Elsayed, Nou han A.; Hilbig, Hanno; Riaz, Sascha; Zibla , Daniel
A icle — Published Ve sion
The Gende ed Pe sis ence o Au ho i a ian Indoc ina ion
B i ish Jou nal o Poli ical Science
P o ided in Coope a ion wi h:
WZB Be lin Social Science Cen e
Sugges ed Ci a ion: Elsayed, Nou han A.; Hilbig, Hanno; Riaz, Sascha; Zibla , Daniel (2025) : The
Gende ed Pe sis ence o Au ho i a ian Indoc ina ion, B i ish Jou nal o Poli ical Science, ISSN
1469-2112, Camb idge Uni e si y P ess (CUP), Camb idge, Vol. 55, pp. 1-11,
h ps://doi.o g/10.1017/S0007123425100720
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/329650
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/
LETTER
The Gende ed Pe sis ence o Au ho i a ian
Indoc ina ion
Nou han A. Elsayed1, Hanno Hilbig2, Sascha Riaz3and Daniel Zibla 4
,
5
1Independen Resea che , Be lin, Ge many, 2Depa men o Poli ical Science, Uni e si y o Cali o nia, Da is, USA,
3Depa men o Poli ical and Social Science, Eu opean Uni e si y Ins i u e, Fiesole, I aly, 4Depa men o Go e nmen ,
Ha a d Uni e si y, USA and 5Di ec o , WZB Be lin Social Science Cen e , Ge many
Co esponding au ho : Hanno Hilbig; Email: [email p o ec ed]
(Recei ed 15 Oc obe 2024; e ised 23 June 2025; accep ed 30 June 2025)
Abs ac
A la ge li e a u e has s udied he e ec s o socializa ion unde au ho i a ianism on poli ical a i udes. In
his esea ch no e, we ex end his li e a u e by demons a ing s iking gende dispa i ies in he pos -
ansi ion pe sis ence o hese e ec s. We s udy he case o au ho i a ian indoc ina ion in he Ge man
Democ a ic Republic (GDR) using a eg ession kink design o causal iden i ica ion. Fi s , we d aw on a
unique su ey ielded igh be o e euni ica ion o show ha educa ion unde au ho i a ianism
subs an ially educed suppo o democ a ic capi alism and euni ica ion wi h he Wes . In he second
s ep, we iangula e mul iple con empo a y da a sou ces o ace he pe sis ence o hese e ec s o e ime.
Mo e han wo decades a e he all o he GDR, he a i udinal e ec s o au ho i a ian socializa ion pe sis
only among men, bu no women. Ou esul s highligh conside able he e ogenei y in he pe sis ence o
au ho i a ian legacies, aising c i ical ques ions abou pos -au ho i a ian ‘ e-socializa ion’and gende ed
adap abili y.
Keywo ds: indoc ina ion; au oc a ic educa ion; GDR; poli ical socializa ion; eg ession kink design
In oduc ion
Does socializa ion in an au ho i a ian egime ha e pe sis en e ec s on poli ical a i udes and
beha io ? Ea lie schola ship on he legacies o au ho i a ianism analyzed mac o-le el ou comes,
such as he s uc u e o he pos -au ho i a ian egime (Jowi 1992; Beissinge and Ko kin 2014;
Simpse e al. 2018). A second gene a ion o wo k has ocused on he mic o-le el a i udinal and
beha io al legacies o au ho i a ian egimes (Alesina and Fuchs-Schündeln 2007; Be nha d and
Ka akoç 2007; Neundo 2010; Pop-Eleches and Tucke 2014; Pop-Eleches and Tucke 2017;
Che u u, 2023; Pop-Eleches and Tucke 2020; see also Neundo and Pop-Eleches 2020).
The la e li e a u e has made he impo an con ibu ions o no only showing ha he poli ical
a i udes and p e e ences ha eme ged ou o an undemoc a ic e a endu e beyond egime change,
bu ha he pe sis ence o hese a i udes a ies wi h how di e en subg oups expe ienced
au ho i a ian socializa ion e o s. While au ho i a ian egimes y o e ec i ely ‘indoc ina e’
hei popula ions, he e is a ia ion in he in ensi y a which di e en g oups a e exposed o he
egime’s e o s, as well as he deg ee o which hei in e es s a e aligned wi h he egime
(Neundo and Pop-Eleches 2020). In addi ion, indi idual cha ac e is ics o ci izens, such as
educa ion, eligion, and gende (see, o example, Pop-Eleches and Tucke 2014; Paglayan 2021;
© The Au ho (s), 2025. Published by Camb idge Uni e si y P ess. This is an Open Access a icle, dis ibu ed unde he e ms o he C ea i e
Commons A ibu ion licence (h ps://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed e-use, dis ibu ion and
ep oduc ion, p o ided he o iginal a icle is p ope ly ci ed.
B i ish Jou nal o Poli ical Science (2025), 55, e124, 1–11
doi:10.1017/S0007123425100720
h ps://doi.o g/10.1017/S0007123425100720 Published online by Camb idge Uni e si y P ess
Pop-Eleches and Tucke 2020), can make some mo e suscep ible o indoc ina ion han o he s. In
pa icula , Pop-Eleches and Tucke (2020) ind ha women we e ma ginally less a ec ed by
au oc a ic indoc ina ion e o s and es a numbe o po en ial explana ions ha ela e o gende
di e ences in li e expe iences unde au ho i a ianism ( o example, highe chu ch a endance).
Taken oge he , o da e, he exis ing li e a u e has demons a ed ha he he e ogenei y o
socializa ion expe iences unde au ho i a ianism d i es pos - ansi ion he e ogenei y o a i udes.
In his a icle, we join his li e a u e (especially Pop-Eleches and Tucke 2017; Neundo and
Pop-Eleches 2020) o highligh wo poin s. Fi s , d awing on e idence om he case o he
o me ly communis Eas Ge many (Ge man Democ a ic Republic [GDR]), we show ha Eas
Ge man poli ical a i udes owa ds capi alism and democ acy e lec his o ical pa e ns o
socializa ion om he communis e a. These esul s mi o he indings by Che u u (2023), who,
using a simila empi ical s a egy, inds ha educa ion unde democ acy inc eased public suppo
o poli ical ins i u ions in Eas Ge many. Rela edly, we ind e idence in suppo o he
‘imp essionable yea s’o ‘ o ma i e yea s’hypo hesis, which s a es ha a i udinal pa e ns
es ablished ea ly in li e end o endu e in o adul hood (see, o example, Neundo e al. 2013;
B um 2018; Neundo and Pop-Eleches 2020).
We come, howe e , o a second inding ha depa s om many exis ing wo ks by showing ha
he a i udinal pa e ns isible a he end o he au ho i a ian pe iod ha e pe sis ed in a s a kly
gende ed way. In pa icula , a i udinal pa e ns es ablished among men du ing he au ho i a ian
pe iod pe sis ed un il decades la e , whe eas o women, hey did no . This sugges s an in iguing
conclusion. Schola s ha e suspec ed ha indi idual-le el pos -au ho i a ian he e ogenei y (based,
o example, on gende o eligion) e lec s di e ing expe iences du ing au ho i a ianism. Ou
e idence shows ha bo h men and women exi ed communism wi h simila a i udinal pa e ns
bu di e ged only a e au ho i a ianism, which sugges s ha while au ho i a ian socializa ion
a ec ed bo h men and women, he p ocess o pos -au ho i a ian ‘compe ing socializa ion’
(Neundo and Pop-Eleches 2020) has been mo e signi ican o women han men.
Ou empi ical analysis iangula es mul iple su ey da ase s o ace he legacy o au ho i a ian
indoc ina ion, as p oxied by educa ion, in he o me GDR. Fi s , we d aw on a unique su ey o
Eas Ge mans collec ed igh be o e and igh a e he dissolu ion o he GDR. Using hese da a,
we obse e a i udes owa ds capi alism, socialism, and euni ica ion wi h Wes Ge many in 1989/
1990. To ace he pe sis ence o hese a i udes o e ime, we d aw on con empo a y su ey da a
on a i udes owa ds capi alism, iews o he GDR egime, and pa y p e e ences. We analyze hese
da a using a eg ession kink design (RKD) –a causal iden i ica ion s a egy ha has a ely been
used in poli ical science bu has ecen ly gained ac ion in labo economics (Ca d e al. 2015;
Ganong and Jäge 2018).
We ind consis en e idence ha au oc a ic educa ion had a p onounced nega i e e ec on
suppo o he democ a ic capi alis o de igh be o e he collapse o he GDR. GDR schooling
educed suppo o capi alism in he GDR in 1989/1990, and likewise educed suppo o
euni ica ion wi h he Wes Ge man Fede al Republic (GFR). These ini ial e ec s a e independen
o gende , ha is, we obse e simila esul s o men and women. In he second s ep, we examine
he pe sis ence o hese e ec s o e ime using con empo a y su ey da a. He e, we ind
p onounced gende di e ences: mo e han wo decades a e he collapse o he GDR, we only ind
e idence o he pe sis ence o au ho i a ian socializa ion among men, bu no women. Men
exposed o mo e GDR schooling iew he GDR in a mo e posi i e ligh , a e mo e c i ical owa ds
he ma ke economy, and show highe le els o suppo o he illibe al, an i-es ablishmen pa y
Al e na i e ü Deu schland (A D). No ably, he e ec s o educa ion unde au oc acy do no
pe sis among women. To add ess conce ns abou o he con ounding ac o s ha a e un ela ed o
au oc a ic indoc ina ion ( o example, age- o coho -speci ic e ec s), we demons a e null
esul s o Wes Ge man men and women ac oss a se ies o placebo es s. In sum, we ind ha he
2 Nou han A. Elsayed e al.
h ps://doi.o g/10.1017/S0007123425100720 Published online by Camb idge Uni e si y P ess
e ec s o educa ion unde au oc acy exhibi p onounced empo al a ia ion, and his a ia ion is
gende ed in na u e.
While a comp ehensi e es o he mechanisms unde lying hese esul s is beyond he scope o
his a icle, we p o ide sugges i e e idence o gende ed di e ences in adap ion o he new social
o de –e idence ha poin s he way owa ds u he sys ema ic esea ch on his opic. We
conside a p ocess whe eby women’s iews di e ge om men’s in a pos -au ho i a ian con ex
h ough he ‘compe ing socializa ion’ amewo k (Neundo and Pop-Eleches 2020). In his
amewo k, Neundo and Pop-Eleches (2020) a gue ha he long- e m e ec s o au oc acies on
he a i udes and beha io s o hei subjec s depend on he na u e and s eng h o he a ious
exis ing socializing o ces, which ei he ein o ce o compe e wi h one ano he in socializing
indi iduals. In his amewo k, au oc a ic ins i u ions can ha e di e en ial impac s on di e en
demog aphic g oups. We de elop his insigh u he o con end ha in euni ied democ a ic
capi alis Ge many, women we e ei he mo e exposed o compe ing mechanisms o an i-
au ho i a ian socializa ion han men, o adap ed mo e success ully o changes in poli ical
condi ions and he coun e -socializa ion o he new sys em. Empi ically, we ind ha a e
euni ica ion, Eas Ge man women inc easingly ou pe o med men in he democ a ic educa ion
sys em, sugges ing highe adap abili y o new social ins i u ions among women.
Ou a icle con ibu es o he li e a u e on au ho i a ian legacies in se e al dis inc ways. Fi s ,
we in oduce a new iden i ica ion s a egy, he eg ession kink design, o s udy how li ing unde
communism le an endu ing imp in on o e s’ iews o poli ics and economics (Wi enbe g
2006; Pop-Eleches and Tucke 2017; Neundo and Pop-Eleches 2020). Second, we pinpoin one
impo an sou ce o his socializa ion, namely schooling, and, like Dinas and No hmo e-Ball
(2020), show e idence ha schools a e a key si e o indoc ina ion in au ho i a ian egimes. Thi d,
like Pop-Eleches and Tucke (2017) and Pop-Eleches and Tucke (2020), we ind ema kable
pe sis ence o hese e ec s. Bu he e we depa company wi h exis ing accoun s by poin ing o a
s a k gende di ide in pe sis ence ha he exis ing li e a u e has no o da e ecognized. We
conclude by explo ing some causal mechanisms ha migh explain why women migh be mo e
a ailable o a pos -au ho i a ian ‘ e-socializa ion’, p oposing his as a p omising esea ch agenda
mo ing o wa d.
Da a and Empi ical S a egy
Da a
In line wi h p io esea ch on au ho i a ian legacies (Neundo and Pop-Eleches 2020), we seek o
measu e he e ec s o au ho i a ian e o s a indoc ina ion on suppo o wha we e m
democ a ic capi alis o de . To do his, we iangula e mul iple da a sou ces and selec ed ou comes
ha p oxy o a i udes owa ds bo h libe al democ acy and capi alism, bo h igh be o e he
dissolu ion o he GDR and oday. We elabo a e on ou da a sou ces and he a iables we
use below.
A i udes owa ds euni ica ion and capi alism be o e euni ica ion
We measu e a i udes owa ds democ a ic capi alism igh be o e he all o he GDR using a
su ey o abou 5,000 Eas Ge mans in 1989/1990. The su ey was conduc ed by he
Zen alins i u ü Jugend o schung (ZIJ; Cen al Ins i u e o You h Resea ch). The aim o he
su ey was o measu e a i udes owa ds he ‘ enewal’ aking place in he GDR and i s u u e
poli ical and economic de elopmen . Speci ically, we obse e (1) suppo o capi alism, (2)
suppo o ‘ e o med’socialism, and (3) opposi ion o he uni ica ion o he GDR and Wes
Ge many. We p o ide mo e de ails on his da a sou ce and he su ey i ems we use in sec ion
A.2.1 in he Supplemen a y In o ma ion (SI).
B i ish Jou nal o Poli ical Science 3
h ps://doi.o g/10.1017/S0007123425100720 Published online by Camb idge Uni e si y P ess
Con empo a y a i udes owa ds he GDR and capi alism
We d aw on wo da a sou ces. Fi s , we use 2017 su ey da a om he su ey company Ci ey o
measu e a i udes owa ds he GDR. Speci ically, we obse e whe he esponden s iew he GDR
as a dic a o ship –a p oxy o a i udes owa ds he o me au oc a ic sys em. We p o ide
addi ional de ails on his da a sou ce in SI sec ion A.2.2. In addi ion, we use da a om he 2016
Poli ba ome e su ey (Fo schungsg uppe Wahlen 2017). We obse e a bina y indica o o
whe he esponden s a e sa is ied wi h ‘social capi alism’(soziale Ma k wi scha ) in Ge many.
Suppo o he illibe al A D pa y
We le e age indi idual-le el da a on he o e in en ions o mo e han 614,000 eligible o e s
su eyed be ween 2017 and 2019. The da a we e o iginally collec ed by Fo sa, a su ey company
ha conduc s daily elephone in e iews on a c oss-sec ion o abou 500 eligible o e s. Using
hese da a, we code a bina y indica o ha equals one o esponden s who in end o o e o he
illibe al Al e na i e ü Deu schland. A D suppo p oxies o an i-sys em, an i-democ a ic
a i udes.1We p o ide addi ional de ails on hese da a in SI sec ion A.2.3.
Empi ical S a egy
To es ima e he causal e ec o educa ion unde au oc acy on poli ical p e e ences and a i udes,
we use a eg ession kink design. This design le e ages slope changes in he ela ionship be ween
he unning a iable –yea o bi h in ou se ing –and he ea men o iden i y causal e ec s
(Ca d e al. 2015). In con as o he egula eg ession discon inui y design, which es s o a
discon inui y in le els a he cu -o , RKD es s o changes in he de i a i e o he ou come
a iable wi h espec o he unning a iable be o e and a e he cu -o (kink) poin . Below, we
p o ide in ui ion o causal iden i ica ion using he RKD design. We p o ide a mo e o mal
discussion o iden i ica ion in he RKD se up in sec ion A.3 o he SI.
The key insigh behind he RKD app oach in ou se ing is ha he a e age numbe o yea s
indi iduals spen in GDR schools is a de e minis ic and con inuous unc ion o hei bi h yea .
We illus a e his ela ionship be ween he unning a iable and he ea men assignmen in
Figu e A.1 in he SI. The e a e wo key akeaways om his g aph. Fi s , he ela ionship be ween
bi h yea and GDR educa ion is cons an up un il 1971. Rega dless o hei p ecise yea o bi h
be ween 1945 and 1971, indi iduals spen a o al o abou wel e yea s in GDR schools (see sec ion
A.1).2This changes a ound 1971. When he Be lin Wall ell in 1989, s uden s bo n in 1971 we e 18
yea s old and had jus comple ed uppe seconda y educa ion. F om 1971 onwa ds, howe e , each
la e bi h yea co esponds o educed exposu e o GDR educa ion. The a e age numbe o yea s
spen in GDR schools dec eases linea ly be ween he bi h yea s 1971 and 1984. The de i a i e o
GDR educa ion wi h espec o bi h yea changes om ze o o nega i e one a ound he kink poin
in 1971.
The eg ession kink design le e ages his kink in he ela ionship be ween bi h yea and GDR
educa ion o causal iden i ica ion and cap u es he ollowing in ui ion: while he ela ionship
be ween yea o bi h (coho ) and poli ical p e e ences cap u es a a ie y o mechanisms p io o
and a e 1971, he only one o hese explana o y ac o s ha changes a ound he kink poin in
1971 is au ho i a ian educa ion in he GDR.3Acco dingly, he RKD allows us o sepa a e he
1A 2022 Ge man cou uling a i med ha he A D can be moni o ed by s a e in elligence se ices due o conce ns ha i
may unde mine he libe al-democ a ic o de (Blanke e al. 2022).
2In he GDR, schooling las ed a maximum o wel e yea s. Mos s uden s eached his by combining en yea s o seconda y
educa ion wi h wo yea s o oca ional school (Be iebsbe u sschule) o by comple ing g ades ele en and wel e a an uppe
seconda y school (E wei e e Obe schule).
3To u he subs an ia e his assump ion, we conduc a placebo popula ion es using Wes -Ge man esponden s (see
Table 3).
4 Nou han A. Elsayed e al.
h ps://doi.o g/10.1017/S0007123425100720 Published online by Camb idge Uni e si y P ess
e ec s o GDR educa ion om gene al age o coho e ec s (see sec ion A.3.3). Empi ically, he
causal e ec is es ima ed by (1) es ima ing he de i a i e o he ou come a iable wi h espec o
he unning a iable (yea o bi h) sepa a ely o he le and igh o he kink poin , (2) calcula ing
he di e ence be ween he es ima ed de i a i es le and igh o he kink poin , and (3) scaling
his di e ence by he slope change in he ea men assignmen unc ion a e he kink poin
(1 in ou case). We p o ide de ails on each o hese s eps in sec ion A.3 in he SI. In a nu shell,
he RKD es s whe he a disc e e slope change in he ela ionship be ween he unning a iable
and he ea men coincides wi h a change in he ela ionship be ween he unning a iable and
he ou come.
We no e ha ou es ima ion s a egy depends on he na u e o ou da a sou ce. When we ha e a
su icien numbe o esponden s, we es ima e he RKD using local polynomial eg ession
Ca aneo e al. (2019). This is he case o he Ci ey da a on con empo a y a i udes owa ds he
GDR and he Fo sa da a on o e in en ions. Fo he Poli ba ome e and ZIJ 1989/1990 su ey
da a, we es ima e he slope change a ound he 1971 kink poin using a pa ame ic o dina y leas
squa es (OLS) speci ica ion. We do his because (i) hese wo da ase s only con ain ela i ely ew
obse a ions a ound he cu -o poin , and (ii) we only obse e bins o he yea o bi h unning
a iable o he Poli ba ome e da a. We no e ha his app oach equi es s onge unc ional o m
assump ions and uses a ixed, manually selec ed bandwid h. The in e en ial esul s ob ained om
hese speci ica ions hus con ain mo e unce ain y and should be in e p e ed wi h g ea e cau ion.
We p o ide addi ional de ails on his app oach in sec ion A.3.2 o he SI.
Be o e mo ing on o discuss ou esul s, we no e wo limi a ions o ou empi ical s a egy. Fi s ,
we es ima e local a e age ea men e ec s –simila o he egula eg ession discon inui y se up.
Mo e speci ically, we es ima e he e ec o an addi ional yea o GDR educa ion o indi iduals
bo n nea he kink poin (1971). Consequen ly, ou indings may no gene alize o coho s a
om he cu -o ( o example, hose bo n in he 1940s). In addi ion, we cau ion agains
ex apola ing ou es ima es a away om he cu -o poin , as exposu e o Ma xis heo y and
s a e ideology was mo e in ense in he la e school yea s compa ed o lowe g ades. Fo example,
he S aa sbü ge kunde subjec –a key pilla o socialis indoc ina ion in GDR schools –was
augh only in he seconda y educa ional ack (see sec ion A.1). As a esul , ou es ima es a e
likely highe han he a e age ea men e ec ac oss he en i e wel e-yea school jou ney.
Second, we only obse e esponden s’yea o bi h a he han exac bi hdays. This lack o
g anula i y means ha ou unning a iable akes on only a mode a e numbe o unique alues.
This may bias ou es ima es i he local linea polynomial we i on bo h sides o he cu -o is a bad
i o he ue unde lying condi ional expec a ion unc ion (Lee and Ca d 2008). The la ge-scale
su ey da ase s we use o measu e con empo a y a i udes allow us o choose a ela i ely small
(op imal) bandwid h, which should limi his po en ial sou ce bias o hese analyses (Kolesá and
Ro he 2018). As no ed p e iously, he es ima es om he pa ame ic RKD speci ica ions using a
la ge bandwid h should be in e p e ed wi h mo e cau ion.
Resul s
Below, we i s p esen he es ima ed e ec o educa ion unde au oc acy on poli ical p e e ences
measu ed jus be o e he dissolu ion o he GDR, and hen assess he pe sis ence o hese e ec s
among he same coho s mo e han wo decades la e .
E ec s o Educa ion Measu ed Righ be o e Reuni ica ion
We begin by assessing how educa ion unde au oc acy a ec ed suppo o he democ a ic
capi alis o de jus be o e he collapse o he GDR, using he ZIJ da a desc ibed in he sec ion
Da a. We p esen he esul s in Table 1. We dicho omize ou comes such ha a gi en ou come
akes he alue one i a esponden (i) suppo s capi alism, (ii) opposes euni ica ion, o
B i ish Jou nal o Poli ical Science 5
h ps://doi.o g/10.1017/S0007123425100720 Published online by Camb idge Uni e si y P ess
(iii) suppo s e o med socialism ( o mo e in o ma ion on hese su ey i ems, see sec ion A.2.1).
We mul iply poin es ima es by 100, such ha he es ima es in Table 1can be in e p e ed as
pe cen age poin changes in suppo o opposi ion. Ac oss six speci ica ions, we ind ha
educa ion unde au oc acy educes suppo o he democ a ic capi alis o de , bo h o men and
o women. The e ec size magni ude is sizable, anging om −4.08 pe cen age poin s o 6.66
pe cen age poin s.4
Like o he pape s in his li e a u e, we u ilize coho -speci ic a ia ion in exposu e o educa ion
unde au oc acy. Typically, such analyses ely on esponden s su eyed yea s o decades a e
democ a ic ansi ions and a e he e o e no able o dis inguish whe he he e ec s a e d i en by
mo e educa ion in an au ho i a ian sys em o by less educa ion in he democ a ic successo s a e.
No ably, ou analyses shown in Table 1can add ess his issue since hey a e de i ed om su ey
esponses ha we e collec ed in la e 1989 and 1990, be o e he GDR egime collapsed. Fo hese
esponden s, less schooling in he GDR does no imply mo e schooling in democ a ic Ge many.
The analysis compa es eigh een-yea -old Eas Ge mans o younge Eas Ge mans who ecei ed
less exposu e o au ho i a ian educa ion bu had no ye ecei ed educa ion ollowing he Wes
Ge man cu iculum. As shown in Table 1, we ind ha his compa ison al eady p o ides e idence
o an indoc ina ing e ec o educa ion in he GDR. Ou analyses, he e o e, allow us o sepa a e
he e ec s o educa ion in he wo sys ems. We conclude ha mo e educa ion in au oc a ic sys ems
can indoc ina e ci izens, e en when i is no accompanied by less educa ion in democ a ic
successo s a es.5
Con empo a y A i udes and Vo e In en ions
In he p e ious sec ion, we obse ed ha educa ion unde au oc acy had p onounced nega i e
e ec s on suppo o he democ a ic capi alis o de jus be o e he demise o he GDR. We now
examine whe he hese pa e ns pe sis a e Ge man euni ica ion. In Table 2, we p esen esul s
based on o e in en ions, a i udes owa ds he GDR, and sa is ac ion wi h he capi alis ma ke
economy. We ind ha he nega i e ela ionship be ween educa ion unde au oc acy and suppo
o he democ a ic capi alis o de pe sis s among men, bu no among women. As be o e, all
ou comes a e bina y, and we mul iply poin es ima es by 100, which means ha he es ima es can
be in e p e ed as pe cen age poin changes.
Table 1. E ec s o au ho i a ian educa ion on a i udes measu ed p io o euni ica ion
Ou come Sample ˆ
τSRK 95% CI n
Suppo capi alism Men −4.08*** [−5.78, –2.38] 3,267
Suppo capi alism Women −2.57*** [−3.91, –1.23] 3,184
Oppose euni ica ion Men 6.12*** [4.26, 7.98] 4,218
Oppose euni ica ion Women 6.66*** [4.80, 8.52] 4,381
Suppo e o med socialism Men 5.16*** [3.10, 7.22] 3,267
Suppo e o med socialism Women 4.80*** [3.02, 6.58] 3,184
No es: Resul s om pa ame ic RKD speci ica ions es ima ed using OLS (see sec ion A.3.2 o mo e de ails). The sample includes esponden s
bo n be ween 1960 and 1975. We p o ide addi ional de ails on he su ey da a in sec ion A.2.1 in he SI. Signi icance codes: ***: 0.001, **: 0.01,
*: 0.05.
4Figu e A.2 p esen s condi ional ou come means. Figu e A.3 shows es ima es om an addi ional speci ica ion compa ing
pos -1970 coho -speci ic ou come a e ages o he mean o he h ee ZIJ ou comes o coho s bo n in 1970 o ea lie .
Consis en wi h he esul s in Table 1, we ind ha , compa ed o ully exposed coho s, hose only pa ially exposed o he GDR
educa ion sys em a e less suppo i e o e o med socialism, mo e suppo i e o capi alism, and less likely o oppose
euni ica ion. This pa e n is mos p onounced o he 1974 and 1975 coho s and does no a y by esponden gende .
5We no e ha ou e idence does no imply ha educa ion a e he ansi ion has no e ec on a i udes (see also ou
discussion o di e ences in he pos - ansi ion ajec o y o men and women in sec ion mechanisms). Ra he , we a gue ha a
leas some o he o al a ia ion can be explained by p e- ansi ion educa ion.
6 Nou han A. Elsayed e al.
h ps://doi.o g/10.1017/S0007123425100720 Published online by Camb idge Uni e si y P ess
Fo men, we ind ha educa ion unde au oc acy inc eases suppo o he adical- igh A D
and lowe s sa is ac ion wi h he ma ke economy. The magni ude o he nega i e poin es ima e
o sa is ac ion wi h he ma ke economy is compa able o he poin es ima e o suppo o
capi alism igh be o e euni ica ion in Table 1, sugges ing ha he an i-capi alis e ec o
au oc a ic educa ion pe sis s o e ime. We u he obse e pe sis en p o-GDR e ec s o men –
educa ion unde au oc acy ende s men signi ican ly less likely o espouse he iew ha he GDR
was a dic a o ship. In con as , poin es ima es o women a e consis en ly s a is ically
insigni ican , be ween i e and en imes smalle , and much close o 0.6We u he es whe he
he di e ence be ween he es ima ed coe icien s o men and women is s a is ically signi ican .
P- alues o he di e ence in coe icien s a e: (i) 0.037 o he A D suppo ou comes, (ii) 0.074 o
he GDR dic a o ship ou come, and (iii) 0.039 o he ma ke economy sa is ac ion ou come. The
o me wo p- alues a e calcula ed assuming coe icien s a e independen , ha is, he a iance o
he di e ence is assumed o be he sum o he a iances o he coe icien es ima es. This is because
he local polynomial app oach (Ca aneo e al. 2019) does no allow us o es ima e he co a iance.
The hi d p- alue is ob ained om a pa ame ic in e ac ion speci ica ion ha mi o s he main
speci ica ion o his ou come a iable (see sec ion A.3.2). We p o ide addi ional de ails on how
p- alues a e calcula ed in SI sec ion A.3.4.
Summa izing he esul s, we ind s ong e idence ha au oc a ic educa ion ‘wo ked’in he
sense ha i had a p onounced nega i e e ec on suppo o he democ a ic capi alis o de igh
be o e he collapse o he GDR. The ini ial e ec s o indoc ina ion a e independen o gende –we
obse e simila e ec s o bo h men and women. Using con empo a y da a, we hen show ha his
e ec pe sis s among men bu no among women. Fo men, educa ion unde au oc acy con inues
o educe suppo o he ma ke economy, induces mo e posi i e iews o he GDR, and inc eases
suppo o he illibe al A D pa y. Fo women, we obse e ha he ini ial e ec s o indoc ina ion
in he GDR educa ion sys em do no pe sis o e ime.
Robus ness Tes s
Fi s , we examine whe he he esul s o Table 2could be d i en by composi ional di e ences due
o selec i e ou -mig a ion a e he all o he GDR. In sec ion A.4.5 o he SI, we ely on addi ional
panel da a o show ha women who we e bo n in he GDR do no become mo e o less likely o
mo e o Wes Ge many i hey we e bo n jus a e he 1971 coho cu -o . This sugges s ha he
Table 2. Con empo a y e ec s o educa ion unde au oc acy
Ou come Sample ˆ
τSRK 95% CI hMSE n
Vo ing
A D suppo Men 2.32 [0.39, 4.26] 7 6,136
A D suppo Women −0.22 [−1.63, 1.18] 7 6,184
A i udes
GDR was a dic a o ship Men −6.96** [−11.14, –2.78] 4 3,066
GDR was a dic a o ship Women −1.18 [−5.96, 3.6] 6 1,818
Sa is ac ion: ma ke economy Men −3.54 [−6.77, –0.31] 4,691
Sa is ac ion: ma ke economy Women 0.73 [−2.8, 4.25] 4,281
No es: We p esen poin es ima es om he sha p eg ession kink design (see sec ion ‘Da a and Empi ical S a egy’). Each poin es ima e
ep esen s he e ec o one addi ional yea o educa ion unde au oc acy o Eas Ge mans. Poin es ima es can be in e p e ed as pe cen age
poin changes. We es ima e sepa a e bandwid hs o each ou come and sample, as ecommended by Ca aneo e al. (2019). Rows 1–2 a e
es ima es based on he Fo sa da a, while ows 3–4 p esen es ima es based on Ci ey da a. Rows 5–6 use Poli ba ome e da a. The
Poli ba ome e esul s s em om a pa ame ic RKD speci ica ion (see sec ion A.3.2). ***: 0.001, **: 0.01, *: 0.05.
6Among nei he men no women do we ind inc eased suppo o Die Linke ( he Le Pa y), he successo o he socialis
uling pa y o he GDR (see SI Table A.3).
B i ish Jou nal o Poli ical Science 7
h ps://doi.o g/10.1017/S0007123425100720 Published online by Camb idge Uni e si y P ess
obse ed lack o pe sis ence among women is no d i en by selec i e ou -mig a ion a ound he
1971 kink poin .
Second, we conduc a ‘placebo popula ion es ’using Wes Ge man esponden s ollowing
Egge s e al. (2024). We do his o es whe he ou esul s could be d i en by o he unobse ed
ac o s ( o example, policies o coho e ec s) ha exhibi a kink a ound he 1971 bi h yea cu -
o . To he ex en ha such unobse ed ac o s Ua e common o all Ge mans, we would expec
signi ican ea men e ec es ima es ˆ
τSRK in he placebo popula ion (Wes Ge mans).
Reassu ingly, we do no ind s a is ically signi ican e ec s among Wes Ge man men o women
(Table 3), who we e no subjec o au ho i a ian indoc ina ion in he GDR. We in e p e his as
e idence agains he idea ha ou indings a e d i en by o he , unobse ed ac o s ha a y
a ound he 1971 cu -o .
Finally, we es o po en ial discon inui ies in co a ia e le els a ound he 1971 coho cu -o
using a s anda d eg ession discon inui y design. We use he con empo a y Fo sa da a and
es ima e sepa a e models o men and women. While we ind no e idence o such discon inui ies,
we no e ha almos all co a ia es a ailable o us (excep gende ) a e measu ed pos - ea men and
could be a ec ed by exposu e o GDR educa ion. The e o e, he esul s o his es should be
ega ded wi h cau ion.
Mechanisms
A key conclusion o ou analysis is ha gende ed a ia ion in he e ec s o indoc ina ion e o s
likely s ems om expe iences a e he democ a ic ansi ion. Gi en he esea ch no e o ma , an
exhaus i e and sys ema ic discussion o po en ial mechanisms ha explain his gende ed
he e ogenei y is beyond he scope o his a icle. Poin ing he way o u u e esea ch, we b ie ly
discuss a numbe o po en ial mechanisms below. We p o ide addi ional de ails and
co esponding e idence in sec ion A.6 in he SI.
Fi s , i migh be he case ha he he e ogenei y we obse e be ween men and women is d i en
by di e en economic ajec o ies pos - euni ica ion. Howe e , we do no ind e ec s o GDR
educa ion on ei he employmen s a us, income, o wel a e dependence among ei he gende in
he 1990s, 2000s, o 2010s. Second, we a gue ha compulso y mili a y se ice canno accoun o
ou esul s, as we do no obse e gende di e ences igh be o e euni ica ion in 1989/1990.
Thi d, Eas Ge man women may ha e adap ed mo e quickly o he new poli ical and social
o de a e euni ica ion. They migh ha e been less a ached o he au ho i a ian alues o he
GDR, making hem mo e open o he democ a ic capi alis sys em – ha is, mo e ecep i e o
capi alis -democ a ic coun e -socializa ion a e 1990 (see also ou discussion o he ‘compe ing
socializa ion’ amewo k [Neundo and Pop-Eleches 2020] in he In oduc ion). We ind
sugges i e e idence in suppo o his mechanism in he ealm o educa ion. Al hough men and
Table 3. Placebo es s using Wes Ge man esponden s
Ou come Sample ˆ
τSRK 95% CI hMSE n
Vo ing
A D suppo Men 0.04 [−0.36, 0.45] 9 47,977
A D suppo Women 0.18 [−0.31, 0.68] 6 28,963
A i udinal ou comes
GDR was a dic a o ship Men −0.02 [−0.72, 0.67] 8 23,422
GDR was a dic a o ship Women −1.16 [−3.01, 0.69] 6 5,816
Sa is ac ion wi h ma ke economy Men 2.09 [−0.24, 4.42] 7,430
Sa is ac ion wi h ma ke economy Women −0.51 [−3.06, 2.04] 6,606
No es: We p esen poin es ima es om he sha p eg ession kink design (see sec ion ‘Da a and Empi ical S a egy’). Poin es ima es can be
in e p e ed as pe cen age poin changes. We es ima e sepa a e bandwid hs o each ou come and sample (Ca aneo e al. 2019). The
Poli ba ome e esul s s em om a pa ame ic RKD speci ica ion (see sec ion A.3.2). ***: 0.001, **: 0.01, *: 0.05.
8 Nou han A. Elsayed e al.
h ps://doi.o g/10.1017/S0007123425100720 Published online by Camb idge Uni e si y P ess
