. As GVC_FL was no included in he model in he na u al loga i hm o m (gi en i s small changes and small alues), he in- e cep o GVC_FL shows ha he change o he lagged dependen a iable by a pe - cen age poin is associa ed wi h he change o GVC_FL by 1%. These esul s con i m he dynamic na u e o he da a and he choice o he dynamic panel es ima o . The esul s o he analysis show ha independen a iable FDI s ock in manu ac u - ing indus y has shown o be s a is ically sig- ni ican a he 1% le el in he speci ica ions wi h GVC_BL and GVC_TOT as dependen a i- ables. A 1% change in FDI s ock in he manu- ac u ing indus y is associa ed wi h 0.12% change in GVC_TOT and GVC_BL in ad anced ma nu ac u ing indus ies. Simila esul s in hese wo speci ica ions can be associa ed wi h e y low le els o he GVC_FL, making he GVC_BL a iable p e ailing in he calcula ion o GVC_TOT in his se o coun ies. The same independen a iable has no shown o be s a is ically signi ican in he case o GVC_FL. So, inwa d FDI s ock in manu ac u ing indus y has shown o be mo e impo an o acili a ing he GVC backwa d in eg a ion o ad anced ln GVC_TOT ln GVC_BL GVC_FL ln L1 0.66** (0.07) 0.65** (0.07) 1.00** (0.02) ln FDImanu 0.12** (0.03) 0.12** (0.03) –0.01 (0.02) ln RDinGDP 0.08* (0.04) 0.09* (0.04) 0.00 (0.02) ln Manu GDP –0.16* (0.07) –0.18* (0.07) 0.04 (0.05) ln REER 0.10 (0.15) 0.12 (0.15) 0.14 (0.11) ln GDPpc 0.04 (0.05) 0.05 (0.05) –0.03 (0.03) ln GDP –0.11** (0.04) –0.12** (0.04) 0.02 (0.02) Time dummies Yes Yes Yes F-s a is ic 31,292.41 29,931.23 17,795.02 G oups/ins umen s 66/32 66/32 66/58 AR (2) 0.45 0.46 0.65 Hansen s a is ic 0.79 0.79 0.11 No e: ** and * deno e 1% and 5% s a is ical signi icance le els espec i ely; s anda d e o s a e s a ed in pa en hesis; all a iables excep he GVC_FL and GVC_FL lagged alue (gi en he cha ac e is ics o da a wi h low alues and e y low changes yea o e yea ) a e exp essed in na u al loga i hms. Sou ce: own Tab. 3: Two-s ep sys em GMM esul s, obus 12 2023, olume 26, issue 4, pp. 4–16, DOI: 10.15240/ ul/001/2023-4-001 Economics manu ac u ing indus ies in his se o coun- ies, making i one o he de e minan s o hei downs eam posi ion in in e na ional p oduc ion ne wo ks. Gi en he impo ance o he GVC_BL o calcula ing GVC_TOT, i.e., due o small alues o GVC_FL, mos o he esul s o bo h model es ima ions a e highly simila . The sha e o e- sea ch and de elopmen in GDP has shown o be s a is ically signi ican a he le el o 5%, bu wi h a ela i ely low coe icien . Inc ease in he sha e o R&D in GDP by 1% is associa ed wi h he change o GVC_TOT by 0.08% and wi h he change o GVC_BL by 0.09% in ad anced manu ac u ing indus ies, which po en ially indi- ca es ha an inc ease in inno a i eness migh somewha lead o seeking e iciency h ough in- c easing backwa d GVC in eg a ion wi hin hese indus ies. The sha e o manu ac u ing indus y in GDP has shown o be nega i ely associa ed wi h bo h GVC_TOT and GVC_BL a he le el o signi icance o 5%. The 1% inc ease o manu ac u ing indus y sha e in GDP is as- socia ed wi h a dec ease o GVC_BL by 0.18% and a dec ease o GVC_TOT by 0.16% in ad- anced manu ac u ing indus ies. This shows ha he change in indus ializa ion le els is ela ed o a e e se change o he GVC_BL (and GVC_TOT). Also, GDP has shown o be nega i ely associa ed wi h he GVC_TOT and GVC_BL a he le el o signi icance o 1%. The inc ease o GDP by 1% is associa ed wi h he dec ease o GVC_BL by 0.12% and GVC_TOT by 0.11%. GDP pe capi a, used as a measu e o de elopmen , has no shown o be signi ican in any o he models, as well as he eal exchange a e which was used as a measu e o cos compe i i eness. Besides he lagged alue o GVC_FL, none o he inde- penden a iables ha e shown o be signi ican in he model wi h he GVC_FL as he dependen a iable. As no ed in he beginning, in hese ad anced manu ac u ing indus ies, his can be due o e y low le els and pe sis ence o his indica o ac oss ime. 3.2 Discussion The esul s o he analysis in GVC_TOT and GVC_BL models a e la gely in line wi h p e i- ous esea ch explo ing he FDI in low/inwa d s ock as a d i e o GVC pa icipa ion using di e en se s o coun ies, le els o analysis, ime pe iods and da a (Ada o & S eh e , 2021; E ogo e al., 2021; Ke san-Skabic, 2019). As p e iously no ed, Ke san-Skabic (2019) has shown FDI in lows o be a s a is ically signi i- can p edic o o GVC in eg a ion on he coun- y le el in EU CEE coun ies. Impo ance o FDI o he inc ease in GVC pa icipa ion is poin ed ou in ecen pape by Ada o and S eh e (2021), which also show ha FDI s ock as a sha e o GDP is s a is ically signi ican and posi i e in he case o GVC backwa d and GVC o al pa icipa ion a he agg ega e coun y-le el, using he WIOD in he sample o Eu opean coun ies in he pe iod om 2000 o 2014. Also, he esul s o he p esen ed model a e in line wi h E ogo e al. (2021) who explo ed he FDI in low e ec s on GVCs in he sample o de eloping coun ies in he pe- iod om 2010 o 2019 using he UNCTAD-Eo a da abase. Rega dless o he sec o (p ima y, seconda y o e ia y), E ogo e al. (2021) show he impo ance o FDI in low o backwa d GVC pa icipa ion. Thus, he es ima ion esul s p esen ed in his pape u he expand p io e- sea ch s essing he impo ance o FDI in lows/ inwa d s ock in shaping he GVCs. This analysis p o ides some aluable insigh s o policymake s. To bols e he in e- g a ion o ad anced manu ac u ing indus ies in global alue chains, hey should ac i ely p omo e FDI in he manu ac u ing sec o and alloca e esou ces o suppo esea ch and de elopmen . In his con ex i is impo an o highligh he indings o Va a e al. (2021). Thei esea ch shows ha i ms in CEE coun- ies ace challenges in achie ing inno a i e ou - pu s due o a ious ac o s such as inno a ion policies (a bo h na ional and company le el), manage ial a i udes owa ds isk and mis us be ween a ious s akeholde s. Ye , impac o in- e nal R&D on i m p oduc inno a ions is o en posi i e e en o hese coun ies. Addi ionally, Dob zanski (2018) shows ha inc easing spend- ing on inno a ion some imes ails o yield p opo - ional e ec s in he CEE egion, whe eas he old EU membe s a es mo e e ec i ely alloca e he R&D esou ces. Thus, inno a ion policies should be ca e ully designed in his se o coun- ies in o de o achie e he desi ed goals. This is especially impo an as os e ing globally compe i i e ad anced manu ac u ing indus ies can yield echnological ad ancemen s ol- lowed by economic g ow h and an inc eased demand o highe -skilled jobs wi h co espond- ingly highe wages. In his con ex he quali y o human capi al is pa icula ly impo an and