Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=fjcs21
East European Politics
ISSN: 2159-9165 (Print) 2159-9173 (Online) Journal homepage: https://www.tandfonline.com/loi/fjcs21
Potent executives: the electoral strength of prime
ministers in Central Eastern Europe
Jan Berz
To cite this article: Jan Berz (2019): Potent executives: the electoral strength of prime ministers in
Central Eastern Europe, East European Politics, DOI: 10.1080/21599165.2019.1662400
To link to this article: https://doi.org/10.1080/21599165.2019.1662400
Published online: 11 Sep 2019.
Submit your article to this journal
Article views: 58
View related articles
View Crossmark data
Potent executives: the electoral strength of prime ministers in
Central Eastern Europe
Jan Berz
Leuphana University of Lüneburg, Lüneburg, Germany
ABSTRACT
Prime ministers in Central Eastern Europe (CEE) are commonly
perceived to be weaker and less “presidential” than their Western
European colleagues. In contrast, I argue that the post-communist
legacy of CEE provides a more favourable context for prime
ministers to develop a central characteristic of “presidential” prime
ministers: a strong personal inﬂuence on voting behaviour in
parliamentary elections, which provides prime ministers with the
opportunity to gain autonomy within their party and cabinet. I
theorise that party system instability, as well as the comparatively
lower media independence and freedom, gives greater electoral
potential to prime ministers in CEE than to their Western
European counterparts. Consequently, prime ministers in CEE have
a stronger inﬂuence on vote choice than West European prime
ministers. These hypotheses are investigated using survey data
from several waves of the CSES project.
ARTICLE HISTORY
Received 5 March 2019
Accepted 17 August 2019
KEYWORDS
Prime minister;
presidentialisation;
performance; voting
behaviour; leader eﬀect;
Central Eastern Europe
1. Introduction
Are prime ministers in Central Eastern Europe (CEE) weaker than prime ministers in
Western Europe (WE)? The notion that prime ministers in CEE are weaker is widely
accepted (Baylis 2007; Elgie 2012) and relies on the operationalisation of prime ministerial
strength as duration in oﬃce. In this paper, I challenge this conclusion by comparing prime
ministers in CEE and WE over another dimension of strength: their personal inﬂuence on
individual vote choice in parliamentary elections as measure of electoral strength. Throughout
this paper, I refer to this personal inﬂuence on citizens’ vote choice with the commonly
used term leader eﬀect (Mughan 2015). The increasing personalisation of electoral behaviour
and personal inﬂuence of leading candidates on voting behaviour has received
growing attention in recent years (Bittner 2011, 2018; Costa and Ferreira da Silva 2015; Ferreira
da Silva 2018; Garzia 2014; King 2002b; Lobo and Curtice 2014), as parliamentary elections
tend to more closely resemble presidential elections. I argue that a prime minister’s
personal eﬀect on voting behaviour in parliamentary elections, their electoral strength, can
be used as a measure of a prime minister’s performance.1
In addition, the more prime ministers
gain a “presidential” role in parliamentary elections, the more authority and autonomy
they may gain within their party and cabinet (Poguntke and Webb 2005; Worthy
2016).
© 2019 Informa UK Limited, trading as Taylor & Francis Group
CONTACT Jan Berz jberz@leuphana.de
EAST EUROPEAN POLITICS
https://doi.org/10.1080/21599165.2019.1662400
Prime ministers in CEE are more likely to hold a presidential connection to voters than
prime ministers in WE, because the comparatively lower voter alignment and party system
stability in CEE provide a favourable context for presidentialised parliamentary elections
(Bértoa and Mair 2012; Mair 1997; Poguntke and Webb 2005). Furthermore, elites in CEE
are able to exert more inﬂuence over institutions of mass communications (Bairett 2015;
Bajomi-Lázár 2013; Dragomir 2018; Garzia 2017a).
I use several waves of Comparative Study of Electoral Systems (CSES) data to study the
leader eﬀect of prime ministers across European countries. The hypothesis that will be
tested is that prime ministers in CEE have a stronger direct personal eﬀect on vote
choice (leader eﬀect) than prime ministers in WE. I employ multilevel models to test
this hypothesis, analysing whether the leader eﬀect of prime ministers varies across
countries and if this variance can be attributed to a systematic diﬀerence between CEE
and WE countries.
I ﬁnd that the leader eﬀect of prime ministers in parliamentary elections is signiﬁcantly
larger in CEE than in WE. This ﬁnding also remains robust to the exclusion of single
countries and prime ministers who the literature considers to be outliers. A further analysis
shows that diﬀerences between CEE and WE can be explained by diﬀerences in media
freedom and party system stability. Furthermore, this paper discusses to what extent
the presented ﬁndings present a puzzle for the presidentialisation thesis, since prime ministers
in CEE seem unable to use their electoral strength to secure their survival in oﬃce.
The remaining part of the paper proceeds as follows: ﬁrst, I discuss why a prime minister’s
personal inﬂuence on individual voting behaviour constitutes an important dimension
of strength. Second, I argue that prime ministers in CEE may be electorally stronger
than prime ministers in WE. Third, I discuss the selection of my cases and data. Fourth, I
provide information on my statistical model and review their ﬁndings. Fifth, I draw my conclusions
and outline opportunities for further research.
2. Prime ministers in Central Eastern and Western Europe
In this section, I argue that a prime minister’s personal impact on vote choice, their leader
eﬀect, fosters their performance. The ability of prime ministers to personally inﬂuence vote
choice is an essential aspect of “presidential” prime ministers and can provide prime ministers
with the opportunity to gain autonomy within their party and cabinet.
How well individual prime ministers perform and what factors may explain diﬀerences
in performance have been a long-standing interest and debate in the study of democratic
government (Helms 2017; Kaarbo and Hermann. 1998; Strangio, Hart, and Walter 2013).
However, it is generally accepted that prime ministers in CEE are weaker and less “presidential”
than prime ministers in WE (Baylis 2007; Elgie 2012). Although the new democracies
in CEE adapted well to the new system of government (Blondel et al. 2007), prime
ministers in CEE stay in oﬃce for shorter periods than prime ministers in WE. This duration
gap still holds true to the present day, although duration varies within CEE and in some
countries prime ministers stay in oﬃce over periods that are comparable to periods in
WE (Grotz and Müller-Rommel 2015). Baylis (2007) argues that a prime minister’s duration
in oﬃce approximates his potential to implement policies and therefore indicates his or
her strength. However, a prime minister’s duration in oﬃce remains a rather rough
approximation of their policy performance. While a short-serving prime minister has
2 J. BERZ
little time to enact policies, nothing guarantees that a long-serving prime minister will
achieve more. Unfortunately, the use of policy achievements themselves is met by two
major diﬃculties. First, assessing policy achievements comparatively over a long period
of time. Second, deciding whether these achievements are attributable to the prime minister
or rather to the cabinet as a whole since democratic governance in parliamentary
systems is based on joint-decision making (Blondel and Müller-Rommel 1993).
In light of these limitations, I turn to another accessible measure of strength and “presidential”
position in oﬃce: a prime minister’s personal inﬂuence on citizen’s vote choice in
parliamentary elections, their leader eﬀect.
In their seminal work on the concept of presidentialisation, Poguntke and Webb (2005)
describe how parliamentary elections start to resemble presidential elections and increasingly
focus on the competition among leading candidates. Together with the internationalisation
of politics and growing complexity of the state these changes enable chief
executives in parliamentary systems to gain increasing control and autonomy similar to
presidents, even under the continuation of existing formal parliamentary procedures.
Under the growing presidentialisation of parliamentary elections, it becomes essential
for prime ministers to inﬂuence voters to vote in favour of their party. This inﬂuence is
commonly described as leader eﬀect: the direct personal inﬂuence on vote choice in parliamentary
election (Garzia 2017b; King 2002a; Mughan 2015). Weakening social cleavages
and changes in communication give greater weight to the competition between leading
candidates and their personality (Bittner 2011, 2018; Blondel and Thiébault 2009; Garzia
2014; Rahat and Sheafer 2007). Performing on this dimension becomes an essential
duty of prime ministers. As citizens continue to base their vote choice to a larger extent
on their impression of the prime ministers and leaders who become the “chief means
of engaging the political interest of publics” (Poguntke and Webb 2005, 21), parties will
rely more heavily on the prime minister for political success.
In terms of recent conceptions of executive authority as leadership capital (Bennister,
Hart, and Worthy 2015; Helms 2016), leader eﬀects reach beyond reputational sources
like popularity. Leader eﬀects evidence a relational change in the leader–follower relationship
(Garzia 2011) between prime ministers and the electorate, which alters the role of
prime ministers and other leaders in the political processes of parliamentary systems.
Curtice and Holmberg (2005) show empirically that leader eﬀects are not a reﬂection of
a leader’s popularity.2
Under the concept of presidentialisation, a prime minister’s presidential role in parliamentary
elections may impact their position within party and cabinet (Poguntke and
Webb 2005, 17) and allows them to govern past their parties. It is plausible for prime ministers
to use this increased importance in parliamentary elections to foster intra-party and
intra-executive presidentialisation to their own beneﬁt. The more parties need to rely on
prime ministers to engage potential voters, the more likely they will give way to the prime
minister’s policy goals in order to maintain the presidential connection between the prime
minister and the electorate. Prime ministers who hold a presidential connection with the
electorate also hold a stronger bargaining position, because they are able to reach past
their party and steer the electorate against intra-party or intra-cabinet opposition. This
entails the plausible assumption that prime ministers will prefer to work towards and campaign
with the policies which they themselves favour and can argue for with conviction.
Worthy (2016) similarly argues that elections provide a basic measure of a prime minister’s
EAST EUROPEAN POLITICS 3
performance and grant them authority. Hence, I understand a prime minister’s leader
eﬀect as a direct and indirect measure of strength:
(1) Prime minister’s leader eﬀects, their electoral strength, are a direct measure of their
“presidential” inﬂuence in parliamentary elections and relationship with the electorate.
Prime minister’s performance on this dimension is an essential function in increasingly
presidentialised parliamentary systems.
(2) Larger leader eﬀects enable prime ministers to foster their autonomy within cabinet
and party and govern past their party. Prime ministers with greater autonomy
within their cabinet and party are likely to secure their policy goals against opposing
goals of colleagues. This constitutes the additional indirect contribution of a prime
minister’s electoral strength to their overall strength.
In summary, a prime minister’s leader effects provide a good indication of their strength
and “presidential” position. This conception of prime ministerial strength complements
duration in ofﬁce as common operationalisation of performance by an additional and
empirically observable component. It follows that a prime minister’s strength is best
understood as a combination of factors. A prime minister with a long duration in ofﬁce,
but without sufﬁcient autonomy from their cabinet may achieve less in terms of policies,
than an autonomous prime minister with a shorter tenure. While it is undeniable that
many prime ministers in CEE have the disadvantage of their comparatively short period
in ofﬁce, they may enjoy advantages in other dimensions of strength. In this paper, I set
out to explore differences in the electoral aspect of prime ministerial strength between
CEE and WE empirically. In the next section, I theorise why such a difference between
prime ministers in CEE and WE may exist.
3. Presidentialisation of parliamentary elections
In the following paragraphs, I discuss why leader eﬀects of prime ministers are likely to be
systematically larger in CEE than in WE. I argue that diﬀerences in social cleavages and
resulting party system instability, as well as in media independence provide leading candidates,
and in particular prime ministers, with a favourable playing ground to personally
inﬂuence vote choice and discuss whether these measures diﬀer between CEE and WE.
Poguntke and Webb (2005) identify the increasing focus on competing candidates in
parliamentary elections as a key driver of presidentialisation and increasing autonomy
of prime ministers. Over the past decades, electoral research has overwhelmingly
secured such an increasing electoral impact of voters’ perception of leading candidates
in parliamentary elections (Bittner 2018; Garzia 2014, 2017b; Lobo and Curtice 2014;
Mughan 2015; Wattenberg 1991). The presidentialisation of parliamentary elections is
driven by societal changes and changes in communication technology, namely the
erosion of societal cleavages in Western democracies and the success of television as
medium of mass communication. Changes in media technology have frequently been
identiﬁed as drivers of electoral presidentialisation and give individual candidates
greater importance (Lenz and Lawson 2011; Mughan 2000; Sartori. 1989). Recent research
has shown that especially the consumption of television moderates the inﬂuence of
leaders in parliamentary elections (Garzia 2017a). The more television voters’ consume
4 J. BERZ
the greater is the eﬀect of leading candidate’s personality on their vote choice. Electoral
presidentialisation therefore shares common causes with personalisation (Rahat and
Sheafer 2007), a process in which eroding societal cleavages and the changing structure
of mass communication increase the weight of individual political actors relative to political
groups. With regard to these two causes of electoral presidentialisation, media structure
and cleavages, CEE diﬀers from WE.
First, the newer democracies in CEE faced several simultaneous challenges of political
transformation (Lewis 2002; Tworzecki 2003; Millard 2004). Departing from the absence of
political competition, dealignment has, therefore, been the starting hypothesis of party
systems in these post-communist countries (Mair 1997) and this hypothesis has largely
held true (Bértoa and Mair 2012). Although there is evidence for increasing organisational
consolidation of party systems in CEE, alignment between voters and parties is still lacking
(Weßels and Klingemann 2006). These contextual factors foster the presidentialisation of
parliamentary elections and provide prime ministers in CEE with greater potential to
inﬂuence vote choice. Since the link between parties and voters is weaker than in WE,
voters are likely to consider individual leaders and their personalities to a greater extent
than in WE when casting their ballot (Grotz and Weber 2017). Voters are more likely to
identify with a speciﬁc leading candidate in a given parliamentary election than with
the candidate’s party as a whole. In turn the leader eﬀects of prime ministers in CEE will
likely be larger, compared to the leader eﬀects of prime ministers in WE where the alignment
between voters and parties is stronger.
Second, elites in CEE are able to exert greater inﬂuence over the media than elites in
WE. Like a transformation of the party system, a transformation of media institutions in
CEE was necessary after the breakdown of communism (Jakubowicz 2001). The new
democracies in CEE needed to reform party- and state-owned media systems to a democratised
media (Dobek-Ostrowska 2015; Splichal. 1992). Despite this transformation, a
greater inﬂuence of political elites on media institutions remains (Jakubowicz 2001;
Örnebring 2012) in which the media is to some extent a resource of elites and clientelistic
networks. Hanretty (2010) ﬁnds that among a large number of European public broadcasters,
and a few broadcasters outside of Europe, broadcasters in CEE are evaluated to be
the least independent. In line with these ﬁndings (Bairett 2015) shows how executives in
CEE actively reduce media freedom to avoid public scrutiny of their actions in oﬃce.
Hungary provides a particular visible case (Bajomi-Lázár 2017) in which the strategy
has been described as a party colonialisation of the media (Bajomi-Lázár 2013). The
pattern is not limited to countries in CEE. Dragomir (2018) shows that after the economic
crisis regimes around the world have made increasing use of government funding to
exert inﬂuence over media outlets. Overall, existing research supports the notion that
PM are aware how media coverage of them can aﬀect their perception among the electorate
negatively. In summary, not only do media structures diﬀer between WE and CEE,
executives in CEE also continually seek to alter the public playing ﬁeld in their favour. I
argue that this higher degree of media inﬂuence and lower degree of media independence
and freedom allow prime ministers in CEE to exert greater inﬂuence on voters
than prime ministers in WE. Prime ministers should be able to use their elite position
at the top of the executive to shape mass communication on their personal image and
importance to their liking. Although other leading candidates may be likewise able to
exert their elite inﬂuence over the media, the executive position of prime ministers
EAST EUROPEAN POLITICS 5
provides them with a clear advantage in political systems departing from widespread
state control over media.
In conclusion, I hypothesise that due to the discussed diﬀerences in media independence
and party institutionalisation the electoral strength of prime ministers will be
larger in CEE than in WE:
H1: Leader eﬀects of prime ministers in Central Eastern Europe are larger than in Western
Europe.
While leading candidates without a position as prime minister may have fewer
resources to shape media reporting to their liking, they still proﬁt from the weakened cleavages
and resulting volatile party systems in CEE democracies to the same extent. I
hypothesise that their leader eﬀect is also likely to be larger than their counterparts in
Western Europe:
H2: Leader eﬀects of other leading candidates are larger in Central Eastern Europe than in
Western Europe.
King (2002a) argues that voters consider party leader when casting their vote because
party leaders may choose to follow their personal policy preferences over their party line
and that the personality of prime ministers may inﬂuence the decisions made by their government.
Consequently, the weight of prime ministers in voters’ calculus may be moderated
by their impact on decision making in parliamentary systems, therefore, the leader
eﬀect of prime ministers in semi-presidential systems may be weaker as executive
power (Baylis 1996; Sedelius and Mashtaler 2013) is shared:
H3: Leader eﬀects of prime ministers and other leading candidates in semi-presidential
systems are smaller.
Since voters may also like a prime minister or party leader, simply because they belong
to a speciﬁc party (Curtice and Holmberg 2005) I will include voters’ party identiﬁcations
(Campbell et al. 1966) as the main explanatory factor opposing the leader eﬀect of prime
ministers. I also include voters’ education level, gender and age as further control variables.
In the following section, I discuss the selection of my cases and the available data.
4. Data and case selection
I use the harmonised Comparative Study of Electoral Systems (CSES) ﬁle (Giebler et al.
2016) as well as the fourth wave (CSES 2018) to test whether the electoral strength of
prime ministers diﬀers between CEE and WE. These data sources allow for the broadest
comparison across diﬀering countries and over time to minimise potential selection
biases. Unfortunately, elections in the second CSES wave have to be excluded since
voters’ evaluation of leaders are not available. I further exclude presidential systems, following
the deﬁnition of Elgie (2018) and non-European countries. In addition, parliamentary
elections in which the prime minister did not run as a leading candidate or in which
the government was led by a caretaker are excluded as well.3
I also exclude the 2009 parliamentary
election in Iceland in which the prime minister came into oﬃce at less than 2
months prior to the election. It is probable that such a short time period is insuﬃcient to
prepare an electoral campaign designated to the prime minister. The ﬁnal selection
6 J. BERZ
encompasses 22 countries and 44 individual prime ministers. A list of included countries,
elections and prime ministers can be found in Table A1 in the Appendix.
My dependent variable measures whether respondents voted for the prime minister’s
party (1) or did vote for any other party (0). This negates diﬀerences between the parties
who do not hold the oﬃce of prime minister, but eases country-comparison and allows for
a more parsimonious model which focuses on the vote choice of interest. I include a
number of individual-level predictors of vote choice: ﬁrst, voters’ thermometer evaluations
of prime ministers and of party leaders of opposing parties. These thermometer evaluations
range from dislike (0) to like (10). I summarise all party leaders opposing the
prime minister in one variable by taking a respondents’ highest evaluation of a party
leader. I furthermore measure whether respondents hold a party identiﬁcation with the
party of the prime minister or any of the other parties. Lastly, respondents’ gender, age
and education level are included as control variables. To test H3 a country-level variable
is created that diﬀerentiates between parliamentary and semi-presidential systems,
using the classiﬁcation by Elgie (2018).
The following section discusses my model strategy and results. I provide graphical
results to facilitate interpretation and give full regression results in the Appendix
(Tables A2 and A3).
5. Analysis
I model the contextual factors of interest by employing a logistic multilevel model with a
individual- and a country-level that allows for varying intercepts by country in which individual
respondents are nested.4
The model includes a random slope for voters’ evaluation
of prime ministers, which allows for the inﬂuence of voters’ evaluation of prime ministers
to vary between each country. Whether there is a systematic variation in prime ministerial
inﬂuence between CEE and WE (H1), regions who diﬀer in media independence and party
institutionalisation, is tested by a cross-level interaction (Eval. prime minister x CEE). I also
allow for the slope of voters’ evaluations of other leaders to vary by country to test
whether the inﬂuence of other leaders (H2) signiﬁcantly varies between CEE and WE.
Non-dichotomous individual-level variables are centred on country means and standardised
by two standard deviations to ease the comparison of eﬀects. Two models are
ﬁtted: ﬁrst, a main model that tests the cross-level interactions between voters’ evaluation
of prime ministers and CEE, as well as voters’ evaluation of other leaders and CEE. Second,
a semi-presidential model in which I also test whether the inﬂuence of prime ministers
and other leaders signiﬁcantly varies between parliamentary and semi-presidential
systems (H3).5
Figure 1 shows the estimated median eﬀects from the main regression model. All explanatory
variables, voters’ evaluation the prime minister, party identiﬁcation with the prime
minister’s party, have a statistically signiﬁcant and positive eﬀect on vote choice for the
prime minister’s party. In contrast voters who identify with a diﬀerent party or evaluate
other party leaders positively are signiﬁcantly less likely to vote for the prime minister’s
party. The control variables age, gender and education level show no relevant inﬂuence
on vote choice.
Figure 1 also shows that the eﬀect of voters’ evaluation of the prime minister on vote
choice, prime ministers electoral strength, diﬀers systematically between CEE and WE (H1).
EAST EUROPEAN POLITICS 7
The respective cross-level interaction “Eval. prime minister x CEE” is signiﬁcant. The eﬀect is
signiﬁcantly larger in CEE than in WE, therefore, prime ministers in CEE have a larger personal
impact on individual vote choice than prime ministers in WE. Similarly the eﬀect of
other party leaders on vote choice is also larger in CEE than in WE (H2), although the conﬁdence
interval is closer to zero. The results of both interaction eﬀects match the previously
derived expectations.
Figure 2 provides a visual representation of the interaction eﬀect between voters’ evaluation
of prime ministers and CEE. The more positively voters evaluate the prime minister,
the more likely they are to vote for the prime minister’s party. This eﬀect holds true for
prime ministers in WE and in CEE, but the eﬀect is larger in CEE. Prime ministers in CEE
are, therefore, more “presidential” than their counterparts in WE as they have a larger
direct eﬀect on voters’ decision in parliamentary elections. Due to this larger leader
eﬀect of prime ministers, parliamentary elections in CEE more closely resemble presidential
elections than parliamentary elections in WE. However, it may be the case that the
diﬀerence in prime ministers leader eﬀects between the two regions is driven by diﬀerences
between semi-presidential and parliamentary systems. As stated in hypothesis,
three leader eﬀects of prime ministers may be smaller in semi-presidential systems than
in parliamentary systems. I test this assumption in an additional model, but ﬁnd no signiﬁcant
diﬀerence of voters’ evaluation of the prime minister between semi-presidential and
parliamentary systems. Table A2 provides the results for this additional model. The
inclusion of this additional interaction does also not improve how well the model ﬁts
the data. Most importantly the interaction between voters’ evaluation of the prime minister
and CEE remains signiﬁcant.
I further substantiate these ﬁndings and test my theoretical assumptions with two
additional models that analyse a moderation of leader evaluations by measures of party
Figure 1. Estimated median eﬀects on vote choice for the prime minister’s party, 95% conﬁdence
intervals.
8 J. BERZ
system stabilisation and media freedom. These models go beyond regional diﬀerences
and explain variation in the electoral inﬂuence of prime ministers between individual
countries. In order to measure the degree of media freedom in a country I use the
freedom of the press index (FOTP) by Freedom House (2002–2017). This index has been
shown to perform well and track other measures of media freedom (Bairett 2015) and
covers all necessary time points and countries. In political system with lower freedom of
the press, prime ministers and other leaders should have more opportunities to control
public perceptions and narratives of them and in turn exert more inﬂuence over voters’
decisions. I rescale the original Freedom House measure from 0–100 to 0–1 and also
reverse its direction so that 1 indicates maximum freedom of the press while 0 indicates
no freedom. Country scores of media freedom are calculated by taking the Freedom House
score that covers the election years in the data and averaging them for each country in
which more than one election is observed. The studied CEE countries have on average
a 0.14 lower media freedom score, see Table A5. However, within region variance exists
and some WE countries rank lower on media freedom than the CEE average.
As a measure of party system stabilisation, I employ the indicator of party system
closure calculated by Bértoa and Enyedi (2016), which measures the degree of stability
in government formation by observing access and alteration in government, as well as
whether new or familiar government coalitions emerge. Following the logic of Mair
(1996, p. 102–105) parties can more easily claim the loyalty of voters when patterns of
inter-party competition and government formation are established, therefore, alignments
of voters should be more frequent and stable. In contrast, more open party systems should
give greater opportunities for prime ministers and other leaders to re-deﬁne party brands
Figure 2. Interaction eﬀect of prime minister evaluation and Central Eastern Europe on vote choice for
the prime minister’s party, 95% conﬁdence intervals.
EAST EUROPEAN POLITICS 9
and previous modes of party cooperation and competition. Table A6 shows that party
systems in CEE countries are considerably more open than in WE, although, as with
media freedom, variance within the regions exists.
To test whether the inﬂuence of prime ministers and other leaders on vote choice is
moderated by these substantial measures of regional diﬀerence I re-estimate the previous
main model, replacing the dichotomous CEE indicator with the continuous measures of
party system stability and press freedom. Since the resulting cross-level interaction in
these additional models interacts two continuous variables on a small number of countries
they are estimated in separate models and do not include random slopes, which would
substantively increase complexity and in turn standard errors. Like all other non-dichotomous
variables the measure of press freedom and party system closure have been mean
centred and standardised by two standard deviations and are centred on the grand mean.
The full regression results of these two models are provided in Table A3.
The moderating eﬀect of press freedom on leader eﬀects is provided in Figure 3. As
visible in the interaction term “Eval. prime minister x Press freedom” if countries score
one standard deviation higher on press freedom prime ministers have a 0.5 point lower
inﬂuence on vote choice for their party. Reversely, prime minister exert a stronger
inﬂuence in countries with lower media freedom. The inﬂuence of other leaders is likewise
moderated by press freedom. In countries with higher scores of press freedom the eﬀect of
other leaders on vote choice for their party is reduced, while their inﬂuence increases with
decrease in press freedom. I also ﬁnd that in countries with higher levels of press freedom
voters tend to be less likely to vote for the ruling party of the prime minister, although this
eﬀect is not signiﬁcant on the 5% level.
Figure 4 shows the results for a moderation by party system closure. The results follow the
theorised pattern: the more closed and therefore stabilised party systems are the lower is the
inﬂuence of prime ministers on vote choice (Eval. prime minister x System closure), as well as
the inﬂuence of other leaders. Unlike the degree of press freedom party system closure
shows no tendency to inﬂuence vote choice on its own and only works as moderator.
Figure 3. Press freedom model. Estimated median eﬀects on vote choice for the prime minister’s party,
95% conﬁdence intervals.
10 J. BERZ
I have shown that the leader eﬀects of prime ministers systematically vary between CEE
and WE. Prime ministers in CEE have a larger leader eﬀect in parliamentary elections than
prime ministers in WE, leading to a more presidential relation between prime ministers
and the electorate in CEE than between prime ministers and the electorate in WE. A
further analysis shows that drivers of this diﬀerence are likely to be caused by diﬀerences
in party systems and media freedom between CEE and WE. Prime ministers enjoy signiﬁcantly
larger leader eﬀects in countries with lower freedom of the press and less stabilised
party systems.
In the following section, I test whether my ﬁnding on the systematic diﬀerence in leader
eﬀects between prime ministers in CEE and WE remains robust to additional tests.
6. Robustness checks
The tested cross-level interaction only entails a small number of cases on the country
level, therefore, I test the robustness of my main model by excluding single countries
from my analysis when estimating the cross-level interaction between voters’ evaluation
of the prime minister and CEE to ensure that the ﬁndings are not driven by voting behaviour
in a single country. Figure 5 shows that the ﬁndings are robust and not driven by a
single country under analysis. Excluding the Czech Republic or Hungary from the analysis
leads to the strongest reduction in p-values,6
but the cross-level interaction remains
signiﬁcant.
I also test whether my ﬁnding that leader eﬀects of opposing leaders are stronger in CEE
than in West Europe is equally robust (see Figure A1). I ﬁnd that the interaction is somewhat
less robust and turns insigniﬁcant when Romania or the Czech Republic are excluded.
In his article, Baylis (2007) notes a series of prime ministers in CEE who may be considered
exceptions to the prime ministerial weakness rule in the region. Since the CSES
data does not cover every election between 1996 and 2016 it may be the case that
Figure 4. Party system closure model. Estimated median eﬀects on vote choice for the prime minister’s
party, 95% conﬁdence intervals.
EAST EUROPEAN POLITICS 11
they include an over proportional amount of prime ministers who are considered to be
outliers. In this case, these exceptions as identiﬁed by Baylis may drive the interaction.
However, the selection of CEE prime ministers in the used CSES data only encompasses
one of the exceptions identiﬁed by Baylis: the parliamentary election in 1996 in which
Václav Klaus ran as prime minister. As the robustness test in Figure 5 shows the crosslevel
interaction holds even when the whole country is excluded, therefore, it is unlikely
that the ﬁnding is driven by a biased selection of prime ministers either.
Overall the ﬁnding that prime ministers in CEE have a stronger personal electoral performance
than prime ministers in WE remains robust. In the following section, I discuss
how future studies on prime minister’s performance and the presidentialisation of
voting behaviour in parliamentary elections may depart from the presented ﬁndings.
7. Conclusion
How strong are prime ministers in Central Eastern Europe compared to their counterparts
in Western Europe? I have addressed this issue by comparing their personal inﬂuence on
citizens’ voting behaviour in parliamentary elections. Comparing European prime ministers
over this presidential characteristic, I ﬁnd that prime ministers in CEE are electorally
Figure 5. Leave-one-out validation of the main model. The estimated median eﬀect of the interaction
between prime minister evaluation and Central Eastern Europe on vote choice is displayed by excluded
country, 95% conﬁdence intervals.
12 J. BERZ
stronger than prime ministers in WE. This pattern can be explained by the increased
inﬂuence of prime ministers and other leaders in countries with less stable party
systems and lower press freedom.
These ﬁndings suggest that judgements about diﬀerences in overall prime ministerial
strength between CEE and WE should be based on multiple dimension. Prime ministers in
CEE may on average have less time to achieve policies, especially in previous decades, but
they hold a more presidential position in parliamentary systems which should strengthen
their position. Rather than focusing on oﬃce duration as sole indicator of strength, this
study emphasises that prime ministers have to perform on other dimensions as well
and that several dimensions will beneﬁt their policy goals. Such a view does not
exclude the consideration of oﬃce duration. The time prime ministers spend in oﬃce
may still limit the amount of policies they can achieve and could be understood as a
necessary condition to perform. Other factors, like the here presented personal impact
in parliamentary elections, support networks within the party, as well as previous executive
or legislative experience may then determine how much prime ministers can
achieve during their time in oﬃce. Moreover, in times of presidentialised parliamentary
elections a prime minister’s direct personal impact on voters is an increasingly important
performance dimension – independently of policy achievements. The changes caused by
presidentialisation could also provide incentives to prime minister to further reduce media
independence of public broadcasters to control their public image during elections – a
trend that is already visible in some countries (Rupnik 2016).
In addition, presented evidence contributes to the study of electoral behaviour and
supports two assumptions of the electoral presidentialisation literature. First, the increased
importance of leaders in less established party systems in which voters alignment is lower.
Second, the role of media to give greater weight to leaders in parliamentary elections.
Both factors provide leaders in CEE with more favourable conditions to personally aﬀect
voting behaviour. Especially prime ministers may beneﬁt from a media system that is
more susceptible to the inﬂuence of elites and which they can actively pursue to
change in their favour.
The presented ﬁndings are in line with theoretical arguments of the electoral face of
presidentialisation and also present a puzzle. Even though prime ministers in CEE are electorally
stronger, their survival in oﬃce is in most cases shorter than the survival of prime
ministers in WE. According to the presidentialisation thesis electorally strong prime ministers
should be able to use their prominent position in parliamentary elections to gain
autonomy within their party and their cabinet. This should plausibly make their survival
in oﬃce more probable. What explanations may solve this apparent puzzle?
It may be the case that other factors enable or hinder prime ministers to use their
electoral strength to gain autonomy within their party and cabinet. These factors are
likely to be speciﬁc to individual prime ministers and not covered by the present analysis
that focuses on diﬀerences at the regional- and country-level and does not test how
eﬀectively individual prime ministers with a presidential role can gain autonomy. In
more institutionalised parties, prime ministers could struggle to gain a strong position
within their party, even if they have a presidential connection to the electorate. Cabinets
composed of coalitions may also moderate the opportunity of prime ministers
to gain autonomy within their cabinet. Lastly, personal characteristics of prime ministers
could further inﬂuence to what extent they are able to use their presidential role in
EAST EUROPEAN POLITICS 13
elections to their beneﬁt. Future studies may choose to investigate such mechanisms
more closely on the individual level and reﬁne our understanding of how prime ministers
use their changing role in parliamentary elections to gain leverage within party and
cabinet to pursue policies and remain in oﬃce. Studying individual prime ministers indepth
under a comparative research design may provide crucial insights.
This paper has set out to extend our knowledge of systematic diﬀerences in prime minister’s
strength across Europe. Central Eastern European prime ministers may on average
serve shorter terms in oﬃce, but they enjoy a more presidential inﬂuence in parliamentary
elections than prime ministers in Western Europe. The presented evidence highlights that
a prime minister’s strength is best perceived as a multi-faceted concept in which prime
ministers in CEE and WE may hold diﬀering advantages.
Notes
1. Throughout the paper, I follow (Baylis 2007) and use the terms performance and strength
interchangeably.
2. Table A4 in the Appendix shows that the mean popularity of prime ministers in this study
diﬀers only slightly between CEE and WE in favour of prime ministers in WE.
3. A few elections are excluded since the CSES covers a leader other than the prime minister (e.g.
the president in France).
4. The following software is employed for statistical modelling: Bates et al. (2014) and Knowles
and Frederick (2016).
5. The formula of both models is provided in the Appendix.
6. The p-values are 0.036 under the exclusion of the Czech Republic and 0.04 if Hungary is
excluded.
Acknowledgements
I would like to thank Ferdinand Müller-Rommel for critical feedback and encouraging remarks to this
paper, as well as the anonymous reviewers for their time and suggestions. I am also indebted to Mary
Stegmaier for supportive comments on this paper and to Corinna Kröber, Michael Jankowski, Markus
Tepe, as well as other participants of the Nordeutsches Kolloqium für Sozialwissenschaften for their
insightful comments on an early version of this project.
Disclosure statement
No potential conﬂict of interest was reported by the author.
Funding
This research did not receive any speciﬁc grant from funding agencies in the public, commercial or
not-for-proﬁt sectors.
Notes on contributors
Jan Berz is a research fellow and doctoral candidate at Leuphana University Lüneburg,
where he is currently completing his doctoral thesis on prime ministers and accountability
in democratic elections. He previously studied at the University of Mannheim, the European
University Institute in Florence and the Jagiellonian University in Krakow.
14 J. BERZ
References
Bairett, Richard L. 2015. “Executive Power and Media Freedom in Central and Eastern Europe.”
Comparative Political Studies 48 (10): 1260–1292.
Bajomi-Lázár, Péter. 2013. “The Party Colonisation of the Media: The Case of Hungary.” East European
Politics and Societies 27 (1): 69–89. doi:10.1177/0888325412465085.
Bajomi-Lázár, Péter. 2017. “Between Neutrality and Engagement: Political Journalism in Hungary.”
Central European Journal of Communication 10 (18): 48–63.
Bates, Douglas, Martin Mächler, Ben Bolker, and Steve Walker. 2014. “Fitting Linear Mixed-Eﬀects
Models Using LME4.” arXiv preprint arXiv:1406.5823.
Baylis, Thomas A. 1996. “Presidents Versus Prime Ministers: Shaping Executive Authority in Eastern
Europe.” World Politics 48 (3): 297–323.
Baylis, Thomas A. 2007. “Embattled Executives: Prime Ministerial Weakness in East Central Europe.”
Communist and Post-Communist Studies 40 (1): 81–106.
Bennister, Mark, Paul’t Hart, and Ben Worthy. 2015. “Assessing the Authority of Political Oﬃceholders:
the Leadership Capital Index.” West European Politics 38 (3): 417–440.
Bértoa, Fernando Casal, and Zsolt Enyedi. 2016. “Party System Closure and Openness:
Conceptualization, Operationalization and Validation.” Party Politics 22 (3): 265–277. doi:10.
1177/1354068814549340.
Bértoa, Fernando Casal, and Peter Mair. 2012. “Party System Institutionalisation Across Time in PostCommunist
Europe.” In Party Government in the New Europe, edited by F. Müller-Rommel and
H. Keman, 104–131. London: Routledge.
Bittner, Amanda. 2011. Platform Or Personality?: The Role of Party Leaders in Elections. Oxford: Oxford
University Press.
Bittner, Amanda. 2018. “The Personalization of Parliamentary Elections?” In The Personalization of
Democratic Politics and the Challenges for Political Parties, 39–55. London: ECPR Press.
Blondel, Jean, and Ferdinand Müller-Rommel, eds. 1993. Governing Together. The Extent and Limits of
Joint Decision-Making in Western European Cabinets. New York: St. Martin’s Press.
Blondel, Jean, Ferdinand Müller-Rommel, Darina Malová, Katja Fettelschoss, Alenka Krašovec, and
Matej Kovačič, eds. 2007. Governing New European Democracies. London: Palgrave Macmillan.
Blondel, Jean, and Jean-Louis Thiébault. 2009. Political Leadership, Parties and Citizens: The
Personalisation of Leadership. New York: Routledge.
Campbell, Angus, Philip E. Converse, Warren E. Miller, and E. Donald. 1966. The American Voter.
New York: Wiley.
Costa, Patrício, and Frederico Ferreira da Silva. 2015. “The Impact of Voter Evaluations of Leaders’
Traits on Voting Behaviour: Evidence From Seven European Countries.” West European Politics
38 (6): 1226–1250.
CSES. 2018. “The Comparative Study of Electoral Systems (www.cses.org). CSES Module 4 Full Release
[dataset and documentation].” doi:10.7804/cses.module4.2018-05-29.
Curtice, John, and Sören Holmberg. 2005. “Party Leaders and Party Choice: A Comparative Study of
Modern Democracies.” In The European voter, edited by Jaques Thomassen, 91–107. Oxford:
Oxford University Press.
Dobek-Ostrowska, Bogusława. 2015. 25 Years After Communism: Four Models of Media and Politics in
Central and Eastern Europe. Bern, Schweiz: Peter Lang. https://www.peterlang.com/view/
9783653981964/xhtml/chapter 001.xhtml.
Dragomir, Marius. 2018. “Control the Money, Control the Media: How Government Uses Funding
to Keep Media in Line.” Journalism 19 (8): 1131–1148. doi:10.1177/1464884917724621.
Elgie, Robert. 2012. “Political Leadership in Old and New Democracies.” In Comparative Political
Leadership, 272–291. London: Palgrave Macmillan.
Elgie, Robert. 2018. “Semi-Presidentialism, Premier-Presidentialism and President-Parliamentarism –
A New Country-Years Dataset [Blog Post, 3 April].” http://presidential-power.com/?p=7869.
Ferreira da Silva, Frederico. 2018. “Fostering Turnout?: Assessing Party Leaders’ Capacity to Mobilize
Voters.” Electoral Studies 56: 61–79.
EAST EUROPEAN POLITICS 15
Freedom House. 2002–2017. “Freedom of the Press (Annual Reports).” https://freedomhouse.org/
report-types.
Garzia, Diego. 2011. “The Personalization of Politics in Western Democracies: Causes and
Consequences on Leader – Follower Relationships.” The Leadership Quarterly 22 (4): 697–709.
Garzia, Diego. 2014. Personalization of Politics and Electoral Change. Basingstoke: Palgrave Macmillan.
Garzia, Diego. 2017a. “Personalization of Politics Between Television and the Internet: Leader Eﬀects
in the 2013 Italian Parliamentary Election.” Journal of Information Technology & Politics 14 (4):
403–416.
Garzia, Diego. 2017b. “Voter Evaluation of Candidates and Party Leaders.” In The SAGE Handbook of
Electoral Behavior, 633–653. London: SAGE Publications.
Giebler, Heiko, Josephine Lichteblau, Antonia May, Reinhold Melcher, Aiko Wagner, and Bernhard
Weßels. 2016. “CSES Module 1-3 Harmonized Trend File” (dataset). May 31, 2016 Version. doi:
10.7804/cses.trendﬁle.2016-05-31.
Grotz, Florian, and Ferdinand Müller-Rommel. 2015. “Schwache Regierungschefs? Politische
Erfahrung Und Amtsdauer Von Premierministern in Mittel-und Osteuropa.” ZParl Zeitschrift für
Parlamentsfragen 46 (2): 310–327.
Grotz, Florian, and Till Weber. 2017. Prime Ministerial Tenure in Central and Eastern Europe: The Role of
Party Leadership and Cabinet Experience, 229–248. Wiesbaden: Springer Fachmedien Wiesbaden.
Hanretty, Chris. 2010. “Explaining the De Facto Independence of Public Broadcasters.” British Journal
of Political Science 40 (1): 75–89.
Helms, Ludger. 2016. “The Politics of Leadership Capital in Compound Democracies: Inferences from
the German Case.” European Political Science Review 8 (2): 285–310.
Helms, Ludger. 2017. “When Less Is More: ‘Negative Resources’ and the Performance of Presidents
and Prime Ministers.” Politics 39 (3): 269–283.
Jakubowicz, Karol. 2001. “Rude Awakening Social and Media Change in Central and Eastern Europe.”
Javnost-The Public 8 (4): 59–80.
Kaarbo, Juliet, and Margaret G. Hermann. 1998. “Leadership Styles of Prime Ministers: How Individual
Diﬀerences Aﬀect the Foreign Policymaking Process.” The Leadership Quarterly 9 (3): 243–263.
King, Anthony. 2002a. “Do Leaders’ Personalities Really Matter?” In Leaders’ Personalities and the
Outcomes of Democratic Elections, edited by Anthony King, 1–43. Oxford: Oxford University Press.
King, Anthony. 2002b. Leaders’ Personalities and the Outcomes of Democratic Elections. Oxford: Oxford
University Press.
Knowles, Jared E., and Carl Frederick. 2016. “merTools: Tools for Analyzing Mixed Eﬀect Regression
Models.” R package version 0.3. 0.
Lenz, Gabriel S., and Chappell Lawson. 2011. “Looking the Part: Television Leads Less Informed
Citizens to Vote Based on Candidates’ Appearance.” American Journal of Political Science 55 (3):
574–589.
Lewis, Paul. 2002. Political Parties in Post-Communist Eastern Europe. London: Routledge.
Lobo, Marina Costa, and John Curtice. 2014. Personality Politics?: The Role of Leader Evaluations in
Democratic Elections. Oxford: Oxford University Press.
Mair, P. 1996. “Party Systems and Structures of Competition.” In Comparing Democracies: Elections
and Voting in Global Perspectives, edited by LeDuc Lawrence, Richard G. Niemi and Pippa
Noriss, 83–106. London: SAGE Publications.
Mair, Peter. 1997. Party System Change: Approaches and Interpretations. Oxford: Oxford University
Press.
Millard, Frances. 2004. Elections, Parties, and Representation in Post-Communist Europe. London:
Palgrave Macmillan.
Mughan, Anthony. 2000. Media and the Presidentialization of Parliamentary Elections. London:
Palgrave Macmillan.
Mughan, Anthony. 2015. “Parties, Conditionality and Leader Eﬀects in Parliamentary Elections.” Party
Politics 21 (1): 28–39.
Örnebring, Henrik. 2012. “Clientelism, Elites, and the Media in Central and Eastern Europe.” The
International Journal of Press/Politics 17 (4): 497–515.
16 J. BERZ
Poguntke, Thomas, and Paul Webb. 2005. The Presidentialization of Politics: A Comparative Study of
Modern Democracies. Oxford: Oxford University Press.
Rahat, Gideon, and Tamir Sheafer. 2007. “The Personalization(s) of Politics: Israel, 1949–2003.” Political
Communication 24 (1): 65–80.
Rupnik, Jacques. 2016. “Surging Illiberalism in the East.” Journal of Democracy 27 (4): 77–87.
Sartori, Giovanni. 1989. “Video-power.” Government and Opposition 24 (1): 39–53.
Sedelius, Thomas, and Olga Mashtaler. 2013. “Two Decades of Semi-presidentialism: Issues of
Intra-executive Conﬂict in Central and Eastern Europe 1991–2011.” East European Politics 29 (2):
109–134.
Splichal, Slavko. 1992. “Media Privatization and Democratization in Central-Eastern Europe.” Gazette
(Leiden, Netherlands) 49 (1–2): 3–22.
Strangio, Paul, Paul’t Hart, and James Walter. 2013. Understanding Prime-Ministerial Performance:
Comparative Perspectives. Oxford: Oxford University Press.
Tworzecki, Hubert. 2003. Learning to Choose: Electoral Politics in East-Central Europe. Stanford
University Press.
Wattenberg, Martin P. 1991. The Rise of Candidate-Centered Politics: Presidential Elections of the 1980s.
Cambridge: Harvard University Press.
Weßels, Bernhard, and Hans-Dieter Klingemann. 2006. “Parties and Voters–Representative
Consolidation in Central and Eastern Europe?” International Journal of Sociology 36 (2): 11–44.
Worthy, Ben. 2016. “Ending in Failure? The Performance of ‘Takeover’ Prime Ministers 1916–2016.”
The Political Quarterly 87 (4): 509–517.
EAST EUROPEAN POLITICS 17
Appendix
Model speciﬁcations:
– Main model
Pr(voteij =1)=logit−1
(p0jcountryij +p1jeval PMij +p2jevalOther Leaderij +p3PIDprimeminister Partyij+
p4PIDother Partyij +p5ageij +p6genderij +p7educationij)
p0j = g00 + g01CEEj + z0j
p1j = g10 + g11CEEj + z1j
p2j = g20 + g21CEEj + z2j
where z0j N(0, s2
0), z1j N(0, s2
1) and z2j N(0, s2
2)
as well as the covariances rs0s1, rs0s2, rs1s2
– Semi-presidential model
Pr(voteij =1)=logit−1
(p0jcountryij +p1jeval PMij +p2jevalOther Leaderij +p3PIDprimeminister Partyij+
p4PIDother Partyij +p5ageij +p6genderij +p7educationij)
p0j = g00 + g01CEEj + g02semi Presidentialj + z0j
p1j = g10 + g11CEEj + g12semi Presidentialj + z1j
p2j = g20 + g21CEEj + g22semi Presidentialj + z2j
where z0j N(0, s2
0), z1j N(0, s2
1) and z2j N(0, s2
2)
as well as the covariances rs0s1, rs0s2, rs1s2
Figure A1. Leave-one-out validation of the main model. Estimated median eﬀect of the interaction
between other party leaders and Central Eastern Europe on vote choice are displayed by excluded
country, 95% conﬁdence intervals.
18 J. BERZ
Table A1. List of countries, elections and prime ministers under
study.
Country Election Prime minister
Austria 2013 Faymann
Croatia 2007 Sanander
Czech Republic 1996 Klaus
2006 Paroubek
Denmark 1998 N. Rasmussen
2007 F. Rasmussen
Estonia 2011 Ansip
Finland 2007 Vanhanen
2011 Kiviniemi
2015 Stubb
Germany 1998 Kohl
2005 Schröder
2009 Merkel
2013 Merkel
Greece 2009 Karamanlis
2015 Samaras
Hungary 1998 Horn
Ireland 2007 Ahern
Iceland 1999 Oddsson
2007 Haarde
2013 Sigurdardottir
Latvia 2010 Dombrovskis
2011 Dombrovskis
2014 Straujuma
Netherlands 1998 Kok
2006 Balkenende
2010 Balkenende
Norway 1997 Jagland
2009 Stoltenberg
2013 Stoltenberg
Poland 1997 Cimoszewicz
2007 Kaczyński
2011 Tusk
Portugal 2002 Guterres
2009 Sócrates
2015 Coelho
Romania 2012 Ponta
Slovakia 2010 Fico
2016 Fico
Slovenia 1996 Drnovsek
2008 Rop
2011 Pahor
Spain 1996 González
2000 Aznar
2008 Zapatero
Sweden 1998 Persson
2014 Reinfeldt
United Kingdom 1997 John Major
2015 Cameron
EAST EUROPEAN POLITICS 19
Table A2. Regression results, generalised linear model with logit link.
Main model Semi pres. model
Fixed eﬀects
(Intercept) −1.30*** −1.52***
(0.27) (0.30)
Eval. prime minister 2.76*** 2.71***
(0.23) (0.27)
Eval. other leaders −1.52*** −1.49***
(0.13) (0.15)
PID PM party 2.64*** 2.64***
(0.05) (0.05)
PID other party −1.52*** −1.52***
(0.04) (0.04)
Age 0.05 0.05
(0.03) (0.03)
Female 0.09** 0.09**
(0.03) (0.03)
Education level −0.06 −0.06
(0.03) (0.03)
CEE 0.27 0.17
(0.42) (0.41)
CEE × Eval. prime minister 0.96** 0.94*
(0.37) (0.37)
Eval. other leaders × CEE −0.48* −0.47*
(0.21) (0.22)
Semi pres. 0.58
(0.41)
Eval. prime minister × Semi pres. 0.12
(0.37)
Eval. other leaders × Semi pres. −0.08
(0.21)
Random eﬀects
Var: Country (Intercept) 0.95 0.87
Var: Country Eval. prime minister 0.63 0.63
Var: Country Eval. other leaders 0.20 0.19
AIC 29650.31 29654.07
BIC 29800.42 29830.68
Log Likelihood −14808.15 −14807.04
Num. obs. 50532 50532
Num. countries 22 22
Note: Standard errors in parentheses, ∗∗∗
p , 0.001, ∗∗
p , 0.01, ∗
p , 0.05.
Table A3. Regression results, generalised linear model with logit link.
Press freedom model Party system model
Fixed eﬀects
(Intercept) −1.27*** −1.22***
(0.20) (0.21)
Eval. prime minister 2.89*** 2.91***
(0.04) (0.04)
Eval. other leaders −1.53*** −1.53***
(0.03) (0.03)
PID PM party 2.64*** 2.64***
(0.05) (0.05)
PID other party −1.52*** −1.52***
(0.04) (0.04)
Age 0.04 0.05
(0.03) (0.03)
Female 0.09** 0.09**
(0.03) (0.03)
(Continued)
20 J. BERZ
Table A3. Continued.
Press freedom model Party system model
Education level −0.05 −0.05
(0.03) (0.03)
Press freedom −0.63
(0.34)
Eval. prime minister × Press freedom −0.52***
(0.09)
Eval. other leaders × Press freedom 0.44***
(0.07)
System closure −0.30
(0.35)
Eval. prime minister × System closure −0.75***
(0.09)
Eval. other leaders × System closure 0.35***
(0.07)
Random eﬀects
Var: Country (Intercept) 0.77 0.87
AIC 29905.76 29884.53
BIC 30011.73 29990.50
Log Likelihood −14940.88 −14930.27
Num. obs. 50532 50532
Num. countries 22 22
Note: Standard errors in parantheses, ∗∗∗
p , 0.001, ∗∗
p , 0.01, ∗
p , 0.05.
Table A4. Mean evaluation of prime ministers in the regions (0, dislike – 10, like).
Region Mean Standard deviation Min. Max.
Central Eastern Europe 5.04 0.78 3.26 6.44
Western Europe 5.61 0.89 2.32 7.25
Table A6. Mean party system closure score in the regions (0, open – 100, closed).
Region Mean Standard deviation Min Max
Central Eastern Europe 38.22 9.63 21 53
Western Europe 94.62 2.71 88.5 98.4
Table A5. Mean score of media freedom in the regions (0, not free – 1, free).
Region Mean Standard deviation Min. Max.
Central Eastern Europe 0.71 0.07 0.58 0.82
Western Europe 0.85 0.08 0.62 0.92
EAST EUROPEAN POLITICS 21