The complex regional effects of macro-institutional change: evidence from EU enlargement over three decades

Mi ze, Timo; B eidenbach, Philipp
A icle — Published Ve sion
The complex egional e ec s o mac o-ins i u ional
change: e idence om EU enla gemen o e h ee
decades
Re iew o Wo ld Economics
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: Mi ze, Timo; B eidenbach, Philipp (2024) : The complex egional e ec s o
mac o-ins i u ional change: e idence om EU enla gemen o e h ee decades, Re iew o Wo ld
Economics, ISSN 1610-2886, Sp inge , Be lin, Heidelbe g, Vol. 160, Iss. 4, pp. 1443-1475,
h ps://doi.o g/10.1007/s10290-024-00528-6
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/315080
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h p://c ea i ecommons.o g/licenses/by/4.0/
Vol.:(0123456789)
Re iew o Wo ld Economics (2024) 160:1443–1475
h ps://doi.o g/10.1007/s10290-024-00528-6
ORIGINAL PAPER
The complex egional e ec s o mac o‑ins i u ional
change: e idence omEU enla gemen o e  h ee decades
TimoMi ze1,2 · PhilippB eidenbach2
Accep ed: 13 Feb ua y 2024 / Published online: 3 May 2024
© The Au ho (s) 2024
Abs ac
The EU enla gemen p ocess has h us EU in e nal bo de egions in o he spo ligh
o he Eu opean single ma ke . This s udy explo es how his speci ic mac o-
ins i u ional change has impac ed hei socio-economic de elopmen . T acking
ou comes ac oss ou EU enla gemen wa es om 1986 o 2007, we iden i y
in eg a ion e ec s ac oss EU NUTS3 egions. Pooled o e all wa es and bo de
egions, posi i e in eg a ion e ec s eme ge o pe capi a GDP, labo p oduc i i y,
pa en s pe capi a, and nigh ligh emissions in bo de egions compa ed o non-
bo de a eas. These e ec s diminish wi h inc easing spa ial dis ance om he
enla gemen bo de . A a de ailed le el, s uc u al he e ogenei ies become e iden
ac oss enla gemen wa es and egion ypes. In e nal bo de egionsin es ablished
EU membe coun ies bene i ela i ely in e ms o GDP pe capi a and labo
p oduc i i y bu expe ience ela i e declines in employmen a es and popula ion.
In con as , bo de egions in new membe coun ies, pa icula ly du ing he 2004
and 2007 eas e n enla gemen s, gain om deepening economic in eg a ion in e ms
o popula ion and employmen g ow h. Sec o -speci ic es ima ions indica e pos -
enla gemen specializa ion o economic ac i i ies in bo de egions in line wi h
s anda d ade heo ies.
Keywo ds Economic in eg a ion· EU enla gemen · In e nal bo de egions·
Regional de elopmen · T ea men e ec es ima ion
JEL Classi ica ion C23· F15· O47· R11
* Timo Mi ze
[email p o ec ed]
1 Depa men o Economics, Uni e si y o Sou he n Denma k, Campus ej 55, DK-5230Odense,
Denma k
2 RWI – Leibniz Ins i u e o Economic Resea ch, Essen, Ge many
1444
T.Mi ze, P.B eidenbach
1 In oduc ion
Regions loca ed along he bo de o economically in eg a ing coun ies a e highly
exposed o his mac o-ins i u ional change and we add ess he ques ion i his
exposu e ansla es in o speci ic in eg a ion e ec s o he egions’ socio-economic
de elopmen pa h. As bo de s a e a na u al ba ie o economic in e ac ion
(Capello e al., 2018a), he disman ling o bo de impedimen s h ough economic
in eg a ion may, on he one hand, imp o e hei ma ke access by shi ing bo de
egions om he coun y’s pe iphe y o he hea o he newly o med economic
block (e.g., Pe coco, 2015). While his a gumen speaks in e ms o posi i e EU
in eg a ion e ec s in bo de is-à- is non-bo de egions, he e a e also a gumen s
o a ela i e weake pe o mance o bo de egions: The ‘pa h dependency’ o ‘lock
in’ a gumen , o ins ance, s a es ha when bo de s ha e pe ained o a long ime,
bo de egions may ha e su e ed om a g adual p ocess o ma ginaliza ion ha
dep i es hem o he abso p i e capaci ies and scale e ec s needed o bene i om
economic in eg a ion mo e han hei be e endowed, agglome a ed non-bo de
coun e pa s (Floe kemeie e al., 2021; Pe akos & Topaloglou, 2008).
A signi ican g ow h p emium in bo de egions would also be absen i ade
cos s we e su icien ly low so ha close geog aphical p oximi y o new ma ke s
i sel is no a decisi e ac o o eaping he bene i s o open bo de s and economic
in eg a ion as i is, o ins ance, p edic ed in K ugman (1991)- ype co e-pe iphe y
models. And inally, socio-economic de elopmen le els in coun ies on bo h
sides o he in eg a ion bo de may di e in such a way ha gains om economic
in eg a ion in es ablished (old) and new membe s a es, i.e., e ec s s emming om
widening and deepening EU in eg a ion, a e une enly dis ibu ed ac oss bo de
egions so ha o e all e ec s in bo de egions a e di icul o measu e. Thus, wha
seemed s aigh o wa d on a i s glance, namely, o iden i y he ea men e ec s o
economic in eg a ion o bo de egions along he in eg a ion bo de may, in ac , be
qui e complex and subjec o s uc u al, empo al, and spa ial he e ogenei ies. In his
pape , we ake his ‘complexi y’ pe spec i e as s a ing poin o an in-dep h s udy
o he in eg a ion e ec s associa ed wi h ou consecu i e enla gemen wa es o he
Eu opean Union be ween 1986 and 2007. Essen ial esea ch ques ions a e:
1. Do we ind e idence o common in eg a ion e ec s o EU enla gemen o
in e nal bo de egions in e ms o key socio-economic ou comes such as GDP,
employmen , and popula ion g ow h?
2. Wha ypes o s uc u al, empo al, and spa ial he e ogenei y de e mine he
di ec ion, magni ude, and du a ion o in eg a ion e ec s ac oss enla gemen wa es
and sub-g oups o bo de egions conside ed?
P o iding answe s o bo h esea ch ques ions shall help policy make s o be e
assess he bene i s and cos s associa ed wi h EU economic in eg a ion in he con ex
o egional commonali ies bu also he e ogenei ies.
When i comes o he selec ion o EU in e nal bo de egions as ea men g oup
in ou empi ical in es iga ion, we a gue ha his ocus is well dese ed ou o
1445
The complex egional e ec s o mac o‑ins i u ional change:…
signi icance and ele ance conside a ions. Fi s , as s a ed by he EU Commission
(2017), bo de egions accoun o app oxima ely one hi d o EU popula ion and
a simila agg ega e p oduc ion sha e in he EU.1 A he same ime, bo de egions,
on a e age, ha e a weake economic pe o mance, lowe le els o labo ma ke
in eg a ion and public se ice p o ision compa ed o non-bo de egions in he EU
Commission (2017), which is why EU egional policy suppo s he socio-economic
de elopmen in bo de egions h ough di e en unding p og ammes (mos no ably
h ough In e eg p ojec unding and he b-solu ions ini ia i e, see EU Commission
(2021).
Second, beyond hei s a us as being a speci ic (disad an aged) egional g oup
wi hin he wide in e nal economic geog aphy o he EU, he EU Commission
sees bo de egions as impo an “li ing labs o Eu opean in eg a ion” (EU
Commission, 2021). The idea o li ing labs is ha hey allow o s udy in eg a ion
e ec s in bo de egions unde he magni ying glass and ha indings ob ained he e
p o ide impo an gene al insigh s on he o e all p og ess o EU in eg a ion and
coope a ion a la ge. Ex ending his logic, we a gue ha EU in e nal bo de egions
a e pa icula ly well-sui ed o in es iga e he economic e u ns o EU economic
in eg a ion a he egional le el as hey ha e been pa icula ly exposed o associa ed
shi s in he EU’s in e nal economic geog aphy, while he enla gemen p ocess i sel
can be seen as an exogenous sou ce o a ia ion o hei de elopmen pa h.
The exogenei y o he enla gemen ‘shock’ a he small-scale egional le el can
be mo i a ed by he ac ha poli ical decisions o EU enla gemen we e made
a he na ional and sup a-na ional le el wi h goals no speci ically ailo ed o he
needs and economic condi ions o bo de egions. The same logic applies o he
alloca ion o he bulk o EU egional unding olumes, which ocus on he egions’
de elopmen s a us i espec i e o hei geog aphical loca ion wi hin a coun y
(B eidenbach e  al., 2019), so ha EU egional unding alone canno explain a
po en ial g ow h p emium associa ed wi h economic in eg a ion in bo de egions.
We acco dingly a gue ha he ocus on bo de egions enables us o s udy he e ec s
o EU in eg a ion in a quasi-expe imen al manne .
P io empi ical e idence on a po en ial g ow h and de elopmen p emium associ-
a ed wi h EU accession and he economic eedoms associa ed wi h he Eu opean
single ma ke has emained inconclusi e. While, e.g., Campos e al. (2019) epo
signi ican posi i e income g ow h e ec s o EU membe ship a he coun y le el
(wi h ew excep ions), Ande sen e al. (2019) gene ally do no ind e idence o an
EU membe ship g ow h p emium. Wi h espec o he ocus o his s udy, he e is
also a knowledge gap on how he po en ial gains om economic in eg a ion a e dis-
ibu ed ac oss he di e en egions wi hin in eg a ing coun ies (Niebuh & S ille ,
2004, B aakmann and Vogel, 2010, and Heide , 2019). I is gene ally supposed
ha egional abili y o eap wel a e gains om EU in eg a ion chie ly depend on
a egion’s ela i e compe i i eness d i en by indus y composi ion and se lemen
s uc u e, i s ins i u ional se up, ade in ensi y as well as size and geog aphical
1 Based on a de ini ion o in e nal bo de egions o he EU-28 (including he UK) and also including
EU bo de egions o EEA coun ies (see EU Commission, 2017).
1446
T.Mi ze, P.B eidenbach
p oximi y o he enla gemen bo de (e.g., B akman e  al., 2012; B ülha e  al.,
2012, 2018; McCallum, 1995).
While ea lie s udies ha e mainly ocused on GDP g ow h as sole ou come
a iable, a no el y o ou analysis is ha we conduc a comp ehensi e empi ical
analysis o he complex bo de egional e ec s associa ed wi h he EU enla gemen
p ocess du ing he 1980s, 1990s and 2000s. Speci ically, we employ a b oad se
o ou come a iables co e ing egion-speci ic ime pa e ns o pe capi a GDP,
(sec o al) labo p oduc i i y, esea ch and de elopmen (R&D) and inno a ion
ac i i y, employmen , popula ion de elopmen , nigh ligh emissions. As he e is no
in o ma ion on he s ock o public (and also p i a e) in as uc u e a he le el o
NUTS3 egions ( he obse a ional uni used in his pape ), nigh ligh da a ill an
impo an gap. An inc ease in nigh ligh emissions e lec s changes o he public
in as uc u e (such as s ee s o public buildings) bu also changes o p i a e
ac i i ies such as housing s ocks o i m densi y/ac i i y.2 I can hus be ega ded
as a gene al measu e o agglome a ion ends and has p e iously been used o map
he de elopmen o popula ion and i m densi y ac oss egions (e.g., Mellande
e al., 2015). Also, p io empi ical analyses ha e used nigh ligh da a o measu e
p ocesses o economic in eg a ion, g ow h, and con e gence, especially when o he
economic da a a e missing (see, e.g., Hende son e al., 2012; Galimbe i, 2020). Bu
e en o geog aphical a eas wi h ai ly good da a p o ision, Mellande e al. (2015)
as well as Lessmann and Seidel (2017) ha e shown ha nigh ligh da a s ill deli e s
impo an insigh s on economic de elopmen ends and di e ences.
Fou EU enla gemen pe iods a e co e ed in ou analysis: Fi s , he EU accession
o Spain and Po ugal in 1986 ( hi d enla gemen wa e); second, he EU membe ship
o Aus ia, Sweden, and Finland in 1995 ( ou h enla gemen wa e); hi d, he so- a
la ges EU enla gemen o mos ly cen al and eas e n Eu opean coun ies in 2004
( i h enla gemen wa e); and ou h, he accession o Romania and Bulga ia o he
EU in 2007 (six h enla gemen wa e). To iden i y ea men e ec s o EU in eg a ion
in bo de egions is-à- is non-bo de egions du ing hese enla gemen pe iods,
we conduc an analysis o he 1289 NUTS3 egions o he EU-27 (including he
UK bu no C oa ia) o e he ime pe iod 1981–2014. We apply s a ic and dynamic
di e ence-in-di e ence (DiD) es ima ion o iden i y pooled and g oup-speci ic
e ec s. The DiD app oach has shown a high deg ee o lexibili y and obus ness
when p e iously been applied o spa io- empo al analyses o bo de egional g ow h
e ec s such as o he di ision and euni ica ion o Ge many (Redding & S u m,
2008) and economic ans o ma ions a e he all o he i on cu ain (B ülha e al.,
2012, 2018) among o he applica ions.
In he es ima ions we pa icula ly accoun o he ac ha he dis ibu ion o
in eg a ion e ec s may be uzzy wi h ega d o spa ial and empo al aspec s. As such,
we explici ly con ol o he ci cums ance ha EU enla gemen canno be ea ed
2 The NUTS classi ica ion (Nomencla u e o e i o ial uni s o s a is ics) is a hie a chical sys em o
di iding up he economic e i o y o he EU. The NUTS egula ion mi o s he e i o ial adminis a i e
di ision o he EU membe s a es and de ines minimum and maximum popula ion h esholds o he size
o egions. A he NUTS Le el 3 egions ha e a popula ion size o be ween 150,000 and 800,000 inhabi -
an s ( o de ails see h ps:// ec. eu opa. eu/ eu os a / web/ nu s/ p inc iples).

1447
The complex egional e ec s o mac o‑ins i u ional change:…
as p ecisely imed e en . Fo example, in 2004, EU accession ollowed a p ocess
co e ing ea ly ag eemen s be ween old and new EU membe s a es ini ia ed in he
a e ma h o he collapse o he So ie sys em (Dange ield, 2006). This po en ially
esul s in ‘ea ly an icipa ion’ e ec s ha weaken he powe o s a ic DiD es ima ions,
which ely on a p ecise classi ica ion o a single p e- and pos - ea men pe iod.3
To accoun o hese me hodical challenges, we apply a lexible DiD app oach ha
es ima es ime-he e ogeneous coe icien s o he di e en s ages a ound he iming
o EU enla gemen . In addi ion, we accoun o he con ounding ac o s, which may
ei he a ec all EU coun ies equally, such as he deepening o economic in eg a ion
h ough he EU single ma ke in 1992, o is con ined o indi idual coun ies and
coun y g oups, such as he in oduc ion o he Eu o cu ency in 1999, by adding
a mul idimensional ‘ ixed e ec s’ s uc u e o ou DiD speci ica ions. We also un
se e al obus ness es s o see i he ob ained esul s hold o a ia ions in he da a
and eg ession speci ica ion.
The emainde o his pape is o ganized as ollows: Sec . 2 ou lines he
unde lying heo y ela ed o bo de egional g ow h e ec s o economic in eg a ion.
This sec ion also summa izes p io empi ical indings o he economic e ec s o
EU enla gemen and iden i ies esea ch gaps in he li e a u e. Sec ion3 desc ibes ou
empi ical s udy design, which is ollowed by a desc ip ion o he da a and a iables
used in Sec .4. Sec ion5 epo s ou empi ical esul s o pooled and he e ogeneous
ea men e ec s o EU economic in eg a ion oge he wi h a se ies o obus ness
es s. Finally, Sec .6 discusses policy implica ions and concludes he pape .
2 Bo de egional e ec s o economic in eg a ion: heo y, e idence
andgaps
2.1 A complexi y pe spec i e o economic in eg a ion
Models o egional g ow h, in e na ional ade, and economic geog aphy s ess
he ole o ade ela ed o ma ke size, ma ke access and anspo cos o
egional de elopmen (e.g., K ugman & Venables, 1990; Pe coco, 2015). I can
be conjec u ed ha bo de egions gain om EU enla gemen due o hei unique
geog aphic loca ion and he associa ed imp o emen o ma ke access. These e ec s
may, howe e , be pa ly o ully o se by sus aining bo de impedimen s, lacking
abso p i e capaci ies and his insu icien scale economies in bo de egions,
which may lock egions in a pe iphe al posi ion (Capello e al., 2018a; Pe akos &
Topaloglou, 2008).
Ou concep ual app oach, which akes hese opposing ac o s in o accoun s a s
wi h a ai ly gene al speci ica ion o a egional p oduc ion unc ion de ined as
Y=A(K𝛼L𝛽
𝐍
φ)
, whe e
Y
is a measu e o egional ou pu ( ypically GDP o GVA),
3 A simila p oblem would a ise i e ec s only g adually ‘phase in’ o e ime because o ins i u ional
a angemen s such as he 2 + 3 + 2 egula ion, which allowed es ablished membe coun ies o empo a ily
p o ec hei labo ma ke s om ee labo mobili y associa ed wi h EU accession o new membe s. Acco d-
ing o he 2 + 3 + 2 egula ion he maximum p o ec ion pe iod amoun ed o se en yea s ( o de ails see
h ps:// www. eu o ound. eu opa. eu/ en/ eu op ean- indus ial- ela ions- dic i ona y/ mobil i y- wo ke s).
1448
T.Mi ze, P.B eidenbach
A is echnology, K is capi al, L deno es labo inpu and
𝐍
is a ec o o u he
inpu s; α,
𝛽
and
φ
a e he espec i e ou pu elas ici ies. I we w i e his egional
p oduc ion unc ion as g ow h speci ica ion in in ensi e o m, we ge
whe e
Δyi
is as measu e o pe wo ke (o pe capi a) ou pu g ow h o egion
i a ime ,
y=Y∕L
,
k=K∕L
and simila o he emaining inpu s (
n =N ∕L
). In
an ea lie analysis wi h na ional da a o he EU-15, Badinge (2005) has ocused
on wo po en ial channels how economic in eg a ion a ec s
Δyi
as: i) a ech-
nology channel
(
ΔA
i
=𝛾
A0
+𝛾
A1
ΔINT
i )
and ii) a physical in es men chan-
nel
(
Δk
i
=𝛾
k0
+𝛾
k1
ΔINT
i )
wi h
Δ
INT being an indica o o changes in he
le el o in eg a ion a ime ;
𝛾A0
and
𝛾k0
a e exogenous componen s o echno-
logical p og ess and capi al o ma ion, espec i ely. This logic can be s aigh -
o wa dly ex ended o he in eg a ion e ec s o o he inpu s such as o inpu as
(
Δn
,i
=γ
,n0
+γ
,n1
ΔINT
i )
and we can measu e he ela i e pe o mance o bo -
de egions o hese inpu s sepa a ely.
Al e na i ely, he inpu channels can be agg ega ed o an o e all e ec o
economic in eg a ion on pe capi a income g ow h as
wi h
𝛿0
=
�
𝛾
A0
+𝛼𝛾
k0
+
∑
=1
𝜑𝛾
,n0�
and
𝛿
1=
�
𝛾A1+𝛼𝛾k1+
∑
R
=1𝜑𝛾 ,n1
�
.
Gi en ou ocus on bo de egions, Eq.(2) can be ex ended by inco po a ing a
spa ial componen in o he analysis o g ow h e ec s om economic in eg a ion as
whe e
(
1
DIST𝜃
i)
measu es p oximi y o each egion i o he newly in eg a ed uni
(wi h
DISTi
being some dis ance measu e o he in eg a ion bo de o a speci ic
poin o in e es ac oss he bo de ). Equa ion(3) hus spli s he g ow h e ec s o
in eg a ion
𝛿1
in o a gene al non-spa ial componen
𝜌1
and a g ow h p emium o
egions wi h close p oximi y o he bo de (
𝜌2
) wi h
𝛿1=𝜌1+𝜌2
. While dis ance/
p oximi y can be measu ed in di e en dimensions (Boschma, 2005), we e ine o
geog aphical dis ance as a ca ch-all e m o o he o ms such as cul u al, social, and
his o ical p oximi y. This ex ension e lec s ha bene i s om economic in eg a ion
do no a ec each egion equally bu p edic s ha egions close o he in eg a ed
ma ke ecei e la ge bene i s as ypically ound in g a i y- ype models o in e -
egional ade such as in McCallum (1995).4 The pa ame e θ shown in Eq. (3)
exp esses he powe o dis ance decay. Fo ins ance, o su icien ly high alues o θ,
(1)
Δ
yi =ΔAi +𝛼Δki +
R
∑
=1
𝜑 Δnl,
i
(2)
Δyi =𝛿
0
+𝛿
1
ΔINTi
(3)
Δ
yi =𝛿0+
(
𝜌1+𝜌2
(
1
DIST𝜃
i))
ΔINT
i
4 Re e ing o he a gumen o in ensi ied in e - egional ade a e EU enla gemen , Figu eA1 in he
Supplemen a y Online Ma e ials p o ides an o e iew o ade lows be ween Ge man NUTS1 egions
and hei wo neighbo ing Eas e n coun ies (Poland and Czechia). Wha can be seen is ha hose Ge -
1449
The complex egional e ec s o mac o‑ins i u ional change:…
we expec o only obse e a spa ial g ow h p emium o egions di ec ly adjacen o
he enla gemen bo de . Ways o empi ically p oxy he spa ial p oximi y o he
enla gemen bo de will be p esen ed below.
The ole o dis ance decay as a ac o de e mining ade cos and e en ually ou pu
e ec s om economic in eg a ion is also s essed in models o he New Economic
Geog aphy (NEG). K ugman and Venables (1990), o ins ance, show o an NEG
model applica ion o he EU single ma ke in 1992 ha wi h educed anspo cos s
mo e i ms may ind i a ac i e o eloca e o he pe iphe y as a way ake ad an age
o ac o p ice di e en ials be ween coun ies. O he NEG models simila ly p edic
ha egions wi h a lowe dis ance and hus anspo cos o in e na ional ma ke s
eap he la ges bene i s om economic in eg a ion (B ülha e al., 2004; C oze &
Koenig, 2004).
Beh ens e  al. (2007) and Mon o and Nicolini (2000) show in NEG model
se ings ha a coun y’s in e nal economic geog aphy cons i u es a signi ican
condi ioning ac o o he egional economic e ec s o in e na ional economic
in eg a ion. Fo ins ance, Rauch (1991) p esen s a model in which cos al bo de
egions a e he main ade hub o a coun y. In his case, bo de egions can
pa icula ly bene i om ade in eg a ion. O e man and Win e s (2006), s udy
he impac o UK accession o he la ge Eu opean ma ke and ind e idence o
his se up indica ing ha coas al (bo de ) egions hos ing a po wi h be e ma ke
access o expo s and in e media e inpu s expe ience highe employmen compa ed
o o he simila egions.
I bo de egions su e om loca ional disad an ages, model p edic ions
may di e , hough. Wi hou scale e ec s emana ing om loca ional ad an ages,
consume s ypically ha e o pay highe p ices and i ms can only supply goods o
he ma ke a highe cos when being loca ed in a bo de egion (Niebuh & S ille ,
2004). Inc eased p oximi y o o eign ma ke s o in eg a ing coun ies hen only
allows bo de egions o g ow as e han non-bo de egions i hey possess speci ic
e i o ial asse s (Capello e  al., 2018a). I such asse s a e missing, he e is he
isk o a ‘ unnel e ec ’, i.e., a bypassing o bo de egions a e in eg a ion, which
could u he ma ginalize bo de egions i ade pa e ns a e EU enla gemen a e
domina ed by cen al co e egions (Pe akos & Topaloglou, 2008). In his case,
𝜌2
can be expec ed o be ze o o e en nega i e.
2.2 P io empi ical e idence and emaining esea ch gaps
Se e al empi ical con ibu ions ha e been conce ned wi h he iden i ica ion o
g ow h e ec s o economic in eg a ion – p edomina ely a he na ional le el (e.g.,
man NUTS1 egions loca ed in geog aphical icini y o he enla gemen bo de expe ienced a much
s onge expo g ow h o Poland and Czech Republic a e 2004 han o he Ge man egions. In line
wi h his s ylized inding and wi h ega d o he expec ed economic e ec s o EU in eg a ion o bo de
egions, i can be hypo hesized ha imp o ed c oss-bo de exchange inc eases he egions’ po en ial o
economic de elopmen (see also EU commission, 2001; B ülha e al., 2004; Hanson, 2005; B ülha ,
2011).
Foo no e 4 (con inued)
1450
T.Mi ze, P.B eidenbach
Ande sen e al., 2019; Badinge , 2005; Campos e al., 2019; Hen ekson e al., 1997).
B idging he gap be ween he a ailable na ional and sca ce egional-le el e idence,
Monas i io is e al. (2017) analyze he spa ial e ec s o EU in eg a ion o Cen al
and Eas e n Eu opean (CEE) egions. Using an e en -s udy app oach, he au ho s
ind ha he p ocess o EU accession has pa icula ly s eng hened agglome a ion
o ces in CEE coun ies a o ing egions wi h a high ma ke po en ial, indus y
concen a ion and egional specializa ion in inc easing e u ns sec o s.5
B ülha e al. (2012) and B ülha e al. (2018) analyze he wage and employmen
e ec s o ade libe aliza ion caused by he all o he i on cu ain o Aus ian bo de
owns. Thei empi ical esul s indica e ha imp o ed access o Eas e n ma ke s has
a posi i e impac on employmen and nominal wages in hese egions is-à- is he
es o he coun y. The esul s in B ülha e  al. (2018) addi ionally sugges ha
la ge ci ies bene i mo e s ongly om he bo de shock in e ms o wages, whe eas
smalle ci ies expe ience la ge employmen e ec s wi h a peak o owns wi h a
popula ion o a ound 150,000. Taken oge he , hei e idence sugges s ha esiden s
o medium-sized owns gain he mos om a gi en opening o c oss-bo de ade.
B akman e al. (2012) ocus on he popula ion e ec s o EU in eg a ion in EU
bo de egions. Analyzing da a o 1457 egions and 2410 ci ies since 1973, he
au ho s ind e idence o posi i e popula ion g ow h e ec s in bo de egions is-
à- is non-bo de egions. This e ec is signi ican a he egional and u ban le el
wi hin a 70km adius om na ional bo de s. I holds o bo h sides o he in eg a ion
bo de . Rela edly, Heide (2019) ocusses on he popula ion g ow h e ec s o
Ge man and Polish bo de own in he cou se o he EU enla gemen in 2004. The
au ho inds e idence o posi i e popula ion g ow h e ec s o Ge man bu no o
Polish bo de owns.
While he majo i y o s udies hus epo s posi i e popula ion and economic
e ec s o ade libe aliza ion and economic in eg a ion in bo de egions o he
EU (pa icula ly in he EU15), he e is also empi ical e idence o insigni ican o
nega i e e ec s as, o ins ance, epo ed in B aakmann and Vogel (2011) o Ma in
(2011). Using da a o i ms loca ed in Eas Ge many close o Ge many’s eas e n
bo de , B aakmann and Vogel (2011) ind no sho - un employmen e ec s o he
EU enla gemen in 2004 excep o i ms ac i e in wholesale and e ail ade, ho els,
and es au an s. Nega i e wage e ec s a e ound o skilled wo ke s in consul -
ing, esea ch, and ela ed ac i i ies. This poin s o sec o -speci ic e ec s in bo de
egions subjec o EU enla gemen . S udying employmen g ow h om he pe spec-
i e o i ms in Cen al and Eas e n Eu opean Coun ies (CEECs), Se wicka e al.
(2022) ind a signi ican inc ease in o eign in es men and employmen g ow h
a e he 2004 EU enla gemen .
While he p io li e a u e has s a ed o shed ligh on egional e ec s o EU in e-
g a ion o selec ed ou comes, mainly income le els and indi idual enla gemen
5 Niebuh (2008) adds o his inding by s udying he income e ec s o EU enla gemen in 2004 using a
h ee- egion economic geog aphy model calib a ed wi h p e-accession da a o 1995–2000. The simula-
ion esul s indica e ha bo de egions ealize highe in eg a ion bene i s han non-bo de egions wi h
he s onges e ec s ound o Cen al and Eas e n Eu opean (CEE) egions along he o me ex e nal
EU15 bo de .
1457
The complex egional e ec s o mac o‑ins i u ional change:…
Table 1 De ini ions and summa y s a is ics o a iables used in he empi ical analysis o 1289 NUTS3 pe iods du ing 1981–2014
Va iable Desc ip ion Mean S.D Min Max
GDPpc Pe capi a GDP (in 1000 Eu o, in 2005 p ices) 20.114 11.070 1.083 188.679
Yp od Labo p oduc i i y de ined as g oss alue added (GVA) pe employee (in 1000 Eu o) 40.974 17.292 1.238 310.492
Yp od (ag icul u e) > Ag icul u e (NACE Re . 2, Sec o code A) 22.777 35.174 0.002 2048.154
Yp od (cons uc ion) > Cons uc ion sec o (F) 37.553 18.350 0.209 420.587
Yp od (indus y excl. cons uc ion) > Indus y excl. cons uc ion (B-E) 50.004 37.242 0.538 1157.735
Yp od (WR se ices, I&C) > Wholesale, e ail, anspo , accommoda ion & ood se ices, in o ma ion, and
communica ion (G-J)
35.210 15.872 1.044 299.409
Yp od ( inancial & business se ices) > Financial & business se ices (K-N) 94.973 62.386 1.094 2028.800
Yp od (non-ma ke se ices) > Non-ma ke se ices (O-U) 33.515 14.325 0.485 287.712
Pa en Pa en applica ions pe 1000 inhabi an s in egion 0.081 0.225 0 5.729
Emp a e Employmen pe popula ion in egion (1 = 100%) 0.433 0.108 0.121 1.224
Emp a e (ag icul u e) > Ag icul u e (NACE e . 2, sec o code A) 0.037 0.054 0 0.816
Emp a e (cons uc ion) > Cons uc ion sec o (F) 0.033 0.013 0 0.188
Emp a e (indus y w/o. Cons uc ion) > Indus y excl. cons uc ion (B-E) 0.092 0.052 0.001 0.539
Emp a e (WR se ices, I&C) > Wholesale, e ail, anspo , accommoda ion & ood se ices, in o ma ion, and
communica ion (G-J)
0.108 0.042 0.007 0.362
Emp a e ( inancial & business se ices) > Financial & business se ices (K-N) 0.046 0.033 0.000 0.397
Emp a e (non-ma ke se ices) > Non-ma ke se ices (O-U) 0.118 0.048 0.006 0.417
Empsha e (ag icul u e) Sec o al employmen sha e o ag icul u e in o al egional employmen (1 = 100%) 0.090 0.116 0 0.920
Empsha e (cons uc ion) Sec o al employmen sha e o cons uc ion in o al egional employmen (1 = 100%) 0.077 0.032 0 0.442
Empsha e (indus y excl. Cons uc ion) Sec o al employmen sha e o indus y excl. cons uc ion (manu ac u ing sec o s) in
o al egional employmen (1 = 100%)
0.212 0.100 0.002 0.799
Empsha e (WR se ices, I&C) Sec o al employmen sha e o wholesale and e ail se ices, in o ma ion, and
communica ion in o al egional employmen (1 = 100%)
0.249 0.063 0.017 0.610
Empsha e ( inancial & business se ices) Sec o al employmen sha e o inancial and business se ices in o al egional
employmen (1 = 100%)
0.101 0.054 0.000 0.726

1458
T.Mi ze, P.B eidenbach
Table 1 (con inued)
Va iable Desc ip ion Mean S.D Min Max
Empsha e (non-ma ke se ices) Sec o al employmen sha e o non-ma ke se ices in o al egional employmen
(1 = 100%)
0.272 0.081 0.022 0.644
Pop Popula ion le el o NUTS3 egion (in 1000 pe sons) 370.476 428.355 6.748 6418.41
Nligh Nigh ligh emission le el pe NUTS3 egion 17.251 15.536 0 63
Bo de egions (EU enla gemen 1986) Bina y dummy o di ec bo de egions o EU enla gemen in 1986 (see Fig.1) 0.008 0.088 0 1
Bo de egions (EU enla gemen 1995) Bina y dummy o di ec bo de egions o EU enla gemen in 1995 (see Fig.1) 0.026 0.160 0 1
Bo de egions (EU enla gemen 2004) Bina y dummy o di ec bo de egions o EU enla gemen in 2004 (see Fig.1) 0.046 0.209 0 1
Bo de egions (EU enla gemen 2007) Bina y dummy o di ec bo de egions o EU enla gemen in 2007 (see Fig.1) 0.007 0.083 0 1
Da a a he le el o NUTS3 egions; all da a (excep o nigh ligh emissions and pa en applica ions) ha e been ga he ed om he Eu opean egional da abase (ERD) o
Camb idge Econome ics ( e sion 2017). Nigh ligh emissions ha e been ex ac ed om: h ps:// ligh ends. ligh pollu ionm ap. in o; pa en s ex ac ed om he OECD
RegPa da abase, see Ma au e al. (2008)
1459
The complex egional e ec s o mac o‑ins i u ional change:…
Non-bo de egions
Indi ec [300km]
Indi ec [250km]
Indi ec [200km]
Indi ec [150km]
Indi ec [100km]
Di ec bo de egions
Fig. 3 Di ec and indi ec bo de egions o EU enla gemen 1986, 1995, 2004 and 2007. In o ma ion
on he e i o ial bo de s o EU-27 (including UK, wi hou C oa ia) NUTS3 egions has been ob ained
om he GISCO s a is ical uni da ase a ailable a : h ps:// ec. eu opa. eu/ eu os a / web/ gisco/ geoda a/ e e
ence- da a/ admin is a i e- uni s- s a i s ical- uni s/ nu s. Maps o bo de egions by EU enla gemen wa e
and egion ype (es ablished s new) a e p o ided in Figu eA2 in he appendix
Gi en he empo al dis ibu ion o EU enla gemen e en s h oughou ou sam-
ple pe iod 1981–2014, we can es ima e dynamic ea men e ec s o a maximum
o i e yea s p io o and se en yea s a e he ins i u ional changes o all ou EU
accession wa es (excep o nigh ligh emissions, which is only a ailable om 1992
1460
T.Mi ze, P.B eidenbach
onwa ds).10 While i would be p e e able o ex end he da a o pe iods beyond 2014
and also include C oa ia’s EU accession, he e a e also easons o es ic he sam-
ple o 2014. Pa icula ly, he EU mig a ion c isis o 2015 and 2016 wi h subs an ial
mig a ion lows o bo de egions o se e al EU coun ies may bias a leas all ou -
comes on “pe capi a” le els.
To measu e he deg ee o spa ial he e ogenei y and neighbo hood e ec s, we
de ine indi ec bo de egions based on hei geog aphical dis ance om he bo -
de . To do so, we calcula e o all egions no classi ied as di ec bo de egions he
geog aphical dis ance om he egion’s cen oid o he closes loca ion a he bo -
de . Using 50km h eshold dis ances g wi h k = {100km, 150km, …, 300km},
we hen build addi ional ea men g oup dummies o egions wi hin hese 50km
dis ance bel s om he bo de and es o spa ially dis ibu ed in eg a ion e ec s
(wi h k = 0km being di ec bo de egions along he in eg a ion bo de ).11 A g aphi-
cal o e iew o di ec and indi ec bo de egions o ou sample o 1289 NUTS3
egions o all ou EU enla gemen wa es is gi en in Fig.3.
5 Empi ical esul s
5.1 Baseline es ima es
Table2 epo s he es ima ion ou pu o ou pooled s a ic DiD speci ica ion acco d-
ing o Eq.(4). Accoun ing o he mul idimensional ‘ ixed e ec s’ s uc u e including
coun y-speci ic ime ends as mos gene al speci ica ion o accoun o la en ime-
a ying con ounding ac o s, h ee signi ican indings eme ge: Fi s , bo de egions
ela i ely inc ease hei labo p oduc i i y ela i e o non-bo de egions a e ea men
(Panel A) and ha e highe le els o nigh ligh emissions (Panel E) as a gene al measu e
o agglome a ion e ec s (Mellande e al., 2015). While he speci ica ions shown in
columns (I) and (II) he eby use he a e age de elopmen o non-bo de egions in he
sample as benchma k, he inclusion o coun y-speci ic ime ends igh ens he bench-
ma k o non-bo de egions in he espec i e coun y o bo de egions conside ed.
E ec size poin s o a oughly 4–5% ela i e inc ease in he le els o labo p oduc i i y
and nigh ligh emissions. In e ms o labo p oduc i i y, his co esponds o an app ox.
inc ease o 1500–2000 Eu o pe wo ke inc ease e alua ed a he sample a e age o
labo p oduc i i y o 40,800 Eu o pe wo ke in non- ea ed egions. Since nigh ligh
emissions le els a e measu ed on a ela i e scale be ween 0 and 63, he abo e epo ed
pe cen age inc ease is di icul o in e p e . Howe e , we can illus a e i s magni ude
wi h he help o an example. E alua ed agains he sample a e age o nigh ligh emis-
sions in non- ea ed egions o abou 20, a 4–5% inc ease means an addi ional nigh
10 Fo his eason, we also exclude he 1981 EU accession o G eece as addi ional ea men . Besides, by
he ime o EU accession, G eece did no sha e any e i o ial bo de wi h an es ablished EU coun y so
ha no ea men g oup can be iden i ied he e.
11 All dis ances a e calcula ed based on he egions’ cen oids. We me ge he i s wo slices o 50km
and 100km dis ances as he e a e ew indi ec bo de egions wi h a maximum dis ance o 50km o he
enla gemen bo de .
1461
The complex egional e ec s o mac o‑ins i u ional change:…
ligh emission le el o 0.8–1. The la e co esponds o an accumula ion o nigh ligh
in ensi ies o g owing me opoli an egions such as Mad id and Hambu g o abou
Table 2 Baseline ea men
e ec s o EU enla gemen o
di ec bo de egions
***,**,* = deno e signi icance a he 1%, 5% and 10% c i ical le el;
obus s anda d e o s clus e ed a he egional le el a e gi en in
b acke s. Sample pe iod 1981–2014; 1289 NUTS3 egions. See
Table 1 o de ails on ou come a iable de ini ions. Va iables
a e log ans o med; in he case o pa en s pe capi a a box-cox
ans o ma ion has been applied as log ans o ma ion did no mee
he no mali y assump ion. The di e ence in obse a ions be ween
columns (I), (II) and column (III) is due o missing alues o
egional con ols (sec o al employmen sha es)
Speci ica ion (I) (II) (III)
Panel A: GDP pe capi a
EU enla gemen 0.0475***
(0.01215)
0.0313**
(0.01274)
0.0258**
(0.01280)
R20.65 0.78 0.79
Obs 42,012 42,012 41,974
Panel B: labo p oduc i i y
EU enla gemen 0.0719***
(0.01330)
0.0530***
(0.01350)
0.0492***
(0.01256)
R20.52 0.72 0.74
Obs 42,012 42,012 41,974
Panel C: pa en s pe capi a
EU enla gemen 0.0281***
(0.00552)
0.0115**
(0.00537)
0.0108**
(0.00539)
R20.32 0.37 0.37
Obs 42,012 42,012 41,974
Panel D: employmen a e
EU enla gemen − 0.0252**
(0.00998)
− 0.0268***
(0.00979)
–0.0277***
(0.00948)
R20.15 0.38 0.46
Obs 42,012 42,012 41,974
Panel E: popula ion
EU enla gemen − 0.0086
(0.00822)
− 0.0012
(0.00845)
− 0.0064
(0.00784)
R20.18 0.41 0.43
Obs 42,012 42,012 41,974
Panel F: nigh ligh emissions
EU enla gemen 0.0804***
(0.01620)
0.0474***
(0.01247)
0.0427***
(0.01203)
R20.51 0.80 0.81
Obs 27,065 27,065 27,041
Region FE Yes Yes Yes
Time FE Yes Yes Yes
Time × C y FE No Yes Yes
Regional con ols No No Yes
1462
T.Mi ze, P.B eidenbach
5–7yea s. Panel C also epo s an inc ease in egional pa en applica ions pe capi a
ollowing EU enla gemen , which poin s o he wo king o he echnology channel
ΔAi
o economic in eg a ion as ou lined in Eq.(1). Howe e , he de elopmen o he
employmen a e alls behind he o e all EU- end du ing he sample pe iod by a ound
3%-poin s ( o an a e age employmen a e o app ox. 43% in ou da a sample). No sig-
ni ican e ec s a e obse ed o egional popula ion de elopmen .
5.2 S uc u al he e ogenei y
Di e ences in ea men e ec s may be d i en by s uc u al he e ogenei y ac oss
he g oup o bo de egions and unde lying composi ional e ec s associa ed wi h
egion-sec o combina ions, which may no be ully cap u ed by ou se o egional
con ols (sec o al employmen sha es). To gain a deepe unde s anding o he unde -
lying mechanisms a play, we disagg ega e e ec s by enla gemen wa es (Panel
A o Table3) and coun y g oups, ha is, we dis inguish be ween e ec s o bo -
de egions in old (es ablished) and new membe s a es o each EU enla gemen
wa e (Panel B o Table3) as ou lined in Fig.1. Especially he enla gemen wa es
in 2004 and 2007 saw la ge s uc u al di e ences be ween es ablished EU membe
coun ies and CEECs in hei ansi ion om planned o ma ke economies a e he
all o he i on cu ain. This mean ha pe capi a income le els, labo p oduc i i y
and labo ma ke pa ame e s we e signi ican ly di e en in es ablished EU membe
coun ies and newly joining CEECs in 2004 and 2007.
As Panel A shows, he es ima ed posi i e ea men e ec s o labo p oduc i i y and
nigh ligh emissions a e mainly d i en by he 1995 and 2004 enla gemen wa es. Nega-
i e e ec s on he employmen a e a e simila ly ound o he 1995 and 2004 enla gemen
wa es in pa icula . The la e also show a decline in popula ion de elopmen and egional
inno a i eness in bo de egions. A posi i e de elopmen in e ms o egional inno a-
i eness measu ed h ough pa en s pe capi a is ound o he 1986 enla gemen wa e.
Panel B o Table3 u he indica es ha es ima ed ea men e ec s no only di e ac oss
enla gemen wa es bu a e also de e mined by he egions’ de elopmen le el and, hence,
he egion’s abso p i e capaci y a he iming o in eg a ion. Posi i e p oduc i i y e ec s
o he 1995 and 2004 enla gemen wa e a e cap u ed by bo de egions in es ablished
EU coun ies. Fo hese egions he 1995 enla gemen also induced gene al agglome a-
ion e ec s measu ed in e ms o a posi i e popula ion de elopmen and inc eases in nigh
ligh emissions ela i e o non-bo de egions.
This e ec is less signi ican o bo de egions in es ablished EU membe
coun ies du ing he i h enla gemen wa e in 2004. Rela i e popula ion le els
a e obse ed o decline in he p ocess o EU in eg a ion. On he o he hand, bo -
de egions in new EU membe s a es g ow in e ms o inno a i eness (1986) and
gene al agglome a ion e ec s (nigh ligh emissions) in 2004 oge he wi h a s ong
inc ease in he employmen a e by app oxima ely 17% in 2007. The la e e ec
is likely d i en by pe sis en wage di e ences be ween es ablished EU coun ies
(G eece) and he new membe s a es Bulga ia and Romania who join in 2007. Only
du ing he 1986 EU enla gemen do bo de egions in new membe coun ies see a
ela i e p oduc i i y inc ease and a imp o emen s in inno a i eness.

1463
The complex egional e ec s o mac o‑ins i u ional change:…
Table 3 T ea men e ec s o di e en EU enla gemen wa es and coun y g oups
Speci ica ion (I) (II) (III) (IV) (V) (VI)
Ou come GDP pe capi a Labo p oduc i i y Pa en s pe capi a Employmen a e Popula ion Nigh ligh emissions
Panel A: symme ic e ec s ac oss bo de egions in es ablished and new EU membe s a es
Bo de egion × 0.0298 0.0326 0.0515*** 0.0030 0.0091 n.a
Enla gemen 1986 (0.01856) (0.02164) (0.01585) (0.01378) (0.03198)
Bo de egion × 0.0324 0.0532** 0.0079 –0.0309** 0.0371*** 0.0867***
Enla gemen 1995 (0.02370) (0.02201) (0.0092) (0.01235) (0.00950) (0.02896)
Bo de egion × 0.0207 0.0469*** 0.0112 − 0.0274*** − 0.0409*** 0.0326**
Enla gemen 2004 (0.01593) (0.01300) (0.00825) (0.01046) (0.00956) (0.01344)
Bo de egion × − 0.0319 − 0.0049 − 0.0215** − 0.0331 − 0.0084 0.0250
Enla gemen 2007 (0.03533) (0.01310) (0.01033) (0.07558) (0.02406) (0.03904)
R20.79 0.74 0.37 0.46 0.44 0.81
Obs 41,974 41,974 41,974 41,974 41,974 27,041
Panel B: asymme ic e ec s ac oss bo de egions in es ablished and new EU membe s a es
Bo de egion (old) × 0.0379 0.0031 0.0140 0.0134 0.0344 n.a
Enla gemen 1986 (0.03066) (0.02905) (0.01384) (0.01176) (0.04181)
Bo de egion (old) × 0.0764** 0.0978*** 0.0109 − 0.0268* 0.0308** 0.0969***
Enla gemen 1995 (0.03176) (0.02774) (0.01045) (0.01538) (0.01248) (0.03437)
Bo de egion (old) × 0.0370* 0.0543*** 0.0166* − 0.0175 − 0.0582*** 0.0216
Enla gemen 2004 (0.01944) (0.01612) (0.01002) (0.01142) (0.01218) (0.01592)
Bo de egion (old) × − 0.0451 0.1085 − 0.0229 − 0.1778** − 0.0049 − 0.0216
Enla gemen 2007 (0.05206) (0.10202) (0.01648) (0.08244) (0.02981) (0.04854)
Bo de egion (new) × 0.0214 0.0633** 0.0904*** − 0.0078 − 0.0171 n.a
Enla gemen 1986 (0.01993) (0.02683) (0.01737) (0.02473) (0.04609)
Bo de egion (new) × − 0.0526*** − 0.0345 0.0023 − 0.0371* 0.0487*** 0.0666
Enla gemen 1995 (0.02037) (0.02488) (0.01754) (0.01998) (0.01268) (0.05150)
1464
T.Mi ze, P.B eidenbach
***,**,* = deno e signi icance a he 1%, 5% and 10% c i ical le el; obus s anda d e o s clus e ed a he egional le el a e gi en in b acke s. Sample pe iod 1981–2014;
1289 NUTS3 egions. See Tables1, 2 o u he de ails on he de ini ion o ou come a iables. The e ms (old) and (new) e e o bo de egions in es ablished and new
EU s a es, espec i ely
Table 3 (con inued)
Speci ica ion (I) (II) (III) (IV) (V) (VI)
Ou come GDP pe capi a Labo p oduc i i y Pa en s pe capi a Employmen a e Popula ion Nigh ligh emissions
Bo de egion (new) × − 0.0341 0.0141 − 0.0033 − 0.0533** 0.0038 0.0558**
Enla gemen 2004 (0.02425) (0.02179) (0.01275) (0.02082) (0.00735) (0.02381)
Bo de egion (new) × − 0.013 − 0.1738*** − 0.0193** 0.1816*** − 0.014 0.0865
Enla gemen 2007 (0.04175) (0.06490) (0.00792) (0.05834) (0.03987) (0.05343)
R20.79 0.74 0.37 0.46 0.44 0.81
Obs 41,974 41,974 41,974 41,974 41,974 27,041
Region FE Yes Yes Yes Yes Yes Yes
Time FE Yes Yes Yes Yes Yes Yes
Time FE × C y FE Yes Yes Yes Yes Yes Yes
Regional con ols Yes Yes Yes Yes Yes Yes
1465
The complex egional e ec s o mac o‑ins i u ional change:…
5.3 Tempo al he e ogenei y
S a ic ea men eg essions may be biased i es ima ed e ec s show signi ican pa -
e ns o ea ly an icipa ion o g adual phasing-in. Figu e4 he e o e plo s he esul s
o a lexible DiD app oach, which es ima es yea ly ea men e ec s ela i e o he
iming o EU enla gemen . The p e-enla gemen yea ( − 1) is used a e e ence yea
agains which p e- and pos -enla gemen e ec s a e e alua ed. The esul s la gely
con i m he s a ic ea men e ec es ima es in e ms o posi i e and signi ican
e ec s o labo p oduc i i y and nigh ligh emissions. In addi ion, Panel A o Fig.4
also epo s a posi i e and signi ican ela i e GDP pe capi a de elopmen in EU
in e nal bo de is-à- is non-bo de egions. Maximum e ec size o he 7-yea lag
pe iod conside ed is an GDP pe capi a inc ease o abou 2% (compa ed o 2.6% in
he baseline s a ic es ima ion app oach). Annual ea men e ec s o labo p oduc-
i i y le els a e ound o ange be ween 2 and 4% du ing he i s se en yea s a e
EU enla gemen . The empo al dis ibu ion o GDP and p oduc i i y e ec s poin
a a le elling ou o addi ional g ow h e ec s a e app oxima ely 5–7yea s, which
suppo s he iew o a medium- e m g ow h bonus associa ed wi h EU in eg a ion
(Baldwin & Wyplosz, 2015; in’ Veld, 2019).
Annual ea men e ec s p io o EU enla gemen a e s a is ically insigni ican
and do no poin o ea ly an icipa ion e ec s associa ed wi h po en ial con ounding
ac o s a ound ea men s a . E ec s u n signi ican wi h a ime lag o 4–5yea s
a e he enla gemen e en . This indica es ha posi i e economic e ec s om eco-
nomic in eg a ion need o un old un il hey a e ully isible in he egional economy.
Likely easons o his g adual phasing-in p ocess a e ha associa ed p i a e and
public in es men e ec s ypically only show up o e ime (B eidenbach e al., 2019;
Ebe le e al., 2019). The di e ence be ween he s a is ically signi ican s a ic es ima-
ion esul s o he de elopmen o pa en s pe capi a and he insigni ican annual
dynamic e ec s in he i s se en yea s a e enla gemen unde line he ole o g ad-
ual phasing in e ec s and he ime needed o ans o m egional inno a ion sys ems
in ea ed egions (Isaksen & T ippl, 2016).
Bu no only did echnology ans o ma ion and he adap ion o p oduc ion sys-
ems akes ime, also labo ma ke opening a e EU enla gemen ollowed a g adual
pa e n, pa icula ly o he EU eas e n enla gemen wa es in 2004 and 2007, de e -
mined, o ins ance, by he 2 + 3 + 2 egula ion ha es ic ed employmen access
in (some) incumben EU membe s a es by ci izens o new EU membe a e an up
o se en yea s ansi ion pe iod.12 Also, he Schengen en y o new membe coun-
ies ollowed EU accession wi h a empo al lag o abou h ee yea s. Compa ed
o he s a ic baseline case, he lexible DiD es ima es show nega i e, albei ma gin-
ally s a is ically insigni ican e ec s o EU enla gemen on he employmen a e in
in e nal bo de egions (Panel C, e alua ed a 95% con idence in e als). Finally, he
dynamic es ima es con i m posi i e and s a is ically signi ican inc eases in nigh
ligh ac i i y as gene al agglome a ion e ec in bo de egions.
12 Fo he case o he la ges enla gemen wa e in 2004, only wo s a es (Ge many and Aus ia) ha e u i-
lized he ull 7-yea du a ion o he 2 + 3 + 2 ule.
1466
T.Mi ze, P.B eidenbach
Panel A: GDP pe capi a Panel B: Labo p oduc i i y
Panel C: Pa en s pe capi a Panel D: Employmen a e
Panel E: Popula ion le el Panel F: Nigh ligh emissions
Fig. 4 Dynamic ea men e ec s o EU enla gemen on bo de egions. Diamonds show poin es ima es
o annual ea men e ec s o EU enla gemen in bo de egions oge he wi h 95% con idence in e als
( e ical lines; based on obus s anda d e o s clus e ed a he egional le el). The e ical dashed line
indica es he p e-enla gemen yea ( − 1) used a e e ence yea agains which p e- and pos -enla gemen
e ec s a e e alua ed. Unde lying lexible DiD es ima es include egion FE, yea FE, coun y-yea FE and
egional con ols. Fo u he de ails see main ex . Sample pe iod 1981–2014; 1289 NUTS3 egions
1473
The complex egional e ec s o mac o‑ins i u ional change:…
Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License,
which pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long
as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e
Commons licence, and indica e i changes we e made. The images o o he hi d pa y ma e ial in his
a icle a e included in he a icle’s C ea i e Commons licence, unless indica ed o he wise in a c edi line
o he ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons licence and you in ended
use is no pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will need o ob ain pe mis-
sion di ec ly om he copy igh holde . To iew a copy o his licence, isi h p://c ea i ecommons.o g/
licenses/by/4.0/.
Re e ences
Ah end, R., Fa chy, E., Kaplanis, I., & Lembcke, A. (2017). Wha makes ci ies mo e p oduc i e? E i-
dence om i e OECD coun ies on he ole o u ban go e nance. Jou nal o Regional Science, 57,
385–410.
Ande sen, T. B., Ba slund, M., & Vanhuysse, P. (2019). Join o p ospe ? An empi ical analysis o EU
membe ship and economic g ow h. Kyklos, 72(2), 211–238.
Ashen el e , O. (1978). Es ima ing he e ec o aining p og ams on ea nings. Re iew o Economics and
S a is ics, 6(1), 47–57.
Badinge , H. (2005). G ow h e ec s o economic in eg a ion: E idence om he EU membe s a es.
Re iew o Wo ld Economics, 141(1), 50–78.
Baldwin, C., & Wyplosz, C. (2015). Economics o Eu opean in eg a ion (5 h ed.). McG aw-Hill.
Beh ens, K., Gaigne, C., O a iano, G., & Thisse, J. (2007). Coun ies, egions and ade: On he wel a e
impac s o economic in eg a ion. Eu opean Economic Re iew, 51(5), 1277–1301.
Bo usyak, K., Ja a el, X. (2020). Re isi ing e en s udy designs. Wo king pape , Ha a d Uni e si y
(Ap il 25, 2020). A ailable a : h ps:// schol a . ha a d. edu/ bo us yak/ publi ca io ns/ e is i ing- e en -
s udy- desig ns.
Boschma, R. (2005). P oximi y and inno a ion: A c i ical assessmen . Regional S udies, 39, 61–74.
Boske , M., B akman, S., Ga e sen, H., & Sch amm, M. (2010). Adding geog aphy o he new economic
geog aphy: B idging he gap be ween heo y and empi ics. Jou nal o Economic Geog aphy, 10(6),
793–823.
B aakmann, N., & Vogel, A. (2010). The impac o he 2004 EU enla gemen on he pe o mance o
se ice en e p ises in Ge many’s eas e n bo de egion. Re iew o Wo ld Economics, 146(1), 75–89.
B aakmann, N., & Vogel, A. (2011). How does economic in eg a ion in luence employmen and wages in
bo de egions? The case o he EU enla gemen 2004 and Ge many’s eas e n bo de . Re iew o Wo ld
Economics, 147(2), 303–323.
B akman, S., Ga e sen, H., an Ma ewijk, C., & Oume , A. (2012). The bo de popula ion e ec s o EU
in eg a ion. Jou nal o Regional Science, 52(1), 40–59.
B eidenbach, P., Mi ze, T., & Schmid , C. M. (2019). EU egional policy and he neighbou ’s cu se: ana-
lyzing he income con e gence e ec s o ESIF unding in he p esence o spa ial spillo e s. Jou nal
o Common Ma ke S udies, 57(2) 388–405.
B ülha , M. (2011). The spa ial e ec s o ade openness: A su ey. Re iew o Wo ld Economics, 86(1),
59–83.
B ülha , M., Ca è e, C., & Robe -Nicoud, F. (2018). T ade and owns: He e ogeneous adjus men o a bo -
de shock. Jou nal o U ban Economics, 105, 162–175.
B ülha , M., Ca è e, C., & T ion e i, F. (2012). How wages and employmen adjus o ade libe aliza ion:
Quasi-expe imen al e idence om Aus ia. Jou nal o In e na ional Economics, 86(1), 68–81.
B ülha , M., C oze , M., & Koenig, P. (2004). Enla gemen and he EU pe iphe y: The impac o changing
ma ke po en ial. The Wo ld Economy, 27(6), 853–875.
CallawaySan ’Anna, B. P. (2021). Di e ence-in-di e ences wi h mul iple ime pe iods. Jou nal o Econome -
ics, 225(2), 200–230.
Campos, N. F., Co icelli, F., & Mo e i, L. (2019). Ins i u ional in eg a ion and economic g ow h in Eu ope.
Jou nal o Mone a y Economics, 103, 88–104.

1474
T.Mi ze, P.B eidenbach
Capello, R., Ca agliu, A., & F a esi, U. (2018a). B eaking down he bo de : Physical, ins i u ional and cul-
u al obs acles. Economic Geog aphy, 94(5), 485–513.
Capello, R., Ca agliu, A., & F a esi, U. (2018b). Compensa ion modes o bo de e ec s in c oss-bo de
egions. Jou nal o Regional Science, 58(4), 759–785.
Capello, R., Ca agliu, A., & F a esi, U. (2018c). Measu ing bo de e ec s in Eu opean c oss-bo de egions.
Regional S udies, 52(7), 986–996.
Capello, R., Ca agliu, A., & Panze a, E. (2022). Economic cos s o COVID-19 o c oss-bo de egions.
Regional Science Policy & P ac ice, 15(8), 1688–1701.
Cla ke, D. (2017). Es ima ing di e ence-in-di e ences in he p esence o spillo e s. MPRA discussion pape
no. 81604, h ps:// mp a. ub. uni- muenc hen. de/ 81604/.
Eu opean Commission. (2001). On he impac o enla gemen on egions bo de ing candida e coun ies.
Communi y ac ion o bo de egions, B ussels.
Eu opean Commission. (2017). Boos ing g ow h and cohesion in EU bo de egions. A ailable a : h p://
ec. eu opa. eu/ egio nal_ policy/ en/ in o ma ion/ publi ca io ns/ commu nica ions/ 2017/ boos ing- g ow h- and-
cohes ion- in- eu- bo de - egio ns.
Eu opean Commission. (2021). EU bo de egions: Li ing labs o Eu opean in eg a ion. A ailable a : h ps://
ec. eu opa. eu/ egio nal_ policy/ en/ in o ma ion/ publi ca io ns/ epo s/ 2021/ eu- bo de - egio ns- li ing- labs-
o - eu op ean- in eg a ion.
C oze , M., & Koenig, P. (2004). EU enla gemen and he in e nal geog aphy o coun ies. Jou nal o Com-
pa a i e Economics, 32, 265–279.
Dange ield, M. (2006). Sub egional in eg a ion and EU enla gemen : Whe e nex o CEFTA? Jou nal o
Common Ma ke S udies, 44, 305–324.
Ea on, J., & Ko um, S. (2002). Technology, geog aphy, and ade. Econome ica, 70(5), 1741–1779.
Ebe le, J., B enne , T., & Mi ze, T. (2019). A look behind he cu ain: Measu ing he complex economic
e ec s o egional s uc u al unds in Ge many. Pape s in Regional Science, 98(2), 701–735.
Floe kemeie , H., Spa a o a, N., Venables, A. (2021). Regional dispa i ies, g ow h, and inclusi eness. IMF
W21/38.h ps:// doi. o g/ 10. 5089/ 97815 13569 505. 001.
Galimbe i, J. K. (2020). Fo ecas ing GDP g ow h om ou e space. Ox o d Bulle in o Economics and S a-
is ics, 82, 697–722.
Goodman-Bacon, A. (2021). Di e ence-in-di e ences wi h a ia ion in ea men iming. Jou nal o Econo-
me ics, 225(2), 254–277.
Hanson, G. (2005). Ma ke po en ial, inc easing e u ns and geog aphic concen a ion. Jou nal o In e na-
ional Economics, 50(2), 259–287.
Heide , B. (2019). The impac o EU Eas e n enla gemen on u ban g ow h and decline: New insigh s om
Ge many’s Eas e n bo de . Pape s o Regional Science, 98(3), 1443–1468.
Hende son, V., S o eyga d, A., & Weil, D. (2012). Measu ing economic g ow h om ou e space. Ame ican
Economic Re iew, 102(2), 994–1028.
Hen ekson, M., To s ensson, J., & To s ensson, R. (1997). G ow h e ec s o Eu opean in eg a ion. Eu opean
Economic Re iew, 41(8), 1537–1557.
in Veld, J. (2019). The economic bene i s o he EU single ma ke in goods and se ices. Jou nal o Policy
Modeling, 41(5), 803–818.
Isaksen, A., & T ippl, M. (2016). Pa h de elopmen in di e en egional inno a ion sys ems. In M. D. Pa -
illi, R. D. Fi ja , & A. Rod iguez-Pose (Eds.), Inno a ion d i e s and egional inno a ion s a egies.
London: Rou ledge.
Kashiha, M., Depken, C., & Thill, J. C. (2017). Bo de e ec s in a ee- ade zone: E idence om Eu opean
wine shipmen s. Jou nal o Economic Geog aphy, 17(2), 411–433.
K ugman, P. (1991). Inc easing e u ns and economic geog aphy. Jou nal o Poli ical Economy, 99, 483–499.
K ugman, P., & Venables, A. (1990). In eg a ion and he compe i i eness o pe iphe al indus y. In C. Bliss
& J. B aga De Macedo (Eds.), Uni y wi h di e si y in he Eu opean economy: The communi y’s Sou h-
e n on ie . Camb idge: Uni e si y P ess.
Lechne , M. (2011). The es ima ion o causal e ec s by di e ence-in-di e ence me hods. Founda ions and
T ends in Econome ics, 4(3), 165–224.
Lessmann, C., & Seidel, A. (2017). Regional inequali y, con e gence, and i s de e minan s—A iew om
ou e space. Eu opean Economic Re iew, 92, 110–132.
Ma au , S., De nis, H., Webb, C., Spiezia, V., & Guellec, D. (2008). The OECD REGPAT da abase: A p es-
en a ion. OECD science, echnology and indus y WP No. 2008/02, OECD, Pa is, h ps:// doi. o g/ 10.
1787/ 24143 71441 44.
1475
The complex egional e ec s o mac o‑ins i u ional change:…
Ma in, D. (2011). The opening up o eas e n eu ope a 20: Jobs, skills and e e se maquilado as in aus ia
and ge many. In M. Jo ano ic (Ed.), In e na ional handbook on he economics o in eg a ion (Vol. 2,
pp. 296–323). Chel enham: Edga Elga .
McCallum, J. (1995). Na ional bo de s ma e : Canada–US egional ade pa e ns. Ame ican Economic
Re iew, 85, 615–623.
Mellande , C., Lobo, J., S ola ick, K., & Ma heson, Z. (2015). Nigh - ime ligh da a: A good p oxy measu e
o economic ac i i y? PLoS ONE, 10(10), e0139779.
Monas i io is, V., Kallio as, D., & Pe akos, G. (2017). The egional impac o Eu opean union associa ion
ag eemen s: An e en -analysis app oach o he case o Cen al and Eas e n Eu ope. Regional S udies,
51(10), 1454–1468.
Mon o , P., & Nicolini, R. (2000). Regional con e gence and in e na ional in eg a ion. Jou nal o U ban
Economics, 48, 286–306.
Niebuh , A. (2008). The impac o EU enla gemen on Eu opean bo de egions. In e na ional Jou nal o
Public Policy, 3(3–4), 163–186.
Niebuh , A., & S ille , S. (2004). In eg a ion e ec s in bo de egions: A su ey o economic heo y and
empi ical s udies. Re iew o Regional Resea ch, 24, 3–21.
O e man, H., & Win e s, A. (2006). T ade shocks and indus ial loca ion: The impac o EEC accession on
he UK. CEP discussion pape no. 588.
Pe coco, M. (2015). Highways, local economic s uc u e and u ban de elopmen . Jou nal o Economic
Geog aphy, 16(5), 1035–1054.
Pe akos, G., & Topaloglou, L. (2008). Economic geog aphy and Eu opean in eg a ion: The e ec s on he
EU’s ex e nal bo de egions. In e na ional Jou nal o Public Policy, 3, 146–162.
Pinko skiy, M. (2017). G ow h discon inui ies a bo de s. Jou nal o Economic G ow h, 22(2), 145–192.
Rauch, J. E. (1991). Compa a i e ad an age, geog aphic ad an age, and he olume o ade. Economic Jou -
nal, 101, 1230–1244.
Redding, S., & S u m, D. (2008). The cos s o emo eness: e idence om ge man di ision and euni ica ion.
Ame ican Economic Re iew, 98(5), 1766–1797.
Rubin, D. (1977). Assignmen o ea men g oup on he basis o a co a ia e. Jou nal o Educa ional S a is-
ics, 2, 1–26.
Schä le , J., Hech , V., & Mo i z, M. (2017). Regional de e minan s o Ge man FDI in he Czech epublic:
New e idence on he ole o bo de egions. Regional S udies, 51(9), 1399–1411.
Schmidheiny, K., & Siegloch, S. (2019). On e en s udy designs and dis ibu ed-lag models: equi alence,
gene aliza ion and p ac ical implica ions. IZA DP No. 12079, IZA Ins i u e o Labo Economics.
Se wicka, I. E., Jones, J., & W en, C. (2022). Economic in eg a ion and FDI loca ion: Is he e a bo de e ec
wi hin he enla ged EU? Annals o Regional Science. h ps:// doi. o g/ 10. 1007/ s00168- 022- 01190-2
Sun, L., & Ab aham, S. (2021). Es ima ing dynamic ea men e ec s in e en s udies wi h he e ogeneous
ea men e ec s. Jou nal o Econome ics, 225(2), 175–199.
Publishe ’s No e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps and
ins i u ional a ilia ions.