ON THE DYNAMICS OF RADICAL-RIGHT PARTIES IN EUROPE

FROM RISE TO RULE:

By

Nicolas Dimitri Tewﬁk Bichay

A DISSERTATION

Submitted to

Michigan State University

in partial fulﬁllment of the requirements

for the degree of

Political Science—Doctor of Philosophy

2021

ABSTRACT

FROM RISE TO RULE:

ON THE DYNAMICS OF RADICAL-RIGHT PARTIES IN EUROPE

By

Nicolas Dimitri Tewﬁk Bichay

Why have we seen such a dramatic rise of radical-right parties recently, and what have

the consequences been? This dissertation traces the three main stages of the rising radical-

right: ﬁrst, how have they been succeeding in elections? Second, after winning election,

why is their participation in government coalitions becoming more common? And third,

what have the substantive eﬀects of this government participation been on the quality and

stability of liberal democracy?

I begin by examining the dynamics of the radical-right’s electoral success in relation

to the rising trend of publicly funding electoral campaigns at the national level. While public

ﬁnancing is often thought of as a way to “level the playing ﬁeld” of elections, I argue that

such policies have had unintended second-order consequences by disproportionately aiding

the rise radical-right parties. I further demonstrate that this public funding policy severs

the tie of radical-right performance to economic growth. This is a counter-intuitive result,

as public campaign ﬁnance is often considered a priority of the left to restore power to the

majority, and yet, it seems to aid those on the right fringes of the political spectrum.

Given this increased electoral success, chapter two investigates why these parties have

found themselves in more and more governing coalitions recently. Until now, the literature

was without a succinct explanation for why this is occurring—instead characterizing the

trend as “politics as usual.” I develop a new theory characterizing such moves as the co-

optation of a growing political rival in an eﬀort to minimize electoral threat. That is, as the

radical platform rises in salience, and radical parties are more threatening to the electoral

success of a moderate party, they will invite the party into their government, thereby staving

oﬀ said threat.

Finally, I look at the impacts of said coalitions on liberal democracy. While many

argue radical-right parties to be one of the largest contemporary threats to democracy, the

evidence remains largely anecdotal.

I empirically demonstrate that the implications of a

growing radical-right are far more severe than previously assumed. Besides being the source

of a potential for democratic breakdown when they take power, radical-right parties bring

with them real decreases in the level of liberal democracy not only when holding executive

control, but also as they merely participate in government coalitions. The speciﬁc eﬀects

diﬀer depending on the type of power held. When holding the prime ministership, we see

signiﬁcant deteriorations to both institutional constraints on the executive and mass civil

liberties. When limited to junior coalition-partner status, we only see the latter aﬀected.

Overall, this dissertation aims to trace the three stages of the rising radical-right

phenomenon. It begins with a study on how they win elections, moves to an analysis of their

entrance into ruling coalitions, and ends with an investigation of the eﬀects of this rise on

liberal democracy.

Copyright by
NICOLAS DIMITRI TEWFIK BICHAY
2021

To my Nona, for always inspiring me.

v

ACKNOWLEDGEMENTS

I have had the great opportunity to draft this dissertation while traveling to several

beautiful places (which, to any comparativist, is simply called ‘working’). I have especially

fond memories of crafting the ﬁrst (very) rough draft of chapter two in several Danish

coﬀee shops throughout Aarhus, and dramatically redrafting chapter one at a beach bar in

Agia, Greece (chapter three, rather unimpressively, was mostly drafted in my oﬃce in East

Lansing). But, it is not all fun and games (it is probably less than 5% fun and games).

Unique to my fellow class of graduates this year, and quite stress-inducing at the time, much

of the work at the end was conducted in the quarantine lockdown in response to the COVID-

19 global pandemic. Thus, I am especially grateful to everyone for oﬀering such help and

support during such stressful and uncertain times. I could not have ever hoped to cross the

ﬁnish line had it not been for all these wonderful people.

First and foremost, my most sincere thanks goes to my parents. Throughout my

whole life they have oﬀered nothing but love and support. Thank you for your lifetime of

guidance, encouragement, and reassurance of my ability to do things I was certain I couldn’t.

They oﬀered counsel when I felt lost, inspiration when I felt dispirited, and perhaps most

importantly, box-loads of food when I needed them most. Without you both, I am certain

I would never be here. Second, to my wonderful siblings, Tina, Alexi, and Keith. You can’t

pick your family—but I am certain you three would be my choice if you could. Our late

nights of laughter whilst listening to the same Paul Simon record 15 times in-a-row gave me

the energy needed to complete this. I thank the three of you for your love, laughter, and

words of encouragement. I have, above all else, been blessed with a large, loud—perhaps

obnoxious—but always loving, supportive, and hilarious Greek-Egyptian family and I cannot

thank everyone enough for the role they played.

I was lucky enough to work in a department full of friendly and collaborative col-

leagues. I thank Caleb Lucas, Jonathan King, Nate Smith, Shane Wery, and Crystal Williams

vi

for their friendship and encouragement. I want to give an extra special thank you to Stephen

Anderson and Marcella McCalmont for becoming two of my best friends. Our late night

study sessions kept me prepared, while our frequent trips to The Peanut Barrel and Los

Tres Amigos kept me sane. I give Steve special recognition for putting up with me as an

oﬃce-mate for six years and pretending that I don’t talk to myself “that much.” Without

him, this dissertation would be riddled with typos and nonsensical sentence structures.

I truly feel as though I hit the academic jackpot with my dissertation committee. Eric

Chang, Erica Frantz, and Cory Smidt have been extraordinarily generous with both their

time and wisdom, and I thank them for their always helpful input that has no doubt made

me a better scholar. I have to give a very very large thank you to my wonderful advisor,

Christian Houle. He guided me through this process ensuring I never felt lost, overwhelmed,

or like I could not make it through. Thank you for being always willing to share your advice

and experience, for reading roughly 30 drafts of each of my papers, and for constantly being

one of the nicest, friendliest people I have met. You, along with my entire committee, truly

made my graduate experience much more enjoyable than it had any right to be.

Fortunately for me Michigan State is very much not the department where guidance

stops with your committee, and as such I want to give a sincere thanks to the rest of the

faculty, especially Shahryar Minhas, for always being willing to help and oﬀer comments on

ongoing research. I also want to thank the wonderful department of political science at Grand

Valley, and especially my undergraduate mentor John Constantelos, for ﬁrst interesting me

in academia, and for guiding me through the graduate admissions process.

Last, but most certainly not least, a very loving thank you to my wonderful partner

Elizabeth. You pulled double duty over the last couple years oﬀering endless support both

emotionally and academically. You are always the ﬁrst to read a new draft of my work, and

the ﬁrst to cheer me up when I need it most. I appreciate your frequent insistences that,

perhaps I should practice my talk, when I was convinced “winging it” would be just ﬁne.

Thank you for being my biggest supporter.

vii

TABLE OF CONTENTS

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

x

LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

CHAPTER 1

PUBLIC CAMPAIGN FINANCING AND THE RISE OF
RADICAL-RIGHT PARTIES . . . . . . . . . . . . . . . . . . . . . .
1.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2 The Rise of Radical Parties Over Time . . . . . . . . . . . . . . . . . . . . .
1.3 Theoretical Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.1 Why the Right? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4 Data and Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.1 Measurement
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.2 Model
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.3 Time-trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.5.1 Robustness Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.6 Motivating Case: Austria . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

CHAPTER 2 COALITION AS COOPTATION: THE RADICAL-RIGHT’S

2.3.1 Coalition Invitation and Acceptance

2.2.1 The Costs of Radical Parties in Coalition Government

ENTRANCE TO GOVERNMENT . . . . . . . . . . . . . . . . . . .
2.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 Coalition Formation in Europe . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . .
2.3 Radical Cooptation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . .
2.4 Measuring Platform Salience . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5 Data and Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6.1 Additional Direct Tests . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7 Comparing Out-of-Sample Prediction of Diﬀerent Theories . . . . . . . . . .
2.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

CHAPTER 3 RADICAL-RIGHT PARTIES AND THE DETERIORATION OF

3.1
3.2 The Inﬂuence of Radical-Right Parties

LIBERAL DEMOCRACY . . . . . . . . . . . . . . . . . . . . . . . .
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . .
3.2.1 The Radical-Right and Liberal Democracy . . . . . . . . . . . . . . .
3.3 Radical-Right Impact on Liberal Democracy . . . . . . . . . . . . . . . . . .
3.4 Data and Modeling Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . .

viii

7
7
10
14
18
20
20
23
26
27
33
34
36

39
39
41
42
44
46
48
51
55
59
62
65

68
68
71
72
75
81

3.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5.1 Eﬀects of Radical-Right Coalitions
. . . . . . . . . . . . . . . . . . .
3.5.2 Robustness Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

85
89
92
95

CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

99

APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
APPENDIX TO CHAPTER 1 . . . . . . . . . . . . . . . . . . 104
APPENDIX TO CHAPTER 3 . . . . . . . . . . . . . . . . . . 112

APPENDIX A
APPENDIX B

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

ix

LIST OF TABLES

Table 1.1: Ordinal Coding of Finance Laws . . . . . . . . . . . . . . . . . . . . .

Table 1.2: Distribution of Finance Laws . . . . . . . . . . . . . . . . . . . . . . .

Table 1.3: Results of Tobit and Beta Regressions . . . . . . . . . . . . . . . . . .

Table 2.1: Coalitions in Europe: 1989-2018 . . . . . . . . . . . . . . . . . . . . .

Table 2.2: Results from Multi-level Probit Regressions . . . . . . . . . . . . . . .

Table 2.3: Results of Direct Tests . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 2.4: Summary of Out-of-Sample Prediction . . . . . . . . . . . . . . . . . .

Table 3.1: Eﬀect of a Radical-Right Prime Minister on Institutional Rule-of-Law

Table 3.2: Predicted Values of Institutional Rule-of-Law . . . . . . . . . . . . . .

Table 3.3: Eﬀect of a Radical-Right Prime Minister on Civil Liberties

. . . . . .

Table 3.4: Predicted Values of Civil Liberties . . . . . . . . . . . . . . . . . . . .

Table 3.5: Eﬀect of a Radical-Right Coalition on Institutional Rule-of-Law . . . .

Table 3.6: Eﬀect of a Radical-Right Coalition on Civil Liberties . . . . . . . . . .

Table 3.7: Predicted Values of Civil Liberties . . . . . . . . . . . . . . . . . . . .

Table 3.8:

Institutional Rule-of-Law, Radical-Right and Almost Radical-Right
Regimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 3.9: Civil Liberties, Radical-Right and Almost Radical-Right Regimes . . .

Table 3.10: Civil Liberties, Radical-Right and Almost Radical-Right Coalitions . .

22

24

28

53

56

61

63

86

86

88

88

90

91

91

95

95

95

Table A.1: Results on Far-Left vote-Share . . . . . . . . . . . . . . . . . . . . . . 105

Table A.2: Results of Alternative Speciﬁcations . . . . . . . . . . . . . . . . . . . 106

Table A.3: Results of Fixed-Eﬀects Tobit on Imputed Dataset . . . . . . . . . . . 108

Table A.4: Results of Fixed-Eﬀects Tobit for Outlier Sensitivity . . . . . . . . . . 109

Table B.1: Models With Alternative DVs

. . . . . . . . . . . . . . . . . . . . . . 113

Table B.2: Institutional Rule-of-Law, Hungary Omitted . . . . . . . . . . . . . . 116

Table B.3: Institutional Rule-of-Law, Poland Omitted . . . . . . . . . . . . . . . 117

x

Table B.4: Civil Liberties, Hungary Omitted . . . . . . . . . . . . . . . . . . . . . 118

Table B.5: Civil Liberties, Poland Omitted . . . . . . . . . . . . . . . . . . . . . . 119

xi

LIST OF FIGURES

Figure 1.1: Extremism Over Time

. . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 1.2: Marginal Eﬀect of Public Financing . . . . . . . . . . . . . . . . . . .

Figure 1.3: Conditional Eﬀect of Economic Growth . . . . . . . . . . . . . . . . .

Figure 1.4: Eﬀect of Finance on Marginal Eﬀect of Growth . . . . . . . . . . . . .

Figure 1.5: Results in Austrian Elections . . . . . . . . . . . . . . . . . . . . . . .

Figure 2.1: Substantive Eﬀects on P (Radical-Right in Government) . . . . . . . .

Figure 2.2: Eﬀect of Joint Platform Salience With Very Low Seat-Share

. . . . .

Figure 2.3: Substantive Eﬀects on P (Radical-Right in Government) . . . . . . . .

27

30

31

32

35

57

58

62

Figure A.1: Tobit vs. Beta Regression: Direct Eﬀects . . . . . . . . . . . . . . . . 110

Figure A.2: Tobit vs. Beta Regression: Marginal Eﬀects

. . . . . . . . . . . . . . 111

Figure B.1: Eﬀect of Radical-Right Prime Minister Over Time . . . . . . . . . . . 114

Figure B.2: Jackknife Prime Minister Parameter Estimates, Rule-of-Law . . . . . 115

Figure B.3: Jackknife Prime Minister Parameter Estimates, Civil Liberties . . . . 115

xii

INTRODUCTION

Throughout time, we have seen a dramatic increase in radical and populist parties

globally. Those who once stood opposed to such parties have found themselves increasingly

willing to form new partnerships. This is happening quickly, and with special vigor across

Europe with regard to the radical-right speciﬁcally—for example: the Sweden Democrats in

Sweden, Danish People’s Party in Denmark, and the National Front in France are but a few

instances of those who have seen large gains in popularity. In the latest German election, the

Alternative for Germany gained 94 seats, enough to make them the largest opposition party.

However, as is becoming clear, their participation is not limited to government opposition.

Currently, radical-right parties serve in governing coalitions in seven European nations, and

hold the prime minister’s oﬃce in two.

The rise of these parties should be somewhat unsettling.

In the aftermath of the

European debt crisis, the Greek neo-Nazi party Golden Dawn saw a surge in popularity; last

year its leaders were found guilty by a criminal court of running a criminal organization,

while several party deputies were jailed for murder and attempted murder. Meanwhile in

Eastern Europe the tenure of ruling parties like Fidesz and Law and Justice have brought

with them major declines in the level of democracy in Hungary and Poland, respectively.

Clearly, there is cause for concern. While some argue this trend has reached a peak, and is

now on the decline, others warn the radical-right is here to stay.

As this trend manifests throughout time and space, the causes remain unclear. Why

have we seen such an accelerated rise from what was long considered a “niche” and “fringe”

party-family? How is it that other parties deem them acceptable partners and agree to rule

in government alongside them? And, perhaps most importantly, what real eﬀects has this

worrisome trend brought with it in terms of democratic stability and the civil liberties of its

1

citizenry? Throughout this dissertation I shed some much-needed light on the dynamics of

radical-right parties, demonstrate how they come to power, and importantly, illustrate the

long-lasting eﬀects of their rule on society.

Overall, I present three related pieces of research, one from each of the three phases of

the rise of radical parties: (1) how they win elections, (2) how they enter into government, and

ﬁnally (3) how their victory has far-reaching implications in the countries they rule. Through

the next three chapters, I intend to analyze every stage of the radical-right phenomenon,

from how they rise in popularity, transition to governing, and ﬁnally impact society. Scholars

agree that the rise of these parties is the largest contemporary threat to liberal democracy.

As such, given their current upward trend, this topic is especially important and topical.

I begin by analyzing the growing electoral success of these parties. Beginning with

Jackman and Volpert (1996), the most basic rule radical-right studies has is this: they

tend to excel in times of economic distress. The reasoning for this empirical relationship is

somewhat straightforward: these parties thrive in protest (Powell Jr 1981). They build a

name and a following for themselves by continuously decrying the prevailing conditions in

their country. Clearly, such bemoaning is more likely to take hold when times are tough. It

is rather hard to advocate such large-scale change to individuals who are doing quite well.

Conversely, when the economy begins to stagnate, people want change. They look for a

promise of something better, and importantly for these parties—they look for someone to

blame. Enter the radical-right party, who is eager to make unrealistic promises, and more

eager still to align blame squarely on the ruling elites and ethnic minorities (Rydgren 2018).

In this way, one can characterize a poor economy as “pushing” voters towards the radical-

right. This is incredibly important for radical-right parties, as they traditionally do quite

poorly in terms of private donations, and thus struggle vis-`a-vis mainstream parties with

campaigning. Times of economic downturn, in this way, act almost as free advertising for

these parties, providing an increased salience and audience for their platform they need not

pay for.

2

What happens, I ask, if instead of waiting for a bad economy to “push” voters towards

them, governments were unknowingly moving towards systems that more easily allowed

radical-right parties to “pull” voters in their direction? Such is the basis of chapter one. I

examine the rising trend of publicly funding electoral campaigns, and argue that this public

campaign ﬁnancing has had the unintended consequence of dramatically aiding radical-right

parties. Further, I argue this aforementioned well-regarded ﬁnding tying radical performance

to the economy only holds in cases of privately-funded elections. This comes about largely

due to the fact that radical parties are severely disadvantaged in terms of private donations,

as well as the propensity of public funds to erode incumbency advantage and dilute the voice

of special interests. Interestingly, many advocates of publicly funding elections tend to be

those on the left side of the political spectrum, who argue for these public ﬁnancing provisions

as a way to reallocate political power to the majority. Here, however, I ﬁnd such provisions

to be provide a poor return-on-investment for these advocates, in that such provisions have

instead reallocated power to the political fringe of the ideological spectrum.

The recent trends in history, however, have not ended with the radical-right’s in-

creased election to national legislature. Rather, they have also found themselves more and

more in ruling coalitions, generally serving alongside another party to govern the nation

and run cabinet ministries. This too is no longer rare, as radical-right parties have joined

government coalitions in nearly half of European nations. This is truly striking, for it is here

that signiﬁcant change may be made. Here, one is able to have an inﬂuence on the direction

and agenda of the government, and further, it is here where leadership over government

ministries is executed.

And yet, to date all research regarding the radical-right in government has indicated

this trend is “politics as usual,” and not altogether too surprising. I argue this is simply not

the case. The theories of minimum winning coalitions (Riker 1962) and minimum connected

winning coalitions (Axelrod 1970; de Swaan 1973), while incredibly informative, are simply

not capable of explaining what we have seen transpire over the last several decades on their

3

own. We have witnessed parties who win an election and have options for more moderate,

more ideologically close, less “costly” political partners still opt to form a coalition with a

radical-right party despite most evidence on the topic telling us such a scenario would not

happen.

In chapter two I formulate a new theory of coalition formation with radical-right

parties. I borrow from the literature on authoritarian cooptation and develop a characteri-

zation of the coalition-building process that diverges from preexisting theories. Rather than

a simple numbers game, I argue such processes represent something much more basic: fear.

Such a move to join in partnership with a radical-right party can be best described as the

cooptation of a growing political rival in an eﬀort to minimize future electoral threat. That

is, as the radical-right platform rises in salience, and radical parties are more threatening to

the electoral success of a mainstream party, they will invite the party into their government,

in an eﬀort to stave oﬀ said threat. It does not matter how you measure this fear, whether

in terms of prevailing economic conditions, immigration inﬂows, voter opinion on policy, or

even simply based on how well similar radical parties are doing in nearby countries, the eﬀect

is the same. We are seeing radical-right parties enter into government not simply because

of their ideology or policy proposals, and certainly not because of their worthiness as an

eﬀective government leader. Rather, this phenomenon is occurring due to the threat their

newfound success poses to mainstream politicians.

After addressing how these parties get into parliament, and further how they ﬁnd

themselves in actual governing coalitions, the next question presents itself: “so what?” What

happens to a country when these rising trends take hold? And similarly, how worried should

we be? Such is the focus of the last chapter. While many have previously argued the answer

to this question is that their impact is quite limited, due in no small part to their historically

poor electoral performance, I again diverge in opinion. As the ﬁrst two chapters show, these

parties are ﬁnding ways to capitalize and become relevant policy-makers. They have taken

advantage of public funds and found ways to win elections. They have taken advantage of

4

their opponents’ feeling their threat, and bargained into positions of power. Why now should

we assume this is where their story ends?

Much of our knowledge today regarding the radical-right’s relationship with liberal

democracy is centered on case studies in the few countries where the radical-right has won

the prime ministership or presidency. This presents two problems. First, this of course

makes an inherent assumption that radical-right parties can only negatively aﬀect liberal

democracy while controlling the head-of-government. Second, it is still unclear in which

areas of democracy the radical-right’s potential eﬀects are most concentrated, even in these

limited cases. In the third and ﬁnal chapter, I attempt to right these wrongs.

I argue, ﬁrst and foremost, radical-right parties can and will have signiﬁcant impacts

on democracy and repression, even if they simply ﬁnd success as a junior coalition partner.

Indeed, when we consider this possibility, and break away from the previous notion that

democracy will only suﬀer if a radical-right party wins total control of the government, we

see these eﬀects to be far more dire than previously understood. In this chapter, I begin

by analyzing the disaggregated eﬀects radical-right parties have when they do indeed hold

executive power in a country, and demonstrate the aspects of democracy most threatened

by these governments. Beyond this, however, I illustrative the diﬀerent, yet still signiﬁcant,

impacts these parties have been having even while limited as merely a junior coalition partner.

Taken together, these ﬁndings are stark. Mainstream parties seem to be allowing

these radical-right opponents to serve with them in coalition due to the fact that they fear

what their rise in popularity will do to their own seat-share. Once in, however, these radical

parties again capitalize on their newly-found situation and begin a slow dismantling of liberal

democracy itself.

Overall, my hope is that this dissertation provides important insight into each stage of

the rising trend of radical-right parties in government. First, I focus on their initial election,

and analyze how the recent policy trend of publicly funding elections have aided their rise.

Second, I consider what happens after candidates from such parties get elected. How do they

5

take the next step and go from opposition party to ruling party, and what part do ruling

mainstream parties play in this transition? And ﬁnally, I look at the substantive eﬀects to

liberal democracy when such a transition is taken, by disaggregating the eﬀects radical-right

parties are able to have on diﬀerent aspects of democracy while in government, whether that

includes the oﬃce of prime minister, or not. Taken together, I provide some much needed

clarity on the dynamics of radical-right parties in Europe, all the way from rise to rule.

6

CHAPTER 1

PUBLIC CAMPAIGN FINANCING AND THE RISE
OF RADICAL-RIGHT PARTIES

1.1 Introduction

In recent years, the world has seen a quickening trend of rising extremism, for example:

the Sweden Democrats in Sweden, Danish People’s Party in Denmark, and the National Front

in France have seen large gains in popularity. Recently, the Alternative for Germany gained

94 seats in the latest German election, enough to make them the largest opposition party.

At the same time in Greece the neo-Nazi Golden Dawn party saw a surge in popularity,

while the Greek parliament was for several years run by a coalition of the radical-left Syriza

and the far-right Independent Greeks (extremists stick together, I suppose). More generally,

in recent years we have seen radicalism rise throughout Europe and the Americas. There is

even evidence of this trend in the United States, as some argue the election of right-populist

candidate Donald Trump is a further example of rising extremism.

A growing literature has made great strides in explaining the myriad conditions that

aﬀect the popularity of the radical-right. This paper adds to this literature by investigating

the eﬀect of public campaign ﬁnancing provisions on the ability of these parties to attract

and maintain public support. While public ﬁnancing is generally seen as “leveling the playing

ﬁeld” for all political parties, I argue it also has the (perhaps unintended) eﬀect of easing

This article was published in Electoral Studies, 66, Nicolas Bichay, “Public Campaign Financing and

the Rise of the Radical-Right,” pp 1-12, Copyright Elsevier (2020).

7

the rise of the extreme-right.1

Increasingly, public ﬁnancing provisions are a policy supported largely by the political

left (Dennis 1998) and are argued to be a mechanism to ensure more fair elections that

increase the power of the majority vis-`a-vis special interests and the wealthy (Esenberg 2010;

Panagopoulos 2011). However, and rather counter-intuitively, these policies have largely

unintended consequences. First, rather than provide a return for their left-wing supporters,

public ﬁnancing instead provides a large beneﬁt to the right-wing. And further, rather than

restore power to the majority, it reallocates power towards those on the radical-right fringe

of the political spectrum.

First, I argue that, due to speciﬁc features of extreme-right parties and the precarious

nature of their coalition of support, policies providing for the public funding of electoral

campaigns will serve to disproportionately aid them, relative to other parties. Speciﬁcally,

public monies will have a direct eﬀect of increasing the vote-share of the radical-right. This

largely comes about as the radical-right has a harder time remaining competitive, raising

funds, and attracting a coalition of voters.

For example in 2016, the year before the surge in support of the Alternative for

Germany (AfD), the AfD received over EUR 6 million in public funding, the highest as a

percent of total revenue of any party in the Bundestaag (Deutscher Bundestag 2018). Many

in the media questioned whether public funding would be exceptionally beneﬁcial to the

AfD, allowing them the means to go build a coalition of support from the general public.2

Beyond the direct eﬀect, I ﬁnd public ﬁnancing severs the inherent ties between

economic performance and extreme-right support. The literature on the extreme-right has

found that economic downturn greatly aides their vote-share, while economic growth hurts

1Throughout this paper I use the terms “extreme-right,” “far-right,” and “radical-right” interchangeably

to refer to the general party group on the far-right of the political spectrum.

2See for example “Germany’s far-right party will get millions more in taxpayer cash” in CNN Money or

“German far right’s vote surge means ﬁnancial bonanza” in Politico EU.

8

it. However this appears to no longer be the case in publicly funded elections. For one,

public money ensures a suitable stream of funds regardless of economic conditions (making

the radical-right no longer dependent on a poor economy to increase the salience of their

platform). And second, public funds drastically diminish incumbency advantage, whose

eﬀect is maximized during periods of economic growth.

Importantly, these same eﬀects are not shared by extreme-left parties. Due to impor-

tant and unique characteristics of far-right parties, public campaign ﬁnancing seems to have

a disproportionate and exclusive eﬀect on them. Thus, this study contributes by furthering

our understanding of the important and oft-ignored diﬀerences between the two ideological

wings of extremism.

Interestingly, there has been very little research done in this area. Two somewhat

related pieces of literature utilize very small sample sizes (n = 2, n = 5) made up only of US

states. They analyze policies that introduce bias into their measurement, and further, they

arrive at contradictory results. Importantly, no study has attempted to analyze this topic

from a cross-national point of view. I examine this question by analyzing a dataset of twenty

industrialized democracies over the past sixty-ﬁve years, employing a quantitative analysis

of over 320 legislative elections.

This study goes a long way in helping to understand the dynamics of extremism, and

perhaps why their support is more prevalent in Europe than elsewhere, given Europe’s much

higher prevalence of public ﬁnancing in elections, which has been rising quickly over the last

few decades.3

The question of why these parties are rising is not entirely clear, yet it is of great

substantive importance. First, scholarly work has found their rise serves as a critical threat to

modern liberal democracy and regime stability. Overwhelmingly, extreme-right candidates

are “populist” not only in their rhetoric, but in their leadership styles. Populism on its own

3According to data from the Varieties of Democracy database.

9

has been characterized as one of the great threats to democracy (see for e.g. Muller 2016).

Populists have been linked with eroding the constraints on executives (Houle and Kenny

2018), pushing countries into authoritarianism (Levitsky and Loxton 2012), and causing

deep polarizations of society (Mudde and Kaltwasser 2012). Above all, there is a strong

trend of populist leaders, once elected, subverting democratic institutions and consolidating

power for themselves (see Mudde and Kaltwasser 2012).

Beyond the perils of populism, many have noted extreme-rightists’ tendencies of inex-

perience in terms of policy-making, and of presiding over particularly unstable governments,

with frequent mass rioting (Powell Jr 1981). Further, it is evident historically that extrem-

ism brings with it serious eﬀects on personal, ﬁnancial, and emotional health. Any insight

in to what makes these possibilities a reality is worthy of careful study.

1.2 The Rise of Radical Parties Over Time

It is, no doubt, diﬃcult to deﬁne any term that aims to encompass such a large group

of separate political parties. Numerous scholars have directly pointed out this diﬃculty

(see Mudde 1996). Many have noted the extreme-right’s views on immigration, nationalism,

anti-tolerance, and trade and deﬁning characteristics. Others have noted their penchant for

violence and anti-democratic tendencies. Overall, however, extreme-rightists tend to share

an “anti-system” view of the government. That is, they seek not to simply change the

composition of the government, but to change its very structure.

Many of these contemporary deﬁnitions follow from Sartori (1976), where he argues

an anti-system party “would change, if it could, not the government, but the system of

government.” While there are certainly non right-wing anti-system parties, right-wing parties

that are also anti-system make up the “extreme right-wing” (see De Bromhead, Eichengreen,

and O’Rourke 2012), a term which includes both contemporary neofascist and right-populist

parties (Golder 2003).

Given this deﬁnition, what do we know about their rise in popularity? The ex-

10

tant literature regarding support for the extreme-right can be broken down into two main

approaches: that of the supply-side, and of the demand-side. The demand-side approach

analyzes extremist vote-share as a function of citizens’ desire to vote for such a party or

candidate, while the supply-side looks instead to institutional features that allow such a rise

to occur.

Ever since Jackman and Volpert (1996), a lasting and robust ﬁnding of the demand-

side literature argues that decreased economic growth is a main vehicle for driving in-

creased extreme-rightist support (Br¨uckner and Gr¨uner 2010; De Bromhead, Eichengreen,

and O’Rourke 2012; Funke, Schularick, and Trebesch 2016). These works all suggest that

as an economy worsens, we can expect public support for far-rightism to increase; thus, a

healthily growing economy serves as an important deterrent against extremism.

This general causal story posits that extreme-rightists tend to blame foreigners, immi-

grants, minorities, and other “outsider” groups for the ills that exist in the country (Rydgren

2018). This scapegoating, of course, is much more prevalent in times of economic downturn

when there exist poor conditions to blame on these “outsiders.” In good times (i.e. during

periods of economic growth) there are no ills for the “outsiders” to bear. Given that extreme-

rightists are characterized as “anti-system,” it is hard to build a convincing argument for

voting against the current system (i.e. for an extreme-rightist) when you are well-oﬀ. Much

in the same way incumbents perform better in good times (Hibbing and Alford 1981). In

times of struggle, however, voting to change the current system has a much more compelling

argument. Importantly, evidence suggests that these features are unique to extreme-right

parties, and not general to niche-parties as-a-whole (Br¨uckner and Gr¨uner 2010; Funke,

Schularick, and Trebesch 2016).

Indeed, much research analyzing the linkage between the economy and the diﬀusion

of populism and radical parties has found rich results. For example, Algan et al. (2017)

ﬁnd that increases in unemployment are strongly correlated with voting for non-mainstream

parties, due largely to the economic insecurity felt by voters. Many others have found similar

11

results for populism in general and economic crisis (e.g. Guiso et al. 2017). In many cases, this

result is described as a cultural backlash (Rodrik 2018), whereby the “losers” of development

turn to other means of representation (Betz 1994). Much in the same way British citizens

who lived in areas with high levels of imports were more likely to vote to exit the European

Union (Colantone and Stanig 2018), a depressed economy incites fears of economic wellbeing,

prompting those who have not beneﬁtting from the changing global economy to fall out of

support of the mainstream political parties, and instead ﬁnd a champion of an alternative

system. Increasingly, they have seen the radical right as this champion (Loch and Norocel

2015). More recently, this has been described as a problem of social integration, where

those marginalized by society (in part by losing out on the beneﬁts of modernization) feel

animosity towards traditional political elites, and thus go looking for alternatives (Gidron

and Hall 2020).

Beyond economic conditions, other factors shown to inﬂuence the demand of extreme-

rightism include immigration (Arzheimer 2009; Golder 2003; Knigge 1998; Swank and Betz

2003), unemployment (Golder 2003), inequality (Han 2016), welfare provision (Swank and

Betz 2003), and education (Mayer and Perrineau 1992).These studies all paint a similar

picture: when people feel threatened– whether that be in terms of a bad economy, rising

immigration, and so on– they become more likely to approve of more “extreme” manners to

return to security.

Apart from citizen demand for extremism, the ability of the political system to trans-

fer this demand into actual vote-share stands as a second major topic of study. The empirical

implication along this avenue of work ﬁnds that the more permissive an election (that is,

multipartyism, proportional systems, and low electoral thresholds), the higher the vote-share

for extreme-right candidates (e.g. Jackman and Volpert 1996). The mechanism here is clear:

high permissiveness eases the barriers of entry for radical-right parties. I argue that public

ﬁnancing can be added to this list of electoral permissiveness that aids right-extremism.

Before understanding why, it is important to understand the pressures public ﬁnanc-

12

ing of elections exerts on electoral outcomes. While the topic of the macro-economy’s eﬀect on

extremism has largely been covered, the implication diﬀering electoral regimes may have on

extremists’ ability to gain support is lacking, especially in the sub-area of campaign ﬁnance.

Several studies have found confounding results relating to public ﬁnancing and extremism or

polarization. One such study (Hall 2014) examines campaign ﬁnance in ﬁve US states, ﬁnd-

ing that public funding leads to increased candidate polarization and divergence. Generally,

they argue public ﬁnancing tends to ease the process of extreme candidates being elected in

these states that have adopted public campaign ﬁnance. This is largely due to the erosion of

incumbency advantage that usually serves as a barrier of entry to extreme challengers. Con-

versely, another study of two US States (Masket and Miller 2015) ﬁnds no diﬀerence between

ideologies of those politicians elected from traditionally– or publicly–funded campaigns.

Not only have the two studies that have been done in this area arrived at diﬀering

results, they have focused solely on the United States, comparing a very small number of

cases (n = 2, n = 5). Further, these studies suﬀer from a fault in research design. Two of

the ﬁve US states included in these previous studies have only partial-funding of campaigns

available from the public purse, thus attenuating whatever result we may expect to see,

and reducing the true number of states with public ﬁnancing being studied to a mere three.

Additionally, and arguably more problematically, candidates in all these states may either

opt-in or opt-out of public funding. This option inherently creates a large self-selection

problem, biasing the presented results. Further, the focus on the United States also serves

to ignore Europe, the region not only most aﬀected by the recent surge in extremism, but

also a much more prevalent source of public campaign funding.

More interestingly, some previous work has argued that public funding should actually

serve to aid the more dominant parties in an election (Jones 1981), calling it a heavily biased

policy aimed to protect incumbents (Winter and Bolton 1973). Thus, given the contrary

results, limited samples, theoretical disagreement, and potentially biased research design,

this puzzle of public ﬁnancing’s eﬀect on an election remains in the literature, and it is

13

therefore evident that more research is required on this topic. In the next section I develop

a theory arguing that public ﬁnancing will serve to disproportionally aid far-right parties,

relative the mainstream and other niche parties.

1.3 Theoretical Framework

In general, it is hard for extremist candidates to be competitive.

Indeed, in my

own sample, extreme-right parties receive an average 4.13% of the vote in parliamentary

elections, and only cross 15% of the vote-share in less than 1 in 10 elections. One reason

for this is their noted diﬃculty in raising campaign funds relative more mainstream can-

didates (e.g. McCarthy, Poole, and Rosenthal 2006; Oklobdzija 2016). Not only does this

lack of fundraising make mobilization and communication with voters diﬃcult, but unlike

mainstream parties who adjust their policy platform to reﬂect changes in public opinion,

extreme-right parties (as well as other niche parties) do not, and in fact tend to fare worse

when doing so (Adams et al. 2006).

This speciﬁcally puts extreme-right parties in an exceedingly precarious place. Not

only is their platform too rigid to adjust to the prevailing public opinion, but they do not

have the resources to build a coalition of support around this platform. Rather, their ability

to attract an audience is dependent on an economic downturn. It is well-established that

economic recession increases the support of extremist candidates, while economic prosperity

serves as a deterrent against them. This comes about largely as issues included in the

extremists’ platform (the struggling economy) rise in salience as the economy worsens (e.g.

Singer 2011). Ordinarily, a healthy economy reduces the salience of economic issues, limiting

the audience of an extremist platform. Given the hardship these parties have at raising

money, the decreased audience that comes with a healthy economy drastically hinders their

ability to gain support from the electorate.

This eﬀect may be magniﬁed by the strategic considerations (e.g. Abramson et al.

2009) those who do share the extreme-right ideology make in deciding to vote for a more

14

competitive party, as a strong economy may serve to signal a lack of viability of extreme-right

parties. Thus, some people may not be voting for the far-right not because of an ideological

incongruence, or even an ignorance of their platform, but rather due to the perception of

non-competitiveness (a sort-of collective action problem).

However, public ﬁnancing allows an extreme-right party to overcome some of this

barrier. It aﬀords the party resources to spread their message (in the form of campaigning,

advertising, mobilization, etc.), as they traditionally struggle from raising their own funds.

Beyond spreading their platform, public ﬁnancing erodes incumbency advantage, diminishing

(and, in the case of full public ﬁnancing: closing) the fundraising gap between incumbent

and challenger (Mayer and Werner 2007; Miller 2008, 2011). This makes it easy for once-

uncompetitive challengers to mount a successful campaign.

Along with an erosion of incumbency advantage, public ﬁnancing severely dilutes

the inﬂuence of special interest groups, simultaneously amplifying the power of the masses

(Baron 1994; Hall 2014; Jones and Borris 1985; Malhotra 2008). For example, Baron (1994)

argues candidate positioning is closer to the median-voter in publicly funded elections, as

candidates are more independent from interest groups (which tend to be more moderate

than individual voters).

Thus, I argue that public ﬁnancing will have a direct impact in helping extreme-right

parties by aﬀording them funds they would not otherwise have, by eroding incumbency

advantage, and by reducing advantages more moderate parties tend to have in the form

of support from special interests. This eﬀect follows from the extreme-right’s diﬃculty in

terms of fundraising, and that individuals (who are more ideologically extreme than interest

groups) gain voice. This is especially important as radical-right voters tend to be more

economically disadvantaged (Ignazi 2006; Inglehart and Norris 2016; Kitschelt and McGann

1997), and thus likely not able to donate signiﬁcant amounts.

In addition to this direct relationship, I expect public ﬁnancing to mediate the eﬀect

economic growth has on limiting extremist support. Beyond providing a mechanism for

15

message spreading, public ﬁnancing’s eﬀect of limiting incumbency advantage should serve

to not only partially overcome the collective-action problem, but to allow the extreme-right

to remain competitive even at higher levels of economic growth, as compared to an election

with no public funding.

Historically the radical-right has not been very successful in terms of winning elec-

tions. This noted diﬃculty in building a coalition of support is mitigated when the nation

economy begins to slow. In this case, rather than building a coalition by using their limited

resources to mobilize, the failing economy greatly increases the salience of their platform, in

eﬀect delivering them an audience “for free.” Public funding, however, allows these parties

to go out and spread their platform even when a struggling economy does not deliver them

this audience. In other words, public ﬁnancing makes their viability no longer conditional

on economic downturn. Thus, even while their platform may not be as salient in a strong

economy as it would otherwise in a poor economy, increased resources allow them the op-

portunity to spread their message, increase advertising throughput, and mobilize voters. In

other words, where a poor economy serves to “push” voters to the extreme, public funding

should allow extreme parties to “pull” voters to them.

Beyond this, it is important to note voters do not simply vote based on economic

considerations. For example, experimental work (Dinas et al. 2019) has shown that similarly

situated areas (i.e. with the same socio-economic proﬁle) with diﬀering experiences regarding

the refugee crisis can lead to very diﬀerent levels of far-right support. Or, take the election of

Donald Trump, whose rise to power coincided with a strong US economy. But nonetheless,

his abundant media coverage provided a strong mechanism for his platform to take hold

(see for e.g. Wells et al. 2016). In this same way, public funding should provide an ample

platform for the far-right, evoking a desire for the extreme platform, even in an economy

which may not directly foster it.

Further, like in the previous hypothesis, it is important to note again that a poor

economy serves to erode incumbency advantage (Erikson 1989; Kayser and Peress 2012),

16

beneﬁting those who are not incumbents (which extreme-rightists rarely are).

In a good

economy, conversely, incumbents tend to be rewarded, hurting those who are not already

in oﬃce (again, extreme-rightists). Public ﬁnancing, however, mitigates this incumbency

advantage on its own, even at high levels of economic growth. Thus, given that public

ﬁnancing erodes incumbency advantage, and incumbency advantage is most prevalent during

a strong economy, it follows that public ﬁnancing should increase the competitiveness of the

radical-right in times of economic prosperity.

In a way, public ﬁnancing allows an extreme-right party to behave more like a main-

stream party, receiving a suitable ﬂow of contributions at all levels of economic growth, and

thus able to compete beyond the condition of economic distress. Thus, my two hypotheses:

HYPOTHESIS 1: As public ﬁnancing of political campaigns in-

creases in a country, radical-right parties will see their vote-share

grow.

HYPOTHESIS 2: As public ﬁnancing of political campaigns in-

creases in a country, the role economic growth plays in decreas-

ing radical-right support will be mitigated.

In eﬀect, there are two forces at work here: (1) a stable and uniform increase in

funding, which should act to increase competitiveness at all levels of economic growth. (2)

an erosion of a strong economy’s inherent incumbency advantage, coupled with the ability

of an extreme-rightist to maintain its audience irrespective of economic improvement.

Of course, it is wholly diﬃcult to ﬁnd direct evidence supporting these exact hypothe-

sized causal mechanisms. For example, there may well be non-policy reasons for seeing these

same results. The rise of the cartel party (Katz and Mair 1995), whereby mainstream parties

utilize state resources to protect their own longterm survival, may also lead to a less direct

mechanism through which the radical-right sees beneﬁt from these public funding policies.

As mainstream parties devise new strategies to ensure their own success, they simultaneously

17

provide their smaller challengers with more plausible claims of corruption and elite-serving

behavior, creating an environment in which the radical-right’s cries of anti-elitism can better

take hold (Katz and Mair 1995). Thus, as public ﬁnancing policies are passed, they may

incite backlash from voters who view these acts as self-serving or potentially corrupt, thus

driving them to seek an alternative government in the radical-right, who often espouse anti-

elite anti-corruption values (Zaslove 2004b; Ziller and Sch¨ubel 2015). Through the use of the

ensuing case examples, I provide some support for the proposed causal mechanisms, but it

is important to note there may be other forces at work as well.

1.3.1 Why the Right?

It may seem, at ﬁrst read, that these two eﬀects would work uniformly to aid all

small/niche parties. However, I expect these two forces to disproportionately increase the

vote-share of the radical-right, relative other parties for a few reasons. In terms of hypothesis

one: it is known extremist parties have traditionally had a harder time raising money from

donors. Extreme-right voters tend to be most economically disadvantaged (Ignazi 2006;

Inglehart and Norris 2016; Kitschelt and McGann 1997), thus their voters are unable to

contribute in large amounts. Therefore, guaranteed public funds should disproportionately

aid them, as other parties would simply be able to raise the money anyway. Further, the far-

right speciﬁcally performs much worse than the far-left in parliamentary elections.4 Thus,

the erosion of incumbency advantage is much more beneﬁcial to them.

Indeed, data from German party funding illustrates this well. The Alternative for

Germany (AfD) was founded in 2013 as a center-right party focused almost solely on the

abandonment of the Euro common-currency (Arzheimer 2015). At this time, they were not

considered a radical-right party (Arzheimer 2015; Rooduijn et al. 2017), nor were they made

up of particularly economically disadvantaged voters (Berbuir, Lewandowsky, and Siri 2015).

It was not until Frauke Petry ascended to the leadership and began realigning the party in

4In observed elections, the far-left receives on average 50% higher vote-share than the far-right.

18

2015 that it took on the radical-right characterization it is known for today.

Thus, the 2017 national election represents an important time in German politics,

as it is the ﬁrst federal election following their radical realignment.

In this election, the

average AfD voter was not so similar to the average mainstream voter, but rather mirrored

the proﬁle of radical-right voters throughout Europe (that is, those who feel they have been

underserved by economic modernization) (Arzheimer and Berning 2019; Lux 2018).

Indeed, in the year prior to this election, government reports show the AfD was by far

the least funded party campaigning. In 2016 the AfD raised a total of only EUR 15.6 million.

Compare this to other German niche parties Linke (the German far-left) who raised nearly

EUR 30 million, and the Greens, who raised over EUR 40 million (Deutscher Bundestag

2018).

Couple these numbers with the fact that, compared to all other German political

parties the AfD was the most reliant on public sources for its funding (that is, government

funding made up the highest percentage of total party revenue for the AfD, relative all other

parties). Thus, it becomes clear that this really is a story of disproportionate help to the

far-right, as they stand to gain the most from guaranteed funding from the public purse.

It is additionally important to note that the average far-right and far-left voter are

quite diﬀerent. The demographic proﬁle of an extreme voter of the far-right diﬀers with those

on the far-left in important areas such as wealth, gender, employment status, and urban/rural

residency (Rooduijn et al. 2017). Where far-right supporters tend to be less-educated than

median-party voters, supporters of the far-left are actually more highly educated than median

voters (Rooduijn et al. 2017; Visser et al. 2013). Thus, given diﬀerent support bases, it is

evident diﬀerent factors should impact the rise of each party-family.

For example, left- and right-wing parties observe important diﬀerences with regard

to group solidarity, with supporters of left-wing parties tending to exhibit much higher levels

(Kiess and Trenz 2019). In turn, voters with higher levels of group solidarity are more likely

to speciﬁcally vote for radical-left parties (Rooduijn et al. 2017). It is plausible, thus, that

19

left-wing parties are more able to invoke this value of group solidarity in a way to elicit

increased small contributions from the masses, as research has shown left-wing parties have

an advantage in terms of individual donations (Ponce and Scarrow 2011).

For hypothesis two, extreme-right parties tend to do well during economic recession,

when their platform of economic issues (including unemployment, immigration, etc.) be-

comes more salient, allowing their message to gain a larger audience. Evidence suggests this

is unique to the far-right, and not shared by other party types, namely the far-left (Br¨uckner

and Gr¨uner 2010; M¨uller-Rommel 1998; Visser et al. 2013). Thus, it follows that public

ﬁnancing can only mediate the eﬀect of the economy on the far-right, as there is no eﬀect to

mediate on the far-left.

I directly test these premises by re-running all main models presented to test the

same eﬀects on the radical-left (table A.1 in supplementary material). The results indicate

these arguments are indeed unique to radical-right parties due to the unique characteristics

of this party-family.

1.4 Data and Methodology

1.4.1 Measurement

I assess these hypotheses by testing a model of radical-right vote-share in lower-house

legislative elections. Data on electoral outcomes come from Funke, Schularick, and Trebesch

(2016). The sample covers historical election results in legislatures for 20 industrialized

democracies.5 I utilize these electoral returns going back to 1950,6 for a total of n = 328

5The countries included are: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany,
Greece, Ireland, Italy, Japan, the Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, the United
Kingdom, and the United States.

6For most of the sample, though due to data availability some countries enter the sample later.

20

elections.7 If two elections were held in the same year, the latter election was used. Each

country in the sample ranges from 9–30 observed elections (mean = 16.4).

The deﬁnition used for categorizing a party as “extremist” follows from De Bromhead,

Eichengreen, and O’Rourke (2012), which itself is taken from Sartori’s deﬁnition (see section

2). This deﬁnition allows for the inclusion of not simply fascist and anti-democratic parties,

but also the “new breed” of far-right nationalist parties. For example, far-right parties in

Greece would include not only Golden Dawn, but also the newly created Independent Greeks

(ANEL), among others. Beginning with the De Bromhead et al. (2012) coding, parties are

tracked over time to include splinter groups that emerge. Further, country-speciﬁc sources

are used to code new parties that are created after the sample. See Funke, Schularick, and

Trebesch (2016) for a more extensive treatment of the coding and a list of country-speciﬁc

sources.

Thus, in all models the dependent variable is the total vote-share won by extreme-

right parties in a national election. An alternative may have been a party-level analysis,

where the unit-of-analysis is party, and thus the dependent variable is vote-share each party

receives in a national election. However, a national-level analysis is preferable to a party-

level analysis for a few reasons. For one, the theoretical implications of this paper focus on

the vote-shares of radical-right parties versus all other parties. However, a party-level model

would instead compare votes for a speciﬁc radical-right party against all other alternatives

(potentially including other radical-right parties)8. This type of analysis would examine

individual party-choice, when in fact the quantity of interest is party-family choice. There

are also several methodological concerns with such a model. For one, vote-shares of individual

parties in a given election are not independently distributed, but rather rely on the vote-

720 elections removed due to missing independent variables. See the supplementary material for a model

on a multiple imputation dataset of all 348 elections.

8Thus, in any scenario where more than one far-right party is contesting an election, the substantive

eﬀect of the independent variables would be substantially downwardly biased.

21

Table 1.1: Ordinal Coding of Finance Laws

Level Deﬁnition
0:

No public ﬁnancing of elections

1:

2:

3:

4:

Small/restricted amount of public ﬁnancing

Some public ﬁnancing

Campaigns are partly publicly ﬁnanced

Public ﬁnancing makes up a signiﬁcant share of campaign spending

shares received by other parties in that election.9 It is also unclear how such a speciﬁcation

would take into account new parties created within the sample (as well as mergers, splits, and

so on). Finally, such a speciﬁcation makes an empirical test of hypothesis two exceedingly

diﬃcult (outside of a potential three-way interaction term). Thus, I follow what has been

the norm in the literature (e.g. Br¨uckner and Gr¨uner 2010; De Bromhead, Eichengreen, and

O’Rourke 2012; Funke, Schularick, and Trebesch 2016; Golder 2003; Jackman and Volpert

1996; Swank and Betz 2003), and utilize total vote-share by all extreme-right parties in the

election (that is, an aggregate of far-right vote in a given election).

The main independent variable of interest is the degree of public ﬁnancing a country

has adopted for its electoral system. This measure comes from the Varieties of Democracy

database (version 6.2), hereafter “V-Dem.” This scale starts out as ordinal from 0–4, mea-

sured via expert surveys (no less than ﬁve country-experts code each country-year). See table

1.1 for the ordinal categories used. This measurement is then converted into an interval-level

measurement through the V-Dem measurement tool, using Bayesian item response theory

models (Pemstein et al. 2015). After this conversion, the new scaled variable has a range of

-1.81 – 4.06.

Along with this measure of ﬁnance, I include an interaction term of Degree of Public
Finance × GDP Growth Rate. Such a term will demonstrate the conditional eﬀect public

9For example, if party A gains votes, it means at least one other party must lose votes (i.e. all party-level

votes in a given country always sum to 100).

22

ﬁnance and economic growth have on extreme-right success, allowing an empirical assessment

of hypothesis two.

Given the vast eﬀect the economy has on extremism, I include several variables to

control for macro-economic conditions. First, I include GDP growth rate. As has been

discussed, level of economic growth plays an important role in determining the level of

success for extremist candidates. Second, I include inﬂation rate. Taken together, GDP

growth and inﬂation tend to provide an adequate measure of overall economic conditions.

I also add a control for government debt (as a %GDP), as high debt may serve as another

factor raising the salience of economic issues, pushing the electorate towards conservative

parties (Buhr 2013). This is especially vital given the European debt crisis that occurred

during the time-period of the sample (further, I directly test for robustness to the debt crisis,

see section 2.3).

Extreme-right party support coincides with electoral permissiveness, thus my ﬁnal

control measures the number of parties in parliament. A high number of parties tends to

ease the entry of new parties into parliament (van Biezen and Rashkova 2014). Further,

controlling for number of parties in the legislature gives additional power to the test of these

hypotheses aﬀording disproportionate beneﬁt to extreme-right parties.

1.4.2 Model

Causal inference in determining the potential eﬀects of a given policy (in this case

public campaign ﬁnancing) is complicated by a potential for bias caused by country take-

up of that policy. It is possible, in other words, that some conﬂating factor increases the

likelihood of a country to adopt public ﬁnancing, and that that same factor also aﬀects the

level of extremism in a country.

Therefore, I opt to take advantage of V-Dem’s unique measurement of public ﬁnancing

as an interval-level measure, and analyze within-country variation across time.10 This is

10Though I present pooled and random-eﬀects models in the supplementary material.

23

Table 1.2: Distribution of Finance Laws

Level
0: None

1: Small/Restricted

2: Some

3: Partly

Frequency (%)

26.2

9.8

13.7

10.7

4: Signiﬁcant
Frequency measured as a percent of country-years in the sample
(ordinal-level measurement).

39.6

conceptualized as a ﬁxed-eﬀects panel regression, with the unit of analysis of country-year

measured at each country’s parliamentary election. Utilizing this measure of public ﬁnancing

allows even small changes in public ﬁnancing schemes within a country to be quantiﬁed.

Indeed, public ﬁnancing is not simply a static measure of a countries’ electoral system,

but rather is dynamic over time. The ﬁxed-eﬀects approach allows all time invariant measures

to be held constant, particularly useful for a study such as this where history, culture, and

institutions all have a large role to play in inﬂuencing both supply-side (obstacles of extremist

entry) and demand-side (citizens’ extremist predispositions) factors of extreme-right vote-

share.

The data displays a large amount of variation, suﬃcient to warrant a time-series

approach. Only one country in the sample, Switzerland, exhibits a constant level of public

ﬁnance. The average standard deviation of degree of ﬁnance within a country is 1.1, where

the overall measure across countries ranges from -1.81 to 4.06.

Further, there appears to be a moderate balance of levels of public ﬁnancing spread

throughout the sample, which provides strong data for robust estimates at all levels of public

ﬁnancing (nearly 70% of observations are concentrated at the extremes, which makes compar-

ison between zero– and full–funding especially strong) See table 1.2 for the full distribution

of laws within the sample (for presentation I present the distribution of the original ordinal-

level measurement of campaign ﬁnance laws, however, in all models I utilize the interval-level

measurement).

24

Studies have noted the problematic nature of using vote-share measurements in esti-

mations of extremism. Namely, the problem that the data is left-censored at zero (Golder

2003; Jackman and Volpert 1996; Swank and Betz 2003). This creates issues, as there is

no doubt some level of extremist support in every country. However, many elections in the

sample (160 out of 328) show zero-percent of the vote-share going to extreme-right parties.

These may be countries that (1) simply do not have extremist parties, (2) have an extremist

base, yet due to strategic considerations they vote for another party, or (3) simply have too

low a level of support to report. Therefore, I follow the widely-used practice of utilizing a

Tobit maximum-likelihood estimator, left-censored at zero.11

Tobit models, while largely intended for censored or truncated data, have been shown

to be quite useful in overcoming this exact issue. Speciﬁcally, they provide a helpful appli-

cation to overcome a ‘corner problem,’ that is when a dependent variable is continuously

distributed for positive values of y, yet for a sizable proportion of observations P (yi = 0) > 0

(Wooldridge 2010, 2012). For example, Wooldridge (2012) cites an example of estimating in-

dividual monthly expenditure on alcohol. He notes that several diﬃculties present themselves

when using a model that ignores the censored nature of the data, and recommends the use-

fulness of a Tobit to correct them. The Tobit speciﬁcation in this case takes into account the

fact that a consequential portion of observations are clustered at the zero-bound. Addition-

ally, Tobit models are very common in the literature on radical-party politics (De Bromhead,

Eichengreen, and O’Rourke 2012; Doleˇzalov´a 2015; Golder 2003; Jackman and Volpert 1996;

March and Rommerskirchen 2015; Swank and Betz 2003), and thus have the added beneﬁt

of additional comparability with a major share of the research done on the topic.

A Tobit speciﬁcation, thus, should yield unbiased and consistent estimators in the

11Though some note potential issues with ﬁxed-eﬀects Tobit models, Monte-Carlo simulation has shown
it to be both unbiased and consistent with suﬃcient cluster sizes of around 8 (Greene 2004a,b). Cluster sizes
in my sample range from 9 to 30, with a mean of 16.4. Thus, I argue there should be no concerns for this
speciﬁcation, especially compared to the alternatives of treating vote-share as uncensored in OLS, or the
potential omitted variable bias in a non-ﬁxed eﬀects Tobit. Even still, the supplementary material contains
alternative speciﬁcations and illustrates results are not conditional on model speciﬁcation.

25

face of this concentration of values at zero. Yet given potential concerns, the main model is

re-speciﬁed as a beta-regression. Beta-regression is a maximum-likelihood estimator which

treats the dependent variable as a proportion, bounded between 0 and 1 (i.e. beta dis-

tributed), with parameters reported in terms of the mean of y (Ferrari and Cribari-Neto

2004). The Beta distribution, unfortunately, assumes a zero-probability for responses equal

to zero.12 Therein is the main beneﬁt of the Tobit model, which not only assumes a positive

probability density for zero-values, but assumes a non-negligible share of observations are

in fact zero. Results remain nearly identical to the Tobit speciﬁcation and are presented

alongside the Tobit results in the next section.

1.4.3 Time-trends

While it may appear on its front that extremism in one country acts as a contagion,

spreading across borders, this is not likely the case. Figure 1.1 shows the level of extremism

over time within each sample country. There does not appear to be any strongly prevailing

trends acting out across all countries over long periods of time. Peaks of extremism in

one country (or even upward trends) do not correspond with peaks in another. As has

been argued before (e.g. Arzheimer 2009), time-eﬀects that are uniform across countries do

not exist. Rather, any correlation found between countries is much more likely a result of

countries reacting similarly to a global economic shock.

However, it is evident that trends of extremism may have lasting eﬀects within a

country. In many cases, level of public funding received by a party for an election is in-part

determined by the vote-share received in the previous election. Therefore I include a lagged-

dependent variable as a control (in this case, share of the vote won by extreme-right parties

in the previous election).

Many warn of the possibility of introducing bias to a model containing both ﬁxed-

12Given the existence of 0 values I follow Smithson and Verkuilen (2006) and transform values such that

y(cid:48) = (y × (n − 1) + 0.5)/n.

26

Figure 1.1: Extremism Over Time

eﬀects and a lagged-dependent variable (Nickell 1981).

In the robustness check section

after I present results, I outline three alternative speciﬁcations used to test for such bias.

First, I rerun the model as a pooled-tobit (i.e. omitting ﬁxed-eﬀects), second, omitting the

lagged-dependent variable, and third, as a multi-level Tobit (that is, with random eﬀects on

countries). In all cases, the results remain materially unchanged from the ones presented

below.

1.5 Results

Table 1.3 presents the results of the regression models. The ﬁrst three columns present

results from the Tobit regressions. Column one displays a baseline model without degree

of public ﬁnance or its interaction with growth, while column two includes degree of public

ﬁnance, but not the interaction with GDP growth. Finally column three presents the full

model with the interactive eﬀect. Meanwhile, the last two columns present results from the

27

Table 1.3: Results of Tobit and Beta Regressions

Degree Public Finance

GDP Growth

Inﬂation

# of Parties

Government Debt

Lagged DV

Degree Public Finance × GDP Growth

Constant

Sigma

Scale

Tobit

(2)
right
1.773∗∗∗
(0.367)

-0.316
(0.174)
-0.152∗
(0.0705)

-0.0792
(0.219)
-0.0304∗
(0.0143)
0.708∗∗∗
(0.101)

(1)
right

-0.392∗
(0.189)

-0.141
(0.0727)

0.128
(0.211)

0.00152
(0.0199)
0.801∗∗∗
(0.0708)

-3.718∗∗∗
(1.091)
4.589∗∗∗
(0.421)

-3.885∗∗∗
(1.085)
4.381∗∗∗
(0.406)

(3)
right
1.287∗∗∗
(0.320)
-0.657∗∗∗
(0.194)
-0.208∗
(0.0817)

-0.119
(0.215)
-0.0331∗
(0.0142)
0.692∗∗∗
(0.0946)
0.208∗∗
(0.0689)
-2.613∗
(1.297)
4.346∗∗∗
(0.412)

(4)

rightb
0.199∗∗∗
(0.0459)
-0.0461∗
(0.0189)
-0.0324∗
(0.0150)

0.0290
(0.0386)
-0.00482∗∗
(0.00154)
0.0794∗∗∗
(0.0142)

-4.246∗∗∗
(0.207)

Beta

(5)

rightb
0.153∗∗∗
(0.0469)
-0.0674∗∗∗
(0.0170)
-0.0400∗∗
(0.0144)

0.0213
(0.0371)
-0.00538∗∗
(0.00178)
0.0804∗∗∗
(0.0142)
0.0210∗∗
(0.00810)
-4.128∗∗∗
(0.203)

3.708∗∗∗
(0.180)

3.733∗∗∗
(0.181)

N
Pseudo R2
Log Pseudo-Likelyhood
Standard errors clustered by country and appear in parentheses. Country ﬁxed-eﬀects omitted. ∗p < 0.05,
∗∗p < 0.01, ∗∗∗p < 0.001, two-tailed test.

328
0.277
-537.6

328
0.299
-521.5

1019.2

1021.3

328

328

328
0.302
-518.8

beta regression models. The results from the two speciﬁcations are virtually identical.13 In

ﬁgures A.1 and A.2 of the supplementary material I present comparisons in the distribution

of predicted values from both models. Given that the results are same, and the norm of the

literature in utilizing Tobit speciﬁcations, the remaining discussion of results focuses on the

13Appearances in diﬀerences of coeﬃcient magnitude are due to rescaling for the beta regression, sub-

stantively the results do not change (see ﬁgures A.1 and A.2 in the supplementary material).

28

Tobit models.

First, as previous ﬁndings suggest, GDP growth indeed has a negative and signiﬁcant

eﬀect on radical-right vote-share.14 As an economy improves and grows at faster rates,

extreme-right candidates see there vote-share diminished. Interestingly, GDP growth appears

to lose the traditional level of signiﬁcance when public ﬁnance is controlled for. However as

a practical matter, the coeﬃcient and standard error remain quite similar to the baseline

model, and still maintains p = 0.07.

In the non-interactive model, degree of public ﬁnancing has a positive and signiﬁcant

coeﬃcient, providing support for hypothesis one that as public ﬁnancing increases extreme-

rightists can expect to see their vote-shares increase, as well. Figure 1.2 displays the average

marginal eﬀect of moving an election from the 5th-percentile of public ﬁnancing (i.e. the

value of the lowest 5% in the sample) to the 95th-percentile (i.e. the value of the highest 5%

in the sample). Going from a very low level of public funding to a very high level of public

funding takes an extreme-right party, on average, from roughly 2% of vote-share to over 6%,

with all else remaining equal.15 This equates to an increase in vote-share of nearly 190%.

This represents quite a large direct eﬀect of public campaign funds. Before even taking

into account the potential mitigation of the eﬀect of the economy, adopting election public

ﬁnancing provisions will, on average, nearly triple the vote-share of extreme-right parties.

I turn now to the evidence for hypothesis two. Though a graphical interpretation is of

more value, I note the positive and signiﬁcant coeﬃcient on the interaction term would seem

to indicate that the negative eﬀect of GDP growth is indeed mitigated as public ﬁnancing

increases. I examine this eﬀect more clearly by interpreting the results via ﬁgures 1.3 and

1.4. Figure 1.3 displays the relationship between the economy and extreme-right vote share

at diﬀerent levels of public ﬁnancing. Again, I use the 5th– and 95th–percentile as limits of

14I alternatively try including lags of GDP growth by up to three years, but the results remain unchanged.

15For a hypothetical country at mean levels of all control variables.

29

Figure 1.2: Marginal Eﬀect of Public Financing

the simulation. The solid line indicates a hypothetical country with very low public ﬁnancing

(a value of -1.6), and shows the standard response found in the literature: as the economy

grows at faster rates, extreme-rightism falls; as the economy contracts, conversely, extremism

increases.

The dashed line, however, depicts the relationship between the economy and extreme-

right vote-share for a hypothetical country with very high public ﬁnancing (a value of 3.4).

When the economy is weak (the left-side of the graph), extreme-right support is high in

both countries (those with a high level of public ﬁnancing, and those without). Due to the

direct-eﬀect of public ﬁnancing (hypothesis one) the level is higher in the country with a

high level of public ﬁnancing. However, as the economy improves, extreme-right vote share

decreases in cases where public ﬁnancing is low. In cases where it is high, however, there is

no signiﬁcant diﬀerence in predicted vote-share at diﬀerent levels of economic growth.

For example, look what happens when an economy transitions from shrinking at

30

Figure 1.3: Conditional Eﬀect of Economic Growth

2% annually to growing at 4%. In a country (with mean values for all control variables)

with a low level of public ﬁnancing, it is predicted that extreme-right vote-share will fall

from just over 3.5% to roughly 1.5% (a nearly 60% decrease).

In a country with high

public ﬁnancing, however, that same economic transition would have no eﬀect in reducing

extreme-right support. Thus, as hypothesized, there seems to be no eﬀect on extreme-right

vote-share as a country’s economy rapidly expands, or as it goes into recession. Rather,

extreme-rightists are able to maintain their higher level of support regardless of economic

conditions.

I illustrate this more clearly in ﬁgure 1.4. This graph illustrates the changing marginal

eﬀect of growth (i.e. the Tobit coeﬃcient) at diﬀerent levels of public ﬁnancing. It is clear

that, at low levels of public ﬁnance, GDP growth has a signiﬁcant and negative eﬀect. As

level of public ﬁnancing increases, however, the negative eﬀect is diminished. By the time

public ﬁnancing reaches 1.4 (roughly the 55th percentile), the eﬀect of economic growth is

31

Figure 1.4: Eﬀect of Finance on Marginal Eﬀect of Growth

no longer signiﬁcant as a predictor of far-right vote-share.

This is a rather counterintuitive result. Public campaign ﬁnancing has increasingly

become a rallying call for those on the political left,16 and is largely supported by more

left-wing politicians (Dennis 1998). Further, voters and candidates call for public ﬁnancing

as a mechanism to restore power to the majority and ‘level the playing ﬁeld’ (Esenberg

2010; Panagopoulos 2011) . That is, to take power out of the hands of elite and special

interests, and return it to the citizenry. However, I illustrate these policies may actually be

counterproductive to their cause for two reasons.

First, rather than helping the proponents of such measures (that is, the left), public

ﬁnancing actually has the unintended consequence of aiding radical-right parties. Due to

the features of this party-family, public monies serve to disproportionately aide them. Their

16See, for example, the “The Government By the People Act” in the United States.

32

reliance on a poor economy, inability to alter their platform to attract voters, and diﬃculty

raising funds and being competitive is key to their disproportional beneﬁt of public funds

at any level of economic growth. Thus, by promoting such measures, voters on the left may

actually serve to weaken their power. Second, public ﬁnancing does not simply reallocate

power back to the majority. Rather, they allocate power toward the fringe of the electorate,

by amplifying their voice, and continuously issuing a steady stream of cash even after much

of the public loses interest in them.

1.5.1 Robustness Checks

I conduct several checks to ensure robustness of the speciﬁed model. First, as stated

in footnote 11, several studies have noted potential complications with ﬁxed-eﬀects Tobit

models. While, given the sizable unit-cluster sizes, this is not an issue in this case and

preferable to alternative modeling strategies, I respecify the model (1) without ﬁxed-eﬀects

(instead utilizing a pooled-tobit model), (2) without the lagged-vote share, and (3) as a multi-

level Tobit model (with country random-eﬀects). In the case of the pooled and random-eﬀects

model I add controls for proportional systems and electoral threshold (dummies for 1- 2- 3-

and 4% thresholds). In all three cases, the results remain unchanged from the ﬁxed-eﬀects

Tobit presented in this paper. The interaction term in the pooled-model does appear to

lose traditional levels of signiﬁcance, but a substantive and graphical interpretation of the

result ensures there is no existent diﬀerence. This, coupled with the evidence from Greene

(2004a,b), see footnote 11, as well as the beta regression re-speciﬁcations, should lead to

great conﬁdence with the Tobit results.

In every one of these models, both the original and the re-speciﬁcations, 20 elections

were omitted by listwise deletion due to missing data in the control variables. I impute these

missing values by multiple imputation with AMELIA II and re-run the main model on the

full sample of 348 elections (rather than the 328 previously). Again in this case, the results

are unaltered and remain robust.

I utilize this multiple imputation dataset and test robustness against the inclusion of

33

several additional control variables thought to have contributed to the recent rise in extrem-

ism. Namely, I include controls for unemployment, immigration, and economic inequality.

Beyond potentially contributing to rises in extremism, inequality is especially important as

a robustness test. It is possible that a change in public ﬁnancing is not entirely exogenous,

as assumed. Rises in inequality may increase calls for adoption of public ﬁnancing (to re-

store power to the majority). This same rise in inequality may drive up support for the

extreme-right.

However, in all cases the interpretation of the results do not change. That is, in

every case, public ﬁnancing has a signiﬁcant direct eﬀect and serves to mitigate the eﬀect

of economic growth on extreme-right vote-share. I report the results of all these alternative

models and describe the procedures in more detail in the supplementary material (tables A.2

and A.3).

Next, to ensure results are not being driven by outliers, I re-run the ﬁxed-eﬀects Tobit

model with several additional speciﬁcations. First, I re-run the model while omitting extreme

values of the DV and main IV. I run three such speciﬁcations: (1) omitting observations above

the 95th percentile of vote-share for extreme-right candidates, (2) omitting observations

above the 95th percentile of economic growth, and (3) omitting observations below the 5th

percentile of economic growth. Second, I jackknife the model, running it 328 times, each

time with one election dropped. Finally, to ensure results are not being driven by recent

events like the European debt crisis, I (1) drop all observations post-2007, and (2) drop the

four countries most eﬀected by the crisis (Portugal, Italy, Spain, and Greece).

The results of these re-speciﬁcations for outlier sensitivity are reported in table A.4

of the supplementary material. In all cases, the results remain robust and unchanged.

1.6 Motivating Case: Austria

Extremism has seen a puzzling trend in recent history. Take Austria as an example.

Austria dramatically raised their level of public ﬁnancing throughout the 1970s and 1980s.

34

Note: Vertical dashed lines refer to passage of public campaign ﬁnance laws (1975, 1984, 1985).

Figure 1.5: Results in Austrian Elections

In 1970 they sat in the bottom 25th-percentile for public funds, at the same time, radical-

right parties in Austria routinely received around 6% of vote-share. By the end of the 1970s

Austria’s level of public ﬁnancing had jumped up to the top 25th-percentile. This was largely

due to the passage of the Federal Law on the Activities, Financing and Campaigns of Political

Parties of 1975, which granted public funding to any Austrian political party with either 5

seats in the National Council or at least 1% of vote-share in a national election (Greco 2011;

National Council of Austria 1975). Further changes in the early 1980s17 increased the level

of public funding in Austria to an even higher degree.

By 1990 extreme-rightists were receiving well over 15% of the vote-share, more than

double their pre-public ﬁnancing average. After the election of 1999, the extreme-right

17The Federal Act on Public Funding for Political Education and Media Information of 1984 and the

Parliamentary Groups Funding Act of 1985 (Greco 2011).

35

Freedom Party (who throughout the time-sample have met both conditions of the 1975 law)

became the second-largest party in parliament, joining in coalition with the more mainstream

Austria’s People’s Party. Importantly, this all occurred while Austria saw a healthy economic

growth rate of nearly 3% throughout the 1980s and 1990s. Figure 1.5 illustrates these changes

over time. The dashed line shows vote-share of radical-right parties in Austrian national

elections. The vertical lines depict the mentioned changes in Austrian public ﬁnancing

laws, while the solid line depicts changes in GDP for reference. The visualization makes

it clear: Austria’s far-right took oﬀ just after the rapid expansion of their public ﬁnancing

scheme, allowing them to reach new heights, even while the economy grew comfortably. For

comparison, the third line shows far-left vote-share over the same time period, making it

clear that this result is wholly unique to the far-right party family.

The large dip in 2002 is due to the backlash faced by the Freedom Party after they

entered in to a governing coalition (as tends to happen to radical-right parties when they

must transition from protester to governor). However, they quickly recovered and even

re-entered into the government coalition following the 2017 election.

The Austrian example paints the picture of a general trend: public ﬁnancing dispro-

portionately helps far-right parties. Austria is but a single example of a recent growth in

radical-rightism worldwide. I argue some of this was made possible by public ﬁnancing of

political campaigns; the results of several statistical models support this conclusion.

1.7 Conclusion

Extreme-right parties have been shown to fare worse in terms of raising funds com-

pared to their mainstream counterparts. This diﬃculty in fundraising makes it harder for

them to remain competitive in elections, unless economic downturn increases the salience

of their platform, delivering them an audience they need not pay for. Imagine instead an

extreme-right party was guaranteed funding, regardless of economic conditions. I argue that

this would serve to disproportionately aid them relative their competition, aﬀording them a

36

larger popular vote-share, while making their success no longer contingent on poor economic

conditions.

Through a quantitative analysis of over 320 legislative elections in advanced democra-

cies I ﬁnd that, even after accounting for economic conditions, electoral permissiveness, and

a host of other important factors, the degree of electoral public ﬁnancing plays a signiﬁcant

role in predicting the level of extremist support.

In fact, a country transitioning from a

very low level of public funding to a signiﬁcant level of public funding is estimated to nearly

triple the vote-share of extreme-right parties. Such an impact is not just signiﬁcant in the

statistical sense, it is of a signiﬁcantly large and important magnitude.

This is an interesting and counterintuitive result. Often, public ﬁnancing provisions

are advocated by those on the left, and labeled a mechanism to restore power to the “av-

erage citizen.” However, these policies have been demonstrated to have largely unintended

consequences. First, rather than provide a return for their left-wing supporters, public ﬁ-

nancing instead provides a large beneﬁt to the right-wing. And further, rather than restore

power to the majority, they reallocate power towards those on the fringes (that is, to the

extreme-rightists).

The radical-right has been correlated with increased executive turnover, increased

cabinet defections, and higher levels of mass rioting (Powell Jr 1981). They are often made

up of inexperienced politicians and preside over particularly unstable times. Their rise to

power is often seen as a great threat to democracy and regime stability, as they tend to

erode democratic checks on the executive, and often-times engender executive takeovers.

Take Golden Dawn, a recently popular far-right party in Greece. Nikolaos Michaloliakos,

the founder and longtime leader of the party, is currently on trial (with several of his deputies)

for leading a “criminal organization,” charges that were ﬁled after the party was implicated in

the murder of a liberal rapper, as well as the numerous attacks and murders of immigrants.18

18See “Greece Crackdown: Golden Dawn Leader Michaloliakos Charged” in the BBC.

37

Though they have remained largely uncompetitive in elections over time, parties

of the extreme-right have been noted to have profound inﬂuence on politics, even while

losing. Increased success of the radical-right, for example, has tended to lead mainstream

parties to adopt more anti-immigrant policies (Abou-Chadi 2016). Look, for example, at

Mark Rutte and the Dutch People’s Party for Freedom and Democracy changing rhetoric

regarding immigrants during the 2017 election as the far-right Party for Freedom led the

polls.19

Beyond these stark implications, understanding the determinants of their rise to power

is a particularly important task of political science, capable of shedding much light on the

inner workings of government, elections, the economy, and the interplay between them. These

results do just that, by adding a new ﬁnding to the list of the determinants of extremism,

and importantly, changing our understanding of the eﬀect of economic growth on limiting

radical-right support.

19For e.g., “Dutch PM Tells Immigrants: Act Normal or Go Away” in The Washington Post.

38

CHAPTER 2

COALITION AS COOPTATION: THE
RADICAL-RIGHT’S ENTRANCE TO
GOVERNMENT

2.1 Introduction

Following the 2017 Austrian National Council elections, the Austrian People’s Party

opted to enter into coalition with the radical-right Freedom Party of Austria for the second

time in almost twenty years. Yet, according to data from the Chapel Hill Survey, the People’s

Party was ideologically closer to the Social Democratic Party of Austria. Additionally, the

two had been serving in a coalition together for over a decade, and their combined seat-share

would have been nearly identical to the Freedom Party/People’s Party coalition. Our existent

theories of coalition building provide little leverage on their own in understanding such an

outcome. In fact prior to the election, research on coalition formation speciﬁcally mentioned

the low likelihood of a Freedom Party/People’s Party coalition (Nyhuis and Plescia 2017),

especially compared to a continuation of the incumbent Social Democratic/People’s Party

partnership.

The inclusion of radical-right parties in coalitions is noteworthy and surprising as there

are high costs to the moderate party for joining forces with a radical partner. For one, there

will likely be a backlash from betrayed voters (see for e.g. “Austria: Protests in Vienna Over

Far-Right in Coalition” in the BBC) who balk at the idea of their moderate party uniting with

such extreme partners. Second, are costs relating to governing. Radical parties are typically

rather inexperienced governors, and will very likely hurt the performance of the remaining

coalition partner. Third, there are international costs, as allies fear a potential threat to

39

liberal democracy.

In the aftermath of the ﬁrst Freedom Party/People’s Party coalition

of 1999, Austria immediately faced economic sanctions and reduced communications from

European allies (Meret 2010), with the EU proclaiming that “the admission of the FP ¨O into

a coalition government legitimizes the extreme right in Europe.”

Why then, did the People’s Party opt to instead invite the Freedom Party into coali-

tion in 2017, given the dramatic backlash the country faced the ﬁrst time? This is not a

phenomenon unique to Austria, but rather is illustrative of a trend becoming more and more

common. While radical parties1 tend to receive a small minority of votes, their presence

in coalitions is becoming much more common. Since 1990, far-right parties in Europe have

served in coalitions in 14 countries, yet held the Prime Ministership in only three. As of

now, research on the radical-right has not treated their inclusion in governing coalitions as

running counter to any current theories of coalition bargaining. However, recent events show

this assumption to be untenable.

In this paper I explore the conditions under which mainstream parties invite a radical-

right party into a coalition. Much of the literature on coalition formation has characterized

this formation process spatially, as picking partners closest to you in ideological space (Mar-

avall 2010). Yet these theories have been unsuccessful in comprehensively explaining recent

events on their own.

I introduce a novel theory which argues a crucial missing element

to these understandings of coalition formation involves the desire of mainstream parties to

co-opt their radical rivals. When radical parties pose no real electoral threat there is little

reason to pay them any attention.

If however they begin rising in popularity, and start

posing a real threat, inviting them into coalition may become less costly than allowing them

to continue on as a political opponent. First, this serves to assure voters that the major

parties are aware of their discontent, and are taking the issues seriously. Second, as radi-

cal parties can largely be characterized as protest parties, their time in opposition is spent

1Throughout this paper I use the terms “radical-right” and “far-right” interchangeably.

40

railing against the current government and dominant parties. Once they become part of the

government however, the protesting seems to stop, and the radicals’ growing damage to the

existent powers is diminished.

Utilizing evidence from over 200 European coalitions, I ﬁnd that as the radical party’s

platform rises in salience–that is, as conditions are prime for their imminent rise in support–

they are much more likely to be included in a county’s governing coalition, regardless of their

seat-share in parliament and ideological incongruence to the ruling party. This evidence

demonstrates current theories of coalition building do not tell the full story. Where, clearly,

literature has already shown that rising platform salience serves to increase the seat-share

of these parties–this paper demonstrates the eﬀects are far more serious. I show that, even

given very low representation, a radical-right party is quite likely to enter government as

national conditions change to give their platform high salience.

2.2 Coalition Formation in Europe

Much work has been done on the determinants of coalition building in parliamen-

tary democracies. Generally, the literature can be broken down into two main paradigms:

oﬃce-seeking (iconcerned with gaining political power) and policy-seeking (concerned with

enacting policy change). Thus, the two most important factors in predicting coalition forma-

tion are seat-share and ideological positioning (M¨uller and Strøm 2003). The oﬃce-seeking

theory argues some politicians simply have a goal of getting into oﬃce. Riker (1962) ar-

gues that politicians will form minimum winning coalitions, a coalition which passes the

threshold for a majority, but by as small a margin as possible. This gives the coalition a

majority control of government, but divides the rewards of maintaining control amongst the

fewest oﬃce-holders, thus maximizing the reward. Work has shown however, that simple

oﬃce-seeking approaches have not been well empirically supported (see for e.g. M¨uller and

Strøm 2003). Rather, it is apparent that party ideology must be considered, as well.

Thus, many in the literature advanced policy-seeking theories of coalition formation.

41

Axelrod (1970) and de Swaan (1973) add ideological constraints to the minimum winning

coalition theory of formation, changing our conception to instead a minimum connected

winning coalition. In this way, coalition formation is often understood spatially, describing

parties as ideologically placed on a left-right dimension, where coalition formation depends

on the euclidean distance between parties (Maravall 2010). This is due to the simple fact

that parties are not simply interested in forming coalitions to hold majority power, but have

actual policy preferences as well, and thus will form coalitions with parties that have similar

policy preferences (Budge and Laver 1993). This process puts radical-right parties in an

extremely precarious position, as the more extreme a party’s ideological position, the less

likely it is to enter government (D¨oring and Hellstr¨om 2013). This too follows from spatial

models of coalition building, as median parties tend to have a dictatorial role in coalition

formation, they can always choose a more moderate coalition partner who will demand less

policy concessions in return for their support (see for e.g. Bolleyer 2007).

This makes the growing trend of radical parties serving as coalition partners run

counter to what previous theories have believed. A median party will almost always have

an option of a more moderate coalition partner, who will demand less in terms of policy

compromise. Beyond this fact, there are impactful costs to be borne for picking a radical

partner, which I outline below.

2.2.1 The Costs of Radical Parties in Coalition Government

Aside from the dramatically disproportionate balance of power in terms of negoti-

ating, such partnerships remain so surprising due to the strong threats that exist for both

the mainstream and radical partners. These threats can be broken down into two main

categories: domestic and international threats.

In terms of domestic threats, strong costs exist for the mainstream partner. It is likely

moderate party voters will feel betrayed by the inclusion of such an ideologically incongruent

party in government. It is evident, for example, that coalition formation is inﬂuenced in part

by voters’ preferences for preferable ruling partners (Debus and M¨uller 2013; Falc´o-Gimeno

42

2012), and likewise, that vote-choice is partially determined by prospective ideas on possible

emergent coalitions (Bowler, Karp, and Donovan 2010; Herrmann 2014). Thus, parties

should expect some sort of a backlash when these preferences are ignored.

Further, radical parties tend to be very inexperienced governors. They are generally

characterized as “protest parties,” thus their transition into a governing party brings with it

many diﬃculties a mainstream party will have to endure. Their governance is correlated with

high levels of executive turnover, and an elevated rate of cabinet defections (Powell Jr 1981),

both likely unwanted by their governing partner. Take for instance the aforementioned 1999

coalition of the radial-right Freedom Party of Austria (FP ¨O) and the moderate Austrian

People’s Party ( ¨OVP). Problems relating to the transition from protestor to governor led

to party inﬁghting, organizational troubles, and other setbacks for the party (Luther 2007;

Meret 2010). Cabinet defections set in as several prominent FP ¨O ministers resigned, and a

new election was called in 2002. By the time the election occurred, support for the FP ¨O had

plummeted 65%, leading to their lowest seat-share in nearly 20 years.

Finally, there are clear international costs to the government when a party on the

radical fringes joins a coalition. Largely, this is due to radical parties’ tenuous relationship

with liberal democracy. Much work has shown how big of a threat radical parties, who tend

to be quite populist in nature (see Rooduijn and Akkerman 2015), are to liberal democracy

(Houle and Kenny 2018; Kendall-Taylor and Frantz 2016). As international partners and

IGOs have a clear interest in maintaining democracy, it is evident some costs may be borne

by governments who form with the inclusion of a radical party. Yet again, this is exactly

what transpired in Austria. Following the 1999 inclusion of the FP ¨O in government, the

European Union initiated sanctions against Austria, as all member-states cut oﬀ non-formal

communication (Luther 2007; Meret 2010).

In sum, existing literature suggests coalitions partnerships with radical-right parties

should be very unlikely. Yet, this is exactly what has been transpiring lately. Why would

mainstream parties agree to form these coalitions given the diﬃculties that would present

43

themselves? It is easier to understand the calculus for the radical party. For one, entering

into coalition aids in granting legitimacy, and likewise helps remove the so-called stigma of

being merely a ‘protest party’ (Dunphy and Bale 2011). However, for the moderate party,

the reasoning is much less clear.

And yet, previous literature that has dived in to this question has concluded that this

is simply a process of “politics as usual,” with existing theories of coalition building capable

of explaining this ongoing phenomenon (e.g. De Lange 2012). This, however, no longer ap-

pears to be the case. Rather, we see time and time again that the major theories of coalition

building alone are incapable of explaining many contemporary government outcomes. There

is a motivation of these moderate parties to utilize coalition building as a tool of coopta-

tion. The aforementioned oddities of recent history are explained comprehensively when this

motivation is taken into consideration alongside previous theories of coalition formation.

2.3 Radical Cooptation

This study oﬀers a new logic for understanding the trend of mainstream-right and

radical-right coalition partnership, borrowing in a way from the literature on authoritar-

ian government’s cooptation of political rivals by their admittance to national legislature.

With no doubt mainstream parties have taken notice of the recent rise in radical-right suc-

cess. Mainstream parties have begun changing their behavior to remain competitive against

the growing threat, altering policy positions (Abou-Chadi and Krause 2018; Williams and

Ishiyama 2018) and emphasizing diﬀerent issues (Abou-Chadi 2016). Indeed, it is apparent

the radical-right has grown immensely as a political threat in recent history, and center-right

parties in particular are most vulnerable to this threat (Akkerman 2012).

Importantly, the radical party thrives as a protest party (Mayer and Perrineau 1992).

Characterized as ‘anti-system’ (De Bromhead, Eichengreen, and O’Rourke 2012; Jackman

and Volpert 1996; Mudde 2006; Sartori 1976), they make a name for themself by railing

against the system of government, dominant parties, and national conditions. Understand-

44

ably, this is not ideal for mainstream parties, who rely on the current government system,

and the people’s trust in said system, to maintain power. This ‘anti-system’ rhetoric takes

hold as voters of the radical-right view themselves as ‘forgotten’: they were left out of the

monetary spoils of modernization (Betz 1994), and increasingly feel left out of society at-

large (Gidron and Hall 2020). As a result, they take hold in the radical-right’s promise as an

alternative to the mainstream political elites (Loch and Norocel 2015). Thus, by engaging

and agreeing to work with these radical parties, moderate parties may well be signaling to

voters that (1) they hear them, but (2) you do not need to abandon us for them, as we can all

get along and work together. In this way, rising discontent in a voter for the current state of

aﬀairs may slow once they see their government incorporating these alternative ideas. This,

of course is helpful to moderate parties as it prevents voters from ﬂeeing them in favor of a

radical-right party.

Beyond simply trying to provide a signal to to voters, moderate parties are likely

inviting these opponents into a coalition to stop the barrage of attacks. It is much harder

for a radical party to criticize so vehemently the government administration when it is itself

a member of said administration. What was once a source of such criticism and electoral

threat is disarmed in government and forced to tone down their critiques and focus only on

reasonably doable policy proposals (Heinisch 2003). Thus, I argue mainstream parties are

strategically inviting a radical opponent into coalition with them as a method of cooptation

to minimize the potential for electoral threat. In Finland, for example, the largest party

Centre’s decision to include the radical-right Finns Party was largely viewed as a move to

minimize the threat of the far-right, with The Guardian referring to the remaining parties

leaders’ agreement that this was the best option, lest they allow them to “gain ground in

opposition,”2 while the Economist simultaneously characterized Centre’s move as an eﬀort

2“Far Right Set to Enter Government Coalition in Finland.” 2015. The Guardian.

45

to “tame” the Finns Party.3

In sum, only when the costs associated with such a coalition (that is, the backlash

they will face) are outweighed by the costs of allowing them to continue their barrage of

attacks in opposition will they be invited into coalition. This happens when the national

conditions are ripe for the radical platform to take hold, and voters are more persuaded by

their platform. In this case, the threat posed by the radical-right far outweighs the potential

costs for a backlash. For one, high platform salience increases the potential for a ﬂight of

voters from the moderate party to the radical party. For two, this salience increases the

audience to the radical-right’s decrying of the current governing system, and those in control

of it. Thus, my ultimate expectation: as the salience of the radical-right platform (which

I deﬁne in the succeeding section) increases, the probability of seeing a radical-right party

enter a governing coalition will also increase, even while their seat-share is unchanged.

2.3.1 Coalition Invitation and Acceptance

Up until now, the process of the radical right’s entry to government has been charac-

terized as one-sided. Yet, this is not the real-world case. While the larger partner generally

“invites” the smaller to form a coalition (Lees 2001), it is not simply that a government is

formed once the moderate party extends such an invitation to a radical party. Rather, there

is a necessary second step: said radical party must accept the moderate’s oﬀer.

First, why would a radical party have any interest in serving in such a coalition,

given they are the ones who pay some of the aforementioned costs? For one, serving in

a government aﬀords them some form of legitimacy they did not have (Dunphy and Bale

2011). As simply a protest party, it may be hard for radical parties to make the case to

voters that they are able to govern eﬀectively. In government however, they gain a kind of

‘oﬃcial’ status. It is evident that they are now seen by a mainstream, large political party

as a worthy partner, capable of being trusted to oversee government ministries on their own.

3“The Right-Wing Finns Party Does Well in Finland’s Election.” 2019. The Economist.

46

Second, like any party, radical parties are policy-seeking. That is, they do not sim-

ply seek to hold political oﬃce, rather they seek to implement the policies they advocate.

Nikolaos Michalolakos, leader of the radical-right Golden Dawn in Greece once exclaimed at

a rally, “No. We are not after just some parliamentary seats. We are after battles like the

one in Thermopylae,” claiming their party is not simply after “securing some high posts.”

No doubt, it is easier to implement change from a place of power. Along with policy-seeking

motivations, vote-seeking behavior is especially important to small parties, as they must pay

close attention to to their vote-share vis-a-vis electoral thresholds that exist (Bolleyer 2007).

Entry to government, and the increase in legitimacy that comes with it, may be seen as a

way to make their impact on policy more visible, and expand their electoral base.

Thus, for the radical party, the beneﬁts of joining a government coalition can be

thought of as constant. Unlike for the mainstream partner, they are not a function of the

salience of their platform. Rather, they simply exist as a forum to garner the party more

legitimacy and help them eﬀect the policy change they desire. It would be hard for the small

junior-partner of a coalition to be similarly strategic. For one, they are almost never the ﬁrst-

actor, but rather have to be invited (Lees 2001). Second, there is a nearly insurmountable

power-balance dynamic in the coalition bargaining process whereby the large party holds all

the cards (Bolleyer 2007). The small party only gets what the large party wants to provide.

Thus, it seems a fair assumption to make that the calculus for a radical party remains fairly

constant: accept an invitation to join a coalition, as the beneﬁts almost always outweigh

the costs. Conceptualized as a formal game: the dominant strategy for the moderate party

depends on the salience of the radical platform, whereas for the radical party, the dominant

strategy does not change.

If this assumption remains unconvincing, it is important to note that such an erro-

neous assumption would only lead to modeling biases if the radical party’s decision to accept

a coalition oﬀer is correlated with the salience of their platform, otherwise the error would

be randomly distributed and thus can be safely ignored. It seems unlikely that such a corre-

47

lation exists. Yet if it were, it would almost certainly be the case of an inverse relationship

whereby the radical party may be less likely to accept an invitation in the context of high

salience of their platform. It is possible, perhaps, that when the radical platform is more

salient the party may well want to “go it alone,” remain in opposition and hope for a chance

to form a majority government of their own in a subsequent election.

In this case, how-

ever, the methodology outlined in the following section would actually lead to conservative

estimates, failing to account for the coalition invitations that were in fact denied.

The only potential problem, thus, for treating such a bargaining process as one-sided

rests on the potential of radical parties to be more likely to accept an invitation when their

platform is more salient. Yet, this intuitively makes little sense, and there does not appear to

be any apparent theoretical reason to expect such a relationship. Thus, given these realities,

it seems entirely reasonable to treat the beneﬁts for a radical party of entering into a coalition

with a moderate party to be constant, whereas for a moderate party entering into coalition

with a radical party is instead partly a function of the threat the radical party poses on the

moderate’s future electoral standing.

2.4 Measuring Platform Salience

Giving this new theory regarding coalition probability and platform salience, the next

step is not an easy one. How do we operationalize the platform salience of the radical-right?

A major feature of radical-right parties is that they tend to ﬁnd relevance as the national

economy declines (Br¨uckner and Gr¨uner 2010; De Bromhead, Eichengreen, and O’Rourke

2012; Funke, Schularick, and Trebesch 2016; Jackman and Volpert 1996). That is, in times

of economic growth, radical support is greatly limited. As the economy worsens, however,

and becomes much more salient to voters (Singer 2011), the very tenants of the radical

platform (immigration, unemployment) become much more persuasive to voters, and thus,

the radical party becomes much more threatening to the mainstream politicians. Even while

some scholars refute the correlation between economic decline and radical-right success, it

48

is the case that these periods of recession provide a greater opportunity for radical-right

parties to experience a performance surge, as support for moderate parties and incumbents

falters (see for e.g. Downes and Loveless 2018). Thus, as radical parties have similar inﬂuence

over policy whether inside or outside of oﬃce (Duncan 2010), the strategic response is to

invite them into coalition where they are no longer able to criticize the government, build a

following of opposition voters, and further threaten the incumbent party system.

Thus, my ﬁrst expectation is that, holding seat-share constant, a radical-right party

is much more likely to ﬁnd itself in a governing coalition in times of economic downturn,

when they pose a bigger threat to mainstream politics due to the enhanced salience of their

platform, and voters increased disposition to vote for them.

HYPOTHESIS 1: As the economy slows in a country, the far-

right is more likely to enter government, even as their seat-share

remains constant.

Beyond economic growth, radical parties are characterized by their nativism (Mudde

2007). That is, their belief that their home nation should be made up only of members

of the native group. Thus, anti-immigration policy tends to be the main focus of most of

these parties (Knigge 1998), with the overwhelming majority of radical-right voters sup-

porting them due to anti-immigration considerations (Akkerman 2012; Arzheimer 2018; Van

Der Brug, Fennema, and Tillie 2000). Centre-right parties no doubt have caught on to this

immigration connection. In fact, novel work has shown far-right success has driven moderate

parties to adopt more anti-immigration policy stances (Abou-Chadi and Krause 2018). And

yet, their shifting of policy positions is not enough to completely stave oﬀ the threat. In

fact, a more restrictive ‘immigration climate’ (i.e. when more parties in a given country have

more anti-immigration policy preferences) still leads to a growth in support for radical-right

parties (Lubbers, Gijsberts, and Scheepers 2002). In other words, merely changing policy

positions to remain competitive is simply not enough. Previous work has uncovered direct

49

links between macro-level measures of foreign inﬂows and issue salience (e.g. Dennison 2020;

Hatton 2021). Hatton (2021) ﬁnds that increases in both immigration ﬂows and asylum

ﬂows serve to increase the salience of the immigration issue. That is, as inﬂows increase,

citizens are much more likely to hold the opinion that immigration is the most important

threat facing the country. Thus, as is the case in a poor economy, mainstream parties will

likely see increases in foreign inﬂows as a signal of a potential future rise in popularity of

the radical-right. Given that they cannot simply change their policy positions to remain

electorally secure, in response to this threat I expect them to invite a radical-party to join

them in coalition to (1) appear to be taking the issue seriously to voters, and (2) to stave

oﬀ the electoral threat of said radical party remaining in opposition where it may criticize

and continue to decry the current political system’s handling of the issue.

HYPOTHESIS 2: As foreign inﬂows grow in a country, the far-

right is more likely to enter government, even as their seat-share

remains constant.

Indeed, I have opted to operationalize this threat by looking at the national-level

conditions during the time of coalition formation. This may seem a rather indirect test.

It may be, for example, that these conditions do not constitute a direct link to salience of

the radical platform (though work like Hatton 2021 suggests they do). Yet, using a more

direct measure of the popularity of the radical-right is problematic. For example, the most

obvious choice of such a measure would simply be the vote-share won by these parties. Yet,

vote-share and seat-share tend to be nearly perfectly correlated. Given the simple fact that

parties are more likely to enter government as their seat-share grows, such a measure is

inappropriate. Rather, the main contribution and originality of this cooptation theory rests

on the fact that, given the same representation in the legislature, a radical-right party is more

likely to enter a governing coalition due to the future threat they impose to the dominant

party system. Thus, a more abstract measure of this threat is required.

50

Literature on radical-right parties has found contextual factors (i.e. the supply-side

factors) to be immensely important in radical-right support (e.g. Arzheimer and Berning

2019), and further found that center-right parties likely recognize the beneﬁt conditions like

economic crisis brings to the radical-right, and thus the threat it brings them (Downes and

Loveless 2018). Using these macro-level measures of radical-right popularity allow for an

approximation of the threat mainstream parties are feeling, while side-stepping any method-

ological concern that our independent variable may have direct eﬀects (i.e. eﬀects not related

to mainstream party threat) on the dependent variable.

Still, to ensure robustness, following the presentation of the main results based on

these contextual variables, I provide two additional studies that alternatively test the coop-

tation theory. First, I utilize survey data from Europe and show that as opinions towards

immigrants become more hostile, a radical-right party is more likely to be included in a gov-

erning coalition. Second, I utilize spatial data and measure domestic threat as the success

of radical-right parties in a given countries’ neighbors. In both cases, the implication of the

results remain unchanged from the main methodology presented.

2.5 Data and Methodology

In this section I outline the empirical strategies used to test this theory of cooptation.

I identify parties that qualify as “radical” via Rooduijn et al. (2019). Their own deﬁnition

for parties of the radical-right follows from Mudde (2007), who identiﬁes two basic charac-

teristics: radical-right parties are nativist and authoritarian. That is, they believe (1) their

country should be ethnically homogenous, consisting only of people of the native group, and

(2) laws are strictly enforced, with any transgression greatly penalized. Parties identiﬁed by

Rooduijn et al. (2019) are coded for all of Europe dating back to 1989 by country-experts

and have been peer-reviewed by over 80 scholars of radical and populist parties.

I code every governing coalition for 30 European countries as to whether or not they

include a radical-right party for the time-period of 1989–2018. This totals 449 cabinets

51

coded over the sample, of which roughly 12% include a radical-right party. Given that the

quantity of interest is the probability a radical-right party enters government, I only include

elections where these parties won at least one seat in the legislature, leaving 254 observations.

Observations thus are at the country-year level, yet only years where a cabinet turnover exists

are included. Changes in cabinets are identiﬁed by the ParlGov dataset (D¨oring and Manow

2019), and deﬁned as any time either (1) any party membership in the governing coalition

changes, (2) the prime minister changes, or (3) any parliamentary election occurs. Table 2.1

includes a list of all countries and their respective number of coalitions included (both the

total number of coalitions coded, and the number of coalitions where radical-right parties

(RRP) won at least one seat). The table also displays the number of governing coalitions

that include a radical-right party, the average seat-share won by radical-right parties, and

the number of radical-right prime ministers who have led the country. In sum, fourteen of

the included countries have separately experienced the inclusion of a radical-right party in

government, and in all but three countries, they have served exclusively as junior partners.

To test hypothesis one, I include a measure of GDP growth (%) in a country. Previous

work has shown long-term economic growth is more likely to have an eﬀect on radical-right

success than the rate of change in one given year (e.g. De Bromhead, Eichengreen, and

O’Rourke 2012). In this way, it is likely mainstream party leaders will be more threatened

by a sustained economic recession than a sudden slow-down. Thus, I follow these ﬁndings

and use a three-year moving average of economic growth (that is, the mean growth rate over

the preceding three years)4. Beyond being more theoretically relevant, this has the added

methodological beneﬁt of reducing outlier sensitivity. This is especially important in an

unbalanced panel such as this, where observations come every few years only when a cabinet

turns over. I expect, therefore, greater long-run decreases in GDP to be correlated with a

higher probability of a governing coalition including a radical-right party.

4This is calculated as (1/3)[(gdpt0 - gdpt−3)/gdpt−3]. I alternatively use the calculation (growtht0 +

growtht−1 + growtht−2)/3 and the results remain the same.

52

Table 2.1: Coalitions in Europe: 1989-2018
Total

Coalitions w/ Coalitions w/ Avg. RRP

RRP

Coalitions RRP in Parl

RRP in Govt

Seat-Share Prime Ministers

0.21
0.09
0.06
0.27
0.00
0.03
0.03
0.12
0.00
0.05
0.08
0.00
0.00
0.03
0.03
0.23
0.00
0.00
0.10
0.01
0.00
0.02
0.06
0.15
0.17
0.00
0.07
0.11
0.23
0.04
0.07

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
0
0
0
0
0
0
0
0
5
0
0
0
2
0
10

Country
Austria
Belgium
Bulgaria
Switzerland
Cyprus
Czechia
Germany
Denmark
Spain
Estonia
Finland
France
United Kingdom
Greece
Croatia
Hungary
Ireland
Iceland
Italy
Lithuania
Luxembourg
Latvia
Netherlands
Norway
Poland
Portugal
Romania
Slovakia
Slovenia
Sweden
Total
Note: “total” value for average seat-share refers to overall mean.

14
15
15
8
14
18
9
13
11
16
16
21
10
18
12
12
11
14
21
19
9
25
14
12
22
10
25
18
17
10
449

14
15
8
8
1
7
2
13
0
10
16
4
2
10
12
10
0
0
21
8
0
5
10
12
17
0
13
15
17
4
254

5
0
1
7
0
0
0
0
0
2
1
0
0
1
0
5
0
0
5
0
0
2
0
3
5
0
1
8
6
0
52

For hypothesis two, there are several possibilities in terms of operationalization. Many

have noted that recent anti-immigration sentiment is due to the ongoing refugee crisis in

Europe (e.g. Arzheimer and Berning 2019), with exposure to the refugee crisis shown to

be a driver in radical-right support (Dinas et al. 2019). Thus I choose to test the second

hypothesis by utilizing a measure of asylum inﬂows, as others have done (Swank and Betz

2003), normalized as a percent of total population. This measure, on top of being most

theoretically relevant, also suﬀers least from missing data problems that often arise when

utilizing immigration data.

53

I include a host of controls in these tests. First, it is clear that as the radical platform

rises in salience, they are likely to win more seats, and that greater seat-share in turn has a

direct eﬀect on the probability of entering government. Thus, the ﬁrst control measures the

seat-share of radical-right parties in the election that led to the coalition. My hypotheses

argue that radical-right parties will be more likely to enter government when their platform

is more salient, and therefore more threatening to mainstream parties. This of course must

be the case even given constant seat-share. It is also evident that institutional factors have

a large role to play (M¨uller and Strøm 2003), and as such I include a control for whether the

country is a parliamentary or presidential system, and a measure of the eﬀective number of

parties in the parliament. Additionally important is the level of democracy in the country,

thus I include the county’s level of democracy (from Polity) as a control. I also include the

country’s GDP per capita (logged).

Finally, given the important work on minimum connected winning coalitions (Axelrod

1970; de Swaan 1973), it is evident that parties whom are ideologically closer together are

more likely to form a coalition with each other. To account for this, I include a measure of

the diﬀerence in left-right ideology between the majority coalition party (i.e. the party that

ends up holding the prime ministership) and the radical-right party contesting the election.

If more than one radical-party contests an election, I utilize the seat-share weighted average

of their ideological scores, that is their seat-share as a proportion of total radical-right seat-

share (see equation 2.1).

Ideology measures for parties come from D¨oring and Hellstr¨om

(2013).5 Each party’s ideology is measured on a 0-10 scale, where higher numbers indicate

further rightist ideology. Thus, the ideological range for election e between radical-right
party r ∈ R and election winner g is given by equation 2.1.

5The measures from D¨oring and Hellstr¨om (2013) themselves are taken from Benoit and Laver (2006);

Castles and Mair (1984); Huber and Inglehart (1995); and the Chapel Hill Expert Survey.

54

(cid:34) R(cid:88)

r=1

(cid:35)

− Ig,e

× Ir,e

Vr,e
VR,e

Ideological Rangee =

(2.1)

Where I is party ideology, V is vote-share, e refers to the election index, g is the party that held the

prime ministership following the election, r is a speciﬁc radical-right party contesting the election, and R all

radical-right parties contesting an election.

The modeling strategy takes the form of multilevel probit regression, conceptualized

as coalitions nested within countries. Thus, I allow for country random-intercepts. I include

cubic-restricted time splines, to account for temporal trends in the popularity of radical-

parties. Finally, given missing data in the controls, I rerun the model utilizing multiple

imputation to impute missing data.

2.6 Results

Table 2.2 presents results from the multilevel probit regressions. Column one presents

a simple model with no controls other than radical-right seat-share, column two presents the

full model, while column three presents the same model utilizing a multiple imputation

(MI) dataset. Utilizing the MI data increases the number of coalitions analyzed from 163

to 255. As there are not any signiﬁcant diﬀerences between the two, I focus the rest of

the interpretation of results on the original speciﬁcation. Year cubic restricted splines are

omitted from the table for presentation, but are included in all models.

First, we see as expected that a radical-right party is much more likely to enter a

coalition as its seat-share increases. Indeed, the average seat-share of a radical-right party

in coalition tops 24% (compared to just 13% when out of government). Similarly, the closer

ideologically the radical-right party is to the prime minister’s party, the more likely they are

to form a coalition. In this case, the average ideological distance of the two when they do

form a coalition successfully is less than half (1.14) of the distance when a coalition is not

55

Table 2.2: Results from Multi-level Probit Regressions

Right Coalition Right Coalition Right Coalition (MI)

(1)

−18.205∗∗
(6.450)
2.586∗
(1.271)

7.977∗∗∗
(1.954)

GDP Growth

Asylum Inﬂows

Ideological Range

Radical-Right Seat-share

Eﬀective Number of Parties

Parliamentary System

GDP Per Capita (Logged)

Polity

Constant

Country-Level Variance

−2.716∗∗∗
(0.578)
0.776

(2)

−19.692∗
(9.587)
5.026∗
(2.552)
−0.647∗
(0.289)
17.431∗∗
(6.760)
−0.054
(0.327)
−0.782
(1.547)
−1.362
(1.627)
−0.248
(0.511)

10.954
(14.898)

7.528

(3)

−13.654∗
(6.659)
5.194∗
(2.116)
−0.473∗∗
(0.151)
9.734∗∗∗
(2.907)

0.144
(0.155)

0.508
(0.622)
−1.136
(0.794)

0.119
(0.246)

5.836
(6.863)

.

182

Observations
Log Likelihood
Akaike Inf. Crit.
Standard errors appear in parentheses. Year cubic splines omitted. ∗p<0.05, ∗∗p<0.01,
∗∗∗p<0.001, two tailed test.

−63.263
142.525

163

−41.950
109.901

255

.
.

formed (2.37).

I turn now to evidence for hypothesis one, that even as seat-share and ideological

distance are accounted for, a radical-right party is more likely to join the governing coalition

as economic conditions deteriorate. This hypothesis is supported by the data. I calculate

substantive eﬀects of GDP growth by randomly simulating 1000 draws of model coeﬃcients
based on their variance. I simulate along a GDP growth rate sequence of −6% to 6%, and

hold all other variables at their observed values. Results of the simulation are displayed in

ﬁgure 2.1. Overall, a one standard deviation change in GDP growth corresponds with a

56

Figure 2.1: Substantive Eﬀects on P (Radical-Right in Government)

nearly seven percentage-point change in the probability of a radical-right party in coalition.6

More generally, going from a very strong economy (growing at average 6% annually) to very

poor economic conditions (contracting at an average rate of 6% annually) nearly triples the

probability of a radical-right entering a coalition, from 12% to 33%.

Hypothesis two sees support as well. As the inﬂows of asylum seekers increase, radical-

right parties are signiﬁcantly more likely to enter government. Similar to economic growth,

a one standard deviation increase in asylum inﬂows corresponds to a roughly six percentage-

point increase in the probability of a radical-right party in coalition. Here, a general transi-

tion from zero asylum seekers to a high rate of 0.5% of population increases their probability

of entering government from roughly 15% to over 41%. These results are illustrated in ﬁgure

2.1, simulations were conducted in the same manner as previously described.

Importantly, we see these results as the context of the election are held constant. That

is, such dramatic increases in the probability of a radical-right party’s inclusion in coalition

6The marginal eﬀect is calculated for a one standard deviation increase in the IV centered about its

median value.

57

0.10.20.30.40.50.6−6−4−20246GDP Growth (%)Probability of Radical−Right in Coalition0.10.20.30.40.50.60.00.10.20.30.40.5Asylum Inflows (% Population)Figure 2.2: Eﬀect of Joint Platform Salience With Very Low Seat-Share

transpire even as seat-share and ideological distance are controlled for. The overall average

seat-share of a radical-right party throughout the sample is nearly 13%. That number jumps

to 24% when looking at coalitions with a radical-right partner. I ask then, is platform salience

alone enough to increase the probability of their entry into government with a very low seat-

share? Thus, I compute one more simulation, this time holding seat-share constant at a

very low 5%. I simulate a realistic scenario: the simultaneous increase of asylum inﬂows and

decrease of economic growth, conditions fairly typical of many European countries during

the debt crisis. I dub the radical platform to be of “low salience” when immigration is low

and GDP growth high, and conversely of “high salience” when immigration is high and GDP

growth low. I utilize the 10th- and 90th-percentiles of GDP growth and immigration for the

the limits of this simulation. Results are displayed in ﬁgure 2.2.

Overall, the change in platform salience increases the probability of a radical-right

coalition from 4% to 23% (a nearly 6x increase in probability). This is a truly striking

58

0.00.10.20.30.4High SalienceLow SaliencePlatform SalienceProbability of Far−Right in Coalitionchange, especially given the high probability of radical-right entry into government given

such a low seat-share of only 5%. I argue that such a pattern is transpiring in the real world

as moderate parties are feeling more and more threatened by the rises in salience of the

radical-right platform. Rather than allow them to continue to gain ground and turn voters

against them in opposition, moderate parties strategically invite them into coalition as a

way to co-opt them, and decrease the level of threat they pose.

2.6.1 Additional Direct Tests

One may argue that the preceding tests are rather indirect. While I argue this is in

fact beneﬁcial, and necessary due to the nature of the question, here I provide additional,

more direct tests of my theory for robustness. Namely, I respecify the tests of the cooptation

theory utilizing both survey data overtime, and spatial data from the European continent.

In both tests, the results remain signiﬁcant and show evidence for the fact that moderate

party utilize coalition building as a way to co-opt radical parties.

One option for a more direct test of this hypothesis is to utilize survey data. Rather

than measure the salience (and thus threat) of a radical-right party via the contextual factors

of a nation, one could instead simply measure the degree to which voters hold opinions in line

with the radical-right platform. Most scholars of radical-right parties agree that the largest

motivator of a vote for these parties are concerns of immigration (Akkerman 2012; Arzheimer

2018; Van Der Brug, Fennema, and Tillie 2000). Evidence suggests a large component of

this is due to “ideational concerns,” or that immigration is a threat to national identity and

culture (Golder 2003). Thus, as a more direct measure of radical-right platform salience I

utilize a question from the European Social Survey (ESS) that asks respondents to rank on

a scale from 0 to 10 whether immigration makes their country a better (10) or worse (0)

place to live.

I utilize all nine rounds of the ESS for a time period of 2002-2018.

I code any

response under ﬁve as “hostile” and any response above ﬁve “not hostile.” Observations

are then collapsed (incorporating respondent probability weights) to arrive at a country-

59

year level measure of the population’s feelings towards immigrants (i.e. the proportion of

citizens who believe immigration makes their nation a worse place to live). Because the ESS

is conducted biannually in even-number years, I use linear interpolation to estimate values

for coalitions that were formed in odd-number years. This leaves 80 coalition observations

for the 16 year time period. Otherwise, all controls remain the same as the main models

presented. Given the very small sample size, I also indicate statistical signiﬁcance at the

P < 0.10 level. Indeed, even seat-share of the party is only signiﬁcant at such a level. Results

are presented in column 1 of table 2.3. While the proportion of people with hostile views to

immigration is only signiﬁant at the P < 0.10, a ﬁrst-diﬀerence test indicates a change from

the 25th-percentile to the 75th-percentile is signiﬁcant at the P < 0.05 level.

Second, I measure threat to mainstream parties via a spatial voting data.

It is

likely that political elites view the rise of the radical-right around them as a premonition

of potential changes coming to them. While it is still unclear if any direct diﬀusion exists,

radical-right success does seem to come in “waves,” and is likely due to similar external

shocks (economy, immigration) aﬀecting neighbors similarly (Arzheimer 2009; Bichay 2020).

Thus, I argue we may measure the threat faced by political elites by measuring the success of

radical-right parties in nations close by. For each year a coalition was formed in each nation, I

calculate the average share of seats held by radical-right parties in the legislatures of countries

that share a border with that nation.7 The results for this test also yield signiﬁcant results.

I simulate the substantive eﬀects from these tests in the same manner as previously

described. In terms of public opinion, a nations’ electorate shifting opinion on immigration

from very welcoming to very hostile has dramatic eﬀects on the probability of a radical-

right coalition. Going from the tenth-percentile (0.48) to the ninetieth-percentile (0.75)

roughly triples the probability of their inclusion from 11% to 37%. We see similar results in

terms of radical-right proximity. A country surrounded by neighbors with no radical-right

7I deﬁne two countries as sharing a border if the minimum distance between them is less than twenty

kilometers.

60

Table 2.3: Results of Direct Tests

Right Coalition

Right Coalition

Immigration Hostility

Neighbor Radical-Right

Ideological Range

Radical-Right Seat-Share

Eﬀective Number of Parties

Parliamentary System

GDP Per Capita (Logged)

Polity

Constant

Observations
Log Likelihood
Akaike Inf. Crit.

(1)
9.020†
(5.451)

−0.603∗
(0.259)
5.898†
(3.060)
−0.161
(0.232)
−0.902
(1.357)

0.785
(0.895)
4.791†
(2.534)
−67.375†
(34.809)

80

−18.419
60.838

(2)

4.808∗
(2.369)
−0.244∗∗
(0.093)
6.684∗∗∗
(1.796)

0.055
(0.141)
−0.253
(0.657)

0.081
(0.597)
−0.006
(0.239)
−2.830
(4.823)

193

−56.057
136.113

Standard errors appear in parentheses. Year cubic splines omitted.
†p<0.1, ∗p<0.05, ∗∗p<0.01, ∗∗∗p<0.001, two-tailed test

representation has a relatively low chance of seeing a radical-right party enter government,

roughly 11%. By the time the surrounding area sees an average seat-share of 20% (the

90th-percentile), that probability more than doubles to 26%.

These tests paint the same picture as before. When conditions lead to the far-right

imposing a great threat on mainstream parties, those parties are much more likely to invite

them into a governing coalition, even as their seat-share does not change. This holds true

regardless of how we measure “threat:” either as the contextual factors in a nation are ripe

for the radical-right, as public opinion begins to line up with the radical platform, or even

61

Figure 2.3: Substantive Eﬀects on P (Radical-Right in Government)

as party leaders see the radical-right grow in nearby countries. However you measure it,

the implication is clear: as the radical-right’s threat grows, so too does the probability of

inviting them into coalition.

2.7 Comparing Out-of-Sample Prediction of Diﬀerent Theories

Finally, I test the usefulness of this new theory—that is, is there any predictive power

gained by not only accounting for minimum winning coalitions (MWC) and minimum con-

nected winning coalitions (MCWC), but for taking into account this cooptation hypothesis,

alongside them? To test this, I build three models, each representing one theory of coalition

formation. Each model is run on a random sample of 75% of my original data.

I then

utilize out-of-sample prediction by predicting values for the remaining 25% of the data. I

then compare predictive power of each of these models and show that accounting for the

cooptation hypothesis leads to more accurate predictions.

The MWC theory implies that the most likely coalitions are those that (1) gain at

least 50% of seats, and (2) remain as close to 50% as possible (Laver 1974; Riker 1962).

62

0.00.20.40.60.40.50.60.70.8Hostility to ImmigrationProbability of Radical−Right in Coalition0.00.20.40.60.00.10.20.3Radical−Right Seat−Share in Neighboring CountriesProbability of Radical−Right in CoalitionTable 2.4: Summary of Out-of-Sample Prediction

Main IV(s) Included

Distance to minimum winning coali-
tion

Area Under

Area Under

ROC

0.670

PR

0.533

Minimum Winning

Minimum Connected Winning

Distance to minimum winning coali-
tion, Ideological Range

0.710

0.556

Cooptation Theory

Distance to minimum winning coali-
tion, Ideological Range, GDP Growth,
Asylum Inﬂows

0.732

0.628

Thus, to base predictions on this theory I construct a variable that measures how close to a

perfect MWC the radical-right party would provide the winning party.8 This is calculated as

the absolute value of the two parties seat-share subtracted by one more than half the total

number of seats in the legislature, see equation 2.2. In other words, the number of seats

away from an ideal MWC seat-share of one-half plus one.

(cid:12)(cid:12)(cid:12)(cid:12)[Sg + Sr] −

(cid:21)(cid:12)(cid:12)(cid:12)(cid:12)

+ 1

(2.2)

(cid:20)Stot

2

Distance to MWC =

Where S is seat share, g is the party that held the prime ministership following the election, r is a

speciﬁc radical-right party contesting the election, and Stot the total number of seats in legislature.

Alternatively, the MCWC theory argues party ideology matters (Axelrod 1970; de

Swaan 1973). Thus, for this theory I add in the Ideology Range variable, which measures the

diﬀerence in ideology between the winning party and the radical-right party (see equation

2.1). Finally, to test the cooptation theory, I retain the distance and ideology variables, but

add in GDP growth and asylum inﬂows as predictors of radical-right coalitions. In all three

models the same controls from the previous analysis are included, including year splines and

country-level random intercepts. I use these three models to calculate probabilities for the

8In the rare instance of more than one radical-right party competing I utilize the seat-share of the larger

party.

63

inclusion of a radical-right party in coalition. The accuracy of these models’ predictions is

assessed in table 2.4.

I ﬁrst present areas under the receiver operating characteristic (ROC) curve. ROC

curves plot a binary model’s true-positive rate of prediction against its false positive rate,

with higher values (closer to 1) indicate better classiﬁcation power of a model. However,

many argue that precision-recall (PR) curves, which instead plots a binary model’s positive

predictive value against its true positive rate, are more appropriate for assessing the useful-

ness of a classiﬁer when the phenomena in question is relatively low-frequency (e.g. Saito

and Rehmsmeier 2015). As before, values closer to 1 indicate higher classiﬁcation power. I

present both metrics for all three models in table 2.4.

As expected, the inclusion of ideology (going from MWC to MCWC) leads to large

increases in predictive power, as the area under the ROC curve increases by 6%, while the

area under the PR curve increases by just over 4%. Meanwhile, subsequently including the

cooptation variables similarly leads to an increase in predictive performance. ROC and PR

both increased by roughly 3% and 13%, respectively, demonstrating that the salience of the

radical-right platform is an important predictor coalition outcome. In support of this, we see

that the model representing the cooptation theory is a much better predictor than either of

the models that represent the preexisting theories alone. Clearly, when it comes to radical-

right parties’ inclusion in coalitions, the process is not simply “business as usual,” but rather

is predicated on an alternative mechanism not included in previous theories. Overall, the

cooptation model provides a 18% increase in predictive power over the MWC model, and a

13% increase over that of the MCWC, in terms of area under the PR curve.

Substantively, we see this diﬀerence in useful ways. Take, for instance, the 2015

Finnish election of which I previously spoke. The simple MWC model calculates a 14.0%

chance of radical-right inclusion, where the MCWC model gives a slightly higher 17.0%

64

chance of inclusion.9 Meanwhile, the full model taking mainstream threat into account

calculates a 72.4% chance of radical-right’s participation in government. This is quite a

meaningful diﬀerence, illustrating how important it is to take such cooptation into account

when considering the coalition that may emerge in the aftermath of a national election.

2.8 Conclusion

These results have demonstrated that as mainstream parties begin to feel the “threat”

of a rising radical-right competitor, they are more likely to invite that competitor to join

them in a governing coalition. This shows our prior understanding of coalition formation was

incomplete. Whereby previously large moderate parties were thought to pick the smallest

ideologically close party as a partner (minimum connected winning coalition), this does not

tell the whole story. Rather, in many cases we see moderate parties picking partners with

quite distant ideologies. It is clear that this is not simply a decision of ideological proximity.

Further, it implies the consequences of increasing platform salience of the radical-right are

much more impactful than previously assumed. While it is clear that literature has already

shown increasing radical-right platform salience will lead to higher vote-shares for radical-

right parties, these results demonstrate that they are much more likely to enter a government

coalition, as well. Crucially, this is true even while their seat-share (1) does not change and

(2) is limited to a very low level.

Indeed, there is much more involved in the calculus of government formation than

previous theories of coalition formation have considered. The question is not as simple

as previously thought, but rather a process of weighing costs and beneﬁts. There will be

costs to forming a government with a radical party, as history shows. Allies are scared by

the threat, voters are angered by the betrayal, and a potential for governmental instability

9For these case calculations, I rely on in-sample prediction, or the full models trained on all available

data.

65

increases. And yet, when the radical platform rises in salience, the beneﬁts outweigh these

costs. For one, governments send a signal to voters that they take these issues seriously, and

are interested in working with radical-right parties to solve them. Second, in government the

radical party is no longer able to focus all their time and energy on criticizing the dominant

party structure. They no longer get to sit by and protest, but rather must attempt to govern.

For the radical party, this oﬀers legitimacy, for the moderate it oﬀers insurance. Even as

seat-share of the radical party remains constant, as conditions favoring their platform arise

(poor economy, high immigration), they are much more likely to enter government, as the

moderate party looks to quell a growing electoral threat.

There is evidence of this mechanism working. Following the Finnish election of 2015,

the Centre party formed a governing coalition with the radical-right Finns Party. Many in

the media then characterized this move as a way to attenuate their recent growth in pop-

ularity, while some called it an eﬀort to “tame” them. “Taming” a radical-right party by
inviting them into coalition is similarly how the ¨OVP−FP ¨O partnership following the Aus-
trian election of 1999 has been described, as some noticed the FP ¨O’s abandonment of many

radical proposals, adoption of neo-liberal economic policies, and a newfound commitment to

democratic values after entering government (Minkenberg 2001).

Substantively, these dynamics have far reaching eﬀects on government. Coalitions

made up of ideologically diverse partners tend to not last nearly as long as ideological ho-

mogenous ones (Axelrod 1970; de Swaan 1973), in part due to the fact that radical parties

have a measurable impact increasing government instability (Powell Jr 1981). Overall, the

characteristics that allow them to gain popularity in opposition serve to doom them once

in oﬃce, and these problems tend to be greatly exacerbated when radical parties serve in

coalition with more mainstream partners (Duncan 2010; Heinisch 2003).

Still, given our previous understanding of coalition bargaining, it seems strange to see

so many moderate parties agreeing to form governments with such radical partners, especially

given the high costs associated with such a move, both to the party and nation as-a-whole.

66

In explaining the logic underlying the inclusion of radical parties in governing coalitions, this

study sheds light on this recent trend that has been transpiring across Europe. The rise of

radical parties in governing coalitions suggests that cases such as Austria and Finland are

not aberrations, but rather indicative of a new calculus mainstream parties utilize to pick

their ruling partners.

Interestingly, once it’s all said and done, mainstream conservative parties have gener-

ally been the largest beneﬁciaries of such processes (Heinisch 2003), generally emerging as an

option to provide capable governing alone and an agenda similar, yet more reasonable, than

the radical-right party’s agenda that had been so high in demand. Where these partnerships

once seemed so strange, once taking into account the threat faced by these parties, and the

ability to quell this threat by forming a strategic partnership, it seems moderate parties

know exactly what they are doing.

67

CHAPTER 3

RADICAL-RIGHT PARTIES AND THE
DETERIORATION OF LIBERAL DEMOCRACY

3.1 Introduction

After winning power in Hungary in 2010, Prime Minister Viktor Orban did not

waste time consolidating power. Beginning with the once-independent judiciary, he and

his deputies moved quickly to usurp authority across the country, including the media, gov-

ernment bureaucracy, and the process of free and fair elections. Many of the constitutional

changes adopted in 2013 eroded institutional rule-of-law, and vastly increased the powers

of the executive (Urbinati 2019). Beyond institutional changes, the Orban government has

greatly curtailed civil liberties, causing great harms to freedom of speech and association.

This trend is not new, but rather a common characteristic of radical-right parties: when they

win power, liberal democracy may be jeopardized (see for e.g. Mudde and Kaltwasser 2012).

In Poland, for example, the radical-right Law and Justice party moved even quicker to usurp

control with leader Jaroslaw Kaczynski, no doubt inspired by Orban, exclaiming “we will

have Budapest in Warsaw.”1 Changes in Poland included collapses of judicial independence

and press freedom, and attacks on electoral fairness (Sadurski 2018).

Yet, the damage a radical-right party inﬂicts on democratic quality on average is

not well known. Much of the work looking at such relationships and processes remains

anecdotal, relying on democratic backsliding in one or few case studies. The reasoning for

1See “Poland’s new government ﬁnds a model in Orban’s Hungary” (2016) by Buckley and Foy in

Financial Times.

68

this, of course, rests with the fact that radical-right parties have not led many governments.

Since 1990, only a handful of European countries have seen prime minsters from radical-right

parties, and all have experienced some level of democratic backsliding in the aftermath of

their elections (Bugariˇc 2015; Cianetti, Dawson, and Hanley 2018). Outside of these cases,

many have argued radical-right parties have been limited in terms of impact on the quality

of liberal democracy (De Lange and Akkerman 2012). This of course, is largely due to their

lack of electoral success, as Mudde and Kaltwasser (2012) argue, it is hard for such parties

to inﬂict democracy while not holding executive power.

Yet, radical-right parties can indeed have proliﬁc inﬂuences on the state of internal

politics. For one, they have an indirect eﬀect by changing the political landscape. As these

parties grow in success and inﬂuence, the salient issues of domestic politics begin to change;

in many cases this includes increased attention paid to immigration and national institutions.

Second, by winning seats and joining coalitions they directly aﬀect politics by inﬂuencing

the national agenda and taking control of cabinet ministries.

In this paper, I argue the literature on the radical-right has dramatically underesti-

mated the eﬀects of these parties. This argument rests on two points. One, looking only at

aggregate, or high-level indices of democracy has missed many of the smaller, disaggregated

eﬀects these parties have while in power. There is no clear picture as to which aspects of

democracy are most eﬀected, and to what degree we should expect to see declines. And two,

by only looking at cases where radical-right parties have won the prime ministership (or

presidency),2 we miss their potential eﬀect while in a governing coalition. The later point

rests on my argument that we will see decreases in several metrics of liberal democracy not

just when a radical-right party wins executive power, but even as it increases representation

in a governing coalition. It is, I argue, simply the case that these eﬀects are (1) diﬀerent than

2Though I commonly refer to the executive as “prime minister,” it is not the case that I exclude or do not
expect similar results in presidential systems. Rather, within this sample it is simply the case that Cyprus is
the only nation that qualiﬁes as a purely presidential system. I discuss this more in the methodology section.

69

the eﬀect of radical prime ministers, and (2) not known as they have been under-analyzed

in the literature. Such analysis of coalitions rather than merely prime ministers allows us

to dramatically increase the number of governments in consideration; while radical-right

parties have only held the prime ministership in few countries, they have participated in

governing coalitions in nearly half of Europe. A large-N quantitative analysis of them allows

us to move on beyond speciﬁc case analysis—which, no doubt have been extremely helpful

and insightful—and allow for a more in-depth analysis of more incremental disaggregated

changes over time.

I examine 30 European countries since 1989 and analyze the disaggregated eﬀects of

both radical-right prime ministers, and radical-right representation in governing coalitions

on several metrics of liberal democracy and civil liberties. I argue that liberal democracy can

be broke into two domains: “institutional rule-of-law” and “mass civil liberties.” I expect

that, due to practical constraints and the decision-making calculus of the party, we will only

see institutional rule-of-law harmed in the cases of radical-right prime ministers. This is in

line with what others have previously argued on the topic of the real impacts of these parties.

However, I argue we should indeed see mass civil liberties negatively aﬀected in cases not

only where the radical-right serve as prime minister, but also as they serve as mere junior

coalition partners. This premise has yet to be empirically scrutinized. Overall, I ﬁnd support

for both these arguments. While the mere inclusion of a radical-right party in government

may not lead to the severe democratic backsliding seen in the more pronounced cases like

Hungary and Poland, their inclusion does lead to increased levels of repression in arenas such

as freedom of speech, freedom of the press, and academic freedom.

Beyond this dichotomy, I also ﬁnd the areas most aﬀected by radical-right prime

minsters tend to be those regarding elites and institutional constraints to executive power.

This makes sense, as radical parties tend to be highly populist in nature, and focus much

of their electoral rhetoric on ridding the country of the “corrupt elite,” while promising to

return power to the masses. Even still, their reign of power does bring smaller, yet still

70

signiﬁcant, negative changes to mass civil liberties and freedoms.

The assumption that radical-right parties cannot do damage unless they control the

government has led to a severe underestimation of their true threat to liberal democracy.

By ignoring cases where they enter into a government coalition and thus exert control over

government ministries, we have looked past the true threat they pose. Take, for example, the

sweeping changes the far-right Alliance for the Future of Austria was able to make regarding

asylum law while holding the Ministry of Justice under Chancellor Sch¨ussel. Rather, this

paper contributes to the literature in two ways. First, by disaggregating and quantifying

the actual real-world eﬀect radical-right parties have had on liberal democracy in Europe.

And second, by demonstrating the far-reaching eﬀects radical-right parties have on society

through their mere participation in national government.

3.2 The Inﬂuence of Radical-Right Parties

Much discussion has centered on the limited impact of radical-right parties in the

legislature. (Mudde 2013b). As a party generally relegated to the opposition, and in limited

cases junior coalition partners, it has been hard for them to acquire power most consider

needed to inﬂuence real change. Yet, at the same time, it is clear we have seen far-reaching

impacts of their rise.

One of the most researched impacts of their rise, for example, rests on their inﬂuence

in changing immigration laws (Abou-Chadi 2016; Abou-Chadi and Krause 2018). The rise

of radical-right parties has been shown to cause moderate-right parties to signiﬁcantly alter

their policy positions regarding immigration, adopting more restrictive preferences. Thus,

on the one hand these parties clearly are able to have an indirect eﬀect on domestic politics,

by forcing those who are in power to move towards them in terms of policy preferences. Yet,

several cases have shown them to have a clear direct eﬀect on policy as well, even while

cut-oﬀ from majority power. For example, in 2005 the Alliance for the Future of Austria

(BZ ¨O) entered into Austria’s governing coalition. While in power, the BZ ¨O was given control

71

over the powerful Ministry of Justice and pushed through many changes to Austria’s asylum

law, making deportations easier, legalizing force-feeding, and making it easier to criminally

prosecute lawyers and NGOs who aimed to help asylum seekers (Fallend 2012).

These parties’ rise also changes the issues that are at the forefront of political debate,

often forcing their mainstream opponents to emphasize diﬀerent issues in order to remain

competitive (Abou-Chadi 2016). Oftentimes, these issues include structural reforms, minor-

ity rights, law and order, and the democratic process itself (De Lange and Akkerman 2012).

These speciﬁc issues, of course, are at the forefront of liberal democracy. Thus, in this way

radical-right parties are able to exert their newly-found inﬂuence in a way to dramatically

change the conversation of domestic politics. They are able to bring the issues of their

agenda to the forefront and force mainstream parties to engage in dialogue with them.

Thus, it is clear that even outside of oﬃce radical-right parties have found a way

to sway politics in the countries they operate in. And further, they clearly are able to

have inﬂuence on policy while serving as minority coalition partners. In the next section

I examine then radical-right’s relationship with democracy, then present a theory joining

these two ideas together, and test for their eﬀects on democracy and repression when they

see increased electoral success, both inside and outside the oﬃce of prime minister.

3.2.1 The Radical-Right and Liberal Democracy

Radical-right parties have a troubled history with liberal democracy. It is the very

characteristics that make a party “radical” that stand directly opposed to the notion of

liberal democracy (Mudde 2007). Importantly, it is not that these parties wish to do away

with the democratic system altogether,3 but rather that the guiding principals of the party

family are simply incompatible with the core tenants of liberal democracy. Generally, we

can see this characterization manifest in two distinct ways: the parties’ relationship with

3Such parties who do remain opposed to democracy itself are considered “extreme-right,” rather than

“radical-right.”

72

institutional rule-of-law, and with mass civil liberties.

Institutionally, the radical-right’s incompatibility with liberal democracy can be ﬁrst

seen by examining their rhetoric. Radical-right parties exhibit quite an extensive overlap

with populist parties. Of the parties in Europe classiﬁed as “far-right,” for example, over

75% are also considered populist (Rooduijn et al. 2019). Along with their populist rhetoric

and leadership, radical-right parties tend to be very anti-pluralist in nature (Muller 2016;

Rydgren 2017). This means they tend to espouse the belief that they are the only ones

capable of legitimate rule, and they alone can ﬁx the “ills” facing the nation. Take for

example, President Erdo˘gan speaking of his Justice and Development party, claiming “We

are the people. Who are you?” (Muller 2016), or President Trump, speaking on terrorism

in 2016, arguing “I am your voice, I alone can ﬁx it.”4

This of course is at odds with liberal democracy, which requires pluralism. Indeed,

much of the threat causing fear among scholars of the radical-right rest on their historic

tendency to erode institutional rule-of-law and constraints on their executive power. While

such actions pose a serious threat to liberal democracy, at the same time we have not seen

radical-right parties push for an outright abandonment of democracy. Rather, much of

radical-right populist rhetoric focuses on taking power away from the corrupt elite and the

“out-groups” and rightfully returning it to the masses (Muller 2016), or to the “real people.”

This ﬁne line can be seen by examining the Belgian Vlaams Belang (VB) party. While

claiming to be a party staunchly in support of democracy, the VB also supports institutional

reforms aimed at consolidating the power of the executive, eliminating Belgian provinces,

and abolishing the upper house of the Belgian National Parliament (De Lange and Akkerman

2012).

Problematically, there are cases which show this incongruence to extend beyond mere

rhetoric. Take, for example, Erdogan’s recent actions in Turkey, considered as leading to

4See “I Alone Can Fix It” (2016) by Yoni Appelbaum in The Atlantic.

73

a breakdown of democracy. His constitutional changes dramatically increased the power

of the executive and abolished the oﬃce of prime minister. Even still, these changes did

little to aﬀect voting rights in Turkey (in fact, the changes were enacted via a popular

referendum). This may seem a subtle diﬀerence, but it is an important one. The rhetoric

of these parties is often categorized as anti-pluralist (Muller 2016), with both voters and

candidates believing that their party is the only one capable of solving the nation’s current

problems. As Muller argues, when these parties demonstrate values counter to democracy,

the people simply believe this is being done for their own beneﬁt (them, being the voters,

the only “true people”).

Similarly, we can see evidence of the radical-right having an equally tenuous rela-

tionship with mass civil liberties. Liberal democracy requires people of all diﬀerent beliefs,

positions, and backgrounds to matter equally in the democratic process. Rather, radical-

right parties seek to exclude those who do not line up with their idea of morality or support

their own party. Again, the Vlaams Belang in Belgium is illustrative of this. In several party

publications the VB has openly shown hostility toward ideas of gender equality (Vlaams Blok

2004b) and referred to the idea of multiculturalism as a “mistake” (Vlaams Blok 2004a).

Clearly, one of the biggest groups these parties seek to exclude are immigrants (Bar-

On 2018; Mudde 2007). Radical-right candidates tend to blame immigrants as the scapegoats

to their nations problems, and argue for strict controls of their intake (Zaslove 2004b).

Overwhelmingly, this anti-immigrant rhetoric has manifested as islamophobia (Kallis 2018),

with radical-right parties such as the Dutch Party for Freedom, the German Alternative

for Germany, and Greece’s Golden Dawn paying much attention to anti-islamic rhetoric.

Recently, for example, several leaders of Golden Dawn were found guilty of numerous criminal

charges in Greece, including attempted murder for the violent assault of a group of Egyptian

ﬁshermen in the port of Piraeus. Xenophobia to this end has by itself been characterized as

completely averse to liberal democracy (Betz 2005).

Along with this, radical-right parties have historically proven perfectly capable of

74

deteriorating the basic rights of certain segments of society, arguing such action to be con-

gruent with the general-will of the majority. In other words, claiming the restriction of rights

of minorities and other ’out-groups’ are democratic as doing so lines up with the desire of

the majority (see for e.g. Mudde 2013a). For instance, changes made in Hungary in 2013

by radical-right Fidesz limited freedoms of speech, religion, association, and voting, among

other basic rights, for many of what they perceive as ’out-groups’ (Urbinati 2019), describing

these moves as a “natural” use of the “authority it received in democratic elections.”5

3.3 Radical-Right Impact on Liberal Democracy

Given what we already know, that (1) radical-right parties are able to have real eﬀects

while in government, and (2) they have a troubled (to say the least) relationship with liberal

democracy, what do we know about the real-world impacts of radical-right parties regarding

liberal democracy speciﬁcally? Indeed, many have argued that their eﬀects are, above all,

“limited” (see for e.g. De Lange and Akkerman 2012; Mudde and Kaltwasser 2012), largely

an artifact of their limited electoral success and tendency to, when successful, serve only as

junior coalition members (Fallend 2012). On its face, this may seem promising. In 2013,

Mudde (2013b) argued the observed impacts of the radical-right were hopeful, as no country

had fallen to “illiberal democracy” due to their inﬂuence. There are however countries

certainly moving in that direction today, with Hungary and Poland among the most clear

examples (Cianetti, Dawson, and Hanley 2018; Huber and Schimpf 2017; Moder 2019), where

radical-right rule has led to an undermining of checks and balances and a dramatic increase

in the power of the executive.

It is clear, at least in Europe, that the potential for a broad relationship between

radical-right rule and the slow underpinning of liberal democracy exists in more widespread

fashion than previous literature has let on. Beyond the aforementioned cases of Hungary

5Gergely Gulyas, Fidesz deputy leader, quoted in the BBC.

75

and Poland, even as a radical-right party is constrained to junior-partner status in a ruling

coalition, it is quite likely we should see small, yet persistent, changes in the many metrics

of liberal democracy available. Even-still, no large-scale disaggregated evidence to this point

exists. Rather, most evidence remains quite anecdotal and limited in nature. It is clear,

at least, that the fear such a relationship exists is certainly real. When the radical-right

Freedom Party of Austria entered into government coalition in 1999, the European Union

initiated sanctions against the country, while the rest of EU member-states unanimously

cutoﬀ communication (Meret 2010). This, of course, occurred even as the Freedom Party

was merely a junior coalition partner.

And yet, virtually all studies focusing on the impacts to democracy of these parties

have focused exclusively on the fairly rare case of a majority radical-right government (or,

the case where the head-of-government is of the radical-right). Yet, as has become apparent,

both the minority coalition partner and the minister given the relevant cabinet portfolio

have inﬂuence over policy enactment and agenda-setting (Laver and Shepsle 1996; Martin

and Vanberg 2011). We have seen this very reality play out in Austria, as mentioned with

regard to immigration and asylum law in the case of the BZ ¨O rule of the Ministry of Justice.

Interestingly, it seems the Ministry of Justice is one of the most common ministries for a

radical-right junior coalition member to be given, second only to Defense (Bichay 2021).

Thus, it is possible we may see important impacts on liberal democracy not only when the

radical-right hold executive power, but also as it is a part of government (i.e. a member of

the ruling coalition). Kenny (2020), for example, ﬁnds persistent negative eﬀects of populists

as minority coalition partners in terms of press freedom, censorship and bias, harassment of

journalists, and freedom of expression in Asia and Latin America. I consider both possibilities

in the succeeding empirical tests.

That being said, it is certain that radical-right prime ministers and radical-right

coalition partners would exert diﬀering levels of inﬂuence on changes in democracy. Thus, I

break the potential eﬀects these parties may have into two components: institutional rule-of-

76

law, and civil liberty. This disambiguation is a logical extension of the premise that liberal

democracy is by its nature institutional, or minimalist, democracy, with the added compo-

nents of acknowledgment and respect for crosscutting cleavages, toleration for conﬂict and

opposition, equal protection under the law for all, and a protection of minority rights (see for

e.g. Pappas 2019). This disambiguation further echos work such as Kaltwasser (2012), who

adopts Dahl’s (1971) bifurcation of democracy into contestation and inclusiveness, noting

populism may harm the former while aiding the latter. This work, of course, diﬀers in that

it ﬁrst focuses on populism writ-large, and not radical-right parties generally, and second

considers only their introduction into the political landscape, and not the case of them ris-

ing to a position of political rule. While these categories of rule-of-law and civil liberties

that I propose are not a perfect match for Dahl’s categorization, they do provide a better

empirical avenue for testing the proposed mechanisms in that (1) they oﬀer categories of

potential change that more closely align with the sorts of declines in democracy that have

been experienced in Europe recently, and (2) provide, as I discuss below, categories that

represent changes a party would be interested in implementing only while holding executive

power in the ﬁrst domain, or would have an interest in altering even if simply conﬁned to

junior coalition status in the second.

Perhaps more straightforward of these domains is the institutional rule-of-law.

In

this case, the theoretical expectation is that radical-right parties bring with them decreases

in the level of constraints on executive power. As anti-pluralists, radical-right politicians

increasingly view their rule as the only avenue of success for the country. As the “true

voice” of the “true people” they fully advocate to be the sole method to enact this true

will into policy ends. Such was the theme of Viktor Orban’s now infamous speech on a new

“illiberal democracy,” where he enumerated policy goals that had not been met by previous

liberal democratic regimes while noting to Hungarians that “a trending topic in thinking

is understanding systems that are not Western, not liberal, not liberal democracies, maybe

not even democracies.” He rounds out the speech by referring to the principle of all citizens

77

being awarding an equal vote in elections as an “absurd situation” and delivering a utopian

image of what could be possible under his illiberal rule.

In this case, however, I only expect to see eﬀects when the actual oﬃce of prime

minister is held by the radical-right. For one, these are clearly more large-scale changes

requiring much more political capital, not to mention simply more political power, in order

to enact. It is unlikely, as others have argued (e.g. Mudde 2013b), to visualize a way for

a junior coalition partner to posture themselves into a position of such inﬂuence. Further,

it is unclear in the ﬁrst place why a radical-right party would have any interest in eroding

constraints on executive power at a time when it does not hold executive power. Such actions

would only seem to further cut them oﬀ from their desired policy platform.

Theoretically, this implies that as radical-right prime ministers win power, we should

see persistent negative eﬀects to the fairness of elections, as well as more direct judicial and

legislative constraints to their rule. Erdogan in Turkey, Orban in Hungary, and the Law and

Justice party in Poland are the most visible examples of this (e.g. Kaufman and Haggard

2019; Sadurski 2018). If this is truly a common theme of the radical-right rule at-large, we

should expect to see gradual changes over time in the aggregate in cases where they come

to power. Thus, I ﬁrst test the following hypotheses:

HYPOTHESIS 1: Radical-right prime ministers are negatively

associated with the freedom and fairness of elections.

HYPOTHESIS 2: Radical-right prime ministers are negatively

associated with legislative and judicial constraints on the exec-

utive.

Further, many note the radical-right’s tendency to despise and distrust the media

(Ellinas 2018). Donald Trump made famous the phrase “fake news” with regard to main-

stream journalists, calling them the “enemy of the people.” In general, the radical-right

78

tends to hold the belief that the mainstream media is dishonest, corrupt, elitist, and uncon-

cerned with the common people (see for e.g. Bhat and Chadha 2020; Figenschou and Ihlebæk

2019). This has, in many cases led to them creating their own media outlets as a source

of alternative, more favorable information. In places where they come to power, it follows

that we may see a slow undermining of press freedom, and an increase in media censorship

and bias (as media networks are pressured to support the regime, and alternative networks

are created for the purpose of supporting the incumbent government). Thus, I next test the

following hypothesis:

HYPOTHESIS 3: Radical-right prime ministers are negatively

associated with freedom of the press.

Beyond the radical-right’s eﬀects to the institutional rule-of-law, are their impacts

related to civil liberties. Anecdotally, many case studies have noted repressive policies pur-

sued and enacted by radical-right parties in power (e.g. Minkenberg 2001; Urbinati 2019).

Examples include clamping down on immigration (Zaslove 2004a), eliminating rights of asy-

lum seekers (Fallend 2012), ethnocentrism (Kallis 2018), and attacks on freedom of speech,

expression, and free association (Urbinati 2019). In many cases, these widespread changes

were pursued even as radical-right parties were simply coalition partners. Whereas the de-

terioration of rule-of-law can be described as eliminating institutional constraints to power,

the deterioration of civil liberties can best be summed up as both the elimination of mass

constraints to power, and an eﬀort to return society to its traditional form—that is, one

made up of only the native population with high levels of law and order (Mudde 2007).

Mass constraints, unlike institutional constraints, are equally salient whether inside or out-

side of sole executive power. Further, these changes aiming to return to a traditional society

do not represent creating an avenue for enacting the radical-right’s policy platform, rather,

these changes are the radical-right’s policy platform. Thus it follows these eﬀects should

exist even as the radical-right party is merely a coalition partner.

79

In this way it follows that, in the aggregate as radical-right parties rise to power we

should see persistent decreases in civil liberties and increases in the deterioration of freedoms

like expression, speech, and intellectual freedom as the radical-right tries to silence those who

would do them harm and implement its vision for traditional society. At the same time, we

should expect to see crackdowns on “out-group” populations that are not part of this vision.

These eﬀects, unlike institutional constraints, should exist also in cases with a radical-right

party in coalition, even if they do not hold the prime ministership. Thus, I also test the

following hypotheses:

HYPOTHESIS 4: Radical-right prime ministers and coalition

partners are negatively associated with the civil liberties of the

population.

HYPOTHESIS 5: Radical-right prime ministers and coalition

partners are positively associated with repression of societal

“out-groups.”

Why should we expect such drastic measures from the radical-right? As Mudde (2004)

notes regarding populists in general, their “opponents are not just people with diﬀerent

priorities and values, they are evil!” The radical-right does not view opponents as simply

people with diﬀering beliefs, but rather as part of the problem. They view themselves as

the only path forward for the country, and have increasingly been willing to adopt radical

means to ensure this path takes hold. Beyond this, these hypotheses are based on the case

knowledge we have from radical-right’s time in government. These expectations are not

simply based on their populist rhetoric—but on the real outcomes observed when they win

elections. Unfortunately, as of now, that limited case knowledge is all we have. In the next

section I outline a large-scale quantitative analysis of 30 European countries over time to (1)

demonstrate these eﬀects are indeed widespread and common to the entire region, and (2)

disentangle the various ways democracy may deteriorate when they come to power.

80

3.4 Data and Modeling Strategy

I utilize deﬁnitions of radical-right parties from Rooduijn et al. (2019). This classi-

ﬁcation itself is taken from Mudde (2007), who argues radical-right parties are those that

believe both (1) the nation should be made up of only the native ethnic-group, and (2) strong

punishment must exist for violations to law and order. The sample includes 31 European

countries since 1989, and has been peer-reviewed by over 80 country experts.

I code all

governing coalitions in these 31 countries since 1989 for whether they had (1) a radical-right

prime minister, and (2) a radical-right party in the governing coalition. This leaves 851

country-year observations, of which 20 have a radical-right prime minister, and 107 have at

least a radical-right party in coalition government.

I utilize several measures from the Varieties of Democracy (V-Dem) dataset to mea-

sure changes in liberal democracy (for robustness, I utilize measures from alternative sources

in the appendix and results do not change). To test hypothesis one (freedom and fairness

of elections) I draw on V-Dem’s “clean elections index” which itself is based on a bayesian

factor analysis of the following components: election monitoring body’s autonomy and ca-

pacity, quality of voter registry, prevalence of vote buying, voting irregularities, government

intimidation, electoral violence, as well as the expert’s overall opinion of the freedom and

fairness of the national election. It is measured on an interval scale that ranges from 0–1,

with higher numbers indicating fairer elections.

To test hypotheses two, I use V-Dem’s measures of legislative and judicial constraints

on the executive. The legislative constraints variable is comprised of the following mea-

sures: (1) the legislature’s ability to question the executive branch, (2) the likelihood of

an oﬃcial investigation by the legislature into unethical/unconstitutional executive conduct,

(3) the likelihood of an oﬃcial investigation by someone other than the legislature into un-

ethical/unconstitutional executive conduct, and (4) the ability of the opposition to provide

government oversight and investigations. Meanwhile, the measure of judicial constraints in-

81

cludes: (1) the degree to which the executive respects the constitution, (2) the executive’s

compliance with court rulings, and (3) the independence of the court’s decision making.

These measures are also based on bayesian factor analysis and are scaled on a 0–1 interval,

where higher numbers indicate more independence.

No appropriate index exists to test the hypothesis related to freedom of the press,

thus I rely on several variables. My hypothesis posits that, ﬁrst, the government will view the

media as a constraint to its continued rule, and will thus attempt to increase censorship of

the media. To test this, I utilize a question from V-Dem which asks country experts to what

extent the government “directly or indirectly attempt[s] to censor the print or broadcast

media.” As journalists are censored and pressured by the government, and more friendly

government media sources are created, we should also expect to see an outright tilt in

coverage more friendly to the incumbent regime. Thus I also include as DVs measures of the

media’s bias against opposition candidates, and the number of media outlets that routinely

criticize the government. These three media variables are all measured on an ordinal scale

by at least ﬁve country-experts, before being converted to an interval level measure via a

Bayesian item response theory model (see Pemstein et al. 2015). In all cases, higher values

refer to more press freedom.

Next, I introduce the variables used to test the civil liberties hypothesis. While V-

Dem does compute an individual liberty index, which includes important measures such as

equality before the law and freedom of religion, it also includes such measures as political

torture and killing. While the former are very likely to be correlated with radical-right rule,

the latter almost certainly are not (though I note that radical-right rule is indeed a signiﬁcant

indicator of this index, even-still). Thus I instead opt to use the raw measures themselves.

To test for a deterioration of freedom of speech, I utilize V-Dem’s measure of “freedom of

discussion,” which measures the extent to which men are able to engage in private or public

discussion of politics (they oﬀer the same measure for women separately, but results are

identical for either measure).

82

Along with this, academic freedom is also commonly seen as a threat to radical

parties, as universities traditionally, like the media, provide oversight on the mechanics of

government. Take for example, the Hungarian government’s moves against Central European

University in 2017 (Enyedi 2018). Thus I include a measure of academic freedom in my tests

of civil liberties. As radical-right parties tend to be islamophobic and ethnocentric, I also test

for their eﬀect on religious freedom and ability to practice without discrimination. Finally,

I test for eﬀects on access to justice. This index measures the ability to which men are

able to seek fair legal recourse if the government violates their rights, without risking their

own safety or rights (again, the same variable exists for women separately, but results are

identical). All of these measures utilize the same Bayesian item response theory modeling

discussed above, and in each case higher values indicate higher levels of civil liberty.

Lastly I test hypothesis 5, which argues societal “out-groups” will be repressed by

radical-right governments. To test this I again draw on V-Dem and utilize their “equal

protection index.” This Bayesian factor analysis index measures the extent to which cer-

tain groups of society enjoy equal rights and freedoms, are allowed to participate, and are

not threatened by other societal groups. It incorporates measures of the diﬀerence in civil

liberties between the rich and the poor as well as diﬀering ethnic groups.

I include a standard host of controls in all models. To control for prevailing economic

conditions I include a log of GDP per capita.

I also include Gini coeﬃcient to measure

inequality within the country, as inequality has been shown to inﬂuence radical-right success

(Han 2016), and may well an eﬀect on civil liberties, certainly those pertaining to the repres-

sion of out-groups. I include a control for parliamentary systems, as leadership and coalition

dynamics diﬀer in presidential and semi-presidential systems.

Indeed Cyprus is the only

presidential system in the sample, and has never experienced a radical-right party in coali-

tion. Yet, this variable—as deﬁned by the Database of Political Institutions (Scartascini,

Cruz, and Keefer 2018)—also includes semi-presidential systems, which may similarly have

divergent baselines in terms of radical success.

83

I also control for whether the country can be classiﬁed as an autocracy (deﬁned as

those with a Polity score of less than seven), as the baseline levels regarding repression, rule

of law, and civil liberties are clearly very diﬀerent in an autocracy. Finally, given the myriad

historical, cultural, and institutional factors that have been shown to eﬀect radical-right

success I opt to include country ﬁxed-eﬀects, allowing instead an analysis of within-country

changes over time.

All models are conceptualized as lagged dependent-variable regressions, as level of

democracy within a country is clearly a dynamic process. Further, it is quite possible na-

tions with lower rule-of-law or civil liberties are more susceptible to the rise of the radical-

right in the ﬁrst place. This changes our interpretation of the results from the eﬀect of

the radical-right on level of democracy, to their eﬀect on changes in democracy, arguably

more theoretically relevant, as well. I note there is a strong possibility of introducing bias

to a model containing both ﬁxed-eﬀects and a lagged-dependent variable (Nickell 1981). In

this case, however, the bias is towards more conservative estimation. In other words, this

methodology leads to a quite cautious test of the potential eﬀects on my dependent vari-

able with respect to my independent variables. This makes the potential ﬁnding of strong

eﬀects particularly interesting in this regard. For robustness I respecify all main models as

random-eﬀects models instead, that is with coalitions nested in countries. The results remain

materially unchanged, and if anything, show even stronger eﬀects–especially in the case of

radical-right junior coalitions. I also specify models that utilize ﬁxed-eﬀects with no lagged

dependent variable, here too we see stronger eﬀects. However, as I view the ﬁxed-eﬀects

lagged dependent variable speciﬁcation as the more methodologically appropriate choice, I

continue with this strategy.

I ﬁrst test models where the main independent variable of interest is a dichotomous

measure of whether executive power was held by a radical-right party (that is, a radical-

right prime minister). Next, following the theory that radical-right parties are able to exert

inﬂuence as mere partners in government, I run models on the same set of dependent variables

84

utilizing instead an independent variable measuring the percent of coalition seats held by the

radical-right party, as a percent of total coalition seat-share, see equation 3.1. All models

include robust standard errors clustered by country.

Radical-Right Government Seat-Share =

Sr

Sr + Sg

(3.1)

Where S is seat-share, r is the radical-right coalition partner, and g are all other coalition parties.

I utilize this seat-share measure rather than a simple dichotomous measure of whether

the radical-right party is in coalition as their inﬂuence within government grows as they make

up a larger share of government seats, and control more government ministries (Browne and

Franklin 1973; Gamson 1961).

It would be problematic, for example, to equate Greece’s

2009 cabinet where the radical-right Popular Orthodox Rally held just over 5% seat-share

and one cabinet post with the 2017 Austrian coalition where the Freedom Party of Austria

won 45% of the government seat-share and several cabinet posts.

3.5 Results

I ﬁrst discuss results related to the eﬀect of a radical-right prime minister on insti-

tutional rule-of-law. These results are presented in table 3.1. Columns 1–3 present results

on the eﬀect on free and fair elections, as well as judicial and legislative constraints on the

executive. Interestingly, it appears that while radical-right rule signiﬁcantly deteriorates leg-

islative and judicial rule on executive constraints, it has no discernible eﬀect on the freedom

and fairness of elections. This may intuitively make sense. Much of the radical-rights’ pop-

ulist rhetoric focuses on taking power away from the corrupt elite and rightfully returning

it to the masses (Muller 2016). In this way, it is not necessarily that they openly advocate

removing people from the decision making process. Rather, their anti-democratic eﬀorts

seemed to be more concerned with removing constraints to their rule, once democratically

85

Table 3.1: Eﬀect of a Radical-Right Prime Minister on Institutional Rule-of-Law

Free/Fair
Elections Constraints Constraints Censorship

Legislative

Judicial

Media

Media Critical Media Bias
Against Opp

of Govt

(1)

Radical-Right PM −0.033
(0.025)

(2)

−0.036∗∗∗
(0.009)

log(GDP/Capita)

Autocracy

Gini

Parliamentary

Lagged DV

Constant

Observations
R2
Adjusted R2

0.009
(0.007)
−0.029
(0.023)

0.002
(0.001)

0.027
(0.022)

0.235
(0.152)
0.552∗∗∗
(0.109)

738
0.933
0.930

0.002
(0.003)

0.023
(0.022)

0.002
(0.001)
−0.012
(0.008)
0.764∗∗∗
(0.104)

0.149
(0.118)

738
0.950
0.947

(3)

−0.049∗
(0.024)
0.018∗∗
(0.006)

0.045
(0.027)

0.002
(0.002)
−0.023∗∗
(0.009)
0.501∗∗∗
(0.105)
0.216∗
(0.095)

738
0.929
0.925

(4)

−0.666∗∗∗
(0.127)
−0.024
(0.042)

0.025
(0.192)

0.015
(0.012)
−0.064
(0.065)
0.690∗∗∗
(0.083)

0.723
(0.451)

738
0.906
0.901

(5)

−0.388∗∗∗
(0.031)
−0.032
(0.035)
−0.019
(0.032)

0.008
(0.010)
−0.114∗∗
(0.042)
0.766∗∗∗
(0.076)

0.750
(0.384)

738
0.908
0.904

(6)

−0.285∗∗
(0.104)
−0.073
(0.038)

0.044
(0.123)

0.006
(0.009)
−0.113
(0.082)
0.665∗∗∗
(0.058)
1.398∗∗∗
(0.422)

738
0.832
0.824

Standard errors are clustered by country and appear in parentheses. Country ﬁxed-eﬀects omitted
for presentation. ∗p<0.05; ∗∗p<0.01; ∗∗∗p<0.001

Table 3.2: Predicted Values of Institutional Rule-of-Law

Non-Radical-Right Radical-Right

1-Year Diﬀerence

Free/Fair Elections
Judicial Contraints
Legislative Contraints
Media Censorship
Media Critical of Govt
Media Bias Against Opposition

0.93
0.91
0.87
2.54
2.27
2.10

0.90
0.87
0.82
1.87
1.88
1.81

-0.03
-0.04
-0.05
-0.67
-0.39
-0.29

4-Year Diﬀerence

-0.13
-0.14
-0.19
-2.66
-1.55
-1.14

elected. Meanwhile, columns 4–6 display results of radical-right prime minsters on press free-

dom. Here too we see persistent negative eﬀects associated with radical-right rule. We see

signiﬁcant increases in media censorship, as well as decreases in critiques of the incumbent

government, and a concurrent increase in anti-opposition bias.

86

To calculate the substantive eﬀect of these measures, I estimate their predicted values

for both radical-right and non-radical-right prime ministers. They are displayed in table 3.2.

All other variables are held at their mean or median value, thus as simulations of democratic

quality for both radical-right and non-radical-right countries begin at the same value, and

the previous year’s respective level is controlled for, these simulations show the diﬀerence

after a single year of radical-right rule (see for e.g. Houle and Kenny 2018). In the ﬁnal

column of table 3.2 I present the estimated change after a four-year term of radical-right

rule. In every case, four years of radical-right rule is associated with at least a one standard

deviation decrease in institutional rule-of law; in the case of press freedom the eﬀects are

larger than two standard deviations.

I turn next to analyzing the eﬀect of radical-right prime ministers on civil liberties

and repression. The results of these models is displayed in table 3.3. As before, radical-right

prime ministers have a signiﬁcant negative eﬀect while in power. Each year the radical-right

holds executive power sees an associated decrease in freedom of speech, religion, academia,

access to justice, and a decrease in equal protection and civil liberties of societal “out-groups.”

While the eﬀect on speech does not appear signiﬁcant at traditional levels, it does maintain

p = 0.051. As before, in table 3.4 I present simulated eﬀects of annual and four-year changes

in these measures when the radical-right is in power.

The results here are more varied. A four-year term of radical-right rule is associated

with a larger than one standard deviation decrease in both freedom of speech and religion, a

two standard deviation decrease in academic freedom, a roughly one-third standard deviation

decrease in access to justice, and a decrease in equal protection of out-groups of roughly two-

thirds of a standard deviation.

Overall, it seems radical-right prime minister’s have the starkest eﬀect on judicial and

legislative constraints on their power, freedom of the press, and academic freedom. These

variables, likely no coincidence, represent the “elite” of the country. The rest of government,

the media, and academics are those most likely to criticize radical-right candidates, and

87

Table 3.3: Eﬀect of a Radical-Right Prime Minister on Civil Liberties

Freedom of Religious Academic Access to Equal Protection
for Out-Groups

Freedom Freedom

Justice

Speech

Radical-Right PM

log(GDP/Capita)

Autocracy

Gini

Parliamentary

Lagged DV

Constant

Observations
R2
Adjusted R2

(1)

−0.156
(0.080)

0.008
(0.044)

0.139
(0.168)
−0.014
(0.011)
−0.088
(0.047)
0.781∗∗∗
(0.041)
0.956∗∗
(0.294)

738
0.889
0.884

(2)

(3)

(4)

0.090
(0.135)

(0.149)
−0.030
(0.041)

−0.166∗∗ −0.386∗∗ −0.060∗∗∗
(0.016)
(0.056)
−0.027
−0.039
(0.022)
(0.028)
−0.070
−0.019
(0.053)
(0.145)
−0.011
−0.001
(0.006)
(0.006)
−0.180
−0.033
(0.123)
(0.045)
0.786∗∗∗
0.832∗∗∗
(0.062)
(0.041)
1.146∗∗∗
1.046∗∗
(0.318)
(0.221)

0.007
(0.010)
−0.130
(0.097)
0.774∗∗∗
(0.043)
0.862∗∗
(0.294)

738
0.919
0.915

738
0.899
0.894

738
0.970
0.969

(5)

−0.009∗∗
(0.004)
−0.001
(0.002)

0.005
(0.005)

0.0002
(0.001)
−0.007
(0.004)
0.807∗∗∗
(0.040)
0.202∗∗∗
(0.045)

738
0.962
0.960

Standard errors are clustered by country and appear in parentheses. Country
ﬁxed-eﬀects omitted for presentation. ∗p<0.05; ∗∗p<0.01; ∗∗∗p<0.001

Table 3.4: Predicted Values of Civil Liberties
1-Year Diﬀerence

Non-Radical-Right Radical-Right

4-Year Diﬀerence

Speech
Religion
Academic
Justice
Equal Protection

2.54
1.93
2.64
2.70
0.92

2.39
1.76
2.26
2.64
0.91

-0.16
-0.17
-0.39
-0.06
-0.01

-0.62
-0.67
-1.55
-0.24
-0.04

those most likely to stand as barriers to their continued power. It has been noted many

times over in the literature that the radical-right, and populist rhetoric at-large, focuses on

the need to reclaim power from the corrupt elite, and return it to the “rightful majority”

(Muller 2016). Thus, while we do see smaller, persistent negative eﬀects on the civil liberties

of those they do not consider to be the “rightful majority,” the eﬀects on those who they

88

view as the “corrupt elite” appear to be much more pronounced.

3.5.1 Eﬀects of Radical-Right Coalitions

In this section, I run models on the same dependent variables as above, this time

looking at the eﬀect of radical-right parties in coalition government. Given their ability to

aﬀect change on policies while merely in a government coalition, it follows that they may

indeed have profound eﬀects on the quality of liberal democracy, as well. However, due to

both practical constraints, and the will and preferences of the party, I argue we will only

see real changes for mass civil liberties, and not institutional constraints. The dependent

variable in these cases, again, is the share of government seats held by the radical-right party.

Crucially, in these models I add a control for whether the radical party holds the prime

ministership, as the argument suggests they should eﬀects negative eﬀects to democracy

even when outside of executive power. The results of radical-right coalition on institutional

rule of law are presented in table 3.5. As expected, the results do not hold. In general, it

appears the radical-right is unable to aﬀect institutional rule-of-law of a country when they

remain relegated to junior coalition-partner status. Only media bias seems to suﬀer from

radical-right parties in coalition.

Indeed, this may make sense regarding constraints to power. For one, these are largely

more large-scale changes requiring much more executive power to enact. It is unlikely a junior

partner of any kind would be able to have such inﬂuence. Further, it is entirely unclear why

a radical-right party would have any interest in eroding constraints on executive power at a

time when it does not hold said executive power. Thus, the only real result we see during

their time in coalition is a signiﬁcant increase of media bias against political rivals, with a

four-year change larger than one standard deviation. This intuitively could make sense as

well, as while in oﬃce radical-right politicians may wield more inﬂuence in stopping negative

stories while also having more incentive and resources to set up favorable media outlets of

their own.

Finally, I present results on the eﬀect of radical-right coalitions on civil liberties. The

89

Table 3.5: Eﬀect of a Radical-Right Coalition on Institutional Rule-of-Law

Free/Fair
Elections Constraints Constraints Censorship

Legislative

Judicial

Media

Media Critical Media Bias
Against Opp

of Govt

Rad-Right Govt Seat

Radical-Right PM

log(GDP/Capita)

Autocracy

Gini

Parliamentary

Lagged DV

Constant

Observations
R2
Adjusted R2

(1)

−0.039
(0.034)
−0.013
(0.021)

0.010
(0.007)
−0.029
(0.023)

0.001
(0.001)

0.027
(0.023)

0.234
(0.152)
0.550∗∗∗
(0.109)

738
0.934
0.930

(2)

−0.019
(0.018)
−0.026
(0.014)

0.002
(0.003)

0.023
(0.022)

0.002
(0.001)
−0.012
(0.008)
0.761∗∗∗
(0.105)

0.149
(0.119)

738
0.950
0.947

(3)

−0.062
(0.043)
−0.017
(0.024)
0.020∗∗
(0.006)

0.045
(0.027)

0.002
(0.002)
−0.023∗
(0.009)
0.496∗∗∗
(0.104)
0.214∗
(0.096)

738
0.930
0.926

(4)

−0.535
(0.372)
−0.396∗
(0.180)
−0.012
(0.043)

0.025
(0.189)

0.013
(0.012)
−0.065
(0.065)
0.686∗∗∗
(0.083)

0.681
(0.465)

738
0.907
0.902

(5)

−0.101
(0.147)
−0.337∗∗∗
(0.089)
−0.030
(0.036)
−0.019
(0.032)

0.007
(0.010)
−0.114∗∗
(0.043)
0.765∗∗∗
(0.076)

0.742
(0.389)

738
0.908
0.904

(6)

−0.570∗
(0.286)

0.005
(0.141)
−0.060
(0.038)

0.046
(0.122)

0.003
(0.009)
−0.115
(0.090)
0.664∗∗∗
(0.057)
1.354∗∗
(0.426)

738
0.834
0.825

Standard errors are clustered by country and appear in parentheses. Country ﬁxed-eﬀects omitted for
presentation. ∗p<0.05; ∗∗p<0.01; ∗∗∗p<0.001

results of these models are displayed in table 3.6. This time, as with radical prime ministers,

we see signiﬁcant negative eﬀects. Aside from equal protection of social-groups, radical-right

parties in coalition are associated with signiﬁcant decreases to mass civil rights, even as the

party of the prime minister is controlled for. Table 3.7 presents the predicted values and

changes for these metrics. In this case, I analyze the ﬁrst-diﬀerence change from no radical-

right party in cabinet, to the case of a radical-right party with 30% of the government

seat-share (note, this implies their overall parliamentary seat-share is still much lower). A

four-year term of radical-right coalition is associated with nearly a 3/4 standard deviation

90

Table 3.6: Eﬀect of a Radical-Right Coalition on Civil Liberties

Freedom of Religious Academic Access to Equal Protection
for Out-Groups

Freedom Freedom

Speech

Justice

Rad-Right Govt Seat

Radical-Right PM

log(GDP/Capita)

Autocracy

Gini

Parliamentary

Lagged DV

Constant

Observations
R2
Adjusted R2

(1)

−0.290∗
(0.121)
−0.010
(0.104)

0.015
(0.044)

0.137
(0.169)
−0.015
(0.011)
−0.090∗
(0.043)
0.777∗∗∗
(0.040)
0.942∗∗
(0.299)

738
0.890
0.884

(2)

−0.373∗
(0.151)

0.020
(0.073)
−0.030
(0.027)
−0.077
(0.145)
−0.013∗
(0.005)
−0.031
(0.045)
0.774∗∗∗
(0.045)
1.134∗∗∗
(0.216)

(3)

−0.585∗
(0.229)
−0.094
(0.161)
−0.015
(0.038)

0.091
(0.134)

0.005
(0.010)
−0.128
(0.097)
0.765∗∗∗
(0.044)
0.810∗∗
(0.276)

(4)

−0.200∗
(0.100)

0.042
(0.052)
−0.023
(0.022)
−0.019
(0.052)
−0.002
(0.006)
−0.180
(0.122)
0.830∗∗∗
(0.062)
1.035∗∗
(0.323)

738
0.920
0.916

738
0.900
0.895

738
0.970
0.969

(5)

−0.003
(0.007)
−0.008∗
(0.003)
−0.001
(0.002)

0.005
(0.005)

0.0002
(0.001)
−0.007
(0.004)
0.807∗∗∗
(0.040)
0.202∗∗∗
(0.045)

738
0.962
0.960

Standard errors are clustered by country and appear in parentheses. Country ﬁxed-
eﬀects omitted for presentation. ∗p<0.05; ∗∗p<0.01; ∗∗∗p<0.001

Table 3.7: Predicted Values of Civil Liberties
1-Year Diﬀerence

Non-Radical-Right Radical-Right

4-Year Diﬀerence

Speech
Religion
Academic
Justice
Equal Protection

2.55
1.94
2.66
2.71
0.92

2.47
1.83
2.49
2.65
0.92

-0.09
-0.11
-0.18
-0.06
-0.00

-0.35
-0.45
-0.70
-0.24
-0.00

decrease in freedom of speech, and a roughly one standard deviation decrease in freedom of

of religion and academic freedom. Meanwhile, access to justice sees a 1/3 standard deviation

decrease, while there again is no eﬀect on equal protections.

While these changes are, naturally, smaller than the eﬀects exhibited during the term

91

of a radical-right prime minister, it is consequential to note the stark eﬀects these parties

are able to have by merely joining a government coalition. It is not the case, as previously

thought, that the eﬀect of the radical-right on liberal democracy is limited by their relegation

as junior members. Rather, they are able to wield their inﬂuence and control of government

ministries to bring with them real deterioration to the level of liberal democracy within their

country.

Thus, in sum, we see strong evidence for hypotheses 2–3. While the freedom-and

fairness of elections does not appear to be hurt by the radical-right’s ascension to prime

minister (hypothesis 1), we do see important deterioration of the constraints to their power,

namely through harms to judicial and legislative checks on power, as well as press freedom.

We also see strong support for hypothesis 4. While the radical-right in coalition government

is not strong enough to alter these institutional constraints, they do seem to have signiﬁcant

eﬀects on civil liberties in the country. This is likely due to their increased inﬂuence as

a cabinet party, and newfound authority in leading certain ministries. Finally, hypothesis

5 receives mixed support. While radical-right prime ministers do seem to harm the equal

protection of societal “out-groups,” mere radical-right coalitions do not.

3.5.2 Robustness Checks

I conduct several tests to ensure results are robust to various speciﬁcations. First,

there may be concerns as to the utilization of V-Dem’s indicators, which rely on the responses

of country-experts. Though each case is reviewed by no less than ﬁve experts, and these

responses are then used in Bayesian IRT models to arrive at a robust country-year interval

level measure, I use alternative measures to ensure models are not sensitive to this speciﬁ-

cation. Namely, I ﬁrst utilize the World Bank’s Rule-of-Law estimate from their Worldwide

Governance dataset. This measure takes into account the degree to which the contract en-

forcement, property rights, the police, and the courts abide by the laws. Results remain

robust in terms of the eﬀect of radical-right prime ministers and coalition seat-share.

In

terms of civil liberties, I rely on Freedom House’s civil liberties measure. While this measure

92

includes metrics such as press freedom and independence of the judiciary (characteristics of

institutional rule-of-law, rather than civil liberties), many of the metrics do indeed refer to

freedoms of speech, assembly, expression, as well as academic freedom. In terms of both the

World Bank and Freedom House measures, results remain largely unchanged from the main

models. The results of these alternative dependent variable models are presented in table

B.1 of the appendix

Next, the extrapolation to calculate the estimated change in democracy for a four-

year term of radical-right rule in the preceding sections crucially relies on the assumption

that the eﬀect of a radical-right government is linear and constant over time. I directly test

this by rerunning the preceding models with an alternative dependent variable. Instead of

a dichotomous measure for whether a radical-right prime minister is in oﬃce, I instead use

a measure of the number of years the radical-right prime minister has been in oﬃce (i.e.

a 1 their ﬁrst year, 2 their second year, and so on). I include a square of this variable to

allow the possibility of nonlinear eﬀects. Importantly, these models do not include lagged

dependent variables. The results are displayed in ﬁgure B.1 in the appendix. I conﬁrm that

(1) the results do seem to be linear over time, and (2) the calculated eﬀects of a four-year

term roughly match those presented above.

Additionally, there may well be concerns that, due to the low number of countries

that do experience a radical-right prime minister, results may be driven by one or a few

observations. I note ﬁrst that while only a handful of cases (20 country-year observations)

do experience radical-right executive rule, it is not simply the case of periods of uninterrupted

rule. Rather it is common for periods of radical-right rule, followed by periods of moderate

rule, followed again by another period of radical rule. This is beneﬁcial in identifying a

model seeking to analyze within-country changes over time. Even still, in the appendix I

provide two major robustness tests to ensure results are not driven by only one country.

First, I perform a clustered jackknife for each main model. That is, I re-run every

model 30 times, each time omitting one country. I plot the distribution of parameter esti-

93

mates of the radical-right prime minister variable in the appendix ﬁgure B.2 for institutional

rule-of-law models, and ﬁgure B.3 for civil liberties models. Overall, the results appear quite

stable. Only in the case of eﬀects to free and fair elections (which already was insigniﬁcant

in the main model) and equal protection under the law does a parameter from the jackknife

approach 0. In other words, radical-right prime minister is always a meaningful predictor

of level of democracy (with exception to the two aforementioned variables) in every case,

regardless of which country is dropped from the sample. Second, I present full model outputs

on both institutional rule-of-law and civil liberties in the cases where Poland and Hungary

are dropped, as these are the most prominent cases of radical-right executive control, and

make up the vast majority of positive observations. Results are displayed in tables B.2 – B.5

of the appendix and indicate results are not driven by one case.

Finally, I address concerns of endogeneity. That is, the possibility that some nations

were already on the path to a deterioration of democracy, and that this very path is in turn

what led them to radical-right rule. While this is addressed in part by the inclusion of a

lagged dependent variable, I adopt a second strategy here to ensure robustness.

I follow

Houle and Kenny (2018) and employ a quasi-experimental discontinuity design to test for

the diﬀerence in demcoracy where a radical-right candidate won, compared to the case where

they almost won. I adopt Houle and Kenny’s deﬁnition and deﬁne the case of almost winning

as those where the radical-right party received over 20% of vote-share in the election leading

to the coalition. While certainly not a quintessential regression discontinuity design, this

sheds some additional light on whether there is something unique about radical rule that

harms democracy, or rather if the eﬀects seen are simply symptoms of the greater systemic

conditions that may endogenously bring a higher probability of radical-right rule, as well.

Table 3.8 displays the results of the quasi-experiment with regard to institutional rule-of-law,

while table 3.9 presents the same with regard to civil liberties. For each measure, the case

where the radical-right candidate actually won is correlated with a lower level of democracy.

In each case this diﬀerence is signiﬁcant at the p < 0.001 level.

94

Table 3.8: Institutional Rule-of-Law, Radical-Right and Almost Radical-Right Regimes

Free/Fair
Elections Constraints Constraints Censorship

Legislative

Judicial

Media

Media Crit. Media Bias
Against Opp

of Govt

Almost Radical-Right
Radical-Right

0.95
0.88
-0.07∗∗∗
Diﬀerence
∗p<0.05, ∗∗p<0.01, ∗∗∗p<0.001.

0.95
0.83
-0.12∗∗∗

0.91
0.79
-0.13∗∗∗

2.97
1.13
-1.84∗∗∗

2.38
0.93
-1.45∗∗∗

2.35
1.05
-1.30∗∗∗

Table 3.9: Civil Liberties, Radical-Right and Almost Radical-Right Regimes

Freedom of

Speech

2.60
2.03
-0.56∗∗∗

Religious
Freedom

2.02
1.54
-0.48∗∗∗

Academic
Freedom

2.84
1.61
-1.23∗∗∗

Access to

Justice

2.87
1.82
-1.05∗∗∗

Equal Protection
for Out-Groups

0.94
0.87
-0.07∗∗∗

Almost Radical-Right
Radical-Right

Diﬀerence
∗p<0.05, ∗∗p<0.01, ∗∗∗p<0.001.

I further conduct this discontinuity test looking at cases where radical candidates

were excessively popular but did not join the government coalition compared to cases where

they did enter a coalition. These diﬀerences with regard to civil liberties are presented in

table 3.10. Again, we see signiﬁcant negative diﬀerences, indicating it indeed is the eﬀect of

their participation into government, and not endogenous conditions of the nation, driving

these results.

Table 3.10: Civil Liberties, Radical-Right and Almost Radical-Right Coalitions

Freedom of

Speech

2.60
2.48
-0.12∗

Religious
Freedom

2.18
1.79
-0.39∗∗∗

Academic
Freedom

2.90
2.42
-0.48∗∗∗

Access to

Justice

2.86
2.49
-0.38∗∗∗

Equal Protection
for Out-Groups

0.94
0.91
-0.03∗∗∗

Almost Radical-Right
Radical-Right

Diﬀerence
∗p<0.05, ∗∗p<0.01, ∗∗∗p<0.001.

3.6 Conclusion

Much discussion of the last decade has focused on the major threats posed by radical-

right parties and candidates. While many scholars consider them to be one of the largest

contemporary threats to liberal democracy, this threat has yet to be empirically quantiﬁed.

95

Much of our knowledge, until now, has relied on limited case studies in the few nations they

have risen to take control of executive power. Overall, these results paint a clear picture.

Radical-right parties do indeed exhibit negative eﬀects to liberal democracy when in oﬃce.

It is a persistent eﬀect we see throughout Europe over time; as a country sees a radical-right

party come to power, we can expect liberal democracy to suﬀer in several ways. With the

utilization of disaggregated measures, we are able to see exactly where these eﬀects are more

prevalent.

In this paper, I operationalize liberal democracy as existing in two distinct domains:

institutional rule-of-law, and mass civil liberties. This disaggregation allows for an approach

to separate the eﬀects of cases where the radical-right holds the prime ministership, as

compared to cases where they serve as junior coalition partners. I ﬁrst argue and ﬁnd that

radical-right parties only have real impacts to institutional rule-of-law when serving as prime

minister. This makes sense, as (1) these are generally large-scale changes that require a large

amount of power and inﬂuence to enact, and (2) it follows that candidates would only have

an interest in eliminating executive constraints when they hold executive power. Even still,

while prime minister the eﬀect the radical-right has is stark. Namely, while a radical-right

prime minister’s four-year term has no discernible eﬀect on electoral freedom, it is associated

with large decreases in legislative and judicial constraints on the executive, and even larger

decreases in terms of press freedom.

This intuitively makes sense. Much of the radical-right’s often populist rhetoric fo-

cuses on taking power away from the corrupt elite and rightfully returning it to the general

masses (Muller 2016). In this way, where they generally are considered “anti-democratic”

(e.g. Mudde 1996), it is not necessarily that they openly advocate removing people from

the decision making processes. Rather, their anti-democratic eﬀorts seem to be far more

concerned with removing constraints to their own political rule, once democratically elected.

After all, in order to be in a position to aﬀect change, they must ﬁrst operate within the

status quo of democracy. As Muller (2016) argues, when these parties demonstrate values

96

counter to democracy, the people simply believe this is being done for their own beneﬁt

(them being their voters, the only “true people”). Perhaps it makes sense, then, that the

bulk of the radical-right’s impact on liberal democracy occurs in the realm of disrespecting

constraints to their power, as these constraints inherently exist only in a world where the

radical-right is already in power. This also explains the lack of eﬀect in terms of radical

coalitions on these institutional constraints.

In terms of mass civil liberties, however, the calculus is diﬀerent. Rather, these

represent true policy preferences of the radical-right, whether in a position of executive

power or not. In one sense they are changes that bring the society more in line with their

vision of morality. In the other, they represent constraints to their power that exist even

when out of power, in terms of those who would criticize them and provide a “check” on

their behavior. These also represent changes that are, in a sense, “easier” to make, and do

not require the level of power that comes with outright executive control. While a junior

coalition member, the BZ ¨O is Austria was able to make widespread changes to the justice

and immigration system, for example. Thus, it follows that we should see a deterioration of

mass civil liberties both when the radical-right holds the prime ministership, but also (if to

a lesser degree) when they serve as mere coalition partners.

In a disconcerting way, this is exactly what I ﬁnd. Radical-right parties are corre-

lated with a deterioration of the quality of civil liberties even while merely junior coalition

partners. While work has been done quantifying the eﬀect they have on social policy as

coalition partners, no work has been oﬀered to quantify their impact on liberal democracy

while in such a position. I ﬁnd that for a standard four-year term in oﬃce, a transition from

no radical-right coalition partner, to a coalition where the radical-right holds 30% of gov-

ernment seats is associated with sizable decreases in freedom of speech, freedom of religion,

academic freedom, and access to justice.

Clearly, it seems the impact of the recent rise of the radical-right has brought with it

implications much starker than previously understood. Future work would be well served to

97

examine the altering ways in which radical-right parties are (1) able to ﬁnd themselves in a

ruling coalition, and (2) the extent of their ability to impact policy at-large while there.

98

CONCLUSION

“Both the far-right and the nationalists have their own strategy that is clear and timeless.
The division of society, the targeting of particular social groups, the cultivation of fear, the
invocation of a national grandeur, where only the chosen ones can ﬁt.”

Alexis Tspiras, Prime Minister of Greece, 2018
(while in governing coalition with right-populist and nationalist party Independent Greeks)

When thinking about radical-right parties in recent history, three key questions come

to bear. First, why are they getting elected? Second, once elected do they spend their time

in opposition, or do they ﬁnd themselves participating in government—and if so, how? And

third, what have the real eﬀects of this rise in popularity been, speciﬁcally to the durability

and stability of liberal democracy? These three questions have guided the research of this

dissertation.

In each chapter, I endeavor to contribute to one of these questions. In the ﬁrst, I

analyze the growing electoral success these parties have seen. I contend that no small part

of their election has been due to the rising trend in Europe of providing public funding for

electoral campaigns.

Indeed, I ﬁnd that publicly funding elections has served to dispro-

portionally aid parties of the radical-right. On average, following the adoption of public

campaign ﬁnancing, countries see roughly a tripling in the vote-share won by the radical-

right. Further, and interestingly, in countries with public ﬁnancing provisions, we no longer

see the well-deﬁned rule whereby radical-right success is dependent on economic growth.

Rather, public funding grants the parties resources to mobilize a base of support, regardless

of prevailing economic conditions. These results are surprising and counterintuitive, as pub-

lic ﬁnancing provisions are generally a cause supported by those on the political left of the

ideological spectrum.

99

Second, I delve into the question of what happens when elected. More and more lately,

it seems these parties are not conﬁned to only serving in opposition. Rather, in nearly half

of European countries, they have been invited to serve in parliamentary coalitions alongside

a more moderate party. Where research to date has argued nothing abnormal is going

on, I contend that this phenomenon is striking, and that current theories of parliamentary

coalition building are unable to explain this trend. Utilizing cross-national, survey, and

spatial data, I demonstrate that mainstream parties are utilizing coalition partnership in

response to the threat they feel of allowing radical-right parties to continue in opposition and

hurt their future probability of electoral success. This theory borrows from the literature

on authoritarian cooptation and contends that coalition partnership is less costly than a

radical-right rival in times when the radical-right platform is high in salience, and thus the

threat they impose the highest.

Finally, I shine some much needed light on the third question, which asks what the

actual eﬀects of this rise and rule has been. To date, most research agrees that while the

threat is high, the actual eﬀects have been small. This, problematically, is due to a faulty

premise that radical-right parties have little inﬂuence when serving as a junior coalition

partner. Rather, I argue that when mainstream parties invite the radical-right to serve

alongside them due to the threat they pose, we may see signiﬁcant eﬀects to the quality

of liberal democracy in the country. Somewhat disconcerting, this is exactly what I ﬁnd.

While the eﬀects are not the same as when the prime minister is themselves a member of

the radical-right, radical-right junior coalition members do indeed have a profound negative

eﬀect on the civil liberties of the citizens they come to rule.

In the introduction of this dissertation, I ask if the growing trend of radical-right

success is, as some occasionally call it, on the decline, or rather if they are here to stay. The

radical-right, I fear, is a global phenomenon that is here to stay. Many in the media take

advantage of a single election where a once-popular radical-right party sees support dwindle,

and asks an unanswerable question, “is this a sign of the coming end of their popularity?”

100

Sadly, the answer is likely always a resounding “no.”

Following the Austrian 1999 election, which saw the Freedom Party of Austria re-

ceive record-high votes, leading to their entrance into the governing coalition, their support

plummeted. In the next election their vote-share collapsed by 65%, leading to their worst

election result since the 1980s. Again, in the subsequent election that followed, their support

remained stagnant. By 2017, however, they rebounded, matched their initial 1999 result in

terms of vote-share, and reentered a governing coalition. This, it seems, is the new normal,

and the radical-right is not going anywhere. Rather, they have become like any mainstream

party-family; while their support may ebb and ﬂow in the short-term, they are forever a

consequential player in modern politics.

In many cases, steps have been taken to attenuate this growing trend. In the late

1980s, Belgium responded to the growing radical-right Vlaams Blok by introducing a cordon

sanitaire around the party. The term cordon sanitaire originated in the early 1800s as a

reference to a fortiﬁed border set up to prevent infectious disease pandemics from spread-

ing across national borders. The Belgians applied the term to describe mainstream parties’

agreement to refuse cooperation with radical-right opponents. Since then, similar agreements

have been entered into in Germany against the Alternative for Germany, and in the Nether-

lands against the Party for Freedom, among other examples. Yet, in eﬀect, these agreements

are simply words that parties are very willing to break if need be. As the currently ongoing

COVID-19 pandemic has shown, borders set up to stop spread will eventually fail. In 2020,

the ruling Christian Democrats in Germany cooperated with the Alternative for Germany

to remove a regional prime minister and replace him with a more favorable candidate. Such

agreement was referred to as a political “earthquake” in Germany and described as an action

having “torn up” the cordon.6 Indeed, in many cases politicians ﬁnd themselves having to

make partners with these parties they promised to isolate.

6See “Far-right party becomes German regional kingmaker” in the Financial Times by Guy Chazan

(2020).

101

The quote that precedes these concluding remarks was made by the then-prime min-

ister of Greece Alexis Tsipras in 2018, warning of the rise of right-populist and nationalist

parties. Prime Minister Tsipras did not even need a new election to go back on his word and

partner-up with his once sworn enemy; he made this statement while already in a coalition

with them. Such is the reality of politicians’ seeming willingness to work with partners they

once vehemently criticized. Cas Mudde, similarly, has noted that in much of Europe these

parties are now considered Koalitionsf¨ahig, or “acceptable coalition partners.”7

As I show in the last chapter, the eﬀects of this rise have been stark. Clearly, in the

few cases where the radical-right has found itself in the prime minister’s oﬃce, a decaying of

institutional checks on power was always imminent. Starker still, however, is the fact that

deep repercussions on civil liberties and basic human rights are felt even as they merely enter

government as a junior coalition member. Such was the importance of the cordon sanitaire.

And yet, as I show in chapter two, mainstream parties seem very willing to work with such

radical partners and jeopardize democracy—not for any noble reason or to pass important

legislation—but simply because they are scared of losing their own power.

Thus is the importance of deepening our understanding of the dynamics of these

parties. Where I focus much of this dissertation on understanding their electoral rise, it is

imperative to know, above all else, that this rise has occurred, they are here, and they are

not going anywhere.

7See “Nativism is driving the far-right surge in Europe—and it is here to stay” by Cas Mudde in The

Guardian (2019).

102

APPENDICES

103

APPENDIX A

APPENDIX TO CHAPTER 1

To illustrate whether or not the argued eﬀects are indeed unique to the far-right, I

re-run all presented models on far-left parties. The sample, included variables, and coding of

parties are all identical to models presented in the main paper.8 The models are presented

in table A.1. As expected, results indicate that their is no evidence of an eﬀect of public

ﬁnancing on far-left success, neither directly or in terms of a mediation on economic growth

(further, as others have found, there does not appear to be any eﬀect of economic growth

itself on far-left success). Rather, the eﬀects of public ﬁnancing are, due to distinct party-

family characteristics, unique to the far-right.

Table A.2 displays results for alternative speciﬁcations of the presented models. Col-

umn 1 is simply the main model presented in the paper. Column 2 omits the lagged-DV

while column 3 omits ﬁxed-eﬀects, and instead runs a pooled-tobit. In this speciﬁcation I

add controls for PR systems, as well as electoral thresholds (I omit threshold dummies from

the table for space). Column 4 re-speciﬁes the model as a random-eﬀects multilevel Tobit.

Again, in this speciﬁcation I include controls for electoral threshold and PR systems. In all

three models, degree of public ﬁnancing has a positive and signiﬁcant eﬀect on vote-share

of the extreme right. Further, the interaction term remains in the hypothesized direction

in all speciﬁcations as well, and graphical interpretations show the same eﬀect holds as

demonstrated previously (even in model 3, the coeﬃcient is signiﬁcant at the p < 0.10).

Next, in column 1 of table A.3 I re-run the main ﬁxed-eﬀects Tobit model from

8With the exception that “Lagged Vote-Share” now refers to far-left vote-share in the previous election

104

Table A.1: Results on Far-Left vote-Share

Degree Public Finance

GDP Growth

Inﬂation

# Parties

Government Debt

Lagged Vote-Share

Degree Public Finance × GDP Growth

Constant

Sigma

(1)
left

0.0223
(0.205)

(2)
left

0.0992
(0.401)

0.0309
(0.191)

(3)
left

-0.224
(0.514)

-0.0895
(0.167)

0.126

0.127

0.115

(0.0710)

(0.0703)

(0.0598)

-0.0904
(0.143)

-0.104
(0.128)

-0.147
(0.138)

-0.0136
(0.0166)
-0.168∗∗
(0.0598)

-0.0147
(0.0159)
-0.170∗∗
(0.0599)

-6.632∗∗∗
(1.579)
3.757∗∗∗
(0.490)

-6.632∗∗∗
(1.577)
3.755∗∗∗
(0.489)

-0.0172
(0.0142)
-0.179∗∗
(0.0628)

0.122
(0.126)
-6.087∗∗∗
(1.269)
3.723∗∗∗
(0.469)

328
N
pseudo R2
0.265
Log Pseudo-Likelyhood
-733.5
Standard errors clustered by country and appear in parentheses. Country
ﬁxed-eﬀects omitted. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, two-tailed test.

328
0.262
-736.0

328
0.262
-735.9

105

Table A.2: Results of Alternative Speciﬁcations

Degree Public Finance

GDP Growth

Inﬂation

# Parties

Government Debt

Lagged Vote-Share

Degree Public Finance × GDP Growth

Fixed-Eﬀects?

PR System

Other Controls?

Constant

Sigma

Country-Level Variance

Tobit

(2)
right

1.937∗∗∗
(0.565)
-1.248∗∗∗
(0.307)
-0.363∗∗
(0.134)

(3)
right

0.950∗
(0.428)
-0.421∗
(0.185)

-0.0969
(0.0693)

0.265
(0.486)

0.192
(0.193)

-0.0385
(0.0295)

0.375∗∗∗
(0.110)

-0.00798
(0.0145)
0.903∗∗∗
(0.0830)

0.120

(0.0714)

(1)
right

1.287∗∗∗
(0.320)
-0.657∗∗∗
(0.194)
-0.208∗
(0.0817)

-0.119
(0.215)
-0.0331∗
(0.0142)
0.692∗∗∗
(0.0946)
0.208∗∗
(0.0689)

Multilevel Tobit

(4)
right

1.446∗∗∗
(0.401)
-0.634∗∗
(0.238)
-0.208∗
(0.0812)

-0.264
(0.155)
-0.0324∗
(0.0137)
0.664∗∗∗
(0.106)
0.184∗
(0.0890)

Yes

Yes

No

Random

-2.613∗
(1.297)
4.346∗∗∗
(0.412)

-3.490
(1.799)
5.574∗∗∗
(0.582)

1.691
(1.569)

Threshold
-3.696∗
(1.861)
4.575∗∗∗
(0.501)

-0.505
(3.000)

Threshold

1.338

(2.0609)

19.43
(12.28)

N
Pseudo R2
Log Pseudo-Likelyhood
Standard errors clustered by country and appear in parentheses. Country ﬁxed-eﬀects and
threshold dummies (models 3 & 4) omitted for presentation. ∗p < 0.05, ∗∗p < 0.01,
∗∗∗p < 0.001, two-tailed test.

328
0.302
-518.8

348
0.247
-580.6

296
0.231
-465.6

296

-451.1

106

the paper using multiple-imputation, rather than the default listwise-deletion. Multiple

imputation is done through AMELIA II in R (see King et al. 1998). I compute 5 separate

imputed datasets. Further, I include a squared time polynomial, and interact these with

cross-sectional units, to allow time trends to behave diﬀerently across countries.9 Again,

there appear to be no material diﬀerences between the models.

Columns 2-5 of table A.3 present results with additional control values. These test

robustness against several variables argued to have contributed to the recent rise in extrem-

ism. Column 2 adds a control for percent of the population unemployed, column 3 adds a

control for immigration, measured as the annual inﬂow of foreign-born population (in thou-

sands of people).10 Column 4 includes a control for inequality, measured as the country’s

Gini coeﬃcient. Finally, in column 5 I include all three controls together. Again, the results

do not change in any of these cases (though interpreting interaction terms via statistical

signiﬁcance is not recommended, I note as a practical matter they all remain signiﬁcant at

the 0.1 level. More importantly, a graphical interpretation of results does not change.).

The results in table A.4 report results of several speciﬁcations testing robustness for

outlier sensitivity. The model speciﬁcations are as follows: model 1 omits observations at

the top 5th percentile of radical-right vote; model 2 omits the bottom 5th percentile of

economic growth, while model 3 omits the top 5th percentile of the same; model 4 presents

a jackknife-1 model; model 5 omits countries hit hardest by the debt crisis (Portugal, Italy,

Greece, Spain); ﬁnally model 6 omits observations after the debt crisis begins (all elections

post-2007).

In all cases, the results remain materially unchanged from the main results

presented. As before, where the interaction term appears to lose signiﬁcance in the jackknife

model, graphical interpretations of the results do not change.

Finally, ﬁgures A.1 and A.2 display comparative predictions from the Tobit and beta

9That is, the “intercs” option.

10I alternatively use measures of inﬂows of asylum seekers, and stock of foreign-born population (rather

than inﬂows). The results are constant across all speciﬁcations.

107

Table A.3: Results of Fixed-Eﬀects Tobit on Imputed Dataset

Degree Public Financing

GDP Growth

Inﬂation

# Parties

Government Debt

Lagged Vote-Share

Degree Public Finance × GDP Growth

Unemployment

Immigration (thousands)

Inequality

Constant

Sigma

(1)
right
1.413∗∗∗
(0.374)
-0.648∗∗∗
(0.173)

-0.190
(0.0966)

0.0219
(0.236)

-0.0224
(0.0151)
0.669∗∗∗
(0.0861)

0.141

(0.0816)

(2)
right
1.058∗∗
(0.383)
-0.653∗∗∗
(0.185)

-0.226
(0.129)

-0.0555
(0.283)

-0.0188
(0.0249)
0.661∗∗∗
(0.0833)
0.172∗
(0.0697)

0.421
(0.247)

(3)
right
1.447∗∗∗
(0.377)
-0.615∗∗∗
(0.173)

-0.173
(0.108)

(4)
right
1.261∗∗∗
(0.369)
-0.618∗∗∗
(0.156)
-0.202∗
(0.0977)

-0.00120
(0.227)

0.0157
(0.243)

-0.0278
(0.0172)
0.659∗∗∗
(0.0901)

-0.0241
(0.0158)
0.663∗∗∗
(0.0847)

0.136

0.135

(0.0803)

(0.0821)

(5)
right
0.842∗
(0.393)
-0.538∗∗
(0.166)

-0.204
(0.163)

-0.128
(0.287)

-0.0377
(0.0309)
0.625∗∗∗
(0.0850)
0.162∗
(0.0687)
0.580∗
(0.243)

0.00284
(0.00448)

0.00812
(0.00427)

-3.185∗
(1.491)
4.678∗∗∗
(0.351)

-5.010∗∗
(1.856)
4.582∗∗∗
(0.338)

-3.524∗
(1.705)
4.664∗∗∗
(0.346)

-0.0846
(0.0619)

-0.102
(0.0628)

-0.232
(2.441)
4.672∗∗∗
(0.347)

-3.153
(2.330)
4.501∗∗∗
(0.317)

N
Standard errors clustered by country and appear in parentheses. Country ﬁxed-eﬀects omitted for
presentation. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, two-tailed test.

348

348

348

348

348

108

(3)
right

Table A.4: Results of Fixed-Eﬀects Tobit for Outlier Sensitivity
(5)
(6)
right
right
1.290∗∗∗ 1.275∗∗∗
(0.335)
(0.320)
-0.634∗∗
-0.551∗
(0.206)
(0.217)
-0.198∗
(0.0845)

(2)
(4)
right
right
1.084∗∗ 1.248∗∗∗ 1.287∗∗
(0.410)
(0.410)
(0.327)
-0.657∗
-1.085∗∗∗ -0.672∗∗
(0.203)
(0.247)
(0.316)
-0.239∗∗
-0.195∗
-0.208
(0.0917) (0.0793) (0.125)

(1)
right
1.389∗∗∗
(0.292)
-0.543∗∗
(0.172)
-0.195∗
(0.0752)

-0.182
(0.112)

Degree Public Financing

GDP Growth

Inﬂation

# Parties

Government Debt

Lagged Vote-Share

Degree Public Finance × GDP Growth

Constant

Sigma

(0.0135)

-0.251
(0.209)

-0.119
(0.194)

-0.110
(0.215)

-0.0834
(0.251)

-0.0292
-0.171
(0.217)
(0.228)
-0.0468∗∗∗ -0.0422∗∗ -0.0289∗ -0.0331 -0.0509∗∗∗ -0.0349∗
(0.0160) (0.0136) (0.0187)
(0.0151)
(0.0133)
0.674∗∗∗ 0.694∗∗∗ 0.692∗∗∗ 0.754∗∗∗ 0.695∗∗∗
0.480∗∗∗
(0.102)
(0.0895) (0.0953) (0.103)
(0.0806)
0.180∗∗
0.320∗∗∗ 0.186∗∗
0.205∗
0.208
(0.0624)
(0.0990)
(0.0933) (0.0694) (0.129)
-2.062∗
(0.967)
3.684∗∗∗
(0.380)

-0.524
-2.347
(1.368)
(1.367)
4.328∗∗∗ 4.351∗∗∗ 4.346∗∗∗ 4.171∗∗∗ 4.232∗∗∗
(0.415)
(0.415)

(0.0850)
0.226∗∗
(0.0754)

-2.626∗
(1.297)

-2.764
(1.582)

-2.613
(2.447)

(0.419)

(0.401)

(0.466)

N
Pseudo R2
Log Pseudo-Likelyhood

313
0.307
-481.9
Standard errors clustered by country and appear in parentheses. Country ﬁxed-eﬀects omitted for
presentation. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, two-tailed test.

312
0.297
-449.7

311
0.303
-487.1

314
0.298
-507.3

328
0.302
-518.8

283
0.316
-429.1

109

Figure A.1: Tobit vs. Beta Regression: Direct Eﬀects

regression models. Figure A.1 presents predictions from the direct eﬀect of public campaign

ﬁnance, where the black line represents Tobit predictions while the grey represents the beta

model. The predictions and conﬁdence intervals are nearly identical between the two models.

Meanwhile, ﬁgure A.2 displays predictions of the interactive eﬀect. Again, black

represents the Tobit model, while grey indicates the beta. Further the dashed lines displays

the relationship between GDP growth as vote-share at high levels of public ﬁnancing, while

the solid lines indicate the same at low levels of public ﬁnancing. Again, the results are the

same.

110

Figure A.2: Tobit vs. Beta Regression: Marginal Eﬀects

111

APPENDIX B

APPENDIX TO CHAPTER 3

Table B.1 Re-runs models using alternative dependent variables. Columns 1 and

2 present the eﬀects of radical-right prime minister and coalition seat-share on the World

Bank’s rule-of-law indicator. Results remain robust to this dependent variable, in that

radical-right prime ministers have a signiﬁcant negative eﬀect, while mere coalition member-

ship does not. Columns 3 and 4 in turn rerun models using Freedom House’s Civil Liberties

measure. Results again are materially unchanged from the main models presented. The

prime minister’s eﬀect on civil liberties is signiﬁcant at the 1% level (p = 0.07), likely due

in part to the lower number of observations, as Freedom House’s disaggregated data begins

is only available post-2006. Note, due to the decreased timeline, autocracy drops out of the

model as no countries in the sample experienced a polity score below 7 during this more

recent sample.

112

Table B.1: Models With Alternative DVs

World Bank
Rule-of-Law

Freedom House
Civil Liberties

(1)

(2)

(3)

(4)

Radical-Right PM −0.089∗∗∗ −0.054 −0.750
(0.414)

(0.022)

Rad-Right Govt Seat

log(GDP/Capita)

Gini

Parliamentary

Lagged DV

Constant

(0.035)

(0.756)

(0.035)
−0.069
(0.061)

0.193
(0.459)
−1.872∗
(0.785)
0.131∗∗∗ 0.132∗∗∗ −0.975 −0.901
(0.036)
(0.758)
−0.010 −0.010 −0.091 −0.082
(0.006)
(0.081)
(0.006)
0.064∗∗∗ 1.153∗∗∗ 1.169∗∗∗
0.064
(0.006)
(0.098)
(0.006)
0.715∗∗∗ 0.712∗∗∗ 0.607∗∗∗ 0.591∗∗∗
(0.071)
(0.127)
−0.654∗ −0.666∗ 34.548∗∗ 34.455∗∗
(11.419)
(0.297)

(11.414)

(0.128)

(0.100)

(0.291)

(0.071)

(0.082)

Observations
R2
Adjusted R2

406
0.990
0.989

406
0.990
0.989

290
0.974
0.971

290
0.974
0.971

Standard errors are clustered by country and appear in
parentheses. Country ﬁxed- eﬀects omitted for presentation.
∗p<0.05; ∗∗p<0.01; ∗∗∗p<0.001

113

Figure B.1 below displays the substantive eﬀects of radical-right prime ministers over

time. These results are based on models where the main independent variable is the number

of years in oﬃce.

Figure B.1: Eﬀect of Radical-Right Prime Minister Over Time

114

Free/Fair ElectionsLegislative ConstraintsMedia Critical of GovtJudicial ConstraintsEqual ProtectionRelgious FreedomMedia Bias against Opp.Media CensorshipAcademic FreedomAccess to JusticeSpeech0123401234012341.82.02.22.42.61.01.52.02.50.860.880.900.922.32.42.52.61.501.752.000.800.840.880.800.851.21.62.02.41.21.41.61.82.01.01.52.00.840.860.880.900.920.94Years of Radical−Right Prime MinisterLevel of DemocracyFigures B.2 and B.3 below plot the parameter distribution from a country-clustered

jackknife procedure. The x-axis refers to the calculated coeﬃcient of the radical-right prime

minister variable, while the y-axis refers to the frequency of that parameter.

Figure B.2: Jackknife Prime Minister Parameter Estimates, Rule-of-Law

Figure B.3: Jackknife Prime Minister Parameter Estimates, Civil Liberties

115

Judicial ConstraintsFree/Fair ElectionsLegislative ConstraintsMedia Bias against Opp.Media CensorshipMedia Critical of Govt−0.8−0.6−0.4−0.20.0−0.8−0.6−0.4−0.20.0−0.8−0.6−0.4−0.20.00102001020Parameter EstimateDensityRelgious FreedomEqual ProtectionAcademic FreedomAccess to JusticeSpeech−0.4−0.20.0−0.4−0.20.0−0.4−0.20.001020300102030Parameter EstimateDensityTable B.2 presents the eﬀects of a radical-right prime minister on institutional rule-

of-law with Hungary omitted. Results largely remain unchanged.

Table B.2: Institutional Rule-of-Law, Hungary Omitted

Free/Fair
Elections Constraints Constraints Censorship

Legislative Media Media Critical Media Bias
Against Opp

Judicial

of Govt

(1)

Radical-Right PM −0.001
(0.003)

(2)

−0.034∗∗
(0.012)

log(GDP/Capita)

Autocracy

Gini

Parliamentary

Lagged DV

Constant

0.011
(0.007)
−0.028
(0.023)

0.002
(0.001)

0.026
(0.022)

0.221
(0.150)
0.544∗∗∗
(0.108)

0.002
(0.003)

0.024
(0.022)

0.002
(0.001)
−0.012
(0.008)
0.767∗∗∗
(0.109)

0.146
(0.119)

(3)

(4)

−0.017∗ −0.769∗∗∗
(0.127)
(0.008)
−0.038
0.019∗∗
(0.041)
(0.007)

0.043
(0.026)

0.027
(0.194)

0.002
(0.002)
−0.023∗∗
(0.009)
0.480∗∗∗
(0.118)
0.226∗
(0.103)

0.015
(0.012)
−0.061
(0.061)
0.709∗∗∗
(0.085)

0.810
(0.437)

Observations
R2
Adjusted R2

712
0.940
0.937

712
0.950
0.947

712
0.930
0.926

712
0.904
0.899

(5)

−0.388∗∗∗
(0.012)
−0.035
(0.035)
−0.020
(0.031)

0.007
(0.010)
−0.110∗
(0.044)
0.775∗∗∗
(0.081)
0.778∗
(0.378)

712
0.901
0.896

(6)

−0.319∗
(0.161)
−0.074
(0.038)

0.043
(0.123)

0.006
(0.009)
−0.113
(0.083)
0.665∗∗∗
(0.058)
1.401∗∗∗
(0.425)

712
0.803
0.794

Standard errors are clustered by country and appear in parentheses. Country ﬁxed-eﬀects
omitted for presentation. ∗p<0.05; ∗∗p<0.01; ∗∗∗p<0.001

116

Table B.3 presents the eﬀects of a radical-right prime minister on institutional rule-

of-law with Poland omitted. Results remain largely unchanged. While radical-right prime

minister appears to lose signiﬁcance in regard to legislative constraints, as a practical matter

both the coeﬃcient and standard error are nearly the same as in the main model, and

maintains p < 0.1.

Table B.3: Institutional Rule-of-Law, Poland Omitted

Free/Fair
Elections Constraints Constraints Censorship

Legislative Media Media Critical Media Bias
Against Opp

Judicial

of Govt

(1)

(2)

Radical-Right PM −0.045 −0.026∗∗∗
(0.005)

(0.031)

log(GDP/Capita)

Autocracy

Gini

Parliamentary

Lagged DV

Constant

0.012
(0.007)
−0.017
(0.016)

0.002
(0.001)

0.028
(0.023)

0.205
(0.152)
0.554∗∗∗
(0.116)

0.001
(0.004)

0.003
(0.008)

0.002
(0.001)
−0.013
(0.007)
0.838∗∗∗
(0.048)

0.096
(0.072)

(3)

−0.047
(0.027)

0.015
(0.007)
0.027∗
(0.013)

0.001
(0.001)
−0.022
(0.011)
0.619∗∗∗
(0.078)
0.163∗
(0.066)

(4)

−0.501∗∗∗
(0.086)
−0.039
(0.036)
−0.135
(0.088)

0.015
(0.009)
−0.075
(0.042)
0.766∗∗∗
(0.066)

0.681
(0.387)

Observations
R2
Adjusted R2

711
0.937
0.934

711
0.967
0.965

711
0.946
0.943

711
0.928
0.924

(5)

−0.377∗∗∗
(0.051)
−0.015
(0.031)
−0.019
(0.032)

0.004
(0.010)
−0.112∗
(0.045)
0.777∗∗∗
(0.079)

0.629
(0.389)

711
0.913
0.909

(6)

−0.171∗∗∗
(0.042)
−0.080∗
(0.034)
−0.062∗
(0.030)

0.004
(0.008)
−0.105
(0.086)
0.701∗∗∗
(0.051)
1.438∗∗∗
(0.410)

711
0.845
0.837

Standard errors are clustered by country and appear in parentheses. Country ﬁxed-eﬀects
omitted for presentation. ∗p<0.05; ∗∗p<0.01; ∗∗∗p<0.001

117

Table B.4 presents the eﬀects of a radical-right prime minister on civil liberties with

Hungary omitted. Results again remain quite similar, with the exception that the eﬀect on

academic freedom is no longer signiﬁcant, though the substantive size of the eﬀect remains

unchanged.

Table B.4: Civil Liberties, Hungary Omitted

Freedom of Religious Academic Access to Equal Protection
for Out-Groups

Freedom Freedom Justice

Speech

(1)

Radical-Right PM −0.133
(0.116)

log(GDP/Capita)

Autocracy

Gini

Parliamentary

Lagged DV

Constant

0.020
(0.044)

0.141
(0.166)
−0.016
(0.011)
−0.090∗
(0.045)
0.778∗∗∗
(0.041)
0.882∗∗
(0.298)

(2)

(3)

(4)

0.088
(0.135)

−0.104∗ −0.344 −0.047∗∗
(0.016)
(0.043)
(0.222)
−0.028
−0.038 −0.034
(0.022)
(0.029)
(0.042)
−0.065
−0.018
(0.053)
(0.146)
−0.012∗
−0.002
0.007
(0.006)
(0.006)
(0.010)
−0.183
−0.035 −0.131
(0.126)
(0.097)
(0.046)
0.792∗∗∗
0.780∗∗∗
0.840∗∗∗
(0.060)
(0.041)
(0.045)
1.141∗∗∗
0.891∗∗
1.032∗∗∗
(0.311)
(0.296)
(0.224)

Observations
R2
Adjusted R2

712
0.892
0.887

712
0.921
0.917

712
0.896
0.891

712
0.970
0.969

(5)

−0.004∗∗
(0.002)
−0.001
(0.002)

0.006
(0.005)

0.0002
(0.001)
−0.007
(0.004)
0.801∗∗∗
(0.045)
0.204∗∗∗
(0.048)

712
0.962
0.960

Standard errors are clustered by country and appear in parentheses. Country
ﬁxed-eﬀects omitted for presentation. ∗p<0.05; ∗∗p<0.01; ∗∗∗p<0.001

118

Table B.5 presents the eﬀects of a radical-right prime minister on civil liberties with

Poland omitted. Results remain unchanged.

Table B.5: Civil Liberties, Poland Omitted

Freedom of Religious Academic Access to Equal Protection
for Out-Groups

Freedom Freedom Justice

Speech

(1)

Radical-Right PM −0.084
(0.065)

log(GDP/Capita)

Autocracy

Gini

Parliamentary

Lagged DV

Constant

0.009
(0.040)

0.016
(0.078)
−0.014
(0.009)
−0.090
(0.051)
0.817∗∗∗
(0.030)
0.836∗∗∗
(0.247)

(2)

(3)

(4)

(0.027)

(0.131)
−0.030
(0.038)

−0.154∗ −0.259∗ −0.056∗∗
(0.021)
(0.070)
−0.026
−0.044
(0.028)
(0.022)
−0.177∗∗∗ −0.029 −0.064∗∗
(0.023)
(0.041)
−0.009
−0.001
(0.006)
(0.005)
−0.185
−0.042
(0.127)
(0.049)
0.811∗∗∗
0.842∗∗∗
(0.059)
(0.035)
1.105∗∗∗
0.994∗∗
(0.323)
(0.222)

0.006
(0.009)
−0.147
(0.101)
0.810∗∗∗
(0.028)
0.801∗∗
(0.278)

Observations
R2
Adjusted R2

711
0.904
0.899

711
0.927
0.924

711
0.916
0.912

711
0.969
0.968

(5)

−0.012∗∗∗
(0.003)
−0.002
(0.002)

0.0005
(0.002)

0.0004
(0.001)
−0.007
(0.004)
0.819∗∗∗
(0.037)
0.192∗∗∗
(0.043)

711
0.963
0.961

Standard errors are clustered by country and appear in parentheses. Country
ﬁxed-eﬀects omitted for presentation. ∗p<0.05; ∗∗p<0.01; ∗∗∗p<0.001

119

REFERENCES

120

REFERENCES

Abou-Chadi, Tarik. 2016. “Niche Party Success and Mainstream Party Policy Shifts - How
Green and Radical Right Parties Diﬀer in Their Impact.” British Journal of Political
Science 46(2): 417–436.

Abou-Chadi, Tarik, and Werner Krause. 2018. “The Causal Eﬀect of Radical Right Success
on Mainstream Parties’ Policy Positions: A Regression Discontinuity Approach.” British
Journal of Political Science 41: 1–19.

Abramson, Paul R, John H Aldrich, Andr´e Blais, Matthew Diamond, Abraham Diskin,
Indridi H Indridason, Daniel J Lee, and Renan Levine. 2009. “Comparing Strategic Voting
Under FPTP and PR.” Comparative Political Studies 43(1): 61–90.

Adams, James, Michael Clark, Lawrence Ezrow, and Garrett Glasgow. 2006. “Are Niche
Parties Fundamentally Diﬀerent from Mainstream Parties? The Causes and the Electoral
Consequences of Western European Parties’ Policy Shifts, 1976-1998.” American Journal
of Political Science 50(3): 513–529.

Akkerman, Tjitske. 2012. “Comparing Radical Right Parties in Government: Immigration
and Integration Policies in Nine Countries (1996–2010).” West European Politics 35: 511–
529.

Algan, Yann, Sergei Guriev, Elias Papaioannou, and Evgenia Passari. 2017. “The European

Trust Crisis and the Rise of Populism.” Brookings Papers on Economic Activity.

Arzheimer, Kai. 2009. “Contextual Vote Factors and the Extreme Right in Western Europe,

1980 - 2002.” American Journal of Political Science 53(2): 259–275.

Arzheimer, Kai. 2015. “The AfD: Finally a Successful Right-Wing Populist Eurosceptic

Party for Germany?” West European Politics 38(3): 535–556.

Arzheimer, Kai. 2018. “Explaining Electoral Support for the Radical Right.” In The Oxford

Handbook of the Radical Right, ed. Jens Rydgren. Oxford University Press.

Arzheimer, Kai, and Carl C. Berning. 2019. “How the Alternative for Germany (AfD) and

Their Voters Veered to the Radical Right, 2013–2017.” Electoral Studies 60: 102040.

Axelrod, Robert M. 1970. Conﬂict of Interest: A Theory of Divergent Goals with Applications

to Politics. Chicago: Markham Publishing Company.

Bar-On, Tamir. 2018. “The Radical Right and Nationalism.” In The Oxford Handbook of

the Radical Right, ed. Jens Rydgren. Oxford University Press.

Baron, David P. 1994. “Electoral Competition with Informed and Uninformed Voters.”

American Political Science Review 88(1): 33–47.

Benoit, Kenneth, and Michael Laver. 2006. Party Policy in Modern Democracies. Routledge.

121

Berbuir, Nicole, Marcel Lewandowsky, and Jasmin Siri. 2015. “The AfD and Its Sympathis-
ers: Finally a Right-Wing Populist Movement in Germany?” German Politics 24(2): 157–
178.

Betz, Hans-George. 1994. Radical Right-Wing Populism in Western Europe. Palgrave

Macmillan.

Betz, Hans-George. 2005. “Against the System: Radical Right-Wing Populism’s Challenge
to Liberal Democracy.” In Movements of Exclusion: Radical Right-Wing Populism in the
Western World, ed. Jens Rydgren. Nova Science Publishers.

Bhat, Prashanth, and Kalyani Chadha. 2020. “Anti-Media Populism: Expressions of Me-
dia Distrust by Right-Wing Media in India.” Journal of International and Intercultural
Communication 13(2): 166–182.

Bichay, Nicolas. 2020. “Public Campaign Financing and the Rise of Radical-Right Parties.”

Electoral Studies 66: 102159.

Bichay, Nicolas. 2021. “Portfolio Distributions of Radical-Right Cabinet Ministers.” Center

for Analysis of the Radical Right.

Bolleyer, Nicole. 2007. “Small Parties: From Party Pledges to Government Policy.” West

European Politics 30(1): 121–147.

Bowler, Shaun, Jeﬀrey A. Karp, and Todd Donovan. 2010. “Strategic Coalition Voting:

Evidence from New Zealand.” Electoral Studies 29(3): 350–357.

Browne, Eric C., and Mark N. Franklin. 1973. “Aspects of Coalition Payoﬀs in European

Parliamentary Democracies.” American Political Science Review 67(2): 453–469.

Br¨uckner, Markus, and Hans Peter Gr¨uner. 2010.

“Economic Growth and the Rise of

Political Extremism: Theory and Evidence.” CEPR Discussion Papers.

Budge, Ian, and Michael Laver. 1993. “The Policy Basis of Government Coalitions: A

Comparative Investigation.” British Journal of Political Science 23(4): 499–519.

Bugariˇc, Bojan. 2015. “A Crisis of Constitutional Democracy in Post-Communist Europe:
’Lands in-between’ Democracy and Authoritarianism.” International Journal of Constitu-
tional Law 13(1): 219–245.

Buhr, Renee L. 2013. “Seizing the Opportunity: Euroscepticism and Extremist Party Success

in the Post-Maastricht Era.” Government and Opposition 47(04): 544–573.

Castles, Francis G., and Peter Mair. 1984. “Left-Right Political Scales: Some ’expert’ Judge-

ments.” European Journal of Political Research 12(1): 73–88.

Cianetti, Licia, James Dawson, and Se´an Hanley. 2018. “Rethinking “Democratic Backslid-
ing” in Central and Eastern Europe–Looking beyond Hungary and Poland.” East European
Politics 34(3): 243–256.

122

Colantone, Italo, and Piero Stanig. 2018. “Global Competition and Brexit.” American Po-

litical Science Review 112(2): 201–218.

Dahl, Robert A. 1971. Polyarchy: Participation and Opposition. Yale University Press.

De Bromhead, Alan, Barry Eichengreen, and Kevin O’Rourke. 2012. “Right-Wing Political

Extremism in the Great Depression.” National Bureau of Economic Research.

De Lange, Sarah L. 2012. “New Alliances: Why Mainstream Parties Govern with Radical

Right-Wing Populist Parties.” Political Studies 60(4): 899–918.

De Lange, Sarah L., and Tjitske Akkerman. 2012. “Populist Parties in Belgium: A Case
of Hegemonic Liberal Democracy?” In Populism in Europe and the Americas: Threat or
Corrective for Democracy?

de Swaan, Abram. 1973. Coalition Theories and Cabinet Formations. A Study of Formal
Theories of Coalition Formation Applied to Nine European Parliaments after 1918. Else-
vier Scientiﬁc.

Debus, Marc, and Jochen M¨uller. 2013. “Do Voters’ Coalition Preferences Aﬀect Government

Formation?” West European Politics 36(5): 1007–1028.

Dennis, Christopher. 1998. “Support for Campaign Spending Limitations in the U. S. Sen-
ate: The Role of Party, Ideology and Electoral Security.” Journal of Socio-Economics
27(5): 641–649.

Dennison, James. 2020. “How Issue Salience Explains the Rise of the Populist Right in

Western Europe.” International Journal of Public Opinion Research 32(3): 397–420.

Deutscher Bundestag. 2018.

“Bekanntmachung von Rechenschaftsberichten Politischer
Parteien F¨ur Das Kalenderjahr 2016 (1. Teil- Bundestagsparteien).” Drucksache 19/2300.

Dinas, Elias, Konstantinos Matakos, Dimitrios Xefteris, and Dominik Hangartner. 2019.
“Waking Up the Golden Dawn: Does Exposure to the Refugee Crisis Increase Support for
Extreme-Right Parties?” Political Analysis 27(2): 244–254.

Doleˇzalov´a, Jitka. 2015. “Economic Crisis and Growth in Vote Share for Extreme Left and

Extreme Right Parties.” Review of Economic Perspectives 15(3): 269–290.

D¨oring, Holger, and Johan Hellstr¨om. 2013. “Who Gets into Government? Coalition For-

mation in European Democracies.” West European Politics 36(4): 683–703.

D¨oring, Holger, and Philip Manow. 2019. Parliaments and Governments Database (Parl-

Gov): Information on Parties, Elections and Cabinets in Modern Democracies.

Downes, James F., and Matthew Loveless. 2018. “Centre Right and Radical Right Party
Competition in Europe: Strategic Emphasis on Immigration, Anti-Incumbency, and Eco-
nomic Crisis.” Electoral Studies 54: 148–158.

123

Duncan, Fraser. 2010. “Immigration and Integration Policy and the Austrian Radical Right

in Oﬃce: The FP ¨O/BZ ¨O, 2000–2006.” Contemporary Politics 16(4): 337 354.

Dunphy, Richard, and Tim Bale. 2011. “The Radical Left in Coalition Government: Towards

a Comparative Measurement of Success and Failure.” Party Politics 17(4): 488–504.

Ellinas, Antonis A. 2018. “Media and the Radical Right.” In The Oxford Handbook of the

Radical Right, ed. Jens Rydgren. Oxford University Press.

Enyedi, Zsolt. 2018. “Democratic Backsliding and Academic Freedom in Hungary.” Perspec-

tives on Politics 16(4): 1067–1074.

Erikson, Robert S. 1989. “Economic Conditions and the Presidential Vote.” American Po-

litical Science Review 83(2): 567–573.

Esenberg, Richard M. 2010. “The Lonely Death of Public Campaign Financing.” Harvard

Journal of Law and Public Policy 33(1): 283–332.

Falc´o-Gimeno, Albert. 2012. “Preferences for Political Coalitions in Spain.” South European

Society and Politics 17(3): 487–502.

Fallend, Franz. 2012. “Populism in Government: The Case of Austria (2000–2007).” In
Populism in Europe and the Americas: Threat or Corrective for Democracy?, eds. Cas
Mudde, and Cristobal Rovira Kaltwasser. Cambridge University Press.

Ferrari, Silvia L.P., and Francisco Cribari-Neto. 2004. “Beta Regression for Modelling Rates

and Proportions.” Journal of Applied Statistics 31(7): 799–815.

Figenschou, Tine Ustad, and Karoline Andrea Ihlebæk. 2019. “Challenging Journalistic Au-
thority: Media Criticism in Far-Right Alternative Media.” Journalism Studies 20(9): 1221–
1237.

Funke, Manuel, Moritz Schularick, and Christoph Trebesch. 2016. “Going to Extremes:

Politics after Financial Crises, 1870–2014.” European Economic Review 88: 227–260.

Gamson, William A. 1961. “A Theory of Coalition Formation.” American Sociological Review

26(3): 373.

Gidron, Noam, and Peter A. Hall. 2020. “Populism as a Problem of Social Integration.”

Comparative Political Studies 53(7): 1027–1059.

Golder, Matt. 2003. “Explaining Variation In The Success Of Extreme Right Parties In

Western Europe.” Comparative Political Studies 36(4): 432–466.

Greco. 2011.

“Third Evaluation Round: Evaluation Report on Austria Transparency of

Party Funding.” Council of Europe.

Greene, William. 2004a. “The Behaviour of the Maximum Likelihood Estimator of Limited
Dependent Variable Models in the Presence of Fixed Eﬀects.” The Econometrics Journal
7(1): 98–119.

124

Greene, William. 2004b. “Fixed Eﬀects and Bias Due to the Incidental Parameters Problem

in the Tobit Model.” Econometric Reviews 23(2): 125–147.

Guiso, Luigi, Helios Herrera, Massimo Morelli, and Tommaso Sonno. 2017. “Demand and

Supply of Populism.” Bankpedia Review 7(1-2): 7.

Hall, Andrew B. 2014. “How The Public Funding Of Elections Increases Candidate Polar-

ization.” Unpublished Manuscript.

Han, Kyung Joon. 2016. “Income Inequality and Voting for Radical Right-Wing Parties.”

Electoral Studies 42: 54–64.

Hatton, Timothy J. 2021. “Public Opinion on Immigration in Europe: Preference and

Salience.” European Journal of Political Economy 66: 101969.

Heinisch, Reinhard. 2003. “Success in Opposition-Failure in Government: Explaining the
Performamce of Right-Wing Populist Parties in Public Oﬃce.” West European Politics
26(3): 91–130.

Herrmann, Michael. 2014. “Polls, Coalitions and Strategic Voting under Proportional Rep-

resentation.” Journal of Theoretical Politics 26(3): 442–467.

Hibbing, John R., and John R. Alford. 1981. “The Electoral Impact of Economic Conditions:

Who Is Held Responsible?” American Journal of Political Science 25(3): 423–439.

Houle, Christian, and Paul D. Kenny. 2018. “The Political and Economic Consequences of

Populist Rule in Latin America.” Government and Opposition 53(2): 256–287.

Huber, John, and Ronald Inglehart. 1995. “Expert Interpretations of Party Space and Party

Locations in 42 Societies.” Party Politics 1(1): 73–111.

Huber, Robert A, and Christian H Schimpf. 2017. “On the Distinct Eﬀects of Left-Wing
and Right-Wing Populism on Democratic Quality.” Politics and Governance 5(4): 146 20.

Ignazi, Piero. 2006. Extreme Right Parties in Western Europe. Oxford University Press.

Inglehart, Ronald, and Pippa Norris. 2016.

“Trump, Brexit, and the Rise of Populism:

Economic Have-Nots and Cultural Backlash.” HKS Working Paper RWP16-026.

Jackman, Robert W, and Karin Volpert. 1996. “Conditions Favouring Parties of the Extreme

Right in Western Europe.” British Journal of Political Science 26(04): 501.

Jones, Ruth S. 1981. “State Public Campaign Finance: Implications for Partisan Politics.”

American Journal of Political Science 25(2): 342.

Jones, Ruth S, and Thomas J Borris. 1985. “Strategic Contributing in Legislative Campaigns:

The Case of Minnesota.” Legislative Studies Quarterly 10(1): 89.

Kallis, Aristotle. 2018. “The Radical Right and Islamophobia.” In The Oxford Handbook of

the Radical Right, ed. Jens Rydgren. Oxford University Press.

125

Kaltwasser, Crist´obal Rovira. 2012. “The Ambivalence of Populism: Threat and Corrective

for Democracy.” Democratization 19(2): 184–208.

Katz, Richard S., and Peter Mair. 1995. “Changing Models of Party Organization and Party

Democracy: The Emergence of the Cartel Party.” Party Politics 1(1): 5–28.

Kaufman, Robert R., and Stephan Haggard. 2019. “Democratic Decline in the United
States: What Can We Learn from Middle-Income Backsliding?” Perspectives on Politics
17(2): 417–432.

Kayser, Mark Andreas, and Michael Peress. 2012. “Benchmarking across Borders: Elec-
toral Accountability and the Necessity of Comparison.” American Political Science Review
106(3): 661–684.

Kendall-Taylor, Andrea, and Erica Frantz. 2016. “How Democracies Fall Apart.” Foreign

Aﬀairs.

Kenny, Paul D. 2020. ““The Enemy of the People”: Populists and Press Freedom.” Political

Research Quarterly 73(2): 261–275.

Kiess, Johannes M., and Hans J¨org Trenz. 2019. “Ties of Solidarity and the Political Spec-
trum: Partisan Cleavages in Reported Solidarity Activity Across Europe.” American Be-
havioral Scientist 63(4): 459–474.

King, Gary, James Honaker, Anne Joseph, and Kenneth Scheve. 1998. “Listwise Deletion Is

Evil: What to Do About Missing Data in Political Science.” Unpublished Manuscript.

Kitschelt, Herbert, and Anthony J McGann. 1997. The Radical Right in Western Europe:

A Comparative Analysis. University of Michigan Press.

Knigge, Pia. 1998. “The Ecological Correlates of Right-Wing Extremism in Western Europe.”

European Journal of Political Research 34(2): 249–279.

Laver, Michael. 1974. “Dynamic Factors in Government Coalition Formation.” European

Journal of Political Research 2(3): 259–270.

Laver, Michael, and Kenneth A. Shepsle. 1996. Making and Breaking Governments: Cabinets

and Legislatures in Parliamentary Democracies. Cambridge University Press.

Lees, Charles. 2001. “Coalitions: Beyond the Politics of Centrality?” German Politics

10(2): 117–134.

Levitsky, Steven, and James Loxton. 2012. “Populism and Competitive Authoritarianism:
The Case of Fujimori’s Peru.” In Populism in Europe and the Americas: Threat Or Cor-
rective for Democracy?, eds. Cas Mudde, and Crist´obal Rovira Kaltwasser. Cambridge
University Press.

Loch, Dietmar, and Ov Cristian Norocel. 2015. “The Populist Radical Right in Europe.” In
Europe’s Prolonged Crisis, eds. Virginie Guiraudon, Carlo Ruzza, and Hans-J¨org Trenz.
Palgrave Macmillan.

126

Lubbers, Marcel, M´erove Gijsberts, and Peer Scheepers. 2002. “Extreme Right-Wing Voting

in Western Europe.” European Journal of Political Research 41(3): 345–378.

Luther, Kurt Richard. 2007. “Electoral Strategies and Performance of Austrian Right-Wing

Populism 1986-2006.” KEPRU Working Paper 24.

Lux, Thomas. 2018. “The AfD and the Low Social Strata. A Research Note on Holger
Lengfeld’s Study The “Alternative Fur Deutschland”: A Party for Losers of Societal Mod-
ernisation?” Kolner Zeitschrift fur Soziologie und Sozialpsychologie 70(2): 255–273.

Malhotra, Neil. 2008. “The Impact of Public Financing on Electoral Competition: Evidence

from Arizona and Maine.” State Politics and Policy Quarterly 8(3): 263–281.

Maravall, Jos´e Mar´ıa. 2010. “Accountability in Coalition Governments.” Annual Review of

Political Science 13(1): 81–100.

March, Luke, and Charlotte Rommerskirchen. 2015. “Out of Left Field? Explaining the
Variable Electoral Success of European Radical Left Parties.” Party Politics 21(1): 40–53.

Martin, Lanny W, and Georg Vanberg. 2011. Parliaments and Coalitions: The Role of

Legislative Institutions in Multiparty Governance. Oxford University Press.

Masket, Seth E, and Michael G Miller. 2015. “Does Public Election Funding Create More
Extreme Legislators? Evidence from Arizona and Maine.” State Politics and Policy Quar-
terly 15(1): 24–40.

Mayer, Kenneth R, and Timothy Werner. 2007. “Electoral Transitions in Connecticut: The
Implementation of Clean Elections in 2008.” Paper presented at the Annual Meeting of
the American Political Science Association, Chicago, IL.

Mayer, Nonna, and Pascal Perrineau. 1992. “Why Do They Vote for Le Pen?” European

Journal of Political Research 22(1): 123–141.

McCarthy, Nolan, Keith T Poole, and Howard Rosenthal. 2006. Polarized America: The

Dance of Ideology and Unequal Riches. MIT Press.

Meret, Susi. 2010. The Danish People’s Party, the Italian Northern League and the Aus-
trian Freedom Party in a Comparative Perspective: Party Ideology and Electoral Support.
Doctoral Dissertation, Aalborg University.

Miller, Michael G. 2008. “Gaming Arizona: Public Money and Shifting Candidate Strate-

gies.” PS - Political Science and Politics 41(3): 527–532.

Miller, Michael G. 2011. “After the GAO Report: Wht Do We Know About Public Election

Funding?” Election Law Journal 10(3): 273–290.

Minkenberg, M. 2001. “The Radical Right in Public Oﬃce: Agenda-Setting and Policy

Eﬀects.” West European Politics 24(4): 1–21.

127

Moder, Clara Maria. 2019. “What Happened to Poland? On the Ongoing Crisis of Democ-
racy.” In Old and New Cleavages in Polish Society, eds. Ulrike Gu´erot, and Michael Hun-
klinger. Edition Donau-Universit¨at Krems.

Mudde, Cas. 1996. “The War of Words Deﬁning the Extreme Right Party Family.” West

European Politics 19(2): 225–248.

Mudde, Cas. 2004. “The Populist Zeitgeist.” Government and Opposition 39: 541 563.

Mudde, Cas. 2006. “Anti-System Politics.” In Developments in European Politics, eds.
Paul M. Heywood, Erik Jones, Martin Rhodes, and Ulrich Sedelmeier. Palgrave Macmil-
lan.

Mudde, Cas. 2007. Populist Radical Right Parties in Europe. Cambridge University Press.

Mudde, Cas. 2013a. “Are Populists Friends or Foes of Constitutionalism?” The Foundation

for Law, Justice and Society.

Mudde, Cas. 2013b. “Three Decades of Populist Radical Right Parties in Western Europe:

So What?” European Journal of Political Research 52(1): 1–19.

Mudde, Cas, and Crist´obal Rovira Kaltwasser. 2012. Populism in Europe and the Americas:

Threat Or Corrective for Democracy? Cambridge University Press.

Muller, Jan-Werner. 2016. What Is Populism? University of Pennsylvania Press.

M¨uller-Rommel, Ferdinand. 1998. “Explaining the Electoral Success of Green Parties: A

Cross-National Analysis.” Environmental Politics 7(4): 145–154.

M¨uller, Wolfgang C., and Kaare Strøm. 2003. Coalition Governments in Western Europe.

Oxford University Press.

National Council of Austria. 1975. “Federal Act on the Functions, Financing and Election

Campaigning of Political Parties.” Government of Austria.

Nickell, Stephen. 1981. “Biases in Dynamic Models with Fixed Eﬀects.” Econometrica

49(6): 1417–1426.

Nyhuis, Dominic, and Carolina Plescia. 2017. “The Nonideological Component of Coalition

Preferences.” Party Politics 24(6): 686–697.

Oklobdzija, Stan. 2016. “Closing Down and Cashing In: Extremism and Political Fundrais-

ing.” State Politics & Policy Quarterly 17(2): 201–224.

Panagopoulos, Costas. 2011. Public Financing in American Elections. Temple University

Press.

Pappas, Takis. 2019. Populism and Liberal Democracy: A Comparative and Theoretical

Analysis. Oxford University Press.

128

Pemstein, Daniel, Kyle L Marquardt, Eitan Tzelgov, Yi-ting Wang, and Farhad Miri. 2015.
“The V-Dem Measurement Model: Latent Variable Analysis for Cross-National and Cross-
Temporal Expert-Coded Data.” Varieties of Democracy Institute Working Paper 21.

Ponce, Aldo F., and Susan E. Scarrow. 2011. “Who Gives? Partisan Donations in Europe.”

West European Politics 34(5): 997–1020.

Powell Jr, G Bingham. 1981. “Party Systems and Political System Performance: Voting
Participation, Government Stability and Mass Violence in Contemporary Democracies.”
American Political Science Review 75(04): 861–879.

Riker, William H. 1962. The Theory of Political Coalitions. Yale University Press.

Rodrik, Dani. 2018. “Populism and the Economics of Globalization.” Journal of International

Business Policy 1(1-2): 12–33.

Rooduijn, M., S. Van Kessel, C. Froio, A. Pirro, S. De Lange, D. Halikiopoulou, P. Lewis,
C. Mudde, and P. Taggart. 2019. The PopuList: An Overview of Populist, Far Right, Far
Left and Eurosceptic Parties in Europe.

Rooduijn, Matthijs, and Tjitske Akkerman. 2015. “Flank Attacks: Populism and Left-Right

Radicalism in Western Europe.” Party Politics 23(3): 193 204.

Rooduijn, Matthijs, Brian Burgoon, Erika J van Elsas, and Herman G van de Werfhorst.
2017. “Radical Distinction: Support for Radical Left and Radical Right Parties in Europe.”
European Union Politics 18(4): 536–559.

Rydgren, Jens. 2017. “Radical Right-Wing Parties in Europe What’s Populism Got to Do

with It?” Journal of Language and Politics 16(4): 485–496.

Rydgren, Jens. 2018. The Oxford Handbook of the Radical Right. Oxford University Press.

Sadurski, Wojciech. 2018.

“How Democracy Dies (in Poland): A Case Study of Anti-

Constitutional Populist Backsliding.” SSRN Electronic Journal.

Saito, Takaya, and Marc Rehmsmeier. 2015. “The Precision-Recall Plot Is More Informative
than the ROC Plot When Evaluating Binary Classiﬁers on Imbalanced Datasets.” PLoS
ONE 10(3): 1–21.

Sartori, Giovanni. 1976. Parties and Party Systems. Cambridge University Press.

Scartascini, Carlos, Cesi Cruz, and Philip Keefer. 2018. “The Database of Political Institu-

tions 2017 (DPI2017).” Inter-American Development Bank.

Singer, Matthew M. 2011. “Who Says “It’s the Economy?” Cross-National and Cross-
Individual Variation in the Salience of Economic Performance.” Comparative Political
Studies 44(3): 284–312.

129

Smithson, Michael, and Jay Verkuilen. 2006. “A Better Lemon Squeezer? Maximum-
Likelihood Regression with Beta-Distributed Dependent Variables.” Psychological Methods
11(1): 54–71.

Swank, Duane, and Hans-George Betz. 2003. “Globalization, the Welfare State and Right-

Wing Populism in Western Europe.” Socio-Economic Review 1(2): 215–245.

Urbinati, Nadia. 2019. “Political Theory of Populism.” Annual Review of Political Science

22: 111–127.

van Biezen, Ingrid, and Ekaterina R Rashkova. 2014. “Deterring New Party Entry? The Im-
pact of State Regulation on the Permeability of Party Systems.” Party Politics 20(6): 890–
903.

Van Der Brug, Wouter, Meindert Fennema, and Jean Tillie. 2000. “Anti-Immigrant Parties
in Europe: Ideological or Protest Vote?” European Journal of Political Research 37(1): 77–
102.

Visser, Mark, Marcel Lubbers, Gerbert Kraaykamp, and Eva Jaspers. 2013. “Support for
Radical Left Ideologies in Europe.” European Journal of Political Research 53(3): 541–558.

Vlaams Blok. 2004a. “Beginselverklaring.” Vlaams Blok.

Vlaams Blok. 2004b. “Een Toekomst Voor Vlaanderen.” Vlaams Blok.

Wells, Chris, Dhavan V. Shah, Jon C. Pevehouse, Jung Hwan Yang, Ayellet Pelled, Frederick
Boehm, Josephine Lukito, Shreenita Ghosh, and Jessica L. Schmidt. 2016. “How Trump
Drove Coverage to the Nomination: Hybrid Media Campaigning.” Political Communica-
tion 33(4): 669–676.

Williams, Christopher, and John Ishiyama. 2018. “Responding to the Left: The Eﬀect
of Far-Left Parties on Mainstream Party Euroskepticism.” Journal of Elections, Public
Opinion and Parties 28(4): 443–466.

Winter, Ralph K, and John R Bolton. 1973. Campaign Financing and Political Freedom.

Washington, DC: American Enterprise Institute for Public Policy Research.

Wooldridge, Jeﬀrey M. 2010. Econometric Analysis of Cross Section and Panel Data. Second

ed. MIT Press.

Wooldridge, Jeﬀrey M. 2012. Introductory Econometrics: A Modern Approach. Fifth ed.

Cengage Learning.

Zaslove, Andrej. 2004a. “Closing the Door? The Ideology and Impact of Radical Right
Populism on Immigration Policy in Austria and Italy.” Journal of Political Ideologies
9(1): 99–118.

Zaslove, Andrej. 2004b. “The Dark Side of European Politics: Unmasking the Radical Right.”

Journal of European Integration 26(1): 61–81.

130

Ziller, Conrad, and Thomas Sch¨ubel. 2015. ““The Pure People” versus “The Corrupt Elite?”
Political Corruption, Political Trust and the Success of Radical Right Parties in Europe.”
Journal of Elections, Public Opinion and Parties 25(3): 368–386.

131