Public administrators as politicians in office

Hessami, Zohal; Häcke , Timo; Thomas, Maximilian
A icle — Published Ve sion
Public adminis a o s as poli icians in o ice
In e na ional Tax and Public Finance
Sugges ed Ci a ion: Hessami, Zohal; Häcke , Timo; Thomas, Maximilian (2025) : Public adminis a o s
as poli icians in o ice, In e na ional Tax and Public Finance, ISSN 1573-6970, Sp inge US, New
Yo k, Vol. 32, Iss. 6, pp. 1960-1987,
h ps://doi.o g/10.1007/s10797-025-09882-z
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/333227
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-nc-nd/4.0/
Vol:.(1234567890)
In e na ional Tax and Public Finance (2025) 32:1960–1987
h ps://doi.o g/10.1007/s10797-025-09882-z
Public adminis a o s aspoli icians ino ice
ZohalHessami1· TimoHäcke 2· MaximilianThomas3
Accep ed: 17 Janua y 2025 / Published online: 30 Janua y 2025
© The Au ho (s) 2025, co ec ed publica ion 2025
Abs ac
We analyze whe he mayo s’ p io occupa ion in he local public adminis a ion ma -
e s o hei pe o mance. In heo y, mayo s’ p o essional backg ound may shape
hei compe ence in bu eauc a ic asks. We use he example o g an eceip s o
isible in es men p ojec s o which mayo s mus submi an ex ensi e applica ion
o he s a e go e nmen . Ou da ase includes 1,933 mayo elec ions (1993-2020) in
he Ge man s a e o Hesse o which we apply a sha p RD design o close mixed-
backg ound aces. Mayo s’ backg ound on a e age has no e ec on g an eceip s.
Ye , public adminis a o mayo s do a ac mo e g an s han ou side mayo s when
hey a e ideologically aligned wi h he council, aising he mo i a ion o apply o
g an s in he i s place. We conclude ha he compe ence o public adminis a o
mayo s only ma e s when hey a e mo i a ed o use i , i.e. his is an example whe e
incen i es a e necessa y o he e ec s o poli ical selec ion o ma e ialize.
Keywo ds Public adminis a ion· P o essional backg ound· Adminis a i e skills·
Poli ical selec ion· Mayo s· In es men g an s
JEL Classi ica ion D73· D78· H71· H77· H83
1 In oduc ion
Wha does i ake o be a success ul mayo ? A g owing li e a u e p o ides insigh s
on he e ec o a ious poli ician cha ac e is ics such as age, gende o educa ion
on policy ou comes (Alesina e al., 2018; Baska an & Hessami, 2025; Besley e al.,
2011). Ye , we know li le abou speci ic compe ences, skills o expe iences ha
may help mayo s do a good job in o ice. Many mayo du ies in ol e bu eauc a ic
* Maximilian Thomas
maximilian. homas@ ub.de
1 Ruh Uni e si y Bochum, CESi o Munich, IZA Bonn, Bochum, Ge many
2 LSE, ci i y Managemen Consul an s, Hambu g, Ge many
3 Facul y o Social Science, Chai o Social Policy andPublic Economics, Ruh Uni e si y
Bochum, Uni e si ä ss . 150, 44801Bochum, Ge many
1961
Public adminis a o s aspoli icians ino ice
asks and equi e ons o pape wo k and coo dina ion wi hin he local public admin-
is a ion and i s many subdi isions. Expe ience in he local public adminis a ion
may hus indeed be help ul.1
In some coun ies, a backg ound in public adminis a ion is a equi emen o
speci ic public o ices. In he US, he e exis wo main o ms o go e nmen : coun-
cil-manage and mayo -council go e nmen s. In bo h sys ems, a ci y manage will
be appoin ed o whom i is cus oma y o hold a Mas e o Public Adminis a ion
(MPA) deg ee,2 In con as , in mos coun ies, e.g. in Ge many, a deg ee o p io
wo k expe ience in his ield is no equi ed o local public o ices. Ye , desc ip-
i ely, many mayo candida es (and elec ed mayo s) happen o ha e a public admin-
is a ion backg ound.3 The ques ion a ises whe he expe ise in local public adminis-
a ion is ele an o policy choices and mayo s’ pe o mance in o ice.
We analyze his ques ion in he con ex o 421 municipali ies in he Ge man s a e
o Hesse whe e mayo s a e elec ed e e y six yea s. Hessian mayo s a e di ec ly
elec ed bu ha e ela i ely ew powe s/compe ences (i.e. Hesse applies a mix u e
be ween he US council-manage and mayo -council sys em). In Hesse (and o he
pa s o Ge many), al hough he e is no educa ional equi emen many mayo s ha e
a backg ound in public adminis a ion – indica ed by a Diplom-Ve wal ungswi /in
deg ee ob ained om a Uni e si y o Applied Sciences – and a e o en employed in
he local public adminis a ion be o e being elec ed in o o ice.
In a eg ession discon inui y se ing wi h close mayo elec ions be ween a candi-
da e wi h a backg ound in local public adminis a ion and a candida e wi h a non-
adminis a i e backg ound, we analyze he e ec o mayo s’ backg ound on munici-
pal ou comes. I hese wo ypes o candida es win he elec ion, we e e o hem as
public adminis a o mayo s o ou side mayo s.4 We ocus on an ou come a iable
whe e mayo s in Hesse ha e disc e ion and bea he sole esponsibili y. Mos local
policies in Hesse a e se by local councils. Howe e , in es men g an s om he
s a e le el a e acqui ed h ough a edious applica ion p ocedu e o which mayo s
ha e o pu oge he an ex ensi e applica ion package (Hessami, 2018).
1 A con a y a gumen could be made by saying e.g. ha an ou side , i.e. a mayo ha ge s elec ed wi h-
ou his expe ience, may exe ex a e o in hese asks o signal compe ence (Rogo & Sibe , 1988;
Rogo , 1990).
2 In he council-manage sys em o e s elec he council, which is he p ima y legisla i e body o he
ci y. The council in u n appoin s a ci y manage who ca ies ou he di ec i es o he council. In some
owns, he council may addi ionally ha e an elec ed mayo . Howe e , in he council-manage o m o
go e nmen , he mayo has ew o no powe s abo e and beyond hose o he o he councilo s. In he
mayo -council sys em, he e is also an elec ed council and an elec ed mayo , bu he mayo ypically is
equipped wi h mo e powe s. A mayo -council go e nmen will also appoin a ci y manage . Rega ding
e idence on e ec s o hese sys ems on policymaking, see Coa e and Knigh (2011), Deno and Mehay
(1987), Enikolopo (2014), Vlaicu and Whalley (2016).
3 In ou sample, his is he case o 50.7% o mayo candida es and 63.9% o elec ed mayo s (see
Table1).
4 36.1% o Hessian mayo s a e ou side s wi h backg ounds in business/ inance (10.9%), law (7.1%),
eaching/social wo k (5.2%), mili a y/police (5.0%), indus y (3.6%), sel -employmen (2.8%), as shown
in Table1.
1962
Z.Hessami e al.
Ou se ing is ideal o s udy he e ec o mayo s’ occupa ional backg ound on
hei pe o mance. Fi s , Hessian mayo s ha e ela i ely ew powe s and compe-
ences (compa ed o e.g. Sou he n Ge man mayo s). Thus, we can ocus on a spe-
ci ic ou come (in es men g an s) a he han analyzing a la ge se o ou comes. Sec-
ond, as Hessami (2018) inds, he swi ch om mayo appoin men s by local councils
o di ec mayo elec ions be ween 1993 and 1998 has boos ed he di e si y o may-
o s in e ms o hei educa ional/p o essional backg ound.5 This gi es ise o a su -
icien ly la ge numbe o mayo candida es in ou sample ha do no ha e a back-
g ound in public adminis a ion. Thi d, compa ed o o he Ge man s a es, mayo
elec ions in Hesse do no ake place a a uni o m da e.6 Thus, i we do ind any
signi ican e ec s, hese a e unlikely o be d i en by empo al pa e ns in in es men
g an alloca ion.
Ou pape is p ima ily ela ed o RD s udies on he e ec o local poli icians‘ p o-
essional backg ound on iscal ou comes. Beach and Jones (2016) ind ha "manage-
ial" skills o ci y council candida es (p oxied by business expe ience) in Cali o -
nia do no ha e a causal e ec on a ious local ou comes (expendi u es, e enues,
unemploymen , adminis a ion sala ies, o e sha es). Simila ly, Kuliomina (2021)
inds no e ec o Czech local councilo s’ en ep eneu ial backg ound on municipal
budge alloca ion, de ici s o deb . Ye , Szakonyi (2021) shows ha Russian mayo s
wi h a business backg ound p omo e business- iendly policies by p io i izing eco-
nomic in as uc u e spending (e.g. building/main aining oads). Likewise, Ki kland
(2021) inds ha US businesspe son mayo s alloca e mo e esou ces o in as uc-
u e p ojec s. O e all, he esul s a e mixed and appea o depend on he speci ic
con ex .7
In con as o ou pape , he abo e li e a u e ocuses on he business backg ound
o local poli icians. Since mos asks o mayo s a e, howe e , bu eauc a ic, i is na u-
al o ask whe he mayo s’ public adminis a ion backg ound a ec s municipal ou -
comes. One excep ion in he li e a u e is Bo dignon e al. (2020) which applies a
di e ence-in-di e ences design o an I alian e o m in he 1990s ha g an ed some
ci ies highe local iscal au onomy. The au ho s show ha mo e a luen egions
5 Be o e 1993, he mayo was appoin ed by he council ia a wo- hi ds majo i y o e. The council was
able o emo e he appoin ed mayo om o ice. Since he e o m, he mayo is elec ed by he people.
6 Sou ces o his a ia ion a e ini ial asynch onous da es o mayo appoin men s in 1946/1947 and
1970s municipal me ge s. In line wi h he Hessian municipal code (§42, HGO), he appoin men /elec-
ion o a mayo mus ake place h ee o six mon hs be o e he incumben s eps down. A hi d eason is
ha some mayo s do no se e a ull e m (dea h, illness, e i emen ). I a e m ends p ema u ely, he
appoin men /elec ion mus ake place wi hin ou mon hs. Hence, u he a ia ion is due o he au on-
omy ha municipali ies ha e in choosing he da e.
7 Rela ed li e a u e s udies he e ec o mayo s’ educa ion le el (no ield) on policy ou comes. A el-
laneda (2009) inds ha mo e educa ed mayo s collec mo e p ope y ax e enues and inc ease social
spending in Columbia. Alpino e al. (2022) show ha college-educa ed mayo s in I aly a e mo e likely
o implemen p og essi e ax e o ms. In Ge many, F eie and Thomasius (2015) ind no e ec o may-
o s’ educa ion on municipal inances. Ea lie li e a u e a he c oss-coun y le el shows ha he le el
(Besley e al. (2011)) and ield o educa ion (D ehe e al., 2009) o poli ical leade s a ec s economic
de elopmen . Fu he , a la ge li e a u e s udies poli ician gende (see Baska an and Hessami (2025) o a
ecen s udy and Hessami and daFonseca (2020) o a li e a u e e iew) and mo e ecen ly poli ician age
(Alesina e al., 2018; Baska an e al., 2024).
1963
Public adminis a o s aspoli icians ino ice
a ac adminis a i e expe s, while poo e egions elec mayo s wi h s ong poli i-
cal skills. Thus, adminis a i e skills a e likely ele an o becoming a mayo in ce -
ain con ex s. Do adminis a i e skills/expe ience also ma e o being a success ul
mayo ?
A second ela ed s and o li e a u e s udies he ole o poli icians e sus bu eau-
c a s in policymaking. Bu eauc a s a e guided by ca ee conce ns, i.e. hey s i e
o mo e up he ca ee ladde wi hin he local adminis a ion, o en holding posi-
ions whe e hey ha e high job secu i y and enjoy inde e mina e appoin men s (see
e.g. Vlaicu and Whalley (2016)). The ques ion o how o alloca e di e en ypes o
asks ac oss bu eauc a s and poli icians has been discussed in p e ious li e a u e and
may depend on a ailable in o ma ion and cha ac e is ics o hese asks (Alesina &
Tabellini, 2007). In ou con ex , a bu eauc a ic ask wi h ela i ely isible (posi i e)
wel a e e ec s is assigned o an elec ed mayo . In some sense, a public adminis a-
o mayo was a bu eauc a be o e and hen became a poli ician, whe eas an ou side
mayo is a pu e poli ician in line wi h he de ini ions abo e. We, he e o e, a gue
ha he main di e ence be ween public adminis a o and ou side mayo s is hei
compe ence in bu eauc a ic asks, which may lead o di e en le els o success in
acqui ing in es men g an s.
We make ou subs an i e con ibu ions o he li e a u e. Fi s , ou unique da a-
se – combining o icial municipal elec ion and iscal da a wi h hand-collec ed
indi idual in o ma ion on op- wo mayo candida es in Hesse – and ou RD design
allow o a c edible iden i ica ion o he causal e ec o mayo s’ occupa ional back-
g ound. Second, we employ a new ou come a iable o analyze iscal policy choices.
Whe eas mos s udies concen a e on spending o axes, we ocus on in es men
g an eceip s. Mayo s in Hesse a e in cha ge o in es men g an applica ions and
hese g an s inance p ojec s ha a e isible o o e s. Thi d, ou hand-collec ed
da ase allows us o de e mine mayo s’ backg ound wi h e.g. he e ia y deg ee Dip-
lom-Ve wal ungswi /in speci ically dedica ed o public adminis a ion. Ou pape is
he i s o e alua e he ele ance o mayo s‘ backg ound in public adminis a ion o
municipal ou comes. Fou h, we also in es iga e how he e ec on in es men g an
eceip s a ies wi h he poli ical en i onmen (e.g. uni ied/di ided local go e nmen ,
poli ical alignmen wi h s a e go e nmen ). Since some p e ious RD s udies do no
ind signi ican e ec s o mayo s’ cha ac e is ics, he speci ic poli ical se ing and
esul ing incen i es likely ma e .
We ind ha mayo s’ public adminis a ion backg ound has no causal e ec on
in es men g an s eceip s. A numbe o obus ness es s con i m his null esul .
Fi s , we show ha he esul s a e no sensi i e o he scaling o he ou come. Sec-
ond, he inclusion o con ols does no a ec he esul s, con i ming he success ul
andomiza ion in ou RD design. Thi d, beyond he da a-d i en op imal bandwid hs
ha we use in ou main es ima ions, we also ob ain a null esul using a ious band-
wid hs be ween 5 and 15%. Finally, we conduc powe calcula ions indica ing ha
ou s udy is su icien ly powe ed o de ec mode a e o la ge e ec sizes.
In wo ex ensions, we analyze he e ogenei y in he ea men e ec acco ding o
he ideological composi ion o councils and alignmen o he mayo ‘s pa y wi h he
s a e go e nmen . Fi s , ou esul s sugges ha poli ical alignmen wi h he s a e
go e nmen plays no ole, con i ming ha he s a e go e nmen and i s minis ies

1964
Z.Hessami e al.
ha e li le disc e ion in ejec ing/app o ing applica ions o in es men g an s based
on poli ical p e e ences. As desc ibed in Hessami (2018), applica ions a e ypically
success ul i he e is money le in a speci ic yea and i he applica ion is comple e
and documen s compliance wi h he unding guidelines.
Second, we ind ha public adminis a o mayo s ecei e mo e g an s when hey
belong o he ideological camp ha has he council majo i y (i.e. uni ied local go -
e nmen s). Public adminis a o mayo s a ac mo e han wice as many in es men
g an s han ou side mayo s when hey ha e he suppo o he council majo i y. This
co esponds o 38 Eu os pe capi a mo e pe yea o a median-sized Hessian munic-
ipali y wi h abou 10,000 inhabi an s and can he e o e be conside ed as a subs an-
ial e ec . While his e ec is no based on explici andomiza ion o ho izon al
alignmen and should hus be in e p e ed wi h cau ion, we can show ha ho izon al
alignmen is no sys ema ically co ela ed wi h a ious municipali y cha ac e is ics.
We conclude ha he compe ence o public adminis a o mayo s appea s o only
make a di e ence o hei pe o mance when hey a e su icien ly (poli ically) mo i-
a ed o use i . As in es men ans e s a e highly isible o o e s and a he same
ime i is di icul o o e s o co ec ly a ibu e ou comes o poli ical bodies, one
in e p e a ion is ha public adminis a o mayo s ha e a g ea e incen i e o a ac
addi ional unds i hey a e suppo ed by he council. Any elec o al bene i in he
nex elec ion will acc ue o bo h he mayo and will spill o e o his/he poli ically
aligned council.8 Ou side mayo s, on he o he hand, ha e less compe ence/expe i-
ence in dealing wi h bu eauc a ic applica ion p ocedu es and hus, e en i he coun-
cil is aligned, i does no ma e o hei e o /success in a ac ing g an s.
To es o mechanisms, we un addi ional es ima ions. Gi en di e se his o ical
ajec o ies, mayo elec ions may ake place a any poin in ime h oughou council
e ms. Council e ms, on he o he hand, a e synch onized ac oss all Hessian munic-
ipali ies. This allows us o es whe he mayo s a e mo i a ed o a ac addi ional
unds o ensu e he eelec ion o ideologically aligned councils.9,10 Ou esul s show
8 An addi ional explana ion could be ha public adminis a o mayo s a e mo e mo i a ed o apply o
in es men g an s when he council is poli ically aligned because i is easie o ag ee on p ojec s o in e -
es and because hese g an s a e ma ching g an s and o e e y app o ed g an he council needs o p o-
ide addi ional inancing. Gi en ou da a, we a e un o una ely no able o es his explana ion.
9 We also es o poli ical cycles in g an eceip s ac oss mayo e ms. Fo mayo s poli ically aligned
wi h councils, we do no obse e highe g an eceip s be o e mayo elec ions (see TableA.10 in he
online appendix). Thus, mayo s in ou sample a e no so conce ned abou hei own eelec ion, which
is no su p ising gi en ha he eelec ion a e o econ es ing incumben mayo s is 87.3% and 78.9%
o public adminis a o and ou side mayo s, espec i ely. In con as , Hessami (2018) inds ha elec ed
mayo s (compa ed o appoin ed mayo s) in Hesse a ac mo e in es men g an s when a mayo elec ion
is imminen . The main di e ence is ha she applies a DiD app oach including all mayo elec ions and
ocuses on he di e ence be ween he wo ypes o selec ion o mayo s, while we ely on a much smalle
RD sample and compa e public adminis a o wi h ou side mayo s.
10 A e e ee poin ed ou ha he highe elec o al success o public adminis a o mayo s may posi i ely
co ela e wi h council alignmen , c ea ing endogenei y. Fo his o be he case, public adminis a o may-
o s would need o be mo e o en a ilia ed wi h one o he majo pa ies likely o hold council majo i-
ies. In ou comple e da ase , public adminis a o mayo s (ou side mayo s) a e a ilia ed wi h one o he
majo pa ies in 62.4% (61.1%) o he cases (di e ence no s a is ically signi ican , p- alue 0.57), while
hey un as independen candida es in 33.9% (31.5%) o he elec ions. The emaining sha e (<5%) a e
1965
Public adminis a o s aspoli icians ino ice
ha indeed in he subsample o uni ied go e nmen s public adminis a o mayo s
a ac mo e g an s when a council elec ion is imminen . In a second se o es ima-
ions, we also show ha hese g an eceip s a e posi i ely co ela ed wi h he ee-
lec ion likelihood o exis ing council majo i ies.
Ou indings ha e impo an implica ions o ou unde s anding o he ole o
poli icians’ cha ac e is ics o hei mo i a ion and pe o mance in o ice. While
on a e age i appea s ha a backg ound in local adminis a ion is i ele an , he e
a e ci cums ances ha gi e ise o di e ences in he ex en o which mayo s make
an e o and success ully a ac in es men g an s. E en hough we only look a
one speci ic ou come, mos o he du ies o mayo s in Hesse (o elsewhe e) in ol e
bu eauc a ic asks. Hence, p io expe ience wi h local adminis a ion may indeed
be help ul o doing a good job in o ice. Since desc ip i ely he sha e o public
adminis a o candida es and mayo s is al eady high (69.3% o elec ed mayo s ha e
a public adminis a ion backg ound), i does, howe e , no ollow ha a compulso y
equi emen o a public adminis a ion backg ound is needed. Vo e s appea o be
awa e o his ad an age and seem o ake his in o accoun in hei o ing decision.
As a mo e gene al con ibu ion o he li e a u e, ou esul s highligh ha compe-
ence and expe ience in bu eauc a ic asks (o b oadly speaking poli ical selec ion,
i.e. poli icians’ cha ac e is ics) only ma e o poli icians’ pe o mance and munici-
pal ou comes when poli icians a e su icien ly (poli ically) mo i a ed o use hem.
Thus, ou esul s can be seen as an example whe e poli ical selec ion and (indi ec )
elec o al incen i es in e ac , while he exis ing li e a u e ypically s udies hese con-
cep s in isola ion.
Finally, we illus a e ha poli icians’ incen i es may go beyond hei own eelec-
ion mo i e. In a se ing whe e some deg ee o collabo a ion be ween wo (o mo e)
poli ical bodies is help ul o be success ul, poli icians (who in any case enjoy a high
incumbency bonus) may be mo i a ed o c ea e cycles in hei pe o mance ha c e-
a e posi i e spillo e s o ideologically aligned poli ical bodies. While p e ious li -
e a u e a gues, ha hese ypes o spillo e s may exis wi hin one o ice and wi hin
one pa y, we show ha hey also exis ac oss o ices and po en ially ac oss pa ies
(bu wi hin an ideological camp in he le / igh sense).11
11 Fo ins ance, Fowle and Hall (2014) and Lopesda Fonseca (2017) dis inguish be ween pe sonal and
pa isan incumbency ad an ages in he con ex o US s a e legisla u e and Po uguese mayo elec ions.
Bo h ind ha pa isan incumbency ad an ages (i.e. spillo e s o candida es o he han he incumben )
a e no empi ically ele an . No e, howe e , ha in hese se ings incumbency ad an ages a e much
smalle han in ou con ex .
Foo no e 10 (con inued)
mino pa ies no consis en ly aligning wi h ei he o he majo pa ies. Thus, he p oposed mechanism
likely does no h ea en ou iden i ica ion.
1966
Z.Hessami e al.
2 Backg ound
2.1 Municipal inances
Loca ed in he cen e o Ge many, Hesse ep esen s he Ge man a e age in e ms o
size and p ospe i y. In 2019, he popula ion size was 6.3 million equalling 7.5% o
he Ge man popula ion. Hesse consis s o 421 egula municipali ies, ou o which
416 a e o ganized in o 21 coun ies, and i e la ge ci ies which hold a special s a us.12
Local adminis a ions subs an ially impac ci izens’ daily li es. Municipali ies
p o ide a wide ange o manda o y and olun a y public goods. Compulso y public
goods include municipal childca e, p ima y schools, and ci il p o ec ion. Volun a y
p o ision includes swimming pools and spo s enues, cul u al and his o ical si es,
hea e s o public lib a ies. The p o ision o public goods a he municipal le el is
inanced h ough a combina ion o municipal axes, use ees, and ans e s om
highe le els o go e nmen .
S a e-le el ans e s make up a la ge sha e o municipal e enues. Ye , munici-
pali ies a e also g an ed conside able iscal au onomy o gene a e own ax e enues
by se ing local business and local p ope y ax a es o in oducing new axes (e.g.
seconda y home ax, ou is ax, dog ax). In 2019, Hessian municipali ies in o al
had 52.2 billion Eu os a hei disposal.
2.2 S a e‑le el in es men g an s
The Hessian s a e go e nmen p o ides municipali ies wi h se e al ypes o ule-
based and disc e iona y g an s. In his pape , we ocus on in es men ans e s
(Zuweisungen und Zuschüsse ü In es i ionen und In es i ions ö de ungsmassnah-
men) in ended o speci ic p ojec s, such as he cons uc ion/mode niza ion o local
in as uc u e (e.g., hospi als, childca e acili ies, municipal elde ly ca e acili ies,
public anspo a ion, oad cons uc ion, municipal d inking wa e sys ems, lib a -
ies, museums, and spo acili ies). In 2019, he Hessian s a e go e nmen alloca ed
301.4 million Eu os in o al as in es men g an s.
In es men g an s a e alloca ed annually ia a o mal applica ion p ocedu e
(An ags- und Bewilligungs e ah en). In he second hal o he p e ious yea , he
s a e go e nmen publishes calls o p oposals ia he s a e go e nmen gaze e
(S aa sanzeige ), de ailing p og am objec i es, eligible p ojec s, and equi ed docu-
men a ion. To ecei e unding, mayo s mus submi comple e applica ions. G an s
a e awa ded i eligibili y equi emen s a e me and unds a e a ailable, wi h deci-
sions made by he esponsible minis y wi hin weeks. Though ea lie applica ions
a e mo e likely o be success ul, un unded bu eligible applica ions can be ca ied
o wa d o he nex yea . Municipali ies may ecei e g an s o se e al p ojec s pe
yea .13
12 These i e ci ies (Da ms ad , F ank u , Kassel, O enbach, Wiesbaden) ul ill bo h municipal and
coun y esponsibili ies a e he e o e excluded om ou analysis.
13 Fo mo e de ails on he g an applica ion p ocedu e, please e e o Hessami (2018).
1967
Public adminis a o s aspoli icians ino ice
Mayo s a e esponsible o he p epa a ion and submission o g an applica ions
and can inc ease he likelihood ha hei municipali y ecei es an in es men g an
(Hessami, 2018). Fi s , mayo s can s ay in o med abou new unding oppo uni-
ies by egula ly checking he calls o g an s. Second, p epa ing and submi ing an
applica ion (o se e al ones) in he i s place is necessa y o ob aining any g an .
Thi d, mayo s can impac he likelihood o g an eceip by ensu ing ha he appli-
ca ion is comple e and mee s all unding guidelines.14
While in es men g an s on a e age only accoun o 19.4% o o al ans e s,
indi idual paymen s o municipali ies can be subs an ial. Fo example, in 2019, he
highes amoun g an ed o a municipali y was 435 Eu os pe capi a. Fo a munici-
pali y wi h 10,000 inhabi an s, his amoun s o 4.35 million Eu os. Despi e being
p ojec -speci ic, hese g an s clea ly enhance he inancial possibili ies o a ecei ing
municipali y.
2.3 Mayo s andlocal poli ics
The e a e wo poli ical ins i u ions a he local le el: he mayo and he council. Vo -
e s elec a municipal council a a uni o m s a e-wide da e e e y i e yea s in Ma ch.
Council size depends on municipal popula ion size and a ies which be ween 15 and
105 sea s. The municipal code o Hesse (Hessische Gemeindeo dnung) s ipula es
ha he local council moni o s he local adminis a ion and decides au onomously
on all policy a eas assigned by he ede al and s a e cons i u ions o he municipal
ie . Council sea s a e alloca ed o pa ies ia a p opo ional ule.
The pa y landscape in Hessian local council and mayo elec ions is domina ed
by he cen e- igh Ch is ian Democ a ic Union (CDU) and he cen e-le Social
Democ a s (SPD). Se e al smalle pa ies also compe e in local council elec ions,
including he G eens (en i onmen al, socially libe al), he FDP ( iscally conse a-
i e, socially libe al), and he Le Pa y (socialis s). Also, se e al a - igh /-le pa -
ies, independen candida es and candida es backed by local o e ini ia i es con-
es in local elec ions. Mul iple o e ini ia i es may un in a single municipali y
and end o be success ul, in a e cases secu ing sea majo i ies. Typically, mul iple
pa ies win sea s, equi ing coali ions o secu e majo i ies. E en absen o malized
coali ions, pa ies mus coope a e o each council decisions (Baska an & Hessami,
2018).
Councilo s hold signi ican poli ical powe , while he mayo emains an impo -
an poli ical igu e. Since 1993, mayo s in Hesse a e di ec ly elec ed o a six-yea
e m.15 A s a e-wide elec ion da e does no exis . Ins ead, a mayo elec ion is held
15 A e iewe has poin ed ou ha he ansi ion phase (1993-1997) om appoin men s o di ec elec-
ions o mayo s alls wi hin he obse a ion pe iod. Howe e , ou close-elec ion RD design ensu es ha
we exclusi ely compa e municipali ies go e ned by di ec ly elec ed mayo s. I migh in p inciple be ha
du ing he ansi ion phase di ec ly elec ed adminis a i e mayo s pe cei ed heigh ened compe i ion wi h
neighbo ing municipali ies led by appoin ed mayo s, po en ially inc easing hei eelec ion incen i es
and in luencing hei decision o apply o in es men g an s. To add ess his conce n, we conduc an
14 No e ha he alloca ion o hese g an s does no include a buil -in equaliza ion scheme ac oss munici-
pali ies.
1974
Z.Hessami e al.
densi y o he ma gin o ic o y a he h eshold, sugges ing ha he e is no manipu-
la ion o he ea men s a us.
4.2.2 Balance es s o municipali y cha ac e is ics
A second assump ion is ha he selec ion o candida es close o he cu o is unco -
ela ed wi h municipal cha ac e is ics ha may ma e o in es men g an eceip s.
Thus, municipali ies in which a public adminis a o candida e ba ely wins agains
an ou side candida e should no sys ema ically di e om municipali ies whe e an
ou side candida e ba ely wins agains a public adminis a o candida e. To o mally
es he success o ou local andomiza ion a he municipal le el a ound he cu o ,
we epea ou baseline es ima ions wi h eigh di e en municipali y cha ac e is ics
as dependen a iables (see TableA.3). None o he di e ences be ween municipali-
ies wi h a public adminis a o and an ou side mayo a e signi ican .
4.2.3 Uncon oundedness wi ho he mayo cha ac e is ics
A u he conce n a e con ounding e ec s due o o he pe sonal a ibu es o elec ed
mayo s. Fo ins ance, public adminis a o candida es migh be concen a ed in spe-
ci ic pa ies. Hence, he esul s ha we ob ain may be due o mayo s’ pa y a ilia ion
a he han mayo s’ backg ound. We es whe he in close aces ic o ious public
adminis a o candida es di e om ic o ious ou side candida es in a ious a ib-
u es. Al hough "abili y" canno be measu ed di ec ly, i is a leas pa ly ela ed wi h
obse able cha ac e is ics such as educa ion. In TableA.4, we es o a discon inu-
i y a he h eshold o  he ollowing cha ac e is ics o elec ed mayo s: educa ion,
gende , age, and pa y a ilia ion. We do no ind any signi ican di e ences be ween
public adminis a o and ou side mayo s who we e elec ed in close aces.
Fig. 2 Desc ip i e s a is ics: In es men g an s pe capi a. This igu e illus a es desc ip i e s a is ics o
in es men g an s pe capi a. Sub igu e (a) shows he dis ibu ion o pe -capi a in es men g an s ac oss
municipali ies. Sub igu e (b) shows he de elopmen o a e age in es men g an s du ing he 1993–2020
pe iod o municipali ies go e ned by ou side mayo s (dashed line) and municipali ies go e ned by pub-
lic adminis a o mayo s (solid line)

1975
Public adminis a o s aspoli icians ino ice
5 Resul s
5.1 Baseline
We i s g aphically explo e he ela ionship be ween mayo s’ backg ound and
in es men g an eceip s. Figu e4 plo s in es men g an eceip s a di e en ma -
gins o ic o y. Posi i e alues o he ma gin o ic o y imply ha he public admin-
is a o candida e wins he elec ion, whe eas nega i e alues imply ha he ou side
candida e wins he elec ion. The do s a e binned a e ages o he ou come a iable.
The solid lines plo lexible polynomials on bo h sides o he cu o wi h 95%-con i-
dence in e als shaded in g ey. The e is no discon inui y in in es men g an eceip s
be ween municipali ies wi h and wi hou a public adminis a o mayo .
Nex , we explo e he e ec o public adminis a o mayo s on in es men g an s
wi h eg essions based on Equa ion (2). Table2 collec s RD es ima es o bo h non-
pa ame ic (Model 1 and 2) and pa ame ic (Model 3) speci ica ions. Again, he e is
no signi ican di e ence.28
In he online appendix (see Table A.5), we also epo esul s om OLS es i-
ma ions based on he ull sample. While he simple bi a ia e ela ionship be ween
public adminis a o mayo s and in es men g an eceip s is signi ican ly posi i e,
including con ols (e.g. popula ion size, popula ion densi y, e c.) and municipali y
and yea FE leads o insigni ican esul s. This indica es ha OLS es ima ion leads
o an upwa d bias in he ela ionship. As s a ed in Sec .4, his migh o ins ance
be due o o e s in poo e o mo e u al municipali ies s a egically elec ing public
adminis a o candida es as mayo s as hey a e likely be e able o secu e g an s o
in es men p ojec s. The ac ha he signi ican ela ionship disappea s once popu-
la ion con ols a e included indica es ha hese cha ac e is ics play a key ole.
5.2 Robus ness
5.2.1 Bandwid h choice
We es he sensi i i y o ou esul s o bandwid h selec ion in Figu e A.1 in he
online appendix based on a a ie y o bandwid hs, anging om 5% o 15% ma -
gins o ic o y. All es ima ed coe icien s o he public adminis a o mayo dummy
a e close o ze o and insigni ican .
28 E en hough he es ima es a e a om con en ional signi icance le el, he coe icien s a e sligh ly
posi i e and o mode a e size, anging be ween 11% and 26% addi ional g an s being a ac ed by pub-
lic adminis a o mayo s. In Sec .5.2.4, we conduc u he es s o in es iga e whe he ou analyisis is
unde powe ed and whe he we would be able o de ec small, mode a e and la ge e ec s gi en ou RD
sample.
1976
Z.Hessami e al.
5.2.2 Scaling o ou come a iable
We es whe he he null e ec o in es men g an eceip s is sensi i e o he scal-
ing o he ou come a iable using i e al e na i e scalings (see Table A.7 in he
online appendix). Models (1)–(5) p esen esul s o he ex ensi e ma gin (dummy
= 1 i a municipali y ecei ed non-ze o g an s in a yea ), o al in es men g an s,
pe capi a in es men g an s, o al pe capi a in es men g an s acqui ed du ing he
en i e e m, and he di e ence in pe capi a in es men g an s be ween he las and
i s yea in o ice. None o he coe icien s a e s a is ically signi ican .
5.2.3 Inclusion o con ols
We also e un ou baseline RD model by including municipal con ols (i.e., popula-
ion size, popula ion densi y, emale popula ion sha e and sha e o wo king-age pop-
ula ion) in TableA.8. As expec ed, adding con ols educes s anda d e o s. Model
(2) is ou p e e ed speci ica ion including ixed e ec s and con ols. The coe icien
is close o ze o, con i ming he baseline null esul .29
Fig. 3 RD alidi y: Densi y es s o unning a iable. This igu e p esen s a his og am o he unning
a iable "Ma gin o ic o y" (Sub igu e (a)) and he esul s o local densi y es ima ion acco ding o
McC a y (2008) and Ca aneo e al. (2018) (see Sub igu es (b) and (c)). The solid line in Sub igu es (b)
and (c) ep esen s he densi y o he unning a iable, while he shaded a easindica econ idence in e -
als
29 We also epea ou baseline es ima ion wi h addi ional economic con ols cap u ing he po en-
ial demand o in es men g an s, i.e. local employmen sha e, local business ax e enues pe capi a,
numbe o i ms pe capi a, municipal deb pe capi a, and o al g an s ecei ed. The esul s emain
1977
Public adminis a o s aspoli icians ino ice
Fig. 4 Baseline RD plo . This igu e shows he RD plo o he log o pe capi a in es men g an eceip s
pe municipali y when he mayo in o ice has a public adminis a o s. ou side backg ound. The
o cing a iable is he ma gin o ic o y o he public adminis a o candida e o he mayo ’s o ice in
mixed-backg ound aces (whe e he op- wo candida es ha e a backg ound in public adminis a ion s.
ou side o i ). Obse a ions o he igh o he cu o imply ha a public adminis a o holds he mayo ’s
o ice. Each do is he binned local a e age o in es men g an s aised by he municipali y o he ma gin
o ic o y. The solid lines a e om a local linea smoo h o he unde lying obse a ions. The g ey-shaded
a eas a e he 95% con idence in e als (Colo igu e online)
Table 2 Baseline: Public adminis a o mayo s and in es men g an s
This able epo s es ima es om sha p RD es ima ions ha ela e he backg ound o elec ed mayo s o
in es men g an eceip s. The sample co e s all mayo elec ions in Hesse whe e he op- wo candida es
we e o opposi e backg ounds ("mixed-backg ound elec ions"). The dependen a iable is he log o
in es men g an s pe municipali y, measu ed in cons an (2015) pe -capi a alues. The ea men a i-
able is a dummy which equal o one i he candida e wi h an adminis a i e backg ound won he las
elec ion. We epo local linea (Model 1 and 2) and quad a ic (Model 3) eg essions applying MSERD
op imal bandwid hs p oposed by Calonico e al. (2019). He e oscedas ici y- and clus e - obus s anda d
e o s a e in pa en heses. The uni o clus e ing is he municipali y o he candida e. S a s indica e sig-
ni icance le els a 10% (*), 5% (**) and 1%(***)
Dep. a .: Log(in g an s p.c.) (1) (2) (3)
Public admin mayo 0.110 0.264 0.189
(0.368) (0.328) (0.518)
Bandwid h ype MSE op imal MSE op imal MSE op imal
Bandwid h size 11.712 11.712 11.712
Polynomial Linea Linea Quad a ic
Time FE No Yes No
N 605 605 605
Elec ions 105 105 105
Municipali ies 103 103 103
Mean (SD) 3.17 (1.52) 3.17 (1.52) 3.17 (1.52)
Foo no e 29 (con inued)
unchanged (a ailable upon eques ). Since hese economic a iables a e no consis en ly a ailable ac oss
all pe iods, hei inclusion educes he numbe o obse a ions. Fo his eason, we do no include hese
addi ional con ols in ou baseline es ima ion.
1978
Z.Hessami e al.
5.2.4 S a is ical powe
One conce n could be ha ou analysis lacks s a is ical powe . The e o e, we conduc
ex-pos powe calcula ions o RDDs ollowing Ca aneo e al. (2019b) in TableA.9
in he online appendix. In mo e conse a i e scena ios e.g., i he ue e ec o
mayo s’ backg ound a he h eshold we e 20% o he s anda d de ia ion, we would
be able o de ec i wi h a 9.7% p obabili y. Howe e , we ha e su icien s a is ical
powe o iden i y la ge e ec sizes. Fo ins ance, i he ue e ec we e 80% o he
s anda d de ia ion, we would de ec i wi h a 72% p obabili y. Thus, al hough we
canno ule ou mino e ec s o mayo s’ backg ound on in es men g an s, ou s udy
is adequa ely powe ed o de ec mode a e o la ge e ec s.
6 E ec he e ogenei y
We explo e ex ensions o ou main eg ession o es o e ec he e ogenei y. We
e un ou baseline RD model es ima ing ully in e ac ed speci ica ions, i.e. we in e -
ac "public adminis a ion mayo " and he "ma gin o ic o y" wi h a dummy ha
ela es o he poli ical en i onmen .30
6.1 Ve ical alignmen : s a e go e nmen s
S a e go e nmen s migh alloca e g an s o poli ically aligned municipali ies o
elec o al easons. Baska an and Hessami (2017) ind ha aligned municipali ies
ecei e mo e disc e iona y ans e s in Hesse, speci ically when he deg ee o local
suppo o he s a e go e nmen is s ong. Quinckha d (2023) shows o eigh Ge -
man s a es ha pa y alignmen be ween a mayo and he s a e go e nmen inc eases
ans e s o municipali ies by 20% pe yea .
One could hus hypo hesize ha municipali ies wi h a public adminis a o mayo
who is om he same pa y as he s a e go e nmen ecei e mo e g an s. We explo e
he ole o pa isan alignmen wi h he s a e go e nmen and e un ou baseline RD
model including and in e ac ing a dummy equal o one i he mayo ‘s pa y is in he
s a e go e nmen wi h "public adminis a o mayo " and "ma gin o ic o y". The
esul s a e collec ed in Table3.31
The in e ac ion e ec is insigni ican in all models and his null inding is obus
ac oss di e en bandwid h choices (see Figu eA.2 in he online appendix). Hence,
30 No e ha he poli ical en i onmen is o cou se no andomly assigned in he ollowing analysis. Bo h
s a e go e nmen alignmen and uni ied local go e nmen s a e mo e likely o eme ge in a se ing whe e
he e is less pola iza ion and speci ic pa isan p e e ences a e mo e p onounced wi hin a s a e o munici-
pali y. Ne e heless, i is o in e es o analyze whe he he baseline e ec a ies wi h he poli ical en i-
onmen .
31 O e all, o 177 municipali y-yea obse a ions in ou close mixed-backg ound aces, he mayo ’s
pa y is poli ically aligned wi h he s a e go e nmen (i.e. 33% o he sample). 103 and 74 obse a ions
e e o public adminis a o mayo s and ou side mayo s, espec i ely.
1979
Public adminis a o s aspoli icians ino ice
in es men g an eceip s do no depend on pa y ies be ween he mayo ’s pa y and
he s a e go e nmen ,32
These esul s ep esen co ela ional e idence. In online appendix Table A.13,
we show ha e ically aligned municipali ies a e s a is ically di e en om non-
aligned municipali ies in se e al obse able cha ac e is ics (popula ion size, popula-
ion densi y, deb , and local ax a es). Al hough we explici ly con ol o popula ion
size and densi y in Model(3), we acknowledge ha (un)obse ed con ounde s could
s ill in oduce endogenei y.33
6.2 Ho izon al alignmen : local councils
Mayo s wi h a public adminis a ion backg ound may be mo e mo i a ed o a ac
in es men g an s i hei pa y also holds he majo i y in he council (uni ied go -
e nmen ). On he o he hand, public adminis a o mayo s may no be willing o le -
e age hei adminis a i e skills i aced wi h an opposing council majo i y (di ided
go e nmen ). Ou mixed-backg ound RD sample (
≤
10% ma gin o ic o y) includes
51 municipali y-yea obse a ions (24 and 27 obse a ions o public adminis a o
and ou side mayo s) whe e he mayo ’s pa y is aligned wi h he ideological camp
o he council majo i y (
≈
9.5% o he sample).34
One conce n is ha municipali ies in which he mayo and he council majo i y
align ideologically migh di e in hei cha ac e is ics om municipali ies wi hou
ho izon al alignmen and hese di e ences could a ec g an eceip s. To add ess
his issue, we compa e municipali ies wi h aligned local go e nmen s o all o he
municipali ies in TableA.12: he only signi ican di e ence is in e ms o popula ion
age s uc u e, which we con ol o in ou model. The e a e no di e ences in o he
cha ac e is ics, pa icula ly iscal indica o s. No e ha ou esul s s ill need o be
in e p e ed wi h cau ion and as mos ly co ela ional since we canno explici ly es
o poli ical s a egies and o he unobse ables de e mining ho izon al alignmen .
Nex , we e un ou baseline RD model including and in e ac ing a dummy equal
o one i he mayo belongs o he ideological camp (le / igh ) which holds he
33 In addi ional es ima ions (a ailable upon eques ), we also con ol o he iscal a iables men ioned
he e. The esul s a e unchanged. Since hese a e simul aneously de e mined (endogenous) a iables con-
side ed as “bad con ols”, we p e e no o include hem in he main es ima ions.
34 As mayo and council elec ions a y in hei iming, council majo i ies may change du ing a mayo
e m.
32 One possible eason why, in con as o Baska an and Hessami (2017), we do no ind a signi ican
di e ence in he amoun o ans e s be ween municipali ies acco ding o s a e go e nmen alignmen
is ha we use in es men g an s while Baska an and Hessami (2017) ocus on budge suppo ans e s
(Zuweisungen und Zuschüsse ü lau ende Zwecke). Mayo s do no apply o budge suppo ans e s and
he e o e he pe sonal e o and compe ence o mayo s plays no ole o he alloca ion o hese ans e s.

1980
Z.Hessami e al.
council majo i y wi h "public adminis a ion mayo " and "ma gin o ic o y".35 The
esul s a e collec ed in Table4.
The posi i e and s a is ically signi ican coe icien o he in e ac ion sugges s
ha public adminis a o mayo s a ac mo e han wice as many in es men g an s
han ou side mayo s when hey ha e he suppo o he council majo i y. This co e-
sponds o 38 Eu os pe capi a mo e pe yea o a median-sized Hessian municipal-
i y wi h abou 10,000 inhabi an s. The coe icien o he in e ac ion e ec is obus
Table 3 E ec he e ogenei y I: Aligned s a e go e nmen s
This able epo s es ima es om sha p RD es ima ions ha ela e he backg ound o he elec ed mayo
o in es men g an eceip s. The sample co e s all mayo elec ions in Hesse whe e he op- wo candi-
da es we e o opposi e backg ounds ("mixed-backg ound elec ions"). The dependen a iable is he log
o in es men g an s pe municipali y, measu ed in cons an (2015) pe -capi a alues. The ea men a i-
able is a dummy equal o one i he candida e wi h an adminis a i e backg ound won he las elec ion.
We in e ac he ea men dummy as well as he con ol unc ion wi h a dummy o whe he he mayo ’s
pa y belongs o he s a e go e nmen . O e all, in 177 municipali y-yea obse a ions, he mayo ’s pa y
is poli ically aligned wi h he s a e go e nmen . O hese, 103 (74) obse a ions e e o public admin
(ou side) mayo s espec i ely. We epo local linea (Model 1, 2, and 3) and quad a ic (Model 4) eg es-
sions using MSERD op imal bandwid hs p oposed by Calonico e al. (2019). Municipal con ols include
popula ion size, popula ion densi y, and he wo king-age popula ion sha e. He e oscedas ici y- and clus-
e - obus s anda d e o s a e in pa en heses. The uni o clus e ing is he municipali y o he candida e.
S a s indica e signi icance le els a 10% (*), 5% (**) and 1%(***)
Dep. a .: Log(in g an s p.c.) (1) (2) (3) (4)
Public admin mayo 0.021
−
0.112
−
0.060 0.335
(0.435) (0.298) (0.335) (0.607)
S a e go alignmen 0.095
−
0.135
−
0.159 0.492
(0.601) (0.330) (0.368) (0.801)
Public admin mayo x 0.239 0.293 0.284
−
0.388
S a e go alignmen (0.759) (0.497) (0.532) (1.031)
Bandwid h ype MSE op imal MSE op imal MSE op imal MSE op imal
Bandwid h size 11.712 11.712 11.712 11.712
Polynomial Linea Linea Linea Quad a ic
Time FE No Yes Yes No
Con ols No No Yes No
N 605 605 605 605
Elec ions 105 105 105 105
Municipali ies 103 103 103 103
Mean (SD) 3.17 (1.52) 3.17 (1.52) 3.17 (1.52) 3.17 (1.52)
35 We ca ego ize he SPD and he G eens as le -wing, and he CDU and he FDP as igh -wing. E en
hough smalle pa ies and o e communi ies con ibu e o coali ion building, he ideological posi ions
a he local le el a e o en mo e "p agma ic" and canno be clea ly classi ied in o a pa icula camp. The
dummy a iable is equal o one i he camp o he mayo ’s pa y also has he majo i y (50%+
𝜂
) o sea s
in he council. No e ha by de ini ion independen mayo s can no be aligned wi h a local council majo -
i y.
1981
Public adminis a o s aspoli icians ino ice
o including ime ixed e ec s (Model (2)), and municipal con ols (Model (3)).36 In
addi ion, Figu eA.3 in he online appendix shows ha he posi i e in e ac ion e m
is obus o di e en bandwid hs.37,38
7 Mechanism: Mayo s’ indi ec elec o al incen i es
In his sec ion, we un addi ional es ima ions o analyze why public adminis a-
o mayo s ha ace an ideologically aligned local council a ac mo e in es men
g an s.
7.1 In es men g an cycles o e council e ms
Mayo s in Hesse (and in o he Ge man s a es) a e eelec ed a a e y high a e and
once in o ice hey o en s ay he e un il e i emen . This makes i unlikely ha
mayo s a e p ima ily mo i a ed by hei pe sonal elec o al incen i es o apply o
in es men g an s. Mo eo e , his would no explain why he compe ence o pub-
lic adminis a o mayo s (compa ed o ou side mayo s) only makes a di e ence o
g an eceip s when he council is poli ically aligned. Wha seems mo e likely is ha
mayo s ha e an in e es in helping he ideologically aligned local council s ay in
powe . This does no only e lec hei pe sonal ideological p e e ences as ci izens
o he municipali y bu also ma e s o he di ec ion o u u e local policy choices.
We es whe he he amoun o in es men g an s ha public adminis a o mayo s
a ac when aced wi h an aligned council a ies o e he council e m, i.e. whe he
hey a e highe jus be o e he nex council elec ion. Table5 collec s he esul s o
speci ica ions whe e we include a dummy o one yea and wo yea s be o e he
council elec ion and in e ac his wi h he public adminis a o mayo dummy.39
E en hough he sample is o cou se smalle and we he e o e ha e o be ca e ul
in in e p e ing he esul s, we ind ha he baseline e ec he e ogenei y is d i en
38 A e iewe has poin ed o he nega i e Council alignmen coe icien in Model (3) in Table4. This
esul is no di ec ly ela ed o ou main hypo hesis bu is indeed in e es ing. One po en ial explana ion
could be ha ou side mayo s join ly go e ning wi h an aligned council a e complacen and e en educe
hei e o in aising addi ional unds. Howe e , no e ha he e ec is only signi ican a he 10%-le el in
one o he ou models (Model 3), and coe icien signs a y ac oss speci ica ions. The e o e, any in e -
p e a ion mus be ea ed wi h cau ion.
39 Figu eA.4 in he online appendix illus a es his mo e clea ly in an e en -s udy plo going back up o
ou yea s be o e council elec ions: he e ec indeed only kicks in one yea be o e council elec ions.
36 Fo he quad a ic polynomial (Model (4), he in e ac ion is insigni ican . Howe e , due o he conce ns
exp essed by Gelman and Imbens (2018), we ocus on he linea speci ica ion as ou p e e ed model.
37 Fo he sake o comple eness, we also explo e di ided local go e nmen s (see TableA.6 in he online
appendix). Ou RD sample includes 53 municipali y-yea obse a ions (
≈
9.8% o he sample) wi h a
di ided go e nmen . The esul s a e, as expec ed diame ically opposed o hose o uni ied go e n-
men s: public adminis a o mayo s who ace an opposing council majo i y a e less success ul in secu -
ing in es men g an s. The in e ac ion is obus o he inclusion o ixed e ec s (Model (2)) and con ols
(Model (3)), and he use o quad a ic polynomials (Model (4)). Figu eA.3 illus a es he obus ness o
di e en bandwid h choices.
1982
Z.Hessami e al.
by highe g an eceip s one yea be o e council elec ions. The ele an in e ac ion
e m coe icien is signi ican a he 10% le el. Mayo s may be unce ain abou when
g an s will be ecei ed and he e o e may submi applica ions “ oo ea ly” which in
some cases may lead o ecei ing he money ea lie . Fo his eason, we also con-
side he e ec wo yea s be o e he council elec ion: he coe icien is posi i e bu
insigni ican .40
7.2 In es men g an s and he eelec ion o council majo i ies
In a second s ep, we es whe he his hike in in es men g an eceip s ma e s o
he elec o al success o aligned councils. This equi es ha o e s indeed a e no
Table 4 E ec he e ogenei y II: Uni ied local go e nmen s
This able epo s es ima es om sha p RD es ima ions ha ela e he backg ound o he elec ed mayo
o in es men g an eceip s. The sample co e s all mayo elec ions in Hesse whe e he op wo candi-
da es we e o opposi e backg ounds ("mixed-backg ound elec ions"). The dependen a iable is he log
o in es men g an s pe municipali y, measu ed in cons an (2015) pe -capi a alues. The ea men a i-
able is a dummy equal o one i he candida e wi h an adminis a i e backg ound won he las elec ion.
We in e ac he ea men dummy as well as he con ol unc ion wi h a dummy o whe he he mayo
belongs o he ideological camp (le / igh ) which holds he council majo i y. O e all, in 51 municipali y-
yea obse a ions, he mayo ’s pa y is poli ically aligned wi h he council majo i y. O hese, 24 (27)
obse a ions e e o public admin (ou side ) mayo s espec i ely. We epo local linea (Model 1, 2 and
3) and quad a ic (Model 4) eg essions using MSERD op imal bandwid hs p oposed by Calonico e al.
(2019). Municipal con ols include popula ion size, popula ion densi y, and he wo king-age popula ion
sha e. He e oscedas ici y- and clus e - obus s anda d e o s a e in pa en heses. The uni o clus e ing is
he municipali y o he candida e. S a s indica e signi icance le els a 10% (*), 5% (**) and 1%(***)
Dep. a .: Log(in g an s p.c.) (1) (2) (3) (4)
Public admin mayo 0.020 0.159
−
0.254 0.199
(0.390) (0.341) (0.292) (0.560)
Council alignmen
−
0.203
−
0.210
−
0.558* 0.782
(0.505) (0.413) (0.322) (0.545)
Public admin mayo x 1.333** 1.667** 1.102** 1.155
Council alignmen (0.656) (0.663) (0.463) (0.903)
Bandwid h ype MSE op imal MSE op imal MSE op imal MSE op imal
Bandwid h size 11.712 11.712 11.712 11.712
Polynomial Linea Linea Linea Quad a ic
Time FE No Yes Yes No
Con ols No No Yes No
N 605 605 605 605
Elec ions 105 105 105 105
Municipali ies 103 103 103 103
Mean (SD) 3.17 (1.52) 3.17 (1.52) 3.17 (1.52) 3.17 (1.52)
40 On he o he hand, i mayo s do no conside po en ial lags, i could be ha he money is acciden ally
ecei ed in he yea o he council elec ion o e en one yea la e . I any hing, his would mean ha we
unde es ima e he ue e ec as he in es men g an eceip s in he “con ol” pe iod a e in la ed.
1983
Public adminis a o s aspoli icians ino ice
awa e o who has w i en and submi ed he g an applica ion and he e o e o e s
gi e he incumben council majo i y c edi o his success. In Table6, we in es-
iga e he ela ion be ween he in es men unds aised and he p obabili y o ee-
lec ion o he uling council majo i y. We use a linea p obabili y model, wi h a
dummy as dependen a iable equal o one i he council majo i y de ends i s majo -
i y in he subsequen council elec ion. The independen a iable is he log o in es -
men g an s pe capi a; he es ima ed coe icien s can he e o e be in e p e ed as
semi-elas ici ies.
We ind ha a 1% inc ease in in es men g an s leads o a 4.4% inc ease in he
eelec ion p obabili y o he council majo i y.41 This sugges s ha he mayo indeed
inc eases he eelec ion p obabili y o he aligned local council majo i y in he
Table 5 Mechanism: Budge cycles in in es men g an s ac oss council e ms
This able epo s es ima es om sha p RD es ima ions ha ela e he backg ound o elec ed mayo s o
in es men g an eceip s. The sample co e s candida es all mayo elec ions in Hesse whe e he op- wo
candida es we e o opposi e backg ounds ("mixed-backg ound elec ions"). The sample is limi ed o he
subsample o uni ied go e nmen s. i.e. whe e he mayo and he majo i y o councilo s belong o he
same ideological camp. The dependen a iable is he log o in es men ans e pe municipali y, meas-
u ed in cons an (2015) pe -capi a alues. The ea men a iable is a dummy equal o one i he candi-
da e wi h an adminis a i e backg ound won he las elec ion. We in e ac he ea men dummy as well
as he con ol unc ion wi h a dummy o whe he he local council is in he las wo yea s (Model 3
and 4) o inal yea o he legisla i e pe iod (Model 1 and 2). We epo local linea eg essions using
MSERD op imal bandwid hs p oposed by Calonico e al. (2019). Municipal con ols include he popu-
la ion size, popula ion densi y, and he wo king-age popula ion sha e. He e oscedas ici y- and clus e -
obus s anda d e o s a e in pa en heses. The uni o clus e ing is he municipali y o he candida e. S a s
indica e signi icance le els a 10% (*), 5% (**) and 1%(***)
Dep. a .: Log(in g an s p.c.) (1) (2) (3) (4)
Pe iod -1 -1 -2 & -1 -2 & -1
Public admin mayo 0.419 0.435 0.372 0.320
(0.389) (0.449) (0.585) (0.594)
Pe iod be o e council elec ion
−
0.734
−
0.591
−
0.157
−
0.088
(0.517) (0.487) (0.500) (0.514)
Public admin mayo x 0.938 1.026* 0.462 0.681
Pe iod be o e council elec ion (0.689) (0.616) (0.784) (0.722)
Bandwid h ype MSE op imal MSE op imal MSE op imal MSE op imal
Bandwid h size 12.229 12.229 12.229 12.229
Polynomial Linea Linea Linea Linea
Con ols No Yes No Yes
N 102 102 102 102
Elec ions 28 28 28 28
Municipali ies 27 27 27 27
Mean (SD) 2.92 (1.44) 2.92 (1.44) 2.92 (1.44) 2.92 (1.44)
41 In addi ional es ima ions, we epea he analysis using a bina y p obi model. The es ima ed e ec s a e
simila . See TableA.11 and Figu eA.5 in he online appendix.