Xia, He; Yu, Haijing; Wang, Senhao; Yang, Hong
A icle
Digi al economy and he u ban- u al income gap: Impac ,
mechanisms, and spa ial he e ogenei y
Jou nal o Inno a ion & Knowledge (JIK)
P o ided in Coope a ion wi h:
Else ie
Sugges ed Ci a ion: Xia, He; Yu, Haijing; Wang, Senhao; Yang, Hong (2024) : Digi al economy and he
u ban- u al income gap: Impac , mechanisms, and spa ial he e ogenei y, Jou nal o Inno a ion &
Knowledge (JIK), ISSN 2444-569X, Else ie , Ams e dam, Vol. 9, Iss. 3, pp. 1-12,
h ps://doi.o g/10.1016/j.jik.2024.100505
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/327408
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-nc-nd/4.0/
Digi al economy and he u ban− u al income gap: Impac , mechanisms,
and spa ial he e ogenei y
He Xia
a,1
, Haijing Yu
b,1
, Senhao Wang
a
, Hong Yang
a,
*
a
College o Economics and Managemen , Xinjiang Ag icul u al Uni e si y, U umqi, China
b
Business School, Xiang an Uni e si y, Xiang an, Hunan, China
ARTICLE INFO
A icle His o y:
Recei ed 20 No embe 2023
Accep ed 9 June 2024
A ailable online 29 June 2024
ABSTRACT
This s udy examines how digi al economy de elopmen in China impac s he u ban− u al income gap. We
cons uc a digi al economy de elopmen index using panel da a om 30 p o inces and ci ies in China om
2013 o 2021. By combining a spa ially a ying coe ficien model wi h a chain−media ed e ec model, we
quan i y he impac o digi al economy on he u ban− u al income gap and examine i s spa ial he e ogenei y.
The esul s show ha he digi al economy influences he u ban− u al income gap h ough ou di e en
pa hways, each o which exhibi s significan spa ial a ia ion. As hese pa hs o se each o he , he digi al
economy de elopmen in Beijing, Inne Mongolia, Shanxi, Henan, Hubei, Jiangxi, Jiangsu, Guangdong, and
o he p o inces has widened he u ban− u al income gap, esul ing in a digi al di ide e ec . Howe e , in
mos a eas o China’s no heas , eas coas , and wes e n egions, he digi al economy de elopmen has na -
owed he income gap, esul ing in a digi al di idend e ec . This s udy in es iga es he ele an deba es
among schola s and p o ides aluable insigh s and ounda ions o s a egic decision making o educe he
u ban− u al income gap in a ious egions.
© 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:
Digi al economy
U ban− u al income gap
Spa ial he e ogenei y
MGWPR−SDM
Chain−media ed e ec
JEL Classifica ion:
O11
O18
R11
R12
In oduc ion
Wi h changes in majo social con adic ions, he issue o imbal-
anced and insu ficien de elopmen has become he main ac o
es ic ing people’s g owing need o a be e li e. Among hese
issues, he u ban− u al income gap emains se e e (Zhu & Liu, 2022).
Al hough he income gap be ween u ban and u al a eas in China
na owed o e he pas decade, i emains much highe han he
in e na ional a e age (Yuan e al., 2020;Sicula e al., 2007). In 2023,
he pe capi a disposable income o u ban households in China was
51,800 yuan, whe eas he pe capi a disposable income o u al esi-
den s was only 21,700 yuan, much lowe han ha o u ban esiden s.
Na owing he u ban− u al income gap and p omo ing coo dina ed
de elopmen be ween u ban and u al a eas as well as de eloped
and unde de eloped a eas can p omo e social s abili y, educe social
inequali y, imp o e people’s li ing s anda ds, and p omo e he eco-
nomic p ospe i y o he en i e coun y (Wen e al., 2014).
Simul aneously, China’s policies o p omo e digi al ans o ma ion
and digi al economy de elopmen a e con inuously deepening. The
na ion’s digi al economy de elopmen has had a s ong influence on
a global scale (Li e al., 2020;Mu hy e al., 2021), becoming a new
d i ing o ce o p omo ing u ban− u al de elopmen (Zhao e al.,
2023). Howe e , his has es ablished de elopmen gaps be ween
u ban and u al a eas due o echnological i e a ions and di e ing
g ow h e ec s, a ec ing u ban and u al esiden s’income dis ibu-
ion pa e ns. The digi al di ide can esul in income dispa i ies
be ween egions and g oups (Qiu e al., 2021;Yin e al., 2021), and
con o e sy exis s conce ning he impac o digi al echnology on
u al de elopmen (E diaw−Kwasie & Alam, 2016), aising he ques-
ion o whe he digi al economy de elopmen b idges o expands
u ban− u al income inequali y (Peng & Dan, 2023). The e o e, i is
essen ial o in es iga e whe he digi al economy de elopmen can
e ec i ely na ow he u ban− u al income gap and i s impac pa h.
This s udy has impo an p ac ical significance o de eloping s a e-
gically a ge ed policies o na ow he income gap be ween u ban
and u al esiden s and ad ance common p ospe i y.
Cu en ly, esea ch on he impac o digi al economy on he u ban
− u al income gap has been inc easing and en iching he li e a u e.
Al hough schola s ha e eached a consensus on he pe spec i e ha
digi al economy de elopmen can aise incomes, di e ences emain
* Co esponding au ho a : 311 Nongda Eas Road, Shayibak Dis ic , U umqi Ci y,
Xinjiang Uygu Au onomous Region, China.
E-mail add ess: [email p o ec ed] (H. Yang).
1
These au ho s con ibu ed equally o his wo k.
h ps://doi.o g/10.1016/j.jik.2024.100505
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) 100505
Jou nal o Inno a ion
&Knowledge
h ps://www.jou nals.else ie .com/jou nal-o -inno a ion-and-knowledge
ega ding he ela ionship be ween digi al economy de elopmen
and he u ban− u al income gap, which can be summa ized in o
h ee ca ego ies: he digi al economy expands and sh inks u ban
− u al income; he ela ionship be ween he wo is nonlinea , exhib-
i ing a U-shaped ela ionship o fi s sh inking and hen expanding;
o he ela ionship has an in e ed U-shaped end o fi s inc easing
and hen dec easing (Li & Li, 2022a;Yu, 2022;Fan e al., 2022;Si & Li,
2023). This con o e sy is a ibu ed o incomple e esea ch on he
mechanisms o impac o digi al economy de elopmen on he u ban
− u al income gap. A majo i y o s udies ha e neglec ed o conside
egional he e ogenei y (Wang & Xiao, 2021). Fu he mo e, adi ional
eg ession analysis me hods gene ally assume ha indi iduals a e
independen and iden ically dis ibu ed, when, in eali y, digi al
economy de elopmen and u ban− u al income gaps ha e spa ial
au oco ela ion. Cu en ly, esea ch on he impac o digi al economy
on u ban− u al income inequali y is limi ed.
To compensa e o his deficiency, he main con ibu ions o his
s udy a e as ollows. Fi s , in con as o p e ious s udies, we adop a
new a iable coe ficien spa ial Du bin model (MGWPR−SDM) p o-
posed by Yu e al. (2021), which elaxes he adi ional implici
assump ions o independen homogeneous dis ibu ion and all
egions being equal and can simul aneously conside spa ial co ela-
ion and he e ogenei y o eflec he ac ual impac o digi al economy
on he u ban− u al income gap mo e accu a ely. Second, p e ious
esea ch has la gely igno ed he ela ionships be ween media ing
a iables and does no conside spa ial he e ogenei y, which can
educe he c edibili y o he conclusions ob ained by p oducing
biased and inconsis en esul s. This s udy b oadens he scope o pol-
icy ecommenda ions by including hese conside a ions. We embed
he MGWPR−SDM in o a chain−media ing e ec model o examine
he impac mechanisms and spa ial he e ogenei y o digi al economy
de elopmen on he u ban− u al income gap, e ec i ely add essing
his issue. Thi d, his s udy esponds o he cu en academic deba es
and en iches he esea ch on he impac o digi al economy de elop-
men on he u ban− u al income gap, p o iding a heo e ical e e -
ence o ele an go e nmen depa men s o de elop s a egically
a ge ed policies o na ow he u ban− u al income gap ha conside
local condi ions. The e o e, his s udy has academic and p ac ical sig-
nificance.
The emainde o he pape is s uc u ed as ollows. The second
pa p esen s he heo e ical mechanism and esea ch hypo heses
based on he e iew o ele an li e a u e, and he con o e sies and
sho comings o ele an esea ch iewpoin s a e summa ized. The
hi d pa is model cons uc ion, which mainly in oduces he p ocess
o embedding he MGWPR-SDM in o he chain−media ed e ec
model. The ou h pa is abou a iable selec ion and da a desc ip-
ion, which in oduces he selec ion o a iables and da a desc ip ion.
Among hem, we calcula e he le el o digi al economy de elopmen
in a ious egions by cons uc ing a digi al economy de elopmen
index sys em. The fi h pa is empi ical esul analysis, which dis-
plays spa ial he e ogenei y esul s h ough maps. The las pa is he
conclusion and policy implica ions.
Theo e ical analysis and esea ch hypo heses
The digi al economy is a new o m o he economy ha eme ged
wi h he de elopmen and applica ion o digi al echnology. In 1996,
Don Tapsco , he a he o he digi al economy, fi s p oposed he
concep o “digi al economy”(Tapsco , 1996). The o ficial in e p e a-
ion o he digi al economy in China can be aced back o 2016 when
he G20 Hangzhou Summi passed he G20 Digi al Economy De elop-
men and Coope a ion Ini ia i e, p o iding a clea defini ion o he
digi al economy. The digi al economy e e s o he use o digi al
knowledge and in o ma ion as key p oduc ion ac o s, a se ies o eco-
nomic ac i i ies ha use mode n in o ma ion ne wo ks as an impo -
an ca ie , and he e ec i e use o in o ma ion and communica ion
echnology as an impo an d i ing o ce o e ficiency imp o emen .
Economic s uc u e op imiza ion has been widely ecognized by a i-
ous sec o s o socie y.
Impac o digi al economy on he u ban
−
u al income gap
Cu en ly, no consensus has eme ged in he academic communi y
ega ding he impac o digi al economy de elopmen on he u ban
− u al income gap. As b iefly desc ibed abo e, he esea ch conclu-
sions can be oughly di ided in o h ee ca ego ies.
(1) Resea ch on he expanding e ec o he digi al economy on he
u ban− u al income gap. Fu uhol and K is iansen (2007) empha-
sized ha he digi al economy, which is cha ac e ized by he
in e ne , big da a, a ificial in elligence, and o he o ms, may
esul in imbalanced de elopmen be ween u ban and u al a eas,
esul ing in a digi al di ide (Lo ence e al., 2006), which is un a-
o able o low−income u al popula ions (Go ski & Cla k, 2002).
Tan e al. (2017) a gued ha a gap is e iden be ween u ban and
u al esiden s in digi al economy accep ance and applica ion.
Ru al esiden s ha e limi ed esou ces, educa ion, and business
en i onmen s han u ban esiden s (He & Xu, 2019). Fu he mo e,
u al esiden s canno iden i y and use in o ma ion e ec i ely o
ully enjoy he ad an ages o in e ne di usion (L , 2021). In con-
as , u ban esiden s a e in a posi ion o in o ma ion ad an age
and can mo e con enien ly access and use digi al in o ma ion
h ough he in e ne . Fu he mo e, u ban esiden s’digi al awa e-
ness and capabili ies a e con inuously enhanced, which ul ima ely
widens u ban− u al weal h dispa i ies, pa icula ly income (Li &
Ke, 2021). This iewpoin highligh s he possibili y ha he digi al
economy may widen he income gap be ween u ban and u al
a eas, eminding policymake s o pay a en ion o he po en ial
inequali y issues. Howe e , his iewpoin may be oo pessimis ic
and may o e look some o he posi i e impac s o he digi al econ-
omy.
(2) Resea ch on he educ ion e ec o digi al economy on he u ban
− u al income gap. This pe spec i e a gues ha digi al economy
de elopmen con ibu es o na owing he u ban− u al income
gap. Some schola s ha e demons a ed ha he digi al economy
can inc ease a me s’income and na ow he u ban− u al income
gap (Bha nani e al., 2008;Ake & Dial, 2011). Acemoglu and
Res epo (2016) p oposed ha digi al echnologies such as he
in e ne b eak down spa io empo al ba ie s, p o iding labo e s
wi h mo e employmen choices and enhancing ma ke esou ce
alloca ion e ficiency. Liu e al. (2021) a gued ha digi al economy
de elopmen has acili a ed significan achie emen s wo ldwide,
imp o ing impo e ished popula ion’s li ing condi ions. Mo e-
o e , he digi al economy has been ound o di ec ly educe he
u ban− u al income gap (Liu, 2021). Digi al economy de elop-
men can le e age inclusi eness and esou ce sha ing h ough
deep in eg a ion o p oduc ion, li e, and ecology (Fu, 2020). The
digi al economy unc ions as an endogenous d i ing o ce ha
o e comes geog aphical cons ain s, elimina es in o ma ion ba -
ie s, expands employmen oppo uni ies, and ad ances g ow h
in unde de eloped egions. I also d i es op imized ac o alloca-
ion, p omo es he di ision o labo and collabo a ion, imp o es
labo p oduc i i y, enhances economies o scale, and acili a es
coo dina ed egional de elopmen (Duan e al., 2020;Hu e al.,
2017;Song, 2012). This iewpoin emphasizes ha he digi al
economy may p o ide oppo uni ies o na owing he u ban
− u al income gap and a possible way o achie e income balance.
I can mo i a e go e nmen s and en e p ises o inc ease hei
in es men in ela ed fields and p omo e he coo dina ed de el-
opmen o u ban and u al economies. Howe e , his iewpoin
may be oo idealis ic, igno ing he new inequali y issues ha he
H. Xia, H. Yu, S. Wang e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100505
2
de elopmen o he digi al economy may b ing; i may also o e -
look he complex and di e se impac o he digi al economy on
di e en egions and popula ions, which may no di ec ly lead o
he na owing o he u ban− u al income gap.
(3) Resea ch on he nonlinea e ec o he digi al economy on he
u ban− u al income gap. A nonlinea ela ionship exis s be ween
digi al economy de elopmen and he u ban− u al income gap.
Ash a e al.( 2021) a gued ha as in o ma ion and communica-
ions echnology (ICT) is p omo ed and adop ed, income inequal-
i y dec eases o e ime. Jiang e al. (2022) e ealed a U-shaped
ela ionship be ween he digi al economy and he u ban− u al
income gap, wi h he gap na owing in ea ly s ages and widening
in la e s ages. Zheng and Li (2022) p oposed ha he impac o
digi al economy on he u ban− u al income gap can ha e simul a-
neous expanding and na owing e ec s, and he o e all impac
depends on he ela i e magni ude o hese e ec s. Conside ing
ha he digi al economy and he u ban− u al digi al di ide a e in
a con inuous dynamic de elopmen p ocess, hei ela ionship is
likely o exhibi ime− a ying cha ac e is ics, exhibi ing a com-
plex nonlinea ela ionship. Wang and Xiao (2021) ound ha he
de elopmen o he digi al economy has a U-shaped ela ionship
wi h he u ban− u al income gap, ini ially na owing and hen
widening. A simila pe spec i e was p esen ed by Chen and Wu
(2021), who p oposed ha while he ini ial s ages o digi al econ-
omy de elopmen can p omo e u baniza ion and na ow he
u ban− u al income gap, u he de elopmen and he esul ing
digi al di ide nega i ely impac u al su plus labo mig a ing o
ci ies o employmen , widening he u ban− u al income gap.
Con e sely, Li and Li (2022a), and Cheng and Zhang (2019) opined
ha he impac o digi al economy on he u ban− u al income
gap shows an o e all in e ed U-shaped end, wi h an ini ial
inc ease and subsequen dec ease. They p oposed ha in he ini-
ial s age o digi al economy de elopmen , due o he dispa i y in
human capi al and he exis ence o he digi al di ide be ween
u ban and u al a eas, u al esiden s canno ully u ilize digi al
economy esou ces o inc ease hei income compa ed o u ban
esiden s, esul ing in an expanding u ban− u al income gap.
Howe e , la e on, due o he law o diminishing ma ginal u ili y,
he ma ginal impac o u ban digi al economy de elopmen on
income will g adually be su passed by u al “la ecome s”, leading
o a na owing o he u ban− u al income gap. Suppo ing his
iew, Li and Li (2022b) a gued ha he impac o digi al economy
de elopmen on he u ban− u al income gap ollows an in e ed
U-shaped pa e n, and a h eshold e ec exis s in he ela ionship
be ween digi al economy de elopmen and he u ban− u al
income gap. This iew is close o eali y, ecognizing ha he
impac o digi al economy on u ban and u al incomes is complex
and a ied, no single−linea , and may show di e en s ages o
changing ends. Howe e , g asping he specific pa e n o nonlin-
ea ela ionships accu a ely is di ficul , and mo e empi ical
esea ch and da a suppo a e equi ed.
The di e gen findings o he a o emen ioned s udies can be
a ibu ed o h ee main easons. Fi s , he impac o digi al economy
de elopmen on he u ban− u al income gap exhibi s spa ial he e o-
genei y, while exis ing esea ch models o en assume a mean eg es-
sion, assuming ha he impac coe ficien s a e he same o all
egions. Fo example, Wang and Xiao (2021) s udied 30 p o inces in
China om 2013 o 2019 and ound ha he impac o digi al econ-
omy de elopmen on na owing he u ban− u al income gap was
mainly concen a ed in he cen al and wes e n egions, wi h a sligh
widening e ec in he eas e n egion and unclea e ec s in he no h-
eas e n egion. Second, adi ional econome ic models assume ha
esea ch subjec s a e independen ly and iden ically dis ibu ed, bu
he le els o digi al economy de elopmen and he u ban− u al
income gap in di e en egions may exhibi spa ial spillo e e ec s,
iola ing his assump ion. The de elopmen o he digi al economy
and he u ban− u al income gap in each p o ince, ci y, and au ono-
mous egion is no only influenced by local ac o s bu also in e ac s
wi h he de elopmen o he digi al economy o he u ban− u al
income gap in neighbo ing egions. Se e al s udies ha e demon-
s a ed spa ial au oco ela ion (Si & Li, 2023;Wei & Chen, 2020), indi-
ca ing he need o use spa ial econome ic models ins ead o
adi ional eg ession models. Thi d, mos s udies only conside he
di ec impac o digi al economy de elopmen on he u ban− u al
income gap, wi h less conside a ion o i s indi ec e ec s; e en ewe
s udies in ol e he spa ial he e ogenei y o hese indi ec e ec s. I
he di ec ions o he di ec and indi ec e ec s a e opposi e, he o e -
all e ec a e agg ega ion is unce ain.
In summa y, a comp ehensi e unde s anding o he impac o dig-
i al economy de elopmen on he u ban− u al income gap equi es
simul aneous conside a ion o spa ial co ela ion, spa ial he e ogene-
i y, and di ec and indi ec e ec s. Howe e , limi ed s udies ha e
aken his app oach. The e o e, his s udy p oposes he ollowing
hypo hesis:
H1. The influence o he digi al economy on he income gap be ween
u ban and u al a eas is spa ially he e ogeneous, and he influence
in di e en egions has expanding, na owing, o unce ain ela-
ionships.
Impac mechanisms o he digi al economy on he u ban
−
u al income
gap
Digi al economy, human capi al, and he u ban
−
u al income gap
Con empo a y apid digi al economy de elopmen , ep esen ed
by he mobile in e ne , and he ole o in e ne and ICT de elopmen
in co ec ing esou ce misalloca ion, expanding u al esiden s’
access o in o ma ion channels, and enhancing u al esiden s’
human capi al has been confi med by a g owing numbe o schola s.
Knowledge and in o ma ion a e non−compe i i e, and he es ablish-
men and imp o emen o digi al in as uc u e exposes he u al
popula ion o high−quali y online educa ion. The u al labo o ce
can access a weal h o in o ma ion and knowledge h ough he in e -
ne , which can imp o e he b ead h and dep h o he u al knowl-
edge base, enhancing u al esiden s’human capi al (Zhu & Liu,
2022;Mi & Qu, 2022) and imp o ing u al esiden s’abili y o
inc ease incomes, e ec i ely b idging he u ban− u al income gap
(Li & Li, 2022a;Jin & Deng, 2022;Xu & Feng, 2022). Con e sely, some
schola s ha e a gued ha he inc easing co e age and dep h o he
digi al economy gene a es new employmen oppo uni ies ha a o
knowledge-based alen , placing highe demand on human capi al.
Non-ag icul u al employees may encoun e s uc u al unemploy-
men isks, u he widening he u ban− u al income gap (Fan e al.,
2022).
Addi ionally, s udies ha e shown ha p ima y u al human capi-
al is mos ly concen a ed in wes e n p o inces, whe eas in e medi-
a e u al human capi al is mos ly concen a ed in p o inces wi h
ela i ely de eloped mode n ag icul u e. A eas wi h high concen a-
ions o ad anced u al human capi al ha e expe ienced a dynamic
shi om he No heas egion o he Yang ze Ri e Del a egion
(Yao & Deng, 2020). Fan and Cui (2018) showed ha human capi al
has egional he e ogenei y in i s impac on he u ban− u al income
gap; an inc ease in he p opo ion o high−skilled human capi al
widens he u ban− u al income gap in he eas e n and cen al
egions, while e ec i ely na owing he u ban− u al income gap in
he wes e n egions.
In summa y, al hough a deba e emains, he in e ne and o he
digi al echnologies a e impo an influencing ac o s in imp o ing
human capi al, which is a key ac o o he u ban− u al income gap
(Song & Gao, 2022). Ru al human capi al has a significan mode a ing
H. Xia, H. Yu, S. Wang e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100505
3
ole in he impac o digi al economy de elopmen on he u ban u-
al income gap (Xu e al., 2023). Fu he mo e, Li and Zhang (2024)
poin ed ou ha he upg ading o human capi al s uc u e can na -
ow he income gap be ween u ban and u al a eas. The e o e, ample
e idence has sugges ed ha he digi al economy a ec s he u ban
− u al income gap by influencing human capi al de elopmen . Con-
side ing his media ing e ec , we p opose he ollowing hypo hesis:
H2. Human capi al de elopmen media es he impac o digi al econ-
omy on he u ban− u al income gap, a ec ing he income gap
h ough human capi al and exhibi ing spa ial he e ogenei y.
Digi al economy, indus ial s uc u e, and he u ban
−
u al income gap
F om he impac mechanism pe spec i e, he digi al economy
gi es ise o new economic models, p oduc ion ac o s, and echno-
logical suppo o indus ial upg ades (Jing & Sun, 2019). I educes
in o ma ion asymme y and ansac ion cos s (Liu e al., 2023),
s eng hens indus ial links, p omo es indus ial in eg a ion, and
es ablishes a coope a i e and mu ually beneficial indus ial ecosys-
em (Hui & Yang, 2022). The digi al economy can e ec i ely op imize
he e ia y indus ies in he na ional economy and coo dina ion
be ween indus ies, significan ly p omo ing indus ial s uc u e
a ionaliza ion (Liu & Chen, 2021), while also exhibi ing ob ious
egional he e ogenei y (Gao e al., 2023).
The digi al economy de elopmen has significan ly p omo ed he
upg ading o indus ial s uc u e (Liu & Chen, 2021), and exhibi s
egional he e ogenei y, wi h a mo e p onounced e ec on he
upg ading o indus ial s uc u e in he cen al and wes e n egions
(Wang, 2023). An op imized indus ial s uc u e can na ow he
u ban− u al income gap by abso bing he u al labo o ce (Zhou &
Wang, 2013). Howe e , some schola s a gue con e sely, sugges ing
ha because o eno mous scien ific esea ch and inno a ion ac i i ies
in ci ies, digi al echnology con inues o pene a e seconda y and
mode n se ice indus ies, g ea ly imp o ing hei p oduc ion e fi-
ciency. Fo ci ies wi h he main indus ial s uc u e o he second and
hi d indus ies, he u ban income le el will also inc ease, which may
lead o u he expansion o he u ban− u al income gap (Yu, 2022).
Some schola s also a gue ha he impac o indus ial s uc u e
upg ading on he u ban− u al income gap exhibi s a U-shaped cha -
ac e is ic, bu in he eas e n egion, he e is a endency o widen he
u ban− u al income gap, while in he wes e n egion, i is beneficial
o na ow he u ban− u al income gap, while in he cen al egion, i
is gene ally loca ed nea he ansi ion poin (Xu & Liu, 2015).
P e ious esea ch has widely ecognized ha he digi al economy
can a ec he u ban− u al income gap by al e ing he indus ial
s uc u e; he e o e, his s udy p oposes he ollowing hypo hesis
conce ning his media ing e ec :
H3. The indus ial s uc u e media es he impac o digi al economy
on he u ban− u al income gap; howe e , he spa ial impac o
he digi al economy on he income gap is unce ain.
Model cons uc ion
The baseline and media ion e ec models in his s udy a e
based on he Mixed Geog aphically Weigh ed Panel Reg ession
model (MGWPR−SDM) p oposed by Yu e al. (2021).Thismodel
combines he ad an ages o he semipa ame ic spa ially a ying
coe ficien Geog aphically Weigh ed Panel Reg ession model
(MGWPR) and he Spa ial Du bin Model (SDM). Whe eas, he
MGWPR−SDM is a gene alized model ha can be degene a ed
in o a ious spa ially a ying coe ficien panel models, such as
Spa ial Au o eg essi e (SAR) and Spa ial E o models (SEM), o
mee di e en esea ch needs. Howe e , his model is an
imp o emen o e adi ional eg ession and spa ial econome ic
models. Classic spa ial econome ic models, such as he SDM,
only conside he spa ial co ela ion o a iables and add spa ial
lag e ms o dependen and independen a iables as explana o y
a iables. Howe e , hey belie e ha he coe ficien s be ween
each egion a e equal, igno ing spa ial he e ogenei y. Geog aphi-
cally Weigh ed Reg ession (GWR) models ocus on spa ial he e o-
genei y and neglec spa ial s a iona i y and co ela ions. The
MGWPR−SDM used in his s udy combines he ad an ages o he
classical SDM and GWR class models while conside ing spa ial
s a iona i y, spa ial co ela ion, and spa ial he e ogenei y (Yu &
Zhu, 2023).
Spa ial weigh ma ix
The spa ial weigh ma ix is no only he bigges di e ence
be ween spa ial econome ic models and adi ional mean e e sion
models bu also a key ac o in geog aphically weigh ed eg ession
and i s ex ension models. In bo h models, spa ial weigh s a e used o
ep esen he in e dependence be ween indi iduals o egions,
which equi es a uni o m spa ial weigh se ing.
Unlike p e ious spa ial weigh se ings, his s udy d aws on he
concep o g a i a ional po en ial o cons uc an asymme ic spa ial
weigh ma ix ha inco po a es geog aphical and economic dis an-
ces. Po en ial g a i a ional ene gy (F¼Gmi¢mj
dij )is used o exp ess he
mu ual g a i a ional o ce be ween wo objec s, whe e Fis he alue
o he po en ial g a i a ional ene gy be ween he wo objec s, Gis a
g a i a ional cons an alue o 6:67 10 11Nm2=kg2,miand mj
deno e he masses o objec s iand j, espec i ely, and dij is he dis-
ance be ween he cen al mass o he wo objec s. I he nume a o
is eplaced by an economic o al o he wo egions, i combines eco-
nomic and geog aphical dis ance. Using his app oach o cons uc a
spa ial weigh ma ix is mo e accu a e han only using economic o
geog aphical dis ances. Howe e , he spa ial weigh ma ix is sym-
me ic, and he mu ual influence be ween he wo egions is he
same, which ob iously does no con o m o eali y. In eali y, he
impac o egions wi h la ge economic ou pu on egions wi h small
economic ou pu is g ea e han ha o egions wi h small economic
ou pu on egions wi h la ge economic ou pu . The e o e, we e e -
ence Newey and Powell (1987), cons uc ing asymme ic leas -
squa es es ima o s, ew i ing he po en ial g a i a ional ene gy o -
mula, and es ablishing an asymme ic spa ial weigh . The specific
o ms a e as ollows:
wij ¼
Imi
‾>mj
‾
ðÞ
¢
mi ¢mj
dij
ð1Þ
whe e wij is he influence o spa ial uni jon ia ime in he i h ow
o he spa ial weigh ma ix. To main ain gene ali y, he cons an Gin
he o iginal po en ial g a i a ional ene gy o mula is se o 1. j
Iðmi
‾>mj
‾Þjcan be called he asymme ic coe ficien , whe e
¼maxðmi
‾;mj
‾Þ=ðmi
‾þmj
‾Þ, and mi
‾and mj
‾ espec i ely ep e-
sen he means o miand mjo e he en i e ime in e al. Iðmi
‾>mj
‾Þ
o an indica i e unc ion, when he index condi ion is me , he alue
o he unc ion is 1, o he wise i is 0.
Ve i ying he supe io i y o he asymme ic po en ial g a i a-
ional ene gy spa ial weigh p oposed in his s udy equi es
compa ison wi h adi ional spa ial weigh s cons uc ed based
on he ecip ocal o economic dis ance (1=jmimjj)andgeo-
g aphical dis ance (1=dij). Table 1 es ima es he alues o he Akaike
in o ma ion c i e ion (AIC) and Bayesian in o ma ion c i e ion (BIC)
co esponding o he h ee weigh s based on he di ec e ec s
model in Eq. (2). The esul s e eal ha he AIC and BIC alues o
he asymme ic po en ial g a i a ional ene gy spa ial weigh a e
he smalles , indica ing ha hey a e supe io o he adi ional spa-
ial weigh ma ix cons uc ed based on geog aphic and economic
dis ances.
H. Xia, H. Yu, S. Wang e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100505
4
Di ec e ec s model
Acco ding o he MGWPR−SDM, we cons uc he ollowing
global smoo h SDM:
Gapi ¼a0þ 1WGapi þa1Digii þS
kakXk;i þS
λ C ;i
þS
quqWXq;i þuiþn þei K≥2ðÞ ð2Þ
whe e Gapi ep esen s he u ban− u al income gap, WGap
i
ep esen s
he spa ial lag e ms o he dependen a iable. Digii ep esen s digi al
economy de elopmen , Xk ep esen s he k- h media ing a iable, and
C ep esen s he - h con ol a iable. The co esponding coe ficien s
o media ing and con ol a iables a e akand λ ; espec i ely;WXq
ep esen s he spa ial lag e ms o he explana o y a iables, including
co e, media ing, and con ol a iables; he e o e, q=k+ +1.uq ep e-
sen s he coe ficien o he spa ial lag e m o he q- h explana o y a i-
able, ui ep esen s p o incial fixed e ec s, n ep esen s yea fixed
e ec s, and ei ep esen s he andomized pe u ba ion e m. All coe fi-
cien s in he abo e o mula can be fixed o a iable and equi e es ing
o de e mine he specific o mo heMGWPR−SDM.
Media ing e ec model
P e ious esea ch has demons a ed ha human capi al exe s a
decisi e influence on upg ading indus ial s uc u e (Romalis, 2004;
Hausmann e al., 2007). Chen and Yang (2014) de e mined ha
human capi al significan ly con ibu es o upg ading China’s indus-
ial s uc u e. Addi ionally, egional di e ences a e e iden conce n-
ing he impac o human capi al he e ogenei y on indus ial s uc u e
upg ading. High skilled human capi al significan ly p omo es he
upg ading o indus ial s uc u e in he eas e n egion, while i s p o-
mo ion e ec is no significan in he wes e n egion (Zhang e al.,
2011). To u he analyze he impac mechanism o he digi al econ-
omy on he u ban− u al income gap, we e e ence Yu and Zhu
(2023) cons uc ing a chain-media ing e ec model.
Hci ¼b0þ 2WHci þb1Digii þX
λ C ;i þX
q
uqWXq;i þui
þn þei ð3Þ
S ui ¼g0þ 3WS ui þg1Digii þg2iHci þX
λ C ;i
þX
q
uqWXq;i þuiþn þei ð4Þ
whe e Hci and S ui ep esen he media ing a iables, Hci quan ifies
human capi al, and S ui ep esen s he indus ial s uc u e. Cis he
ma ix o con ol a iables, and i s co esponding coe ficien ma ix is
λ;WXinEqs. (3) and (4) deno e he ma ix o spa ially lagged a iables
o explana o y a iables o he han he spa ially lagged e ms o he
explana o y a iables, and u ep esen s he co esponding coe ficien
ma ix; ui ep esen s p o incial fixed e ec s; n ep esen s yea fixed
e ec s; and ei ep esen s he andomized pe u ba ion e m.
Fig. 1 shows ha he indi ec impac o digi al economy de elop-
men (Digi) on he u ban− u al income gap (Gap) has h ee pa hs. (1)
Digi has an impac on he Gap by a ec ing human capi al (Hc), wi h
an e ec size o b1*a2. (2) Digi a ec s he Gap by influencing he
indus ial s uc u e (S u), wi h an e ec size o g1*a3. (3) Digi has an
impac on he S u by influencing Hc, which a ec s he Gap, wi h a
magni ude o b1*g2*a3.
Va iable selec ion and da a desc ip ion
Va iable selec ion
Explained a iable
Gap is he explained a iable in his s udy, e e encing Wang and
Xiao (2021), who adop ed he pe capi a disposable income a io o
u ban and u al esiden s as an al e na i e indica o o he income
gap be ween u ban and u al esiden s.
Co e explana o y a iable
Digi is he co e explana o y a iable. Cu en ly, no common
au ho i a i e indica o has been de eloped o measu ing digi al
economy de elopmen . The measu emen o he digi al economy
de elopmen index in he exis ing li e a u e gene ally includes he
“In e ne plus”digi al economy index designed by he Tencen
Resea ch Ins i u e (Jiang & Sun, 2020) and sel −cons uc ed indica-
o s. Al hough he “In e ne plus”digi al economy indica o s co e a
ela i ely comp ehensi e ange, due o he dynamic adjus men o
subdi ision indica o s and weigh s e e y yea , his o ical da a is
incompa able and can only be used as c oss−sec ional da a, which
canno be applied o panel da a. Re e ing o Yu and Zhu (2023),Guo
e al. (2020),Pan e al. (2021), and Wang e al. (2023), and conside -
ing he da a a ailabili y, his s udy cons uc s a China P o incial digi-
al economy index ha includes digi al in as uc u e, digi al
indus y de elopmen , and digi al financial se ices. The model
includes 3 p ima y indica o s, 7 seconda y indica o s, and 53 e ia y
indica o s, as shown in Table 2. D awing on Wang e al. (2021), he
en opy me hod is used o de e mine he weigh s o he indica o s,
and a comp ehensi e sco e is calcula ed using he weigh ing unc ion
me hod o indica e he le el o digi al economy de elopmen .
Media ing a iables
(1) Hc: We measu e pe capi a educa ion yea s, e e encing Li and Li
(2022a).Hu and Lu (2019) sugges ed ha a me s can enhance
hei lea ning capabili ies h ough in e ne in o ma ion echnol-
ogy, which is beneficial o imp o ing o e all Hc and inc easing
u al incomes.
(2) S u: Re e encing Wang and Xiao (2021), we use he a io o he
ou pu alue o he e ia y indus y o ha o he seconda y
indus y o measu e S u. The se ice-o ien ed economic s uc u e
d i en by in o ma ion echnology is an essen ial ea u e o S u
upg ading as he g ow h a e o he e ia y indus y is as e han
ha o he seconda y indus y, which is a ypical ac in he p o-
cess o China’s se ice-o ien ed economic ansi ion (Wu, 2005).
Table 1
Compa ison o fi ing e ec s o h ee weigh
ma ices.
Wg a Wdis Wecon
AIC 11,712.28 12,624.54 12,292.91
BIC 30,312.52 31,224.78 30,893.15
No es: Wg a ep esen s he asymme ic g a i a-
ional po en ial ene gy spa ial weigh ma ix p o-
posed in his a icle, Wdis and Wecon ep esen
spa ial weigh ma ices cons uc ed based on geo-
g aphic dis ance and economic dis ance, espec-
i ely.
Fig. 1. Diag am o he chain media ing e ec model.
H. Xia, H. Yu, S. Wang e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100505
5
As a measu e o indus ial s uc u e upg ading, he a io o he
ou pu alue o he e ia y indus y o ha o he seconda y
indus y clea ly eflec s he se ice−o ien ed endencies o he
economic s uc u e. I he a io is upwa d, he economy is ad anc-
ing in a se ice−o ien ed di ec ion, and he indus ial s uc u e is
upg aded (Gan e al., 2011). Ji (2023) showed ha al hough he
de elopmen o an ad anced indus ial s uc u e will no di ec ly
ha e a significan widening e ec on he u ban− u al income gap,
i will ha e a ce ain e e se egula o y e ec on he p ocess o
digi al economy de elopmen , a ec ing he u ban− u al income
gap.
Con ol a iables
(1) We quan i y economic de elopmen (Pgdp) using he pe cap-
i a GDP o each egion. (2) The s udy measu es u baniza ion le el
(U ban) as he a io o he numbe o pe manen u ban esiden s a
he end o he yea o he o al popula ion. (3) We measu e he
deg ee o openness o he ou side wo ld (Open) using he p opo ion
o o al impo s and expo s o GDP. (4) T anspo a ion de elopmen
(T a ) is measu ed as a e age oad mileage, which is he o al oad
mileage/land a ea o each p o ince, e e encing Wang e al. (2021).
(5) The s udy quan ifies ag icul u al mechaniza ion (Ag i_ a e) as he
a io o he o al machine y inpu in each egion o he numbe o
u al esiden s. (6) We measu e fiscal suppo o ag icul u e (Sub_-
a e) as he p opo ion o fiscal expendi u e on ag icul u e, o es y,
and wa e o he o al fiscal expendi u e a he co esponding le el.
Da a desc ip ion and explana ion
Table 3 p esen s he s a is ical indica o s, including minimum
alue, in e qua ile, median, mean, in e qua ile, maximum alue,
and s anda d de ia ion o a iables. The minimum alue o Gap is
1.840 and he maximum alue is 3.800, indica ing significan di e -
ences in Gap ac oss China’s egions. The minimum alue Digi is 0.015
and he maximum alue is 0.509, which also indica es a digi al di ide
in digi al economy de elopmen in a ious egions o China.
Da a ela ed o e−comme ce ha e only been a ailable since 2013
and do no include da a om Hong Kong, Macao, Taiwan, and Xizang.
The e o e, his s udy uses panel da a om 30 p o inces in China om
2013 o 2021 o empi ical analysis. The da a a e sou ced om he
“China Popula ion and Employmen S a is ical Yea book” om 2014
o 2022, he o ficial websi e o he Na ional Bu eau o S a is ics, and
he “Peking Uni e si y Digi al Inclusi e Finance Index Repo ”pub-
lished by he Peking Uni e si y Digi al Finance Resea ch Cen e .
Empi ical esul s analysis
Model selec ion
We fi s de e mine whe he a spa ial econome ic model is
equi ed. Mo an’s Index (Mo an’s I) is an impo an indica o o
measu ing spa ial co ela ion, and uses he Mo an’s I s a is ic o es
whe he sample indi iduals a e spa ially independen o one ano he .
The Mo an’s I es esul s in Table 4 ha when Gap, Hc, and S u a e
used as explana o y a iables, he p- alues co esponding o he
Mo an’s I s a is ic a e all 0.000, which significan ly ejec s he null
hypo hesis ha he a iables do no ha e spa ial co ela ion a he
5 % le el. All h ee a iables ha e significan spa ial au oco ela ion;
he e o e, a spa ial lag e m should be added o he model as an
explana o y a iable, employing a spa ial econome ic model.
Table 2
Indica o Sys em o he De elopmen Le el o he Digi al Economy.
P ima y indica o s Seconda y
indica o s
Te ia y indica o
Digi al
In as uc u e
T adi ional digi al
in as uc u e
In e ne b oadband access po s
Leng h o fibe op ic lines
Numbe o In e ne domain
names
Numbe o In e ne si es
Numbe o In e ne pages
New digi al
in as uc u e
Local exchange capaci y
Mobile In e ne access a fic
Numbe o cell phone base
s a ions
Ip 4/Ip 6 add ess numbe
Digi al indus y
de elopmen
digi al
indus ializa ion
Re enue om elecommunica-
ion se ices
Mobile Phone P oduc ion
IC p oduc ion
P oduc ion o mic ocompu ing
equipmen
Re enue om so wa e
ope a ions
Re enue om in o ma ion ech-
nology se ices
Pe cen age o In o ma ion Tech-
nology Employees
Indus ial
digi iza ion
exp ess deli e y olume
E−comme ce ansac ion alue
Numbe o websi es pe 100
en e p ises
Pe cen age o en e p ises wi h e
−comme ce ading ac i i ies
Digi al financial
se ices
B ead h o co e age Re e o Guo e al. (2020)
Dep h o applica ion Re e o Guo e al. (2020)
Digi iza ion le el Re e o Guo e al. (2020)
Table 3
Desc ip i e s a is ics o a iables.
Min Q1 Median Mean Q3 Max SD
Gap 1.840 2.293 2.485 2.549 2.758 3.800 0.383
Digi 0.015 0.039 0.068 0.098 0.126 0.509 0.087
Hc 7.474 8.824 9.253 9.348 9.618 12.681 0.906
S u 0.570 0.940 1.150 1.324 1.380 5.300 0.717
Pgdp 2.315 4.185 5.398 6.223 7.213 18.398 2.943
U ban 0.379 0.530 0.593 0.609 0.665 0.896 0.115
Open 0.010 0.090 0.140 0.241 0.308 1.270 0.244
T a 0.097 0.567 0.946 0.986 1.336 2.237 0.523
Ag i_ a e 0.330 1.170 1.610 1.806 2.220 6.450 0.961
Sub_ a e 0.040 0.093 0.120 0.116 0.140 0.200 0.034
No e: Q1 and Q3 ep esen he 25 h and 75 h pe cen iles, espec i ely, wi h SD
being he abb e ia ion o s anda d de ia ion.
Table 4
Mo ans’I es esul s.
Gap Hc S u
2013 0.100 0.063 0.009
(0.000) (0.000) (0.000)
2014 0.084 0.056 0.007
(0.000) (0.000) (0.000)
2015 0.087 0.052 0.008
(0.000) (0.000) (0.000)
2016 0.087 0.052 0.015
(0.000) (0.000) (0.000)
2017 0.086 0.041 0.016
(0.000) (0.000) (0.000)
2018 0.085 0.041 0.020
(0.000) (0.000) (0.000)
2019 0.084 0.044 0.007
(0.000) (0.000) (0.000)
2020 0.086 0.049 0.007
(0.000) (0.000) (0.000)
2021 0.087 0.049 0.011
(0.000) (0.000) (0.000)
The p− alues co esponding o Mo ans’I
s a is ic a e in pa en heses.
H. Xia, H. Yu, S. Wang e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100505
6
Second, spa ial econome ic models ha e a ious o ms such as
he spa ial au o eg ession (SAR), he spa ial e o model (SEM), and
he SDM. Model selec ion equi es app op ia e es ing. This s udy
uses he likelihood a io app oach o es whe he he SDM degene -
a es in o SAR o SEM, and Table 5 summa izes he esul s. The p- al-
ues co esponding o he chi-squa e s a is ics o Models (1)−(3) a e
all less han 5 %, indica ing ha all h ee models ejec he o iginal
assump ion ha SAR and SEM models a e ue; he e o e, he SDM
wi h bidi ec ional fixed e ec s should be chosen.
Va iable selec ion: spa ial a ying o fixed coe ficien a iable
As some a iable coe ficien s may occu in Models (1)−(3), i is
necessa y o iden i y he coe ficien s ha change wi h spa ial a ia-
ion h ough es ing. Rega ding he a iable selec ion p oblem in he
mixed geog aphic weigh ed eg ession model, Mei e al. (2016) p o-
posed applying a boo s ap me hod o a iable coe ficien selec ion,
which is mo e obus han he F- es p oposed by B unsdon e al.
(1999).Table 6 p esen s he es esul s ob ained using his me hod.
Combining he esul s o a iable selec ion in Table 6 and com-
pa ed wi h he model in Equa ion (11) om Yu e al. (2021) e eals
ha Models (1) and (2) ha e some explana o y a iables wi h insig-
nifican boo s ap es s, and he co esponding p- alues a e mo e sig-
nifican han 0.05, indica ing ha hey a e fixed coe ficien s and he
o he explana o y a iables a e a iable coe ficien s. This confi ms
sui abili y o employing he MGWPR−SDM because he spa ial lag
e m WY o he explana o y a iables in Model (1) is a a iable coe fi-
cien , and sui able o applying he MGWPR−SDM (1, kc, k ), while
he spa ial lag e m WY o he explana o y a iables in Model (2) is a
fixed coe ficien , which is sui able o applying he MGWPR−SDM (0,
kc, k ). Simila ly, he p- alues co esponding o he explana o y a i-
ables in Model (3) a e less han 0.05, indica ing ha all he
explana o y a iables, including WY, a e a iable coe ficien s; he e-
o e, he MGWPR−SDM (0, 0, k ) is applicable.
Va ying coe ficien eg ession esul s
As his s udy ocuses on he co e explana o y a iable Digi and he
media ing a iables o Hc and S u, we do no in e p e he emaining
con ol a iables. Table 7 e eals ha only S u in Model (1) has a
fixed coe ficien ha is s a is ically significan ly posi i e a he 1 %
le el, sugges ing ha upg ading he S u will widen he Gap, and i s
impac a ies less ac oss egions. The S u o neighbo ing egions
(w_S u) in Model (1) and he co e explana o y a iable Digi, he
media ing a iable Hc, and S u in he o he wo models a e a iable
coe ficien s (Table 7).
Va ying coe ficien eg ession esul s
Table 8 p esen s he media ion esul s o Model (1), e ealing
ha he coe ficien s o Digi a e bo h posi i e and nega i e, wi h a
maximum alue o 0.094 and a minimum alue o 0.145, indica ing
ha digi al economy de elopmen widens he Gap in some egions,
while he opposi e is ue o o he egions. Simila ly, he coe ficien s
o Hc a e bo h posi i e and nega i e, sugges ing ha he enhance-
men o human capi al in some egions educes he Gap, while he
opposi e is ue o o he egions.
Table 9 p esen s he esul s o media ion Model (2), e ealing ha
he coe ficien s o Digi a e bo h posi i e and nega i e, wi h a maxi-
mum alue o 1.698 and a minimum alue o 1.196, indica ing ha
digi al economy de elopmen in some egions enhances human capi-
al, while he opposi e is ue o o he egions. Fu he mo e, digi al
economy de elopmen in he neighbo ing egions (w_Digi) also
exhibi s bo h a posi i e and nega i e impac on he egion’s human
capi al, e ealing ha some neighbo ing egions can p omo e human
capi al imp o emen , whe eas o he s ha e he opposi e e ec .
Table 5
LR es esul s.
Model (1) Model (2) Model (3)
SDM o SAR 27.57 75.76 24.29
(0.002) (0.000) (0.004)
SDM o SEM 31.32 61.07 30.65
(0.000) (0.000) (0.000)
Table 6
Boo s ap es esul s (P− alue).
Va iables Model (1) Model (2) Model (3)
In e cep 0.051 0 0.001
WY 0.008 0.088 0
Digi 0.002 0.001 0
Hc 0.033 −0
S u 0.24 −−
Pgdp 0.194 0.004 0
U ban 0.062 0 0.002
Open 0.008 0 0
T a 0.032 0 0
Ag i_ a e 0.027 0 0
Sub_ a e 0.051 0.002 0.011
w_Digi 000
w_Hc 0.025 −0.004
w_S u 0.002 −−
w_Pgdp 0.148 0.007 0
w_U ban 0.109 0 0
w_Open 000
w_T a 0.01 0 0
w_Ag i_ a e 0.025 0.004 0
w_Sub_ a e 0.033 0 0
WY is he spa ial lag e m o he explained a iables o
each model, he a iable s a ing wi h “w_” ep esen s
he spa ial lag e m o he a iable, and “-” ep esen s he
missing alue. Same as he ollowing ables.
Table 7
Fixed coe ficien esul s.
Va iables Model (1) Model (2)
WY y−0.769***
(0.003)
Digi yy
Hc y−
S u 0.164*** −
(0.000))
Pgdp −0.01 y
(0.310)
U ban −3.231*** y
(0.000)
Open yy
T a yy
Ag i_ a e yy
Sub_ a e 0.043 y
(0.938)
w_Digi yy
w_Hc y−
w_S u y−
w_Pgdp −0.105 y
(0.123)
w_U ban −2.697 y
(0.335)
w_Open yy
w_T a yy
w_Ag i_ a e yy
w_Sub_ a e yy
R
2
0.304 0.038
“y”indica es a ying coe ficien a iables.
“−” ep esen s he missing alue. *, **, and
*** indica e s a is ically significance a he
0.1, 0.05, and 0.01 le els, espec i ely, wi h
s anda d de ia ions in pa en heses.
H. Xia, H. Yu, S. Wang e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100505
7
Table 10 p esen s he media ion esul s o Model (3), e ealing
ha he coe ficien s o Digi a e bo h posi i e and nega i e, wi h a
maximum alue o 0.130 and a minimum alue o 0.303. This indi-
ca es ha digi al economy de elopmen in some egions can p omo e
he S u in eg a ion and upg ading and he opposi e occu s in o he
egions. Fu he mo e, he impac o he digi al economy de elop-
men in neighbo ing egions (w_Digi)onS u is posi i e and nega i e.
Digi in neighbo ing egions (w_Digi) has posi i e and nega i e e ec s
on he local S u; ha is, some neighbo ing egions p omo e local
S u upg ading, whe eas o he s ha e he opposi e e ec . The coe fi-
cien o Hc is bo h posi i e and nega i e, wi h a maximum alue o
0.058 and a minimum alue o 0.055, indica ing ha enhancing
human capi al in some egions p omo es S u in eg a ion and
upg ading, while he opposi e is ue in o he egions. The influence
o Hc in neighbo ing egions (w_Hc)onS u upg ading is also bo h
posi i e and nega i e; ha is, some neighbo ing egions p omo e
S u upg ading, while o he s do no .
Media ing e ec analysis
To clea ly illus a e he spa ial he e ogenei y o he mechanism o
impac o Digi on Gap, his s udy p esen s each media ing e ec in
he o m o maps. The le sides o Figs. 2−4p esen he e ec alues
and he igh sides show he p- alues co esponding o he Sobel
(1982) es .
Fig. 2 e eals significan spa ial he e ogenei y in Digi’s e ec on
Gap h ough Pa hway 1. The coe ficien s o Jilin, Inne Mongolia,
Shanxi, Hubei, Guangdong, and o he egions a e posi i e a a 10 %
significance le el, indica ing ha Digi widens he Gap by enhancing
Hc. The coe ficien s o Fujian, Chongqing, and o he egions a e
nega i e, indica ing ha Digi can e ec i ely na ow he Gap by
enhancing Hc. Mos egions in cen al China a e posi i e, and no h-
wes e n egions and Heilongjiang and Liaoning a e nega i e, bu
none o hem a e significan .
Fig. 3 e eals ha Pa hway 2 also exhibi s significan spa ial he -
e ogenei y. The coe ficien s in mos egions o eas , cen al−wes ,
and no heas China a e nega i e a a 10 % significance le el, wi h
Jiangxi and Hubei ha ing he la ges e ec s. This indica es ha Digi
has na owed he Gap in hese egions by upg ading he indus ial
s uc u e. In con as , Beijing, Shanxi, Ningxia, Inne Mongolia, and
o he egions exhibi posi i e coe ficien s, indica ing ha Digi has
widened he Gap in hese a eas by upg ading he indus ial s uc u e.
Fig. 4 illus a es he magni ude and significance o he chain-
media ed e ec s, indica ing ha Digi influences he Gap by al e ing
Hc, which subsequen ly a ec s S u. The coe ficien s in egions such
as Jilin, Tianjin, Hubei, and Guangxi a e posi i e a he 10 % signifi-
cance le el, whe eas egions such as Shanxi, Inne Mongolia, Henan,
and Fujian exhibi nega i e coe ficien s. The coe ficien s in o he
egions ha e bo h posi i e and nega i e alues, bu a e no s a is i-
cally significan .
Di ec and o al e ec s analyses
In addi ion o examining he indi ec e ec s media ing a iables
such as Hc and S u on he Gap h ough, Digi also di ec ly impac s he
Gap.Fig. 5 p esen s he di ec e ec s and hei significance, e ealing
ha he Digi in Jiangxi, Chongqing, and Shaanxi has significan ly wid-
ened he u ban− u al income gap. In con as , he Digi in Jilin, Shan-
dong, Zhejiang, Yunnan, Hebei, Sichuan, Guizhou, Liaoning, Hainan,
and Xinjiang significan ly na owed he Gap. Digi had no significan
impac on he Gap in o he a eas.
The analysis abo e e eals ha he h ee indi ec pa hways
h ough which Digi a ec s he Gap exhibi significan spa ial he e o-
genei y, wi h bo h posi i e and nega i e e ec s. These e ec s end o
o se each o he , making i necessa y o in es iga e he weigh o he
h ee media ing e ec s o examine o e all indi ec impac .
Fig. 6 p esen s he weigh s o he indi ec and o al e ec s. The le
g aph in Fig. 6 shows ha he o al media ing e ec o Digi on he Gap
is nega i e in mos egions. Only 12 p o inces and ci ies exhibi a
posi i e o al media ing e ec , and Jilin P o ince in he no heas had
he la ges e ec , ollowed by Beijing, Inne Mongolia, and Shanxi in
no h China; Henan, Hubei, and Jiangxi in he cen al egion; and
Shandong, Jiangsu, Guangdong, and Hainan in he eas coas a ea,
wi h ela i ely smalle e ec s. The g aph on he igh o Fig. 6
p esen s he spa ial dis ibu ion o he o al e ec s a e combining
Table 8
Va ying coe ficien esul s o model (1).
Va iables Min 1s Qu Median 3 d Qu Max
WY −0.794 −0.020 0.245 0.511 1.683
Digi −0.145 −0.078 −0.020 0.014 0.094
Hc −0.938 −0.115 −0.011 0.107 0.477
Open −0.244 0.087 0.153 0.193 0.310
T a −0.073 −0.046 −0.012 −0.003 0.081
Ag i_ a e −2.493 0.465 1.013 2.632 6.261
w_Digi −0.034 0.294 0.450 0.637 1.011
w_Hc −0.569 −0.345 −0.157 0.009 0.433
w_S u −2.729 −0.597 −0.401 0.066 0.913
w_Open 0.551 0.809 0.971 1.111 1.778
w_T a −0.509 −0.151 −0.053 0.053 0.365
w_Ag i_ a e −9.667 −4.872 −2.228 −0.391 7.486
w_Sub_ a e −0.747 0.215 0.456 0.751 1.251
R
2
0.009 0.083 0.104 0.118 0.148
Table 9
Va ying coe ficien esul s o model (2).
Va iables Min 1s Qu Median 3 d Qu Max
Digi −1.196 −0.564 −0.171 0.074 1.698
Pgdp −0.055 −0.035 0.013 0.003 0.055
U ban −4.244 −2.189 −1.551 −0.730 3.913
Open −1.032 −0.669 −0.331 −0.179 0.355
T a −0.342 −0.040 0.117 0.331 0.620
Ag i_ a e −0.078 −0.020 0.018 0.041 0.127
Sub_ a e −3.055 −0.217 0.887 1.521 4.514
w_Digi −3.141 0.177 2.901 6.091 14.593
w_Pgdp −0.612 −0.221 −0.154 −0.061 0.068
w_U ban −1.531 12.794 15.895 18.866 22.850
w_Open −6.435 −4.545 −3.697 −2.831 1.283
w_T a −1.581 −0.681 −0.325 0.205 1.365
w_Ag i_ a e −1.290 −0.726 −0.369 −0.200 0.183
w_Sub_ a e −32.820 −8.570 −3.944 1.607 13.224
R
2
0.070 0.108 0.138 0.154 0.199
Table 10
Va ying coe ficien esul s o media ion model (3).
Va iables Min 1s Qu Median 3 d Qu Max
WY −2.967 −0.999 −0.363 0.081 0.836
Digi −0.303 −0.107 −0.076 −0.019 0.130
Hc −0.055 −0.030 −0.018 0.002 0.058
Pgdp −10.649 −8.801 −6.507 −5.673 −3.638
U ban −0.652 −0.074 0.130 0.382 0.639
Open 0.091 0.290 0.405 0.546 0.779
T a −0.170 −0.034 0.019 0.085 0.161
Ag i_ a e 0.896 1.879 3.138 3.830 4.315
Sub_ a e −12.055 −1.031 0.523 4.030 9.046
w_Digi −0.467 0.551 0.947 1.204 2.245
w_Hc −0.301 −0.226 −0.098 −0.041 0.277
w_Pgdp −14.957 −6.947 −3.365 −1.772 5.487
w_U ban −1.732 −0.160 1.288 1.986 3.958
w_Open −1.099 0.279 0.619 1.087 1.868
w_T a −1.120 −0.673 −0.450 −0.270 0.199
w_Ag i_ a e −10.407 3.848 7.774 14.595 22.178
w_Sub_ a e −2.848 −1.517 −1.047 −0.699 0.151
R
2
0.206 0.289 0.318 0.332 0.365
H. Xia, H. Yu, S. Wang e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100505
8
