Ellge , Fabio; Hilbig, Hanno; Riaz, Sascha; Tillmann, Philipp
A icle — Published Ve sion
Local Newspape Decline and Poli ical Pola iza ion –
E idence om a Mul i-Pa y Se ing
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: Ellge , Fabio; Hilbig, Hanno; Riaz, Sascha; Tillmann, Philipp (2024) : Local
Newspape Decline and Poli ical Pola iza ion – E idence om a Mul i-Pa y Se ing, B i ish Jou nal
o Poli ical Science, ISSN 1469-2112, Camb idge Uni e si y P ess, Camb idge, Vol. 54, Iss. 4, pp.
1256-1275,
h ps://doi.o g/10.1017/S0007123424000243
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/310003
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/
ARTICLE
Local Newspape Decline and Poli ical Pola iza ion –
E idence om a Mul i-Pa y Se ing
Fabio Ellge 1, Hanno Hilbig2, Sascha Riaz3and Philipp Tillmann4
1
T ans o ma ions o Democ acy Uni , WZB Be lin Social Science Cen e , Ge many,
2
Depa men o Poli ical Science,
Uni e si y o Cali o nia, Da is, CA, USA,
3
Depa men o Poli ical and Social Sciences, Eu opean Uni e si y Ins i u e, I aly
and
4
Independen Schola
Co esponding au ho : Hanno Hilbig; Email: [email protected]
(Recei ed 24 Sep embe 2022; e ised 23 No embe 2023; accep ed 4 June 2024; i s published online 23 Oc obe 2024)
Abs ac
How does he decline o adi ional news ou le s a ec poli ical pola iza ion? We p o ide no el e idence
on his ques ion by examining he link be ween local newspape exi s, media consump ion, and elec o al
beha iou in a mul ipa y se ing. Ou empi ical analysis combines a unique panel o all Ge man local
newspape s be ween 1980 and 2009, elec o al e u ns, and an annual media consump ion su ey o
mo e han 670,000 esponden s. Using a di e ence-in-di e ences design, we demons a e ha local news-
pape exi s inc ease elec o al pola iza ion. Addi ional analysis poin s o changes in media consump ion as
he unde lying mechanism d i ing his esul : ollowing local news exi s, consume s subs i u e local news
wi h na ional abloid news. Ou indings ex end p io esul s in he US con ex o a mul ipa y se ing and
shed new ligh on he causal chain unning om changing local news landscapes o elec o al beha iou .
Keywo ds: pola iza ion; local news decline; newspape s; media consump ion
In oduc ion
How does he decline o adi ional news ou le s a ec poli ical pola iza ion? Pola iza ion has
become a co e challenge o es ablished democ acies. In he Uni ed S a es, pola iza ion has
inc eased he ele ance o pa isan iden i ies and a achmen s. In Eu ope, pola iza ion has c ea ed
new lines o poli ical con lic , pa icula ly ega ding clima e change o immig a ion (Hobol ,
Leepe , and Tilley 2021). Pola iza ion has been linked o a numbe o undesi able consequences,
such as legisla i e g idlock (Jones 2001) as well as de-legi imiza ion and animosi y owa ds pol-
i ical opponen s (Mason 2018), which ul ima ely spills o e in o nonpoli ical ealms (Iyenga
e al. 2018; McConnell e al. 2018).
Social scien is s ha e ecognized how changes in he media en i onmen a e esponsible o
inc easing pola iza ion (see o example Da , Hi , and Dunaway 2018;2021; Le endusky
2013a; P io 2007; P io 2013). Recen wo k also highligh s he ole o local news co e age in
shaping elec o al beha iou (see Hayes and Lawless 2017;2021). In his pape , we p o ide new
e idence on he ela ionship be ween poli ical pola iza ion and he decline o local news in a
mul ipa y sys em. We build on esea ch om he Ame ican con ex whe e p io s udies ha e
examined he link be ween elec o al pola iza ion, b oadband access (T ussle 2021) and exposu e
o na ional s local news (Moskowi z 2021). Mos closely ela ed o his s udy, Da , Hi , and
Dunaway (2018;2021) examine he link be ween local news exi s, exposu e o na ional news,
and pola ized o ing in he Uni ed S a es. We s udy he case o Ge many, whe e cen is pa ies
ha e been losing o es o small pa ies, which end o hold mo e ideologically ex eme posi ions.
© The Au ho (s), 2024. 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 p://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 (2024), 54, 1256–1275
doi:10.1017/S0007123424000243
A he same ime, simila o he Uni ed S a es and o he Wes e n democ acies, newspape ci cu-
la ion has plumme ed o e he pas decades (see Fig. 1).
Going beyond p io wo k, we examine he link be ween local news decline, indi idual-le el
media consump ion beha iou , and poli ical pola iza ion in a mul ipa y sys em o e h ee dec-
ades. We combine a no el, manually collec ed panel on he ci cula ion o all Ge man local news-
pape s be ween 1980 and 2009 wi h coun y-le el elec o al esul s in local and na ional elec ions.
We assess he e ec s o local news exi s on elec o al pola iza ion and suppo o small pa ies.
We ope a ionalize elec o al pola iza ion as a measu e o ideological dispe sion o pa y posi ions
weigh ed by hei espec i e o e sha e. The index hus e lec s bo h changes in elec o al suppo
and pa y ideology o e ime and ac oss coun ies (Dal on 2008). To cons uc his measu e, we
loca e pa y posi ions on he le - igh ideological spec um based on he Compa a i e Mani es o
P ojec (Volkens e al. 2020) and use hei coun y-le el o e sha es as weigh s.
Ou empi ical analysis elies on a di e ence-in-di e ences app oach o pa ial ou ime-
in a ian con ounde s. Fo bo h ede al and municipal elec ions, we ind ha exi s ha e a posi i e
e ec on he a e age ideological dispe sion in a coun y. Simila ly, we demons a e ha he o e
Figu e 1. F ac ionaliza ion o he o e in Ge many and newspape ci cula ion o e ime.
No e: The i s panel shows he o e sha e o he wo la ge Ge man pa ies, he cen e- igh Ch is ian Democ a s (CDU/CSU) and he
cen e-le Social Democ a s (SPD) in compa ison o all o he pa ies in ede al elec ions since 1980. The second panel shows he ci -
cula ion o paid daily newspape s pe capi a in he Uni ed S a es and Ge many o e ime. Da a o he US was collec ed by he Pew
Resea ch Cen e . Da a o Ge many was o iginally published by he In o ma ionsgemeinscha zu Fes s ellung de Ve b ei ung on
We be äge n. We ob ained he da a om S a is a (2020).
B i ish Jou nal o Poli ical Science 1257
sha e o small pa ies, which end o hold posi ions u he om he ideological cen e, inc eases
a e local news exi . In a se ies o addi ional checks, and we con i m ha ou come and co a ia e
ends do no di e be ween coun ies ha e en ually expe ience local news exi s and hose ha do
no . We u he show ha adjus ing o po en ial di e ences in elec o al ajec o ies be ween Eas
and Wes Ge many yields s a is ically signi ican ye smalle e ec es ima es, highligh ing he
impo ance o Eas -Wes he e ogenei y. Taken oge he , ou indings demons a e ha he decline
o local news has con ibu ed o elec o al pola iza ion o e he pas decades.
Fu he , we p o ide new e idence on he mechanisms unde lying he agg ega e elec o al
esul s. In es ablished democ acies like Ge many and he Uni ed S a es, local poli ics is o en pe -
cei ed as mo e consensual and less poli icized compa ed o na ional poli ics, wi h a smalle
emphasis on pa isan di e ences (Bucchiane i 2020; Hol kamp 2017). Acco dingly, he decline
o local news can inc ease pola iza ion by exposing o e s o mo e na ional –and less consen-
sual –poli ics. In pa icula , his may occu when consume s subs i u e local news ou le s
wi h na ional news (Da , Hi , and Dunaway 2018;2021; Hayes and Lawless 2015;2021). To
empi ically es his mechanism, we examine how local news exi s a ec media consump ion
a he indi idual le el. To do his, we d aw on a la ge-scale annual media consump ion su ey
o mo e han 670,000 esponden s ha co e s ou en i e s udy pe iod. We demons a e ha
na ional news consump ion inc eases bo h in ela i e and absolu e e ms a e a local ou le
exi s (B en el 2021; Hagenah, Meulemann, and Akinci 2006). The ou le ha gains he la ges
numbe o eade s is Bild, a abloid ha is known o sensa ionalis and in lamma o y epo ing
on na ional poli ics (A nold 2013).
Ou wo k builds on esea ch conce ning he link be ween newspape closu es and pola iza-
ion, pa icula ly in he Ame ican con ex . In a ecen s udy, Da , Hi , and Dunaway (2018)
concep ualized elec o al pola iza ion as s aigh - icke o ing.
1
They use a ma ching design o
show ha Ame ican local newspape closu es be ween 2009 and 2012 led o highe deg ees o
s aigh - icke o ing in p esiden ial and sena o ial elec ions in 2012. The au ho s a ibu e ising
elec o al pola iza ion o di e ences in epo ing on consensual local and con es ed na ional pol-
i ics. This aligns wi h a gene al end o na ionaliza ion o poli ics –bo h in media epo ing and
cons i uen s’pe cep ions (see Hayes and Lawless 2021; Hopkins 2018; T ussle 2020). In a ecen
book, Da , Hi , and Dunaway (2021) le e age a na u al expe imen o p o ide u he e idence
ha educed exposu e o na ional news dec eases social and a ec i e pola iza ion. Ou empi ical
e idence on he na ionaliza ion mechanism hus builds on he a gumen s made by Da , Hi ,
and Dunaway (2018;2021).
While ou s udy has a simila s a ing poin o p e ious s udies in he Ame ican con ex , ou
main ou come and p oposed mechanism a e, o some ex en , speci ic o a mul ipa y sys em.
S udies in he Ame ican con ex o en ocus on spli - icke o ing as a measu e o pola ized
and pa isan elec ion esul s in a wo-pa y sys em, whe e o e s may de ia e om hei na ional
pa y p e e ences in local aces. In a coun y wi h a mul ipa y sys em such as Ge many, whe e
he e a e pa ies mo e ideologically ex eme han he la ge cen is CDU/CSU and SPD, local
newspape closu es aise he ques ion o whe he cons i uen s swi ch o pa ies u he om
he ideological cen e.
While exposu e o mo e ideologically cha ged epo ing in na ional news may make cons i u-
en s mo e ideologically ex eme in he Ame ican con ex , i should no lead o swi ching, as he e
a e no mo e ex eme ( ele an ) pa ies. Fo example, Republican o e s a e unlikely o swi ch o a
mo e ex eme pa y e en i hey become mo e conse a i e. In he Ge man con ex , howe e ,
o e s o he cen e-le SPD can swi ch o small pa ies ha end o hold mo e ideologically
ex eme posi ions. The e o e, ou analysis, like ha o Da , Hi , and Dunaway (2018), assumes
1
Rela ed s udies ha e examined he e ec o media and poli ical communica ion on a i udinal pola iza ion, measu ed as
he dispe sion o esponden ideologies, g owing pa isan animus, o pa y iden i ica ion (see Boxell, Gen zkow, and Shapi o
2017; Campan e and Hojman 2013).
1258 Fabio Ellge e al.
ha losing a local newspape and swi ching o na ional news means ha consume s a e aced
wi h ‘a media choice se wi h a highe p opo ion o pa isan and con en ious na ional op ions’
(p. 1012). Howe e , ou a gumen ocuses on a consequence unique o a mul ipa y sys em,
which is he possibili y o o e s swi ching o mo e ideologically ex eme pa ies.
Ou s udy makes se e al con ibu ions o p io esea ch on changing media en i onmen s,
poli ical pola iza ion, as well as he na ionaliza ion o news (Da , Hi , and Dunaway 2018;
2021; Hayes and Lawless 2021; Le endusky 2013a; Moskowi z 2021; P io 2013). Ou a icle con-
s i u es he mos ex ensi e s udy o examine he ela ionship be ween media consump ion and
pola iza ion in a mul ipa y se ing. P io wo k demons a ed p onounced di e ences be ween
wo-pa y se ings and mul ipa y democ acies wi h espec o pola iza ion (Reiljan 2020;
Wagne 2021). S udying he link be ween he media and pola iza ion in a mul ipa y democ acy
p o ides he unique ad an age o acing pola iza ion on he basis o ideological dispe sion in
o e beha iou . In he Ame ican con ex , measu ing elec o al pola iza ion as s aigh - icke o -
ing (Da , Hi , and Dunaway 2018) e eals li le in o ma ion abou changes in ideological
ex emi y. By compa ison, ou measu e e lec s changes in o e a i udes based on inc easing
suppo o non-cen is pa ies. By aking bo h o e beha iou and pa y ideology in o accoun ,
his ope a ionaliza ion cons i u es a mo e p ecise measu e o elec o al pola iza ion. In con as o
he Ame ican con ex , we do no ind e idence ha local newspape exi s dec ease spli - icke o -
ing. This sugges s ha in a pa y-cen ed sys em whe e candida es a e less sepa a ed om hei
pa ies (Pogun ke 1994), he loss o in o ma ion abou indi idual candida es due o newspape
exi s does no dec ease spli - icke o ing.
In addi ion, ou s udy sheds ligh on se e al in e media e s eps along he causal pa hway
be ween newspape exi s and elec o al beha iou . T iangula ing a a ie y o di e en da a sou ces,
we ace he indi idual-le el mechanisms unde lying inc eased poli ical pola iza ion a he agg e-
ga e le el. We ind ha na ional news consump ion inc eases a e local news exi s, and hus
empi ically demons a e a key mechanism ha mos p io esea ch has hypo hesized (Da ,
Hi , and Dunaway 2018; Moskowi z 2021; T ussle 2021). We shed new ligh on his subs i u ion
mechanism by examining he cen al ole o consume choice in his con ex . We show ha con-
sume s no only subs i u e local wi h na ional news bu in doing so end o choose pa icula ly
sensa ional and poli icized abloid news. The ichness o ou newspape co e age da a also allows
us o es u he empi ical implica ions o he heo e ical amewo k, including e ec he e ogen-
ei y, depending on how s ongly he exi o a single ou le a ec s he local ma ke o news. In line
wi h ou heo e ical amewo k, we es ablish ha la ge ela i e changes o he local newspape
ma ke – ha is, exi s whe e only one o wo local ou le s emain –ha e mo e p onounced e ec s
on pola iza ion.
Building on p io wo k, ou s udy also con ibu es o an unde s anding o newspape decline
and pola iza ion by ejec ing a numbe o al e na i e mechanisms o he han subs i u ion wi h
na ional news. Fi s , newspape exi s may be accompanied by unde lying demog aphic o eco-
nomic changes, which could also a ec poli ical beha iou . Rela edly, inc easing demand o
na ional news may lead o bo h newspape exi s and changes in elec o al beha iou . Second,
we examine whe he local news exi s a ec poli ical accoun abili y and go e nmen e icacy,
which may, in u n, change poli ical a i udes. Thi d, consume s may subs i u e local news
wi h pa isan media, which could pola ize poli ical a i udes. As we discuss in de ail in sec ion
A.1 in he SI, quan i a i e and quali a i e e idence sugges s ha hese mechanisms a e o
mino impo ance in ou se ing.
Finally, ou panel da a co e ing h ee decades allows us o es ima e he empo al pe sis ence o
agg ega e elec o al e ec s: we documen an inc ease in elec o al pola iza ion o abou eigh yea s
a e local news exi . Ou esul s sugges ha local news exi s ha e long- e m poli ical conse-
quences ha ma e ialize o e mul iple elec ion pe iods. This implies ha p io esea ch on his
opic, which has mos ly ocused on sho e ime ames, may ha e unde es ima ed he o e all
cumula i e e ec o local news exi s on poli ical pola iza ion (Da , Hi , and Dunaway 2018;
B i ish Jou nal o Poli ical Science 1259
2021; Moskowi z 2021). Going beyond p io wo k, he empo al scope o ou da a also allows us
o assess changing local media be o e and a e he widesp ead in oduc ion o he in e ne . We
demons a e ha newspape exi s p ima ily a ec ed pola iza ion p io o 2000 when ci cula ion
was highe and he in e ne was no a p ima y sou ce o in o ma ion. O e ime, changes in he
gene al media landscape may hus mode a e he e ec s o local newspape exi s on poli ical
beha iou (P io 2013).
Theo e ical F amewo k
How do local newspape exi s a ec poli ical beha iou and pola iza ion? Newspape exi s can
lead o mo e na ionalized pa e ns o indi idual news consump ion. Reade s a ec ed by a local
news exi can (i) swi ch o a di e en local ou le , (ii) subs i u e wi h a na ional newspape ,
(iii) inc ease consump ion o o he mo e na ionalized news sou ces ( o example adio o ele i-
sion), o (i ) no subs i u e a all, and he eby educing hei o e all news consump ion. We no e
ha hese o ms o consump ion shi s a e no mu ually exclusi e. The la e h ee esponses will
induce an inc ease in he ela i e sha e o na ional news consump ion, pa icula ly when cons i-
uen s eac o exi s by subsc ibing o a na ional newspape .
How do such shi s om local o na ional news exposu e a ec poli ical p e e ences?
Fi s and o emos , a shi o na ional news consump ion heigh ens he salience o na ional
poli ics (Moskowi z 2021). This easoning builds on heo e ical a gumen s laid ou in p io
wo k by Da , Hi , and Dunaway (2018;2021). While local newspape s in Ge many de o e
some space o co e age o na ional issues, local poli ics cons i u e a la ge sha e o hei
epo ing.
2
By con as , poli ical epo ing in na ional ou le s ocuses almos exclusi ely
on na ional poli ics. We p esen suppo ing e idence on di e ences in epo ing based
on a con en analysis o a subsample o local and na ional newspape s be ween 2009 and
2013 p o ided by he Ge man Longi udinal Elec ion S udy (GLES 2014) in appendix
sec ion A.8.
While epo ing in na ional ou le s does no exhibi s ong ideological biases owa d he le o
he igh , na ional ou le s ypically p esen eade s wi h mo e ideologically ex eme a gumen s
han local news ou le s (see sec ions A.2, A.6, and A.8 o e idence).
3
This mos di ec ly s ems
om he ac ha na ional poli ics is ea u ed mo e p ominen ly in na ional news and is mo e
con en ious han local poli ics (Wehling and Kos 2010). Compa ed o local ou le s, cons i uen s
a e ela i ely mo e exposed o ei he c oss-cu ing o like-minded epo ing when eading
na ional ou le s.
4
Ideological di e ences in Ge man local poli ics a e less p onounced han in na ional poli ics
(Hol kamp 2017). In Ba a ia, o example, 45 pe cen o cu en mayo s a e no a ilia ed wi h
any o he la ge na ional pa ies (Baye isches Landesam ü S a is ik 2021). This s ands in s a k
con as o he na ional pa liamen , whe e each o he 709 MPs is s ic ly a ilia ed wi h a single
pa y (Bundes ag 2021). Compa ed o he na ional le el, local poli ics is ma ked by a lowe salience
o pa isan iden i ica ion (Me ag 2016), mo e consensus ac oss pa y lines (Wehling and Kos
2010), and a highe p opensi y o candida es o un as independen s o on bipa isan icke s.
2
We p o ide a de ailed discussion o local news co e age and he ole o local newspape s in he Ge man media landscape
in sec ion A.2 in he SI.
3
In he Uni ed S a es, ela ed esea ch e eals a simila pa e n: mo e ex eme poli icians a e mo e likely o be ea u ed in
na ional news (Padge , Dunaway, and Da 2019; Wagne and G uszczynski 2018).
4
To make his s a emen mo e p ecise, we can o malize i as ollows: seconda y e idence sugges s ha we can hink o he
ideological posi ion epo ed on in a gi en a icle X
ij
as no mally dis ibu ed, such ha Xij N(
m
j,
s
2
j). Compa ing a
na ional ou le j o a local ou le j*, we p opose ha μ
j
≈μ
j*
, bu
s
2
j>
s
2
j∗– he e o e, he a e age a i udinal ex emi y o a
gi en news s o y iin na ional ou le jis E(|Xij|)=
s
j
2
√/
p
√.
s
j∗
2
√/
p
√=E(|Xij∗|). This means ha eade s will, on
a e age, encoun e mo e ex eme posi ions in he na ional ou le , bu he expec ed ideological slan o he a icles is he
same ac oss he na ional and he local ou le .
1260 Fabio Ellge e al.
When local newspape s exi , pola iza ion can be exace ba ed h ough inc eased exposu e o
ideologically cha ged na ional poli ics (Da , Hi , and Dunaway 2018;2021). This can occu
ei he h ough he ac i a ion o la en ideological disposi ions o h ough a pe suasion mechan-
ism, whe e indi iduals mo e u he away om he ideological cen e. Based on p io indings in
poli ical psychology, we p opose ha g ea e ideological ex emi y can pola ize eade s e en i he
epo ing is, on a e age, no slan ed. Below, we desc ibe hese mechanisms and hen discuss al e -
na i e mechanisms as well as di e en ial e ec s in local and na ional elec ions.
How can mo e ideologically ex eme posi ions in na ional news ou le s pola ize eade s? To
heo ize how messaging in na ional ou le s may pe suade consume s, ou heo e ical a gumen
d aws on he mo i a ed easoning amewo k (Tabe , Cann, and Kucso a 2009). P io esea ch
has shown how exposu e o bo h like-minded and c oss-cu ing in o ma ion can lead o e s o
espouse mo e ideologically ex eme posi ions. This can be d i en by con i ma ion bias – he en-
dency o unc i ically accep ideologically cong uen a gumen s –and discon i ma ion bias – he
endency o ejec and coun e a gue ideologically incong uen posi ions (Di o and Lopez 1992;
Le endusky 2013a; Tabe , Cann, and Kucso a 2009; Tabe and Lodge 2006). Tabe , Cann, and
Kucso a (2009) p esen e idence on he ela ion be ween discon i ma ion bias and a i udinal
pola iza ion: indi iduals become mo e pola ized e en when aced wi h in o ma ion ha is no
cong uen wi h p io belie s. Acco dingly, Tabe , Cann, and Kucso a (2009, 150) ema k ha
‘a i udes […] become mo e ex eme as a consequence o he a i ude cong uence and discon-
i ma ion biases, e en when he o e all se o a gumen s is balanced be ween p o and con’.
Rela ed e idence on he pola izing e ec s o ideologically ex eme in o ma ion ha is ei he con-
g uen o incong uen wi h p io a i udes has been shown by Le endusky (2013b), Tabe and
Lodge (2006), and Bail e al. (2018).
5
The p e ious pa ag aph has ou lined how wo ypes o biased in o ma ion p ocessing can
induce a i udinal pola iza ion h ough pe suasion i eade s a e exposed o mo e ex eme iew-
poin s in he na ional news. This con as s local news, which co e s mo e consensual and o en
non-pa isan local poli ics. We, he e o e, expec ha pola iza ion due o pe suasion is less p ob-
able when indi iduals only consume local news, as he ela i e lack o ideological di e ences in
local poli ics does no ‘a ouse su icien pa isan mo i a ion o induce much-biased p ocessing’
(Tabe and Lodge 2006, 756). We u he co obo a e his h ough a co ela ional analysis, in
which we use he GLES su ey da a o show ha local news consume s a e mo e likely o suppo
cen is pa ies (see sec ion A.5 in he SI).
The heigh ened salience o pa isan posi ions in na ional news ou le s may p ime and
s eng hen poli ical iden i ies, e en wi hou pe suading consume s o change hei opinions
(Lelkes, Sood, and Iyenga 2017). S onge ideological p iming h ough na ional news consump-
ion may inc ease a i udinal ex emi y as indi iduals inc easingly seek ou posi ions ha con i m
hei p io ideological belie s Le endusky (2013b).
E en i indi iduals do no swi ch om a local o a na ional ou le , exi s may ne e heless
exace ba e pola iza ion among hose who p e iously ead he ou le . Local news ou le s can main-
ain communi y iden i y and cohesion by p o iding in o ma ion abou local ma e s (Pa k,
Fishe , and Lee 2022). This is e lec ed in posi i e co e age o social connec ions and solida i y
be ween ci izens, as well as a ou able epo ing on socie al ins i u ions and ci ic pa icipa ion
(Leupold, Klinge , and Ja en 2018). Exposu e o his kind o co e age may educe poli ical pola -
iza ion by (i) spu ing suppo o cen is pa ies ha ep esen a b oade se o in e es s o (ii)
educing he salience o g oup-based iden i ies is-à- is a common, place-based iden i y.
Rela edly, local newspape s may ac as a coo dina ion de ice by in o ming o e s ha he
CDU/CSU and he SDP a e he pa ies mos likely o succeed in o ming go e ning coali ions.
5
We no e ha an impo an scope condi ion o he mo i a e easoning mechanism is ha a leas some local news con-
sume s do no hold s ongly mode a e belie s. Pu di e en ly, a leas some local news consume s need o ha e p io belie s
ha a e su icien ly le o igh o allow o biased p ocessing o ideologically ex eme posi ions in na ional news.
B i ish Jou nal o Poli ical Science 1261
A local newspape exi ha does no esul in swi ching o ano he news ou le may hen esul in
inc eased consump ion o en e ainmen a he expense o in o ma ion abou poli ical pa ies. As
a esul , o e s may now base hei elec o al decisions on exp essi e p e e ences, which may esul
in mo e idiosync a ic choices a he o ing boo h. In pa icula , such idiosync a ic choices can
bene i smalle pa ies i o e s now pu less weigh on pa y cha ac e is ics such as poli ical cen-
ism o elec abili y. Empi ically, his mechanism aligns wi h p io wo k by Campan e, Du an e,
and Sobb io (2018), who demons a ed he endency o en e ainmen consump ion o c owd
ou exposu e o poli ical news in he I alian case. We no e ha a g ea e likelihood o exp essi e
o ing a e newspape exi s is obse a ionally equi alen o poli ical pola iza ion i i bene i s
non-cen is pa ies. Howe e , i may be concep ually di e en i he easons o de ia ions
away om he cen e a e un ela ed o o e ideology o he s eng h o poli ical iden i ies.
Summa izing he a gumen , a shi om local o na ional news may (i) pe suade consume s o
espouse mo e poli ically ex eme posi ions and (ii) ac i a e o s eng hen poli ical iden i ies. In
addi ion, newspape exi s may shi public ocus om poli ical news o en e ainmen , esul ing
in o e s making elec o al choices based on exp essi e p e e ences a he han pa y elec abili y,
po en ially bene i ing smalle pa ies. In addi ion o ou main heo e ical a gumen ou lined
abo e, we now u he discuss (i) e ec s in local and na ional elec ions, (ii) po en ial e ec s on
spli - icke o ing, (iii) al e na i e mechanisms, and (i ) p edic ions ha s em om he RAS
amewo k p oposed by Zalle (1992).
E ec s in local s. na ional elec ions: We a gue ha local news exi s can lead o inc eased elec o al
pola iza ion in bo h ede al and local elec ions. Gi en ha he pa isan poli ical issues discussed
in na ional news ou le s p edominan ly pe ain o ede ally go e ned ma e s ( o example, immi-
g a ion, ax policy, e c.), we expec a i udinal pola iza ion su ounding hese issues o ma e ialize
in ede al elec ions. Howe e , a i udinal pola iza ion migh also ‘spill o e ’ o local elec ions and
esul in inc eased elec o al pola iza ion a his le el. This is because local news exi s dec ease
exposu e o in o ma ion abou local poli ics (Hayes and Lawless 2017), and migh he eby lead
o e s o inc easingly cas ballo s based on na ional issues and pa isanship.
Spli - icke o ing: P io esea ch in he Ame ican con ex sugges s ha local newspape decline
leads o less spli - icke o ing. This is because local newspape exi s lead o a decline in in o -
ma ion abou indi idual candida es. Vo e s, he e o e, ha e o ely on na ional-le el heu is ics
and na ional-le el pa isan iden i ies when hey cas ballo s (Moskowi z 2021). In a supplemen-
a y analysis, we ind no e idence ha newspape exi s in luence spli - icke o ing in Ge man
ede al elec ions, which is likely due o he coun y’s pa y-cen ed poli ical sys em. We elabo a e
on his esul in sec ions 4 and A.16.
Al e na i e mechanisms: We conside se e al al e na i e mechanisms. These a e (i) swi ches o
highly pa isan ou le s o o he local news ou le s, (ii) unde lying economic o demog aphic
changes ha could a ec bo h newspape exi s and elec o al beha iou , and (iii) newspape
exi s and hei epe cussions o poli ical accoun abili y and go e nmen e icacy. We discuss
and es hese mechanisms in mo e de ail in sec ion A.1 in he SI. Howe e , we a gue ha
hese mechanisms a e likely o mino impo ance in ou se ing.
Zalle ’s RAS amewo k: The Recei e-Accep -Sample (RAS) model by Zalle (1992) p esen s an
al e na i e o mo i a ed easoning in p edic ing how cons i uen s espond o changes in he
media landscape. Unlike mo i a ed easoning, he RAS model sugges s ha indi iduals ei he
align wi h he ideological di ec ion o a ecei ed message o ejec i based on p e-exis ing ideo-
logical cues. In speci ic condi ions, RAS p edic s ha ideologically ex eme bu balanced news
does no esul in ideological pola iza ion. Howe e , RAS also sugges s ha exposu e o balanced
o neu al news can lead o pola iza ion, especially when consume s a e poli ically sophis ica ed
1262 Fabio Ellge e al.
and ha e s ong pa isan p io s who selec i ely accep messages ha align wi h hei ideologies.
We discuss he RAS amewo k and i s empi ical implica ions in ou se ing in mo e de ail in
sec ion A.1.1.
Da a and Empi ical S a egy
Elec o al Pola iza ion
Ou main ou come a iable is poli ical pola iza ion measu ed on he basis o elec ion esul s in
ede al and municipal elec ions. Ou da a co e s eigh ede al elec ions held be ween 1980 and
2009 (see Table A.1 o summa y s a is ics). Ou da a on municipal elec ions co e s he pe iod
1990–2009 (Rademache 2018): da a o he 1980s is no cen ally a ailable. A de ailed lis o
all elec ions in ou analysis can be ound in Fig. A.6 in he SI. To map be ween elec o al esul s
and pola iza ion in a mul ipa y se ing, we d aw on he widely-used pola iza ion measu e in o-
duced by Dal on (2008). Fo ede al elec ions, ou main ou come a iable Y
c,
, elec o al pola iza-
ion in coun y cin elec ion , is de ined as:
Yc, =
j
c
j,c, (
j, −
)2
whe e ψ
j,c,
is he o e sha e o pa y jin coun y ca ime ,τ
j,
is he posi ion o pa y ja ime-
pe iod ac oss he le - igh poli ical spec um, and
is he mean pa y posi ion in Ge many a
elec ion pe iod (weigh ed by he numbe o o es). To measu e τ
j,
, he ideological posi ion o
pa ies, we d aw on he le - igh RILE sco es om he Mani es o p ojec (Volkens e al. 2020). To
he bes o ou knowledge, his is he only da a sou ce ha co e s pa y posi ions a elec ions
h oughou ou en i e s udy pe iod s a ing in he 1980s. We isualize he mean RILE sco es
o Ge man poli ical pa ies du ing ou s udy pe iod in igu e A.5 in he appendix.
The pola iza ion index cap u es poli ical pola iza ion in a mul ipa y se ing as i is sensi i e o
(i) changes in pa y ideology and (ii) changes in o e s’le el o elec o al suppo o di e en
pa ies. Theo e ically, pola iza ion can be d i en by he elec o al success o new ex eme pa ies,
he ideological adicaliza ion o exis ing pa ies, o e shi s owa ds al eady exis ing ex eme pa -
ies, o any combina ion o hese dynamics. The pola iza ion index is able o cap u e and
ep esen all hese possible changes ha equen ly occu in a mul ipa y se ing.
6
The measu e
cap u es he in ui ion ha elec o al pola iza ion is he smalles when all o es a e concen a ed
in he cen e o he poli ical spec um o when all pa ies un on he same ideological pla o m.
By con as , he pola iza ion index is maximized when he o e is spli be ween pa ies on he a
le and he a igh . Ma hema ically, ou measu e co esponds o he weigh ed s anda d de i-
a ion o pa y posi ions in coun y ca ime . The weigh s a y ac oss coun ies and co espond
o each pa y’s o e sha e in a gi en coun y. The pa y posi ions on he le - igh poli ical spec-
um a y o e ime bu a e cons an ac oss coun ies wi hin a gi en elec ion yea . We isualize he
dis ibu ion o ou pola iza ion measu e ac oss all ede al elec ions be ween 1983 and 2009 in
Fig. A.2 in he SI. Pola iza ion ends o be highe in Eas Ge many, whe e he Le Pa y a ac s
a sizable numbe o suppo e s. Ou measu emen o he pola iza ion ou come is analogous o
municipal elec ions (see appendix sec ion A.3.3 o mo e de ails).
Ou second ou come is he combined o e sha e o small pa ies (FDP, Le Pa y, G eens) as
an al e na i e elec o al ou come. Small pa ies gene ally hold posi ions u he om he ideo-
logical cen e and o en emphasize single poli ical issues. The G een Pa y, o example, p ima ily
campaigned on en i onmen al p o ec ion du ing ou s udy pe iod. The socialis Le Pa y, on he
o he hand, ypically ocuses on edis ibu ion in hei elec ion campaigns.
6
This di e s om p io measu es in he Uni ed S a es, which mos ly used s aigh - icke o ing as an ou come o in e es
(see Da , Hi , and Dunaway 2018).
B i ish Jou nal o Poli ical Science 1263
Na ional News Consump ion
Ha ing es ablished ha newspape exi s inc ease elec o al pola iza ion, we now p obe whe he
local newspape exi s a ec he consump ion o na ional news.
In Fig. 5, we p esen he esul s om a leads and lags speci ica ion. We es ima e he e ec o
newspape exi s on absolu e and ela i e le els o na ional newspape consump ion o e ime. We
ind ha na ional news consump ion inc eases a e local newspape s exi . This holds o ou
absolu e and ela i e measu es o na ional news consump ion. These esul s sugges ha local
news exi s lead o a dec ease in exposu e o consensual local poli ics. A he same ime, exposu e
o pa isan poli ics a he na ional le el inc eases. Reassu ingly, he iming o hese e ec s is simi-
la o wha we obse e o he elec o al ou comes in Fig. 4. Mo e p ecisely, e ec s inc ease a e
he exi s occu and a e la ges in he second elec ion a e he exi .
Fu he , he esul s in Fig. 4 shed ligh on ends in na ional news consump ion p io o local
news exi s. Feasibly, local news exi s may be d i en by inc eased demand o na ional news, which
is accompanied by a d op in demand o local news. Ou na ional news consump ion da a allows
us o di ec ly assess his hypo hesis. Based on he es ima es o he wo elec ions p io o exi s
shown in Fig. 4, we ind no e idence ha demand o na ional news inc eased di e en ially in
coun ies ha e en ually expe ience local news exi s. In summa y, ou esul s sugges ha he
inc eased consump ion o na ional news does no occu be o e local newspape s exi s. As an add-
i ional obus ness check, we also es ima e he e ec o local news exi s on na ional news con-
sump ion including ime- a ying co a ia es. Reassu ingly, ou esul s emain unchanged when
con olling o changes in popula ion and GDP pe capi a (see Table A.12 in he SI).
The magni ude o he e ec size es ima es may appea small a i s sigh . Howe e , we empha-
size ha ou consump ion da a co e s he en i e popula ion, o which only a small subse is di -
ec ly a ec ed by local newspape exi s. Speci ically, we would only expec a shi in he media die
o indi iduals who p e iously subsc ibed o he exi ing ou le . Fo he absolu e consump ion
measu e, ou e ec size es ima es imply ha wo in one hund ed people in a gi en coun y sub-
sc ibe o a na ional news ou le as a esul o local news exi . Howe e , gi en ha local news ou -
le s only each abou hal o all adul s in Ge many (see sec ion A.2 in he SI) and ha he exi ing
ou le likely held less han 50 pe cen ma ke sha e, he popula ion sha e di ec ly a ec ed by he
exi is likely much lowe han 25 pe cen (see also oo no e 9). Simila ly, o ou ela i e measu e
o na ional news consump ion, ou esul s a e consis en wi h a scena io in which he sha e o
Figu e 5. E ec s o newspape exi s on na ional news consump ion.
No e: The plo shows he es ima ed e ec (τ) o a newspape exi on wo ou comes: absolu e le els o na ional newspape consump-
ion, and he ela i e sha e o na ional ou le s among all newspape s ha indi iduals consume (see sec ion ‘Na ional news consump-
ion’). The x-axis indica es he elec ion ela i e o he ime when he newspape exi occu s. Elec ions labeled −1 and −2 a e p io o
exi , while elec ions labeled 1 and 2 a e a e exi occu s. N= 1,769 o he i s elec ion a e exi . Mo e de ails a e gi en in Table A.12.
1270 Fabio Ellge e al.
na ional news consump ion inc eases om 50 pe cen o 100 pe cen o wo in one hund ed
people in a gi en coun y. Fo example, his would apply o indi iduals who consumed one
na ional and one local ou le each p io o exi and con inue o ead he na ional ou le a e
he local newspape exi occu ed.
11
Conside ing he size o he ea ed popula ion, he e ec
size es ima es a e no only s a is ically signi ican bu also subs an i ely meaning ul.
In he nex s ep, we disagg ega e he na ional news eade ship e ec s by ou le . In Fig. 6,we
p esen he e ec s o local newspape exi s on he likelihood o consuming ou di e en na ional
newspape s. Toge he , hese ou le s accoun ed o mo e han 90 pe cen o he Ge man na ional
newspape ma ke as o 2010 (B and 2020). We obse e la ge and signi ican posi i e e ec s o
Bild, a abloid ha is Ge many’s mos widely ead na ional ou le . Fo he FAZ,SZ and Wel ,
which epo less sensa ionally and mo e se iously, we do no obse e any changes in eade ship
due o local newspape exi s. While we do no ha e da a on indi idual mo i a ions o swi ching
om local newspape s o Bild a he han o he ou le s, his esul could be explained by he ac
ha Bild is mo e eadily a ailable, cheape , and epo s in a mo e accessible manne (A nold
2013). Acco dingly, he eade ship o Bild closely esembles he o e all popula ion in e ms o
socio-demog aphic cha ac e is ics and cu s ac oss pa isan lines (we elabo a e on hese poin s
in appendix sec ions A.6 and A.17).
We a gue ha he shi om local o na ional news pola izes consume s. An al e na i e con-
sequence is ha consume s adop he poli ical leanings o he na ional ou le hey swi ch o. This
scena io p esumes ha na ional ou le s a e su icien ly pa isan o pe suade consume s. While
some o he na ional ou le s we analyze ha e been desc ibed as le - o igh -leaning, da a on
he pa isan composi ion o hei eade ship indica es ha all he majo na ional newspape s
a e ead by people om ac oss he poli ical spec um (see sec ion A.6 in he appendix o
mo e in o ma ion). In pa icula , he ou newspape s desc ibed in Fig. 6 all ha e a la ge sha e
o eade s who do no iden i y wi h any poli ical pa y –in ac , he newspape wi h he highes
sha e o ‘independen ’ eade s is Bild. As a esul , we do no expec s ong pe suasion e ec s in a
speci ic di ec ion due o he p onounced pa isan slan o one o he na ional ou le s.
O e all, hese esul s a e consis en wi h he a gumen ha he pola izing e ec s o newspape
exi s s em om an inc ease in na ional news consump ion. Mo e speci ically, na ional news
exposu e ises because a subse o a ec ed consume s inc ease hei consump ion o na ional ou -
le s, sugges ing ha a shi owa ds na ional news ollowing local newspape exi s is he p ima y
d i e o ou main esul s.
Indi idual-le el Poli iciza ion
Finally, we p esen supplemen a y e idence on he poli icizing na u e o local newspape closu es
in he appendix, sec ion A.19. Poli ical in ol emen and pa isan a achmen a e es ablished
indi idual-le el co ela es o inc easing pola iza ion (see Iyenga e al. 2018; Mason 2018).
Using a la ge-scale panel su ey and an empi ical s a egy ha mimics he i s -di e ences es i-
ma o o he main analysis, igu e A.17 in he SI plo s esponden s’inc easing pa isan iden i i-
ca ion and poli ical in ol emen ollowing local newspape exi s in hei coun y. We discuss he
da a, ope a ionaliza ion, empi ical s a egy, and esul s in mo e de ail in sec ion A.19 o he SI.
Discussion
Does he decline o local news a ec poli ical pola iza ion? Based on a no el da a se on he en y
and exi o all Ge man local newspape s be ween 1980 and 2009, we show ha local newspape
exi s inc ease elec o al pola iza ion. Going beyond p io wo k, we p o ide new e idence on he
11
We no e ha ou agg ega e esul s o na ional news consump ion can a ise om mul iple, he e ogeneous scena ios o
shi ing media consump ion a he indi idual le el. We desc ibe ypical scena ios o illus a e he e ec size magni ude.
B i ish Jou nal o Poli ical Science 1271
link be ween local media ma ke changes in media die s using a la ge-scale annual media con-
sump ion su ey. We show ha bo h absolu e and ela i e consump ion o na ional newspape s
inc eases a e local newspape s exi . Speci ically, he exi s lead eade s o consume he abloid
Bild, which a guably exhibi s he s onges emphasis on con lic and pa isan di e ences
among all majo Ge man na ional newspape s (A nold 2013; Wellb ock 2011). These esul s
a e consis en wi h heo e ical p edic ions ha s em om ei he (i) he mo i a ed easoning
amewo k, (ii) Zalle ’s RAS model unde he condi ion ha consume s ejec coun e -a i udinal
messaging, o (iii) a mechanism whe e a local newspape ac s as a coo dina ion de ice, leading o
exp essi e o ing a e exi s. Taken oge he , ou esul s unde line how changes in local media
ma ke s shape pola iza ion in es ablished democ acies (Da , Hi , and Dunaway 2018,2021;
Moskowi z 2021).
Be o e discussing he wide implica ions o ou indings, we no e ha he e idence we p esen
on he mechanisms d i ing ou esul s emains sugges i e. Because we use sepa a e la ge-scale
da a se s o ack poli ical a i udes, elec o al beha iou , and media consump ion a e local
news exi s, we canno examine he link be ween hese a iables wi hin he same indi idual.
Ins ead, we p esen ed agg ega e (coun y-le el) e idence ha he e ec o local news exi s on pol-
i ical pola iza ion is media ed by inc eased exposu e o na ional news and ha local news ead-
e ship co ela es wi h o ing o cen is pa ies (see Table A.3 in he SI). Fu u e esea ch migh
be able o causally demons a e his mechanism a he indi idual le el. Fo an obse a ional s udy,
his would equi e a single panel su ey ha co e s media consump ion beha iou , pa isanship,
and poli ical a i udes o e a su icien ly long ime span.
Ano he impo an limi a ion o ou s udy is ha we a e no p o iding e idence on changing
local news con en o e ime (Hayes and Lawless 2017;2021). While ou indings demons a e
he consequences o newspape exi s, Hayes and Lawless (2021) p ominen ly showcase he impac
o inc easingly na ionalized news con en in he emaining local ou le s on elec o al engagemen .
In es iga ing he in luence o newspape con en among emaining ou le s in a mul ipa y con ex
cons i u es ano he ui ul a enue o u u e esea ch. Impo an ly, looking a ecen changes in
local con en migh p o ide u he e idence on why we do no obse e e ec s in la e ime
pe iods.
We also no e a numbe o seemingly con adic o y ends a he mac o (na ionwide) and
mic o (coun y) le els in ou se ing. Vo ing o smalle pa ies, o example, inc eased sha ply
a e he yea 2000, ye he e ec s o local news exi s we e mos p onounced in he p e-2000 pe -
iod. These seemingly con adic o y indings can be econciled h ough he lens o Simpson’s
Pa adox, which highligh s ha a ela ionship be ween a iables can di e o e en e e se when
conside ing subg oups o agg ega e da a. A he agg ega e le el, he ela ionship be ween
Figu e 6. E ec s o newspape exi s
on na ional news consump ion, by
ou le .
No e: The plo shows he es ima ed
e ec o a local newspape exi on he
consump ion o ou na ional newspa-
pe s. We employ he same i s -di e ence
speci ica ion as in he main analyses. We
p esen e ec s one elec ion a e exi
(labeled ‘1–4 yea s a e exi ’,
N= 1,769) and wo elec ions a e exi
(labeled ‘5–8 yea s a e exi ’,N= 1,771).
1272 Fabio Ellge e al.
newspape exi s and pola iza ion can be obscu ed due o he p esence o o he con ounding ac-
o s ha may ha e a mo e signi ican impac on elec o al pola iza ion a he na ional le el. The
pos -2000 pe iod is cha ac e ized by a numbe o mac o-le el ends ( o example, inc eased
immig a ion, exposu e o globaliza ion, and cu s o wel a e spending) ha may ha e con ibu ed
o inc eased pola iza ion a he na ional le el. In addi ion, we no e ha ou s udy does no seek o
explain all a ia ions in elec o al pola iza ion. Ra he , we zoom in on one con ibu ing ac o ha
can explain a small bu meaning ul sha e o a ia ion in pola iza ion. In o he wo ds, ou s udy is
mo i a ed by unde s anding he ‘causes o e ec s’, ye ou empi ical analysis ocuses on iden i-
ying he ‘e ec s o causes’(Gelman and Imbens 2013)– ha is he e ec s o local news exi s on
elec o al beha iou .
Going beyond he Ame ican case, ou s udy expands on he link be ween na ionalized news
en i onmen s and pola iza ion in he con ex o es ablished mul ipa y democ acies. In
Ge many, Eu ope’s la ges democ acy, elec o al pola iza ion ises ollowing local news exi s.
Impo an ly, ou s udy shows ha he supply o news al eady a ec ed poli ical pola iza ion
p io o he in oduc ion o social ne wo ks and pa isan online news (see Boxell, Gen zkow,
and Shapi o 2017). E en be o e he ex ensi e a ailabili y o b oadband in e ne , p essu es on
he local news ma ke and subs i u ion wi h na ional abloid media a e de e minan s o mo e di -
isi e poli ics and ele a ed pa isan a achmen s. The mechanism o changing consump ion habi s
also aligns wi h exis ing ecen s udies ha e e ence pa y eli es as impo an causes o mass
pola iza ion (Rogowski and Su he land 2016; Webs e and Ab amowi z 2017). To he ex en
ha elec o al pola iza ion is d i en by inc eased exposu e o con es ed policy deba es and co e -
age o adical pa y leade s in na ional news ou le s (Padge , Dunaway, and Da 2019; Wagne
and G uszczynski 2018), elec o al pola iza ion in mul ipa y sys ems can also be unde s ood as, a
leas pa ially, eli e-d i en (Lenz 2013).
Whe he ou esul s a el o o he con ex s depends on he na u e o local poli ics, he ele-
ance o local news, and di e ences in elec o al sys ems. Wi h ega d o he decline o he news-
pape publishing ma ke , Ge many is one among many de eloped coun ies whe e eade ship
and e enues a e dec easing. Howe e , he elec o al e ec s o local newspape exi s likely a y
ac oss con ex s. In pa icula , we would expec simila esul s in coun ies whe e na ional poli ics
is mo e ad e sa ial, and whe e local news consump ion emains su icien ly salien o he popu-
la ion. In he Eu opean con ex , Ge man consume s a e among he mos likely o s a e ha hey
closely ollow local news. While o e s in o he Eu opean coun ies end o ollow local news less
closely, a b oade in e na ional compa ison e eals ha local news emains highly ele an in a
mul i ude o con ex s, such as sub-Saha an A ica, India, and Aus alia (Mi chell e al. 2018).
Recessions o local newspape ma ke s and ends o inc easingly na ionalized news co e age
may hus ha e elec o al epe cussions in a wide a ie y o se ings, e en beyond adi ional
Wes e n democ acies.
Supplemen a y ma e ial. The supplemen a y ma e ial o his a icle can be ound a h ps://doi.o g/10.1017/
S0007123424000243.
Da a a ailabili y s a emen . Replica ion da a o his a icle can be ound in Ha a d Da a e se a : h ps://doi.o g/10.7910/
DVN/MD0UYZ
Acknowledgemen s. We hank Daniel Bischo , Maca an Humph eys, Ga y King, Shi o Ku iwaki, Heike Klue e , Michael
Olson, Pia Ra le , Lukas S oe ze , Ma kus Wagne , Anna Wilke, S ephane Wol on, Delia Zollinge , audiences a APSA,
Ha a d, HU Be lin, he IPI g oup a WZB Be lin, EUI as well as p io anonymous e iewe s o help ul commen s. We
hank Inga B en el and Jö g Hagenah o assis ance in accessing and cleaning he media consump ion da a.
Financial suppo . This wo k was suppo ed by he Ins i u e o Quan i a i e Social Science a Ha a d Uni e si y (g adu-
a e s uden esea ch g an ).
Compe ing in e es s. The au ho s decla e none.
B i ish Jou nal o Poli ical Science 1273
Re e ences
A nold K (2013) Die Quali ä De BILD-Zei ung. Jou nalis ische Leis ungsk i e ien Und Die Boule a dp esse. Zei sch i ü
Kommunika ionsökologie 15(1), 43–48.
Bail CA e al. (2018) Exposu e o opposing iews on social media can inc ease poli ical pola iza ion. P oceedings o he
Na ional Academy o Sciences 115(37), 9216–9221.
Baye isches Landesam ü S a is ik (2021) E s e Bü ge meis e , Obe bü ge meis e Und Land ä e. A ailable om
h ps://www.s a is ik.baye n.de/wahlen/kommunalwahlen/bgm/
Boxell L, Gen zkow M and Shapi o JM (2017) G ea e in e ne use is no associa ed wi h as e g ow h in poli ical pola -
iza ion among US demog aphic g oups. P oceedings o he Na ional Academy o Sciences 114(40), 10612–10617.
B and M (2020) Axel Sp inge on P in k ise Besonde s Be o en. A ailable om h ps://de.s a is a.com/in og a ik/10376/
e kau e-au lage-uebe egionale - ageszei ungen-in-deu schland/
B en el I (2021) Media-Analyse: P essemedien - Lang is da en (Ve sion 2.0) (Ve sion: 1.0.0).
Bucchiane i P (2020) Pa y compe i ion and coali ional s abili y: E idence om Ame ican local go e nmen . Ame ican
Poli ical Science Re iew 114(4), 1055–1070.
Cagé J (2020) Media compe i ion, in o ma ion p o ision and poli ical pa icipa ion: E idence om F ench local newspape s
and elec ions, 1944–2014. Jou nal o Public Economics 185, 104077.
Campan e FR and Hojman DA (2013) Media and pola iza ion. Jou nal o Public Economics 100,79–92.
Campan e F, Du an e R and Sobb io F (2018) Poli ics 2.0: The mul i ace ed e ec o b oadband in e ne on poli ical pa -
icipa ion. Jou nal o he Eu opean Economic Associa ion 16(4), 1094–1136.
Dal on RJ (2008) The quan i y and he quali y o pa y sys ems: Pa y sys em pola iza ion, i s measu emen , and i s conse-
quences. Compa a i e Poli ical S udies 41(7), 899–920.
Da JP, Hi MP and Dunaway JL (2018) Newspape closu es pola ize o ing beha io . Jou nal o Communica ion 68(6),
1007–1028.
Da JP, Hi MP and Dunaway JL (2021) Home S yle Opinion: How Local Newspape s Can Slow Pola iza ion. Camb idge:
Camb idge Uni e si y P ess.
Deu sche Bundes ag. (2021) Abgeo dne e des Deu schen Bundes ags. A ailable om h ps://www.bundes ag.de/mdb
de Chaisema in C and D’Haul œuille X (2020) Two-way ixed e ec s es ima o s wi h he e ogeneous ea men e ec s.
Ame ican Economic Re iew 110(9), 2964–2996.
Di o PH and Lopez DF (1992) Mo i a ed skep icism: Use o di e en ial decision c i e ia o p e e ed and nonp e e ed
conclusions. Jou nal o Pe sonali y and Social Psychology 63(4), 568–584.
Ellge F e al. (2024) “Replica ion Da a o : ‘Local Newspape Decline and Poli ical Pola iza ion –E idence om a
Mul i-Pa y Se ing’”. A ailable om h ps://doi.o g/10.7910/DVN/MD0UYZ, Ha a d Da a e se, V1.
Gelman A and Imbens G (2013) Why Ask Why? Fo wa d Causal In e ence and Re e se Causal Ques ions. Technical Repo
w19614, Na ional Bu eau o Economic Resea ch Camb idge, MA.
Gen zkow M, Shapi o JM and Sinkinson M (2011) The e ec o newspape en y and exi on elec o al poli ics. Ame ican
Economic Re iew 101(7), 2980–3018.
Gil de Zúñiga H, Jung N and S. Valenzuela S (2012) Social media use o news and indi iduals’social capi al, ci ic engage-
men and poli ical pa icipa ion. Jou nal o Compu e -Media ed Communica ion 17(3), 319–336.
GLES (2014) Long-Te m Media Agenda Analysis (GLES). GESIS Da a A chi e, Cologne. ZA5323 Da a ile Ve sion
2.0.0. h ps://doi.o g/10.4232/1.11866
Hagenah J, Meulemann H and Akinci H (2006) Ge man Media-Analyse (MA): A la ge scale comme cial da a sou ce a ail-
able o seconda y analyses on Media use and social change. Jou nal o Con ex ual Economics –Schmolle s Jah buch
126(1), 129–137.
Hayes D and Lawless JL (2015) As local news goes, so goes ci izen engagemen : Media, knowledge, and pa icipa ion in US
House elec ions. The Jou nal o Poli ics 77(2), 447–462.
Hayes D., and Lawless JL (2017) The decline o local news and i s e ec s: New e idence om longi udinal da a. The Jou nal
o Poli ics 80(1), 332–6.
Hayes D and Lawless JL (2021) News Hole: The Demise o Local Jou nalism and Poli ical Engagemen . Communica ion,
Socie y and Poli ics Camb idge: Camb idge Uni e si y P ess.
Hobol SB, Leepe TJ and Tilley J (2021) Di ided by he o e: A ec i e pola iza ion in he wake o he B exi e e endum.
B i ish Jou nal o Poli ical Science 51(4), 1476–93.
Hol kamp L (2017) Fo men Kommunale Demok a ie. Be n, Swi ze land: Pe e Lang.
Hopkins, DJ (2018) The Inc easingly Uni ed S a es –How and Why Ame ican Poli ical Beha io Na ionalized. Chicago, IL:
Uni e si y o Chicago P ess.
Iyenga S, Lelkes Y, Le endusky M, Malho a N and Wes wood SJ (2018) The o igins and consequences o a ec i e pola -
iza ion in he Uni ed S a es. Annual Re iew o Poli ical Science 22(1), 129–146.
Jones DR (2001) Pa y pola iza ion and legisla i e g idlock. Poli ical Resea ch Qua e ly 54(1), 125–141.
Lelkes Y, Sood G and Iyenga S (2017) The hos ile audience: The e ec o access o b oadband in e ne on pa isan a ec .
Ame ican Jou nal o Poli ical Science 61(1), 5–20.
1274 Fabio Ellge e al.
Lenz GS (2013) Follow he Leade ?: How Vo e s Respond o Poli icians’Policies and Pe o mance. Chicago, IL: Uni e si y o
Chicago P ess.
Leupold A, Klinge U and Ja en O (2018) Imagining he Ci y. Jou nalism S udies 19(7), 960–982.
Le endusky M (2013a) How Pa isan Media Pola ize Ame ica. Chicago, IL: Uni e si y o Chicago P ess.
Le endusky MS (2013b) Why do pa isan media pola ize iewe s? Ame ican Jou nal o Poli ical Science 57(3), 611–623.
Mason L (2018) Unci il Ag eemen : How Poli ics Became Ou Iden i y. Chicago, IL: Uni e si y o Chicago P ess.
McConnell C e al. (2018) The economic consequences o pa isanship in a pola ized e a. Ame ican Jou nal o Poli ical
Science 62(1), 5–18.
Me ag J (2016) Poli ical communica ion and opinion o ma ion in Ge many: A compa a i e s udy o local and na ional
issues. Eu opean Jou nal o Communica ion 31(2), 185–203.
Mi chell A, Simmons K, Ma sa KE and Sil e L (2018) Publics globally wan unbiased news co e age, bu a e di ided on
whe he hei news media deli e s. Pew Resea ch Cen e 11.h ps://www.pew esea ch.o g/jou nalism/2018/01/11/publics-
globally-wan -unbiased-news-co e age-bu -a e-di ided-on-whe he - hei -news-media-deli e /
Moskowi z DJ (2021) Local news, in o ma ion, and he na ionaliza ion o US elec ions. Ame ican Poli ical Science Re iew
115(1), 114–129.
Padge J, Dunaway JL and Da JP (2019) As seen on TV? How ga ekeeping makes he U.S. house seem mo e ex eme.
Jou nal o Communica ion 69(6), 696–719.
Pa k S, Fishe C and Lee JY (2022) Regional news audiences’ alue pe cep ion o local news. Jou nalism 23(8), 1663–1681.
Pogun ke T (1994) Pa ies in a legalis ic cul u e: The case o Ge many. In Ka z R and Mai P (eds), 185–215. How Pa ies
O ganize. Change and Adap ion in Pa y O ganiza ions in Wes e n Democ acies. London. SAGE Publica ions.
P io M (2007) Pos -B oadcas Democ acy: How Media Choice Inc eases Inequali y in Poli ical In ol emen and Pola izes
Elec ions. Camb idge: Camb idge Uni e si y P ess.
P io M (2013) Media and poli ical pola iza ion. Annual Re iew o Poli ical Science 16(1), 101–127.
Rademache C (2018) Local Pa y Sys ems in Ge many. Lokale Pa eiensys eme in Deu schland. GESIS Da a A chi e,
Cologne. ZA6914 Da a ile Ve sion 1.0.0. h ps://doi.o g/10.4232/1.12975
Reiljan A (2020) ‘Fea and loa hing ac oss pa y lines’(Also) in Eu ope: A ec i e pola isa ion in Eu opean pa y sys ems.
Eu opean Jou nal o Poli ical Resea ch 59(2), 376–396.
Rogowski JC and Su he land JL (2016) How ideology uels a ec i e pola iza ion. Poli ical Beha io 38(2), 485–508.
S amm W and Tewes U (1979) STAMM Lei aden Du ch P esse Und We bung. Essen: S amm-Ve lag.
S a is a (2020) Paid ci cula ion o daily and Sunday newspape s in Ge many in selec ed yea s om 1975 o 2019. A ailable
om h ps://www.s a is a.com/s a is ics/410995/daily-and-sunday-pape s-ci cula ion-ge many/
S ockinge B (2019) B oadband in e ne a ailabili y and es ablishmen s’employmen g ow h in Ge many: E idence om
ins umen al a iables es ima ions. Jou nal o Labou Ma ke Resea ch 53(1), 7.
Tabe CS and Lodge M (2006) Mo i a ed skep icism in he e alua ion o poli ical belie s. Ame ican Jou nal o Poli ical
Science 50(3), 755–769.
Tabe CS, Cann D and Kucso a S (2009) The mo i a ed p ocessing o poli ical a gumen s. Poli ical Beha io 31(2),
137–155.
T ussle M (2020) The e ec s o high-in o ma ion en i onmen s on legisla i e beha io in he US House o Rep esen a i es.
Legisla i e S udies Qua e ly 47(1), 95–126.
T ussle M (2021) Ge in o ma ion o ge in o ma ion: The e ec s o high-in o ma ion en i onmen s on legisla i e elec ions.
B i ish Jou nal o Poli ical Science 51(4), 1529–1549.
Volkens A e al. (2020) The Mani es o Da a Collec ion. Mani es o P ojec (MRG/CMP/MARPOR). Ve sion 2020a.
Wissenscha zen um Be lin. A ailable om h ps://mani es op ojec .wzb.eu/
Wagne M (2021) A ec i e pola iza ion in mul ipa y sys ems. Elec o al S udies 69, 102199.
Wagne M and G uszczynski M (2018) Who ge s co e ed? Ideological ex emi y and news co e age o membe s o he U.S.
Cong ess, 1993 o 2013. Jou nalism & Mass Communica ion Qua e ly 95(3), 670–690.
Webs e SW and Ab amowi z AI (2017) The ideological ounda ions o a ec i e pola iza ion in he U.S. elec o a e.
Ame ican Poli ics Resea ch 45(4), 621–647.
Wehling H-G and Kos A (2010) Kommunalpoli ik in de Bundes epublik Deu schland —eine Ein üh ung. In Kos A and
Wehling HG (eds), Kommunalpoli ik in den deu schen Lände n. Wiesbaden: VS Ve lag ü Sozialwissenscha en. h ps://
doi.o g/10.1007/978-3-322-99404-2_1
Wellb ock CM (2011) Die jou nalis ische Quali ä deu sche Tageszei ungen –Ein Ranking. MedienWi scha 8(2), 22–31.
Wissenscha liche Diens e des Deu schen Bundes ages (2007) In e ne nu zung: Globale En wicklung und Da s ellung
empi ische Da en. Deu sche Bundes ag. h ps://www.bundes ag.de/ esou ce/blob/414768/797 a31a4e17ca56be 9ebb6 b0306 c/
wd-10-070-07-pd -da a.pd
Zalle JR (1992) The Na u e and O igins o Mass Opinion. Camb idge: Camb idge Uni e si y P ess.
Ci e his a icle: Ellge F, Hilbig H, Riaz S, Tillmann P (2024). Local Newspape Decline and Poli ical Pola iza ion –E idence
om a Mul i-Pa y Se ing. B i ish Jou nal o Poli ical Science 54, 1256–1275. h ps://doi.o g/10.1017/S0007123424000243
B i ish Jou nal o Poli ical Science 1275
