Lu, Zhaoying
A icle
Human capi al spillo e s om special economic zones:
E idence om Yang ze Del a in China
Economies
P o ided in Coope a ion wi h:
MDPI – Mul idisciplina y Digi al Publishing Ins i u e, Basel
Sugges ed Ci a ion: Lu, Zhaoying (2022) : Human capi al spillo e s om special economic zones:
E idence om Yang ze Del a in China, Economies, ISSN 2227-7099, MDPI, Basel, Vol. 10, Iss. 5, pp.
1-15,
h ps://doi.o g/10.3390/economies10050099
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/328399
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/
Ci a ion: Lu, Z.. 2022. Human capi al
spillo e s om Special Economic
Zones: E idence om Yang ze Del a
in China. Economies 10: 99.
h ps://doi.o g/10.3390/
economies10050099
Academic Edi o s: Helmi Hamdi and
Ral Fendel
Recei ed: 8 Feb ua y 2022
Accep ed: 16 Ma ch 2022
Published: 21 Ap il 2022
Publishe ’s No e: MDPI s ays neu al
wi h ega d o ju isdic ional claims in
published maps and ins i u ional a il-
ia ions.
Copy igh : © 2022 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license (h ps://
c ea i ecommons.o g/licenses/by/
4.0/).
economies
A icle
Human Capi al Spillo e s om Special Economic Zones:
E idence om Yang ze Del a in China
Zhaoying Lu
G adua e School o Economics, Osaka Uni e si y, Osaka 560-0043, Japan; [email p o ec ed]
Abs ac :
This pape e alua es he e ec s o a place-based p og am in he Yang ze Del a o
China—Special Economic Zones (SEZs). Taking in o accoun spa ial p oximi y, his pape quan i ies
he spillo e e ec s o he human capi al in SEZs. One majo inding is ha egional p oduc i i y
bene i s om he human capi al in SEZs. The spillo e e ec s a e no only con ined o hei own
coun ies, bu also neighbo ing coun ies. SEZs con ibu e mo e o he egional p oduc i i y o neigh-
bo ing coun ies han he one o he hos ing coun y i sel . Mo eo e , posi i e spillo e e ec s o he
human capi al in SEZs s ill hold o he g ow h o egional p oduc i i y.
Keywo ds: special economic zone; spa ial analysis; human capi al; spillo e s; Yang ze Del a
1. In oduc ion
A g owing numbe o “place-based” policies a e implemen ed in geog aphically
a ge ed a eas o imp o e economic pe o mances Neuma k and Simpson (2015). Fo
ins ance, En e p ise Zone p og ams in he US and some Eu opean coun ies, and Special
Economic Zone p og ams in de eloping coun ies. Such policies a e h own in o doub as o
whe he hey a e simply a o m o ealloca ing esou ces o a bi age bene i s Glaese and
Go lieb (2008); G eens one and Looney (2010). On he o he hand, some li e a u e suppo s
implemen ing he policies and p o ides a ionales o hem Kline and Mo e i (2014).
The ques ion a ises whe he he a ionales o he policies in de eloped coun ies
such as knowledge spillo e s s ill hold o de eloping coun ies. Knowledge spillo e s
a e expec ed o p oduce posi i e e ec s on he local economy, mo e speci ically, b inging
addi ional people who ha e mo e skills can imp o e he o he ’s p oduc i i y h ough
sha ing o knowledge and as e echnology adop ion Mo e i (2011). Knowledge spillo e s
om addi ional human capi al in a ge ed a eas a e c ucial o policy-make s o unde s and
he e iciency o he policies, especially o de eloping coun ies ha ha e cons ain
esou ces o alloca e.
Among de eloping coun ies, his pape ocuses on place-based policies in he China–
Special Economic Zone (SEZ) p og am. SEZ p og ams a e usually in ended o inc ease
in es men s, and expo ea nings and go e nmen e enues. The objec i e o SEZs in China
includes gene a ing demons a ion e ec s and imp o ing egional p oduc i i y. China
has been implemen ing SEZ p og ams as a i al pa o Chinese economic e o ms since
1978. In ecen decades, he go e nmen has cos a huge amoun o money and alloca ed
millions o ac es o se up SEZs o egional de elopmen s. Empi ical li e a u e inds ha
implemen ing he policies has posi i e impac s o he i ms loca ed wi hin SEZs o he
ci ies hos ing he SEZ p og am Wang (2013); Zheng e al. (2017). Howe e , ew a emp s
we e made o s udy he ex en o geog aphy di usion o he impac s, assuming ha he
spillo e s o SEZs a e only limi ed o hei conside ed egions.
The ela ion o knowledge spillo e s and egional g ow h is in es iga ed in a ge-
og aphical dimension Dö ing and Schnellenbach (2006). A common mechanism o a
ans e o knowledge is he mobili y o indi iduals and he ade o goods, which ca y
p oduc ion- ela ed knowledge wi h hem Ma usik and Hill (1998), and no ably, would also
Economies 2022,10, 99. h ps://doi.o g/10.3390/economies10050099 h ps://www.mdpi.com/jou nal/economies
Economies 2022,10, 99 2 o 15
allow a spilling o e o embodied knowledge. Human capi al e e s o an employee’s skills,
educa ion, and o he implied alues, which is one o he pa s o knowledge spillo e s.
He ea e , he impac s o human capi al a e de ined as human capi al spillo e s.
Human capi al spillo e s a e ega ded as ca alys s o he de elopmen o a ci y Lucas
(1988). Mo e gene ally, he human capi al o o he s in close p oximi y can aise e e yone’s
human capi al and inc ease i m p oduc i i y h ough sha ing o knowledge and as e
adop ion o inno a ions Mo e i (2011). Lo s o li e a u e measu es human capi al by
he educa ional le el, such as he sha e o wo ke s wi h a college deg ee o compa able
educa ion, because posi i e spillo e s a e mo e likely o di use om mo e highly educa ed
wo ke s, due o he knowledge hey possess and pe haps he wo k hey do. Human capi al
spillo e s do no only occu wi hin one pa icula egion, bu also ans e o neighbo ing
ones Fische e al. (2009); López-Bazo e al. (2004). Ramos e al. (2010) in es iga e he
e ec s o human capi al spillo e s on egional p oduc i i y and g ow h, no only o he
conside ed egion, bu also o he neighbo ing ones. Bal agi e al. (2016) es ima e i m-le el
p oduc i i y spillo e s in China’s chemical indus y conside ing he neighbo ’s skilled
labo a io.
In China, he i ms in SEZs a e encou aged o make connec ions wi h local i ms. Wi h
suppo om human capi al o SEZs, local i ms upda e he in o ma ion on ma ke s and
ha e as e adop ion o new echnology. In his way, human capi al o SEZs may ha e
posi i e impac s on he local economy. The human capi al o SEZs is no equal, howe e . In
his pape , we assume ha posi i e impac s a e mo e likely o di use om highly educa ed
wo ke s. In addi ion, we suppose ha he impac s o human capi al may no occu wi hin
one pa icula egion, bu also ans e o neighbo ing ones. In he con ex o he SEZ
p og am, we need o conside spillo e e ec s beyond he a ge ed zones. Some li e a u e
e alua es he impac s o place-based policies in he u he scope Rosen hal and S ange
(2008); Zheng e al. (2017). As a de elopmen policy widely accep ed in many coun ies,
he ex e nali ies om a ge ed zones o he local economy should also be included o
policy-make s o do a cos -bene i analysis. Co esponding p e e en ial policies such as ax
deduc ion and subsidy a e p o ided by he go e nmen o a ac skilled labo in o de
o s imula e egional g ow h. Howe e , one may be conce ned ha en an s o a ge ed
zones ha ha e highe human capi al pu local i ms a disad an age. They may poach
away locally ained alen and c owd ou i ms ou side he zones in he same indus y.
This s udy aims o add ess wo impo an ques ions on human capi al spillo e s
om SEZs o he local economy. Fi s , does he human capi al o SEZs di use posi i e
ex e nali ies o he same egion? Second, will he human capi al o SEZs bene i neighbo ing
egions? Only conside ing he egions ha hos he SEZ p og am may o e es ima e
he magni ude o he ex e nali ies o SEZs. B inging addi ional skilled labo o SEZs is
suspec ed o be simply a o m o ealloca ing esou ces o ano he loca ion. On he con a y,
a ac ing i ms ha ha e ela i ely highe human capi al is likely o spill o e knowledge
o ad anced echnology o s imula e he p oduc i i y o local indus ies.
This pape builds on he esea ch es ima ing he human capi al spillo e s a he
egional le el. In many s udies, hey se he human capi al o a coun y as an explana o y
a iable and e eal he exis ence o human capi al ex e nali ies López-Bazo e al. (2004);
Ramos e al. (2010). To e alua e he spillo e s o SEZs o he local economy, we se he
human capi al o he i ms wi hin SEZs as an explana o y a iable. In addi ion, we ocus
on he human capi al spillo e s in he same indus y a he han ac oss all he indus ies,
supposing ha he obse a ion in p oximi y has high simila i y in he same indus y.
This pape builds on he spa ial econome ic li e a u e, which allows us o examine
spillo e s in h ee channels Elho s (2014); LeSage and Pace (2009). Speci ically, an inc ease
in he human capi al o a SEZ is likely o a ec i s own egion (di ec e ec ) and may
a ec he neighbo ing egions (indi ec e ec ). Mo eo e , spa ial eg ession can measu e
impac s passing h ough he neighbo ing egions and back o he egion i sel ( eedback
loops). These e ec s seek o in es iga e bo h in a- and in e - egional e ec s o human
capi al spillo e s on egional p oduc i i y. In he con ex o he SEZ p og am, i should be
Economies 2022,10, 99 3 o 15
examined ha whe he he ex e nali ies a ising om a SEZ ha e impac s on he geog aphi-
cal p oximal egions. Howe e , o he bes o he au ho ’s knowledge, exis ing li e a u e
ocuses on he impac s o SEZs in an in a- egion dimension ha SEZs exe in luences on
he hos ing egions in China. To ill his gap, he pape adds o he exis ing li e a u e by
conside ing spa ial p oximi y. Bo h in a- and in e - egional examina ions a e necessa y o
policymake s o e ise he cos -bene i analysis o he SEZ p og am.
In his pape , we answe he ques ion o whe he he ex e nali ies o human capi al in
SEZs a e able o aise local p oduc i i y. F om he empi ical esul s, we conclude ha he
human capi al o SEZs con ibu es o di using knowledge and hus imp o ing p oduc i i y
in he same egion. In addi ion, his pape u he add esses whe he he spillo e e ec s o
SEZs can bene i neighbo ing egions in spa ial p oximi y as well. The posi i e spillo e s
o wide geog aphic scope imply ha he ising p opo ion o college-educa ed wo ke s in
SEZs does no ha m he i ms in he p oximi y, ins ead, he d i ing e ec o human capi al
o acili a e local i ms o imp o e p oduc i i y is e iden .
The es o his pape is o ganized as ollows. Sec ion 2p esen s ela ed li e a u e
ha mainly ocuses on place-based policies in de eloping coun ies. Sec ion 3in o-
duces he backg ound o he SEZ p og am eleased in China. Sec ion 4desc ibes he
da a.
Sec ions 5and 6
p esen es ima ion models and empi ical esul s. The conclusion is
summa ized in he las sec ion.
2. Rela ed Li e a u e
Place-based policies in de eloped coun ies di e om hose in de eloping coun ies
on bo h policy goals and implemen measu es. To con as empi ical e idence o simila
ypes, his pape di ec ly ela es o place-based policies ope a ed in de eloping coun ies.
In de eloping coun ies, implemen ing place-based policies is a way o s imula e he
local economy, as well as o p omo e de elopmen wi hin he zones. Johansson and Nilsson
(1997) ind ha SEZs ha e posi i e e ec s on expo s in Hong Kong, Malaysia, Singapo e,
and S i Lanka. In pa icula , hey highligh he pe o mances o SEZs in Malaysia o
a ac ing o eign in es o s who ans e knowledge o local ma ke s. Wang (2013) and
Alde e al. (2016) ind ha SEZs exe posi i e in luences on he local economy wi h espec
o employmen , expo , and o eign di ec in es men in China. Since using mac o-le el
da a, hey e alua e he in luences on he whole SEZ-ope a ing coun ies, do no sepa a e
i ms in zones wi h hose ou side zones in a coun y. Zheng e al. (2017) and Lu e al. (2019)
dis inguish i ms loca ed wi hin SEZs and hose ou side zones by using geocoded i m-
le el da a. Zheng e al. (2017) ind ha in se e al majo ci ies o China SEZs ha e posi i e
spillo e s wi h espec o wage, employmen , and p oduc i i y o non-SEZ i ms loca ed
nea by SEZs. Ci˙
zkowicz e al. (2017) also ind ha implemen ing place-based policies
c ea es jobs o he i ms ou side zones in Poland.
Wi h ega ds o esea ch me hods, exis ing li e a u e s udying place-based policies
can be di ided in o wo ypes. One is li e a u e ha summa izes expe iences and a ionales
o implemen ing place-based policies based on case s udies and in e iews. Zeng (2011)
desc ibes expe iences in China and discusses keys o success and aced oppo uni ies.
Fa ole (2011) in oduces expe iences in A ica and possible easons o hose ou comes.
The o he one is li e a u e ha employs econome ic models o es ima e he e ec s o
place-based policies. Many empi ical s udies e alua e he impac s o place-based poli-
cies by a quasi-expe imen al app oach such as he di e ence-in-di e ence me hod (See
Lu e al. (2019); Zheng e al. (2017)). These s udies ha allow o compa isons o ea men
g oups and app op ia e con ol g oups make alid iden i ica ion o causal e ec s. The
quasi-expe imen al app oach is b oadly applied o in es iga e he impac s o place-based
policies in bo h de eloping and de eloped coun ies. Excep o his kind o app oach,
spa ial econome ic models a e used o assess he spillo e e ec s o place-based policies by
aking in o accoun he geog aphical p oximi y o he obse a ions Ci˙
zkowicz e al. (2017).
Economies 2022,10, 99 4 o 15
3. Backg ound
This sec ion in oduces he backg ound o SEZ p og ams implemen ed in China. SEZs
a e geog aphically designa ed a eas by he go e nmen ha aim o s imula e economic
g ow h in hei ju isdic ion. Gene ally, he e m “SEZ” is a zone ha includes some common
cha ac e is ics such as—i is loca ed in geog aphically designa ed a eas; i o e s bene i o
in es o s wi hin he zone; i has sepa a e cus oms a eas; and i has single adminis a ions
Zeng (2011).
The SEZ p og am has been implemen ed o e he las 40 yea s as a c ucial pa o
Chinese economic e o ms since 1978. Be o e he e o ms, China was an isola ed socialis
s a e domina ed by cen al planning. Ins ead o ca ying on p e ious policies, he c ea ion
o SEZs in he ini ial s age was an expe imen o es o he ma ke -o ien ed economy.
By 1980, he i s ou SEZs we e app o ed o se up in Shenzhen, Zhuhai, Shan ou, and
Xiamen. Since hese egions we e eas e n coas al a eas nea Hong Kong, Macao, and
Taiwan, he designa ion a ha ime ook in o accoun bo h economic and poli ical ac o s.
Wi h he success o hese egions, an addi ional 14 coas al egions we e designa ed as
SEZs in 1984 o gain u he access o o eign ma ke s and in es men . The nex wa e
o SEZs in he 1990s g adually began o ex end om eas e n coas lines owa ds inland
egions, especially a e Deng Xiaoping’s sou he n ou o SEZs in 1992. By 2001, SEZs
we e es ablished one a e ano he , co e ing all p o inces ac oss China. As o 2006, SEZs
numbe 1,568, occupying 9,949 ac es o land in o al Zheng e al. (2017).
By adminis a i e le els, SEZs a e di ided in o wo ca ego ies: na ional-le el and
p o incial-le el. The o me one is conduc ed by he cen al go e nmen , and he la e
one is di ec ed by he p o incial go e nmen s. Na ional-le el SEZs g an mo e au onomy
and enjoy mo e p i ileges han p o incial-le el SEZs. Only au ho ized by he cen al
go e nmen o he go e nmen s ha ha e p o incial adminis a i e le el a e legal SEZs
in China. The o he s ha we e no app o ed by he cen al o p o incial go e nmen s
iola ed ela ed laws and egula ions, and we e abolished in 2006.
Al hough he e a e some di e ences be ween speci ic p i ileges o na ional- and
p o incial-le el, SEZs a e all g an ed ma ke -o ien ed eedom and o e ed p e e en ial
bene i s. The SEZ p og am is gi en g ea e au onomy o adjus ela ed egula ions along
he basic lines o na ional ones by emo ing some cons ain s wi hin he scope o he zones.
Addi ionally, he go e nmen p o ides a se ies o p e e en ial policy packages o o eign
and domes ic in es o s, which en e he zones as he ollowing1:
1. In Gene al, he policy deduc s co po a e income ax a e o 15-24% o i ms in SEZs
ela i e o 33% o o dina y domes ic i ms ou side SEZs. Addi ionally, cus oms du y is
exemp ed and du y- ee allowances o in e media e inpu s a e o e ed o he i ms loca ed
in SEZs.
2. The land is owned by he go e nmen and he land use igh is s ic ly egula ed in
China. Di e en om many coun ies, o icials can alloca e land on a la ge scale and con e
ag icul u al land o indus ial pu poses when necessa y. Land use igh s o indus ial
pu poses a e g an ed o domes ic and o eign in es o s ha en e he SEZs. Besides land
use igh s, land use ees a e discoun ed o en an s ela i e o he i ms ou side zones.
3. The s a e-owned bank loosens lending policies and gi es p io i y o he i ms in
SEZs o apply o a loan.
4. Fo po en ial in es o s who en e he zones, p ocedu es a e simpli ied and app o ed
o high-speed.
5. The go e nmen commi s o he in es o s who en e he zones ha all o hei
p i a e p ope ies a e unde -p o ec ed.
Besides hese p e e en ial policies, he go e nmen also makes a g ea e o o imp o e
in as uc u es o zones such as oads, po s, elec ici y, gas, wa e , elecommunica ions,
and o he se ice acili ies. Fu he mo e, o a ac skilled human capi al, SEZs o e an
ex a pe sonal income ax deduc ion, allowances, and Hukou egis a ion p io i y bene i
o a highly quali ied indi idual2.
Economies 2022,10, 99 5 o 15
Each SEZ has i s adminis a i e commi ee ha pe o ms managemen unc ions
wi hin i s geog aphical scope. Adminis a i e commi ees a e no unde he con ol o
local go e nmen s, hey a e di ec ly con olled by he s a e o he p o incial go e nmen s.
They di ec and adminis e a ai s o SEZs on he behal o he go e nmen such as p ojec
app o al, local axa ion, land managemen , public acili ies planning, inancial e enue,
pe sonnel, en i onmen p o ec ion, and so on. Fo example, adminis a i e commi ees ake
esponsibili y o a ac in es o s om domes ic o ab oad ha mee he s anda ds o local
de elopmen . Each adminis a i e commi ee has he igh o decide which in es o s could
en e he zone. They o e a bundle o p e e en ial policies and nego ia e wi h po en ial
en an s o de ails. A e hese nego ia ions, he in es o s decide whe he o en e he zone
o no .
4. Da a
The main da a is i m-le el da a om he Annual Su ey o Indus ial Fi ms (ASIF)
conduc ed by he Na ional Bu eau o S a is ics o China (NBS). The ASIF da a co e s all
s a e-owned i ms and non-s a e-owned i ms wi h an annual u no e exceeding i e
million yuan (app oxima ely $700,000), and hose i ms occupy o e 90% o he o al
indus ial ou pu s o China in 2004. The ASIF da a con ains mo e han 100 a iables,
p o iding in o ma ion on indus ial ou pu , in e media e inpu , o al employmen , indus y
a ilia ion, and geog aphic loca ion. Since co e ing comp ehensi e a iables, he da a has
been widely used in empi ical li e a u e such as B and e al. (2012).
The ASIF da a in 2004 a e mo e comp ehensi e han o he yea s’ ASIF da a. Besides
he basic in o ma ion, he da a in 2004 include he le el o employee educa ion. To e alua e
he spillo e e ec s o human capi al, we exclusi ely use he 2004 da a
3
. In addi ion,
ou pape ocuses on manu ac u ing i ms in Yang ze Del a, which is made up o h ee
p o ince-le el a eas in eas -cen al China: Shanghai Ci y, Jiangsu, and Zhejiang P o inces.
Al hough occupying abou 2% o he o al land a ea, Yang ze Del a con ibu es mo e han
21% o he GDP in China.
The in o ma ion o SEZs is om no ices o he go e nmen . One is “The e iew o
Special Economic Zones in China”, which was published by he Na ional De elopmen and
Re o m Commission and he Minis y o Land and Resou ces in 2006. I con ains au ho ized
SEZs’ names, app o ed imes, and occupied a eas. The o he one is “ he no ice o ou
di ec ions”, which includes geog aphic in o ma ion on SEZs eleased by he Minis y o
Land and Resou ces o China. I has de ail in o ma ion on he geog aphic bounda y o
SEZs, which co e s speci ic illages, oads, o coas s.
The ASIF da a does no epo any in o ma ion abou SEZs; howe e , i con ains each
i m’s add ess and geog aphic loca ion code. The geog aphic loca ion consis s o a 12-digi
geog aphic code, which p o ides loca ion in o ma ion a he mos disagg ega ed le el. I
consis s o dis ic (o coun y), jiedao (s ee s o a enues), and juweihui (communi ies o
illages). Following Zheng e al. (2017), each i m’s geog aphic loca ion code and he exac
geog aphic bounda ies o SEZs can be used o iden i y whe he a i m is loca ed in SEZs o
no . By compa ing geog aphic loca ion code and bounda ies, he e a e abou 15% o he
i ms loca ed in SEZs.
To in es iga e in a- and in e - egional spillo e s om SEZs, we de ine egion as a
coun y in his pape . In Figu e 1, we p esen he coun ies hos ing SEZ p og ams in Yang ze
Del a. SEZ-hos ing coun ies a e mainly dis ibu ed on he eas coas o along he Yang ze
Ri e . The e a e as many as 108 SEZ-hos ing coun ies in Yang ze Del a in 2004, which
means SEZ-hos ing coun ies occupy abou wo- hi ds o all he coun ies in he a ea.
Economies 2022,10, 99 6 o 15
Figu e 1. SEZ-hos ing coun ies in Yang ze Del a.
In his pape , all a iables a e agg ega ed a he coun y-le el by using i m-le el da a
in o ma ion. In pa icula , we de ine egional p oduc i i y as he log o alue-added pe
employed wo ke o indus y
k
in coun y
i
(
yik
). We assume ha he egional p oduc i i y
unc ion depends on h ee ac o s. Fi s , we employ he le el o educa ion as a measu e
o human capi al.
hik
deno es he sha e o labo who ha e college deg ees o abo e o
he i ms loca ed in he zones o indus y
k
in coun y
i
;
eduik
deno es he a e age numbe s
o schooling yea s o he i ms ou side he zones o indus y
k
in coun y
i
. Second, since
in es men is undamen al o egional p oduc i i y, we use physical capi al in o he unc ion
as a p oxy a iable o desc ibe egional p oduc i i y: he log o capi al s ock pe employed
wo ke o each indus y in coun y
i
(
capik
). Thi d, we con ol o compe i ion index (
comik
)
and indus ial di e si y index (
di ik
). The s eng h o compe i ion o an indus y in a
local ma ke is likely o in luence egional p oduc i i y. Fie ce compe i ion may dec ease
egional p oduc i i y due o diminishing ma ginal e u ns. On he o he hand, in ensi e
compe i ion o ces local i ms o adop ad anced echniques o imp o e p oduc i i y. In
his pape , we de ine a compe i ion index as ollows,
comik =ln(1/Hik)
whe e
Hik =∑ ∈Ωik ( e enueik
e enueik )2
deno es an He indahl index o sales e enue concen a ion
o indus y
k
in coun y
i
;
e enueik
is sales e enue o i m
belongs o indus y
k
in coun y
i
; and
e enueik
deno es he sum o sales e enue o indus y
k
in coun y
i
. Addi ionally,
we con ol o indus ial di e si y index (
di ik
). P oduc i i y o an indus y in a egion can
be in luenced by di e en indus ies, and he mo e di e se, he mo e likely o imp o e
egional p oduc i i y. Following Hende son (1997), an indus ial di e si y index is de ined
as he ollows:
di ik =ln(1/ ∑
k6=k0
(empik0
empi−empik
)2)
whe e
empi
deno es o al employmen in coun y
i
;
empik
deno es he o al employmen o
indus y kin coun y i.
5. Model Speci ica ion
In his pape , we in oduce spa ial dependence o analyze he spillo e e ec s o he
human capi al o SEZs o hei own coun ies and he neighbo ing ones. Conside ing spa ial
dependence o examine he e ec s o SEZs o in a- and in e - egion is plausible because
Economies 2022,10, 99 7 o 15
a posi i e shock on he p oduc i i y o an indus y in a coun y could also be ans e ed
o he same indus y in he nea by coun ies. Mo eo e , he spa ial lag o explana o y
a iables can also be included because ex e nali ies a ising om neighbo ’s human capi al
o o he neighbo cha ac e is ics could also play a ole in de e mining he p oduc i i y o
he local indus y.
The e a e wo ad an ages o employ spa ial eg essions. Fi s , spa ial eg essions can
ake in o accoun he spa ial lag o he dependen a iable and independen a iables o
desc ibe he ou come o in e es . The in oduc ion o spa ial dependence allows obse a-
ions o ha e associa ions wi h each o he and o explo e he ela ionship be ween a coun y
and he neighbo ing coun ies. In addi ion, spa ial eg ession cap u es spillo e e ec s
h ough h ee channels. A 1% inc ease in he humans o SEZs will a ec he SEZ-hos ing
coun y i sel (di ec e ec ) and possibly a ec he neighbo ing coun ies (indi ec e ec ).
Fu he mo e, spa ial eg ession exploi s impac passing h ough he neighbo ing coun ies
and back o he coun ies hemsel es ( eedback e ec ). The inclusion o spa ial in o ma ion
makes i possible o assess he e ec s o he SEZ p og am mo e comp ehensi ely h ough
mul iple channels.
In his pape , we di ide manu ac u ing i ms in o 28 indus ies by 2-digi le el classi i-
ca ion, o he eason ha he same indus y clus e s oge he and has simila cha ac e is ics
in p oximi y. We es ima e human capi al spillo e s o SEZs on he local economy ia a
Spa ial Du bin Model (SDM) as ollows:
y=α+Wyδ+Xβ+WXθ+ε
X= [h edu cap com di ](1)
whe e
y
is egional p oduc i i y o indus y
k
in coun y
i
;
W
is ow-s anda dized in e se
weigh ma ix wi hin 150 km, o he wise ze o;
h
deno es he sha e o employees who ha e
college deg ees o abo e o he i ms loca ed in he zones in coun y
i
;
edu
con ols he
a e age numbe s o schooling yea s o he i ms ou side he zones o indus y
k
in coun y
i
;
cap
is he capi al s ock pe employed wo ke o each indus y in coun y
i
;
com
and
di
deno e compe i ion index and indus ial di e si y index; and
ε
is he e o e m, no mally
dis ibu ed by (0, σ2).
The coe icien o
h
desc ibes he spillo e e ec s om he human capi al o SEZs on
egional p oduc i i y. I he coe icien o
h
is signi ican ly posi i e, which means ha a
1% inc ease in he human capi al o SEZs s imula es egional p oduc i i y. Addi ionally,
i he es ima e
δ
is s a is ically signi ican , i implies ha a spa ial au oco ela ion exis s
and in oducing spa ially lagged dependen a iable is meaning ul. Fu he mo e, i he
coe icien o
Wh
is signi ican ly posi i e, which means ha he human capi al o SEZ in a
coun y is simila o he one in neighbo ing coun ies loca ed in close dis ance.
Wi h espec o he selec ion o models, we compa e SDM models wi h o he spa ial
eg ession models. As no ed by LeSage and Pace (2009), SDM models nes mos o he o he
speci ica ions like Spa ial E o Model (SEM) and Spa ial Au o eg essi e Model (SAR).
Hence, we es ima e he SDM models and hen compa e hem wi h SEM o SAR models by
Likelihood Ra io (LR) es . The es ima ion esul s ob ained om hese models can be used
o es he hypo hesis:
H0:θ=0
H0:θ+δβ=0.
The o me one examines he hypo hesis o whe he he SDM model can be simpli ied
o o ms o SAR model. I he null hypo hesis is ejec ed, i demons a es ha he SDM
model be e desc ibes he da a; he la e one is used o examine be ween SDM and SEM
model. I he null hypo hesis is ejec ed, i means ha he SDM model canno be simpli ied
o he SEM model. Since he SDM model is a gene a ed o m nes ing he o he wo models,
i is mo e a o ed han he wo models in he case ha ei he o he wo hypo heses
is ejec ed.
Economies 2022,10, 99 8 o 15
As sugges ed by LeSage and Pace (2009), we measu e a e age di ec and indi ec
e ec s o explain he ma ginal e ec s o he explana o y a iable. A e age di ec e ec ,
measu ed by aking an a e age o he own de i a i es o he coun ies hemsel es, which
cap u es he e ec s o SEZs o i s own coun y. The a e age indi ec e ec is calcula ed
by he a e age o de i a i es wi h espec o neighbo ing coun ies, which measu es he
spillo e s o SEZs o neighbo ing coun ies. To illus a e, we ake he pa ial de i a i e o
he dependen a iable wi h espec o has ollows:
∂y
∂h=S(W) = (In−δW)−1(Inβ1+Wθ1)(2)
whe e
In
a e 3815
×
3815 idempo en ma ices. F om Equa ion (4), he pa ial de i a i e
is no only depends on
β1
, bu also
θ1
,
δ
, and he spa ial weigh ma ix. By de ini ion, he
a e age indi ec e ec is calcula ed by he a e age o diagonal elemen s o
S(W)
, and
he a e age indi ec e ec is measu ed by he a e age o o -diagonal elemen s o
S(W)
.
F om he abo e equa ion, we can cap u e spillo e e ec s h ough h ee channels. An
inc ease in human capi al in SEZs will a ec he i ms in he conside ed coun y i sel (di ec
e ec ) and a ouse spillo e s passing h ough neighbo ing coun ies and back o he coun ies
hemsel es ( eedback e ec ). Fu he mo e, a g ow h in human capi al in a coun y may
a ec he neighbo ing coun ies (indi ec e ec ).
Ramos e al. (2010) sugges ha one needs o es ima e he e ec s o human capi al on
g ow h a e o p oduc i i y because he human capi al used in mos li e a u e is a s ock
a iable, which may a ouse he p oblem o endogenei y. Following López-Bazo e al. (2004);
Ramos e al. (2010), we ans o m he Equa ion (1) as ollows:
∆y=α+W∆yδ+y−1γ+Xβ+WXθ+ε
X= [h edu cap com di ](3)
whe e
∆y
deno es he g ow h a e o egional p oduc i i y o indus y
k
in coun y
i
o he
pe iod om 2003 o 2004; and
y−1
deno es he egional p oduc i i y o indus y
k
in coun y
iin 2003.
The Equa ion (3) e alua es he e ec s o human capi al on he g ow h o egional
p oduc i i y con olling o he neighbo s, lagged alue o he p oduc i i y, and o he
a iables. I he es ima e
δ
is signi ican ly posi i e, i implies ha he g ow h o egional
p oduc i i y depends on he one in he neighbo s. Addi ionally, he coe icien o
h
de-
sc ibes he spillo e e ec s om he human capi al o SEZs on he g ow h o egional
p oduc i i y when con olling o he lagged p oduc i i y and o he explana o y a iables.
Fu he mo e, i he coe icien o
Wh
is signi ican ly posi i e, i implies ha he human
capi al o SEZ in a coun y is simila o one ano he ha is loca ed in close dis ance.
6. Empi ical Resul s
Table 1 epo s he esul s o spa ial eg ession models using egional p oduc i i y as
he dependen a iable. To diagnose he exis ence o spa ial dependence, we i s es ima e
he Equa ion (1) using linea o dina y leas squa es (OLS) models wi hou spa ial e ec s.
The es ima ed coe icien s o
h
om he OLS o SDM model a e all signi ican ly posi i e.
In addi ion, he es ima ed coe icien s o he spa ially lagged dependen a iable in he
SAR and SDM models a e bo h signi ican a he 1% signi icance le el, which sugges s
he exis ence o spa ial dependence. In he SDM model, he coe icien o
h
amoun s o
0.7936, which implies ha a 1% inc ease in he human capi al o SEZs will lead o 0.7936%
imp o emen in egional p oduc i i y. Fu he mo e, he spa ial lag o
h
is signi ican ly
posi i e in he SDM model, which implies ha he human capi al in a zone is simila o he
one in he neighbo zones.
The coe icien s o
edu
in he OLS and SAR models a e signi ican ly posi i e, howe e ,
he one in he SDM model is insigni ican . One may doub ha he human capi al ou side
he zones has a limi ed impac on he local p oduc i i y. The coe icien s o physical
capi al a e all signi ican ly posi i e. In he hi d column, he coe icien o
cap
amoun s
Economies 2022,10, 99 15 o 15
Bal agi, Badi H., Pe e H. Egge , and Michaela Kesina. 2016. Fi m-le el p oduc i i y spillo e s in China’s chemical indus y: A spa ial
Hausman-Taylo app oach. Jou nal o Applied Econome ics 31: 214–48. [C ossRe ]
B and , Lo en, Johannes Van Bieseb oeck, and Yi an Zhang. 2012. C ea i e accoun ing o c ea i e des uc ion? Fi m-le el p oduc i i y
g ow h in Chinese manu ac u ing. Jou nal o De elopmen Economics 97: 339–51. [C ossRe ]
Ci˙
zkowicz, Pio , Magda Ci˙
zkowicz-P˛ekała, Pio P˛ekała, and And zej Rzo´nca. 2017. The e ec s o special economic zones on
employmen and in es men : A spa ial panel modeling pe spec i e. Jou nal o Economic Geog aphy 17: 571–605. [C ossRe ]
Dö ing, Thomas, and Jan Schnellenbach. 2006. Wha do we know abou geog aphical knowledge spillo e s and egional g ow h?: A
su ey o he li e a u e. Regional S udies 40: 375–95. [C ossRe ]
Elho s , J. Paul. 2014. Spa ial Econome ics: F om C oss-Sec ional Da a o Spa ial Panels. Heidelbe g: Sp inge .
Fa ole, Thomas. 2011. Special Economic Zones in A ica: Compa ing Pe o mance and Lea ning om Global Expe iences. Washing on, DC: The
Wo ld Bank.
Fische , Man ed, Monika Ba kowska, Aleksand a Riedl, Sascha Sa dad a , and And ea Kunne . 2009. The impac o human capi al
on egional labo p oduc i i y in Eu ope. Le e s in Spa ial and Resou ce Sciences 2: 97–108. [C ossRe ]
Glaese , Edwa d, and Joshua Go lieb. 2008. The economics o place-making policies. In B ookings Pape s on Economic Ac i i y.
Washing on, DC: B ookings Ins i u ion.
G eens one, Michael, and Adam Looney. 2010. An Economic S a egy o Renew Ame ican Communi ies. Hamil on P ojec . Washing on,
DC: B ookings Ins i u ion.
Hashiguchi, Yoshihi o, and Kiyoyasu Tanaka. 2015. Agglome a ion and i m-le el p oduc i i y: A bayesian spa ial app oach. Pape s in
Regional Science 94: 95–114. [C ossRe ]
Hende son, Ve non. 1997. Ex e nali ies and indus ial de elopmen . Jou nal o U ban Economics 42: 449–70. [C ossRe ]
Johansson, Helena, and La s Nilsson. 1997. Expo p ocessing zones as ca alys s. Wo ld De elopmen 25: 2115–28. [C ossRe ]
Kline, Pa ick, and En ico Mo e i. 2014. People, places, and public policy: Some simple wel a e economics o local economic
de elopmen p og ams. Annual Re iew o Economics 6: 629–62. [C ossRe ]
LeSage, James, and Robe Kelly Pace. 2009. In oduc ion o Spa ial Econome ics. Boca Ra on: CRC P ess.
López-Bazo, En ique, Es he Vayá, and Manuel A is. 2004. Regional ex e nali ies and g ow h: E idence om Eu opean egions.
Jou nal o Regional Science 44: 43–73. [C ossRe ]
Lu, Yi, Jin Wang, and Lianming Zhu. 2019. Place-based policies, c ea ion, and agglome a ion economies: E idence om China’s
economic zone p og am. Ame ican Economic Jou nal: Economic Policy 11: 325–60. [C ossRe ]
Lucas, Robe , J . 1988. On he mechanics o economic de elopmen . Jou nal o Mone a y Economics 22: 3–42. [C ossRe ]
Ma usik, Sha on F., and Cha les W. L. Hill. 1998. The u iliza ion o con ingen wo k, knowledge c ea ion, and compe i i e ad an age.
Academy o Managemen Re iew 23: 680–97. [C ossRe ]
Mo e i, En ico. 2011. Local labo ma ke s. In Handbook o Labo Economics. Ams e dam: Else ie , ol 4, pp. 1137–313.
Neuma k, Da id, and Helen Simpson. 2015. Place-based policies. In Handbook o Regional and U ban Economics. Ams e dam: Else ie ,
ol. 5, pp. 1197–287.
Ramos, Raul, Jo di Su iñach, and Manuel A ís. 2010. Human capi al spillo e s, p oduc i i y and egional con e gence in Spain.
Pape s in Regional Science 89: 435–47. [C ossRe ]
Rosen hal, S ua S., and William C. S ange. 2008. The a enua ion o human capi al spillo e s. Jou nal o U ban Economics 64: 373–89.
[C ossRe ]
Wang, Jin. 2013. The economic impac o special economic zones: E idence om Chinese municipali ies. Jou nal o De elopmen
Economics 101: 133–47. [C ossRe ]
Zeng, Douglas Zhihua. 2011. How Do Special Economic Zones and Indus ial Clus e s D i e China’s Rapid De elopmen ? Washing on, DC:
The Wo ld Bank.
Zeng, Douglas Zhihua. 2016. Special economic zones: Lessons om he global expe ience. PEDL Syn hesis Pape Se ies 1: 1–9.
Zheng, Siqi, Weizeng Sun, Jian eng Wu, and Ma hew E. Kahn. 2017. The bi h o edge ci ies in China: Measu ing he e ec s o
indus ial pa ks policy. Jou nal o U ban Economics 100: 80–103. [C ossRe ]
