Opinion polls, political connections, and firms

Uni e sidade
do
Minho
Escola
de
Economia
e
Ges ão
B uno Emanuel Pi es Fe nandes
Opinion Polls, Poli ical Connec ions, and
Fi ms
ou ub o
de
2024
Opinion Polls, Poli ical Connec ions, and Fi ms
B uno Fe nandes
UMinho | 2024
Uni e sidade
do
Minho
Escola
de
Economia
e
Ges ão
ou ub o
de
2024
B uno Emanuel Pi es Fe nandes
Opinion Polls, Poli ical Connec ions, and
Fi ms
Tese de Dou o amen o
Dou o amen o em Economia
T abalho
e e uado
sob
a
o ien ação
do(a)
P o esso a
Dou o a Linda Gonçal es Veiga
P o esso Dou o Luís Aguia -Con a ia
ii
DIREITOS DE AUTOR E CONDIÇÕES DE UTILIZAÇÃO DO TRABALHO POR TERCEIROS
Es e é um abalho académico que pode se u ilizado po e cei os desde que espei adas as eg as e
boas p á icas in e nacionalmen e acei es, no que conce ne aos di ei os de au o e di ei os conexos.
Assim, o p esen e abalho pode se u ilizado nos e mos p e is os na licença abaixo indicada.
Caso o u ilizado necessi e de pe missão pa a pode aze um uso do abalho em condições não p e is as
no licenciamen o indicado, de e á con ac a o au o , a a és do Reposi ó iUM da Uni e sidade do Minho.
Licença concedida aos u ilizado es des e abalho
A ibuição-NãoCome cial-SemDe i ações
CC BY-NC-ND
h ps://c ea i ecommons.o g/licenses/by-nc-nd/4.0/
iii
Acknowledgmen s
I g a e ully acknowledge he inancial suppo om he Po uguese Founda ion o Science and Technology
(FCT) h ough he SFRH/BD/138513/2018 ellowship and he suppo o my hos ins i u ion, NIPE.
I am e e nally g a e ul o my supe iso s. Thei knowledge, guidance, and ad ice we e essen ial
o my g ow h as a young esea che . I am indeb ed o Linda Veiga, who has accompanied me since my
mas e ’s hesis as a supe iso . I owe i o he i he e is any quali y in me as a esea che . I am also
indeb ed o Luís Aguia -Con a ia. His s aigh o wa dness and his p ac ical sense helped me o make
p og ess in my esea ch and o o e come obs acles.
I would also like o exp ess my since e g a i ude o all hose who ha e accompanied me on his
jou ney. P o esso s, colleagues, and iends ha e been essen ial o my success and pe sonal and
p o essional g ow h. I mus especially men ion Hélde , João, and Rui, who ha e been wi h me since he
beginning o my unde g adua e s udies. We ha e sha ed g ea memo ies and coun less s udy hou s,
which I conside p iceless.
I am deeply hank ul o my amily, who ha e always suppo ed me and shown me he igh
di ec ion. A special hank you o my g andpa en s, who ha e un o una ely al eady passed away bu who
I belie e would be p oud o my jou ney. I wan o hank my pa en s and sis e o hei uncondi ional lo e
and suppo .
I wan o exp ess my g a i ude o Ma ga ida o e e y hing she has done o me, o he
uncondi ional suppo and mo i a ion, and o sha ing he good and bad imes wi h me. Wi hou he , i
would no ha e been easy o each he end.

i
STATEMENT OF INTEGRITY
I he eby decla e ha ing conduc ed his academic wo k wi h in eg i y. I con i m ha I ha e no used
plagia ism o any o m o undue use o in o ma ion o alsi ica ion o esul s along he p ocess leading o
i s elabo a ion.
I u he decla e ha I ha e ully acknowledged he Code o E hical Conduc o he Uni e si y o Minho.
Sondagens, Conexões Polí icas e Emp esas - Resumo
Um dos emas mais deba idos em Economia Polí ica é o impac o da economia na popula idade
dos líde es polí icos. É amplamen e acei e que es a em impac o na popula idade, no en an o, a
in es igação em p oduzido esul ados ins á eis e, po ezes, con adi ó ios. Os ês p imei os ensaios
des a disse ação p ocu am da oco a po enciais causas des e p oblema.
O p imei o ensaio es uda a di e ença en e a economia “obje i a” e “mediada”. O impac o da
e olução e i icada da economia na ap o ação dos polí icos pode se meno do que o impac o da
economia e a ada nos
media
. Com uma base de dados pa a Po ugal, pa a o pe íodo de 2001 a 2018,
analisamos se a ap o ação melho a quando u ilizamos a in o mação económica que é disponibilizada
ao público. Concluímos que es a é p e is a com maio igo quando usando o c escimen o económico
p oje ado pa a o ano em cu so.
Num sis ema semip esidencial, a ap o ação que dos p imei os-minis os, que dos p esiden es,
pode se in luenciada pela economia. Usando uma base de dados pa a Po ugal en e 1986 e 2018, o
segundo ensaio es uda a possibilidade de os e ei os económicos se em assimé icos, ou seja, pe íodos
de c ise a exe ce em e ei os mais signi ica i os, e se ais e ei os são mi igados em pe íodos de coabi ação.
Os esul ados suge em que a coabi ação e a al a de incen i os à eeleição pa a os p esiden es,
con ibuem pa a obscu ece a esponsabilização dos polí icos pelos esul ados económicos.
Ado ando uma abo dagem di e en e, o e cei o ensaio ap imo a o uso das e amen as
wa ele
na ciência polí ica, ala gando o oco pa a a análise mul i a iada e in oduzindo es imações de equações
usando wa ele s. O ensaio explo a o impac o de ês a iá eis económicas na popula idade do úl imo
p esiden e dos EUA a cump i dois manda os, Ba ack Obama. Os esul ados mos am que a sua
ap o ação es á es u u almen e ligada às axas de desemp ego e de in lação, com di e en es pad ões de
sinc onização, e elacionada de o ma ansi ó ia com o sen imen o do consumido .
Finalmen e, o qua o ensaio abo da a elação en e polí ica e negócios, um ema mul idisciplina .
In es iga as ca ac e ís icas e o impac o das conexões polí icas nas emp esas co adas na bolsa
po uguesa en e 2002 e 2022, es imando a p obabilidade de e conexões polí icas e o impac o nas
a iá eis inancei as. Os esul ados do ensaio suge em que as conexões polí icas são mais comuns em
emp esas de maio dimensão, especialmen e em indús ias eguladas e localizadas na á ea de Lisboa.
No en an o, o impac o inancei o dessas conexões pa ece se limi ado.
Pala as-Cha e: Análise com Wa ele s; Coabi ação Polí ica; Conexões Polí icas; Funções Popula idade;
Va iá eis Económicas
i
Opinion Polls, Poli ical Connec ions, and Fi ms - Abs ac
One o he mos s udied and deba ed opics in he ield o Poli ical Economy is he impac o he
economy on how o e s e alua e hei poli ical leade s. I is widely accep ed ha he economy in luences
poli ical popula i y, ye esea ch has o en p oduced uns able and some imes con adic o y esul s. The
i s h ee essays o his disse a ion aim o shed ligh on he po en ial causes o his p oblem.
The i s essay s udies he di e ence be ween he “objec i e” and he “media ed” economy.
Economic de elopmen s on he g ound may less in luence o e s’ app o al o incumben s han he
economy po ayed by he media. Using e idence om Po ugal om 2001 o 2018, we examine whe he
he abili y o p edic app o al is imp o ed when we employ economic in o ma ion made a ailable o he
public. We conclude ha app o al is bes p edic ed by o ecas ed g ow h o he cu en yea .
In a semi-p esiden ial sys em, he app o al o p ime minis e s and p esiden s can be a ec ed by
economic pe o mance. The second essay uses app o al da a and GDP g ow h igu es om 1986 un il
2018 in Po ugal. We examine he possibili y ha economic e ec s a e
asymme ic
, wi h bad economic
imes exe ing s onge e ec s, and whe he such e ec s a e supp essed in pe iods o cohabi a ion.
Resul s sugges ha cohabi a ion combined wi h a lack o eelec ion incen i es o p esiden s con ibu e
o obscu e accoun abili y o economic ou comes, diminishing he ole o economic pe o mance.
Taking a di e en app oach, he hi d essay enhances he use o he wa ele oolse o poli ical
science by ex ending he analy ical ocus om bi a ia e o mul i a ia e and in oduces eg ession
es ima ion using wa ele s. I explo es he impac o h ee economic a iables on he popula i y o he
mos ecen wo- e m US P esiden , Ba ack Obama. The esul s show ha Obama’s app o al a ing was
s uc u ally linked o unemploymen and in la ion a es, wi h di e en synch oniza ion pa e ns, and
ansien ly ela ed o consume sen imen .
Finally, he ou h essay add esses he ela ionship be ween poli ics and business, a
mul idisciplina y opic. I in es iga es he cha ac e is ics and impac o poli ical connec ions on
Po uguese publicly lis ed i ms be ween 2002 and 2022 and es ima es he p obabili y o poli ical
connec ions and hei e ec on i ms’ inancial ou comes. The esul s sugges ha poli ical connec ions
a e mo e common in la ge i ms, especially in egula ed indus ies and hose loca ed in he Lisbon a ea.
Howe e , he o e all inancial impac o hese connec ions appea s o be limi ed.
Keywo ds: Economic Va iables; Poli ical Cohabi a ion; Poli ical Connec ions; Popula i y Func ions;
Wa ele Analysis.
ii
Table o Con en s
1. In oduc ion ................................................................................................................................ 1
2. Economic Fo ecas s and Execu i e App o al ........................................................................ 5
2.1. In oduc ion ........................................................................................................................ 5
2.2. The “Real” and he “Media ed” Economy ...................................................................... 6
2.3. Da a and Me hods .............................................................................................................. 9
2.4. Resul s ............................................................................................................................... 14
2.4.1. Vin age s. Fi s Values ............................................................................................ 14
2.4.2. The Fo ecas ed Economy......................................................................................... 16
2.5. Conclusion ......................................................................................................................... 18
3. The Economy and Execu i e App o al in a Semi-P esiden ial Regime: he Case o
Po ugal ......................................................................................................................................... 20
3.1. In oduc ion ...................................................................................................................... 20
3.2. Public app o al and he economy in Po ugal .............................................................. 22
3.2.1. Economic Con ex s and Asymme y ....................................................................... 22
3.2.2. Po uguese Semi-P esiden ialism and i s Implica ions ........................................ 23
3.3. Da a, Va iables, and Es ima ion S a egy ..................................................................... 26
3.3.1. Da a and Va iables ................................................................................................... 26
3.3.2. Es ima ion S a egy .................................................................................................. 30
3.4. Resul s ............................................................................................................................... 31
3.5. Conclusion ......................................................................................................................... 36
4. US P esiden ial App o al Ra ing: Time-F equency App oach ............................................ 39
4.1. In oduc ion ...................................................................................................................... 39
4.2. Wa ele analysis and he s udy o popula i y unc ions .............................................. 40
4.2.1. The challenge o ins abili y ...................................................................................... 40
4.2.2. The (pa ial) wa ele cohe ency, gain, and phase-di e ence ............................ 41
4.3. An example ........................................................................................................................ 43
4.3.1. The Da a Gene a ing P ocess ................................................................................. 43
4.3.2. Reading he esul s ................................................................................................... 44
4.4. Da a .................................................................................................................................... 46
3
In imes o economic u bulence, he impac o he economy on popula i y is pe cei ed o be
signi ican . Thus, in Chap e 4, I s udied he p esidency o Ba ack Obama, cha ac e ized by a slow
economic eco e y a e he subp ime mo gage c isis in 2007. To do so, I used wo impo an economic
a iables, he unemploymen a e and he in la ion a e, as p oxies o he s a e o he economy, and an
addi ional a iable able o ca ch o e ’s pe cep ions, which is he consume sen imen , sou ced by he
Michigan Uni e si y. Compa ing he esul s o OLS es ima ions and he esul s o wa ele analysis, unning
es ima es ha a e ime- a ying can shed some ligh on he easons why he ela ionship be ween economy
and poli ical popula i y has no been consis en .
Poli icians ca e abou hei popula i y because hey desi e o win elec ions. Maximizing he
p obabili y o being elec ed is a p io i y o he incumben . The li e a u e on poli ical business cycles is
abundan , s a ing ha poli icians make an e o o accele a e economic pe o mance by educing
unemploymen o os e ing he economy o by inc easing isible expendi u es when he elec ions d aw
nea . One mechanism by which he incumben may use o cause his poli ical cycle is h ough poli ical
connec ions. Poli ically in luenced i ms can bene i incumben s a ound he elec ion ime by os e ing
employmen (Be and
e al.
, 2018) o con ibu ing di ec ly o indi ec ly wi h unds o he campaign
(Claessens, 2008; Sukh anka , 2012).
Howe e , poli icians also ha e in e es s o he han elec o al o poli ical. They ha e hei own
p i a e in e es s. The e o e, i is no di icul o ind poli icians as membe s o i ms’ boa ds o o he
ela ionships o di e en na u es (e.g., lawye s, consul an s, o audi o s). Fi ms also bene i om
es ablishing hese connec ions wi h poli icians in a quid p o quo ela ionship. O he wise, hey would no
exis (Faccio and Hsu, 2017). They can be benign, wi h connec ed i ms eaping he bene i s o ha ing
epu ed people who know how he go e nmen adminis a ion wo ks, o malign. Some i ms a e in e es ed
in es ablishing connec ions o ob ain ad an ages agains hei compe i o s, which could be associa ed
wi h en -seeking ac i i ies o co up ion.
A su ey conduc ed in 2024 by he Eu opean Union (EU) o measu e “[…] he le el o co up ion
pe cei ed and expe ienced by businesses […]” indica es ha 40% ag ee wi h he idea ha in hei coun y,
he only way o succeed in business is o ha e poli ical connec ions (Eu opean Commission, 2024a). The
pe cen age becomes mo e p onounced wi h he Po uguese businesspeople, wi h 65% ag eeing wi h he
s a emen ( he highes alue among he EU coun ies). In addi ion, he majo i y o he EU businesspeople
(79%) ag ee ha links be ween businesses and poli ics lead o co up ion. The pe cen age inc eases o
82% when we es ic he analysis o Po uguese businesspeople. Ano he su ey conduc ed in 2024 by
he EU o e alua e “[…] he le el o co up ion pe cei ed and expe ienced by Eu opean ci izens […]”

4
highligh s ha 51% ag ee ha poli ical connec ions a e key o succeed in business, and he pe cen age
inc eases o 55% among Po uguese ci izens (Eu opean Commission, 2024b). As in he su ey o
businesses, 75% o Eu opean ci izens ag ee ha links be ween business and poli ics a e associa ed wi h
co up beha io . Wi hin he Po uguese ci izens, he pe cen age inc eases o 81%. The pe cei ed link
be ween poli ical connec ions and co up ion unde mines o e s’ con idence in poli icians and in luences
hei e alua ion o how good o bad poli icians a e handling hei jobs (Ande son and T e do a, 2003).
Howe e , his pe cep ion does no p e en poli icians om being elec ed, e en i hey ha e been accused
o condemned o co up ion (de Sousa and Mo iconi, 2013).
Li e a u e has s udied poli ical connec ions and he complex ne wo k be ween poli ics and
businesses. Many o hese ne wo ks a e associa ed wi h en -seeking beha io and co up ion, which ha e
a pe asi e impac on economic ou comes (Schoenhe , 2019; Khwaja and Mian, 2005). O en, his
ela ionship in ol es a
quid p o quo
dynamic cha ac e ized by ecip ocal exchanges and bene i s:
inc easing poli ical in luence o poli icians o ex ac ing en s, easing egula ions, o c ea ing ba ie s o
en y o incumben i ms. The p esence o poli icians and o me poli icians on company boa ds is a
global phenomenon, no exclusi e o de eloping coun ies.
In Po ugal, i is common o ind o me poli icians on company boa ds. Chap e 5 o his
disse a ion del es in o he exis ence and cha ac e is ics o poli ical connec ions in Po ugal. S udying
hese connec ions in Po ugal is impo an because i allows esea che s o analyze hei impac on he
economy. This chap e examines poli ical connec ions in publicly aded Po uguese companies om
2002-2022. Fi m size and egula ed sec o s a e associa ed wi h poli ical ies, pa icula ly hose
headqua e ed in Lisbon. Addi ionally, igh -wing poli ically connec ed membe s a e common in la ge
i ms, whe eas le -wing poli ically connec ed membe s a e mo e no iceable in i ms whose headqua e s
a e loca ed in he Lisbon a ea. Howe e , he analysis shows no signi ican impac o poli ical connec ions
on inancial pe o mance.
Finally, Chap e 6 summa izes he main indings om he disse a ion, discusses limi a ions, and
o e s sugges ions o u u e esea ch.
5
2. Economic Fo ecas s and Execu i e App o al
1
2.1. In oduc ion
One o he mos es ablished ideas in he s udy o public opinion and elec ions is ha o e s ca e abou
he economy when e alua ing incumben s (Lewis-Beck and S egmeie , 2000 and 2013). Ye , objec i e
economic pe o mance indica o s, such as GDP g ow h, unemploymen , o in la ion, do no seem o be
e y s ongly ela ed o poli ical suppo as exp essed in execu i e app o al o o ing o he incumben .
Al hough “ he ( eal) economy ma e s,” “i s e ec s a e weak and con ain li le explana o y powe ” (Van
de Eijk
e al.
, 2007: 180). “[T]he economy is only a small de e minan o elec o al ou comes (…) and
cap u es a iny p opo ion o he o e all a ia ion in suppo o he lead pa y” (Kayse , 2014: 117). Some
e en a gue ha , a leas o he weal hie and mo e es ablished democ acies, economic g ow h may ha e
no e ec on he elec o al a e o incumben s (B ende and D azen, 2007). How can we econcile he
seemingly ob ious no ion ha he economy is essen ial o o e s wi h he weak explana o y powe o
objec i e economic indica o s?
In a 2015 s udy, Kayse and Leininge (2015) use an imagina i e esea ch design o add ess his
puzzle. Taking he case o he Uni ed S a es, hey es ima e di e en elec ion o ecas ing models employing
in age
da a
on
GDP g ow h ( he e ised his o ical es ima es o economic pe o mance, e lec ing he
mos accu a e measu emen s o economic ac i i y a ailable) and
eal- ime
da a on GDP g ow h ( he i s
es ima es o economic pe o mance eleased immedia ely a e he pe iod o which hey a e e e ing and
ci cula ed in he media). They conclude ha he da a ha mos accu a ely e lec s he s a e o he
economy ( he in age da a) ends up being a
wo se
p edic o o poli ical suppo han he less accu a e
ini ial economic es ima es con eyed by he media ( he eal- ime da a). In o he wo ds, he bes measu e
1
A e sion o his Chap e was published as: Aguia -Con a ia, L., Fe nandes, B., & Magalhães, P. C. (2024). Economic o ecas s and execu i e app o al.
Jou nal o Elec ions, Public Opinion and Pa ies, 34(4), 643-655.
6
o “ eal” economic g ow h and he media ed in o ma ion abou economic g ow h a ailable o o e s do
no necessa ily ma ch, and he la e is mo e poli ically consequen ial han he o me .
This essay d aws inspi a ion om Kayse and Leininge ’s app oach bu o es ye an al e na i e
answe as o why objec i e indica o s abou he ecen pe o mance o he economy may no be e y good
p edic o s o poli ical suppo : he a ailable in o ma ion people a e mo e likely o use o e alua e
incumben s is no abou ecen economic pe o mance, bu ins ead abou expec ed
u u e
pe o mance.
In o he wo ds, o e s ely upon in o ma ion o e en mo e unce ain accu acy han eal- ime da a o
e alua e incumben s, and hey do so p ospec i ely a he han e ospec i ely.
We es his hypo hesis o he case o Po ugal by looking a di e en se ies o economic da a as
p edic o s o p ime minis e ial app o al. Fi s , we conside he mos ecen ly a ailable S a is ics Po ugal
(INE) his o ical ime-se ies da a on he yea -on-yea GDP g ow h a e o each qua e , which ha e been
e ised o accu acy. Howe e , we also conside bo h he p elimina y — less accu a e — es ima es o he
same GDP g ow h announced by S a is ics Po ugal in hei qua e ly publica ions since 2000 and he
OECD’s o ecas s o GDP g ow h ob ained om he o ganiza ions’ Economic Ou look issued in he second
and ou h qua e s o each yea
.
We show ha he in o ma ion made a ailable o e ime abou he
o ecas ed g ow h o he cu en yea
is he bes p edic o o p ime minis e ial app o al. In o he wo ds,
like Kayse and Leininge (2015), we ind indi ec e idence ha o e s espond less o he “objec i e
economy” han hey do o he in o ma ion ha happens o be a ailable abou i . Howe e , in he case o
Po ugal, he mos consequen ial a ailable economic in o ma ion seems o be no abou he pas bu
a he abou he nea u u e.
The chap e is s uc u ed as ollows. Chap e 2.2 discusses he li e a u e on popula i y unc ions
and he media's impac on people’s expec a ions and pe cep ions. Chap e s 2.3 and 2.4 desc ibe he
da a and me hodology used and he econome ic esul s. Chap e 2.5 concludes.
2.2. The “Real” and he “Media ed” Economy
A long adi ion in poli ical science, ini ia ed by Muelle (1970) and Goodha and Bhansali (1970), has
shown ha o e s’ suppo o he execu i e is ela ed o economic pe o mance. GDP g ow h, in la ion,
and le els o /changes in unemploymen a es ha e all been shown, a one ime o ano he , o accoun
o pa o he a iance in he suppo awa ded by he public o he execu i e, as cap u ed by su ey-
based indica o s o by he o e in he incumben pa y o p esiden .
7
Howe e , since hose ounding wo ks, esea ch has inc easingly mo ed away om a concep ion o
“a public ha almos mechanis ically esponds o economic condi ions” (G onke and Newman, 2003:
506). The cen al aspec o his mo emen is he no ion ha wha should ul ima ely be consequen ial o
poli ical suppo is o e s’ subjec i e pe cep ions o economic pe o mance. Those pe cep ions, al hough
hemsel es shaped by economic condi ions (Beche and Donnelly, 2013), a e also a ec ed by a my iad
o o he ac o s besides he “ eal” pe o mance o he economy. These include o e s’ poli ical
p edisposi ions (Wlezien
e al.,
1997; E ans and Ande sen, 2006; Chzhen
e al.,
2014), news co e age,
and people’s exposu e o i (Goidel and Langley, 1995; Sande s and Ga in, 2004; De Boe and Kells ed ,
2004; Boyds un
e al.,
2018), and combina ions o he abo e (Duch
e al.
, 2000).
This basic idea a once cla i ies and complexi ies he ela ionship be ween he economy and
poli ical suppo . On he one hand, i sugges s ha he subs an i ely small co ela ions be ween economic
agg ega es and poli ical suppo ha ha e eme ged in he li e a u e should no be in e p e ed as signaling
ha o e s do no ca e abou he economy. Ins ead, hey sugges ha he s udy o he economic d i e s
o poli ical suppo was ne e p ima ily abou he “ eal” o he “measu ed” economy bu a he abou he
economic in o ma ion made a ailable o di e en o e s by he media — he “media ed economy” — and
how hose o e s deal wi h ha in o ma ion o o m subjec i e pe cep ions (S e enson and Duch, 2013).
On he o he hand, elying on subjec i e economic pe cep ions as p edic o s o poli ical suppo aises
he isk o o e es ima ing how much economic pe cep ions eally ma e o o e s when e alua ing
incumben s, pa icula ly i we conside how much hose pe cep ions may be d i en by poli ical suppo
i sel and how di icul i is o add ess he ela ed measu emen and endogenei y p oblems in obse a ional
s udies (Ande son, 2007, and many o he s). This dilemma emains, un il oday, one o he mos
con en ious in he sub- ield (Hellwig and Ma ino a, 2017).
Howe e , he impo ance o he “media ed economy” can be shown, a leas indi ec ly, in ways
ha ci cum en such h ea s o in e ence. Kayse and Leininge (2015) examine he p edic i e powe o
di e en o ecas ing models o U.S. p esiden ial elec ions ha employ economic a iables. As economic
inpu , hey use wo ypes o indica o s. The i s ,
Vin age
, co esponds o GDP o GNP g ow h es ima es
as e ised by he U.S. Bu eau o Economic Analysis. The second,
Real-Time
, is he ea ly es ima es o
GDP o GNP g ow h o each qua e published one o wo mon hs a e he end o ha qua e . In o he
wo ds, while
eal- ime
da a cap u e he economic in o ma ion a ailable o o e s as each elec ion
app oaches, as con eyed by he media,
in age
da a co espond o enhanced measu emen s ha mo e
accu a ely cap u e economic ac i i y and pe o mance. They ind ha
in age
da a in oduces g ea e
o ecas ing e o s han
eal- ime
da a and ha he e o inc eases as he numbe o e isions inc eases.
8
In o he wo ds, he mo e accu a e he economic da a, and he mo e di e en hey become om he ini ial
es ima es, he wo se p edic o s o poli ical suppo hey become. Indi ec ly, his sugges s ha o e s’
esponse o economic pe o mance is a esponse no so much o he “objec i e” expe ience o how he
economy has pe o med bu a he o in o ma ion abou economic pe o mance as con eyed h ough he
media.
Howe e , aking he media ed economy se iously o ces us o conside ha i is no composed
exclusi ely o in o ma ion abou ecen economic pe o mance. I also con ains in o ma ion abou he
u u e, including o ecas s o economic pe o mance om p o essional economis s, he go e nmen ,
cen al banks, o in e na ional o ganiza ions. Se e al s udies ha e shown ha , a leas in coun ies like
he US, he UK, o Canada, he posi i i y and nega i i y o economic news seems o mi o mo e closely
leading
a he han cu en o pas economic indica o s. In o he wo ds, he one o economic news is
mo e a e lec ion o u u e economic de elopmen s han pas ones (So oka
e al.
, 2015; Wlezien
e al.
,
2017).
Co espondingly, media messages abou he economy seem o shape no only people’s subjec i e
pe cep ions abou how he economy has been doing bu also hei subjec i e expec a ions abou how he
economy is likely o pe o m in he u u e (So oka
e al.
, 2015). Some e en a gue ha media messages’
p ima y ole is o shape expec a ions abou a mo e unce ain u u e a he han a ec ing e alua ions o
an al eady expe ienced pas (Boomgaa den
e al.,
2011; Dams a and Boukes, 2021; Boukes
e al.
,
2019). And he e is, o cou se, conside able e idence ha pas economic pe o mance o people’s
pe cep ion o i a e no necessa ily he bes p edic o s o poli ical suppo . Ins ead, o e s may beha e
like “banke s,” judging go e nmen s based on
expec a ions
ega ding he u u e economy (MacKuen
e
al.
, 1992; E ikson
e al.
, 2000), a leas o some o e s (Al
e al.
, 2016; Lacy and Ch is enson, 2017;
Ace edo
e al.
, 2017), in some con ex s (Singe and Ca lin, 2013), and in some poli ical sys ems (Cohen,
2004).
Tha he “media ed economy” also con ains in o ma ion abou he u u e and ha o e s ca e abou
such in o ma ion is a possibili y no con empla ed by Kayse and Leininge ’s (2015) compa ison o
“ in age” and “ eal- ime” economic indica o s in p edic ing poli ical suppo . In hei app oach, o e s
who ha e “expe ienced he pas ” o a e in o med abou i a e assumed o de i e poli ical suppo om
ha expe ience o in o ma ion. Howe e , i economic in o ma ion is also shaped by economic o ecas s
and/o o e s a e p ospec i e ins ead o e ospec i e, in o ma ion abou he u u e should also ma e o
explain poli ical suppo . This is exac ly wha we p opose o examine in his s udy: besides examining he

9
ole o di e en measu es abou he ecen pe o mance o he economy, we also examine he ole o
o ecas s
o ha pe o mance.
2.3. Da a and Me hods
Ou dependen a iable is a measu e o p ime minis e ial app o al in Po ugal a ailable a he Execu i e
App o al P ojec (Ca lin
e al.,
2019). We use qua e ly da a om 2001 o 2018, esul ing om polls
conduc ed by Po uguese polls e s. We use a smoo hed e sion o he a iable
App o al
(% o posi i e
a ings) a ailable in he Execu i e App o al Da abase 2.0 o Po uguese P ime Minis e s. Figu e
1
plo s
he e olu ion o p ime minis e ial app o al om he i s qua e o 2001 un il he hi d qua e o 2018,
co e ing eigh go e nmen s and six p ime minis e s.
To measu e economic pe o mance, we conside se e al di e en a iables o cap u e he g ow h
a e o Po uguese’s eal GDP (Table 1):
GDP Vin age G ow h
,
GDP Fi s Values G ow h, GDP Real-Time
Figu e 1: Qua e ly da a on app o al a ing o Po uguese P ime Minis e s om he i s qua e
o 2001 o he hi d qua e o 2018.
No e: Each dashed line co esponds o a go e nmen ’s inaugu a ion. Sou ce: Ca lin e al. (2019).
10
G ow h, OECD Fo ecas Cu en , and OECD Fo ecas Nex
.
2
Because he di e ences a e sub le, we will
desc ibe hem in de ail.
GDP Vin age G ow h
ep esen s he mos accu a e da a on he e olu ion o eal
GDP, he e ised alues o each qua e . This in o ma ion is con eyed by he Po uguese S a is ics
Ins i u e (INE). We use he e ised da a published in he i s qua e o 2021, ep esen ing he bes
in o ma ion a ailable ega ding he e olu ion o eal GDP h oughou he en i e 2001-2018 pe iod. Fo
he pu pose o ou econome ic analyses, we ake in o conside a ion ha o e s need ime o in e nalize
economic in o ma ion and hus assume a con en ional sho lag be ween he economic measu es and
app o al (one qua e , -1).
Table 1: Economic a iables employed.
Va iables
De ini ion
Sou ce
GDP Vin age G ow h
-1
2021 e ision o he yea -on-yea eal GDP
g ow h a e a he qua e be o e app o al
was measu ed.
S a is ics Po ugal
(INE)
GDP Fi s Values G ow h
-1
Fi s es ima es made public o he yea -on-
yea eal GDP g ow h a e a he qua e
be o e app o al was measu ed.
S a is ics Po ugal
(INE)
GDP Real-Time G ow h
-1
The mos ecen publicly a ailable qua e ly
yea -on-yea eal GDP g ow h a e a he
qua e be o e app o al was measu ed.
S a is ics Po ugal
(INE)
OECD Fo ecas Cu en
-1
The mos ecen yea ly o ecas o eal GDP
g ow h o he cu en yea publicly a ailable
a he qua e be o e app o al was
measu ed.
OECD Economic
Ou look
OECD Fo ecas Nex
-1
The mos ecen o ecas o eal GDP g ow h
o he ollowing yea publicly a ailable a he
qua e be o e app o al was measu ed.
OECD Economic
Ou look
GDP Fi s Values G ow h
-1
is cons uc ed using he e y i s alues o he yea -on-yea qua e ly
eal GDP g ow h a e disclosed by INE in hei epo s issued since 2000. The di e ence be ween in age
o e ised and i s alues o GDP g ow h a e is obse ed in Figu e 2. A i s glance, he p elimina y
alues end o be mo e conse a i e han e ised alues. Again, in he analysis, we ma ch app o al a es
2
Economic indica o s such as unemploymen o in la ion a es ha e also been equen ly included in (ea ly) s udies o Po uguese go e nmen app o al
unc ions (Veiga 1998; Veiga and Veiga, 2004a). Howe e , while in la ion emained ema kably s able in Po ugal h oughou he pe iod unde examina ion,
unemploymen a es, as Goula and Veiga (2016) a gue, we e somewha decep i e indica o s o economic pe o mance, pa icula ly in pe iods when massi e
emig a ion (such as du ing he 2008-2013 economic and inancial c isis, when i eached an his o ical peak) con ibu ed o mask he ue magni ude o he
c isis as exp essed in unemploymen a es.
11
wi h he alue o
GDP Fi s Values G ow h
o he p eceding qua e . Howe e , i we ake he poin o iew
o when in o ma ion is made a ailable o he public, we should conside he ac ha he GDP g ow h
announced o he public by S a is ics Po ugal in each qua e e e s no o he qua e o publica ion bu
o p e ious ones. The e o e, we also include, o each qua e p eceding he one when app o al is gauged,
he mos ecen publicly a ailable alue o he yea -on-yea GDP g ow h: i is
GDP Real-Time G ow h -1
.
Finally, we conside he ole o economic o ecas s. We ex ac hem om OECD’s epo s, i.e.,
“OECD Economic Ou look.” Those o ecas s a e con eyed wice pe yea , e e y second and ou h
qua e . In hese publica ions, he OECD discloses annual eal GDP o ecas s o each membe coun y
o he cu en and ollowing yea s. Al hough se e al ins i u ions p oduce g ow h o ecas s, he OECD
p o ides he mos ex ended se ies.
3
I s p ojec ions a e always issued using he same iming. They
consis en ly p ojec g ow h o he cu en and he nex yea s and a e a highly epu ed sou ce hea ily
co e ed by he Po uguese media. They a e in luen ial in bo h news con en and opinion a icles. F om
2000 o 2018,
Lusa
, he sole Po uguese Na ional News Agency and a p ima y sou ce o o he media
3
O he op ions would ha e been he p ojec ions o he IMF o he Eu opean Commission. Howe e , while he o me has mo e i egula publica ion da es in
compa ison wi h he OECD, he la e has changed publica ion schedules in bo h 2012 and 2018. Mo eo e , Po ugal ecei ed aid om bo h ins i u ions
be ween 2011 and 2014, making hem signi ican poli ical en i ies, which could a ec he pe cei ed impa iali y o hei o ecas s in compa ison o he OECD.
Figu e 2: Qua e ly da a on Real GDP G ow h Ra e, pe cen age change om he yea ago.
No e: The ull line ep esen s he e y i s alues o qua e ly eal GDP g ow h, and he
dashed line plo s e ised alues o qua e ly eal GDP g ow h. Sou ce: INE.
12
ou le s, p oduced 97 news a icles wi h he ph ase “economic o ecas s” in he headline and nea ly 2500
men ions o his e m wi hin he body o he ex . The speci ic combina ion be ween he same exp ession
and “OECD” occu s in 416 news i ems in Po uguese news sou ces a ailable in Fac i a in he same
pe iod.
4
Ou logic emains he same: we ocus on he in o ma ion a ailable o he public in he qua e
be o e p ime minis e ial app o al is gauged. The e o e, we s a by modeling
PM App o al
a he i s and
second qua e s o each yea as a unc ion o he mos ecen OECD’s g ow h o ecas o ha yea (which
was issued in he las qua e o he p eceding yea ), while
PM App o al
a he hi d and ou h qua e s
o each yea is modeled as a unc ion o g ow h p ojec ion ( o ha yea ) issued in he second qua e :
OECD Fo ecas Cu en
-1. Finally, we also conside he possibili y ha
PM App o al
may be a ec ed by
in o ma ion abou o ecas s o he yea
a e
which app o al is gauged. Co espondingly, his a iable
ollows he same s uc u e as he p e ious one, bu now we conside he GDP g ow h o ecas o he
ollowing yea :
OECD Fo ecas Nex
-1. The plo s o hese o ecas s can be obse ed in Figu e 3. As
suspec ed, o ecas s o he cu en yea end o be mo e accu a e han o ecas s o he yea beyond.
4
We a e g a e ul o Tiago Dias, om
Lusa
, and o Nelson San os, om ICS-ULisbon, o sha ing hese da a.
Figu e 3: Qua e ly da a on annual OECD o ecas s and annual eal GDP inal alues.
No e: The ull line shows p ojec ions o he cu en yea , and he dashed line plo s he
p ojec ions o he nex yea . Sou ce: OECD.
19
Be ween 2001 and 2018, Po ugal expe ienced h ee ecessions. The hi d one, a so e eign deb
c isis, was se e e enough o igge a eques o in e na ional inancial assis ance. As a esul , i is no
un easonable o sugges ha he Po uguese elec o a e may ha e been pa icula ly a en i e o he
economic p ojec ions p esen ed by in e na ional o ganiza ions such as he OECD. To ensu e he
gene alizabili y o ou indings, u he in es iga ions should be conduc ed in a ious egions. A possibili y
o u u e esea ch would be o conduc a panel s udy, aking ad an age o he consis en p oduc ion o
he OECD Economic Ou look ac oss se e al coun ies.
Fu he mo e, a comple e es o his a gumen would ha e o ely on measu emen s o indi idual-
le el pe cep ions, he way hey a e shaped by economic o ecas s and hei media co e age, and, in u n,
how hey d i e execu i e app o al. S ill, hese esul s sugges he impo ance o economic o ecas s in
shaping public a i udes. They also encou age possible new app oaches o modeling in elec ion o ecas s,
an impo an ield o poli ical economy s udies (S egmaie and No po h, 2013). Typically, models ha
include economic a iables eso o “ in age” da a o es ima ion and analysis o i and hen plug “ eal-
ime” da a o p oduce conc e e o ecas s. To he ex en ha he lessons d awn he e abou execu i e
app o al ex end o elec o al suppo o incumben s, economic o ecas s may cap u e be e ci izens’
o e all pe cep ions o he s a e o he economy and how hey use hose pe cep ions o e alua e poli ical
pe o mance.

20
3. The Economy and Execu i e App o al in a Semi-P esiden ial
Regime: he Case o Po ugal
10
3.1. In oduc ion
O e he las wo decades, se e al s udies ha e enhanced ou unde s anding o he ela ionship be ween
he economy and public suppo o he execu i e in Po ugal (Veiga 1998; Veiga and Veiga, 2004a and
2007). The cen al inding is clea : he public app o al o go e nmen s and p ime minis e s is a ec ed
by economic pe o mance. This is no a all su p ising. Le - igh socio-economic policy issues ha e almos
wholly domina ed ci izens' conce ns h oughou Po ugal's democ a ic his o y (Tsa sanis
e al.
, 2014:
527). Despi e a p opo ional ep esen a ion elec o al sys em, pa y sys em agmen a ion in Po ugal has
emained ela i ely low and was, un il ecen ly, declining, a singula end among es ablished democ acies
(Lijpha , 2012: 75). Since 1985, cabine s ha e ei he been single-pa y go e nmen s (o en wi h an
absolu e majo i y) o minimum-winning ideologically cohesi e coali ions. The combina ion o pe sis en ly
high salience o economic issues and high cla i y o esponsibili y o ou comes makes Po ugal a clea -
cu candida e o a solid and s able linkage be ween economic pe o mance and he execu i e's public
app o al.
Howe e , se e al blind spo s emain. We know, o example, ha he ela ionship be ween
objec i e economic pe o mance and execu i e app o al (Ca lin
e al.,
2023) o suppo o incumben
pa ies (F ei e and San ana-Pe ei a, 2012; Lobo and Pannico, 2020) has been s onge in some pe iods
han in o he s in Po ugal. Despi e sha ply con as ing economic condi ions, consecu i e elec ions ha e
yielded simila ly he y punishmen s o go e nmen pa ies (Magalhães, 2017). Fo ecas ing models o
incumben pa y o e ha e been equi ed o inco po a e poli ical a iables o accoun o he insu icien
10
A e sion o his Chap e was published as: Aguia -Con a ia, L., Fe nandes, B., & Magalhães, P. C. (2023a). The economy and execu i e app o al in a semi-
p esiden ial egime: he case o Po ugal. In T. Hellwig, & M. M. Singe (Eds.), Economics and poli ics e isi ed: Execu i e app o al and he new calculus o
suppo (pp. 108–130). Ox o d Uni e si y P ess.
21
p edic i e powe o economic a iables (Magalhães and Aguia -Con a ia, 2009). In o he wo ds, he
ques ions o whe he o e o popula i y unc ions a e s able and o he a iance in public suppo o he
incumben explained by he economy ( he "e- ac ion" o hose unc ions), amously aised by Paldam
(1991), emain ele an and in iguing o he case o Po ugal.
In his chap e , we ad ance wo a gumen s abou how o econcile he e idence abou he
simul aneous impo ance and ins abili y o economic pe o mance as a d i e o execu i e suppo in
Po ugal. The i s poin s o he
economic con ex
i sel and wha i should imply o he cla i y o
pe o mance signals a ailable o o e s. Mo e speci ically, we in es iga e whe he execu i es a e mo e
punished by ad e se economic de elopmen s han hey a e ewa ded by posi i e ones. Al hough he
possibili y o such
asymme y
in he ela ionship be ween economic pe o mance and app o al has been
pa o he li e a u e since i s mode n incep ion (Muelle , 1970), he ela i ely sho democ a ic elec o al
his o y o Po ugal has con ibu ed o lea ing ha hypo hesis unexamined. To add ess ha gap, we
in es iga e whe he he ela ionship be ween economic pe o mance and execu i e app o al in Po ugal
di e s depending on whe he he economy is expanding o con ac ing.
The second ele an con ex , we a gue, is
poli ical
and
ins i u ional
. In Po ugal’s semi-p esiden ial
sys em, p ime minis e s who head he execu i e coexis wi h popula ly elec ed and e m-limi ed p esiden s
who, while no sha ing go e ning esponsibili ies, hold subs an i e powe s and enjoy conside able public
isibili y. We sugges ha he uns able na u e o he poli ical ela ionship be ween p ime minis e s and
p esiden s in Po ugal shapes he economics-app o al link. To be su e, he li e a u e on execu i e app o al
in semi-p esiden ial sys ems has al eady sugges ed his is he case, ea ing pe iods o cohabi a ion —
when he p esiden and p ime minis e come om di e en pa ies o pa y coali ions — as causing a
swi ch o esponsibili y om one incumben o ano he , and he e o e a swi ch in which ac o o e s hold
accoun able o he economy (Hellwig and Samuels, 2008; Elgie, 2018). We a gue, howe e , ha Po ugal
displays di e en ea u es wi h di e en consequences. Wha ma e s in Po ugal’s b and o semi-
p esiden ialism is no cohabi a ion pe se bu a he whe he hese pe iods a e cha ac e ized by a
con lic ual ela ionship
be ween p ime minis e s and p esiden s. In such con lic ual con ex s, we
a gue, cla i y o esponsibili y o policy ou comes should be obscu ed, and he ela ionship be ween
economic pe o mance and he app o al o bo h en i ies weakened. This calls a en ion o he impo ance
o ins i u ional ac o s behind he "ins abili y dilemma" add essed in Ca lin
e al.,
2023.
Taken oge he , he wo a gumen s help us unde s and some o he puzzling ins abili y in he
ela ionship be ween he economy and app o al in Po ugal, hope ully also con ibu ing o he s udy o
public suppo o p ime minis e s (and p esiden s) in semi-p esiden ial sys ems simila o he Po uguese.
22
The chap e is s uc u ed as ollows. Chap e 3.2 discusses asymme y and semi-p esiden ialism
in he li e a u e on popula i y unc ions. Chap e s 3.3 and 3.4 desc ibe he da a and me hodology used
and he econome ic esul s. Chap e 3.5 concludes.
3.2. Public app o al and he economy in Po ugal
3.2.1. Economic Con ex s and Asymme y
The s udy o he ela ionship be ween economic pe o mance and execu i e app o al in Po ugal has
esul ed in ela i ely ew s udies bu wi h b oadly consis en esul s. The pionee ing wo k o Veiga (1998),
using mon hly app o al da a om 1986 o 1996, showed ha p ime minis e ial popula i y, besides
ecei ing a boos immedia ely a e he elec ion (a honeymoon e ec ) and declining wi h ime (cos o
uling), was nega i ely a ec ed by inc eases in he unemploymen a e. La e , e isi ing a mo e ex ended
ime se ies o simila da a, Veiga and Veiga (2004a and 2007) con i med he p e ious indings. O he
subsequen s udies using agg ega e da a, his ime ocusing on he o e unc ion, ha e con i med he
impo ance o economic pe o mance — pa icula ly o unemploymen and GDP g ow h — as d i e s o
incumben suppo (Veiga and Veiga, 2004b and 2010; F ei e and Lobo, 2005; Magalhães and Aguia -
Con a ia, 2009; Goula and Veiga, 2016).
Howe e , has economic pe o mance been equally consequen ial a all imes? Recen ly, Lobo
and Pannico (2020) showed ha "economic o ing" in Po ugal seems o ha e become s onge wi h he
economic c isis o he ea ly 2010s. Simila ly, Ca lin
e al.
(2023) show ha he ela ionship be ween
objec i e economic indica o s and execu i e app o al in Po ugal became s onge in he pe iod ha
encompasses he wo s economic c isis expe ienced by he coun y in decades. This sugges s ha c ises
may ende economic pe o mance mo e poli ically consequen ial. In ac , om a c oss-na ional
pe spec i e, Lewis-Beck and Nadeau (2012: 475) had al eady de ec ed ha economic a iables had a
g ea e impac on incumben suppo in Sou he n han in No he n Eu ope. They concluded ha "[ ]he
g ea e weigh o he economic o e in he PIGS [Po ugal, I aly, G eece, and Spain] coun ies appea s
almos en i ely due o hei poo e economic posi ion, coupled wi h hei less complex links be ween
go e nance and policy execu ion." In sum, he economy seems o "ma e mo e" in con ex s whe e
economic condi ions a e mo e ad e se.
Why should ha occu ? A well-known hypo hesis poin s o an
asymme y
in economic o ing
(Muelle , 1970; Nannes ad and Paldam, 1997; Bélange and Meguid, 2008; Dassonne ille and Lewis-
23
Beck, 2014). As E ans and Ande sen pu i , economic down u ns "elici sha ed and easonably pe cep i e
esponses ha a e no powe ully a ec ed by poli ical condi ioning" (2006: 195). Unde an economic
c isis, he in o ma ion con eyed by he mass media is mo e unequi ocally nega i e han i is posi i e
unde economic good imes (So oka, 2006), and he salience o economic issues inc eases, boos ing he
ele ance o economic pe o mance o go e nmen app o al (Singe , 2011). In con as , du ing pe iods
o g ow h, he economy loses impo ance, and he signals con eyed by he "media ed economy" become
less unequi ocal (E ans and Ande sen, 2006:195; see also Chzhen
e al.
, 2014 and Dicke son, 2016).
The 2011 and 2015 elec ions in Po ugal illus a e his phenomenon. In he o me , a majo economic
c isis led o la ge losses o he incumben cen e -le Socialis Pa y (PS). Howe e , in he la e , a clea
economic eco e y ailed o p e en an almos equally la ge punishmen o he uling cen e - igh
coali ion. Magalhães (2017) shows ha while ci izens' e alua ions o he economy we e undamen ally
exogenous in ela ion o poli ical p edisposi ions du ing he pe iod o economic c isis, in 2015, hey we e
powe ully con amina ed by pa isanship.
This sugges s ha he ins abili y in he o e unc ion in Po ugal may pa ially esul om how
buoyan o s agnan economic condi ions a e di e en ially ansmi ed o incumben suppo . To he
ex en ha hese gene al ideas abou he o e unc ion also apply o he popula i y unc ion, we should
see he app o al o he Po uguese p ime minis e su e ing mo e unde economic down u ns han i is
boos ed by economic eco e ies. In o he wo ds, we should expec he ela ionship be ween economic
pe o mance and p ime minis e ial app o al o be
asymme ic
.
3.2.2. Po uguese Semi-P esiden ialism and i s Implica ions
A second po en ial sou ce o ins abili y in he economics-app o al link conce ns he ins i u ional and
poli ical con ex unde which o e s e alua e incumben s. The idea is a om new. The "cla i y o
esponsibili y" a gumen is p edica ed on he no ion ha he ela ionship be ween economic pe o mance
and public suppo depends on he ex en o which he e is a "pe cei ed uni ied con ol o policymaking
by he incumben go e nmen " (Powell and Whi en, 1993: 398). In con ex s whe e such pe cei ed
con ol is lacking — cha ac e ized by g ea e ho izon al o e ical powe -sha ing — he economics-app o al
link should weaken. In such ci cums ances, i becomes mo e di icul o de e mine who should be
e ospec i ely blamed o ewa ded, and o e s should discoun economic pe o mance as a signal o
go e nmen compe ence (Duch and S e enson, 2008).
24
In he s udy o execu i e app o al in semi-p esiden ial sys ems —— whe e a popula ly elec ed
head o s a e coexis s wi h a p ime minis e accoun able be o e pa liamen (Elgie, 1999) — he no ion
ha con ex a ec s esponsibili y o he economy has also been p esen . Howe e , in o med by he
expe ience o Fi h Republic F ance, he main ocus o ha inqui y has been how semi-p esiden ialism
allows o wo al e na ing pa e ns o economic accoun abili y. Unde a uni ied go e nmen , when F ench
p esiden s a e suppo ed by a pa liamen a y majo i y o hei own pa y, hey become he leade s o ha
majo i y and hus o he execu i e, gi ing he sys em an unmis akable p esiden ial dynamic (Shuga ,
2005). In con as , unde "cohabi a ion," when p esiden s and p ime minis e s belonged o di e en
pa ies, powe shi s o he la e , c ea ing a "pa liamen a y" dynamic. The p esence o absence o
cohabi a ion was hough o de e mine an oscilla ion be ween phases when he sys em became
de ac o
,
espec i ely, pa liamen a y and p esiden ial (Du e ge , 1980; Lijpha , 2012).
This, in u n, gene a ed ano he oscilla ion: on whom o e s held accoun able o economic
pe o mance. In he p esiden ial phases, unde a uni ied go e nmen , he F ench p esiden 's app o al
was s ongly a ec ed by economic pe o mance. Howe e , in he pa liamen a y phases, unde
cohabi a ion, he esponsibili y o economic go e nance shi ed o he p ime minis e . Du ing hese
pe iods, he economy’s in luence on p esiden ial suppo d opped signi ican ly, while he opposi e
occu ed wi h p ime minis e ial app o al (Lewis-Beck, 1997; Lewis-Beck and Nadeau, 2004).
11
In 2002, he synch oniza ion o p esiden ial and legisla i e elec ions made cohabi a ion in F ance
much mo e unlikely, ending his oscilla ion. Howe e , he F ench expe ience s ill molds he hinking abou
economic accoun abili y in semi-p esiden ial sys ems, wi h s udies inding he same kind o oscilla ing
pa e n in coun ies o he han F ance. Looking a elec ion esul s in many semi-p esiden ial sys ems,
Hellwig and Samuels (2008) showed ha p esiden s a e ewa ded and punished o he economy only
unde condi ions o uni ied go e nmen , while he PM's pa y is only ewa ded and punished unde
condi ions o cohabi a ion. Elgie (2018), inding simila esul s, sugges ed he exis ence, in he s udy o
semi-p esiden ial sys ems, o "a basic scien i ic consensus abou he e ec o cohabi a ion on economic
o ing" (Elgie, 2018:103).
Howe e , we p opose ha his pa e n is unlikely o ex end o a semi-p esiden ial sys em like
Po ugal’s. Since 1986, all p esiden s in Po ugal ha e been p ominen poli ical igu es, all o me leade s
o hei pa ies. P esiden s enjoy impo an powe s, including he possibili y o e oing legisla ion o
e e ing bills o s a u es o cons i u ional e iew. P esiden s also employ "going public” ac ics, using
he au ho i y and p es ige o hei o ice o di ec public a en ion o he issues hey p io i ize (Ne o and
11
Bu see also Tu geon
e al.
(2015) showing ha only p ime minis e ial app o al is a ec ed by economic a iables, and only du ing cohabi a ion.

25
Lobo, 2009; Jalali, 2011). And hey e en ha e he (mos ly) uncons ained possibili y o dissol ing
pa liamen in he absence o a p ime minis e ial esigna ion o a success ul mo ion o censu e. This is a
powe hey ha e used wi h pa simony in he las ou decades, bu which ne e heless has been used
and s ands as a ele an shadow o e go e nmen s.
Ye , while ne e becoming pu ely pa liamen a y, he Po uguese sys em's dynamic has ne e
become p esiden ial ei he ,
de ju e
o
de ac o
. Po uguese p esiden s a e no he leade s o hei pa ies.
The head o he execu i e is always he p ime minis e , i.e., he pa y and pa liamen a y majo i y leade
wi h whom p esiden s coexis . Acco dingly, p esiden ial elec ions in Po ugal bea some ai s o a
"second-o de " elec ion (Rei and Schmi , 1980), mos no ably hei lowe le el o u nou in compa ison
wi h legisla i e elec ions (Fo es and Magalhães, 2005; F anco, 2020). The implica ion is ha , a leas in
Po ugal, he e has been no oscilla ion be ween pa liamen a y and p esiden ial phases depending on
cohabi a ion o uni ied go e nmen . This, in u n, sugges s ha an oscilla ion in how o e s hold p ime
minis e s and p esiden s accoun able o he economy should be absen as well.
Howe e , his does no mean ha he ela ionship be ween p esiden s and p ime minis e s is
i ele an o he ole o he economy in he V-P unc ion. Al hough i does no change who uns he
e e yday ope a ion o go e nmen , cohabi a ion has ne e heless been a sou ce o
poli ical con lic
be ween p esiden s and p ime minis e s unde semi-p esiden ialism (P o syk, 2005a and 2005b; Elgie,
2010). The exis ing li e a u e p o ides clea e idence abou he condi ions ha ha e a o ed such
con lic ual ela ionship in Po ugal. Po uguese p esiden s can only se e wo consecu i e e ms o o ice.
Fi s - e m incumben s who seek eelec ion o a second e m in o ice — so a , all o hem — ha e
incen i es o p ese e a non-con lic ual ela ionship wi h p ime minis e s and hei pa liamen a y
majo i ies in o de o secu e he equi ed absolu e majo i y o he o e in hei eelec ion bid. This has
occu ed e en in con ex s o cohabi a ion. Since 1986, p esiden s Má io Soa es, Jo ge Sampaio, Ca aco
Sil a, and Ma celo Rebelo de Sousa success ully a oided majo con on a ions wi h p ime minis e s du ing
hei i s e m, e en when hey belonged o di e en pa ies. By de using he po en ial o con lic , some
incumben p esiden s ha e e en been able o secu e opposing go e ning pa ies' explici o implici
suppo in hei eelec ion bids, as in he case o he Socialis Má io Soa es in 1991 o PSD’s Ma celo
Rebelo de Sousa in 2021.
Howe e , in hei second e m, he combina ion o cohabi a ion and he lack o a eelec ion
incen i e has ypically led p esiden s o engage in a mo e con lic ual ela ionship wi h p ime minis e s. As
Jalali (2011: 170) pu s i , while p esiden s' i s e ms a e cha ac e ized by a "conside ably weake
ac i ism (…), no leas mo i a ed by he need o main ain a majo i y o eelec ion," second e ms ypically
26
show a ise in “legisla i e in e en ionism” and “gue illa wa a e” on he pa o p esiden s. In ac ,
second e ms ha e only been his o ically peace ul when p esiden s ace a go e nmen majo i y cong uen
wi h hem in pa isan e ms (F ei e and San ana-Pe ei a, 2019). Howe e , unde cohabi a ion, p esiden s
in hei second and las e m o o ice ha e used hei e o and cons i u ional e e al powe s mo e
equen ly and ha e engaged in mo e open poli ical con on a ions wi h he go e nmen (Magalhães,
2001; Ne o and Lobo, 2009; Jalali, 2011; F anco 2020).
This sugges s ha , by a o ing in e -ins i u ional con lic be ween p ime minis e s and p esiden s,
second- e m cohabi a ions
should con ibu e o blu ing accoun abili y o he economy. By os e ing
“legisla i e in e en ionism” on he pa o p esiden s and “cons aining he go e nmen s’ capaci y o
poli ical leade ship” (Jalali, 2011: 169), second- e m cohabi a ions a e pe iods when wha Powell and
Whi en de ined as "cla i y o esponsibili y” — he "pe cei ed uni ied con ol o policymaking by he
incumben go e nmen " (1993: 398) — is unde mined. This leads us o expec ha he ela ionship
be ween economic pe o mance and p ime minis e ial app o al, as we hypo hesized ea lie , should be
weake when he p esiden is o a di e en pa y and is se ing in hei second and las e m o o ice.
In summa y, we p opose wo main hypo heses. H1 is ha he ela ionship be ween economic
pe o mance and p ime minis e ial app o al should be con ingen upon he economic con ex i sel :
s onge unde c ises and weake unde expansions (i.e., asymme ic). H2 is ha i should also be
con ingen upon he ela ionship be ween p ime minis e s and p esiden s: s onge unde peace ul
coexis ence, weake unde con ex s ha ha e a o ed poli ical con lic s, i.e., du ing (p esiden ial) second-
e m cohabi a ions.
3.3. Da a, Va iables, and Es ima ion S a egy
3.3.1. Da a and Va iables
Ou da ase encompasses wo measu es o app o al and a se o economic and poli ical a iables. We
ex ac ed wo app o al a ings om he Execu i e App o al P ojec da ase (Ca lin
e al.
, 2019). Ou
dependen a iables a e smoo hed e sions o he qua e ly app o al a ings o he Po uguese p esiden
and p ime minis e (
P esiden ial App o al Ra ing
and
PM App o al Ra ing
). The da a a ailable in he
Execu i e App o al Da abase 2.0 goes om he second qua e o 1986 o he hi d qua e o 2018.
Table 5 lis s he p ime minis e s and p esiden s h oughou his pe iod and indica es whe he
p esiden s we e se ing hei i s o second e m and whe he hey and he p ime minis e s belonged o
27
he same o di e en pa ies. The la e wo aspec s a e c ucial o es ou second hypo hesis and show
conside able di e si y h ough ime.
Po ugal has had i e p esiden s since he p omulga ion o he cu en 1976 democ a ic
cons i u ion. Since he app o al se ies began in 1986, he i s p esiden , Ramalho Eanes, is excluded
om ou analysis. The las we conside is Ma celo Rebelo de Sousa, who was elec ed in 2016. A he
ime o his w i ing, all p esiden s spen en yea s in o ice,
12
se ing wo consecu i e e ms o i e yea s.
Rega ding he execu i e b anch, headed by he p ime minis e , se e al pe sonali ies ha e been in powe
since 1986. Only wo pa ies occupied he p ime minis e ial pos du ing his pe iod: he PS and PSD.
Ano he pa y, CDS-PP, has in eg a ed some o PSD's go e nmen s o achie e a pa liamen a y majo i y.
Table 5: Po uguese P esiden s and P ime Minis e s be ween 1986 and 2018.
Figu e 5 shows he qua e ly app o al a es o Po uguese p ime minis e s (black line) and
p esiden s (g ey line). Shaded a eas ep esen pe iods o cohabi a ion, ei he du ing he p esiden ’s i s
12
The excep ion is he cu en p esiden , Ma celo Rebelo de Sousa, who was elec ed in 2016 and eelec ed in 2021.
Pe iod
P ime Minis e
(Pa y)
P esiden (Pa y)
P esiden ial
Te m
Cohabi a ion
1986-1991
Ca aco Sil a (PSD)
Má io Soa es (PS)
Fi s
Yes
1991-1995
Ca aco Sil a (PSD)
Má io Soa es (PS)
Second
Yes
1995-1996
An ónio Gu e es (PS)
Má io Soa es (PS)
Second
No
1996-2001
An ónio Gu e es (PS)
Jo ge Sampaio (PS)
Fi s
No
2001-2002
An ónio Gu e es (PS)
Jo ge Sampaio (PS)
Second
No
2002-2004
Du ão Ba oso (PSD)
Jo ge Sampaio (PS)
Second
Yes
2004-2005
Ped o San ana Lopes
(PSD)
Jo ge Sampaio (PS)
Second
Yes
2005-2006
José Sóc a es (PS)
Jo ge Sampaio (PS)
Second
No
2006-2011
José Sóc a es (PS)
Ca aco Sil a (PSD)
Fi s
Yes
2011
José Sóc a es (PS)
Ca aco Sil a (PSD)
Second
Yes
2011-2015
Ped o Passos Coelho
(PSD)
Ca aco Sil a (PSD)
Second
No
2015-2016
An ónio Cos a (PS)
Ca aco Sil a (PSD)
Second
Yes
2016-2018
An ónio Cos a (PS)
Ma celo Rebelo de Sousa
(PSD)
Fi s
Yes
28
(ligh -g ey) o second e m (da k-g ey). The da a co e s ou p esidencies and en p ime minis e ial
cabine s. Wi h b ie excep ions in he ea ly 1990s and la e 2015, Po uguese p esiden s ha e been mo e
popula han p ime minis e s h oughou he en i e pe iod. While isual inspec ion sugges s he wo se ies
a e ela ed, he ela ionship is clea ly impe ec , and he e a e e en pe iods whe e ends seem o di e ge,
sugges ing ha hey may be d i en by somewha di e en p ocesses. Table 6 p o ides desc ip i e
s a is ics o bo h he P ime Minis e ial and P esiden ial app o al a ings.
Table 6: Desc ip i e s a is ics o he a iables employed.
Obse a ions
Mean
S d De
Min
Max
PM App o al Ra ing
130
39.80
10.41
18.54
66.13
P esiden ial App o al Ra ing
130
58.59
13.06
32.18
82.35
GDP G ow h
130
2.20
2.83
-4.53
8.71
GDP G ow h Nega i e
130
0.40
1.03
0.00
4.53
GDP G ow h Posi i e
130
2.60
2.21
0.00
8.71
Honeymoon PM
130
0.21
0.55
0.00
2.00
Honeymoon P esiden
130
0.16
0.49
0.00
2.00
Cohabi a ion 1s e m
130
0.38
0.49
0.00
1.00
Cohabi a ion 2nd e m
130
0.25
0.43
0.00
1.00
Uni ied go e nmen
130
0.37
0.48
0.00
1.00
Figu e 5: P ime Minis e ial and P esiden ial App o al Ra ing, qua e ly da a.
No e: Da a om Ca lin e al. (2019).
35
Table 8: P esiden ial and p ime minis e ial app o al, he economy, and cohabi a ion.
While Models 2 and 3 showed ha
GDP G ow h Nega i e -1
was ha m ul o he P ime Minis e ,
Model 4 shows ha such nega i e alue esul s om he agg ega ion o h ee coe icien s o h ee
di e en con ex s.
Cohabi a ion 1s e m
*
GDP G ow h Nega i e -1
and
Uni ied go e nmen * GDP G ow h
Model 3
Model 4
PM App o al
P esiden ial
App o al
PM App o al
P esiden ial
App o al
PM App o al Ra ing -1
0.769***
(0.038)
-
0.771***
(0.038)
-
P esiden ial App o al Ra ing -1
-
0.769***
(0.054)
-
0.773***
(0.053)
Honeymoon PM
3.038**
(0.920)
3.073**
(0.909)
-
Honeymoon P esiden
_
1.638
(1.869)
-
1.634
(1.868)
Cohabi a ion 1s e m
1.940†
(1.037)
3.337*
(1.483)
1.998†
(1.109)
3.476†
(1.796)
Cohabi a ion 2nd e m
0.069
(1.057)
1.675
(1.372)
-0.356
(1.095)
1.889
(1.558)
GDP G ow h Nega i e -1
-1.066***
(0.237)
-
-
-
Cohabi a ion 1s e m* GDP G ow h Nega i e -1
-
-
-1.331**
(0.449)
-
Cohabi a ion 2nd e m * GDP G ow h Nega i e -
1
-
-
0.511
(0.819)
-
Uni ied go e nmen * GDP G ow h Nega i e -1
-
-
-1.138***
(0.272)
-
GDP G ow h -1
-
0.596**
(0.184)
Cohabi a ion 1s e m* GDP G ow h -1
-
-
-
0.566*
(0.264)
Cohabi a ion 2nd e m* GDP G ow h -1
-
-
-
0.442
(0.664)
Uni ied go e nmen * GDP G ow h -1
-
-
-
0.651**
(0.250)
Cons an
8.112***
(1.558)
10.529***
(2.815)
8.044***
(1.591)
10.250***
(2.676)
Va (e.PM App o al)
19.589
(3.108)
19.046
(3.114)
Va (e.P es App o al)
34.984
(6.240)
35.091
(6.232)
Co (e.PM App o al,
e.P es App o al)
13.899
(3.074)
13.567
(3.055)
Obse a ions
129
129
SRMR
0.013
0.013
No e: S anda d-e o s be ween b acke s. S a is ical signi icance: †p<.10; *p<.05; **p<.01; ***p<.001
( wo- ailed es s).

36
Nega i e -1
a e bo h nega i e, signi ican , and simila ly sized. In o he wo ds, economic pe o mance
a ec s he popula i y o p ime minis e s unde condi ions o i s - e m cohabi a ion o uni ied go e nmen ,
and i does so asymme ically, as posed in H1. In con as , he coe icien o he in e ac ion
Cohabi a ion
2nd e m
*
GDP G ow h Nega i e
-1 o p ime minis e s is a om s a is ical signi icance a con en ional
le els. In o he wo ds, unde condi ions ha ha e been cha ac e ized by a mo e con lic ual ela ionship
be ween p ime minis e s and p esiden s, p ime minis e ial app o al becomes insensi i e o economic
pe o mance.
Simila ly, Models 2 and 3 showed he coe icien
GDP G ow h -1
is posi i e o he p esiden .
Model 4 shows hese esul s om
Cohabi a ion 1s e m
*
GDP G ow h -1
and
Uni ied go e nmen * GDP
G ow h -1
, whose coe icien s a e posi i e, signi ican , and simila ly sized: unde condi ions ha a o
peace ul coexis ence wi h p ime minis e s, p esiden ial app o al is (symme ically) a ec ed by economic
pe o mance. Howe e , like wha occu s wi h p ime minis e ial app o al, he in e ac ion
Cohabi a ion 2nd
e m
*
GDP G ow h
-1 o p esiden s is no s a is ically signi ican .
21
In sum, con ex ual condi ions ha combine a lack o eelec ion incen i es o p esiden s and pa isan
di e gence be ween hem and p ime minis e s ha e been inimical o economic accoun abili y: in e -
ins i u ional con lic be ween p esiden s and p ime minis e s makes he popula i y o bo h en i ies less
sensi i e o economic pe o mance, pa icula ly in he case o he la e .
3.5. Conclusion
I has long been suspec ed ha suppo o incumben go e nmen s and pa ies in semi-p esiden ial
sys ems migh be cha ac e ized by somewha di e en dynamics om hose obse ed in pu e
pa liamen a y o p esiden ial sys ems. In he well-s udied F ench case, cohabi a ion seemed o swi ch
esponsibili y o he economy om he p esiden o he p ime minis e , as he leade ship o he execu i e
mo ed om he o me o he la e (Lewis-Beck, 1997). O he s udies ha e e en sugges ed his also
occu s wi h semi-p esiden ial sys ems mo e b oadly, wi h cohabi a ion se ing as he swi ch ha makes
p esiden s (unde he o me ) o p ime minis e s and hei pa ies (unde he la e ) publicly accoun able
o he economy.
21
Howe e , i has he same sign as he o he in e ac ions, sugges ing he con as be ween 2nd e m cohabi a ions and he emaining si ua ions is mo e obus
o he case o he p ime minis e han o he case o he p esiden .
37
We sugges , howe e , ha his dynamic has no been ep oduced in he case o semi-p esiden ial
Po ugal, o a leas no since he mid-1980s. Because cohabi a ion in Po ugal does no swi ch he locus
o esponsibili y o economic policy, i also does no swi ch he locus o esponsibili y o economic
pe o mance om he public’s poin o iew ei he . Vo e s s ongly punish p ime minis e s o bad
economic pe o mance unde ei he cohabi a ion o uni ied go e nmen . The e is, howe e , one c ucial
excep ion:
con lic ual cohabi a ions
, such as hose ha ha e p omp ed p esiden s o make mo e ex ensi e
uses o hei legisla i e and o e sigh powe s o place impo an obs acles o go e nmen s' policy agenda
du ing p esiden ial second (and las ) e ms. Unde such condi ions, public app o al o
bo h
en i ies,
p ime minis e s and p esiden s, has been mos ly insensi i e o economic pe o mance. These indings
a e compa ible wi h he no ion ha con lic s be ween poli ically ele an p esiden s and he p ime
minis e s hey oppose con ibu e o blu ing esponsibili y o he economy. By doing so, he s udy indings
poin o a sou ce o ins abili y in he Po uguese V-P unc ion ha , while oo ed in ins i u ional ac o s —
semi-p esiden ialism, p esiden ial powe s, e m-limi s — is a om s a ic: i depends on he poli ical
con ex shaping he ela ionship be ween p ime minis e s, p esiden s, and hei pa ies, a con ex ha
has a ied signi ican ly h oughou he pe iod unde examina ion (see Table 5).
Fu he mo e, we also show ha al hough he app o al o bo h p ime minis e s and p esiden s
is a ec ed by he economy ou side o second- e m cohabi a ions, ha ela ionship is asymme ic o p ime
minis e s and symme ic o p esiden s. Fu u e s udies may be able o es ablish wi h g ea e ce ain y he
easons o ha di e ence. Howe e , we specula e ha i is consis en wi h he di e en oles played by
bo h en i ies in he poli ical sys em. As pa isan en i ies in cha ge o he execu i e, p ime minis e s a e
a ec ed by he economy in a way ha is compa ible wi h he lowe pa isan con amina ion o economic
e alua ions in bad imes (E ans and Ande sen, 2006; see Magalhães, 2017, o Po ugal), i.e.,
asymme ically. Con e sely, he ela ionship be ween he economy and p esiden ial app o al is less
con amina ed by such p ocess, gi en he lowe ele ance o pa isanship as a cue wi h which o e alua e
p esiden s. The asymme y ound o p ime minis e ial app o al add esses pa o he puzzle conce ning
he appa en ins abili y o he V-P unc ion in he coun y (F ei e and San ana-Pe ei a, 2012; Ca lin
e al.,
2023). In pa icula , i helps accoun ing o he e y s ong connec ion be ween economic pe o mance
and p ime minis e ial app o al du ing Po ugal’s wo s economic imes since he 1970s.
Mo e b oadly, his chap e makes wo main con ibu ions. Fi s , a clea e unde s anding o some
o he easons why, in he case o Po ugal, he link be ween economic pe o mance and execu i e suppo
appea s uns able i app oached in a way ha igno es he a iabili y in oduced by economic and
ins i u ional con ex s. A second con ibu ion consis s o po en ially impo an clues o be explo ed in he
38
s udy o semi-p esiden ial sys ems ha bea simila i ies o he Po uguese. In coun ies such as Bulga ia,
Poland, Li huania, Romania, Slo akia, o Slo enia, o example, popula ly elec ed p esiden s a e nei he
like he F ench heads o he execu i e no like he mos ly ce emonial igu eheads ha can be ound in
he semi-p esiden ialism o Aus ia, I eland, o Iceland. Thei powe s a e o en ele an , and hei
cohabi a ion wi h p ime minis e s who belong o di e en pa ies has been cha ac e ized by peace ul
coexis ence in some cases bu con lic and e en policy s alema es in o he s (P o syk, 2005a and 2005b;
Elgie, 2010). The Po uguese case p o ides a ew es able hypo heses abou how economic pe o mance
should a ec he app o al o bo h p ime minis e s and p esiden s in such con ex s.
39
4. US P esiden ial App o al Ra ing: Time-F equency App oach
22
4.1. In oduc ion
One o he challenges posed by he s udy o popula i y unc ions has been he di icul y in dis inguishing
be ween s uc u al and ansien e ec s on go e nmen app o al o di e en economic a iables while a
he same ime add essing issues o non-s a iona i y and non-linea i y. One o he consequences o his
di icul y has been a ple ho a o inconsis en indings, e en in he case whe e he ela ionship be ween
economic ac o s and go e nmen app o al has been mos exhaus i ely s udied: he Uni ed S a es
(Be lemann and Enkelmann, 2014).
In his wo k, we show how wa ele analysis can be employed in conjunc ion wi h o he
econome ic me hods o add ess hese p oblems. We desc ibe he ad an ages o wa ele ools,
pa icula ly how hey allow us o decompose ime se ies da a in o a sum o wa es o di e en equencies,
p o ide local analyses una ec ed by non-s a iona i y and non-linea i y p oblems, and, wi h ad ancemen s
(Aguia -Con a ia and Soa es, 2014; Aguia -Con a ia
e al.
, 2023), pe o m mul i a ia e analysis. To show
how he use o mul i a ia e wa ele analysis allows o a mo e nuanced unde s anding o he ela ionship
be ween he economy and p esiden ial app o al, we illus a e his app oach, ocusing on he app o al
a ings o P esiden Ba ack Obama and compa ing he esul s wi h OLS es ima ions. To analyze he
ela ionship be ween he economy and he app o al, we u ilized he con en ional economic a iables:
unemploymen and in la ion a es. We also used consume sen imen as a measu e o economic
pe cep ion. As ou ocus is P esiden Obama’s p esidency, ou da a goes om Janua y 2009 o Decembe
2016.
The esul s e eal how Obama’s app o al a ing was s uc u ally in luenced by economic a iables
such as unemploymen and in la ion, wi h di e en synch oniza ions, bu also how hey we e ansien ly
ela ed o consume sen imen , in a esul ha would no eme ge om adi ional econome ic models.
22
A e sion o his chap e has al eady been submi ed o a academic jou nal.
40
The s udy is o ganized as ollows. In Chap e s 4.2 and 4.3, we p o ide a sel -con ained summa y
o wa ele analysis, desc ibing he ad an ages o he ools we use and why hey a e ele an o he s udy
o popula i y unc ions and complemen ing wi h an example. In Chap e 4.4, we desc ibe he da a ha
will be employed o illus a e he po en ial o wa ele analysis o his opic. Chap e 4.5 p esen s he
esul s o bo h a adi ional OLS app oach and wa ele analysis, highligh ing he added alue o he la e .
Chap e 4.6 concludes.
4.2. Wa ele analysis and he s udy o popula i y unc ions
4.2.1. The challenge o ins abili y
E e since Muelle (1970), a p o usion o s udies has empi ically in es iga ed how he economy a ec s
p esiden ial app o al in he Uni ed S a es. While se e al indica o s ha e been used as p oxies o
assessing economic condi ions, unemploymen and in la ion a es ha e been among he mos employed
in he V-P unc ion li e a u e since he 1970s. Vo e s a e assumed o ecognize ha he s a e o he
economy is un a o able when unemploymen is high, and ha high in la ion ac s as a bu den on hei
income, and hey should use ha in o ma ion o assess he pe o mance o incumben s.
Howe e , con a y o ini ial expec a ions, he co ela ion be ween economic pe o mance and
execu i e popula i y has been less han consis en . This ins abili y can be a ibu ed o many ac o s. One
possibili y is he dispa a e eac ions elici ed by le - and igh -wing incumben s. Fo example, le -leaning
execu i es may expe ience in ensi ied c i icism du ing pe iods o heigh ened unemploymen , whe eas
hei igh -leaning coun e pa s ace ampli ied sc u iny amids in la ion (Hibbs, 1977; Powell and Whi en,
1993; Swank, 1998). Mo eo e , he esponses o economic shocks a e no symme ic. As So oka (2006)
no ed, ad e se economic de elopmen s ypically esul in a sha pe decline in popula i y compa ed o he
mild up ick associa ed wi h posi i e economic su p ises.
Howe e , ye ano he possibili y is ha hese e ec s may be mo e o less ansi o y. Fo example,
e en s may lead o e s o se aside economic and poli ical p e e ences empo a ily. In he V-P unc ion
li e a u e, e en s such as wa s o e o is a acks elici a well-known phenomenon e e ed o as he “ ally
‘ ound he lag” e ec (Muelle , 1970; Newman and Fo cehimes, 2010), whe ein indi iduals uni e in
suppo o hei poli ical leade ship, blu ing he ela ion be ween economic undamen als and app o al
a ings. Fo ins ance, despi e he Ame ican economy acing an economic ecession in 2001, P esiden
Geo ge W. Bush’s popula i y expe ienced a signi ican su ge ollowing he 9/11 e o is a ack.

41
Mo e gene ally, as G onke and Newman (2003) highligh , p esiden ial pe o mance e alua ions
a e likely o be inhe en ly uns able. Based on he media p iming li e a u e, hey a gue ha media
emphasis on speci ic issues shapes he c i e ia ci izens employ o e alua e p esiden ial pe o mance. I
public expec a ions o p esiden s a e dynamic and e ol e o e ime, a single s a is ical model canno
uni e sally cap u e he d i e s o app o al a ings. Acco dingly, hey ad oca e o he de elopmen o new
es ima ion me hods ha accommoda e ime- a ying coe icien s. Simila ly, Be lemann and Enkelmann
(2014), while sugges ing ha in la ion and unemploymen seem o exe a long- e m in luence on
p esiden ial app o al in he U.S., show inconsis encies in sho - e m e ec s, a ibu ing hem as well o a
po en ially e ol ing ela ionship be ween economic ac o s and popula i y. Fu he mo e, hey also
ques ion bo h he p e alen use o linea es ima ion in p e ious s udies, p oposing ha nonlinea
ela ionship migh exis , as well as he ac ha many s udies ha e no adequa ely conside ed s a iona i y
issues. In ac , Be lemann
e al.
(2015) and Choi
e al.
(2016) ind an in e dependen and non-linea
impac o he unemploymen and in la ion a es on US p esiden ial popula i y.
These di e en insigh s se e as an in i a ion o ake ad an age o he possibili ies awa ded by
wa ele analysis. E e since Fou ie 's ime, we ha e known ha many ime se ies can be exp essed as a
sum o pe iodic cycles. This has laid he ounda ion o me hods like he Fou ie ans o m and wa ele
ans o m, which enable us o decompose ime se ies da a in o a sum o sinusoidal wa es o a ying
equencies, o e ing insigh s in o he da a's pe iodic pa e ns. The wa ele ans o m has, howe e , a
c ucial ad an age o e he Fou ie ans o m: i allows o pa e ns in he da a and in he ela ionships
be ween a iables o be
ime- a ying
. Acco dingly, as a
local
analysis, i assumes nei he a ime-in a ian
gene a ing p ocess (s a iona i y) no ha ela ionships be ween a iables a e linea . In he ollowing
Subchap e s, we unde sco e he bene i s o employing con inuous wa ele analysis in conjunc ion wi h
adi ional econome ic me hods o add ess he challenge posed by ins abili y.
4.2.2. The (pa ial) wa ele cohe ency, gain, and phase-di e ence
Ea lie s udies ha e al eady delinea ed he me i s o wa ele analysis o poli ical science (Aguia -Con a ia
e al.
, 2012 and 2013). Mo e ecen ly, we ind some applica ions o he bi a ia e analysis in poli ical
science (Wang, 2023a and 2023b). Aguia -Con a ia
e al.
(2014) wen a s ep u he by allowing an
analysis in ol ing mo e han wo se ies – mul i a ia e wa ele analysis. This me hodology has been
applied in di e en opics, such as mac oeconomics and, o a less ex en , in poli ical science (Wang,
42
2023a). Recen ly, some s udies ha e es ima ed equa ions in he ime- equency domain. Aguia -Con a ia
e al.
(2023) p o ide a ecen applica ion, hough i s u iliza ion in poli ical science emains unexplo ed.
The s udies p ima ily u ilize h ee wa ele ools: he wa ele powe spec um, (pa ial) wa ele
cohe ency, and (pa ial) wa ele phase-di e ence. The wa ele powe spec um es ima es a a iable's
a iance in he ime- equency domain, pinpoin ing he con ibu ion o speci ic equencies o o al
a iance and iden i ying signi ican cycles. Wa ele cohe ency, akin o he absolu e alue o he
co ela ion coe icien , es ima es he s eng h o he associa ion be ween wo a iables ac oss di e en
imes and equencies, e ealing a ying co ela ions. Phase-di e ence in o ms he
sequence/synch oniza ion and na u e (posi i e/nega i e) o a iable co ela ions. The ex ensions, pa ial
wa ele cohe ency, akin o he pa ial co ela ion coe icien , and pa ial phase-di e ence, p o ide
he same in o ma ion while accoun ing o hi d a iables. Finally, he mul iple wa ele cohe ency: i s
in e p e a ion is simila o he R-squa ed bu , o cou se, done in he ime- equency domain.
In pa icula , he pa ial phase-di e ence allows us o dis inguish be ween posi i e and nega i e
co ela ion, a e aking in o accoun he e ec s o a hi d a iable, and i also in o ms us on which a iable
is leading which. Le 𝜙𝑥,𝑦|𝑧 be he phase-di e ence be ween 𝑌𝑡 and 𝑋𝑡 a e con olling o 𝑍𝑡. I
𝜙𝑦,𝑥|𝑧 =0, hen he wo a iables’ se ies mo e oge he a a gi en ime and equency. I
𝜙𝑦,𝑥|𝑧 є (0,𝜋
2) hen hey a e in phase (posi i e co ela ion) wi h 𝑌𝑡 leading 𝑋𝑡, and i 𝜙𝑦,𝑥|𝑧 є (−𝜋
2,0),
he se ies a e in phase, wi h 𝑋𝑡 leading. On he o he hand, i 𝜙𝑦,𝑥|𝑧 є (−𝜋,−𝜋
2), se ies 𝑌𝑡 leads 𝑋𝑡
in an ou o phase ela ion (nega i e co ela ion), and i 𝜙𝑦,𝑥|𝑧 є (𝜋
2,𝜋) he se ies 𝑌𝑡 and 𝑋𝑡 a e ou o
phase, wi h 𝑋𝑡 leading. Figu e 7 summa izes his in o ma ion in a ci cle scheme.
Figu e 7: Phase-Di e ence ci cle.
No e: Aguia -Con a ia and Soa es (2014).
43
Al hough highly bene icial, he wa ele ools desc ibed p e iously could no pe o m eg essions in
he ime- equency domain. This limi a ion was add essed by Aguia -Con a ia
e al.
(2023) who ex ended
he concep o wa ele gain, p oposed by Mandle and Scha nagl (2014), o a mul i a ia e con ex . They
demons a ed ha he pa ial wa ele gain o 𝑌𝑡 o e 𝑋𝑡, while con olling o 𝑍𝑡, and 𝑌𝑡 o e 𝑍𝑡,
while con olling o 𝑋𝑡, can be in e p e ed as he coe icien in he linea eg ession o 𝑌𝑡 on he
explana o y a iables 𝑋𝑡 and 𝑍𝑡 a each speci ic ime and equency.
In he ollowing Chap e , we p o ide an in ui i e explana ion, illus a ed by a speci ic example, o
in e p e ing g aphical ep esen a ions and ex ac ing pe inen in o ma ion ela ed o pa ial cohe ency,
pa ial phase-di e ence, and pa ial wa ele gain. Fo he o mulae and igo ous ma hema ical ea men ,
he eade is di ec ed o Aguia -Con a ia
e al.
(2012) and Aguia -Con a ia and Soa es (2014).
4.3. An example
4.3.1. The Da a Gene a ing P ocess
Suppose a a iable 𝑋𝑡 is p ima ily composed o wo pe iodic cycles: a one-yea cycle and a h ee-yea
cycle, along wi h some whi e noise. Assume we ha e 10 yea s o mon hly da a. In his case, he exp ession
o 𝑋𝑡 can be w i en as:
𝑋𝑡=cos(2𝜋𝑡
1)+cos(2𝜋𝑡
3)+𝜀𝑥,𝑡, 𝑓𝑜𝑟 𝑡=0, 1
12,2
12,…,10.
23
Assuming s ill ha we ha e mon hly da a spanning 10 yea s, conside ano he a iable, 𝑍𝑡, which
comp ises bo h whi e noise and a six-mon h cycle, exp essed as:
𝑍𝑡=cos(2𝜋𝑡
1 2
⁄)+𝜀𝑧,𝑡, 𝑓𝑜𝑟 𝑡=0, 1
12,2
12,…,10.
Finally, conside a a iable 𝑌𝑡 ha is a unc ion o he p e ious wo. Howe e , o ully app ecia e
he powe o wa ele es ima ion, le 's in oduce some complexi ies. Fi s , he ela ionship be ween 𝑋𝑡 and
𝑌𝑡 is no cons an ei he ac oss ime o equencies. Fo example, since i is en i ely possible o one
23
The use o cosines migh ini ially appea unusual; howe e , Aguia -Con a ia
e al.
(2012) demons a e ha se e al au o eg essi e p ocesses, as es ima ed
in he poli ical science li e a u e, can indeed be cha ac e ized in his manne .
44
a iable o ha e a posi i e e ec on ano he in he sho e m and dele e ious e ec s in he long e m, we
assume ha he ela ionship be ween 𝑋𝑡 and 𝑌𝑡 is posi i e a highe equencies (co esponding o he
one-yea cycle in ou example), while nega i e a lowe equencies ( he h ee-yea cycle).
Second, o encapsula e ime- a ying ela ionships, we modi y he eg ession coe icien ha
connec s he one-yea cycle be ween 𝑋𝑡 and 𝑌𝑡. I is se a 1 o he i s i e yea s o da a and hen
adjus s o 3 o he emaining pe iod. Las ly, o add ess ansien ela ionships, we p opose ha he
in luence o 𝑍𝑡 on 𝑌𝑡 is empo a y, las ing only o one yea , speci ically when 2≤𝑡<3.
Pu ing all his in o ma ion oge he , we ha e ha
𝑌𝑡=𝛼1,𝑡cos(2𝜋𝑡
1 2
⁄)+𝛼2,𝑡cos(2𝜋(𝑡−1 12
⁄ )
1)+𝛼3,𝑡cos(2𝜋(𝑡−3 12
⁄ )
3)
+𝜀𝑦,𝑡 𝑓𝑜𝑟 𝑡=0, 1
12,2
12,…,10.
Wi h 𝛼1,𝑡 =1 i 2≤𝑡<3 and ze o, o he wise; 𝛼2,𝑡 =1 i 𝑡≤5 and 𝛼2,𝑡 =3 i 5<𝑡≤10;
𝛼3,𝑡 =−2. We' e added ano he wis o conside : a highe equencies, he e is a one-mon h lag o
he impac o 𝑋𝑡 on 𝑌𝑡, and a he h ee-yea equency, he lag ex ends o h ee mon hs. Howe e , he
in luence o 𝑍𝑡 on 𝑌𝑡 is ins an aneous.
4.3.2. Reading he esul s
The esul s o he wa ele analysis o his example a e p esen ed in Figu e 8. On he op, we ha e he
wa ele mul iple cohe ency (cha a). S ill on he le , we ha e he pa ial wa ele cohe ency be ween 𝑌𝑡
and 𝑋𝑡, a e con olling o 𝑍𝑡 (cha c.i), and be ween 𝑌𝑡 and 𝑍𝑡, a e con olling o 𝑋𝑡 (cha d.i). The
colo code o cohe ency (pa ial o mul iple) anges om blue (low cohe ency – close o ze o) o ed
(high cohe ency – close o one). The black con ou s indica e he 5% signi icance le el o he cohe ency
compu ed wi h 5000 Mon e-Ca lo simula ions. A pa abola-like black line indica es he egion a ec ed by
edge e ec s. Resul s should be in e p e ed wi h ca e in his a ea.
24
24
As wi h o he ypes o ans o ms, he Con inuous Wa ele T ans o m applied o a ini e leng h ime se ies su e s om bo de dis o ions due o he ac
ha he alues o he ans o m a he beginning and he end o he ime se ies in ol e missing alues o he se ies, which a e a i icially p esc ibed. One
usually pads he se ies wi h ze os o a oid w apping. These edge-e ec s inc ease o lowe equencies. In his a ea o he ime- equency plane, he esul s
a e subjec o bo de dis o ions and ha e o be in e p e ed ca e ully.
51
unemploymen and in la ion. Then, we expand he model o include consume sen imen . The models
we e es ima ed using he O dina y Leas Squa es.
We i s es ima e a s a ic model. Gi en ha ou a iables do no exhibi a uni oo ,
26
we canno
say ha i co esponds o a coin eg a ion ela ion, bu i s ill cap u es he idea o a long- un ela ion:
𝑂𝑏𝑎𝑚𝑎 𝐴𝑝𝑝𝑟𝑜𝑣𝑎𝑙𝑡=𝛼+ 𝛾 𝑿′𝑡−1+𝛿 𝒁′𝑡+𝜀𝑡.
Second, we add dynamics o he model by including he lagged dependen a iable as a
eg esso .
𝑂𝑏𝑎𝑚𝑎 𝐴𝑝𝑝𝑟𝑜𝑣𝑎𝑙𝑡=𝛼+ 𝛽 𝑂𝑏𝑎𝑚𝑎 𝐴𝑝𝑝𝑟𝑜𝑣𝑎𝑙𝑡−1+ 𝛾 𝑿′𝑡−1+ 𝛿 𝒁′𝑡+𝜀𝑡.
In he abo e equa ions, 𝑂𝑏𝑎𝑚𝑎 𝐴𝑝𝑝𝑟𝑜𝑣𝑎𝑙𝑡−1 is he lagged dependen a iable, 𝑿′𝑡−1 is a
ec o o independen economic a iables, and 𝜀𝑡 is he e o e m. To add ess poli ical ac o s (𝒁′𝑡), we
inco po a ed wo ex a a iables in he econome ic model: a "honeymoon e ec " wi h alues 2 o he
ini ial mon h, 1 o he second, and 0 he ea e , o ac o in he ini ial popula i y su ge pos -p esiden ial
inaugu a ion ( i s and second e m); and a " i s e m" dummy a iable, o accoun o Obama's ini ial
p esiden ial manda e and o e osion o popula i y. We expec ed bo h o ha e posi i e coe icien s.
The esul s a e p esen ed in Table 9. Upon ini ial obse a ion, i seems e iden ha he economy
exe ed an impac on Obama’s popula i y. The long- un equa ions, columns (1) and (2), show ha bo h
he unemploymen a e and he in la ion a e ha e s ong, signi ican , and nega i e e ec s on he app o al
a ing. Howe e , he Consume Sen imen (column 2) is comple ely i ele an (p- alue o 0.94).
27
When one conside s a dynamic model able o cap u e he sho - un dynamics, columns (3) and
(4), he esul s change somewha . The in la ion a e is no longe signi ican , and he consume sen imen
becomes highly signi ican . Subsequen ly, gi en hese esul s, a esea che would p obably es ima e an
E o Co ec ion Model combining bo h equa ions, explici ly modeling he dis inc ion be ween he long-
un ela ion and he sho - un dynamics. We do no pe o m ha analysis he e, as ou pu pose is a
di e en one: o illus a e he ad an ages o complemen ing he adi ional eg ession echniques wi h
wa ele analysis.
26
Using he Augmen ed Dickey-Fulle es o assess he s a iona i y o he dependen a iable, he es s a is ic o he app o al a e is -3.295, which ollows
be ween he 1% c i ical alues o -3.517 and he 5% c i ical alue o -2.894.
27
Some au ho s can a gue ha , o es ima e a long- un ela ionship, unemploymen and in la ion should no be lagged. Including he con empo a y alues has
minimal impac on ou es ima ions. We kep hem lagged jus o consis ency wi h columns 3 and 4.

52
Table 9: Obama’s app o al a e as a unc ion o economic and poli ical a iables, es ima ed by OLS.
(1)
(2)
(3)
(4)
L.App o al
0.878***
0.890***
(0.041)
(0.041)
L.Unemploymen
-1.405***
-1.381**
-0.697***
-0.382**
(0.493)
(0.599)
(0.159)
(0.172)
L.In la ion
-2.311***
-2.309***
-0.044
0.018
(0.377)
(0.393)
(0.189)
(0.177)
Sen imen
0.007
0.083***
(0.090)
(0.029)
Honeymoon
6.408***
6.446***
0.368
0.755*
(1.160)
(1.281)
(0.520)
(0.444)
Fi s _Te m
8.014***
8.049***
2.119***
2.472***
(1.664)
(1.643)
(0.634)
(0.645)
Cons an
57.25***
56.53***
9.927***
0.192
(3.058)
(10.50)
(2.406)
(3.567)
Obse a ions
95
95
95
95
R
2
0.466
0.466
0.914
0.921
F s a is ic
17.74
15.43
167.920
269.229
No e: S anda d-e o s be ween b acke s. S a is ical signi icance: * p<.10; ** p<.05; *** p<.01 ( wo- ailed
es s).
In sho , Table 9 ells us ha he economy exe ed an impac on Obama’s popula i y. Clea ly, he
unemploymen a e is ele an h oughou he en i e pe iod unde conside a ion. The in la ion a e and
he consume sen imen a e also ele an , bu hey a e model dependen . I is also e iden ha he
coe icien s o ou poli ical a iables a e posi i e, as we expec ed. Howe e , ega ding s a is ical in e ence,
he honeymoon e ec is model dependen as well.
The econome ic models employed in ou analysis did no p o ide addi ional in o ma ion
ega ding he speci ic ime pe iods and equencies a which he ela ionship be ween app o al a ing and
economic a iables a ained s a is ical signi icance. Subsequen ly, we will examine how hese a iables
in e ac o e ime and equency using wa ele analysis.
53
4.5.2. Wa ele Analysis
We s a wi h an examina ion o he Pa ial Wa ele Cohe ency, depic ed in Figu e 10. Clea ly,
unemploymen (cha a.i) and in la ion (cha b.i) a es exhibi a signi ican co ela ion wi h app o al,
pa icula ly a low equencies. Du ing Obama’s enu e, a co-mo emen be ween his app o al a ing and
he s a e o he economy is obse ed o equencies be ween 2 yea s o 4 yea s. This inding aligns wi h
ou p e ious discussion, emphasizing he subs an ial in luence o economic ac o s on Obama’s app o al.
The pa ial phase-di e ences in he middle o Figu e 10 p o ide e idence o a nega i e co ela ion
be ween Obama’s app o al a ing and bo h he unemploymen a e and he in la ion a e.
S a ing wi h he unemploymen a e, we obse e a subs an ial and s a is ically signi ican
cohe ence du ing a la ge pa o Obama’s p esidency, pa icula ly wi hin he equency band o 2-4 yea s.
A hose equencies, he phase-di e ence (diag am a.ii) consis en ly alls be ween π/2 and π, indica ing
Figu e 10: Wa ele Analysis: Obama’s App o al, Unemploymen Ra e, and In la ion Ra e.
No e: Top: Rela ion be ween he App o al a e and he Unemploymen a e a e con olling o
In la ion. Bo om: Rela ion be ween he App o al a e and he In la ion a e a e con olling o
Unemploymen . Cha s a.i and b.i: Pa ial Wa ele Cohe ency be ween he App o al a e and one
economic a iable a e con olling o he o he . Diag ams a.ii and b.ii: Pa ial Phase-Di e ence
be ween he App o al a e and one economic a iable a e con olling o he o he in wo equency
bands. Diag ams a.iii and b.iii: Pa ial Wa ele Gain be ween he App o al a e and one economic
a iable a e con olling o he o he in wo equency bands.
54
an in e se ela ionship be ween he wo a iables, wi h unemploymen leading he way.
28
The phase-
di e ence a highe equencies con eys a simila message; howe e , gi en he low signi icance o he
cohe ency, we e ain om a aching impo ance o his esul . As we can see on he diag am a.iii, he
es ima ed coe icien o he 2-4 yea s equency band emains s able, ho e ing a ound 5.
Rega ding he in la ion a e (depic ed in cha b.i), we obse e a simila equency band o
signi ican cohe ence (2-4 yea s), which pe sis s h oughou almos he en i e y o Obama’s p esidency.
Wi hin his equency band, he pa ial phase-di e ence closely app oxima es o π (o –π), indica ing a
nega i e co ela ion be ween he in la ion a e and he app o al a e. Like he ela ionship obse ed
be ween he unemploymen a e and he app o al a e, he in la ion a e leads he app o al a e. Ins ead,
simul aneous inc eases (dec eases) in he in la ion a e coincide wi h co esponding dec eases
(inc eases) in he app o al a e. The es ima ed coe icien o his equency band luc ua es be ween 3.5
and 4. I is no ewo hy ha a highe equencies, he e is a small island o high cohe ency be ween 2015
and 2016. Fo ha pe iod, he phase-di e ence indica es a nega i e ela ionship wi h in la ion leading
again.
In Figu e 11, we add consume sen imen o he analysis. Ou esul s ega ding in la ion and
unemploymen do no change subs an i ely when we use consume sen imen as a con ol a iable. The
mos impo an egions o high cohe ency occu a lowe equencies, and he phase-di e ences and
es ima ed coe icien s a e simila . Visually, he mos impo an change is abou he a ea o he high
cohe ency egions, which is now smalle . Bu ha is o be expec ed, gi en he co ela ion be ween
consume sen imen and he economic a iables.
A no able inding eme ges when one looks a he pa ial cohe ency be ween he app o al a e
and he consume sen imen in Cha C o Figu e 11, which helps o make sense o he OLS esul s in
he p e ious Subchap e . Consume sen imen is signi ican a high cohe encies o e y speci ic pe iods.
Tha is pa icula ly ob ious be ween 2011 and 2012. A e 2014, and a high equencies, al hough no
as high, i becomes signi ican again. This sugges s ha he s a e o he economy is o g ea e impo ance
in he medium o long e m, while he consume sen imen cap u es mo e immedia e e ec s. In con as
o he ela ionship es ablished wi h unemploymen and in la ion, he connec ion be ween consume
sen imen and app o al a ing appea s o be less s able.
28
See Figu e 7.
55
Analyzing he diag am c.ii (p ima ily o highe equencies, which exhibi s a is ical signi icance),
we ind ha be ween 2011 and 2012, he phase-di e ence is nea ly ze o, sugges ing a co-mo emen
be ween he app o al a e and consume sen imen . Subsequen ly, a e 2014, he phase-di e ence alls
be ween –π/2 and 0, implying ha du ing his pe iod, consume sen imen leads changes in he app o al
a ing. Fu he mo e, when examining he es ima ed coe icien , we obse e he same pa e n o ins abili y.
The es ima ed coe icien be ween 2011 and 2012 is lowe han he coe icien obse ed be ween 2014
and 2016.
Figu e 11: Wa ele Analysis: Obama’s App o al, Unemploymen Ra e, In la ion Ra e, and
Consume Sen imen .
No e: Top: Rela ion be ween he App o al a e and he Unemploymen a e a e con olling o
In la ion a e and Consume Sen imen . Middle: Rela ion be ween he App o al a e and he
In la ion a e a e con olling o Unemploymen a e and Consume Sen imen . Bo om: Rela ion
be ween he App o al a e and he Consume Sen imen a e con olling o Unemploymen and
In la ion a es. Cha s a.i, b.i and c.i: Pa ial Wa ele Cohe ency be ween he App o al a e and one
economic a iable a e con olling o he o he s. Diag ams a.ii, b.ii and c.ii: Pa ial Phase-
Di e ence be ween he App o al a e and one economic a iable a e con olling o he o he s in
wo equency bands. Diag ams a.iii, b.iii and c.iii: Pa ial Wa ele Gain be ween he App o al a e
and one economic a iable a e con olling o he o he s in wo equency bands.
56
The wa ele app oach sheds, he e o e, ligh on he OLS esul s. The long- un ela ion ound in
he s a ic equa ion (in which he consume sen imen ’s coe icien was no s a is ically signi ican )
co esponds o he esul s ha we ound o lowe equencies using wa ele s. The sho - un ela ion ha
we ound o he consume sen imen in he dynamical equa ion co esponds o he esul we ound a
high equencies.
No e, howe e , ha he wa ele esul s also ell us ha he ela ion be ween he consume sen imen
and he app o al a e is no s able, sugges ing ha i was pa icula ly s ong be ween 2011 and 2012. In
ac , i one ees ima es Model 4 o Table 9 a e emo ing he obse a ions be ween he end o 2011 and
2012, he coe icien is no longe signi ican a 5%, and i is only ma ginally signi ican a 10% (p- alue o
0.095). This is he ype o in o ma ion ha is almos immedia e o ob ain wi h he con inuous wa ele
ans o m, which would be ha d o ob ain wi h he mo e adi ional echniques.
4.6. Conclusion
One o he co ne s ones o he li e a u e on go e nmen app o al is he idea ha he s a e o he economy
a ec s people’s judgmen o poli icians (Lewis-Beck and S egmaie , 2013). Howe e , he es ima ion o
he impac o he economy on poli ical popula i y has no yielded consis en s a is ical s abili y. As a esul ,
se e al s udies ha e p esen ed di e gen indings. This s udy seeks o con ibu e o he unde s anding o
his issue by explo ing a no el econome ic ool. Wa ele analysis allows us o explo e nuances ha
adi ional econome ic analysis does no . As a complemen , wa ele ools can con ibu e o o e coming
some es ima ion p oblems, ci cum en ing s a iona i y and nonlinea issues, and adding new laye s o
in o ma ion, allowing he s udy o he dynamics be ween wo o mo e a iables in he ime- equency
domain. In pa icula , o he ex en ha he ela ionship be ween he economy and poli ical popula i y is
dynamic and may e ol e o e ime, wa ele analysis is app op ia e o dis inguish long- e m and s able
ela ionships om ansien and uns able ones. Mo eo e , using he wa ele gain in he ime- equency
domain (Aguia -Con a ia
e al.
, 2023), we can e en es ima e coe icien s o mul i a ia e wa ele analysis,
an applica ion ha has so a emained unexplo ed in Poli ical Science.
In his s udy, we illus a e his me hodological app oach o he complex ela ionship be ween he
economy and he popula i y o poli icians by examining he app o al unc ion o P esiden Ba ack Obama
in a pe iod spanning om Janua y 2009 o Decembe 2016, a pe iod ma ked by conside able economic
ins abili y and unce ain y. Using a adi ional econome ic app oach, we ind a nega i e impac o
unemploymen and in la ion a es in long- un models, while consume sen imen is only s a is ically

57
signi ican when we un dynamic models ha cap u e sho - un beha io . Subsequen ly, we complemen
hese esul s by unning he wa ele s ools desc ibed in Chap e 4.2. The pa ial wa ele analysis con i ms
hese esul s and ein o ces he idea ha he ela ionship be ween economic a iables and app o al a ing
is nega i e and dynamic. The unemploymen and in la ion a es exhibi a long- e m ela ion wi h Obama’s
app o al a ing, which is e iden h oughou his p esidency and a lowe equencies. Howe e , he
wa ele analysis adds ano he laye o ele an in o ma ion. We know ha unemploymen and in la ion
ha e a long- un impac on popula i y, bu wi h he phase-di e ence diag ams, we also know ha he ime
o impac is di e en . Al hough bo h he unemploymen and in la ion a es lead he app o al a e, he
synch oniza ion is di e en . The pa ial phase-di e ence diag ams show ha he app o al a e eac ed
mo e quickly o unemploymen a e mo emen s han i did o mo emen s o he in la ion a e.
Fu he mo e, ou analysis also e eals a posi i e and synch onized co-mo emen be ween
consume sen imen and Obama’s app o al a ing, pa icula ly du ing he pe iod spanning 2011 and
2012 and a highe equencies. This ansien and uns able ela ionship be ween consume sen imen
and Obama’s app o al is no deli e ed by he adi ional econome ic models. Mo eo e , he esul o
consume sen imen coincides wi h signi ican e en s such as he Uni ed S a es deb -ceiling c isis and
he subsequen Fede al Go e nmen Shu down, which a e esponsible o a signi ican inc ease in
consume s’ wo iedness.
In conclusion, wa ele analysis is use ul o complemen ing econome ic analysis in social
sciences, pa icula ly in poli ical science. While adi ional econome ic ools a e impo an in analyzing
and quan i ying ela ionships wi hin a se o a iables, hey do no allow a ime- equency analysis as
he wa ele app oach does.
58
5. Poli ical Connec ions in Po ugal
5.1. In oduc ion
Poli ical connec ions ep esen an omnip esen aspec o go e nance and economic sys ems (Faccio,
2006). A hei co e, hese connec ions highligh he complex ne wo k o ela ionships be ween
indi iduals, businesses, and go e nmen al en i ies, wielding signi ican in luence o e esou ce alloca ion,
policy o mula ion, and he o e all socio-economic landscape.
29
Wi hin he domain o poli ical economy,
he s udy o poli ical connec ions eme ges as a c ucial poin o unde s anding he dynamics o powe ,
weal h dis ibu ion, and ins i u ional beha io .
Wi hin he deba e on poli ical connec ions, he e is e idence ha his ela ionship be ween poli ics
and business has a pe asi e impac on economic ou comes (Schoenhe , 2019; Khwaja and Mian,
2005). Rega dless o he poli ical egime o he coun y’s economic and social de elopmen , he abili y
o o ge alliances wi h poli ical eli es can con e subs an ial ad an ages o i ms, anging om p e e en ial
access o public p ocu emen o inancial c edi , egula o y exemp ions o enhanced ma ke in luence,
and p o ec ion om compe i ion. I also con e s subs an ial ad an ages o poli icians, such as an inc ease
in poli ical in luence and elec o al success (Be and
e al.
, 2018). Such connec ions a e o en associa ed
wi h en -seeking beha io , whe eby indi iduals o en i ies seek o ex ac economic en s h ough
le e aging hei poli ical a ilia ions. Mo eo e , he opaci y su ounding hese ela ionships can unde mine
anspa ency, accoun abili y, and he ule o law, os e ing en i onmen s conduci e o co up ion and
c onyism (Ramalho, 2007; Shlei e and Vishny, 1994).
Poli ical connec ions a e no exclusi e o de eloping coun ies. E en in de eloped coun ies wi h
low le els o pe cei ed co up ion, he linkage be ween i ms and poli icians is a eali y (Amo e and
Bennedsen, 2013). Po ugal is no excep ion, and i is easy o ind o me poli icians as membe s o he
boa d o di ec o s. Mo eo e , in a se o su eys conduc ed by he Eu opean Union o businesspeople
29
The li e a u e has also ound connec ions be ween poli icians and o he social g oups, such as eligious leade s (T oncone and Valli, 2024).
59
om all o e Eu ope, mos o he Po uguese businesspeople su eyed consis en ly con i med ha ha ing
poli ical connec ions is he only way o succeed in business (Eu opean Commission, 2014, 2015, 2024a).
As he li e a u e highligh s, his ha m ul impac on i ms and, o e all, he coun y’s economy c ea es an
incen i e o s udy he Po uguese case. The e o e, his Chap e s udies he exis ence and e ec s o
poli ical connec ions in Po ugal. We aim o analyze he cha ac e is ics ha dis inguish connec ed i ms
om non-connec ed i ms and whe he poli ical connec ions a ec hese i ms inancially. Fo ha
pu pose, we cons uc a new da abase using a sample o publicly aded Po uguese companies. This
new da abase encompasses h ee ypes o da a: 1) da a on he membe s o he boa d o di ec o s o he
Po uguese publicly aded i ms be ween 2002 and 2022; 2) da a on all P esiden s o he Republic,
membe s o he go e nmen , and membe s o he pa liamen since 1974; and, inally, 3) annually da a
wi h inancial and co po a e in o ma ion be ween 2002 and 2022.
Ou indings highligh ha i m size, measu ed by o al asse s, eme ges as a signi ican ac o
in luencing he p esence and he numbe o poli icians on co po a e boa ds. Fu he mo e, i ms in
egula ed sec o s and headqua e ed in he Lisbon Me opoli an A ea show a highe p obabili y o poli ical
linkages. We also ake in o conside a ion he poli ical ideology, iden i ying whe he he boa ds a e igh o
le -wing. The esul s sugges ha size explains he p edominance o igh -wing poli ically connec ed
membe s, whe eas le -wing connec ions a e mo e likely in i ms headqua e ed in he Lisbon egion.
Fu he mo e, ou eg ession analysis examines he impac o poli ical connec ions on a ious inancial
and ma ke pe o mance indica o s. The esul s indica e ha poli ical connec ions do no signi ican ly
in luence hese inancial and ma ke pe o mance measu es. This conclusion is obus o di e en
compu a ions o ou poli ical connec ion a iable.
The main con ibu ion o his Chap e is o enhance a ai deba e abou he complex ela ionship
be ween business and poli ics in Po ugal. This ela ionship has endu ed since he XIX cen u y and
con inues o in luence socie y and i ms oday (Cos a
e al.
, 2010). I is impo an o unde s and i s
exis ence and socie al co po a e impac s. Li e a u e has iden i ied po en ial p oblems o igina ing om he
exis ence o his kind o ela ionship, and su eys conduc ed by he Eu opean Union demons a e i .
Po uguese businesspeople ha e conside ed poli ical connec ions as a ehicle ha enhances co up ion
and can hampe business compe i ion (Eu opean Commission, 2024a). The e o e, as he li e a u e has
highligh ed he nega i e impac o co up ion on economic g ow h (G ündle and Po a ke, 2019), i seems
c ucial o deepen he s udy o he exis ence o poli ical connec ions in Po uguese i ms and o analyze
whe he hey a e associa ed wi h en ac i i ies and co up ion.
60
The chap e is s uc u ed as ollows. Chap e 5.2 p esen s a b ie e iew o he li e a u e on
poli ical connec ions, and Chap e 5.3 con ex ualizes he Po uguese poli ical scena io since he XIX
cen u y oge he wi h i s di ec o indi ec pa icipa ion on business. Conce ning he empi ical pa ,
Chap e 5.4 desc ibes he da a collec ed and how he da ase was cons uc ed, Chap e 4.5 discusses
he empi ical s a egy and he hypo hesis o mula ed, and Chap e 4.6 p esen s he esul s. Finally,
Chap e 4.7 concludes.
5.2. Li e a u e Re iew
Connec ions be ween poli icians and i ms a e a global phenomenon, no con ined o poo o de eloping
coun ies, no limi ed o less anspa en o au oc a ic egimes (Faccio, 2006; Faccio, 2010). This is
e idenced by he p esence o poli icians in impo an posi ions wi hin bo h na ional and mul ina ional
companies and sha eholde s o co po a e boa d membe s wi h amilial o social ies o poli ically
mo i a ed indi iduals. Mo eo e , di ec o s o companies wi h no poli ical backg ound may un o poli ical
o ice o ecei e appoin men s o public o ices.
When analyzing poli ical connec edness, i is impo an o conside he ins i u ional and business
amewo k o each coun y. Ce ain na ions a e mo e suscep ible o such connec ions, which a y in
na u e and in ensi y. Gene ally, coun ies wi h a highe incidence o poli ical-business en anglemen end
o exhibi g ea e co up ion, impose es ic ions on o eign in es men ab oad by hei ci izens, and ha e
mo e anspa en sys ems (Faccio, 2006).
30
Anecdo al e idence sugges s a g ea e p e alence o
poli ically connec ed i ms in de eloping coun ies, whe e co up ion le els a e high and ins i u ions a e
weak. Ne e heless, such connec ions a e also ound in de eloped coun ies wi h s ong ins i u ions and
low le els o co up ion (Amo e and Bennedsen, 2013).
The ela ionship be ween business and poli ics is an old s o y. Fi ms y o in luence poli ical
ou comes o gain, o ins ance, compe i i e ad an ages, while poli icians suppo i ms o p i a e o
elec o al gains. Fe guson and Vo h (2008) assessed he alue o Ge man companies associa ed wi h he
Na ional Socialis Pa y (Nazi Pa y) ollowing Hi le ’s appoin men as chancello . They ound ha i ms
ha suppo ed he Ge man dic a o poli ically and inancially enjoyed highe s ock e u ns, ega dless o
hei indus y. Simila ly, Fascis I aly also g an ed high en s o ce ain i ms (Faccio and McConnell,
2023), wi h he all o he egime causing hese i ms o unde pe o mance compa ed o hei pee s,
30
T anspa en sys ems allow an easy iden i ica ion o he exis ing poli ical connec ions. T anspa ency can also p omo e some ole ance ega ding connec ions
because any misconduc “is mo e likely o be punished.”
67
By me ging hese wo da ase s, we iden i ied 100 poli icians, o which only ou a e women. To
minimize he isk o misma ches, we used a ious me ging s a egies and con i med po en ial connec ions
using di e se sou ces, such as online newspape s o di ec o s’ CVs. O e all, we ound 144 ins ances o
poli icians holding company di ec o ships. Si ua ions whe e a di ec o assumed a posi ion in a company
be o e hei poli ical ca ee a e no conside ed connec ions.
Finally, we collec ed da a on i ms’ inancial si ua ion. Following Faccio (2010), we collec ed da a
on To al Asse s ( housands o eu os) (
Asse s
), Ma ke Capi aliza ion ( housands o eu os) (
MCap
), Re u ns
on Equi y (%) (
ROE
), To al Deb o e Ma ke Capi aliza ion (
Le e age
), Book- o-Ma ke , Ma ke Sha e,
P oduc i i y, and Annual S ock Re u ns (
Re u ns
).
39
Ma ke Sha e is calcula ed as he p opo ion o each
company’s ma ke capi aliza ion ela i e o he o al ma ke capi aliza ion o all companies in he same
indus y by yea . Finally, P oduc i i y is compu ed by applying a s anda d Cobb-Douglas p oduc ion
unc ion, whe e we eg essed he na u al loga i hm o e enues (ou pu ) agains he na u al loga i hms o
o al asse s (capi al) and employees (labo ), using OLS es ima es. We hen used he es ima ed esiduals
as a measu e o mul i ac o p oduc i i y. Indus y sec o s we e iden i ied using he Indus y Classi ica ion
Benchma k (
ICB
). All da a was ob ained om he Thomson Reu e s da abase.
Wi hin he li e a u e on poli ical connec ions, he opic o p i a iza ion is pa icula ly pe inen
(Boubak i
e al.
, 2008). A signi ican numbe o pos -p i a ized i ms main ain hei poli ical connec ions,
and he e is a posi i e co ela ion wi h he le el o esidual s a e owne ship. The e o e, we cons uc ed
wo addi ional a iables: a dummy a iable codded as one o all publicly aded Po uguese companies
ha unde wen p i a iza ion o e-p i a iza ion (
P i a ized
),
40
and ano he dummy a iable codded as one
o yea /company combina ions in which he go e nmen held an owne ship o sha eholde posi ion,
including mino i y s akes (
S a e Owne ship
). To accoun o poli ical ideology, empo al changes in poli ical
connec ions, and hei geog aphical dis ibu ion, we inco po a ed addi ional con ols. Dummy a iables
we e cons uc ed o p ime minis e s (
Du ao&San ana
,
Soc a es
, and
PCoelho
), o igh -wing
go e nmen s (
Righ
), and o he pe iod du ing and ollowing he inancial in e en ion and bailou by he
T oika g oup (
T oika
).
41
The 2011 Po uguese Deb C isis and subsequen in e en ion by he T oika had
a posi i e impac on anspa ency, media sc u iny, and egula ion o publicly aded companies.
39
All mone ized a iables we e de la ed using he Consume P ice Index (CPI), wi h 2012 as he base yea . The CPI da a was sou ced by INE.
40
These i ms we e go e nmen -owned i ms.
41
Following he Global Financial C isis in 2007 and he beginning o he Eu opean Deb C isis, P ime Minis e José Sóc a es called o inancial suppo in
Ap il 2011. A ew days la e , a g oup o nego ia o s om he Eu opean Cen al Bank, he Eu opean Commission, and he In e na ional Mone a y Fund, known
as he
T oika
, wen o Lisbon o nego ia e a bailou package. In his package, Po ugal commi ed o equilib a ing he public inances, educing he deb , and
p i a izing some public companies in exchange o inancial aid.

68
Addi ionally, he T oika manda ed he p i a iza ion o se e al companies, educing he go e nmen 's
p esence in a ious economic sec o s. The e o e, he a iable
T oika
is codded as one o he pe iod a e
2011 and ze o o he wise. Fo geog aphic localiza ion, we collec ed he headqua e s add ess o each
publicly aded company and c ea ed a dummy a iable (
Lisbon
), which equals one i he headqua e is
in a municipali y o he Lisbon Me opoli an A ea and ze o o he wise. To con ol o sec o ixed e ec s,
we also c ea ed dummies o he ollowing sec o s: Consume s aples, ene gy, inancial, heal h ca e,
indus ials, eal es a e, echnology, elecommunica ions, and u ili ies. Finally, we c ea ed a dummy
a iable (
Regula ed
) codded one o i ms in sec o s subjec o egula o y o e sigh : ene gy, inancials,
elecommunica ions, and u ili ies.
Table 10 p esen s he desc ip i e s a is ics o he abo e-men ioned a iables.
42
To a oid ou lie s, we
excluded he 0.5 h and 99.5 h pe cen iles om
To al Asse s
,
ROE
, Le e age, B
ook- o-Ma ke
, and
Re u ns
. Rega ding co po a e inance, he a e age company has o al asse s alued a app oxima ely
5,700 million eu os despi e a conside ably high s anda d de ia ion. Fu he mo e, he a e age company
exhibi s a e u n on equi y close o 5.6% and a ma ke capi aliza ion nea 1,400 million eu os. Addi ionally,
Table 11 epo s di e ences in he means o hese co po a e a iables be ween poli ically connec ed i ms
and non-connec ed. Rega ding i m size, measu ed by he o al asse s o he ma ke capi aliza ion, bo h
in loga i hms, connec ed i ms a e signi ican ly la ge . Mo eo e , connec ed i ms a e also mo e p o i able,
ha e ela i ely highe p oduc i i y le els, and a e less le e aged. Table 11 also shows ha ma ke sha e
is highe o connec ed i ms, bu he e is no s a is ical di e ence in e u ns be ween hese wo g oups.
Finally, book- o-ma ke is lowe o he connec ed i ms, sugges ing ha hese i ms a e o e alued by
in es o s.
42
Mo eo e , Table A.4 shows he co ela ions be ween all a iables desc ibed and used in he empi ical analysis.
69
Table 10: Desc ip i e s a is ics.
Obse a ions
Mean
S d-De
Min
Max
Boa d (N)
907
9.49
5.62
2.00
31.00
Connec ion
907
0.42
0.49
0.00
1.00
Connec ion Righ
907
0.22
0.42
0.00
1.00
Connec ion Le
907
0.09
0.28
0.00
1.00
Connec ion (N)
907
0.85
1.40
0.00
9.00
To al Asse s
899
6016451.56
15400121.37
2826.98
94466240.00
Ma ke Cap
902
1469821.35
3041676.12
0.00
20006490.00
ROE
841
5.59
23.40
-171.50
86.54
Le e age
897
3.38
6.03
0.00
54.16
Book o Ma ke
893
1.46
3.34
-18.46
36.50
Ma ke Sha e
902
23.61
29.06
0.00
100.00
P oduc i i y
762
-0.00
0.70
-2.82
1.97
Re u ns
886
4.12
44.84
-77.06
299.27
P i a ized
907
0.21
0.41
0.00
1.00
S a e Owne ship
907
0.09
0.29
0.00
1.00
Righ
907
0.34
0.47
0.00
1.00
Lisbon
907
0.63
0.48
0.00
1.00
T oika
907
0.51
0.50
0.00
1.00
Regula ed
907
0.28
0.45
0.00
1.00
Basic Ma e ials
907
0.13
0.34
0.00
1.00
Consume Disc e iona y
907
0.18
0.38
0.00
1.00
Consume S aples
907
0.07
0.25
0.00
1.00
Ene gy
907
0.02
0.14
0.00
1.00
Financials
907
0.09
0.29
0.00
1.00
Heal h Ca e
907
0.02
0.14
0.00
1.00
Indus ials
907
0.24
0.43
0.00
1.00
Real Es a e
907
0.02
0.12
0.00
1.00
Technology
907
0.07
0.25
0.00
1.00
Telecommunica ions
907
0.11
0.31
0.00
1.00
U ili ies
907
0.06
0.23
0.00
1.00
On a e age, boa ds o di ec o s consis o 9.5 membe s, wi h a ange om 2 o 31 membe s (Table
10). Mo eo e , he a e age numbe o connec ions pe boa d is 0.85, wi h some boa ds ha ing up o 9
poli ically connec ed membe s (
Connec ion (N)
). Figu e 12 illus a es he e olu ion o he a e age numbe
o connec ions om 2002 o 2022. The e was an upwa d end om 2002 o 2011, wi h he a e age
numbe o poli ically connec ed membe s pe boa d inc easing om 0.69 in 2002 o 1.07 in 2011.
Subsequen ly, a e 2011, he e was a signi ican dec ease in he a e age numbe o poli ically connec ed
membe s, plumme ing o 0.31. This decline was spu ed by a educ ion in poli ically connec ed membe s
70
wi hin boa ds despi e no pa allel educ ion in boa d size. Figu e 1313 highligh s his declining end,
pa icula ly e iden om a ound 2006 onwa d and no ably p onounced om 2016 onwa ds.
Table 11: T- es on he di e ences in he means o inancial indica o s be ween poli ically connec ed and
non-connec ed i ms.
Connec ed
Non-Connec ed
Di
Log TA
14.76
12.61
2.15***
Ma ke Cap
13.48
11.06
2.42***
Roe
7.10
4.41
2.69*
Le e age
2.94
3.71
-0.78*
Ma ke Sha e
33.13
16.59
16.54***
Re u ns
3.24
4.78
-1.54
P oduc i i y
0.06
-0.06
0.12**
Book o Ma ke
1.07
1.76
-0.69***
No e: S a is ical signi icance: ***p<.01; **p<.05; *p<.1.
Figu e 12: A e age numbe o boa d membe s poli ically connec ed by yea (2002-2022).
71
Ce ain indus ial sec o s, especially hose in ol ing non- adable goods (Domadenik
e al.
, 2016) and
egula ed indus ies (Boubak i
e al.
, 2008), seem o be mo e p one o poli ical connec ions. Table
12
12
shows ha companies wi hin he Ene gy, U ili ies, Financials, and Telecommunica ion sec o s ypically
ha e boa ds o di ec o s wi h mo e han en membe s and a highe p opo ion o o me poli icians.
U ili ies ha e he highes a e age o poli ically connec ed membe s (3.14), and app oxima ely 16.2% o
U ili y boa d membe s a e o me poli icians, in con as o 7% ac oss all sec o s. Doing an analysis
be ween non- egula ed and egula ed sec o s, Figu e 14 shows ha in all yea s, he egula ed sec o
highligh s highe pe cen ages o poli ically connec ed membe s han he non- egula ed sec o . Mo eo e ,
he pe cen age o poli ically connec ed membe s om he non- egula ed sec o has exhibi ed a declining
end since 2005, o e lapping he declining end obse ed in Figu e 13. Howe e , he pe cen age om
he egula ed sec o main ained abo e 10% un il 2019, when i s a ed o dec ease. In 2022, he
pe cen age o poli ically connec ed membe s om he egula ed sec o was less han 5%.
Figu e 13: A e age numbe o boa d membe s, a e age numbe o boa d membe s poli ically
connec ed, and pe cen age o boa d membe s poli ically connec ed by yea (2002-2022).
72
Table 12: A e age numbe o boa d membe s and poli ical connec ions by indus y.
Indus y
A e age numbe o boa d membe s
Poli ical Connec ions
A e age numbe
(%)
Basic Ma e ials
7.20
0.26
3.34
Consume Disc e iona y
6.63
0.70
8.92
Consume S aples
7.88
0.38
3.99
Ene gy
18.53
2.88
15.39
Financials
14.20
1.76
10.97
Heal h Ca e
7.37
0.07
0.44
Indus ials
7.67
0.48
4.55
Real Es a e
3.57
0.00
0.00
Technology
8.15
0.46
5.67
Telecommunica ions
10.73
0.83
7.46
U ili ies
17.40
3.14
16.18
O e all
9.94
0.99
6.99
Figu e 14: Pe cen age o poli ically connec ed membe s on boa ds in egula ed and non- egula ed
i ms by yea .

73
5.5. Empi ical S a egy
In his Subchap e , we desc ibe he empi ical me hodology used o examine he ela ionship be ween
poli ical connec ions and Po uguese publicly aded i ms, as discussed in he in oduc ion. Ou app oach
aims o analyze how a se o cha ac e is ics, such as i m size, localiza ion, indus y, and egula ion,
in luence he likelihood o a i m ha ing poli ically connec ed boa d membe s. Addi ionally, we assess he
impac o poli ical connec ions on key inancial indica o s.
5.5.1. Fi m Cha ac e is ics and Poli ical Connec ions
To s udy he cha ac e is ics ha could dis inguish i ms wi h and wi hou poli ical connec ions, we employ
p obi eg ession models o es ima e he likelihood o a i m being poli ically connec ed. The dependen
a iable is a dummy a iable ha equals one i a i m’s boa d has a leas one o me minis e , s a e
sec e a y, o membe o pa liamen as a di ec o in a gi en yea and ze o o he wise (see a iable
Connec ion
in Table 10). This a iable iden i ies he p esence o poli ical links wi hin he boa d o each
company and yea .
P obi models a e pa icula ly sui ed o his kind o analysis, as hey allow us o model he p obabili y
o an e en occu ing while accoun ing o he ac ha he dependen a iable is bina y. In his pa icula
case, he p obi model assumes ha he ela ionship be ween he co a ia es ( i m size, localiza ion,
indus y, e c.) and he p obabili y o being poli ically connec ed is nonlinea and ollows a no mal
dis ibu ion. The gene al model o be es ima ed can be ep esen ed by he equa ion (1):
𝑌𝑖,𝑡 =𝛽0+𝛾.𝑿𝑖,𝑡 +𝜇𝑖+𝜆𝑡+𝜀𝑖,𝑡, (1)
whe e 𝑌𝑖,𝑡 is he dependen a iable, which is ou dummy a iable ha cap u es he p esence o poli ical
connec ions o each Po uguese publicly aded i m in yea . 𝑿𝑖,𝑡 is a ec o o i m’s cha ac e is ic, 𝜆𝑡
ep esen s ime- ixed e ec s and 𝜇𝑖 indus y ixed-e ec s.
43
𝛽0 is a cons an , 𝛾 a ec o o pa ame e s,
and 𝜀𝑖,𝑡 he e o e m. As o he i m’s cha ac e is ics, we included he loga i hm o o al asse s o
measu e he i m’s size, a dummy a iable o con ol o he exis ence o a pas p i a iza ion o e-
43
Es ima ing nonlinea models like P obi wi h indi idual ixed e ec s can lead o a phenomenon known as he inciden al pa ame e s p oblem when T is no
su icien ly la ge. To a oid his issue, we used indus y- ixed e ec s ins ead. Fo obus ness, we also es ima ed logi models wi h indus y- ixed e ec s, and
he esul s we e simila .
74
p i a iza ion p ocess o i he i m has a public s ake, and a dummy a iable ha iden i ies hose i ms
ha ha e headqua e s in he Lisbon egion. Fu he mo e, we also included dummy a iables o each
Po uguese go e nmen and a dummy o i ms ope a ing in egula ed indus ies.
The li e a u e has iden i ied he cha ac e is ics ha allow he iden i ica ion o poli ically connec ed
i ms. As explained in Chap e 5.2, size, p i a iza ion o s a e p esence, localiza ion, o egula ion a e
impo an a iables o s udy poli ical connec edness. The e o e, we an icipa e ha la ge i ms o i ms
ha a e p i a ized o ha e a public s ake ha e a highe p obabili y o being poli ically connec ed. Mo eo e ,
i ms whose headqua e s a e in he Lisbon egion, close o he poli ical decision-making cen e , a e
expec ed o ha e mo e o me poli icians on hei boa ds. Finally, i ms om egula ed sec o s (ene gy,
inancials, elecommunica ions, and u ili ies) may exhibi mo e poli ically connec ed boa ds because o
he possibili y o in luencing he decisions ega ding indus y egula ion.
5.5.2. Impac on Co po a e Va iables
The nex s ep is o de e mine whe he he p esence o poli ically connec ed boa d membe s in luences
key inancial ou comes such as p o i abili y, le e age, and ma ke sha e. He e, we explo e he
me hodology used o examine he e ec o poli ical connec ions on he inancial pe o mance o
Po uguese publicly aded i ms. We un o dina y leas squa es (OLS) models o assess he ela ionship
be ween poli ical connec ions and inancial a iables while con olling o he i m’s cha ac e is ics.
The e o e, ou main independen a iable o in e es is he dummy a iable
Connec ion
. Thus, he
es ima ed equa ion is he ollowing:
𝒀𝒊,𝒕 =𝛽0+𝛽1.𝐶𝑜𝑛𝑛𝑒𝑐𝑡𝑖𝑜𝑛+𝛾.𝑿𝑖,𝑡 +𝜇𝑖+𝜆𝑡+𝜀𝑖,𝑡, (2)
whe e 𝒀𝒊,𝒕 is a ec o o dependen a iables, and 𝑿𝑖,𝑡 a ec o o i m’s cha ac e is ics. The a iable
Connec ion
is he mos impo an independen a iable and, he e o e, i s a is ically signi ican , he
es ima e 𝛽1 o each model cap u es di e ences in i ms’ inancial pe o mance depending on he
exis ence o poli ical connec ions. To con ol o i m’s cha ac e is ics, we included i s size and
localiza ion. 𝛽0 is a cons an , 𝛾 a ec o o pa ame e s, and 𝜀𝑖,𝑡 he e o e m. Finally, 𝜆𝑡 ep esen s
ime- ixed e ec s and 𝜇𝑖 indus y ixed-e ec s.
As dependen a iables, we chose a se o inancial a iables commonly used in he li e a u e:
e u ns on equi y (ROE), le e age, book- o-ma ke , ma ke sha e, p oduc i i y, and s ock e u ns. Based
75
on he li e a u e, we expec ha poli ical connec ions may ha e a mixed impac on hese inancial
ou comes. P e ious s udies (Faccio, 2010; Desai and Olo sga d, 2011) sugges ha poli ically connec ed
i ms o en unde pe o m in e ms o p o i abili y (ROE) and p oduc i i y due o ine iciencies a ising om
en -seeking beha io and esou ce misalloca ion. Mo eo e , hese i ms may be mo e le e aged, elying
on easie access o c edi (Bussolo
e al.
, 2022). On he o he hand, poli ical connec ions could lead o
highe ma ke sha es o highe book- o-ma ke (Faccio, 2010). Finally, hese i ms a e also expec ed o
enjoy highe s ock e u ns. Howe e , he li e a u e also highligh s ha hese i ms su e mo e om
poli ical cycles and economic decline, and he esul s can be mixed.
5.6. Resul s
Table 13 p esen s he ma ginal e ec s o he models es ima ed by P obi eg ession models. These
models aim o es ima e he p obabili y o a i m being poli ically connec ed while con olling o a ious
i m’s cha ac e is ics. In Model 1, he p obabili y o a i m ha ing poli ical connec ions inc eases wi h i s
size, measu ed by he na u al loga i hm o o al asse s. Speci ically, he ma ginal e ec o a one pe cen
inc ease in o al asse s on he condi ional p obabili y o ha ing a leas one poli ician is 0.14 pe cen age
poin s. The p esence o he s a e on he sha eholde s uc u e o he i m also has a posi i e ma ginal
e ec on he p obabili y o being connec ed, wi h a coe icien o 0.196.
In Model 2, we eplaced he a iable
S a e Owne ship
wi h he a iable
P i a ized
. I is no ewo hy
ha almos all i ms ha we e p i a ized o ep i a ized wen h ough se e al p i a iza ion phases, which
means ha p i a ized i ms had he s a e as sha eholde s du ing hese pe iods. Addi ionally, some i ms
con inue o ha e he s a e as a sha eholde . Thus, he a iable
P i a ized
se es as an al e na i e p oxy
o con ol o he s a e’s p esence and he po en ial in luence on he i m boa ds. The coe icien is almos
he same in magni ude and s a is ically signi ican , sligh ly s onge han he
S a e Owne ship
’s.
Geog aphic p oximi y o he poli ical decision cen e was also conside ed. Po ugal is conside ably
cen alized a ound i s capi al, Lisbon, whe e all poli ical ins i u ions and egula o s’ headqua e s a e
loca ed.
44
Model 3 in oduces a a iable o con ol o he localiza ion o he i m’s headqua e s (
Lisbon
).
The coe icien is posi i ely signed and s a is ically signi ican , as we expec ed. The e o e, i ms
headqua e ed nea he poli ical decision-making cen e ha e a highe likelihood o poli ical connec ions,
wi h a ma ginal e ec o 0.193. Fi ms’ size con inues o be an impo an cha ac e is ic, whe eas he
S a e
44
The excep ion is he Po uguese Heal h Regula o y Au ho i y (
ERS
), loca ed in Opo o. The e we e some a emp s o mo e he Na ional Au ho i y o Medicines
and Heal h P oduc s (
In a med
) o Opo o and he Po uguese Cons i u ional Cou o Coimb a, bu unsuccess ully.
76
Owne ship
coe icien u ns ou o be s a is ically no signi ican . As a obus ness check, we also eg essed
Model 3 wi h he a iable
P i a ized
ins ead o
S a e Owne ship
(no shown in he Table), and he esul s
we e simila . Thus, he Lisbon localiza ion seems o be a be e p edic o o he p obabili y o a i m ha ing
membe s o he boa d poli ically connec ed. This esul is co obo a ed by he ac ha only one p i a ized
i m was headqua e ed ou side o he Lisbon Me opoli an Region. The e o e, he a iable
Lisbon
is also
likely o cap u e he e ec o he p esence o he s a e as a sha eholde . Consequen ly, he a iables
con olling o he p esence o he S a e on i ms’ owne ship we e d opped in he ollowing models o
Table 13.
To a oid losing a conside able numbe o deg ees o eedom, yea - ixed e ec s we e eplaced wi h
h ee dummy a iables (column 4), each signaling he enu e o a p ime minis e , anging om Du ão
Ba oso/San ana Lopes o Passos Coelho (see Table A.2 in appendix).
45
This adjus men yielded
consis en indings o i m size and localiza ion. The p ime minis e s’ coe icien s a e all posi i e and
s a is ically signi ican , exhibi ing a dec easing end o e ime. This sugges s ha he p obabili y o a i m
ha ing poli ically connec ed di ec o s dec eased o e ime, con i ming he end obse ed in Figu e 13.
Addi ionally, in Model 5, we eplaced he a iables o p ime minis e s wi h a dummy a iable ha
cap u es he oika pe iod. As expec ed, he coe icien is nega i e and s a is ically signi ican .
46
Rega ding he indus y, columns 1 o 6 p esen he es ima ed coe icien s o each sec o .
47
Resul s
indica e ha companies om he sec o s o Consume Disc e iona y, Technology, Telecommunica ions,
and U ili ies a e mo e likely o ha e boa ds wi h poli ical connec ions. As some indus ies a e subjec ed
o ma ke egula ion, in Model 6, ins ead o using indus y- ixed e ec s, we used a a iable ha cap u es
he dependency o a speci ic indus y on egula o y p ocedu es (R
egula ed
). The e is e idence ha being
a i m in a egula ed sec o inc eases he p obabili y o ha ing a poli ical connec ion, as indica ed by a
ma ginal e ec o 0.199, which is s a is ically signi ican .
O e all, size, headqua e s’ localiza ion, and being in a egula ed sec o a e cha ac e is ics ha
augmen he p obabili y o poli ical linkages. These esul s a e consis en wi h he exis ing li e a u e, as
e idenced by Faccio (2010) and Boubak i
e al.
(2008). We also ied he loga i hmic ans o ma ion o
45
The base scena io is he go e nmen s o PS om 2016 o 2022. We did no c ea e wo sepa a e dummy a iables o Du ão Ba oso and San ana Lopes
since San ana’s go e nmen had a sho du a ion, las ing less han a yea , and was om he same pa y as i s p edecesso .
46
We also un a model subs i u ing he dummy a iable
T oika
wi h a ca ego ical a iable, whe e he pe iod be o e 2011 is codded as 1, he pe iod be ween
2011 and 2014 is codded as 2, and he pe iod a e 2014 is codded as 3. Conside ing he pe iod p e- oika as he base ca ego y, he esul shows ha he
p obabili y o ha ing poli ical connec ions is only s a is ically lowe a e 2014. Resul s a e a ailable upon eques .
47
The Ene gy and he Real Es a e sec o s ha e only one company. The e o e, hei coe icien s do no appea in Table 13.
83
5.7. Conclusion
In conclusion, his s udy del ed in o he exis ence and impac o poli ical connec ions in Po uguese
publicly aded companies. By analyzing he composi ion o boa ds o di ec o s and hei o me poli ical
posi ions, we aimed o unco e he cha ac e is ics ha dis inguish connec ed and non-connec ed i ms
and how poli ical connec ions migh in luence a i m’s inancial ou comes. Ou analysis spanned wo
decades, ocusing on i ms lis ed on he Eu onex Lisbon s ock exchange. This esea ch con ibu es o
he b oade unde s anding o he in e ac ion be ween poli ics and business, shedding ligh on he
p e alence and implica ions o poli ical ies wi hin co po a e go e nance in Po ugal.
To conduc his analysis, we cons uc ed h ee new p ima y da ase s. The i s da ase comp ised
he names o all boa d membe s om Po uguese publicly aded companies since 2002. The second
da ase included he names o all poli icians who held o ice since he 1974 coup d’é a un il Ma ch 2022.
This lis encompassed p esiden s o he epublic, membe s o in e im and cons i u ional go e nmen s,
and membe s o Pa liamen . Addi ionally, we ga he ed ex ensi e da a on i ms’ inancial condi ions,
including a iables such as To al Asse s, Ma ke Capi aliza ion, Re u ns on Equi y, Le e age, Book- o-
Ma ke , Ma ke Sha e, and P oduc i i y. These da ase s enabled a ho ough examina ion o he po en ial
in luence o poli ical connec ions on i m pe o mance.
Ou indings e eal se e al key insigh s. Fi s , la ge i ms, i ms in egula ed sec o s, and i ms
headqua e ed in he Lisbon Me opoli an A ea show a highe p obabili y o poli ical linkages. Howe e ,
he numbe o poli ically connec ed di ec o s ends o dec ease o e ime, and di e ences in ha numbe
a e no signi ican ly in luenced by s a e egula ion in he sec o . In e es ingly, while go e nmen ideology
plays a ole, i m size, and headqua e s’ loca ion a e mo e c i ical de e minan s o poli ical connec ions.
Boa ds wi h a p edominance o igh -wing poli ically connec ed membe s a e ypically la ge , whe eas le -
wing poli ically connec ed membe s a e mo e p e alen in i ms headqua e ed in he Lisbon Me opoli an
A ea. No ably, ou eg ession analysis indica es ha poli ical connec ions do no signi ican ly impac key
inancial and ma ke pe o mance indica o s. Thus, while poli ical connec ions exis in Po uguese publicly
aded i ms, al hough hey ha e los s eng h o e ime, hey do no appea o con e a signi ican inancial
ad an age o disad an age. This las esul aligns wi h Kee e (2019), highligh ing ha he e ogeneous
ins i u ional con ex s and speci ic mechanisms may explain he inconsis en esul s ound in he li e a u e.
Mo eo e , his Chap e looks a poli ical connec ions be ween publicly aded i ms, o e seen by he s ock
ma ke egula o and unde sc u iny by he media, wi h “high-p o ile poli ical connec ions.” Fo Akcigi
e

84
al.
(2023), he links wi h local poli icians “[…] a e much mo e pe asi e and can ha e b oade
consequences o he o e all economy.”
This Chap e p o ides aluable insigh s in o he ela ionship be ween poli ics and i ms,
pa icula ly in he con ex o a de eloped coun y like Po ugal. Howe e , se e al limi a ions should be
acknowledged. Fi s , ou da ase is es ic ed o publicly aded companies, which may no ully cap u e
he dynamics o poli ical connec ions in p i a ely held i ms. Mo eo e , he Po uguese s ock ma ke
consis s o a ela i ely small numbe o companies when compa ed o o he Eu opean s ock ma ke s.
Second, ou measu e o poli ical connec ions is based on he explici p esence o o me poli icians on
boa ds, po en ially igno ing o he o ms o poli ical in luence ha may also play a signi ican ole and ha
ha e been discussed in he li e a u e. Fo u u e esea ch, expanding he da ase o include a b oade
ange o i ms and addi ional yea s could p o ide a mo e comp ehensi e pic u e o poli ical connec ions
in Po ugal. Fu he mo e, explo ing o he channels h ough which poli ical in luence ope a es could
deepen ou unde s anding o how hese ies impac i ms’ pe o mance.
85
6. Conclusion
This disse a ion del es in o some opics s udied in he ield o Poli ical Economy. The i s h ee essays
discuss and p esen po en ial solu ions o he uns able and some imes con adic o y impac o he
economy on poli ical popula i y. The las essay analyzes poli ical connec ions, a opic s udied by di e en
economic b anches, including he poli ical economy.
Chap e 2 p esen s a me hodological al e na i e ha uses di e en economic measu es. The
s udies o popula i y unc ions use accu a e economic da a o measu e he impac o he economy on
popula i y. This essay dis inguishes be ween he “objec i e” and “media e” economies. The “objec i e”
economy is he da a ha e lec s he ue de elopmen s o economic g ow h, whe eas he “media e”
economy e lec s he economic in o ma ion a ailable o he public by he media. We expec o e s o use
he economic in o ma ion po ayed by he media a he han he mos accu a e economic in o ma ion.
The essay also in oduces p ojec ed economic g ow h o he cu en and he ollowing yea , which
is in o ma ion made public by he media.
The essay e eals ha p ojec ions o GDP g ow h, a he han ac ual o e ospec i e economic
pe o mance, a e he s onges p edic o o p ime minis e ial app o al in Po ugal, using qua e ly da a
om 2001 o 2018. This inding emphasizes ha o e s ely on media ed economic in o ma ion,
pa icula ly ega ding u u e expec a ions when assessing go e nmen pe o mance. The limi a ions o
his s udy include i s ocus on Po ugal, sugges ing a need o c oss-na ional esea ch o con i m he
applicabili y o hese indings. Addi ionally, u u e esea ch should examine indi idual-le el da a o
unde s and be e how economic o ecas s and media aming shape public pe cep ions and poli ical
judgmen s. This could enhance o ecas ing models in he poli ical economy by inco po a ing eal- ime
p ojec ions, po en ially imp o ing he accu acy o elec ion o ecas s.
Chap e 3 p esen s ano he me hodological al e na i e by conside ing a speci ic poli ical sys em:
he semi-p esiden ialism. I in es iga es whe he he dynamics obse ed in o he semi-p esiden ial
con ex s—whe e cohabi a ion o en shi s economic accoun abili y be ween execu i e leade s—apply in
Po ugal, whe e he p esiden is he head o s a e and he p ime-minis e is he head o he execu i e
86
b anch. The e o e, i assesses whe he he ela ionship be ween bo h poli ical en i ies, unde cohabi a ion
o uni ied go e nmen , changes he esponsibili y o economic de elopmen . Addi ionally, i seeks o
iden i y po en ial asymme ies in how economic e alua ions a ec p ime minis e s e sus p esiden s.
The essay inds ha in Po ugal’s semi-p esiden ial sys em, be ween 1986 and 2018, p ime
minis e s, no p esiden s, a e p ima ily held accoun able o economic pe o mance, whe he unde
cohabi a ion o uni ied go e nmen . A no able excep ion a ises du ing con lic ual cohabi a ions in second
p esiden ial e ms, whe e p esiden ial opposi ion seems o blu economic accoun abili y, making public
app o al o bo h leade s less sensi i e o economic condi ions. We also obse e an asymme y: economic
down u ns impac p ime minis e ial app o al mo e han p esiden ial, e lec ing di e ing pa isan
expec a ions. This insigh in o Po ugal’s V-P dynamics sugges s ha in simila semi-p esiden ial sys ems,
economic accoun abili y may simila ly a y ac oss leade s, especially unde cohabi a ion con lic s. Fu u e
esea ch should explo e hese pa e ns in compa able poli ical con ex s.
Chap e 4 aims o discuss a po en ial solu ion o he uns able esul s o he impac o he
economy on popula i y by u ilizing a new se o ools. This essay explo es how wa ele analysis can help
poli ical economic esea che s ci cum en some adi ional econome ic limi a ions. By applying
mul i a ia e wa ele analysis, which allows a ime- equency domain analysis, his essay seeks o cap u e
bo h s able long- e m pa e ns and ansien luc ua ions in poli ical app o al. To do so, we ocused on
Ba ack Obama’s p esidency, he las wo- e m US p esiden (2009-2017), which was associa ed wi h he
economic eco e y a e he 2009 G ea Dep ession.
The esul s show a he e ogeneous ela ionship be ween economic a iables and Obama’s
app o al a ing. T adi ional models e eal ha unemploymen and in la ion ha e a long- e m nega i e
e ec on Obama’s app o al, wi h consume sen imen a ec ing app o al only in sho - un dynamic
models. The wa ele analysis con i ms hese indings, adding de ail by showing ha unemploymen
in luences app o al a ings mo e quickly han in la ion despi e bo h leading app o al a e changes o e
he long un. Addi ionally, wa ele analysis e eals a posi i e, ye ansien , connec ion be ween consume
sen imen and app o al in 2011-2012, co esponding o he i s Obama deb -ceiling c isis e en . While
wa ele analysis o e s aluable insigh s in o he iming and equency o hese ela ionships, u u e
esea ch could ex end his me hod o examine o he poli ical con ex s in he poli ical economy and poli ical
science ields.
Finally, Chap e 5 aims o in es iga e he poli ical connec ions in publicly aded Po uguese
companies by examining he poli ical backg ounds o boa d membe s and hei po en ial in luence on
i m pe o mance. Focusing on public ade i ms lis ed on he Eu onex Lisbon s ock exchange o e he
87
pe iod 2002-2022, we cons uc ed h ee unique da ase s de ailing boa d membe s, Po uguese
poli icians, and i ms’ inancial ou comes. Ou objec i e is o p o ide insigh s in o he cha ac e is ics o
i ms ha ha e poli ical connec ions and how poli ical ies may in luence inancial ou comes.
Ou indings e eal ha la ge i ms, hose based in Lisbon, and hose in egula ed sec o s a e
mo e likely o ha e poli ical connec ions, hough hese connec ions ha e weakened o e ime and ha e
a limi ed impac on key inancial ou comes. Righ -leaning poli ically connec ed membe s a e mo e
common on la ge boa ds while le -leaning connec ions a e mo e p e alen in i ms headqua e ed in
Lisbon. Impo an ly, poli ical connec ions do no appea o signi ican ly impac inancial pe o mance.
Limi a ions o his s udy include he na ow ocus on publicly aded i ms and he speci ic measu emen
o poli ical connec ions, which may miss o he o ms o poli ical in luence. Fu u e esea ch could expand
his analysis o include p i a ely held i ms and al e na i e poli ical in luence channels o build a mo e
comp ehensi e unde s anding o poli ical ies in co po a e pe o mance and in he o e all economy.
88
Bibliog aphy
Acemoglu, D., Hassan, T. A., & Tahoun, A. (2017). The powe o he s ee : E idence om egyp ’s a ab
sp ing.
The Re iew o Financial S udies
,
31
(1), 1-42.
Acemoglu, D., Johnson, S., Ke mani, A., Kwak, J., & Mi on, T. (2016). The alue o connec ions in
u bulen imes: E idence om he Uni ed S a es.
Jou nal o Financial Economics
,
121
(2), 368-
391.
Ace edo, D. N., Fogleman, C., & Daniel U a, J. (2017). Peasan s and banke s: Educa ion, consume
sen imen , and p esiden ial app o al.
P esiden ial S udies Qua e ly
,
47
(2), 230-244.
Ag awal, A., & Knoebe , C. R. (2001). Do some ou side di ec o s play a poli ical ole?.
The Jou nal o Law
and Economics
,
44
(1), 179-198.
Aguia ‐Con a ia, L., & Soa es, M. J. (2014). The con inuous wa ele ans o m: Mo ing beyond uni‐and
bi a ia e analysis.
Jou nal o economic su eys
,
28
(2), 344-375.
Aguia ‐Con a ia, L., Magalhães, P. C., & Soa es, M. J. (2012). Cycles in poli ics: Wa ele analysis o
poli ical ime se ies.
Ame ican Jou nal o Poli ical Science
,
56
(2), 500-518.
Aguia -Con a ia, L., Magalhães, P. C., & Soa es, M. J. (2013). The na ionaliza ion o elec o al cycles in
he Uni ed S a es: a wa ele analysis.
Public Choice
,
156
(3-4), 387-408.
Aguia -Con a ia, L., Ma ins, M. M., & Soa es, M. J. (2023). The Phillips cu e a 65: Time o ime and
equency.
Jou nal o Economic Dynamics and Con ol
,
151
, 104620.
Akcigi , U., Baslandze, S., & Lo i, F. (2023). Connec ing o powe : poli ical connec ions, inno a ion, and
i m dynamics.
Econome ica
,
91
(2), 529-564.
Al , J. E., Ma shall, J., & Lassen, D. D. (2016). C edible sou ces and sophis ica ed o e s: When does
new in o ma ion induce economic o ing?.
The Jou nal o Poli ics
,
78
(2), 327-342.
Amo e, M. D., & Bennedsen, M. (2013). The alue o local poli ical connec ions in a low-co up ion
en i onmen .
Jou nal o Financial Economics
,
110
(2), 387-402.
Ande son, C. J. (2007). The end o economic o ing? Con ingency dilemmas and he limi s o democ a ic
accoun abili y.
Annual Re iew o Poli ical Science
,
10
, 271-296.
Ande son, C. J., & T e do a, Y. V. (2003). Co up ion, poli ical allegiances, and a i udes owa d
go e nmen in con empo a y democ acies.
Ame ican Jou nal o Poli ical Science
,
47
(1), 91-109.
Beche , M., & Donnelly, M. (2013). Economic pe o mance, indi idual e alua ions, and he o e:
In es iga ing he causal mechanism.
The Jou nal o Poli ics
,
75
(4), 968-979.
Bélange , É., & Meguid, B. M. (2008). Issue salience, issue owne ship, and issue-based o e choice.
Elec o al S udies
,
27
(3), 477-491.
Be lemann, M., & Enkelmann, S. (2014). The economic de e minan s o US p esiden ial app o al: A
su ey.
Eu opean Jou nal o Poli ical Economy
,
36
, 41-54.
Be lemann, M., Enkelmann, S., & Kuhlenkaspe , T. (2015). Un a eling he ela ionship be ween
p esiden ial app o al and he economy: A mul idimensional semipa ame ic app oach.
Jou nal
o Applied Econome ics
,
30
(3), 468-486.
Be and, M., K ama z, F., Schoa , A., & Thesma , D. (2018). The cos o poli ical connec ions.
Re iew
o Finance
,
22
(3), 849-876.
Blau, B. M., B ough, T. J., & Thomas, D. W. (2013). Co po a e lobbying, poli ical connec ions, and he
bailou o banks.
Jou nal o Banking & Finance
,
37
(8), 3007-3017.
Boomgaa den, H. G., Van Spanje, J., Vliegen ha , R., & De V eese, C. H. (2011). Co e ing he c isis:
Media co e age o he economic c isis and ci izens’ economic expec a ions.
Ac a Poli ica
,
46
(4),
353-379.
Boubak i, N., Cosse , J. C., & Sa a , W. (2008). Poli ical connec ions o newly p i a ized i ms.
Jou nal
o Co po a e Finance
,
14
(5), 654-673.

89
Boukes, M., Dams a, A., & Vliegen ha , R. (2021). Media e ec s ac oss ime and subjec : How news
co e age a ec s wo ou o ou a ibu es o consume con idence.
Communica ion Resea ch
,
48
(3), 454-476.
Boyds un, A. E., High on, B., & Linn, S. (2018). Assessing he ela ionship be ween economic news
co e age and mass economic a i udes.
Poli ical Resea ch Qua e ly
,
71
(4), 989-1000.
B ende , A., & D azen, A. (2007). Elec o al iscal policy in new, old, and agile democ acies.
Compa a i e
Economic S udies
,
49
(3), 446-466.
B ody, R. (1991).
Assessing he p esiden : The media, eli e opinion, and public suppo
. S an o d
Uni e si y P ess.
B own, J. R., & Huang, J. (2020). All he p esiden 's iends: Poli ical access and i m alue.
Jou nal o
Financial Economics
,
138
(2), 415-431.
Bussolo, M., de Nicola, F., Panizza, U., & Va ghese, R. (2022). Poli ically connec ed i ms and p i ileged
access o c edi : E idence om Cen al and Eas e n Eu ope.
Eu opean Jou nal o Poli ical
Economy
,
71
, 102073.
Ca lin, R. E., Ha lyn, J., Hellwig, T., Lo e, G. J.,Ma inez-Galla do, C., & Singe , M. M. (2019).
Execu i e
App o al Da abase 2.0
. h p://www.execu i eapp o alda a.o g
Ca lin, R. E., Ha lyn, J., Hellwig, T., Lo e, G. J.,Ma inez-Galla do, C., & Singe , M. M. (2023).
In oduc ion: Ins abili y and go e nmen popula i y in he wen y- i s cen u y. In T. Hellwig, & M.
M. Singe (Eds.),
Economics and poli ics e isi ed: Execu i e app o al and he new calculus o
suppo
(pp. 1–31). Ox o d Uni e si y P ess.
Child, T. B., Massoud, N., Schabus, M., & Zhou, Y. (2021). Su p ise elec ion o T ump connec ions.
Jou nal o Financial Economics
,
140
(2), 676-697.
Choi, J. P., & Thum, M. (2009). The economics o poli ically-connec ed i ms.
In e na ional Tax and Public
Finance
,
16
, 605-620.
Choi, S. W., James, P., Li, Y., & Olson, E. (2016). P esiden ial app o al and mac oeconomic condi ions:
E idence om a nonlinea model.
Applied Economics
,
48
(47), 4558-4572.
Chzhen, K., E ans, G., & Pickup, M. (2014). When do economic pe cep ions ma e o pa y
app o al?.
Poli ical Beha io
,
36
(2), 291-313.
Claessens, S., Feijen, E., & Lae en, L. (2008). Poli ical connec ions and p e e en ial access o inance:
The ole o campaign con ibu ions.
Jou nal o Financial Economics
,
88
(3), 554-580.
Cohen, J. E. (2004). Economic pe cep ions and execu i e app o al in compa a i e pe spec i e.
Poli ical
Beha io
,
26
(1), 27-43.
Cos a, J., Fazenda, L., Honó io, C., Louçã, F., & Rosas, F. (2010).
Os donos de Po ugal: cem anos de
pode económico
(1910-2010)
. Edições A on amen o.
Coulomb, R., & Sangnie , M. (2014). The impac o poli ical majo i ies on i m alue: Do elec o al
p omises o iendship connec ions ma e ?
Jou nal o Public Economics
,
115
, 158-170.
Dams a, A., & Boukes, M. (2021). The economy, he news, and he public: A longi udinal s udy o he
impac o economic news on economic e alua ions and expec a ions.
Communica ion Resea ch
,
48
(1), 26-50.
Dassonne ille, R., & Lewis-Beck, M. S. (2014). Mac oeconomics, economic c isis and elec o al ou comes:
A na ional Eu opean pool.
Ac a Poli ica
,
49
(4), 372-394.
De Boe , S., & Kells ed , P. M. (2004). The poli ical (and economic) o igins o consume con idence.
Ame ican Jou nal o Poli ical Science
,
48
(4), 633-649.
De Sousa, L., & Mo iconi, M. (2013). Why o e s do no h ow he ascals ou ?—A concep ual amewo k
o analysing elec o al punishmen o co up ion.
C ime, Law and Social Change
,
60
, 471-502.
Desai, R. M., & Olo sgå d, A. (2011). The cos s o poli ical in luence: Fi m-le el e idence om de eloping
coun ies.
Qua e ly Jou nal o Poli ical Science
,
6
(2), 137-178.
90
Dicke son, B. (2016). Economic pe cep ions, p esiden ial app o al, and causali y: The mode a ing ole
o he economic con ex .
Ame ican Poli ics Resea ch
,
44
(6), 1037-1065.
Dinç, I. S. (2005). Poli icians and banks: Poli ical in luences on go e nmen -owned banks in eme ging
ma ke s.
Jou nal o Financial Economics
,
77
(2), 453-479.
Domadenik, P., P ašnika , J., & S ejna , J. (2016). Poli ical connec edness, co po a e go e nance, and
i m pe o mance.
Jou nal o Business E hics
,
139
(2), 411-428.
Duch, R. M., & S e enson, R. T. (2008).
The economic o e: How poli ical and economic ins i u ions
condi ion elec ion esul s
. Camb idge Uni e si y P ess.
Duch, R. M., Palme , H. D., & Ande son, C. J. (2000). He e ogenei y in pe cep ions o na ional economic
condi ions.
Ame ican Jou nal o Poli ical Science
, 635-652.
Du e ge , M. (1980). A new poli ical sys em model: Semi‐p esiden ial go e nmen .
Eu opean jou nal o
poli ical esea ch
,
8
(2), 165-187.
Elgie, R. (1999). The poli ics o semi-p esiden ialism
.
In: R. Elgie (Ed.),
Semi-p esiden ialism in Eu ope
(pp. 1-21)
.
Ox o d Uni e si y P ess.
Elgie, R. (2010). Semi-p esiden ialism, cohabi a ion and he collapse o elec o al democ acies, 1990–
2008.
Go e nmen and Opposi ion
,
45
(1), 29-49.
Elgie, R. (2018). Regime ypes, p esiden ial powe , and cla i y o economic esponsibili y. In R. Elgie (Ed.),
Poli ical Leade ship: A P agma ic Ins i u ionalis App oach
(pp. 85-125)
.
Palg a e Macmillan.
E ikson, R. S., MacKuen, M. B., & S imson, J. A. (2000). Banke s o peasan s e isi ed: economic
expec a ions and p esiden ial app o al.
Elec o al S udies
,
19
(2-3), 295-312.
Eu opean Commission (2014). Flash Eu oba ome e 374: Businesses' a i udes owa ds co up ion in
he EU. Eu opean Union. h ps://eu opa.eu/eu oba ome e /su eys/de ail/1077.
Eu opean Commission (2015). Flash Eu oba ome e 428: Businesses' a i udes owa ds co up ion in
he EU. Eu opean Union. h ps://eu opa.eu/eu oba ome e /su eys/de ail/2084.
Eu opean Commission (2024a). Flash Eu oba ome e 543: Businesses' a i udes owa ds co up ion in
he EU. Eu opean Union. h ps://eu opa.eu/eu oba ome e /su eys/de ail/3180.
Eu opean Commission (2024b). Special Eu oba ome e 548: Ci izens' a i udes owa ds co up ion in he
EU in 2024. Eu opean Union. h ps://eu opa.eu/eu oba ome e /su eys/de ail/3217
E ans, G., & Ande sen, R. (2006). The poli ical condi ioning o economic pe cep ions.
The Jou nal o
Poli ics
,
68
(1), 194-207.
Faccio, M. (2006). Poli ically connec ed i ms.
The Ame ican Economic Re iew
,
96
(1), 369-386.
Faccio, M. (2010). Di e ences be ween poli ically connec ed and nonconnec ed i ms: A c oss‐coun y
analysis.
Financial Managemen
,
39
(3), 905-928.
Faccio, M., & Hsu, H. C. (2017). Poli ically connec ed p i a e equi y and employmen .
The Jou nal o
Finance
,
72
(2), 539-574.
Faccio, M., & McConnell, J. J. (2023). Poli ical connec ions cause esou ce misalloca ion: E idence om
he all o ascism in I aly.
ECGI Wo king Pape Se ies in Finance
, (939).
Faccio, M., & Zingales, L. (2022). Poli ical de e minan s o compe i ion in he mobile elecommunica ion
indus y.
The Re iew o Financial S udies
,
35
(4), 1983-2018.
Faccio, M., Masulis, R. W., & McConnell, J. (2006). Poli ical connec ions and co po a e bailou s.
The
Jou nal o Finance
,
61
(6), 2597-2635.
Feijó, R. G. (2021). "Mode a ing powe " and pa isan p esiden s: Two empi ical cases. In R. G. Feijó
(Ed.),
P esiden s in semi-p esiden ial egimes
(pp. 129-149). Palg a e Pi o .
Fe guson, T., & Vo h, H. J. (2008). Be ing on Hi le — he alue o poli ical connec ions in Nazi
Ge many.
The Qua e ly Jou nal o Economics
,
123
(1), 101-137.
Fe nandes, F. (2018). Os emp esá ios de Ma cello Cae ano:
Os g andes anos de in es imen o público e
p i ado, en e 1968-1974, e as gue as emp esa iais
. Casa das Le as.
91
Fisman, R. (2001). Es ima ing he alue o poli ical connec ions.
The Ame ican Economic Re iew
,
91
(4),
1095-1102.
Fo es, B. G., & Magalhães, P. (2005). As eleições p esidenciais em sis emas semip esidenciais:
pa icipação elei o al e punição dos go e nos.
Análise Social
, 891-922.
F anco, V. S. D., 2020.
Semip esidencialismo em Po ugal: Os pode es p esidenciais na in e ação com
o go e no e com a Assembleia da República (1982-2016).
Lisbon: Assembleia da República.
F ei e, A., & Lobo, M. C. (2005). Economics, ideology and o e: Sou he n Eu ope, 1985–2000.
Eu opean
Jou nal o Poli ical Resea ch
,
44
(4), 493-518.
F ei e, A., & San ana-Pe ei a, J. (2012). Economic o ing in Po ugal, 2002–2009.
Elec o al S udies
,
31
(3), 506-512.
F ei e, A., & San ana-Pe ei a, J. (2019). The P esiden 's dilemma: The Po uguese semi-p esiden ial
sys em in imes o c isis (2011‐16).
In e na ional Jou nal o Ibe ian S udies
,
32
(3), 117-135.
Goidel, R. K., & Langley, R. E. (1995). Media co e age o he economy and agg ega e economic
e alua ions: Unco e ing e idence o indi ec media e ec s.
Poli ical Resea ch Qua e ly
,
48
(2),
313-328.
Goodha , C. A. E & Bhansali, R. J. (1970). Poli ical Economy.
Poli ical S udies
,
18
(1), 43-106.
Goula , P., & Veiga, F. J. (2016). Po uguese 2015 legisla i e elec ions: How economic o ing, he
median o e and unemploymen led o ‘ he imes hey a e a’changin’?
Elec o al S udies
,
43
,
197–200.
G onke, P., & Newman, B. (2003). FDR o Clin on, Muelle o?: A ield essay on p esiden ial app o al.
Poli ical Resea ch Qua e ly
,
56
(4), 501-512.
G ossman, E., & Guinaudeau, I. (2023). Cos o uling abo e all else: Explaining p esiden ial popula i y in
F ance. In T. Hellwig, & M. M. Singe (Eds.),
Economics and poli ics e isi ed: Execu i e app o al
and he new calculus o suppo
(pp. 80–107). Ox o d Uni e si y P ess.
G ündle , K., & Po a ke, N. (2019). Co up ion and economic g ow h: New empi ical e idence.
Eu opean
Jou nal o Poli ical Economy
,
60
, 101810.
Hao, Z., Liu, Y., Zhang, J., & Zhao, X. (2020). Poli ical connec ion, co po a e philan h opy and e iciency:
E idence om China’s an i-co up ion campaign.
Jou nal o Compa a i e Economics
,
48
(3), 688-
708.
Ha s ad, B., & S ensson, J. (2011). B ibes, lobbying, and de elopmen .
Ame ican Poli ical Science
Re iew
,
105
(1), 46-63.
Hellwig, T., & Ma ino a, D. M. (2017). Dead ends and new pa hs in he s udy o economic o ing.
In
Ox o d Resea ch Encyclopedia o Poli ics
. Ox o d Uni e si y P ess.
Hellwig, T., & Samuels, D. (2008). Elec o al accoun abili y and he a ie y o democ a ic egimes.
B i ish
Jou nal o Poli ical Science,
38
(1), 65-90.
Hibbs, D. A. (1977). Poli ical pa ies and mac oeconomic policy.
The Ame ican Poli ical Science Re iew
,
71
(4), 1467-1487.
Jalali, C. (2011). The p esiden is no a passenge . In R. Elgie, S. Moes up, & Y.-S. Wu (Eds.),
Semi-
p esiden ialism and democ acy
(pp. 156-173). Palg a e Macmillan.
Johnson, S., & Mi on, T. (2003). C onyism and capi al con ols: e idence om Malaysia.
Jou nal o
Financial Economics
,
67
(2), 351-382.
Kayse , M. A. (2014). The elusi e economic o e. In L. Leduc & R. Niemi (Eds.),
Compa ing democ acies:
Elec ions and o ing in a changing wo ld
(Vol. 4, pp.112-132). Thousand Oaks, CA: Sage.
Kayse , M. A., & Leininge , A. (2015). Vin age e o s: do eal- ime economic da a imp o e elec ion
o ecas s?.
Resea ch & Poli ics
,
2
(3), 1-11.
Kee e, M. O. C. (2019). A heo y o poli ical connec ions and inancial ou comes.
In e na ional Re iew o
Economics & Finance
,
61
, 108-127.
92
Khwaja, A. I., & Mian, A. (2005). Do lende s a o poli ically connec ed i ms? Ren p o ision in an
eme ging inancial ma ke .
The Qua e ly Jou nal o Economics
,
120
(4), 1371-1411.
Lacy, D., & Ch is enson, D. P. (2017). Who o es o he u u e? In o ma ion, expec a ions, and
endogenei y in economic o ing.
Poli ical Beha io
,
39
(2), 347-375.
Leh e , N. D. (2018). The alue o poli ical connec ions in a mul ipa y pa liamen a y democ acy: e idence
om he 2015 elec ions in Is ael.
Eu opean Jou nal o Poli ical Economy
,
53
, 13-58.
Lewis‐Beck, M. S. (1997). Who's he che ? Economic o ing unde a dual execu i e.
Eu opean Jou nal o
Poli ical Resea ch
,
31
(3), 315-325.
Lewis-Beck, M. S. (2004). F ench elec ion heo ies and he 2002 esul s: An in oduc ion. In M. S. Lewis-
Beck (Ed.),
The F ench o e
(pp. 1-11). Palg a e Macmillan.
Lewis-Beck, M. S., & Nadeau, R. (2004). Dual go e nance and economic o ing: F ance and he Uni ed
S a es. In M. S. Lewis-Beck (Ed.), The F ench o e (pp. 136-154). Palg a e Macmillan.
Lewis-Beck, M. S., & Nadeau, R. (2012). PIGS o no ? Economic o ing in Sou he n Eu ope.
Elec o al
S udies
,
31
(3), 472-477.
Lewis-Beck, M. S., & S egmaie , M. (2000). Economic de e minan s o elec o al ou comes.
Annual Re iew
o Poli ical Science
,
3
(1), 183-219.
Lewis-Beck, M. S., & S egmaie , M. (2013). The VP- unc ion e isi ed: a su ey o he li e a u e on o e
and popula i y unc ions a e o e 40 yea s.
Public Choice
,
157
, 367-385.
Lijpha , A. (2012).
Pa e ns o democ acy: Go e nmen o ms and pe o mance in hi y-six coun ies
.
Yale Uni e si y P ess.
Lobo, M. C., & Pannico, R. (2020). Inc eased economic salience o blu ing o esponsibili y? Economic
o ing du ing he G ea Recession.
Elec o al S udies
,
65
, 102141.
MacKuen, M. B., E ikson, R. S., & S imson, J. A. (1992). Peasan s o banke s? The Ame ican elec o a e
and he US economy.
Ame ican Poli ical Science Re iew
,
86
(3), 597-611.
Magalhães, P. C. (2001). As a mas dos acos: O e o polí ico e a li igância cons i ucional do P esiden e
da República. In
Ac as do I Encon o Nacional de Ciência Polí ica
. Bizâncio.
Magalhães, P. C. (2017). A ale o wo elec ions: in o ma ion, mo i a ed easoning, and he economy in
he 2011 and 2015 Po uguese elec ions.
Análise Social
,
52
(225), 736-758.
Magalhães, P. C., & Aguia -Con a ia, L. (2009). G ow h, cen ism and semi-p esiden ialism: Fo ecas ing
he Po uguese gene al elec ions.
Elec o al S udies
,
28
(2), 314-321.
Mandle , M., & Scha nagl, M. (2014). Money g ow h and consume p ice in la ion in he eu o a ea: A
wa ele analysis.
Bundesbank Discussion Pape
, (33).
Ma ins, P. S. (2020). Jobs C onyism in Public-Sec o Fi ms.
GLO Discussion Pape
, (624).
Maydeu-Oli a es, A., Shi, D., & Rosseel, Y. (2018). Assessing i in s uc u al equa ion models: A Mon e-
Ca lo e alua ion o RMSEA e sus SRMR con idence in e als and es s o close i .
S uc u al
Equa ion Modeling: A Mul idisciplina y Jou nal
,
25
(3), 389-402.
Moon, T., & Schoenhe , D. (2022). The ise o a ne wo k: Spillo e o poli ical pa onage and c onyism
o he p i a e sec o .
Jou nal o Financial Economics
,
145
(3), 970-1005.
Muelle , J. E. (1970). P esiden ial Popula i y om T uman o Johnson.
Ame ican Poli ical Science
Re iew
,
64
(1), 18-34.
Nannes ad, P., & Paldam, M. (1997). The g ie ance asymme y e isi ed: A mic o s udy o economic
o ing in Denma k, 1986–1992.
Eu opean Jou nal o Poli ical Economy
,
13
(1), 81-99.
Nannes ad, P., & Paldam, M. (2003). The cos o uling: A ounda ion s one o wo heo ies. In E.
Do ussen & M. Taylo (Eds.),
Economic o ing
(pp. 17-44). Rou ledge.
Ne o, O. A., & Lobo, M. C. (2009). Po ugal's semi‐p esiden ialism ( e) conside ed: An assessmen o he
p esiden 's ole in he policy p ocess, 1976–2006.
Eu opean Jou nal o Poli ical Resea ch
,
48
(2),
234-255.
99
Table A4: Ma ix o co ela ions. Own calcula ions.
Boa d
(N)
Connec
ion
Connec
ion
Righ
Connec
ion
Le
Connec
ion (N)
To al
Asse s
Ma ke
Cap
ROE
Le e ag
e
Book o
Ma ke
Ma ke
Sha e
P oduc
i i y
Re u ns
P i a iz
ed
S a e
Owne s
hip
Righ
Lisbon
T oika
Regula
ed
Boa d (N)
1.000
Connec ion
0.587
1.000
Connec ion Righ
0.420
0.625
1.000
Connec ion Le
0.109
0.362
-0.167
1.000
Connec ion (N)
0.691
0.703
0.494
0.247
1.000
To al Asse s
0.682
0.369
0.384
0.009
0.508
1.000
Ma ke Cap
0.629
0.416
0.232
0.239
0.581
0.492
1.000
ROE
0.122
0.057
0.032
0.073
0.057
0.044
0.178
1.000
Le e age
-0.018
-0.064
-0.044
-0.001
0.038
0.144
-0.142
-0.288
1.000
Book o Ma ke
-0.128
-0.102
-0.056
-0.035
-0.095
-0.041
-0.125
-0.100
0.288
1.000
Ma ke Sha e
0.413
0.282
0.214
0.128
0.337
0.324
0.560
0.192
-0.217
-0.100
1.000
P oduc i i y
0.085
0.085
-0.001
0.076
0.090
-0.215
0.224
0.100
-0.214
-0.155
0.125
1.000
Re u ns
-0.011
-0.017
0.013
0.010
0.008
-0.008
0.095
0.196
-0.228
-0.075
0.052
0.025
1.000
P i a ized
0.509
0.377
0.268
0.125
0.365
0.217
0.448
0.083
-0.060
-0.049
0.319
0.290
-0.050
1.000
S a e Owne ship
0.253
0.240
0.079
0.244
0.290
0.061
0.361
0.044
-0.011
0.024
0.238
0.379
-0.014
0.627
1.000
Righ
-0.018
0.040
0.003
0.005
0.014
0.011
-0.063
-0.132
0.017
0.010
-0.042
0.000
0.107
-0.002
0.005
1.000
Lisbon
0.048
0.194
0.073
0.164
0.118
-0.099
0.092
-0.069
0.021
0.030
0.201
0.135
-0.110
0.364
0.207
0.022
1.000
T oika
0.035
-0.095
-0.032
-0.036
-0.064
-0.005
-0.010
0.007
0.121
0.128
0.064
0.000
0.058
-0.010
-0.133
0.125
-0.039
1.000
Regula ed
0.549
0.413
0.352
0.128
0.427
0.496
0.451
0.044
-0.049
-0.134
0.179
0.008
-0.029
0.332
0.253
-0.016
0.093
0.016
1.000