Knowledge spillovers along the sustainable supply chain of China's listed companies: The role of long-term orientation

Chen, Li eng; Shen, Qi aisong; Yu, Xiaolan; Chen, Xiaohui
A icle
Knowledge spillo e s along he sus ainable supply
chain o China's lis ed companies: The ole o long- e m
o ien a ion
Jou nal o Inno a ion & Knowledge (JIK)
P o ided in Coope a ion wi h:
Else ie
Sugges ed Ci a ion: Chen, Li eng; Shen, Qi aisong; Yu, Xiaolan; Chen, Xiaohui (2024) : Knowledge
spillo e s along he sus ainable supply chain o China's lis ed companies: The ole o long- e m
o ien a ion, Jou nal o Inno a ion & Knowledge (JIK), ISSN 2444-569X, Else ie , Ams e dam, Vol. 9,
Iss. 2, pp. 1-12,
h ps://doi.o g/10.1016/j.jik.2024.100478
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/327383
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 ps://c ea i ecommons.o g/licenses/by/4.0/
Knowledge spillo e s along he sus ainable supply chain o China’s lis ed
companies: The ole o long- e m o ien a ion
Li eng Chen
a,b
, Qi aisong Shen
a,c,
*, Xiaolan Yu
d
, Xiaohui Chen
e
a
School o Business, Hangzhou Ci y Uni e si y, Hangzhou 310015, Zhejiang, China
b
School o Public A ai s, Zhejiang Uni e si y, Hangzhou 310058, Zhejiang, China
c
Resea ch Cen e o Digi al T ans o ma ion and Social Responsibili y Managemen , Hangzhou Ci y Uni e si y, Hangzhou 310015, Zhejiang, China
d
School o Managemen , Zhejiang Uni e si y o Technology, Hangzhou 310014, Zhejiang, China
e
School o Economics, Guangxi Uni e si y, Nanning 531499, Guangxi, China
ARTICLE INFO
A icle His o y:
Recei ed 20 No embe 2023
Accep ed 18 Ma ch 2024
A ailable online 9 Ap il 2024
ABSTRACT
A new indus ial de elopmen pa adigm, he Fi h Indus ial Re olu ion o ‘Indus y 5.0’is expec ed o
humanise g een echnological inno a ion (GTI), social esilience, and sus ainable de elopmen o indus ial
ecosys ems. Emphasising he supplie -cus ome in e ac ion, ‘Indus y 5.0’is echnological e olu ion, which
is a alue-d i en ini ia i e ha d i es knowledge spillo e s along he sus ainable supply chain (SSC). In his
con ex , en e p ises a e inco po a ing mo e GTI ac i i ies in o hei pa ne s’SSC s a egies. This s udy
employs samples o China’s A-sha e lis ed companies co e ing he pe iod om 2010 o 2021 o examine he
spillo e e ec s be ween cus ome s’GTI and supplie s o al ac o p oduc i i y (TFP). Bo h GTI and TFP indi-
ca o s a e impo an ep esen a ions o ‘Indus y 5.0’, wi h ou majo conce ning on he enhancemen o sup-
plie ’s long- e m o ien a ion (LTO). We calcula ed TFP by he Olley-Pakes and Le insohn-Pe in me hods,
di iding GTI indica o s in o h ee le els: subs an ial g een echnological inno a ion (SGTI), non-subs an ial
g een echnological inno a ion (NGTI), and o e all g een echnological inno a ion (OGTI). Ou empi ical
esul s show ha 1) Cus ome ’s GTI con ibu es o supplie ’s TFP, and supplie ’s LTO including en i onmen-
al, social & go e nance (ESG) pe o mance, and R&D inpu can s eng hen his beneficial link; 2) cus ome ’s
NGTI ep esen s a significan di e en ia ion in compa ison wi h SGTI, whe e R&D in es men weakens he
posi i e ela ionship be ween cus ome ’s NGTI and supplie ’s TFP; 3) he e ogenei y analysis indica es ha
he spillo e s o cus ome ’s GTI on supplie ’s TFP a e mo e p onounced among s a e-owned en e p ises,
high- ech en e p ises, and non-pollu ing en e p ises. No ably, he connec edness demons a es dynamic pa -
e ns h ough 2SLS and GMM eg essions, highligh ing he impo ance o en e p ises’SSC and LTO being he
shock ansmi e s o hei TFP. Ou s udy is p one o benefi lawmake s, egula o s, and fi m execu i es
esponsible o analysing and assessing he SSC, p o iding mo e policy implica ions wi h u u e p ospec s o
sus ainable ans o ma ion in he upcoming ‘Indus y 5.0’e a.
© 2024 The Au ho (s). Published by Else ie España, S.L.U. on behal o Jou nal o Inno a ion & Knowledge.
This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/)
Keywo ds:
G een echnological inno a ion
To al ac o p oduc i i y
Sus ainable supply chain
Long- e m o ien a ion
Indus y 5.0
JEL classifica ion:
G18
G30
O30
O32
In oduc ion
Li e a u e e iew
Sus ainable p oduc i i y has ecei ed inc easing a en ion om
bo h go e nmen s and en e p ises since he end o he las cen u y.
The ocus o p oduc i i y g ow h is now di ec ed a Indus ial In e -
ne o Things (IIoT) o ‘Indus y 5.0’−a new indus ial de elopmen
pa adigm ela ed o he humanisa ion o g een echnologies, social
esilience, and sus ainable de elopmen o indus ial ecosys ems.
The e o e, he long- e m impac o echnological p og ess on o al
ac o p oduc i i y (TFP) g ow h should be u he assessed h ough
an in eg a i e amewo k (Colino e al., 2014;Chou e al., 2014;
Edquis & Hen ekson, 2017). The impac o g een echnological inno-
a ion (GTI) ac i i ies on eco-e ficiency is mo e specific and a ge ed
(Chen e al., 2017;Ghise i e al., 2017;Liu e al., 2020), p omo ing
he esponse o he economy o global clima e change (Chen e al.,
2022;Saunila e al., 2018;Yu e al., 2019). The GTI emphasises g een
ans o ma ion in he indus ialisa ion p ocess, which can elie e he
damage caused by indus y on he ecological en i onmen in he con-
ex o ‘Indus y 5.0’. China’s expansion o hea y indus ies o e he
las se e al decades has e e sed he cou se o i s ecological p ocess
(Chen, 2015), and China’s impo compe i ion u he s imula es a
* Co esponding au ho .
E-mail add ess: [email p o ec ed] (Q. Shen).
h ps://doi.o g/10.1016/j.jik.2024.100478
2444-569X/© 2024 The Au ho (s). Published by Else ie España, S.L.U. on behal o Jou nal o Inno a ion & Knowledge. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/)
Jou nal o Inno a ion & Knowledge 9 (2024) 100478
Jou nal o Inno a ion
&Knowledge
h ps://www.jou nals.else ie .com/jou nal-o -inno a ion-and-knowledge
apid g ow h in echnological inno a ion and TFP since he end o he
las cen u y (Bloom e al., 2016). Unde he dual p essu e o sus ain-
able p oduc i i y and en i onmen al p o ec ion, he Chinese go e n-
men has con ibu ed o global ecosys ems h ough indi ec policy
lea ning and GTI di usion (Zhu e al., 2019). Consis en wi h he
esea ch o Song e al. (2022) and Chen e al. (2024), his s udy
selec ed GTI and TFP indica o s as impo an ep esen a ions o
‘Indus y 5.0’.
Fu he mo e, he supplie -cus ome in e ac ion (SCI) can signifi-
can ly a ec s ock p ices (Pandi e al., 2011), CEO u no e (In in oli
e al., 2017), adul e a ion isks (Le i e al., 2020), ea nings announce-
men s (Cho e al., 2020), in es men e ficiency (Chiu e al., 2019), and
inno a ion ac i i ies (Chu e al., 2018). Howe e , exis ing empi ical
esea ch ocuses on he influencing mechanism o en e p ises’inno-
a ion on p oduc i i y a he indi idual le el, wi h a ew conce ning
he SCI (Thomas and Ca s en, 2020). Based on he SCI, a sus ainable
supply chain (SSC) has eme ged as he managemen o ma e ial and
in o ma ion flows wi h he join objec i e o imp o ing he sus ain-
able ou comes (Kobe g & Longoni, 2019), aligned wi h en i onmen al
and social pe o mance (Fa ooque e al., 2019). Eme ging echnolo-
gies ha e also gi en ise o mo e oppo uni ies o sus ainable ope a-
ions wi hin supply chain managemen (Choi, 2019;Choi e al., 2020;
Xu e al., 2023), and he new indus ial de elopmen pa adigm o
‘Indus y 5.0’ u he highligh ing he impo ance o SSC (Wang e al.,
2023). The key aspec is de e mining he influencing mechanism o
pa ne s’sus ainable ope a ions on bo h TFP and GTI pe o mance.
The main p oblem lies in he unce ain y o he in e nal impac o
SSC wi hin he ela ionship be ween he TFP and GTI indica o s. In
addi ion, he e may exis mo e nega i e ex e nali ies o a di e en
kind; he e o e, he sea ch o a comp omise and consis ency o
in e es s is o undamen al impo ance. Inspi ed by he p omo ion o
‘Indus y 5.0’, we di ec ed ou esea ch owa d sea ching o
app oaches o he GTI ac i i ies o pa ne s’TFP in China’s SSC pa a-
digm. Based on he long- e m ela ionship and o ien a ion, his s udy
aims o s imula e empi ical esea ch leading o hough -p o oking
wo ks ha ha e subs an ial ele ance o sus ainable business s a e-
gies and GTI suppo ing sus ainable p oduc i i y along he SSC.
The emainde o his pape is o ganised as ollows: Sec ion
2"Hypo hesis de elopmen " de ails he heo e ical ounda ion unde -
pinning he in es iga ion and p oposes he hypo heses. Sec ion 3"
Me hodology" desc ibes he s udy da a, a iables, and models. Sec-
ion 4"Empi ical esul s" p esen s he benchma k eg ession esul s,
obus ness checks, and a he e ogenei y analysis. Finally, Sec ion
5"Conclusions, enligh enmen s, and limi a ions" concludes wi h
a summa y o he findings and p ac ical implica ions o u u e
p ospec s.
Ou con ibu ions
The implemen a ion o IIoT o ‘Indus y 5.0’a aches a majo con-
ce n o alue c ea ion in he indus ial ecosys ems. This s udy ad an-
ces he li e a u e by iden i ying cus ome s’GTI as an impo an
an eceden o supplie s’TFP o SSC and explo ing he mode a ing
oles o en e p ises long- e m o ien a ion (LTO) indices. B oadly, we
con ibu e o he li e a u e on LTO and he SSC as ollows.
(1) Exis ing esea ch has ocused on he TFP o ocal fi ms (Song e al.,
2022), neglec ing he po en ial influencing mechanism o pa -
ne s’GTI (such as cus ome s) on fi ms’TFP. Meanwhile, he R&D
spillo e e ec o cus ome s’R&D inpu (RDI) on a supplie ’s R&D
ou pu is posi i ely mode a ed by pa ne s’ echnological p oxim-
i y (Isaksson e al., 2016). This s udy explo es he e ec i e pa h-
ways o s imula ing cus ome s’GTI p ac ices on supplie s’TFP
g ow h, while RDI plays an impo an ole in en iching and
ex ending he cu en scope o knowledge spillo e heo y in aca-
demia.
(2) As a comp ehensi e non-financial indica o , ESG (en i onmen al,
social, and go e nance) a ing has been comple ely in es iga ed
o i s causal ela ionship in he econome ics field; howe e , a
ew empi ical s udies ha e cla ified he enhancing and mode a -
ing mechanisms o ESG e o s. Mo eo e , p e ious s udies indi-
ca e ha LTO is highly ela ed o RDI (Flamme & Bansal, 2017;
Mille & Xu, 2020), en i onmen al pe o mance (Du ach & Wien-
ga en, 2017;Dou e al., 2019) and social esponsibili y (Kim e al.,
2020;Wang & Bansal, 2012). Combining a deepe insigh o SCI
and LTO, his s udy in oduces he mode a o o ESG indices in o
he econome ics model. Ou findings con ibu e owa d p esen -
ing he in e nal d i ing mechanisms and impac o LTO pe o -
mance a he mic o en e p ise le el, filling he esea ch gap in he
enhancemen o he o e all ESG p ac ice o he upcoming ‘Indus-
y 5.0’e a.
(3) A sus ainable supply chain (SSC) play a c i ical ole in acili a ing
en i onmen ally and economically sus ainable ansi ions in
human communi ies (Rekli is e al., 2021). Indus y 5.0 has
eme ged as a no el imp o emen in SSC by u ilising ad anced
echnologies o con inuously moni o and manage indus ial p o-
cesses (Wang e al., 2023). Howe e , he cu en li e a u e does
no ex end o knowledge spillo e s in iew o he uniqueness o
cus ome s’GTI ac i i ies along he SSC. This pape cons uc s a
heo e ical analysis amewo k o SSC spillo e s, ocusing on he
in e nal mode a ing impac o LTO. We examine how cus ome s’
GTI s imula es supplie s’TFP, which may a ec he es ablishmen
o long- e m ela ionships wi h co e en e p ises and comple e
exploi a ion o echnical knowledge esou ces wi hin he SSC.
F om he pe spec i e o pa ne s’GTI ac i i ies, ou findings also
con ibu e o he b oade li e a u e on sea ching o app oaches
owa d SSC pa adigm, p o iding a econdi e unde s anding o
‘Indus y 5.0’p ospec s in co po a e finance.
Hypo hesis de elopmen
Cus ome ’s GTI and supplie ’s TFP
Kelle (2002) and Lychagin e al. (2016) examined he influence o
knowledge spillo e s on en e p ise p oduc i i y g ow h a bo h
mac o and mic o le els. Technological inno a ion can acili a e e fi-
cien esou ce u ilisa ion and educe p oduc ion cos s (Gold a b &
Tucke , 2019), and i is posi i ely co ela ed wi h TFP g ow h (Balasu-
b amanian & Si adasan, 2011). Fu he mo e, he imp o emen in TFP
due o upg ades in he echnology-d i en indus ial s uc u e is also
eflec ed in he s eng hening o he indus ial s uc u e based on GTI
ac i i ies (Zhao e al., 2022). The benefi s o GTI ac i i ies include
inc eased p oduc i i y g ow h (Po e , 1991) and educed p oduc ion
cos s (Be one e al., 2013;Sun e al., 2017). The GTI can inc ease TFP
in OECD (O ganiza ion o Economic Co-ope a ion and De elopmen )
membe coun ies (Sohag e al., 2021). Compa ed wi h de eloping
coun ies, he GTI has mo e a significan impac on he TFP o de el-
oped economies (Du & Li, 2019). Howe e , de ailed esea ch a he
mic o-le el is s ill lacking, indica ing a esea ch gap in econome ics.
Wi h ega d o bo h SCI and SSC, he social exchange heo y (Law-
le e al., 1999) posi s ha he majo i y o in e pe sonal in e ac ion
beha iou o igina es om social exchange, and he ‘ ecip oci y e ec ’
is he basic p inciple o a mu ually beneficial ela ionship. Technolog-
ical inno a ion is closely associa ed wi h supply chain pa ne s in
e ms o he ‘ ecip oci y e ec ’o ex e nal ne wo ks (Li e al., 2018).
TFP ep esen s a fi m’s GTI ou pu mos in ui i ely, and is spillable
and quan ifiable wi hin he in e nal and ex e nal ne wo ks o he
indus y (Chen e al., 2021). Bo h cus ome s and supplie s along he
SCI would benefi om e ficien communica ion and specialised
L. Chen, Q. Shen, X. Yu e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100478
2
coope a ion wi hin hei pa ne ships (Dye & Singh, 1998;Ko abe e
al., 2003). The e ec o a fi m’s GTI on TFP can be u he eflec ed in
he concen a ion o inno a ion ac o s in he ex e nal ne wo k
(Zhao e al., 2022). The uniqueness o he SCI mo i a es his ‘ eci-
p oci y e ec ’, due o he high mu ual dependence and low p oduc -
ma ke compe i ion (Pandi e al., 2011). G een p ac ices wi hin he
SSC can inc ease cus ome s’posi i e in en ion (Peng, 2023), and a
highe le el o cus ome GTI ends o o e flow knowledge ia SSC o
s abilise his mu ually beneficial ela ionship (Xu e al., 2022), which
may ul ima ely p omo e he supplie ’s TFP. Based on he SCI, he GTI
is highly associa ed wi h inc eases in he lea ning abili y o en e -
p ises, allowing hem o de elop ad anced p oduc ion echnologies
and managemen expe iences (Zhao e al., 2021), hus enhancing TFP
h ough p oduc e o m and he GTI (Song e al., 2022). As he le el
o GTI ac i i ies and he demand o g een p oduc s con inue o
inc ease, a spon aneous flow o supplie s’p oduc ion ac o s will be
ully exploi ed o cus ome s’g een s a egy. The e o e, a close sup-
plie -cus ome ela ionship can u he mo i a e he spillo e e ec s
o cus ome s’GTI ac i i ies (Zhong e al., 2021). Cus ome s’GTI ac i i-
ies ha e addi ional knowledge spillo e e ec s, and supplie s may
ac i ely imp o e hei TFP by lea ning and imi a ing ad anced g een
s a egies om hei cus ome s. Based on he abo e heo e ical
amewo k, we p opose Hypo hesis 1.
H1. Cus ome ’s GTI is posi i ely ela ed o supplie ’s TFP.
Enhancemen o supplie ’s ESG pe o mance
En e p ises’GTI ac i i ies a e closely ela ed o long- e m o ien a-
ion (LTO) (Sae he e al., 2021) and co po a e social esponsibili y
(CSR) (Hao & He, 2022) indica o s. F om he pe spec i e o s ake-
holde heo y (F eeman e al., 2004), CSR can be ecognised as a pow-
e ul and c ucial ac o in e ms o social esponsibili y. Schola s ha e
widely accep ed ha CSR can enhance a co po a e’s epu a ion and
educe he ola ili y o i s s ock p ice (Hu e al., 2014;Mille e al.,
2020;Wang e al., 2022). Thus, CSR ac i i ies can c ea e ex a alue
o sha eholde s i long- e m in es o s manage s a e p ope ly moni-
o ed by Nguyen e al. (2020). Conside ing s akeholde incen i es,
o ganisa ional LTO can maximise s a commi men o imp o ing sus-
ainabili y pe o mance (Cap a & Ne ille, 2012) which in ol es a
p oac i e en i onmen al s a egy (Du ach & Wienga en, 2017), en i-
onmen al p ac ices, and en i onmen al pe o mance e ficacy (Dou e
al., 2019). Meanwhile, LTO can o se he liabili y o newness by mak-
ing s a egic decisions ha ealise he benefi s o CSR ac i i ies be e ,
leading o p ofi able ou comes o CSR ini ia i es (Wang & Bansal,
2012). On one hand, CSR pe o mance is a ocal indica o in building
long- e m ela ionships o o ien a ion wi h cus ome s (Kim e al.,
2020); on he o he hand, LTO benefi s s akeholde s’ ela ionships
(Flamme & Bansal, 2017), based on he con ibu ion o such ela ion-
ships o long- e m alue c ea ion (Edmans, 2012;Flamme , 2015).
Based on CSR ini ia i es, ESG indica o s a e highly co ela ed wi h
o ganisa ional LTO (Capelle e al., 2019;Sa danelli e al., 2022;Chen
e al., 2024). In eg a ing a posi i e in e ac ion be ween LTO and CSR
a he mic oen e p ise le el, we add ess he impo ance o ESG indi-
ces o discuss he possibili y o enhancemen in he p omo ion o GTI
on TFP. P e ious s udies ha e in es iga ed he impac o ESG pe o -
mance on co po a e financial aspec s such as financial pe o mance
(B una e al., 2022;Zheng e al., 2022), financial isks (Capelle e al.,
2019;Shakil, 2021), and financing cos s (Ng & Rezaee, 2015;Ielasi e
al., 2021). Howe e , li le emphasis has been placed on how ESG
e o s influence co po a e p oduc i i y and TFP. In an SSC, he e a e
h ee p ima y mechanisms h ough which ESG can s eng hen he
posi i e impac o GTI on TFP. Fi s , as a gene al concep ela ed o a
fi m’s p oduc i i y, non-financial in o ma ion is an unigno able ele-
men , and i is na u al o ESG a ings o influence TFP (Deng e al.,
2023). Thus, supplie s’ESG pe o mance can imp o e TFP a he indi-
idual le el. Second, ESG p ac ice has become inc easingly impo an
o financial ins i u ions and indi idual in es o s, especially o ais-
ing s akeholde s’awa eness on ecological issues and mi iga ing en i-
onmen al ex e nali ies om social assessmen s (Be g e al., 2022;
Zheng e al., 2022). Imp o ed ESG pe o mance o e s low-cos
financing channels o ex e nal ins i u ions and go e nmen s (Eliwa
e al., 2021;Khou y e al., 2022), which can alle ia e financial p es-
su e on inno a ion ac i i ies (Ape gis e al., 2022). Tha is, a highe
ESG sco e can alle ia e he financial p essu e on GTI ac i i ies (Chen
e al., 2022), hus s eng hening he beneficial link be ween GTI and
TFP. Thi d, ESG e o s s eng hen employees’ ecogni ion and sa is-
ac ion (He e al., 2022) o p omo e labou p oduc i i y and he com-
pe i i e ad an age o p oduc ion (Hu e al., 2018). Wi h such
incen i es om he cus ome side, supplie s’ESG pe o mance
encou ages cus ome s’GTI mo i a ion, which can educe p oduc ion
cos s and p oduce g eene and di e en ia ed p oduc s compa ed o
hose o compe i o s, he eby gaining supplie s’TFP and ma ke
sha e in he long un. Based on his, we p opose Hypo hesis 2 wi h a
deep insigh in o SSC.
H2: Supplie ’s ESG pe o mance s eng hens he posi i e ela ion-
ship be ween he cus ome ’s GTI and supplie ’s TFP.
Enhancemen o supplie ’s R&D in es men
The uppe echelons heo y (Hamb ick & Mason, 1984) posi s ha
he scope o nega i e a ainmen disc epancy mo i a es fi ms o
ac i ely inc ease R&D inpu (RDI), pa icula ly o fi ms ope a ing
wi h LTO (Chen e al., 2023). As echnological inno a ion ela es
s ongly wi h LTO (Ru io e al., 2014), and exogenous long- e m
incen i es can inc ease he amoun o RDI (Flamme & Bansal, 2017),
o ganisa ional LTO is becoming a c i ical p e equisi e o RDI (Mille
& Xu, 2020). The con ibu ion o a fi m’s RDI o p oduc i i y g ow h
has been in es iga ed ex ensi ely, and mos quan i a i e s udies
ha e shown ha RDI posi i ely a ec s p oduc i i y g ow h (G i fi h
e al., 2004). Higon (2007) e iewed p e ious s udies and concluded
ha he ou pu elas ici y o RDI o TFP g ow h is be ween 0.015 and
0.37. The ecen li e a u e has d awn con o e sial conclusions
ega ding he ela ionship be ween RDI and TFP. Acco ding o Fang e
al. (2020), bo h pa en s ocks and ini ial pa en ing e en s con ibu e
undamen ally o TFP. Xu and Deng (2022), u he indica ing ha
go e nmen -di ec ed R&D expendi u e has a significan impac on
he u ban TFP g ow h. Howe e , Eje mo e al. (2011) and Yu e al.
(2021) find ha RDI canno be con e ed in o high p oduc i i y,
which is known as he ‘Swedish Pa adox’e ec . Gi en he impo an
ole RDI plays in ensu ing he su i al and g ow h o fi ms a e
COVID-19 (Chen e al., 2022), an inc easing numbe o fi ms a e seek-
ing o eco e om LTO and RDI ac i i ies, whe eby i is easonable
o assume ha RDI s imula es TFP g ow h wi h LTO incen i es.
Conside ing SCI and SSC, he high-cos and complexi y o inno a-
ion ac i i ies push companies o le e age ex e nal knowledge
esou ces (Lau sen & Sal e , 2006). Reduced p oduc ma ke compe i-
ion and mu ual dependence u he mo i a e aci knowledge ans-
e du ing SCI (Isaksson e al., 2016). Acco ding o he knowledge
spillo e heo y, knowledge spillo e s eflec unexpec ed, uncom-
pensa ed, and in o mal knowledge ans e s (Acs e al., 2009), and
such knowledge ans e s ensue in bo h ups eam and downs eam
indus ies (Xu e al., 2022). By exploi ing knowledge spillo e s along
he SSC wi h LTO, enhancemen o RDI on he ela ionship be ween
GTI and TFP can be ca ego ised in o he ollowing h ee aspec s. Fi s ,
ups eam R&D ac i i ies a e key inpu s o downs eam R&D incen-
i es (Ha ho , 1996). Compa ed wi h ho izon al R&D spillo e s
among i al fi ms, knowledge exchange in SCI ac s as an addi ional
channel o p omo ing he ansac ion o aluable aci knowledge
(Cassiman & Veugele s, 2002). Thus, cus ome s’RDI con ibu e signi -
ican ly o he supplie ’s RDI, owing o he acili a ed possibili ies o
R&D spillo e e ec s along SCI h ough epea ed in e ac ions (Isaks-
son e al., 2016). Such RDI incen i es along he SCI a e ypically
L. Chen, Q. Shen, X. Yu e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100478
3
associa ed wi h epea ed in e ac ions be ween pa ne s, enabling he
ans e o g een echnical knowledge and pa en s om he SSC. Sec-
ond, knowledge di usion om RDI can s imula e bo h labou p oduc-
i i y and TFP le el (Bin, 2008). By combining insigh s de i ed om
knowledge spillo e s and LTO, supplie s along he SSC may ac i ely
inc ease RDI o enhance hei TFP le els and main ain long- e m ela-
ionships wi h cus ome s. Thi d, he impac o cus ome ’s RDI on sup-
plie s’RDI is posi i ely mode a ed by he du a ion o SCI and
echnological p oximi y (Isaksson e al., 2016). Wi h such incen i es,
supplie s’RDI encou ages en e p ise manage s o conduc R&D ou -
pu and GTI ac i i ies ha educe p oduc ion cos s and p oduce
g eene and mo e di e en ia ed p oduc s han compe i o s. Acco d-
ingly, supplie ’s RDI s eng hens he posi i e impac o he cus om-
e ’s GTI on TFP. Consis en wi h he abo e hypo hesis de elopmen ,
we p opose Hypo hesis 3 and esea ch design o Fig. 1, as shown
below.
H3: Supplie ’s RDI s eng hens he posi i e ela ionship be ween
he cus ome ’s GTI and supplie ’s TFP.
Me hodology
Da a
This s udy selec ed A-sha e lis ed companies on he Shenzhen and
Shanghai S ock Exchange om 2010 o 2020, company-le el da a
we e ob ained om lis ed companies’da abases such as Chinese
Resea ch Da a Se ices (CNRDS), China S ock Ma ke & Accoun ing
Resea ch (CSMAR), and Wind China Financial da abase (WIND). Based
on Isaksson e al. (2016) and Chu e al. (2018), ou samples comp ised
all supplie −cus ome pai s ha can be iden ified in he CSMAR da a-
base du ing he pas decade, whe e he supplie −cus ome -yea
obse a ions o A-X-2015, A-Y-2015, and A-Z-2015 ep esen sup-
plie A, co esponding o mul iple cus ome s X, Y, and Z in 2015,
espec i ely. As cus ome s’annual GTI in o ma ion is only disclosed
in he ollowing yea , he impac on supplie s has 1-yea lag pe iod.
Thus, ou sampled cus ome da a a e om 2009 o 2020, and he co -
esponding supplie samples a e om 2010 o 2021; all con inuous
a iables we e winso ised by 1 % and 99 % quan iles o con ol he
influence o ex eme alues. We conduc ed he ollowing da a p oc-
essing by excluding (1) samples wi h missing da a, (2)financial and
eal es a e samples, and (3) ST and PT samples du ing he esea ch
pe iod. Finally, a o al o 1410 supplie -cus ome -yea obse a ions
we e acqui ed, including 528 supplie s and 408 cus ome s.
Va iables
Explana o y a iables
The numbe o pa en applica ions is closely ela ed o he TFP
indica o (Balasub amanian & Si adasan, 2011), and g een pa en
applica ions can significan ly imp o e he TFP pe o mance o lis ed
Chinese companies (Song e al., 2022). Thus, he numbe o g een pa -
en applica ions is conside ed a p oxy o a fi m’s GTI ac i i ies (Liu e
al., 2022). Consis en wi h p e ious li e a u e, his s udy employs he
numbe o g een pa en applica ions as a p oxy a iable o measu e
GTI ac i i ies, dis inguishing GTI om dimensions such as subs an ial
g een echnological inno a ion (SGTI), non-subs an ial g een echno-
logical inno a ion (NGTI), and o e all g een echnological inno a ion
(OGTI) (aligned wi h Cao e al., 2022;Jiang & Bai, 2022;Zhang e al.,
2021). Conside ing he huge de ia ion and delay e ec s o pa en
applica ions wi hin companies, we conduc ed na u al loga i hm
p ocessing and lag by one pe iod on he pa en coun s.
Explained a iables
The Olley-Pakes (OP) and Le insohn-Pe in (LP) me hods a e
inc easingly adop ing empi ical measu es o he g een bioeconomy
o bio-based economy impac on panel da a (Chen e al., 2021;Mog-
haddasi & Pou , 2016;Zheng e al., 2021). The pa ame ic es ima ion
me hod p oposed by Olley and Pakes (1996) is as ollows:
lnYi ¼aklnKi þailnLi þaaFAi þasFOi þaeEXi
þX
m
dmYea mþX
n
λnRegnþX
k
zkIndkþAi ð1Þ
In o mula (1),Y
i
ep esen s he o al ou pu o company
i
du ing
yea
; and A
i
ep esen s he TFP ha could acili a e he ma ginal
p oduc o all inpu s.; K
i
and L
i
delega e he capi al inpu and labou
inpu o he company; Yea , Reg, and Ind delega e dummy a iables
o he en e p ise yea , egion, and indus y, espec i ely. Addi ion-
ally, dummy a iables o fi m age (FA), fi m owne ship (FO), and
expo beha iou (EX) a e in oduced in he o mula (1). Howe e ,
he Olley-Pakes (OP)me hod may cause he missing samples wi hin
he es ima ion p ocess (Chen e al., 2021), whe eas he Le insohn-
Pe in (LP) me hod can e ec i ely a oid hese endogenei y and un-
ca ion issues (Li e al., 2022). The e o e, he p oduc ion unc ion es i-
ma ion o TFP de eloped by Le insohn and Pe in (2003) is u he
adop ed, as shown below:
lnYi ¼bklnKi þbmMi þWi þAi ð2Þ
lnYi ¼bllnLi þF ðKi þMi ÞþAi ð3Þ
In o mula (2), W
i
ep esen s he in e media e inpu p oduc s,
and M
i
is calcula ed using a mono one inc easing unc ion M
i
=M
(k
i
,W
i
). This s udy uses he LP me hod o o mula (3) o sol e he
da a unca ion and endogenei y issues in TFP calcula ion, and he
TFP sco es measu ed by he LP unc ion a e employed as ou obus -
ness indica o s.
Mode a ing a iables
ESG pe o mance. D awing upon Tang (2022),Deng e al. (2023), and
Zhong e al. (2023), his s udy employs Huazheng’s ESG a ings o
measu e he ESG pe o mance o he en e p ise, which di e en ia es
ESG a ings in o nine scales (C-AAA) and assigns poin s (1−9) o
each obse a ion. The Huazheng ESG da abase comp ises 14 second-
a y-and 26 hi d-le el sco ing sys ems.
R&D in es men . The a io o R&D in es men o ope a ing income
can e ec i ely delega e a fi m’s inno a i e e o s while allowing o
conduc quan i a i e esea ch (Isaksson e al., 2016;Taques e al.,
2021). This s udy employs he a io o R&D inpu o ope a ing income
as a quan i a i e indica o o RDI and examines knowledge spillo e s
along he SSC.
Con ol a iables
Re e ing o Chen e al. (2022),Zheng e al. (2022), and Song e al.
(2022), we in oduce a se ies o cha ac e is ic a iables ha a ec
en e p ises’TFP and GTI, including e u n on o al asse s, fixed asse s
a io, ope a ing income g ow h a io, p ofi abili y a io, boa d di ec-
o s, boa d duali y, independen boa ds, and book alue (see Table 1).
Fig. 1. . Resea ch concep .
L. Chen, Q. Shen, X. Yu e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100478
4

Meanwhile, yea - and indus y-fixed e ec s we e added o educe
he di e ence in esea ch esul s be ween yea s and indus ies.
Resea ch models
Benchma k model
Based on p e ious heo e ical analysis esul s, his s udy es ab-
lishes he ollowing fixed e ec (FE) panel eg ession model:
TFPi; ¼b0þb1GTIi; 1þb2Con olsi; þei; ð4Þ
In Model (4), he subsc ip s iand deno e he indus y fi m and
yea , espec i ely. The explana o y a iable is TFP, and he explained
a iable o GTI includes subs an ial g een echnological inno a ion
(SGTI), and non-subs an ial g een echnological inno a ion (NGTI). C
on olsi; includes all con ol a iables, and ei; is he andom e o
e m o he model.
Mode a ing e ec models
To u he examine he enhancemen e ec o LTO on he posi i e
ela ionship be ween GTI and TFP, he ollowing econome ic models
we e es ablished based on he me hod desc ibed by Shakil (2021)
and Zhao e al. (2021), whe e he coe ficien s o he in e ac ion e m
in Models (5) and (6) ep esen he mode a ing e ec s o LTO and GTI
and include he h ee dimensions o OGTI, SGTI, and NGTI.
TFPi; ¼b0þb1GTIi; 1þb2ESGi; þb3GTIi; 1ESGi;
þb4Con olsi; þei; ð5Þ
TFPi; ¼b0þb1GTIi; 1þb2RDIi; þb3GTIi; 1RDIi;
þb4Con olsi; þei; ð6Þ
Empi ical esul s
Desc ip i e s a is ics and co ela ion ma ix
Table 2 p esen s he desc ip i e s a is ics ha o e concise
insigh s in o he dis ibu ion and cha ac e is ics o he a iables
examined. The mean alue o TFP is 6.729, wi h a s anda d de ia ion
(SD) o 1.008, indica ing su ficien a ie y wi hin he TFP among
en e p ises. The mean alues o GTI indica o s (SGTI, NGTI, and OGTI)
we e 1.366, 1.040, and 2.406, wi h SD o 1.697, 1.405, and 2.950,
espec i ely. This sugges s ha he GTI indica o s can be ully iden i-
fied be ween he samples. The ESG indica o s displayed a medium
le el o a iabili y (SD = 1.064), and he RDI showed wide a iabili y
(SD = 5.775). The absence o ou lie s wi hin he sample en e p ises
indica es ha no ex eme obse a ions a e likely o influence subse-
quen esea ch significan ly.
Table 3 p esen s Pea son’s co ela ion coe ficien s, he coe ficien s
be ween TFP and GTI indica o s (SGTI, NGTI, OGTI) a e 0.133, 0.118,
and 0.133, wi h a s a egically significan 1 %. The OGTI comp ises he
wo dimensions o SGTI and NGTI, so i is easonable o assume a
high co ela ion be ween SGTI, NGTI, and OGTI. Mo eo e , he co e-
la ion coe ficien s o o he a iables in Table 3 a e less han 0.7,
implying ha hese a iables can be ully iden ified, and he e is no
se ious mul icollinea i y in ou panel da a.
Reg ession analysis
Benchma k eg ession
Table 4 epo s he eg ession esul s o he benchma k model o
cus ome ’s GTI on supplie ’s TFP. Non-subs an ial g een echnological
inno a ion (NGTI) can a ec TFP posi i ely a e impo ing he con-
ol a iables. Al hough he posi i e ela ionship exis s in column (2),
his impac is no significan unless he influence o he con ol a ia-
bles is conside ed. Mo eo e , bo h OGTI and SGTI demons a e a s a-
is ically significan posi i e coe ficien , indica ing a s ong causal
ela ionship be ween cus ome ’s subs an i e g een inno a ion ac i i-
ies on supplie ’s TFP, he eby suppo ing Hypo hesis 1.
The mode a ing e ec o ESG and RDI
To ully iden i y he in insic influencing mechanism o he GTI on
TFP, his s udy in oduces ESG and RDI indica o s in o he benchma k
eg ession model. Table 5 epo s he mode a ing e ec s o ESG and
RDI, wi h significan coe ficien s o all he c oss- e ms. The esul s
illus a e cus ome ’s NGTI is significan ly di e en om SGTI, and
Table 1
Va iable desc ip ion.
Type Symbol Va iable Desc ip ion
Explana o y a iables SGTI Subs an ial g een echnological inno a ion Na u al loga i hm o he numbe o o al g een pa en s applied by adding 1
NGTI Non-subs an ial g een echnological inno a ion Na u al loga i hm o he numbe o g een u ili y model pa en s by adding 1
OGTI O e all subs an ial g een echnological inno a ion Na u al loga i hm o he numbe o g een in en ion pa en s by adding 1
Explained a iable TFP To al ac o p oduc i i y OP and LP me hod based on he Cobb Douglas unc ion.
Mode a o a iables ESG En i onmen al, social, and go e nance pe o mance ESG sco e om 1 o 9
RDI R&D in es men Ra io o R&D inpu o ope a ing income
Con ol a iable ROA Re u n on o al asse s Ra io o ne income o o al asse s
FIXED Fixed asse s a io Ra io o fixed asse s o o al asse s
GROW Ope a ing income g ow h a io Ra io o asse g ow h in he cu en yea o o al asse s
LOSS P ofi abili y a io Take 1 i ne p ofi o he cu en yea is less han 0; o he wise, ake 0
BOARD Boa d di ec o Na u al loga i hm o numbe o boa d di ec o s
DUAL Boa d duali y Take 1 i he gene al manage and chai man is he same pe son; o he wise, ake 0
INDEP Independen boa d Na u al loga i hm o numbe o independen boa d di ec o s
BM Book alue Ra io o equi y o ma ke capi alisa ion
No es: Table shows he summa y s a is ics o he a iables.
Table 2
Desc ip i e s a is ics.
Va iables N. Mean Median S. D. Min Max
TFP 1410 6.729 8.322 1.008 4.251 9.162
SGTI 1410 1.366 0.693 1.697 0 7.228
NGTI 1410 1.040 0 1.405 0 6.045
OGTI 1410 2.406 1.099 2.950 0 12.655
ESG 1410 6.376 6 1.064 3 9
RDI 1410 5.846 4.76 5.775 0.001 75.98
ROA 1410 0.048 0.046 0.057 0.250 0.225
FIXED 1410 0.226 0.187 0.162 0.002 0.725
GROW 1410 0.185 0.134 0.388 0.602 4.330
LOSS 1410 0.093 0 0.297 0 1
BOARD 1410 2.157 2.197 0.186 1.609 2.708
INDEP 1410 0.366 0.333 0.051 0.286 0.571
DUAL 1410 0.243 0 0.422 0 1
BM 1410 0.955 0.668 1.036 0.051 10.142
No es: All he a iables a e explained in Table 1.
L. Chen, Q. Shen, X. Yu e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100478
5
supplie ’s RDI weakens he posi i e ela ionship be ween cus ome ’s
NGTI and supplie ’s TFP. Thus, Hypo hesis 2 is ully suppo ed, and
Hypo hesis 3 is pa ial e ified.
Robus ness es ing
To e i y he obus ness o ou esul s om he pe spec i e o TFP
measu emen s, we eplaced he supplie ’s TFP by he LP me hod as
he dependen a iable. Table 6 p esen s he eg ession esul s,
whe e supplie ’s ESG and RDI can enhance he impac o cus ome ’s
GTI on supplie ’s TFP.
Conside ing endogenous p oblems such as measu emen e o s
and missing a iables, we e e o he me hod p oposed by De inge
and Hi h (2020) and Zheng e al. (2022) employing he a e age
numbe o SGTI, NGTI, and OGTI in he indus y and yea as he
ins umen al a iables, ob aining he ins umen al alues o IV_SGTI,
IV_NGTI, and IV_OGTI o in es iga e he influencing mechanism o
cus ome ’s GTI on supplie ’s TFP applying he e ogeneous ools. The
calcula ion o mula is as ollows:
IV_SGTIy;i¼X
n
1
SGTIy;i
!
=nð7Þ
IV_NGTIy;i¼X
n
1
NGTIy;i
!
=nð8Þ
IV_OGTIy;i¼X
n
1
OGTIy;i
!
=nð9Þ
Tables 7 and 8 illus a e he eg ession esul s o he wo-s age leas
squa es (2SLS) app oach using ins umen al a iables, and he coe fi-
cien s o he one-s age eg essions IV_SGTI, IV_NGTI, and IV_OGTI pass
he significance es a he 1 % le el. The p- alue o he Kleibe gen-Paap
Table 3
Co ela ion ma ix.
TFP SGTI NGTI OGTI ESG RDI ROA FIXED GROW LOSS BOARD INDEP DUAL BM
TFP 1
SGTI 0.133*** 1
NGTI 0.118*** 0.808*** 1
OGTI 0.133*** 0.960*** 0.941*** 1
ESG 0.119*** 0.019 0.020 0.020 1
RDI 0.173*** 0.015 0.092*** 0.035 0.002 1
ROA 0.191*** 0.009 0.044 0.026 0.077** 0.032 1
FIXED 0.139*** 0.010 0.098*** 0.053* 0.020 0.197*** 0.212*** 1
GROW 0.078*** 0.017 0.032 0.025 0.048 0.048 0.229*** 0.030 1
LOSS 0.167*** 0.017 0.018 0.018 0.130*** 0.029 0.624*** 0.145*** 0.077** 1
BOARD 0.094*** 0.023 0.050* 0.010 0.178*** 0.004 0.047 0.132*** 0.058* 0.031 1
INDEP 0.024 0.010 0.007 0.009 0.022 0.012 0.045 0.064** 0.020 0.062** 0.472*** 1
DUAL 0.066** 0.050* 0.017 0.037 0.062** 0.114*** 0.082*** 0.143*** 0.063** 0.020 0.261*** 0.214*** 1
BM 0.302*** 0.071** 0.002 0.042 0.173*** 0.237*** 0.274*** 0.141*** 0.072** 0.098*** 0.122*** 0.047 0.139*** 1
No es: All he a iables a e explained in Table 1. The -s a is ics in pa en heses, * p<0.1, ** p<0.05, *** p<0.01.
Table 4
P omo ion o cus ome ’s GTI on supplie ’s TFP.
Va iables TFP
SGTI 0.114*** 0.148***
(3.14) (4.32)
NGTI 0.059 0.084*
(1.17) (1.72)
OGTI 0.067*** 0.081***
(2.71) (3.50)
ROA 4.608*** 3.205*** 4.523***
(7.14) (4.78) (7.00)
FIXED 0.918*** 0.815*** 0.926***
(3.67) (3.08) (3.69)
GROW 0.030 0.026 0.026
(0.42) (0.35) (0.37)
LOSS 0.101 0.132* 0.102
(0.83) (1.03) (0.84)
BOARD 0.296 0.343* 0.262
(1.63) (1.79) (1.44)
INDEP 0.003 0.391 0.071
(0.00) (0.55) (0.10)
DUAL 0.052*** 0.050 0.042
(0.72) (0.67) (0.58)
BM 0.349*** 0.353*** 5.667***
(10.18) (10.19) (10.11)
Cons an 6.573*** 6.668*** 6.568*** 5.556*** 5.818*** 5.667***
(119.74) (116.31) (102.80) (9.94) (9.95) (10.13)
Yea Fix Yes Yes Yes Yes Yes Yes
Indus y Fix Yes Yes Yes Yes Yes Yes
N 1410 1410 1410 1410 1410 1410
Adj. R20.021 0.019 0.022 0.203 0.193 0.205
No es: The -s a is ics alues a e epo ed in pa en heses as * p<0.1, **p<0.05,
***p<0.01.
Table 5
Mode a ing e ec o Supplie ’s ESG and RDI.
Va iables TFP
SGTI 0.192* 0.002
(1.84) (0.07)
NGTI 0.295** 0.091*
(2.31) (1.83)
OGTI 0.088 0.001
(1.44) (0.01)
ESG 0.011 0.035 0.046
(0.026) (0.86) (1.09)
RDI 0.031** 0.016*** 0.032**
(2.52) (2.99) (2.55)
ESG*SGTI 0.042***
(2.73)
ESG*NGTI 0.052***
(2.84)
ESG*OGTI 0.026***
(2.96)
RDI*SGTI 0.006***
(2.67)
RDI*NGTI 0.010**
(2.20)
RDI*OGTI 0.004**
(2.35)
Cons an 5.565*** 6.326*** 6.103*** 4.815*** 5.910*** 4.851***
(9.32) (10.18) (9.75) (5.26) (10.74) (5.29)
Con ol Yes Yes Yes Yes Yes Yes
Yea Fix Yes Yes Yes Yes Yes Yes
Indus y Fix Yes Yes Yes Yes Yes Yes
N 1410 1410 1410 1410 1410 1410
Adj. R20.228 0.191 0.201 0.144 0.213 0.139
No es: The -s a is ics alues a e epo ed in pa en heses as * p<0.1, **p<0.05,
***p<0.01.
L. Chen, Q. Shen, X. Yu e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100478
6
LM s a is ic is less han 0.1, indica ing ha ou ins umen al a iables
can be e ec i ely iden ified. The C agg-Donald Wald F-s a is ic is much
g ea e han he S ock-Yogo c i ical judgmen alue a he 10 % le el,
indica ing ha he e is no weak ins umen al a iable issue in he ol-
lowing eg essions: As shown in Table 8, he eg ession coe ficien s o
all h ee dimensions o GTI can ha e a significan ly posi i e impac on
TFP a e con olling o endogenei y. Fu he , he 2SLS eg ession
esul s indica e ha supplie ’s LTO pe o mance, including ESG and RDI,
can s eng hen he posi i e impac o cus ome ’s GTI on supplie ’sTFP,
which is also consis en wi h he baseline eg ession esul s, hus e i y-
ing he obus ness o ou main es .
Ou pe sonal con ibu ions in ol e a wo-s ep gene alised
me hod-o -momen s (GMM) app oach based on he weak ins u-
men al a iable es . These esul s emphasise he applied scien ific
na u e, which is app op ia e and consis en wi h he ools adop ed in
empi ical esea ch. Bo omed on he ins umen al a iable, we u -
he conduc ed a dynamic es ima ion using he GMM me hod, and all
he esul s p esen ed in Table 9 a e aligned consis en wi h he main
eg ession.
The coe ficien s in columns (1−3) o Table 9 a e significan and
posi i e, indica ing ha a cus ome ’s GTI can imp o e he supplie ’s
TFP significan ly. The e ec o ESG pe o mance on SGTI, NGTI, and
OGTI in columns (4−6) is posi i e, bu he coe ficien o ESG*NGTI is
no significan . The coe ficien s o RDI*NGTI a e significan and nega-
i e; he coe ficien s o bo h RDI*SGTI and RDI*OGTI a e posi i e, con-
sis en wi h he esul s o he main es .
He e ogenei y analysis
(1) As he p ope y na u e o he lis ed companies can de e mine
he alloca ion and u ilisa ion o p oduc ion esou ces such as he GTI,
he e may be disc epancies in knowledge spillo e s be ween cus om-
e s and supplie s. The di ec o indi ec connec ion o s a e-owned
en e p ises (SOEs) wi h he go e nmen can acili a e access o sca ce
esou ces and in o ma ion; he e o e, hei mo i a ion o echnolog-
ical inno a ion is ela i ely weak in compa ison wi h o he s (Chen e
al., 2014). This would a ach mo e impo ance o supplie en e p ises
on he spillo e e ec s o GTI o SOEs. By con as , non-SOEs need o
make mo e e o s on sea ching o knowledge and in o ma ion om
ex e nal ne wo ks o imp o e hei inno a ion le els, hus weaken-
ing he spillo e s o hei GTI ac i i ies. F om he pe spec i e o
mu ual causali y be ween he GTI and TFP, his s udy specula es ha
he GTI o s a e-owned cus ome s has a significan posi i e impac on
hei supplie s’TFP. Panel A o Table 12 shows he eg ession esul s,
indica ing ha cus ome s’GTI has a mo e significan impac on sup-
plie s’TFP among s a e-owned cus ome s.
(2) High- ech en e p ises (HTEs) a e changing he inno a i e p o-
cess o indus ies and people’s li es yle, exhibi ing a mo e compe i-
i e end (Kim & S eensma, 2017). Such inno a i e esou ces a e
cons an ly ga he ed o HTEs h ough he agglome a ion e ec ,
he eby p omo ing companies’independen inno a ion abili y and
p oduc ion e ficiency (Jang e al., 2017). Specifically, among HTEs,
cus ome s’inno a ion abili y has addi ional knowledge spillo e s
and a ec s supplie s’inno a ion (Isaksson e al., 2016). The e o e,
HTEs would ocus on scien ific on ie s o ob aining co esponding
inno a i e esou ces h ough he spillo e e ec s o GTI along he
Table 6
Reg ession esul s o he eplacing a iable.
Va iables TFP
SGTI 0.109*** 0.321*** 0.089**
(3.09) (3.00) (2.21)
NGTI 0.051 0.428*** 0.093*
(1.07) (3.29) (1.85)
OGTI 0.063*** 0.186*** 0.063**
(2.63) (2.95) (2.46)
ESG 0.050 0.031 0.052
(1.17) (0.75) (1.20)
RDI 0.321*** 0.021*** 0.027**
(4.67) (3.80) (4.02)
ESG*SGTI 0.067***
(4.28)
ESG*NGTI 0.073***
(3.89)
ESG*OGTI 0.038***
(4.26)
RDI*SGTI 0.003
(1.00)
RDI*NGTI 0.011**
(2.52)
RDI *OGTI 0.002
(0.14)
Cons an 6.146*** 6.422*** 6.216**** 6.342*** 6.576*** 6.103*** 4.815*** 6.374*** 6.173***
(10.69) (11.27) (10.82) (10.00) (10.39) (9.75) (5.26) (11.38) (10.90)
Con ol/Indi idual/Yea Yes Yes Yes Yes Yes Yes Yes Yes Yes
N 1410 1410 1410 1410 1410 1410 1410 1410 1410
Adj. R20.266 0.248 0.266 0.279 0.262 0.278 0.144 0.267 0.274
No es: The -s a is ics alues a e epo ed in pa en heses as * p<0.1, **p<0.05, ***p<0.01.
Table 7
Es ima ion o weak ins umen al a iables in he 2SLS fi s s age.
Va iables 2SLS fi s s age
SGTI NGTI OGTI
IV_SGTI 0.729***
(5.10)
IV_NGTI 0.681***
(4.40)
IV_OGTI 0.624***
(5.10)
Con ol/Indi idual/Yea Yes Yes Yes
N 822 822 822
Kleibe gen-Paap k LM s a is ic 21.733 14.411 17.316
C agg-Donald Wald F s a is ic 75.647 35.583 38.965
Kleibe gen-Paap k Wald F s a is ic 26.031 19.355 26.053
No es: The -s a is ics alues a e epo ed in pa en heses as * p<0.1,
**p<0.05, ***p<0.01.
L. Chen, Q. Shen, X. Yu e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100478
7
SSC. In his con ex , we assume ha cus ome s’GTI has a mo e signi -
ican posi i e impac on he TFP o high- ech supplie s. Acco ding o
he OECD classifica ion s anda ds, HTEs can be ca ego ised in o six
indus ies engaged in he manu ac u e o : scien ific, elec ical, ae o-
space, compu e , pha maceu ical, and communica ion equipmen .
Meanwhile, “high- ech indus y classifica ion (2018)”o he S a e
Council o China (SCC) u he classified he high- ech indus y in o
nine subca ego ies: R&D se ices, p o essional se ices, in ellec ual
p ope y se ices, e-comme ce se ices, in o ma ion se ices,
inspec ion, es ing se ices, echnology ans e se ices, and
en i onmen al go e nance se ices. D awing upon he me hods
o Wang (2020) and Han and Gu (2021), hiss udyhasclassified
13indus iesasHTEs(seeTable 10). The he e ogeneous esul s
o HTEs a e shown in Panel B o Table 12, demons a ing a mo e
significan impac o cus ome ’s GTI on supplie ’sTFPamong
HTEs han o he s.
Table 8
Reg ession esul s in he 2SLS second s age.
Va iables 2SLS second s age TFP
SGTI 0.408*** 1.708*** 0.659***
(3.60) (2.82) (3.04)
NGTI 0.492* 3.534 0.975
(1.80) (1.35) (1.34)
OGTI 0.203** 2.412* 0.421***
(1.97) (1.72) (2.65)
ESG 0.358** 0.645 0.910
(2.34) (1.38) (1.63)
RDI 0.043 0.066 0.052
(1.11) (0.81) (1.11)
ESG*SGTI 0.255***
(2.65)
ESG*NGTI 0.440
(1.54)
ESG*OGTI 0.194*
(1.72)
RDI*SGTI 0.079**
(2.30)
RDI*NGTI 0.109
(1.25)
RDI*OGTI 0.019*
(1.78)
Con ol/Indi idual/Yea Yes Yes Yes Yes Yes Yes Yes Yes Yes
N 822 822 822 822 822 822 822 822 822
No es: The -s a is ics alues a e epo ed in pa en heses as * p<0.1, **p<0.05, ***p<0.01.
Table 9
Reg ession esul s by GMM me hod.
Va iables TFP
SGTI 0.045* 0.213* 0.005
(1.80) (1.72) (0.16)
NGTI 0.018* 0.540*** 0.120**
(1.75) (2.83) (2.60)
OGTI 0.071* 0.307*** 0.028
(1.94) (3.13) (1.19)
ESG 0.010 0.401*** 0.407***
(0.08) (3.88) (3.44)
RDI 0.181*** 0.025*** 0.030**
(4.50) (3.93) (2.02)
ESG*SGTI 0.038**
(2.03)
ESG*NGTI 0.019
(0.70)
ESG*OGTI 0.014***
(2.97)
RDI*SGTI 0.009*
(1.83)
RDI*NGTI 0.035***
(3.06)
RDI*OGTI 0.004
(0.96)
Con ol/Indi idual/Yea Yes Yes Yes Yes Yes Yes Yes Yes Yes
N 455 455 455 455 455 455 455 455 455
AR (1) 0.217 0.195 0.241 0.207 0.145 0.236 0.226 0.101 0.235
AR (2) 0.534 0.193 0.625 0.739 0.308 0.372 0.656 0.803 0.222
Sa gan 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.072 0.001
Hanse 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
No es: The -s a is ics alues a e epo ed in pa en heses as * p<0.1, **p<0.05, ***p<0.01.
L. Chen, Q. Shen, X. Yu e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100478
8