C RI S ES O F D EM O C RA C Y

Is democracy in crisis? The current threats to democracy are not just

political: they are deeply embedded in the democracies of today, in
current economic, social, and cultural conditions. In Crises of
Democracy, Adam Przeworski presents a panorama of the political
situation throughout the world of established democracies, places it
in the context of misadventures of democratic regimes, and
speculates on the prospects. Our present state of knowledge does
not support facile conclusions. We “should not believe the flood of
writings that have all the answers.” Avoiding technical aspects, this
book is addressed not only to professional social scientists, but to
everyone concerned about the prospects of democracy.

adam przeworski is the Carroll and Milton Professor of

Politics and Economics at New York University. A member of the
American Academy of Arts and Sciences since 1991, he is the
recipient of the 1985 Socialist Review Book Award, the 1998
Gregory M. Luebbert Article Award, the 2001 Woodrow Wilson
Prize, the 2010 Lawrence Longley Article Award, the 2018 Sakip
Sabanci International Award, and the 2018 Juan Linz Prize. In 2010,
he received the Johan Skytte Prize. He recently published Why
Bother with Elections? (2018).
CRISES OF DEMOCRACY

ADAM PRZEWORSKI
New York University
University Printing House, Cambridge cb2 8b s , United Kingdom
One Liberty Plaza, 20th Floor, New York, n y 10006, USA
477 Williamstown Road, Port Melbourne, v ic 3207, Australia
314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre,
New Delhi – 110025, India
79 Anson Road, #06–04/06, Singapore 079906

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781108498807
d o i : 10.1017/9781108671019
© Adam Przeworski 2019
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2019
Printed in the United Kingdom by TJ International Ltd. Padstow Cornwall
A catalogue record for this publication is available from the British Library.
i sbn 978-1-108-49880-7 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of
URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate
 contents

List of Figures page vii

List of Tables ix

Preface xi

1. Introduction 1

Part I. The Past: Crises of Democracy 25

2. General Patterns 29

3. Some Stories 39

4. Lessons from History: What to Look For 78

Part II. The Present: What Is Happening? 81

5. The Signs 83

6. Potential Causes 103

7. Where to Seek Explanations? 123

8. What May Be Unprecedented? 133

v
 contents

Part III. The Future? 143

9. How Democracy Works 145

10. Subversion by Stealth 172

11. What Can and Cannot Happen? 192

References 207

Index 227

vi
 fi g u r e s

3.1. Putative location of parties on the Left–Right

and democratic–authoritarian space page 45
3.2. Unrest by year in Germany, 1919–33 48
3.3. Unrest by year in Chile, 1938–73 61
3.4. Unrest by year in France, 1945–70 74
3.5. Unrest by year in United States, 1919–2012 75
5.1. Proportion of parties that were the two top vote
winners around 1924 that remained in the top
two 85
5.2. Effective number of parties in the electorate
since 1960, in countries that were members of
the OECD as of 2000 86
5.3. Average electoral support for radical Right,
by year 90
5.4. Vote shares of parties by years in countries that
were members of the OECD before 2000 92
5.5. Turnout by year 93
5.6. Turnout and radical Right vote share in ten
developed democracies 94
5.7. Ideological distance between center parties,
by year 95
6.1. Rate of growth of per capita income by year of
countries that were members of the OECD
before 2000 104

vii
 list of figures

6.2. Average Gini coefficient of pre-fisc incomes in

Europe, Japan, Australia, New Zealand,
by year 105
6.3. Average labor share by year among countries
that were members of the OECD before
2000 105
6.4. Average employment by sector over time,
absolute numbers 106
6.5. Real household income at selected percentiles,
1967 to 2011 107
6.6. Average incomes of selected groups,
OECD-2000 countries excluding the United
States 108
6.7. Disconnect between productivity and a typical
worker’s compensation, 1948–2014 110
6.8. Productivity and wage index (G20 advanced
economies) 111
6.9. Union density by year in countries that were
members of the OECD before 2000 112
6.10. Democrats and Republicans more ideologically
divided than in the past 114
6.11. Immigration wasn’t always a partisan issue 115
6.12. European attitudes to immigrants: racial
differences 116
7.1. Wages and productivity in Germany and
France 124
9.1. Proportion of bills passed and riots 169

viii
 tables

2.1. Democracies which experienced at least two

alternations after 1918 and subsequently
fell page 30
2.2. Incidence of economic crises and survival of
democracy 31
2.3. Incidence of political crises and survival of
democracy 33
2.4. Some differences between democracies that fell
and did not before 2008 34
2.5. Probability of democratic breakdown, given the
number of governmental crises and institutional
systems 36
5.1. Share of votes of radical Right (countries that
were members of the OECD as of 2000) 91
8.1. Economic conditions in democracies that fell or
did not fall before 2008 and the post-2008 means
for democracies that survived 135
8.2. Some political features of countries that were
members of the OECD as of 2000, before and
after 2008 139

ix
 preface

Writing an academic book about current events is perilous.

The period between the moment the book is written and the
time it is read is long, while political life does not stop in-
between. Hence, much information contained below must be
read with the caveat “as of such and such date.” Yet if a book is
worth anything, the arguments and the conclusions should
survive the particular events that may have transpired in the
meantime. I say this without much confidence: the very event
that prompted me to plunge into this volume was something
I never expected, the victory of Donald Trump. Yet I think
I learned something in retrospect, namely, that the reasons to
worry about the current state of democracy in the United
States and in some European countries are much deeper than
the contingent events. Had Trump lost, many people who are
now rushing to write books similar to this one, myself
included, would have been occupied by other pursuits. Yet
the economic, social, and cultural conditions that brought
Trump to office would have been the same. This is what
I learned writing this text: that the causes of the current
discontent are deep, that they would not have been alleviated
by accidental events, and that we need to ask what if Clinton
had won or Brexit had lost, and what will happen if and when
whatever governments that are now in office in developed
democracies fail to improve the lives of people who had voted
for them? What then? Where should we seek solutions: in

xi
 preface

economic policies, in political reforms, in discursive strategies

of combating social fragmentation and racism? No answers to
such questions are obvious to me, so there is little I try to
persuade the readers about. All I can do is to formulate
questions, entertain possibilities, and invite the readers to
think together.
I present a panorama of the current political situation
across the world of well-established democracies, place it in
the context of past misadventures of democratic regimes, and
speculate about their prospects. I know that some readers will
be disappointed by how often I fail to arrive at firm conclu-
sions. But one should not believe the flood of writings that
have all the answers. I understand, and share, the quest to find
sense in what is happening around us, and the urge to think
that the diverse developments that surprise us must be some-
how related, that everything must have a cause. Yet establish-
ing what causes what and what matters most is often very
difficult and sometimes impossible. Particularly in our peri-
lous times, it is important to know what we do not know
before deciding how to act. Hence, I hope to encourage
skepticism among those who will read this book only because
they are concerned about the prospects of democracy. At the
same time, I hope that graduate students and my professional
colleagues will find here an agenda for research on questions
that are technically difficult and politically important.
The topic of this book concerns the dangers to
democracy lurking in the current economic, cultural, and
political situation. Yet the greatest danger we face is not to
democracy but to humanity, namely, that unless we do some-
thing drastic now, immediately, our children will be baked or

xii
 preface

flooded. If this danger materializes, all our concerns about

democracy will become moot. Tragically, this specter receives
only scant political attention, and this absence is reflected in
the pages that follow. Yet it casts an ominous shadow over
everything else we may care about.
Some people have already reacted to various parts of
this text, so the current version is indebted to comments by
Carlos Acuna, Jose Antonio Aguilar Rivera, Jess Benhabib,
Pierre Birnbaum, Bruce Bueno de Mesquita, Zhiyuan Cui,
Daniel Cukierman, Larry Diamond, John Dunn, Joan
Esteban, Roberto Gargarella, Stephen Holmes, John
Ferejohn, Joanne Fox-Przeworski, Fernando Limongi,
Zhaotian Luo, Boris Makarenko, Bernard Manin, Jose Maria
Maravall, Andrei Melville, Patricio Navia, Gloria Origgi,
Pasquale Pasquino, Molly Przeworski, John Roemer, Pacho
Sanchez-Cuenca, Aleksander Smolar, Willie Sonnleitner,
Milan Svolik, Juan Carlos Torre, Joshua Tucker, Jerzy
J. Wiatr, and three anonymous reviewers. I am particularly
indebted to John Ferejohn for forcing me to revise the analy-
tical framework.

xiii
 1

Introduction

The crisis consists precisely in the fact that the old is dying
and the new cannot be born; in this interregnum a great
variety of morbid symptoms appear.

(Antonio Gramsci, Prison Notebooks, c.1930: 275–6)

Something is happening. “Anti-establishment,” “anti-system,”

“anti-elite,” “populist” sentiments are exploding in many mature
democracies. After almost a century during which the same
parties dominated democratic politics, new parties are springing
up like mushrooms while the support for traditional ones is
dwindling. Electoral participation is declining in many countries
to historically unprecedented levels. Confidence in politicians,
parties, parliaments, and governments is falling. Even the sup-
port for democracy as a system of government has weakened.
Popular preferences about policies diverge sharply. Moreover,
the symptoms are not just political. Loss of confidence in institu-
tions extends to the media, banks, private corporations, even
churches. People with different political views, values, and cul-
tures increasingly view each other as enemies. They are willing to
do nasty things to each other.
Is democracy in crisis? Is this change epochal? Are we
living through an end of an era? It is easy to become alarmist, so
we need to maintain a perspective. Apocalyptic announcements
of an “end to” (Western Civilization, History, Democracy) or
“death of” (the State, Ideology, Nation-State) are perennial. Such

1
 introduction

claims are titillating but I cannot think of anything on this list

that did end or die. Not yielding to fears, a dose of skepticism,
must be the point of departure. The null hypothesis must be that
things come and go and there is nothing exceptional about the
present moment. After all, it may well be true that, as the
Hungarian Marxist Georg Lukács would have it, “crises are but
an intensification of everyday life of bourgeois society.” Just note
that the Harvard Widener library holds more than 23,600 books
published in the twentieth century in English containing the
word “crisis” (Graf and Jarausch 2017).
Yet many people fear that this time it is different, that
at least some established democracies are experiencing con-
ditions that are historically unprecedented, that democracy
may gradually deteriorate, “backslide,” or even not survive
under these conditions.

1.1 Crises of Democracy

What should we be looking for if we fear that democracy is
experiencing a crisis? To identify crises of democracy, we
need a conceptual apparatus: What is democracy? What is
a crisis? Is the crisis already here or is it only impending? If it
is already here, how do we recognize it? If it is not yet visible,
from what signs do we read the future?
We are repeatedly told that “Unless democracy is X or
generates X, . . .” The ellipsis is rarely spelled out, but it insin-
uates either that a particular system is not worthy of being called
a “democracy” unless some X is present or that democracy will
not endure unless some X is satisfied. The first claim is norma-
tive, even if it often hides as a definition. Skinner (1973: 303), for

2
 introduction

example, thinks that a system in which only some people rule

does not merit being called a “democracy,” even if it is
a competitive oligarchy. Rosanvallon (2009), in turn, claims
that “Now power is not considered fully democratic unless it is
submitted to the tests of control and validation at the same time
concurrent and complementary to the majoritarian expression.”
The second claim is empirical, namely, that democracy may not
endure unless some Xs are present (or absent). If democracy
requires some conditions – say J.S. Mill’s (1977: 99) “high wages
and universal reading” – just to function, then it is vulnerable to
breakdowns when these conditions are absent. A modicum of
economic welfare, some level of citizen’s confidence in political
institutions, or some minimal level of public order are the most
plausible candidates for such conditions.
Thus, one way to think is that democracy experi-
ences a crisis when some features which we consider as
definitional of democracy are absent. Consider a triad of
what Ginsburg and Huq (2018a) consider to be “the basic
predicates of democracy”: competitive elections, liberal
rights of speech and association, and the rule of law.
If we treat this triad as definitional, we get a ready-made
checklist of what we should be looking for to identify
crises of democracy: elections that are not competitive,
violations of rights, breakdowns of the rule of law. Yet if
we believe that democracy may not survive given some
particular situation, we may still be worried that it faces
a crisis even if no such violations are observed. We may
still have a checklist constructed by the definition but now
we also have a set of hypotheses that condition the survival
of democracy on some potential threats, and we are

3
 introduction

directed by these hypotheses to examine the particular

threats. If such hypotheses are valid, if the survival of
democracy depends on some aspects of its performance,
and democracy does not generate the required outcomes,
a crisis occurs – democracy is in crisis.
Note that some features may be treated alterna-
tively as definitional or as empirical. If one defines democ-
racy as Rosanvallon does, to include contramajoritarian
constraints on majority rule, “constitutional democracy,”
then the erosion of judicial independence is prima facie
evidence that something is wrong. But one may also reason
that if the judiciary is not independent, the government will
be free to do whatever it wants, violate the liberal right, or
make elections non-competitive. The problem with adding
adjectives to “democracy” is that not all good things must
go together. The more features – “electoral,” “liberal,”
“constitutional,” “representative,” “social” – we add to the
definition of democracy, the longer the checklist, and the
more crises we will discover. In contrast, the same list can
be treated as a set of empirical hypotheses. We can then
investigate empirically what are the conditions for elections
to be competitive or for rights to be observed or for the rule
of law to prevail. If it is true that elections are competitive
only if rights are observed and law rules, then taking any
one of these features as definitional and treating others as
“preconditions” is coextensive. If they are not coextensive,
then some kind of definitional minimalism is unavoidable:
we must choose one of the potential features as definitional
and treat others as hypothetical conditions under which the
selected feature is satisfied.

4
 introduction

Hence, what we would consider as crises and how we

should go about diagnosing them depends on how we think
about democracy. The view of democracy I adopt is “minim-
alist” and “electoralist”: democracy is a political arrangement
in which people select governments through elections and
have a reasonable possibility of removing incumbent govern-
ments they do not like (authors who held this view include
Schumpeter 1942, Popper 1962, and Bobbio 1987). Democracy
is simply a system in which incumbents lose elections and
leave when they lose. Hence, I investigate the possible threats
to elections becoming non-competitive or inconsequential for
whoever remains in power. To repeat, these threats may
include violations of the preconditions for contested elections
enumerated by Dahl (1971) – the liberal rights and the free-
doms – simply because without them the incumbent govern-
ment could not be defeated. They may also include
breakdowns of the rule of law and erosion of the independent
power of the judiciary, along with loss of confidence in repre-
sentative institutions (as in “representative democracy”),
acute inequality (as in “social democracy”), or the use of
repression to maintain public order (“liberal democracy”).
But I treat these violations as potential threats to the ability
of citizens to remove governments by elections, not as defini-
tional features of “democracy.”
The relation between “democracy” in the minimalist
sense and the “rule of law” is particularly complex. First, there
are both logical and empirical reasons to question whether
supra-majoritarian institutions, such as bicameralism or presi-
dential veto, or counter-majoritarian institutions, such as con-
stitutional courts or independent central banks, are necessary to

5
 introduction

support the rule of law. Gargarella (2003), for example, lists

several mechanisms by which a majority can and would want
to constrain itself even in the absence of such institutions.
As McGann (2006) observes, there are well-established democ-
racies, including the United Kingdom and Sweden, which have
neither a separation of powers nor judicial review of the con-
stitution, and yet in which majorities constrain themselves from
violating rights. Indeed, Dixit, Grossman, and Gull (2000: 533)
demonstrate logically that violations of rights are likely to be
more egregious in the presence of supra-majoritarian institu-
tions once a government enjoys supra-majority support.
Second, I put “rule of law” in quotation marks
because, as Sanchez-Cuenca (2003: 62) astutely put it,
“The law cannot rule. Ruling is an activity, and laws cannot
act.” What is typically seen as a relation between democracy
and the rule of law is in fact a relation between populated
institutions: governments and courts (Ferejohn and Pasquino
2003). Law “rules” when politicians and bureaucrats obey
judges, and whether politicians do or do not comply with
the instructions of constitutional justices is a contingent out-
come of their electoral incentives. Moreover, as will be seen
below, it is often next to impossible to determine if some
particular measures they adopt do or do not conform to
legal or constitutional norms, with individual judgments,
including those of constitutional justices, clouded by parti-
sanship. Under democracy, the only effective device for dis-
ciplining politicians are elections: as Dixit, Grossman, and
Gull (2000: 533) observe, “The ruling individuals must foresee
an appreciable chance that their power will come to an end . . .
And they must foresee a possibility of regaining power once it

6
 introduction

is lost.” There are two possibilities: (1) politicians (and

bureaucrats) obey judges because otherwise they would lose
elections, so that “the law” rules; (2) politicians do not obey
judges because otherwise they would lose elections –
a majority does not want politicians to listen to what the
judges tell them they can or cannot do. The rule of law is
violated but as long as politicians’ actions are motivated by the
fear of losing elections, the system is still democratic by the
minimalist criterion. Democracy is “illiberal” – a term made
fashionable by Zakaria (1997) and embraced by the Hungarian
Prime Minister, Viktor Orbán – but it is illiberal because
politicians expect that otherwise they would lose elections.
Yet, if politicians do not obey the judges even if a majority
would want them to because they do not fear elections, the
regime is not democratic.
Understood in this way, democracy is a mechanism
for processing conflicts. Political institutions manage conflicts
in an orderly way by structuring the way social antagonisms
are organized politically, absorbing whatever conflicts may
threaten public order, and regulating them according to some
rules. An institutional order prevails if only those political
forces that have institutionally constituted access to the repre-
sentative system engage in political activities, and if these
organizations have incentives to pursue their interests
through the institutions and incentives to temporarily tolerate
unfavorable outcomes. Specifically, conflicts are orderly if all
political forces expect that they may achieve something, at the
present or at least in some not too distant future, by proces-
sing their interests within the institutional framework while
they see little to be gained by actions outside the institutional

7
 introduction

realm. Hence, democracy works well when whatever the

conflicts that arise in society are channeled into and processed
through the institutional framework – most importantly elec-
tions, but also collective bargaining systems, courts, and pub-
lic bureaucracies – without preventing anyone from gaining
access to these institutions just because of the substance of
their demands. To put it succinctly, democracy works when
political conflicts are processed in liberty and civil peace.
The conflicts that divide a particular society at
a particular time may be more or less intense and may divide
the society along different lines depending on whether they
concern economic interests, cultural values, symbolic issues,
or just fleeting passions. Their forms, their subjects, and their
intensity depend on the actions of governments and the
alternatives offered by competing political forces. The stakes
entailed in institutionalized conflicts do not simply reflect the
intensity of antagonisms that arise in a society. Institutional
frameworks shape the ways in which social conflicts become
politically organized, some increasing and others limiting the
stakes in the outcomes of political competition. I argue below
(see Chapter 9) that democracy works well when the stakes
entailed in institutionalized conflicts are neither too small or
too large (for a technical version of this argument, see
Przeworski, Rivero, and Xi 2015). The stakes are too low
when results of elections have no consequences for people’s
lives. They are too high when results of elections inflict intol-
erable costs on the losers. When people believe that results of
elections do not make a difference in their lives, they turn
against “das System,” as in Weimar Germany. When the
electoral losers discover that the government pursues policies

8
 introduction

that significantly hurt their interests or values, they become

willing to resist the government by all – including violent –
means, as did the bourgeoisie in Chile under President
Allende. Hence, democracy works when something is at
stake in elections but not too much is at stake.
An often overlooked emphasis of Schumpeter’s (1942:
chapter 23, section 2) “minimalist” view of democracy is that
governments must be able to govern and must govern com-
petently. Later I delve into some historical periods in which
the institutional framework made it difficult for governments
to be able to govern, either because the electoral system led to
government instability, as in Weimar Germany and the
French Fourth Republic, or because the system of separation
of powers generated a stalemate between the executive and
the legislature, as in Allende’s Chile. To govern effectively,
governments must satisfy a majority yet not ignore the views
of intense minorities. When conflicts are intense and a society
is highly polarized, finding policies acceptable to all major
political forces is difficult and may be impossible. There are
limits to what even the best-intentioned and competent gov-
ernments can do.
If this is the standard, when is democracy “in
crisis”? The very word “crisis” originates from ancient
Greek, where it meant “decision.” Crises are situations
that cannot last, in which something must be decided.
They emerge when the status quo is untenable and nothing
has yet replaced it. This is what we mean when we say that
“the situation reached a crisis point”: when doctors say
someone is in a crisis, they mean that the patient will
either recover or die but cannot remain in the current

9
 introduction

state. Crises may be more or less acute: in some a turning

point may be imminent but some crises may linger indefi-
nitely, with all the morbid symptoms.
The intuition of crises conveyed by Gramsci’s motto
is that the current situation is in some ways untenable, that
some threat to democracy has already materialized, yet the
status quo democratic institutions remain in place. While
Marx (1979 [1859]: 43–4) thought that “new superior rela-
tions of production never replace older ones before the
material conditions for their existence have matured within
the framework of the old society,” nothing guarantees that
when the status quo institutions malfunction, some other
institution would descend on earth as a deux ex machina.
What happens when the status quo institutions do not gen-
erate desirable outcomes depends on their properties and on
the alternative institutions – would any do better? – on
exogenous conditions, and on the actions of the relevant
political forces under these conditions. That a disaster is
unfolding under the status quo institutions need not imply
that some other institutions would do better: this was
Winston Churchill’s view of democracy. But even if some
alternatives are feasible, it may well be that given the rela-
tions of political power under the extant institutions, the
situation would linger on and on. Crises are then situations
in which the condition under the status quo institutions is
some kind of a disaster: no change occurs, but it may. This is
what we will be looking for below: whether the current
situation is in some ways threatening and whether there
are signs that the traditional representative institutions are
being affected.

10
 introduction

“Crises of capitalism” deserve a separate comment.

Capitalism – an institution that combines private ownership
of most productive resources with the allocation of resources
and distribution of incomes by markets – periodically gener-
ates “crises,” understood as periods in which incomes fall
sharply and either inflation flares or unemployment soars or
both, as during the “stagflation crisis” of the 1970s,
a combination of high inflation with high unemployment
caused by a jump in prices of raw materials (Bruno and
Sachs 1985). But are economic crises “crises of capitalism”?
They would be if one expects that when the economy is in the
doldrums, capitalism will or at least may collapse. But an
implosion of capitalism is not in the realm of the possible.
When a famous leftist economist, Michal Kalecki (1972
[1932]), asked in 1932, at the worst moment of the Great
Depression, “Is a capitalist exit from the crisis possible?,” his
argument was that, even if the adjustments required to exit
from economic crises are painful and may take time, capital-
ism is a self-correcting system. Prices and wages may be sticky
but eventually supply and demand adjust, the crisis is over,
and capitalism is still here. It can be abolished by a political
revolution – a possibility Kalecki did entertain and
Communists implemented – but not implode. The general
lesson for understanding crises is that some institutions are
impervious to the outcomes they generate, so that crises
which occur under them do not turn into crises of the
institutions.
Disasters that occur under democracy, however, may
turn into crises of democracy. Borrowing their list from
Habermas (1973: 49), disasters are situations in which

11
 introduction

– the economic system does not produce the requisite quan-

tity of consumable values, or;
– the administrative system does not produce the requisite
quantity of rational decisions, or;
– the legitimation system does not provide the requisite
quantity of generalized motivations, or;
– the socio-cultural system does not generate the requisite
quantity of action-motivating meaning.

This list, however, is too abstract to guide research.

The observable candidates for disasters are economic crises,
intense conflicts in society, and political paralyses, situations
in which the government is unable to govern given the parti-
cular form of democratic institutions.
When we think that the situation is in some way
threatening, we look for signals – harbingers of change.
Several countries, ranging from Canada in 1931–3 to Uruguay in
2001–3, experienced profound economic crises with almost no
political repercussions and no signals of democracy being
weakened. Yet in some situations crises in other realms –
whether economic, cultural, or autonomously political (say
corruption scandals, as in Italy in 1993 or in Brazil now) –
manifestly weaken the established democratic institutions.
The visible signals that democracy is in crisis include
a sudden loss of support for established parties, withdrawal of
popular confidence in democratic institutions and politicians,
overt conflicts over democratic institutions, or an incapacity of
governments to maintain public order without repression.
Perhaps the most tangible sign of a crisis is a breakdown of
public order: in the words of Linz (1978: 54), “The most serious

12
 introduction

crises are those in which the maintenance of public order

becomes impossible within a democratic framework.”
Democracy is in crisis when fists, stones, or bullets replace
ballots. Either the incumbents make it impossible for the
opposition to remove them from office and the opposition
has no other avenues than resistance, or the opposition does
not recognize the legitimacy of the government and the gov-
ernment defends itself by repression, or antagonistic political
groups do not accept the outcomes of the institutional interplay
of interests and revert to direct, often violent, confrontations.
When such situations extend over time, public order breaks
down, everyday life becomes paralyzed, and violence tends to
spiral. Such crises become mortal when the design of demo-
cratic institutions generates institutional stalemates, as in
Weimar Germany or in Chile under President Allende.
Institutions may generate outcomes that are intoler-
able for some and wonderful for others. Moreover, people
may differ in their normative attachments: some valuing
liberty more than order, others being willing to sacrifice it
for the promise that trains would run on time (Mussolini
promised they would under fascism, but they did not).
Hence, to understand crises it is necessary to think in terms
of conflicting interests and values. The poor are dissatisfied
when their incomes stagnate, the rich enjoy their wealth and
power, while some people, whether poor or wealthy, may care
about political and economic inequality per se. Solutions to
crises are likely to be controversial and subject to political
conflicts. They depend on what the relevant political actors do
under the circumstances. To this extent, therefore, they are
indeterminate ex ante. Will a reduction of economic

13
 introduction

inequalities restore the political vitality of democracy? Will

restrictions on immigration appease radical Right senti-
ments? Will some tinkering with representative institutions
restore confidence in these institutions? Because the actors in
crisis may choose different courses of actions, with different
consequences, the best we can strive to determine is what is
and what is not possible, perhaps with some cavalier forecasts
about what is most likely.
What, then, are the possible outcomes of crises? Not all
crises are mortal: some end in restoring the status quo ante,
a return to “normalcy.” The sources of a crisis sometimes con-
veniently disappear. Democracy may be in a crisis when society
experiences an economic disaster, but the crisis may dissipate
when prosperity returns. Some crises can be overcome by partial
reforms. The group that benefits under extant institutions can
make concessions to the groups that suffer most under them.
Such concessions have to be credible, because otherwise these
groups will expect that they would be withdrawn once the crisis
is over. Hence, concessions must entail some institutional
reforms: the classical example is the extension of suffrage to
the lower classes, which neutralized the threat of revolution by
changing the income location of the decisive voter (Acemoglu
and Robinson 2000). Yet when we think about democracy what
we fear is the prospect that some political forces would success-
fully claim that the only way to remedy some already occurring
disasters – economic crises, deep-rooted divisions in society,
breakdown of public order – is to abandon political liberty,
unite under a strong leader, and repress pluralism of opinions,
in short autocracy, authoritarianism, or dictatorship, whatever
one wants to call it. The impending cataclysm is that democracy

14
 introduction

would either collapse outright or gradually erode beyond the

point of no return.
The specter that haunts us today, I believe, is the last
possibility: a gradual, almost imperceptible, erosion of demo-
cratic institutions and norms, subversion of democracy by
stealth, “the use of legal mechanisms that exist in regimes with
favorable democratic credentials for anti-democratic ends”
(Varol 2015). Without manifest signs that democracy has broken
down, the line becomes thin, as evidenced by labels such as
“electoral authoritarianism” (Schedler 2006), “competitive
authoritarianism” (Levitsky and Way 2010), “illiberal democ-
racy” (Zakaria 1997), or “hybrid regimes” (Karl 1995, Diamond
2002). “Backsliding,” “deconsolidation,” or “retrogression” need
not entail violations of constitutionality and yet gradually
destroy democratic institutions.
To summarize this concept of “crisis of democracy,”
think schematically as follows. Given some exogenous shocks,
democracy generates some outcomes, positively or negatively
evaluated by people with heterogeneous preferences over
these outcomes and over the democratic institutions per se.
Outcomes that threaten the continued existence of the tradi-
tional democratic institutions constitute “disasters.” Whether
a particular situation qualifies as a crisis must be read from
some manifest signals that democratic institutions are under
threat. We are attentive to such signals because they may
constitute harbingers of democratic collapse or gradual ero-
sion. Yet the potential solutions to crises may include restora-
tion of the institutional status quo, some partial reforms of
traditional representative institutions that still preserve
democracy, as well as its either abrupt or gradual destruction.

15
 introduction

Why would democracies be vulnerable to crises? One

must not forget that democracy is but a speck of human
history, recent and still rare. It was born only in 1788, when
the first national-level election based on individual suffrage
took place in the United States; the first time in history that
the helm of the government changed as a result of an election
was in 1801, also in the United States. Use of force – coups and
civil wars – remained frequent: between 1788 and 2008 poli-
tical power changed hands as a result of 544 elections and 577
coups. Electoral defeats of those in power were rare until very
recently and peaceful changes of governments even less fre-
quent: only about one in five national elections resulted in the
defeat of incumbents and even fewer in a peaceful change in
office. As of today, sixty-eight countries, including the two
behemoths, China and Russia, have never experienced
a change in office between parties as a result of an election.
Democracy is a historical phenomenon. It developed under
specific conditions. It survived in some countries as these
conditions evolved, but can it survive under all conditions?
Two structural conditions, I think, deserve special
attention. The first is that political equality, which democracy
is supposed to be based on, coexists uneasily with capitalism,
a system of economic inequality. The second is the sheer quest
for political power, whether or not based on economic
interests.

1.2 Democracy and Capitalism

The relation between democracy and capitalism is subject to
contrasting views. One claims a natural affinity of “economic

16
 introduction

freedom” and “political freedom.” Economic freedom means

that people can decide what to do with their property and their
labor endowments. Political freedom means that they can pub-
licize their opinions and participate in choosing how and by
whom they will be governed. But equating the concepts of
“freedom” in the two realms is just a play on words. Looking
into history shows that we should be surprised by the coexis-
tence of capitalism and democracy. In societies in which only
some people enjoy productive property and in which incomes
are unequally distributed by markets, political equality com-
bined with majority rule presents a threat to property. Indeed,
beginning with Henry Ireton’s speech in the franchise debate at
Putney in 1647, almost everyone had thought that they could not
coexist. The English conservative historian and politician
Thomas Macaulay (1900: 263) vividly summarized in 1842 the
danger presented to property by universal suffrage:

The essence of the Charter is universal suffrage. If you

withhold that, it matters not very much what else you
grant. If you grant that, it matters not at all what else you
withhold. If you grant that, the country is lost . . . My firm
conviction is that, in our country, universal suffrage is
incompatible, not only with this or that form of
government, and with everything for the sake of which
government exists; that it is incompatible with property
and that it is consequently incompatible with civilization.

Nine years later, from the other extreme of the poli-

tical spectrum, Karl Marx (1952: 62) expressed the same con-
viction that private property and universal suffrage are
incompatible:

17
 introduction

The classes whose social slavery the constitution is to

perpetuate, proletariat, peasantry, petty bourgeoisie, it [the
constitution] puts in possession of political power through
universal suffrage. And from the class whose old social
power it sanctions, the bourgeoisie, it withdraws the
political guarantees of this power. It forces the political rule
of the bourgeoisie into democratic conditions, which at
every moment jeopardize the very foundations of
bourgeois society. From the ones it demands that they
should not go forward from political to social
emancipation; from the others they should not go back
from social to political restoration.

The combination of democracy and capitalism was

thus for Marx an inherently unstable form of organization of
society, “only the political form of revolution of bourgeois
society and not its conservative form of life” (1934 [1852]: 18),
“only a spasmodic, exceptional state of things . . . impossible
as the normal form of society” (1971 [1872]: 198).
These dire predictions turned out to be false. In some –
specifically thirteen countries – democracy and capitalism
coexisted without interruptions for at least a century, and in
many other countries for shorter but nevertheless extended
periods, most of which continue today. Working-class parties
that had hoped to abolish the private property of productive
resources realized that this goal is unfeasible, and learned to
value democracy and to administer capitalist economies when-
ever elections brought them into office. Trade unions, also
originally viewed as a mortal threat to capitalism, learned to
moderate their demands. The outcome was a compromise:
working-class parties and trade unions consented to capitalism,

18
 introduction

while bourgeois political parties and organizations of employ-

ers accepted some redistribution of income. Governments
learned to organize this compromise: regulate working condi-
tions, develop social insurance programs, and equalize oppor-
tunities, while promoting investment and counteracting
economic cycles (Przeworski 1986).
Yet perhaps this compromise is now broken. Unions
lost much of their capacity to organize and discipline workers
and with it their monopoly power. Socialist parties lost their
class roots and with them their ideological as well as policy
distinctiveness. The most visible effect of these changes is the
sharp decline in the share of incomes from employment in the
value added and, at least in the Anglo-Saxon countries, a steep
increase of income inequality. Combined with a slowdown of
growth, rising inequality causes many incomes to stagnate
and income mobility to decline.
Is the coexistence of democracy and capitalism con-
ditional on a continual improvement of material conditions
of broad sectors of the population, either because of growth or
because of increasing equality? History indicates that democ-
racies are solidly entrenched in economically developed
countries and impervious to economic as well as other crises,
even of a large magnitude. But is history a reliable guide to the
future?

1.3 Democracy and the Quest for Power

The second reason democracies may experience crises is
inherent in political competition. The dream of all politicians
is to conquer power and to hold on to it forever. It is

19
 introduction

unreasonable to expect that competing parties would abstain

from doing whatever they can do to enhance their electoral
advantage, and incumbents have all kinds of instruments to
defend themselves from the voice of the people. They are able
to consolidate their advantage because they constitute
a legislative majority and because they direct public bureau-
cracies. Although at times they are constrained by indepen-
dent courts, control over legislation grants incumbents an
opportunity to adopt legal regulation in their favor: just
think of voter registration, manipulation of electoral systems,
or gerrymandering. The courts or some other independent
bodies may invalidate some such attempts but not always
have reasons or the will to do so: there are many ways to
carve districts, each with electoral consequences, which are
not blatantly discriminatory. In turn, as principals of osten-
sibly non-partisan bureaucracies, incumbents can instrumen-
talize them for partisan purposes. Control over the
apparatuses of repression plays a particularly important role
in undermining all or some opposition. Exchange of favors
for financial resources is yet another source of advantage.
And, when all else fails, fraud is the last resort.
The question is why some political leaders use these
methods while others are content with letting the people
decide and being willing to leave office when people do so
decide. Their motives matter and so do the constraints. When
political parties are highly ideological, when they believe that
essential issues or values are at stake, they see their opponents
as enemies who must be prevented from coming to office by
any means. In Poland the ruling party, PiS (Law and Justice),
believes that the very values that constitute Poland as

20
 introduction

a Christian Nation are at stake and all their opponents are

“traitors.” In Hungary, President Orbán thinks that what is at
stake is whether “Europe will remain the continent for
Europeans.” Hence, both attempt to control the media,
restrict freedom of association, pack state agencies with
their partisan supporters, and toy with electoral rules. These
actions are intended to relax the electoral constraints they
face, and to make an electoral victory of the opposition next to
impossible. Yet they still face political, rather than narrowly
electoral, constraint: various forms of popular resistance, such
as mass demonstrations, political strikes, or riots. They face
the danger that political conflicts could spill out of institu-
tional bounds, resulting in a breakdown of public order. They
may or may not take this risk, and if they do, democracy is in
crisis.

1.4 A Preview
How then should we go about determining if democracy is
presently in crisis, or at least if a crisis is impending?
To look into the future, to identify the possibilities
latent in the current situation, we first need to see if we can
learn something from the past. Under what conditions did
democratic institutions fail to absorb and peacefully regulate
conflicts? To answer this question, Part I summarizes the
historical experience of all democracies that have been at
one time or another consolidated, in the sense of having
experienced at least two peaceful alternations in office that
resulted from elections, comparing some observable condi-
tions of the democracies that fell and those that survived. Such

21
 introduction

comparisons, however, are inevitably static, while the out-

comes that emerge under any conditions are highly contin-
gent, depending on who does what when. To develop
intuitions, I delve in more detail into four cases: the Weimar
Republic between 1928 and 1933, and Chile between 1970 and
1973, are two flagrant instances in which democracy suc-
cumbed, while France and the United States in the 1960s are
cases of political repression and breakdown of order that were
resolved institutionally.
Yet history does not speak for itself. Can we trust its
lessons? Lessons from history are relatively reliable when
current conditions imitate those observed in some past, but
iffy when they are unprecedented (King and Zheng 2007).
Hence, to see if history can be our guide, we need to compare
the current situation with those of the past. Do the current
conditions resemble those of democracies that fell or of those
that survived? Or are they unprecedented? Some aspects of
the current situation are new, in particular a rapid destabili-
zation of traditional party systems. So is the stagnation of low
incomes as well as the erosion of the belief in material pro-
gress. But causal links are far from obvious. Is the current
political conjuncture driven by economic trends or by cultural
transformations, or is it autonomous from changes in the
economy and society? At what level should we seek explana-
tions: general trends, such as globalization, or specific situa-
tions of particular individuals, say those who fear losing
decently paying jobs? These are the questions considered in
Part II.
To assess the prospects for the future, we need to
understand how democracy works when it works well,

22
 introduction

which is the subject of the theoretical chapter that opens Part

III. With this understanding, we can consider the foreboding
and uncharted possibility of a gradual erosion of democracy,
its subversion by elected governments. Finally, even if we
cannot tell what is most likely to ensue, we can at least
speculate about what is and what is not possible. Can it
happen here?

23
Part I

The Past: Crises of Democracy

To see what we can learn from history, we need to examine
the experience of democracies that had during some periods
functioned according to institutional rules. These are democ-
racies in which control over governments changed as a result
of elections at least twice without their result being resisted by
force. The reason for limiting the cases in this way is that we
need to examine democracies in which competing parties had
learned that losing elections is not a disaster, that one can lose
and come back to office again, and in which the political
forces behind the electoral parties had a chance to see that
they could protect or advance their interests by directing their
efforts within the institutional framework. The number of
such democracies is quite large: since 1918, there were eighty-
eight democracies that satisfy the criterion of having had at
least two alternations in office.
Thirteen of these democracies collapsed in a palpable
way. Note, however, that the line dividing democracies from
non-democracies, or whatever one wants to call them, is not
always clear. Even thirty years ago, when Alvarez et al. (1996)
tried to classify regimes as democracies and dictatorships,
there was a class of cases about which it was impossible to
tell. They were epitomized by Botswana, a country in which
all the liberal freedoms seemed to be respected but the same
party always won elections, for thirty years and by now for
almost sixty. The solution to this difficulty that subsequently

25
 the past: crises of democracy

became fashionable was to use a trichotomous classification,

introducing “hybrid regimes,” “semi-authoritarianism,” or
“electoral authoritarianism,” but these labels just cover up
the fact that there are situations which we do not know how
to classify. Now, as regimes that allow some opposition but
still assure themselves of winning elections proliferate rapidly,
the problem has become much worse. The central point of
Chapter 10 is that when democracies “backslide” there is no
clear line to cross. I look at classifications of Venezuela in
several data sources, only to learn that no one agrees whether
it is still a democracy and if not as of when. So in the end there
are some cases in which the collapse of democracy is manifest,
marked by some discrete event, but there are some in which
democracy slides down a continuous slope, so not only do we
not have discrete markers but we can reasonably disagree
about whether a particular regime is still democratic or
already past the point of no return. I consider here only
those breakdowns of democracy that have been marked by
some discrete, manifest events, and discuss gradual erosion of
democracy in Chapter 10.
Following Magaloni (2017), among the manifest
deaths of democracy we should still distinguish different
ways in which democracies fall: some are destroyed by mili-
tary coups while others die when politicians who accede to
office in a legal way succeed in removing all institutional
checks on their power and in eradicating all organized oppo-
sition. Coups – at least those that lead to a death of democ-
racy, such as in Chile in 1973 – are conspicuous events.
Usurpations of power by incumbents may be slow and gra-
dual, but in many cases the breakpoints are obvious. The legal

26
 the past: crises of democracy

end of the Weimar democracy was signaled by a discrete

event: the authorization by the Reichstag (the parliament)
on March 23, 1933 for the government to act in non-
constitutional ways. In Estonia, the breakdown of democracy
was marked by the declaration of martial law and the post-
ponement of elections by the prime minister, Konstantin Pats,
on March 12, 1934.
To see if history can direct us to what we should be
paying attention to in analyzing the current situation,
I compare some observable conditions of the democracies
that survived and fell in the past. I focus in particular on the
effects of different kinds of crises: economic, cultural, or
political. These comparisons, however, tell us little about the
contingent dynamics of crises, the denouement of events
under the various conditions. Hence, I delve into some pro-
minent crises under which democracy fell or survived. Finally,
I search for lessons, asking what we should look for in the
current political situation if the past were to repeat itself.

27
 2

General Patterns

The aim of this chapter is to examine whether the collapses

and survival of democracy are associated with some observa-
ble differences between countries that experienced such out-
comes. Obviously I am not the first to do so: the literature on
the topic is voluminous and by now technically sophisticated.
Almost everyone agrees that democracies are unlikely to
collapse in economically developed countries; there is strong
evidence that in less developed countries democracies are
vulnerable to income inequality, and that the longer they
have been around the more likely they are to still exist.
Whether anything else matters – institutional frameworks,
ethno-linguistic or religious fractionalization, educational
levels, etc. (the list is long) – is more controversial. While
limiting the scope to democracies that at some time became
consolidated, the analysis here reproduces some of these
findings. My particular interest is in the effect of different
kinds of crises: economic, broadly political, and narrowly
governmental. The statistical analyses presented below are
purely descriptive, so no inferences about causality should
be drawn. The role of this chapter in the context of the book is
only to arrive at a list of factors that may inform us regarding
what to look for in the present situation.
The consolidated democracies that collapsed are
listed in Table 2.1. Those that are still around include four
African countries (Benin, Cape Verde, Ghana, and

29
 the past: crises of democracy

Table 2.1 Democracies which experienced at least two alternations

after 1918 and subsequently fell

Country Year second Year fell Alternations Mode

Germany 1928 1933 3 From above

Estonia 1932 1934 2 From above
Greece 1951 1967 2 Coup
Chile 1952 1973 4 Coup
Sri Lanka 1960 1977 3 From above
Philippines 1961 1965 2 From above
Solomon Islands 1989 2000 2 Coup
Peru 1990 1990 2 Legal
Ecuador 1992 2000 3 Coup
Thailand 1996 2006 3 Coup
Pakistan 1997 1999 2 Coup
Bangladesh 2001 2007 2 From above
Honduras 2005 2009 2 Coup

Note: In order of dates of the second alternation. “Alternations” is the

number of alternations by the time democracy fell. Mode: “From above” if
power was usurped by a chief executive who came into office in
a constitutional way, “Coup” if democracy fell because of a military coup.
Only the breakdowns marked by discrete events are included. Source: Boix,
Miller, and Rosato (2012) for regime classification, own research for the
mode.

Mauritius), eleven Central and South American countries,

several Caribbean and small Pacific islands, India,
Indonesia, Taiwan, and all the current members of the OECD.
The remaining democracies survived, but it does
not mean that they did not face some kind of disaster.
Several countries where democracy survived experienced
serious economic crises – which I take as periods in which
per capita incomes fell by at least 10 percent during

30
 general patterns

Table 2.2 Incidence of economic crises and survival of

democracy

Crises Survived Fell Total Incidence

None 66 10 76 1/7.6
Yes 9 3 12 1/4.0
Total 75 13 88 1/6.8

Note: Crises are situations in which per capita income fell

by at least 10 percent during consecutive years. Cell entries
are numbers of countries. Sources: Maddison (2011) for
income data, Boix, Miller, and Rosato (2012) for regime
classification.

consecutive years – without major political repercussions:

Canada (1931–3), the United States (1932–4 and 1946–8), the
United Kingdom (1946–7), Jamaica (1976–8), Costa Rica
(1982–3), Finland (1992–3), Venezuela (1980–5), and
Uruguay (2001–3). Indeed, only three consolidated democ-
racies fell following thus-defined economic crises: Germany
in 1933, Ecuador in 2000, and Peru in 1990. Hence, trans-
formations of economic into political crises are far from
automatic. Lindvall (2014) compared the electoral effects of
the economic crises of 1929–32 and 2008–11, finding that
they were very similar. In both periods incumbents lost
votes, and in both periods elections that occurred soon
after the crisis favored the Right, while those that occurred
later generated no swing or a swing to the Left. Hence, if
a democracy survives an economic crisis, its electoral effects
are short-lived.
Several democracies overcame political crises.
In Table 2.2, I count as political crises situations: in which

31
 the past: crises of democracy

there are conflicting claims as to who should govern; where

the competent courts declare the government to have violated
the constitution or its members to be legally unfit to continue
serving (typically accused of corruption); when a conflict
between separate powers renders the government unable to
function; or where a government is forced to resign or repress
the opposition either because of popular pressures or a threat
from the military, rather than by a decision of the competent
body (either the legislature or the courts). Extended negotia-
tions over government formation in parliamentary systems –
the record holder as of now is Belgium, where negotiations
took 353 days in 2011 – are not considered as crises, and neither
are impeachment procedures in presidential systems if suc-
cession follows constitutional rules in a timely manner. Such
crises occurred in ten consolidated democracies that did
survive. In chronological order, they erupted in France in
1958, the United States in 1973–4, Jamaica in 1983, the
Dominican Republic in 1994, Guyana in 1997, Argentina in
2001–3, Romania in 2007, Ukraine in 2014, Mauritius in 2014,
and Guatemala in 2014–15. Most of these crises were termi-
nated by a subsequent election. Notably, the institutional
status quo ante was restored in all cases except for France,
where the crisis resulted in a change of the constitution. Yet,
as Table 2.3 shows, such political crises are dangerous: five out
of fifteen democracies that experienced them collapsed.
What, then, are the differences between those democ-
racies that fell and those that continue to function, including
those that experienced political and economic crises?
Unfortunately, systematic information is scarce. But some
patterns stand out.

32
 general patterns

Table 2.3 Incidence of political crises and survival of

democracy

Crises Survived Fell Total Incidence

None 65 8 73 1/9.1
Yes 10 5 15 1/3.0
Total 75 13 88 1/6.8

Note: Crises as defined in the body of the text. Cell entries

are numbers of countries. Sources: Own research for
crises, Boix, Miller, and Rosato (2012) for regime
classification.

The most striking difference, which will not surprise

students of regime transitions, is in per capita income.
We have known for some time that democracies are impreg-
nable in economically developed countries. Przeworski and
Limongi (1997) observed that the probability of democracy
surviving increases steeply as income increases, and that no
democracy in a country with per capita income higher than
that of Argentina in 1976 has ever collapsed, although it did
collapse in Thailand in 2006 with income slightly higher.
The general pattern, however, remains the same and, as
Table 2.4 shows, it holds strongly for the consolidated democ-
racies as well. Sixty-nine consolidated democracies lasted
a total of 1957 years with incomes higher than that of
Thailand in 2006 and none of them fell.
Economic growth was much slower in democracies that
fell than in those that survived. The difference is big: economies
of countries where democracy fell were almost completely stag-
nant. Using a different source, Maddison (2011), which includes

33
 the past: crises of democracy

Table 2.4 Some differences between democracies that fell and did not
before 2008

Survived Survived Fell Fell Probabilitye

N Mean N Mean
a
GDP/cap 1,484 18,012 103 5,770 1.00
Growtha 1,471 0.031 103 0.011 1.00
Labor sharea 1,397 0.60 96 0.50 1.00
Gini grossb 1,148 42.6 64 44.6 1.00
Gini netb 1,148 33.8 64 44.6 1.00
Regimec 1,739 0.55 124 1.18 1.00
Government 1,689 0.17 140 0.44 1.00
crisesd
Riotsd 1,689 0.53 140 0.73 0.89
d
Strikes 1,689 0.13 140 0.26 0.99
d
Demonstrations 1,689 0.64 140 0.63 0.49

Note: Cell entries are numbers of annual observations (until 2014) and the
average values of the particular variables. (a) From PWT9.0. (b) Gini
coefficients of gross and net incomes, from SWIID (2014). (c) Regime = 0 if
parliamentary, Regime = 1 if mixed, Regime = 2 if presidential, from
Cheibub, Gandhi, and Vreeland (2010). (d) From CNTSDA, Wilson (2017).
(e) The probability that the difference of the means is not a matter of
chance. Based on t-test with unequal variances.

an earlier period and ends in 2008, shows an even larger differ-

ence. Hence, even if short-term economic crises do not threaten
democracy, protracted stagnation of incomes may.
While the number of observations is low, it is also
clear that democracies that fell had a more unequal distribu-
tion of income. In functional terms, labor share was lower in
democracies that fell. The Gini index of gross (market, pre-
fisc) incomes was higher among democracies that fell, as was
the inequality of net (post-fisc) incomes. The comparison of

34
 general patterns

these indices indicates that democracies that survived were

those which redistributed a fair part of incomes, while those
that fell redistributed none.
Moving beyond the economy, another striking differ-
ence is among the democratic systems of institutions: parlia-
mentary, mixed (semi-presidential), and presidential.
The weakness of presidential democracies is manifest. There
were forty-four consolidated parliamentary democracies and
of those six fell, 1 in 7.3; sixteen mixed (or semi-presidential)
systems of which one fell, and twenty-six presidential ones of
which six fell, 1 in 3.7. This difference need not be due to these
systems per se: Cheibub (2007: chapter 6) shows that presi-
dential democracies are brittle when they succeed military,
but not civilian, dictatorships. Yet, given the traditional role of
the military in Latin America, presidential democracies were
particularly vulnerable to crises of governance. The major
difference between parliamentary and presidential systems is
that the former have a built-in mechanism for changing
governments that cannot cope with crises and become
unpopular: the vote of non-confidence in the government.
In presidential systems, however, the chief executive is elected
for a fixed term and appoints his or her cabinet, at most
subject to congressional approval. Unless the president com-
mits illegal acts, he remains in office regardless of his capacity
to govern, even when his popularity reaches single-digit num-
bers and he has no support in the legislature.
The frequency of government crises is also sharply
higher in those democracies that fell. This information has to
be taken with a grain of salt: its source, the Cross-National
Time-Series Data Archive (CNTSDA) (Wilson 2017) provides

35
 the past: crises of democracy

Table 2.5 Probability of democratic breakdown, given the number of

governmental crises and institutional systems

Crisesa Parliamentary Presidential Total

0 0.030 (1,213) 0.097 (496) 0.048 (2,184)

1 0.045 (157) 0.320 (37) 0.087 (242)
2 0.120 (33) 0.333 (9) 0.158 (57)
>2 0.00 (7) 0.430 (7) 0.221 (23)
Total 0.034 (1,410) 0.120 (549) 0.057 (2478)
Probabilitiesb 0.115 0.000 0.000
Gamma 0.154 0.092 0.081

Note: Cell entries are the probabilities that democracy would fall given the
number of crises, with numbers of observations in parentheses. (a) Number
of crises during a particular year. (b) The probabilities are that the value of
the statistic below is higher than a threshold value (statistical significance of
the differences). Sources: Own research for crises, Cheibub, Gandhi, and
Vreeland (2010) for the institutional systems, Boix, Miller, and Rosato
(2012) for regime classification.

only a vague definition of “major government crises” and

warns that the data may not be reliable. Nevertheless, com-
paring the proportions of democracies that fell under differ-
ent institutional arrangements given government crises shows
that presidential systems are highly vulnerable once
a government crisis explodes. Table 2.5 shows that the effects
of such crises are not statistically significant in parliamentary
systems, but they are in systems that have directly elected
presidents.
A puzzling feature of the patterns shown in Table 2.4 is
the difference among different forms of popular mobilization
against the government. Again, one should not put too much
credence in the data, but it is startling that while those

36
 general patterns

democracies that fell had higher frequencies of general strikes

and riots, the frequency of anti-government demonstrations
was the same. It bears emphasis that the mere appearance of
masses of people on the streets need not indicate a crisis.
In some democracies, peaceful demonstrations are a standard
repertoire of democratic politics, a routine way to inform the
government that some people feel intensely about some issues,
whether in support of or in opposition to government policies.
The propensity to hit the streets differs greatly across democ-
racies – it is frequent in France and extremely rare in Norway,
frequent in Argentina and rare in Costa Rica, reflecting perhaps
differences in political culture. What the statistical patterns
seem to indicate is that recourse to anti-government demon-
strations is just an aspect of everyday life in democracies.
A caveat is obvious, however, and will appear prominently
below: as long as these demonstrations do not lead to physical
violence.
Finally, a crucial factor not considered thus far is the
past experience of democracy. Cornell, Møller, and Skaaning
(2017) warn against drawing analogies between the collapse of
democracies in the interwar years and the present situation.
Notably, they show that in spite of the turbulence of the inter-
war period, none of the twelve democracies that existed for at
least ten years before World War I fell, while twelve of the
fifteen democracies that were born right before or after the war
collapsed during the interwar period. More generally,
Przeworski (2015) learned that the probability of a democracy
falling decreases rapidly as a country accumulates the experi-
ence of peaceful alternations in office resulting from elections.
Among the eighty-eight consolidated democracies, one in ten

37
 the past: crises of democracy

collapsed when the particular democratic spell witnessed no

more than three alternations and only one, in Chile, fell when
the number of past alternations reached four.1
Putting these patterns together, here are the lessons
one can draw from these comparisons of democracies that fell
with those that survived. The economy matters: both the
income at which democracies are consolidated and subse-
quent economic growth sharply distinguish the different out-
comes. Inequality, functional and household, also matters.
Presidential democracies are more likely to fall, being parti-
cularly vulnerable to governmental crises. Finally, while riots
and strikes weaken democracy, as long as they are not violent
we should not fear that anti-government demonstrations
would undermine it.

1
Below I use a count of “unrest” several times: the sum of riots,
assassinations, general strikes, and anti-government demonstrations,
from a data collection originated by Banks (1996) and continued by
Wilson (2017), which I refer to as CNTSDA. These data seem comparable
across countries and time during the early period, but appear to suffer
a bias resulting from increased media coverage during the recent period
and from unequal attention to small and large countries.

38
 3

Some Stories

The comparisons presented above are static: they just summar-

ize average conditions of democracies that survived with those
that collapsed. But histories unravel as sequences of contingent
events, and are not uniquely determined by the extant condi-
tions. Not all that did happen had to happen. To capture these
contingent dynamics, I relate four stories: the collapses of
democracy in Weimar Germany and in Chile and the institu-
tionally resolved political crises in the United States and in
France. I include Germany – a case in which a politician came
to office constitutionally and usurped power, still constitution-
ally, while already in office – because the collapse of the
Weimar democracy is widely used as an omen. I include
Chile because mutatis mutandis it was a paradigmatic case in
which the military acted in defense of capitalism, and democ-
racy fell via a military coup. The political crises in the United
States and France shared the combination of intense conflicts
over foreign wars with highly divisive domestic issues. In the
United States under Nixon the danger was the usurpation of
power by the incumbent, while in France at the end of the 1950s
the threat was a military coup, so these cases parallel those of,
respectively, Germany and Chile. In both countries the crises
were institutionally resolved but in contrasting ways: in the
United States the extant institutions overcame the crisis and
remained intact, while in France the solution involved conflict
and entailed major institutional reforms.

39
 the past: crises of democracy

In each case I first briefly describe the past history of

a particular spell of democracy and then follow the schematic
analysis of crises announced above. I identify the “disaster(s)”
that exposed democracy to a threat, I look for the signs that
democracy was being weakened, and finally I describe the
outcomes of the crises. Because each of these histories is
a subject of many existing volumes, the few pages dedicated
to them below must abstract from many complexities.
Moreover, these were dramatic situations that evoked politi-
cal passions not only at the time, but that continue to evoke
controversies, some still intensely partisan, until today.
My only purpose in recounting them is to search for lessons
that may illuminate our current political situation, so their
presentation is inevitably partial and schematic. What I want
to understand is how the crises emerged and how they were
resolved, by force or by recourse to institutions.

3.1 Germany, 1928–1933

1. Democracy. Democracy, or more accurately “Republic,”
was an improvised solution to the external pressure to abolish
the monarchy, from the onset seen by several political forces
as a temporary one. It did enjoy the support of voters: the
“Weimar Coalition” of the Social Democratic Party (SPD), the
Catholic Zentrum Party (Z), and the German Democratic
Party (DDP) received 76 percent of the vote in the election
of 1919. But several sectors of the Right were either monarchist
or authoritarian and never accepted the republican form of
government, while the extreme Left, first the Independent
Social Democratic Party (USPD) and then the Communist

40
 some stories

Party (KPD) was committed to a socialist revolution, if need

be by force.
Nevertheless, until 1930 German governments were
formed as the result of elections. The first alternation occurred
in 1920, when the SPD was forced to leave the government led
by a Zentrum Prime Minister (PM), Konstantin Fehrenbach,
and the second in 1928, when the SPD returned to the govern-
ment with Hermann Muller becoming the PM (even if the
Zentrum remained in the government coalition).

2. Threats. The Weimar Republic was born out of the disaster of

Germany’s defeat in World War I. During the entire life of the
Weimar Republic the terms of the settlement of the war – the
conditions of the Armistice of November 1918 and of the
Versailles Treaty of 1919 (the abandonment of a monarchical
form of government, the loss of territory, the prohibition of
unification with Austria, the prohibition on rearmament, and
the reparations) – continued to be extremely divisive.

Economic disasters struck twice. Income inequality

was not particularly high: according to Jung (2011: 31), the Gini
coefficient was about 0.33, while inferring it from the data on
top incomes (Atkinson, Piketty, and Saenz 2011) yields a Gini
of about 0.36. Yet the history of Weimar was punctuated by
two economic crises: the hyperinflation of 1923 and the unem-
ployment resulting from the crash of 1929. The hyperinflation
redistributed incomes from savers to borrowers, with average
income falling by 17.4 percent. The unemployment rate rose
from about 3 percent in 1925 to 12 percent in mid-1930 and to
25 percent by 1932, with civilian employment falling from

41
 the past: crises of democracy

about 20 million in 1928 to about 13 million in the first quarter

of 1932 (Dimsdale, Horsewood, and Van Riel 2004: figure 1).
Per capita income fell by a cumulative 18.9 percent between
1928 and 1932. Both crises intensified conflicts over economic
policies, particularly unemployment compensation, as well as
reviving disputes over payments of reparations.
German society was intensely polarized both over
democracy and capitalism. The nationalistic Right developed
an interpretation of German defeat in the war as being caused
by the treason of socialists and democrats, a “stab in the back”
(Dochstosslegende) rather than on the battlefields. The parties
that accepted the conditions of the Armistice were dubbed
“November criminals” by the nationalists. Anti-democratic
thought of all kinds, not just Nazi, remained strong during
the entire period (Sontheimer 1966). The Communist Party
swung back and forth between purely insurrectionary and
electoral strategies until 1928, when under the direction of the
Comintern it adopted a “class contra class” posture, expecting
that the economic crisis would mobilize the masses for
a Communist revolution (Flechtheim 1966). The polarization
was so intense that an eyewitness to these years observed that
all one could ever hear was either “hooray” or “death to . . . ”
(Haffner 2002: 57).
Successive governments were unstable and often
unable to govern. Given the structure of political cleavages
and the institutional system – most importantly, the electoral
system – forming majority governments that were sufficiently
homogeneous to be able to govern was next to impossible.
The German electoral system was based on proportional repre-
sentation without a threshold, leading to a proliferation of

42
 some stories

parties in the Reichstag, the parliament. While the election of

1919 was dominated by the Weimar Coalition and generated 4.1
effective parties, their number rose to 6.4 by 1920, 7.4 after the
first election of 1924 and 6.2 after the second election the
same year, 6.2 again in 1928, 7.1 in 1930, and when the Nazis
became the plurality party, 4.3 in the first and 4.8 in the second
election of 1932. The numbers of competing parties were bewil-
dering: from eighteen in 1919 to twenty-six in 1920, twenty-nine
in May 1924, twenty-eight in December 1924, forty-one in 1928,
thirty-seven in 1930, sixty-two in July 1932, and sixty-one
in November 1932. Yet until 1932 the SPD remained the plur-
ality party. The German National People’s Party (DNVP)
was second or third until 1930, when it was passed by the
Nazis and the KPD. The Zentrum was always third or fourth.
The DDP (DStP as of 1930) and DVP (German People’s Party)
were also weakened by the progress of the Nazis and the KPD
by 1930. Finally, the Bavarian People’s Party (BVP, the Bavarian
wing of the Zentrum) remained unaffected by all the changes,
hovering at around twenty seats throughout the entire period.
Hence, at least until 1930, the distribution of parliamentary
seats was not particularly unstable. Government coalitions,
however, were. Between February 11, 1919 and January 30,
1933, when Hitler became the Reichskanzler, the Weimar
Republic experienced twenty-one cabinets, with an average
duration of 243 days. The shortest was the second government
of Stresemann in 1923, the longest was the broad coalition
headed by Muller after the election of 1928.
Lepsius (1978: 41) places the parties on two dimen-
sions: capitalist versus socialist and authoritarian versus
democratic. His ranking on the pro-democracy dimension is

43
 the past: crises of democracy

SPD and DDP, followed closely by Z+BVP, then at some

distance DVP, DNVP, and KPD, and finally NSDAP
(National Socialist German Workers’ Party). His ranking on
the pro-socialist dimension is KPD, SPD, in the center DDP,
Z+BVP, and NSDAP, followed by DNVP, and then by DVP.
The SPD was clearly committed to democracy and leftist on
economic issues: its 1925 Heidelberg Program called for socia-
lization of the means of production and its policies in the
government promoted workers’ rights and various social
policies. The Zentrum and the DDP contained a broad
range of economic positions; the DDP was resolutely demo-
cratic but the Zentrum contained some monarchists.
The BVP, the Bavarian wing of the Zentrum, was more con-
servative and monarchist. The DVP was nationalistic, fiscally
conservative, and lukewarm about democracy. The DNVP
was nationalistic, somewhat less right-wing on economic
issues than the DVP, anti-republican, and pro-monarchy.
Note that Z, DVP, and DNVP moved to the right around
1930. The two parties never included in coalitions were the
KPD and the NSDAP.
Figure 3.1 shows what the issue space may have
looked like. If this mapping is correct, almost all of the coali-
tions were ideologically contiguous except for the Marx IV
government (January 27, 1927 to June 28, 1928) that did not
include the DDP, and the Brüning II coalition (September 15,
1930 to June 1, 1932) that did not include DVP. The center-left
coalition, which included SPD, Z, and DDP, was alone never
majoritarian after 1920. Neither was the center-right coalition
of Z, DDP, and DVP (plus often BVP). Majoritarian coali-
tions would have to include SPD together with DVP, or Z and

44
 some stories

DDP SPD
Democratic 10
9 Z+BVP
8
7
6 DVP
5
4
3
2 KPD
DNVP
1
NSDAP
0
0 1 2 3 4 5 6 7 8 9 10
Socialist

Figure 3.1. Putative location of parties on the Left–Right and

democratic–authoritarian space
Parties: SPD (German Social Democratic Party), Z (Zentrum),
DDP (German Democratic Party; changed name to DStP in
1930), DVP (German People’s Party), BVP (Bavarian People’s
Party), DNVP (German National People’s Party), NSDAP
(National Socialist German Workers Party), KPD (German
Communist Party)

DDP together with DNVP. The first of these coalitions was

formed twice, under Stresemann I (August 13, 1923
to November 3, 1923) and Muller II (June 28, 1928
to March 29, 1930), but in both cases the SPD could not
agree with the DVP over economic issues. The second was
formed once, under Luther I (January 15, 1925 to October 26,
1925), but the DNVP left it over the Locarno Treaties. It seems
that the distance between the SPD and the DVP was too large
on economic issues and between Z and DNVP over nation-
alism, so that any majority coalition was too fragile to last an
entire electoral period.

45
 the past: crises of democracy

The parties themselves were far from homogeneous

and disciplined. In particular, in several instances prime
ministers found that they did not have the support of their
own parties for the various policy compromises with which
they attempted to save coalitions. This situation is best
described by Lepsius (1978: 44; see also Carr 1969: 336):
“The coherence of a government was achieved by the inter-
action of a few personalities who could exert influence
within their parties to make them tolerate the government
from issue to issue . . . The government became more depen-
dent on the prerogatives of the Reichspräsident and saw
itself as an independent agency that had to continue govern-
ing by continuous crisis management despite the fragmen-
ted Parliament.”
3. Signs.1 German democracy emerged from the violence of
World War I, and it emerged violent. From the outset,
“political parties associated themselves with armed and uni-
formed squads, paramilitary troops whose task it was to
provide guards at meetings, impress the public by marching
in military order through the streets, and to intimidate, beat
up and on occasion kill members of the paramilitary units
associated with other political parties. The relationship
between the politicians and the paramilitaries was often
fraught with tension, and paramilitary organizations always
maintained a greater or lesser degree of autonomy; still, their
political colouring was usually clear enough” (Evans 2003:
KL 1707–11). Politically motivated murders were frequent:

1
This subsection draws on Schumann (2009).

46
 some stories

156 democratic politicians were assassinated by right-wing

paramilitaries. Political violence peaked in 1923, with the
bloody suppression of an abortive Communist uprising in
Hamburg, gun battles between rival political groups in
Munich, and armed clashes involving French-backed
separatists in the Rhineland. While the period 1924–8 was
relatively orderly and peaceful, social peace was shattered
again by the crisis of 1929–30 and the subsequent rise of Nazi
stormtroopers. “By the end of 1931,” Carr (1969: 351)
observes, “the centre of gravity of German political life was
rapidly moving away from the Reichstag and the chancellery
to the streets.” According to Evans (2003, KL 5211–16),

In 1930 the figures rose dramatically, with the Nazis claiming

to have suffered seventeen deaths, rising to forty-two in 1931
and eighty-four in 1932. In 1932, too, the Nazis reported that
nearly ten thousand of their rank-and-file had been
wounded in clashes with their opponents. The Communists
reported forty-four deaths in fights with the Nazis in 1930, 52
in 1931 and seventy-five in the first six months of 1932 alone,
while over fifty Reichsbanner men died in battles with the
Nazis on the streets from 1929 to 1933. Official sources
broadly corroborated these claims, with one estimate in the
Reichstag, not disputed by anybody, putting the number of
dead in the year to March 1931 at no fewer than 300.

The summary of these events is shown in Figure 3.2,

which illustrates the incidence of “unrest” (the number of
events per year): again, the sum of riots, assassinations, gen-
eral strikes, and anti-government demonstrations by year.

47
 the past: crises of democracy

20
15
Unrest
10
5
0

1920 1925 1930 1935

Year

Figure 3.2. Unrest by year in Germany, 1919–33

The support for traditional parties, the top four vote-

getters in 1919, eroded sharply over time, from their total share
of 85.7 percent in 1919, to 68.3 in the first election of 1924, to
65.2 in 1928, to 51.6 in 1930, to 41.1 in the first election of 1930.
Turnout fell from 83 percent in 1919 to 75.6 percent by 1928.
Carr (1969: 337) observes that “interest in party politics was
declining . . . Party politics – ‘das System’, as many started to
call it – were clearly falling into disrepute.” Yet the intensifi-
cation of conflicts following the economic crisis resulted in
turnout rising to 82 percent in 1930, when the Nazis brought
previous non-voters to the polls. The number of voters
increased from 31.2 million in 1928 to 35.2 million in 1930,
while the number of votes for the NSDAP jumped from 810
thousand to 6.38 million. According to King et al. (2008), the
main basis of support for the NSDAP was various kinds of

48
 some stories

self-employed, while people who were unemployed tended to

vote Communist.
The extreme instability of economic, cultural, and poli-
tical life of Weimar Germany made many people willing to
embrace all kinds of delusions, including curing the sick by
applications of cottage cheese or making gold from base metals.
As another eyewitness reports, “The shock of Germany’s defeat,
the inflation, the get-rich-quick boom that followed stabiliza-
tion, the influx of foreign money, and the swirl of underground
activity against the ‘internal and external enemy’, had all com-
bined to produce an atmosphere of unreality made to order for
revivalists, quacks, and confidence men” (Delmer 1972: 95).
Until 1930 Hitler was just one among many.
4. Outcome. The final outcome is known, so there is no point
entering into details. Two aspects of Hitler’s rise to office and the
subsequent fall of democracy, however, merit emphasis.
The first is that Hitler came to power legally, through an
“authoritarian gap in the Weimar Constitution” (Bracher 1966:
119): Article 48 that allowed the president to empower the
government to rule by decree. President Hindenburg used this
prerogative in the winter of 1930, when the Reichstag could not
agree on how to cope with the economic crisis and the coalition
government led by the Social Democrat Muller resigned.
Hindenburg appointed Heinrich Brüning as chancellor, making
it clear that the new government could rely on emergency
powers, and when Brüning could not muster a parliamentary
majority in July 1930 he promulgated the budget by decree.
From this point on, no government – Brüning’s, Papen’s, or
Schleicher’s – ruled with the support of a majority in the

49
 the past: crises of democracy

Reichstag. The Reichstag lost its very raison d’être and almost
ceased to meet: according to Evans (2003: KL 5328–30),
“The Reichstag sat on average a hundred days a year from
1920 to 1930. It was in session for fifty days
between October 1930 and March 1931; after that, it only met
on twenty-four further days up to the elections of July 1932.
From July 1932 to February 1933 it convened for a mere three
days in six months.” Hitler was appointed chancellor
on January 30, 1933 with the same powers as his three predeces-
sors. He grabbed dictatorial powers on March 23, 1933 when the
Reichstag passed the “Law for the Relief of the People and the
Reich” by the constitutionally required two-thirds majority,
which allowed the government to issue decrees deviating from
the constitution. While Communist deputies were not allowed
to sit and some Social Democrats were too intimidated to
participate, Hitler was still concerned about obtaining the requi-
site majority, which he did when the Zentrum voted in favor
(Ermakoff 2008). Hence, from the purely legal point of view, the
final coup against the Weimar Constitution was delivered
constitutionally.
The second aspect of these events is that nobody –
not the politicians who facilitated Hitler’s entrance into the
government, not his opponents, and seemingly not Hitler
himself – expected him to monopolize and consolidate
power. The NSDAP electoral success in 1930 was widely
viewed as temporary and their reversal in the second elec-
tion of 1932 as an indication that the Nazi tide was ebbing.
After a Nazi defeat in a local election in 1932, an influential
liberal newspaper proclaimed that “The mighty National

50
 some stories

Socialist assault on the democratic state has been repulsed”

(Frankfurter Zeitung, quoted in Turner 1985: 313).
Moreover, by the second half of 1932 the Nazis were despe-
rately short of money and Hitler was despondent about his
chances of becoming chancellor. When Schleicher decided
in 1932 to undermine Brüning and seek accommodation
with Hitler, he was “assuming that it [the Nazi party] was
a healthy nationalist movement which he could tame and
exploit by adroit political manipulation” (Carr 1969: 352).
Even when Hitler became chancellor, his cabinet included
only three National Socialists as against eight non-Nazi
conservatives. According to Delmer (1972: 117), the general
perception was that “Hitler is Chancellor, but he is
Chancellor in handcuffs. He is the prisoner of Papen,
Hugenberg, and Hindenburg.” Papen declared that
“In less than two months we will have pushed Hitler so far
into the corner that he’ll be squeaking” (quoted in Bracher
1966: 120).
These two features of the events in Germany are
worth retaining for general understanding of crises of democ-
racy. The first is that particular institutional design matters:
the proportional representation system made it difficult to
form governments that could effectively govern, while the
emergency powers embedded in the constitution permitted
a constitutional slide into authoritarianism. The second is that
the final outcome of these events was not foretold and not
even anticipated by the people who ended up generating it.
Contingency and the corresponding uncertainty are inherent
aspects of complex conflicts.

51
 the past: crises of democracy

3.2 Chile, 1970–1973

1. Democracy. The spell of democracy that ended in 1973
began in 1938, with the election of a Radical Party candi-
date Pedro Aguirre Cerda. The Communist Party was
proscribed by President Gabriel González Videla in 1946,
but the ban was overturned six years later by President
Carlos Ibáñez del Campo. Partisan alternations in the
presidency occurred in 1952, 1958, and 1964. In 1958, the
winner, Jorge Alessandri, won the plurality of only
31.6 percent but his victory was recognized by the runner
up, Salvador Allende. Hence, respect for results of elec-
tions was well entrenched.
2. Threats. By the mid-1960s, Chile was a country deeply
divided along economic lines, with a wide spectrum of inten-
sely held ideological beliefs about capitalism versus socialism.
Political divisions based on class were profound. In 1958,
93 percent of “rich people” voted for the right-wing candidate,
Alessandri, while 73 percent of workers voted for Allende
(Prothro and Chaparro 1976: 73). Still, Navia and Osorno
(2017) emphasize the importance of purely ideological divi-
sions, autonomous from class positions. The importance of
ideology is also documented by Prothro and Chaparro (1976:
87), who report that in the 1964 election, 59 percent of respon-
dents gave purely ideological reasons for supporting or
opposing Allende and 45 percent did the same for the winner,
Eduardo Frei. While the evidence is only anecdotal, these
divisions had deep roots in the social fabric, with gossip in
Santiago about bourgeois fathers who expelled their

52
 some stories

daughters from home not because they became pregnant but

because they supported the Allende government.

Income inequality had been traditionally high in Chile.

According to WIDER (UNU-WIDER 2014), the Gini of market
incomes was 46.2 in 1964 and it increased to 50.3 by 1968 (50.5
according to SWIID), which was exceptionally high for this
period. Labor share (from PWT9.0) was particularly low, at
45 percent. According to Lambrecht (2011), the highest decile
received 40.23 percent of all incomes, while the lowest earned
1.45 percent. The upper 50 percent of recipients received 83 per-
cent of national income, while the bottom 50 percent earned
17 percent. Distinguished by sector, 80.4 percent of people
employed in agriculture were in the bottom half of recipients,
41.7 percent of those employed in industry, and 40.9 of those in
services. By occupation, 72 percent of industrial workers
received less than the median income and 89 percent of white-
collar employees gained incomes higher than the median, while
the self-employed were about equally likely to be in either
bracket. In addition to the rural–urban cleavage, there were
large differences across sectors: workers in copper mining and
industry were much better off than the economically marginal
groups that occupied newly formed Santiago slums, popularly
called callampas (“mushrooms,” because of how fast they
grew). The reason for this high inequality was that the
Chilean economy – industry, finance, and agriculture – was
highly concentrated. In 1966, 144 firms owned more than
50 percent of industrial assets, three banks held 44.5 percent
of deposits and received 55.1 percent of profits, while 9.7 percent
of landowners owned 86 percent of cultivable land.

53
 the past: crises of democracy

The crisis of governance, eventually a complete stale-

mate between the president and the Congress, developed in
steps. In the presidential election of September 4, 1970, the Left
coalesced under the candidature of Dr. Salvador Allende, the
Right was represented by Arturo Alessandri (PN), and the
Christian Democrats (PDC) by a left-leaning candidate,
Radomiro Tomic. The results were close, with Allende winning
36.1 percent, Alessandri 35 percent, and Tomic 28 percent.
Given that no one won a majority, the final choice was to be
made by the Congress. To stop Allende, some people around
Alessandri were trying to craft a deal with the Christian
Democrats, promising that if they voted for him, Alessandri
would immediately resign, and that the outgoing Christian
Democratic president, Eduardo Frei, who was eligible to run
in the ensuing election, would win. Frei, however, refused to go
along with this, so the deal collapsed. Alessandri himself called
for the members of his party to vote for Allende, but the
Executive Committee of the PN refused to follow his instruc-
tion and Alessandri received thirty-five votes in the Congress.
Allende was confirmed by the Congress on October 24 with the
votes of seventy-eight members of his coalition and seventy-
four Christian Democrats, assuming office on November 3.
Allende won by a small margin as the head of
a coalition of seven parties, Unidad Popular (UP). This coali-
tion included the center-left, represented by the Radical Party,
which would split in June 1971; the Communist Party, which
had a far-reaching program but was moderate and disciplined
about tactics (Corvalan 2003); Allende’s own Socialist Party;
and several small groups, composed mainly of intellectuals
(on divisions with the Left, see Yocelevzky 2002: chapter 2).

54
 some stories

A radical Left extra-parliamentary group, the Movimiento de

Izquierda Revolucionaria (MIR), rooted in the Santiago
slums, remained outside the coalition. The original cabinet
was composed of four Socialists, three Communists, three
Radicals, two Social Democrats, two members of smaller
parties, and one independent, who was the minister of the
economy.
The new president inherited a House elected in 1969,
in which the Christian Democrats held fifty-six seats, Partido
Nacional thirty-four, the Communists twenty-two, the
Radicals twenty-four, and the Socialists fifteen, of the total
of 150. The composition of the Senate was similar. Hence, the
government coalition held a minority of seats in both cham-
bers and the Christian Democrats were pivotal.
To govern, Allende would have to find a compromise
with the Christian Democrats that would be palatable to the
members of his coalition. No such compromise would be found.
Allende did not even control his own party: he had been pro-
claimed as the party’s presidential candidate, with thirteen votes
in favor and fourteen abstentions. The deeply divided Socialist
Party took a sharp turn toward insurrectionary strategy
in January 1971, when Carlos Altamirano replaced Aniceto
Rodriguez as general secretary. Altamirano spoke in disparaging
terms about the possibility of a “peaceful road to socialism,”
believing that because the bourgeoisie would defend its position
by force, socialism could be achieved only by the armed action of
the working class (Resolution of the Socialist Party adopted in la
Serena in January 1971, Altamirano 1979: 19).
In addition to redistributive measures aimed at stimu-
lating the economy in the short run, the program of the UP

55
 the past: crises of democracy

included continuing land distribution to peasants and complet-

ing the nationalization of copper, both initiated under the pre-
vious government, as well as nationalizing nitrate mining,
banks, and some large industrial firms (see ODEPLAN 1971).
Agrarian reform had already been enabled by a law passed
under Eduardo Frei in 1969. Copper was nationalized by
a unanimous vote in the Congress, while other mining sectors
and the banks were gradually bought by the state, with the
opposition of their United States and Chilean owners, but with-
out requiring legislation. The nationalization of industrial firms,
however, required a law. Note that while some sectors of the UP
coalition pushed for nationalization as a goal in itself, the gov-
ernment portrayed it as limited and instrumental. The minister
of the economy, Pedro Vukovic, argued that given the highly
oligopolist structure of the Chilean economy and the endemi-
cally high inflation, the state should take over one or two large
firms in each of the crucial sectors, and use pricing by these firms
as an instrument to control inflation.
The government negotiated an agreement with the
Christian Democrats, with vague criteria – “firms that operated
activities of primary importance for the economic life of the
country” – according to which firms would be nationalized,
moved to mixed ownership, or remain private. On October 12,
1971, the government sent to the Congress a bill that would
implement this agreement, La Ley de las Areas. On the eve of
the vote, however, a split within the Christian Democratic
Party led it to renege on the original agreement, substituting
the original bill by a law that would allow the Congress to
decide each nationalization one-by-one. While the original
goal of the government was to nationalize 243 firms, later scaled

56
 some stories

down to ninety-one, it was estimated that the Congress could

process with at most eight nationalizations per year (Martner
1988). This law was passed by both houses of the Congress
on February 19, 1972. It included annulling 520 expropriations
made after October 14, 1971. Following a massive demonstra-
tion by the Left against this version of the reform, Allende
vetoed it on April 11, 1972, and the opposition, led by the
president of the Senate, Patricio Aylwin, demonstrated against
the veto the day after. The result was a legal limbo. Under
pressure from several members of the UP as well as from
ordinary workers, Allende reverted to an obscure and never
previously used decree dating to 1932, according to which the
state could intervene in firms “paralyzed by labor unrest.”
Needless to say, workers in many firms, including small family-
held enterprises, were only too eager to paralyze them and have
the state intervene. The result was chaos, which turned out to
be uncontrollable. By October 1972, workers occupied several
Santiago factories. Spontaneous forms of self-organization,
Cordones Industriales and Comandos Comunales, emerged to
replace formal economic and state organizations (for details,
see de Vylder 1974: chapter 6).
Once the Ley de las Areas failed, the executive–legislative
stalemate was complete. To the best of my knowledge, not
a single major bill proposed by the government was subsequently
passed by the Congress and all major laws passed by the
Congress were vetoed by the president. In a Chilean tradition
dating back to the early twentieth century, the Congress raised
constitutional accusations and censored several ministers, to
which Allende’s reaction was always to change their portfolios
and keep them. By September 1, 1972, Senator Hamilton (PDC)

57
 the past: crises of democracy

called for the resignation of the president and the clamor for
impeachment became frequent. Impeachment, however,
required a two-thirds majority, which the opposition could not
muster in the parliamentary elections of March 1973, when
Allende still enjoyed the support of 49.7 percent of survey
respondents in Santiago (Navia and Osorio 2017).
The final constitutional limbo emerged late in
1972. On October 23, 1972 the Congress passed a law which
granted the armed forces the prerogative to search all places
that may store arms and explosives, which according to the
opposition were being secretly amassed by some pro-
government organizations. The president responded by
invoking a legal clause according to which the armed forces
had no right to enter public buildings without his specific
authorization. Up to this time, the position of most leading
generals was that the armed forces should remain apolitical
unless the government violated the constitution. This conflict,
however, rendered the criterion of constitutionality inopera-
tive. On August 22, 1973, the Congress declared that the
government violated the constitution and was illegitimate.
One day later General Pinochet replaced General Prats as
the head of the armed forces, and the road was opened to
the coup of September 11.
3. Signs.2 As soon as Allende was elected, the Right initiated
a US-assisted “campaign of terror,” raising the specter of
Soviet domination. On October 22, 1970, a group of officers
aiming to generate political instability attempted to kidnap

2
The chronology in this subsection is based mainly on Los mil dias de
Allende (1997).

58
 some stories

the head of the armed forces, General Schneider, but ended up

killing him.

Conflicts in agriculture and strikes in industry inten-

sified immediately after the election. While in 1969, the
last year of Frei’s government, there were 1,127 strikes in the
countryside and 148 instances in which peasants occupied
latifundia (“tomas de fundo”), in 1970 there were 1,580 strikes
and 456 occupations, of which 192 took place in the last three
months of 1970 alone (Landsberger and McDaniel 1976).
According to a different source (Martner 1988), there were
1,758 rural strikes in 1971, while occupations increased from
450 in 1970 to 1,278 in 1971. Both the president and the minister
of agriculture spoke publicly against illegal land seizures, but
could not control them. Landowners mobilized, in some cases
importing arms, and organized forcible retakings of land.
They created para-judicial institutions, Tribunales Agrarios,
that ruled in their favor. Owners who earlier had ceded land
when it was legally expropriated joined in the resistance.
In 1969 there were 977 industrial and mining strikes
with 275,000 workers participating; in 1971, the first full year of
Allende, there were 2,709 strikes involving 302,000 workers;
and in 1972 there were 3,289 strikes engaging 397,000 workers.
Note that the strikes extended to smaller firms, as indicated by
the ratio of strikers to strikes.
The Right first occupied the streets on December 1, 1971,
when elite women of Santiago marched banging empty pots
(cacerolas vacias), protected by bodyguards from a fascist group,
Patria y Libertad. They were attacked with stones by a Left
counter-demonstration. Ten months later, in October 1972, the

59
 the past: crises of democracy

country was paralyzed by a strike of truck owners, lockout of

many factories, and work stoppage by professionals. To deal
with the unrest a state of emergency was declared by the govern-
ment, and in November the military entered the government for
the first time, with General Prats becoming the minister of
interior. Another strike of truck owners took place
in August 1973.
The violence escalated from the first shooting
on December 2, 1970, when two left-wing students were
wounded in Concepción. A peasant was killed during a land
occupation on June 9, 1971. The assassination of General
Schneider was followed by the murder, also on June 9, by an
obscure left-wing group (VOP, Vanguardia Organizada del
Pueblo) of a former minister of interior, Edmundo Pérez
Zujovic. A few days later two members of the VOP were killed
in a five-hour shoot-out with the police. By October 28, 1971 El
Mercurio reported seven armed confrontations during land
conflicts in Temuco, with four dead and nineteen wounded.
Another conflict over land ended with one person dead and
three wounded on November 22, 1971. On May 20, 1972 the
shootings moved to Santiago, with a member of MIR killed by
the police. A violent conflict between inhabitants of a Santiago
slum and the police led to one dead and several wounded
on August 5, 1972. Street fighting took place over the govern-
ment’s education bill, with several wounded, on April 26, 1973.
Allende’s military aide, Arturo Araya, was murdered by Patria
y Libertad on June 27, 1973. By mid-June 1973, shootings,
explosions, and fires became daily occurrences: supermarkets
were sacked, right-wing paramilitary groups fired shots from
cars and exploded bombs at several local headquarters of

60
 some stories

15
10
Unrest
5
0

1940 1950 1960 1970 1980

Year

Figure 3.3. Unrest by year in Chile, 1938–73

government parties, and workers from the El Teniente mine

marched to Santiago behind a tractor filled with explosives
(Bitar and Pizarro no date). The government tried to find a far-
reaching accommodation with the opposition, the Catholic
Church tried to mediate, but the spiral of violence could not
be stopped.
As Figure 3.3 shows, the incidence of popular unrest
skyrocketed in 1972. None of these dramatic events affected
the distribution of political attitudes. In 1972, 99 percent of
upper-income respondents said that household staples are
“difficult” to find, while 75 percent of lower-income persons
found it “easy.” Surveys reported by Prothro and Chaparro
(1976: 102 and 104) show that both the upper and lower class
agreed that Chile was “living in a climate of violence,” but
only 7 percent of upper-income respondents attributed it

61
 the past: crises of democracy

exclusively to the opposition while 35 percent of lower-

income respondents did. The vote shares of the five largest
parties in the congressional elections in March 1973 remained
almost the same as in 1969, with a slight gain for the Socialists.
Turnout was higher in 1973 than in 1969, while the potential
electorate increased by almost 40 percent. Hence, there were
no signs either of dissatisfaction with party politics or of
erosion of support for the traditional parties.
The most ominous signals emanated from the military.
While the official stand of the military leadership was that the
armed forces would stay away from politics as long as the
government did not violate the constitution, officers impatient
to depose the government by force were many. General Viaux,
who commanded the operation against General Schneider
in October 1970, had already led a mutiny over military pay
against President Frei in 1969. A failed coup attempt
of June 1973 was led by a lower-ranking officer, Colonel
Souper. Throughout the entire period, the possibility of a coup
led by some group of the military was the most frequently
discussed topic in Santiago. After the Right failed to obtain the
two-thirds of the seats required to impeach the president, in the
election of March 1973, it became clear that Christian Democrats,
by then led by the future democratic president of Chile, Patricio
Aylwin, would welcome a coup. By July of 1973, the only question
was what form the coup would assume. The fear was that if it
were initiated by a navy unit stationed in the south of the country
it would lead to a civil war, but most people expected a golpe
blando, in which the military leadership would put Allende on
a plane to Cuba and call for an election that would be won by
Eduardo Frei. I do not know anyone who expected that the coup

62
 some stories

would be as ferocious and bloody as it eventually turned out

to be, or that the military would stay in power for sixteen
years.
4. Outcome. As Hutchison, Klubock, and Milanic (2013: 348)
remarked, “The tensions between the phased and controlled
revolution from above and the more spontaneous and locally
informed revolution from below were never resolved, con-
stituting a fatal flaw in the Chilean revolutionary process.”
One major aspect of the dynamic of Chilean events was the
inability of the government to control its own supporters.
Allende could not act strategically because groups within his
own party as well as some other members of his coalition,
notably excluding Communists, could not be disciplined to
moderate their demands and to demobilize when the situa-
tion so required. Lord Bevan, a minister in the UK’s post-
war Labour government, once remarked that “We do not
want to be in the position of having to listen to our own
people.” Allende did not have this choice: peasants were
occupying land against the pleading of the government,
some sectors of the coalition were amassing arms against
the injunction by the government, and an important left-
wing armed group was not even a member of the governing
coalition. Ideological passion was intense, discipline non-
existent.
The Chilean events again highlight the importance of
the democratic institutional framework. In contrast to Weimar,
Chile had a purely presidential system. Under this system
a government can become paralyzed when different political
forces control the executive and the legislature: “divided

63
 the past: crises of democracy

government” in the parlance of the United States. Legislation

initiated by the executive was not passed by the legislature;
legislation passed by the Congress was vetoed by the president.
When the paralysis became acute, the president sought emer-
gency powers while the Congress moved to impeach the pre-
sident. In the meantime, incomes were falling, inflation soared,
and shortages became ubiquitous, while the supporters of the
opposite camps tried to impose themselves on the streets.
Finally, the Chilean events speak to the tension between
democracy and capitalism. Writing in 1886, Hjalmar Branting,
the leader of the Swedish Social Democrats, wondered whether
“the upper class would respect popular will even if it demanded
the abolition of its privileges” (quoted in Tingsten 1973: 361).
An SPD leader, August Bebel, thought in 1905 that revolution
may be necessary “as a purely defensive measure, designed to
safeguard the exercise of power legitimately acquired through
the ballot” (quoted in Schorske 1955: 43). Clearly, Allende did not
have a popular mandate for far-reaching social and economic
transformations: he won the presidency by a slim plurality and
his coalition never had a majority in the legislature. He did win
according to the rules and tried to govern according to the rules,
but he was being pushed by the forces behind him to reach
beyond his mandate. The upper class, whose privileges were
being threatened, turned to the military and, not without some
hesitation, the military were willing to oblige.

3.3 France, 1954–1962 and 1968

1. Democracy. The French Third Republic was founded in
1875. Elections were held regularly until 1939, with control

64
 some stories

over government changing partisan hands nine times during

this period. In the immediate aftermath of World War II
a series of provisional governments, the first headed by
General Charles de Gaulle and the last by Leon Blum, gov-
erned the country until a new constitution was adopted in
1946, giving birth to the Fourth Republic. Parliamentary elec-
tions were held under this constitution in November 1946,
June 1951, and January 1956.

2. Threats. The first years after the war witnessed a high degree
of social unrest, culminating in 1947 with a series of insurrec-
tionary strikes, which were bloodily repressed. A massive
strike of the public sector took place in 1953 but there were
no more major labor conflicts during the remaining years of
the Fourth Republic, until May 1958.
From 1946 France was engaged in several colonial wars
against independence movements. The Indochina war ended
in July 1954 after the rout of the French forces at Dien Bien Phu.
However, a few months later a new war began in Algeria, with
profound consequences for the next eight years. This war,
which until 1999 was officially referred to euphemistically as
“the Algerian events,” caused deep divisions within French
society and several rifts between civilian authorities and the
army units stationed in Algeria, with two coup attempts,
in May 1958 and in April 1961, as well as a wave of terrorism.
Government instability and ineffectiveness under the
Fourth Republic mirrored that of Weimar Germany.
The electoral law adopted in 1945 (in France electoral laws are
not a part of the constitution) provided for proportional repre-
sentation at the level of electoral districts. This law was slightly

65
 the past: crises of democracy

modified in 1951 to allow electoral coalitions and to introduce

a dose of majoritarianism, with the intention to reduce the
weight of the Communists and the Gaullists (Bon 1978: 68).
Between January 22, 1947 and June 2, 1958, twenty-four govern-
ments were in office, for an average duration of 173 days, shorter
than the 243 in Weimar, ranging from two days to one year and
four months. Moreover, between October 27, 1946 and June 2,
1958, there were 375 days, more than one year in twelve, during
which there were no governments. All the governments of the
Fourth Republic were coalitions of several parties, from four to
eight. After Paul Ramadier, the first prime minister of the Fourth
Republic, expelled the Communists (PCF) from his government
in October 1947, continuity during the term of the parliament
elected in 1946 was provided by the presence in all the ten
governments of three parties: SFIO (Socialist), MRP
(Radical), and PRS (Radical-Socialist). The last two parties
also participated in all governments following the election
of June 17, 1951, but the SFIO was not included in any of the
eight governments in office during this term. The parliament
was dissolved at the end of 1955, with the election following
on January 2, 1956, and the SFIO joined all the subsequent
governments of the Fourth Republic. The fragility of the
successive governments made them incapable of making
major decisions, while the continuity of their composition
meant that no other political alignment constituted a viable
alternative. As Denquin (1988: 88) observes, “From the point
of view of citizens, the political system was lived as at the
same time odious and natural.”
3. Signs. None of the short-lived governments was able to
build majority consensus over the handling of the Algerian

66
 some stories

conflict. The Mendes France government’s plan of reforms

was defeated in the Assemblée and it resigned on March 3,
1955. The succeeding government of Edgar Faure pro-
claimed a state of emergency in Algeria but did not survive
either. The combination of the intensification of the
Algerian crisis with the incapacity of the successive govern-
ments to cope with it made attempts at forming govern-
ments increasingly frantic. When the government of Guy
Mollet (SFIO), which granted special powers to the military
in Algeria, fell on May 21, 1957, the interregnum lasted three
weeks. It was followed by the government of Maurice
Bourgès-Maunoury (PRS), which lasted 110 days. After
this government fell on September 30, 1957, two attempts
to form a government, by Antoine Pinay (CNIP) and by
Guy Mollet, failed, and the next government was formed
only on November 6. This government, headed by Félix
Gaillard (PRS), proposed a new framework law for Algeria,
which was again defeated in the parliament, and the gov-
ernment resigned after 160 days in office. It took twenty-
nine days to form the next government, of Pierre Pflimin
(MRP), which lasted fourteen days. Altogether, the fall of
Mollet was followed by eighty-nine days in which France
had no government. The last government of the Fourth
Republic was formed by General Charles de Gaulle
on June 1, 1958.
Interpretations of de Gaulle’s accession to the office of
prime minister are highly controversial. It would be presump-
tuous for me to take a side in this controversy, which centers on
the issue of whether de Gaulle was complicit in the events that

67
 the past: crises of democracy

brought him to power. The basic facts, however, are known.

On May 13, 1958, several groups of civilians assaulted and occu-
pied the seat of government in Algiers, with the passive attitude
of the police forces that were guarding it. The leaders of the
insurrection, some but not all of whom were Gaullists, formed
a Committee of Public Safety, headed by an active general,
Jacques Massu, and declared that they would govern Algeria
until a government favorable to maintaining French control
over the colony was installed in Paris. The intention of the
insurrection was to prevent the investiture of Pflimin and force
on the Assemblée the candidature of de Gaulle. On May 24,
a similar Committee of Public Safety was formed in Corsica, so
it appeared that the revolt was spreading toward the metropolis
(Operacion Ressurection). Pflimin was named prime minister
on May 14, but on May 27 de Gaulle broke his silence, declaring
himself a candidate for the office, and ordering the military in
Algeria to obey the orders of their superiors. The declaration
certainly reads like a usurpation of powers the general did not yet
have: “In these conditions, all actions by whatever side that
undermine public order risk to have serious consequences.
While recognizing the circumstances, I will not approve them”
(quoted in Denquin 1988: 171). The Pflimin government resigned
one day later and the parliament invested a government headed
by de Gaulle. It was a broad coalition of six parties plus indepen-
dents and it included three former prime ministers.
The government was granted the power to rule by decrees for
six months and to initiate the project of a new constitution.
Hence, while the procedure by which de Gaulle acceded to office
was constitutional, his advent to power was a response to a coup
d’état in Algeria as well as the opening gambit on his part.

68
 some stories

The project of the new constitution was prepared

rapidly. First, however, the parliament had to modify the
amendment clause of the Constitution of the Fourth Republic,
Article 90, to allow the change of the entire constitution. Having
done this, the new government subjected the constitutional
project to a referendum. Before voting, each household received
two ballots (yes or no?), the text of the constitution, and the text
of a speech by de Gaulle of September 4, arguing for the new
constitution. The constitution was supported by 79 percent of
voters in the metropolis and 95 percent in Algeria. It was for-
mally adopted on September 28, 1958 and de Gaulle was elected
president on December 21, 1958, with Michel Debré becoming
prime minister. The law regulating parliamentary elections was
changed by a decree of the government on October 13, 1958.
The new president’s views on the solution to the
Algerian crisis are again subject to conflicting interpreta-
tions, namely, whether he believed from the outset that the
independence of Algeria was inevitable or changed his
views as events unfolded. In the immediate aftermath of
his accession to office, in a famous speech in which he told
the French in Algeria, “Je vous écoute” (“I hear you”), he let
it be understood that he favored the colonial status quo,
Algérie Française, albeit accompanied by the extension of
political rights to Algerians and massive investment pro-
jects to develop the territory. Yet his stance evolved rapidly,
causing French Algerians to claim that they were being
betrayed. Public opinion in France was evolving as well:
in July 1956, 45 percent of respondents were favorable to
negotiations with “chiefs of the rebellion,” in July 1957 the
proportion was 53 percent, in May 1959 it rose to 71 percent

69
 the past: crises of democracy

(Ageron 1976). As de Gaulle’s language was shifting from

“association,” to “autodetermination,” and then to “Algérie
algérienne,” conflicts erupted within the French commu-
nity in Algeria. Violent Franco–Franco confrontation took
place on February 24, 1960, with several deaths. On June 16,
1960 an organization defending French Algeria (Front
Algérie Française, FAF) was formed; it engaged in several
violent demonstrations, and was banned in December of
that year. A referendum on autodetermination, which took
place on January 8, 1961, was supported by 75.8 percent in
the metropolis and 69.1 percent in Algeria. A terrorist orga-
nization, Organisation Armée Secrète, that was to engage in
many bombings and several assassinations, was formed
in February 1961. On April 22, 1961, several generals pro-
claimed the secession of Algeria from France, under their
government. According to Droz and Lever (1991: 296–313),
this new attempt at a coup generated feverish fears in Paris
of an impending invasion, but it fizzled out when the army
recruits refused to obey the rebellious generals.
The agreement to end the war was concluded
on March 19, 1962, and it was ratified by the referendum
of April 8, 1962.
The economic, social, and military burden of the war
was enormous: altogether 1.75 million Frenchmen served in
Algeria, of whom 1.34 million were conscripts. It deeply divided
French society, not only in Algeria but also in the metropolis,
with several instances in which police bloodily repressed anti-
war demonstrations. After several protestors were killed
on February 8, 1962, their funeral brought 500,000 people to
the streets.

70
 some stories

4. Outcome. The Algerian war was a disaster that could not be

coped with by any government under the institutional frame-
work of the Fourth Republic. Yet, even if the war continued
for another four years, the impasse was broken by the advent
of de Gaulle to office, the broad powers granted to him, and
the institutional change that moved the center of power from
the parliament to the president.

Forget all the details and consider the situation in

abstract terms. A society is experiencing a profound disaster
and the institutional framework does not generate govern-
ments that could effectively cope with it. The military rises in
arms to affect the orientation of a new government. A retired
general, a war hero, takes the initiative of imposing himself as
the head of the government. He demands, and obtains from
the parliament, the power to rule by decree and to change the
constitution. His government controls the press and the
radio, bans several organizations on both sides of the conflict,
and prosecutes several people for “demoralizing the army”
(Denquin 1988: 150). The story sounds familiar, and ominous.
Yet eventually the war ends and democracy survives.
In the aftermath of World War II, the anti-democratic
forces, mainly former supporters of the Vichy regime, were
weak in France. They opposed de Gaulle from the Right but all
the organized political forces were resolutely democratic. This
commitment was enshrined in Article 93 of the 1946 constitu-
tion, which stated that the one aspect of the constitution that
could not be amended was the republican form of government.
The rebellious military were prepared to engage in insurrec-
tionary activities but their demands were restricted to keeping

71
 the past: crises of democracy

Algeria French, not to instituting an authoritarian regime.

Indeed, many of the military rebels were Gaullist. There are,
thus, good reasons to think that any attempt at installing
a dictatorship would have been violently opposed. But the
story of 1958 cannot be told without invoking the personality
of General de Gaulle. While not hesitating to use all the instru-
ments of power, including highly repressive ones, to lead
France through the labyrinth of the war, de Gaulle never
thought to establish a lasting dictatorship. Indeed, when the
opposition raised in May 1958 the specter of de Gaulle becom-
ing a dictator, he responded: “Would one believe that, at the
age of 67, I will begin a career of a dictator?”
He did not. In 1962 the constitution was amended to
allow for the direct election of the president. Yet in 1965, de
Gaulle was humiliated by failing to be elected in the first
round of the presidential election, even if he was re-elected
with 55.2 percent of the vote in the second round. Four years
later, on April 28, 1969, he resigned from office after a defeat in
a minor referendum. The institutions of the Fifth Republic
survived his departure from office and his erstwhile opponent,
François Mitterrand, became president in 1981.
The first lesson of the French crisis is again institutional:
given the relations of political forces, no government formed
under the institutions of the Fourth Republic was able to muster
the majority required to act decisively in the face of a disaster.
This situation was remedied by the constitutional
change, which strengthened the power of the chief executive
and stabilized the governments. The temporary powers granted
to the newly elected president were almost unlimited, yet he was
willing and probably compelled to tolerate opposition and thus

72
 some stories

to preserve democracy. Counterfactual questions again

highlight the role of contingencies: What would have hap-
pened had someone with personal authority over the mili-
tary not been available to become a civilian leader? What
would have happened had the leader been willing to use his
constitutionally acquired powers to eradicate the democratic
opposition? Such questions are unanswerable, but they
demonstrate that the survival of democracy in France may
have been a historical accident, as was its failure in Nazi
Germany.
5. A note on May 1968. The French events of May 1968 were
dramatic but I do not think that they threatened democracy.
They were certainly violent: some thousands of people were
injured, including 1,912 among the police forces (Le Gac,
Olivier, and Spina 2015: 524). Yet with millions of people on the
streets, massive strikes, barricades, and occupations of buildings,
the death toll was minimal (between four and seven).
No significant terrorist groups emerged in the aftermath. Even
if there were some dramatic moments, the situation was quickly
stabilized when the government overwhelmingly won the par-
liamentary elections of June 23–30, 1968.
To complete the story, Figure 3.4 shows the incidence
of unrest in France between 1945 and 1970, with the peaks in
1947, 1960–4, and 1968.

3.4 United States, 1964–1976

1. Democracy. The oldest democracy in the world, the
United States experienced several crises, most profoundly

73
 the past: crises of democracy

20
15
Unrest
105
0

1945 1950 1955 1960 1965 1970

Year

Figure 3.4. Unrest by year in France, 1945–70

the Civil War. Waves of repression occurred intermit-

tently: during the Civil War, in the aftermath of World
War I, and during and after World War II. Yet even the
most bloody war in history did not interrupt the regular
functioning of representative institutions: all elections
occurred on schedule and the Congress never ceased to
function.
2. Threats. Civil strife was widespread during the 1960s,
principally in the form of urban riots (Rochester, Harlem,
Philadelphia in 1964; Watts in 1965; Cleveland, Omaha in
1966; Newark, Plainfield, Detroit, Minneapolis in 1967;
Chicago, Washington, Baltimore, Cleveland in 1968).
Two political assassinations, of Martin Luther King and
of Robert Kennedy, occurred in 1968. The country was
embroiled in the Vietnam War, which was highly divisive

74
 some stories

80
60
Unrest
4020
0

1920 1940 1960 1980 2000 2020

Year

Figure 3.5. Unrest by year in United States, 1919–2012

internally. Mass demonstrations against the war began in

1964 and continued throughout the period. Repression
was widespread. In 1966, the House Committee on Un-
American Activities launched investigations into the
opponents of the war. On May 4, 1970, the National
Guard opened fire on students demonstrating against the
war at Kent State University, killing four and wounding
twelve demonstrators, while a few weeks later the police
killed two students and injured twelve at Jackson State
University. As Figure 3.5 shows, incidence of unrest was
high during the 1960s, culminating in 1968.
3. Signs. Richard Nixon was elected the 37th President of
the United States in 1968, and while the electoral cam-
paign was punctuated by violence, notably the “police

75
 the past: crises of democracy

riot” during the 1968 Chicago Democratic Convention, his

first term showed no signs of a crisis of democratic
institutions. As the election of 1972 approached, however,
Nixon began to use government agencies for partisan
purposes. He ordered surveillance of his potential oppo-
nent, Senator Ted Kennedy. An “enemies list,” with
expanding membership, was made by his staff, with the
purpose of targeting those on it using federal agencies.
One of his counsels, John Dean, made the purpose of this
list clear: “This memorandum addresses the matter of
how we can maximize the fact of our incumbency in
dealing with persons known to be active in their opposi-
tion to our Administration; stated a bit more bluntly –
how we can use the available federal machinery to screw
our political enemies” (memorandum by John Dean to
Lawrence Higby, August 16, 1971). The targeted enemies
were bugged and the Federal Bureau of Investigation, the
Central Intelligence Agency, and the Internal Revenue
Service were utilized to harass them. The operative unit
coordinating this campaign was the Committee for
Reelection of the President, popularly nicknamed CREEP.
Nixon was re-elected in 1972, but as news of the bungled
burglary of the Democratic National Committee in June 1972
was becoming public, the campaign by the administration
against the media and the courts intensified. Even though
the Vietnam War ended on January 27, 1973, the Watergate
scandal continued to embroil people around the president in
violations of legality. The commitment of the Nixon admin-
istration to defend its power by all means is evidenced by
the fact that sixty-nine of his supporters were eventually

76
 some stories

charged and forty-eight were convicted of illegal acts related

to the Watergate scandal, including two attorney generals,
the chief of staff, three White House staffers, the secretary of
commerce, and Nixon’s personal lawyer.
4. Outcome. The remarkable aspect of the crisis in the United
States is that the representative institutions, the system of
checks and balances, worked effectively to stop the abuse of
power by the president. The Senate voted seventy-seven to
zero to open the investigation of the Watergate incident and
subsequently the House initiated impeachment procedures
against the president. It is important to note that both houses
were controlled by the Democrats. Yet some Republican
senators and representatives also voted against the president.
Courts also played a role in the development of the crisis: the
decision of the Supreme Court to force Nixon to release his
tapes brought the administration to the precipice. Facing the
prospect of inevitable impeachment, President Nixon
resigned on August 8, 1976.
The obvious counterfactual question is whether the
institutional system would have counteracted the abuse of
power by the president had Republicans controlled both
houses of Congress. Does the system of checks and balances
work, as James Madison had hoped, because members of
different institutions defend the interests of their institutions,
or only if the powers of the government are divided between
different parties and the representatives defend their partisan
interests?

77
 4

Lessons from History: What to

Look For
Such absolute certainties have eluded me. I have found
only a series of ups and downs and a succession of
unforeseeable contingencies, none of which seem to have
been inevitable.
(Alon 2002: 11, about Weimar Germany)

If the past illuminates, the future depends on whether the

conditions we observe at the present mirror those of the past.
Hence, it is still premature to draw lessons. The past tells us,
however, what we should be looking for, what the signs are
that a democracy is in crisis and what kinds of events may
actually lead to its downfall.
If the past is a guide, we should be looking at eco-
nomic conditions: income, its growth, and its distribution.
We should consider the democratic history of a particular
country: how entrenched democracy is in terms of the habit of
changing governments through elections. We should pay
attention to the intensity of divisions in society: both the
extent of political polarization and of hostility between adher-
ents of different political solutions.
All the four cases in Chapter 3 indicate that we should
also pay attention to the particular forms of democratic insti-
tutions, in particular to whether they are conducive to the
formation of majority governments that can act decisively if
some disasters occur, yet without being able to usurp power.
Yet making counterfactual inferences is perilous. Would

78
 lessons from history: what to look for

democracy have survived in Germany under institutions

more favorable to stable majority governments? Would the
Chilean coup have been avoided under a parliamentary sys-
tem, in which the chief executive could be removed by insti-
tutional procedures? As much as political scientists believe in
the importance of institutions, perhaps the conflicts in these
two countries were just too intense to be peacefully managed
under any institutional arrangement.
Clearly, these directives are biased by what we can
observe. The intuitions from memoirs and even novels may
be as illuminating as from systematic data: they tell us how
individuals perceived and experienced the dramatic events in
which they were protagonists and, in the end, it is their
actions that determined the outcomes of crises.
Conditions do not determine the outcomes; actions
of people under the conditions do. Even when the conditions
are given, outcomes are not unique. For example, Stern (1966:
xvii) thinks that “By 1932, the collapse of Weimar had become
inevitable; the triumph of Hitler had not.” But perhaps by 1925
even the fall of the Republic was not inevitable: had the KPD
or the Bavarian People’s Party (BVP) voted for the Zentrum
candidate, Marx, in the second round of the presidential
elections, Hindenburg with his anti-democratic instincts
would not have been there in 1932 and who knows what
would have happened. One can entertain similar fantasies
about Chile: after the change of leadership of the Christian
Democrats resulting in the rejection of the Ley de las Areas,
perhaps the fall of Allende was inevitable but the brutal coup
was not. Conversely, one may wonder what would have hap-
pened in France had a leader with impeccable military

79
 the past: crises of democracy

credentials not been available or not been a democrat. I do not

think that we can answer such questions, so we must allow
that the lessons from history are paltry, and that the future is
not uniquely determined by present conditions – that it is
uncertain.

80
Part II

The Present: What Is Happening?

A candidate claiming to be a billionaire, who advocates low-
ering taxes and reducing social programs, is supported by the
working class, while a candidate who wants to tax the rich is
supported by the Wall Street Journal. A thrice-married man
who prides himself on unwanted sexual advances receives
almost unanimous support from religious groups committed
to “family values.” A lot of people believe any kind of apparent
nonsense. The incumbent party loses an election when the
economy is the best it’s been in recent decades. An election in
which almost all parties, including the victorious one, cam-
paign against “the establishment” generates a parliament that
is even more elitist than the outgoing one. A blow against
globalization, the free flow of capital and commodities, is
inflicted by parties on the right wing of the political spectrum.
Nationalists form international alliances. None of this makes
sense.
What is going on and why? What do we need to make
sense of if we suspect that democracy may be in crisis? I want
to make sense of the current political, economic, social, and
cultural transformations: what, if anything, do they add up to?
Yet “making sense” is a deceptive endeavor, guided by what
Pangloss tells us in Voltaire’s Candide: that there must be
a “because” for everything, everything must be logically con-
nected. As intellectuals, we seek hidden logical connections
among appearances: in Marx’s words, “If essence and

81
 the present: what is happening?

appearance coincided, no science would be necessary.” But

the danger is that we may overdo it, finding causal connec-
tions where none exist. While the quest for sense is inexor-
able, finding it is always perilous: our beliefs are replete with
false positives.
Moreover, it is not always obvious of what we should
be making sense, what are the “facts.” As Leo Goodman once
said, “A fact in fact is quite abstract.” Facts are constructed,
subject to interpretation, and often disputed. Which parties
should be considered to be radical Right? Does automation
reduce the demand for labor or are jobs that are substituted by
machines replaced by other jobs? Is there a “hollowing of the
middle class”? What is the marginal product of the CEOs?
Not only explanations but even facts cannot be taken for
granted.
In what follows, I invert the schema used to analyze
the past. I first describe the signs that a crisis may be here: the
collapse of traditional parties, the rise of the radical Right, and
of attitudes supporting it. Then I venture into possible expla-
nations: economic, cultural, and autonomously political.
Subsequently, I address the issues entailed in looking for
causality and focus on micro-level explanations. Finally,
I ask whether and which of the current conditions may be
historically unprecedented and ominous.

82
 5

The Signs

The signs that we may be experiencing a crisis include: (1) the

rapid erosion of traditional party systems; (2) the rise of xeno-
phobic, racist, and nationalistic parties and attitudes; and (3) the
decline in support for “democracy” in public opinion surveys.

5.1 Erosion of Traditional Party Systems

Party systems that endured without much change during almost
a century are eroding in many countries. The systems that
emerged in Western European and Anglo-Saxon countries in
the aftermath of World War I were typically dominated by two
parties, one left and one right of center. Parties bearing social
democratic, socialist, or labor labels occupied the space of the
moderate Left. The labels were more varied on the Right, but
each country had at least one major party located right of center.
These systems have remained almost ossified until recently.
While at times they changed labels, merging and splitting, they
survived not only the turmoil of the interwar period and World
War II but also the profound economic, demographic, and
cultural transformations of more than fifty years following
the war.
However we characterize this stability, it is astonishing.1
Very few parties that did not receive at least 20 percent of the

1
The following numbers and figures are based on countries that were
members of the OECD as of 2000, except for Greece, Italy, Portugal, and

83
 the present: what is happening?

vote in the elections closest to 1924 have broken this barrier

since. Liberals in 1929 in the United Kingdom and NSDAP in
1932 in Germany were the only ones to do so before 1939.
The immediate aftermath of World War II witnessed an upsurge
of the Left vote (Communist in France in 1945, Finnish People
Democratic League in 1945, Socialist in Japan in 1947). Between
1951 and 1978 only two parties, in Belgium and in France, crossed
for the first time the threshold of 20 percent. Yet from 1978 until
the moment this text is being written seventeen new parties
broke this barrier. One way to see this stability and its erosion
is that, in spite of the upheaval following World War II, a new
party crossed the threshold once every 7.6 years between 1924
and 1977 and once every 2.3 years after 1977.
Another way to characterize this stability and its
erosion is to consider the percent of the two top vote-getters
in each country around 1924 that remained in the top two in
the subsequent elections. Except for NSDAP in 1930, the two
top vote-getters remained in this position in all the countries
under consideration during the entire period until 1945.
The aftermath of the war shook their positions somewhat,
but almost 90 percent of the two 1924 leaders remained in the
top two until the late 1990s. A major destabilization in 1999
was largely overcome by 2007, but the 2008 financial crisis led
to another major shake-up. Chiaramonte and Emanuele
(2017) show that the movement of voters across parties has

Spain. The total number of countries is nineteen. Given the changes of

names, mergers, and splits it is sometimes necessary to make decisions
about which parties are heirs of the already existing ones and which are
new. The data cover the period through 2014.

84
 the signs

1
.9
Proportion
.8
.7

1920 1940 1960 1980 2000 2020

Year of observation

Figure 5.1. Proportion of parties that were the two top vote
winners around 1924 that remained in the top two

increased in the most recent period and that electoral volati-

lity is due mainly to the entry and exit of parties. Figure 5.1
illustrates these patterns.
This picture still underestimates the original sta-
bility as well as its recent erosion. It underestimates the
stability because several countries had a three- or even
four-party system in which the vote margins between the
parties were small, so that it was easy for them to change
places. But considering only party labels, rather than their
programs, does not take into account the general ideolo-
gical drift to the right, both of the center-left and the
center-right parties (see Maravall 2016). If we were to
consider programs, the recent destabilization would
appear more pronounced.

85
 the present: what is happening?

4
3.8
3.4 3.6
Number
3.2
3

1960 1980 2000 2020

Year of observation

Figure 5.2. Effective number of parties in the electorate since

1960, in countries that were members of the OECD as of 2000
Source: Armigeon et al. 2016 CDPS Lowess smooth

Finally, as Figure 5.2 shows, the effective number of

2
parties in the electorate has increased since the early 1980s,
again with an upturn during the past few years.
All these trends indicate that the traditional party sys-
tems are crumbling. But an argument can be made that this is
not a sign of a crisis but just a routine partisan realignment that
will result in a rejuvenation of democracy. Hopefully we may
still learn ex post that this is what it was. But at the moment all
we see is that the old party system, which has ossified over
seventy-five years, is crumbling, and that no stable new pattern

2
“Effective number of parties” is an index that weighs parties by their vote
(or seat) shares. Specifically, it is measured by 1/vi2, where vi is the vote
share of party i. For example, if the vote shares of three parties are 0.5, 0.4,
and 0.1, the effective number is 1/0.42 = 2.38.

86
 the signs

has yet crystallized. Hence, this is a crisis: the old is dying and the
new is not yet born. Moreover, if a realignment does ensue, it
will include the rise of xenophobic parties that have little
patience for democratic norms. As Piketty (2018) emphasizes,
given multidimensional divisions of the electorate, different
coalitions may emerge. Specifically, he speculates that in
France and the US the most likely realignment is one of “glob-
alists” against “nativists,” while in Britain a “two-elite model” –
wealthy against educated – is likely to persist. Note that this
phenomenon is almost universal among developed democra-
cies, so something strange is going on.

5.2 The Rise of Right-Wing Populism

The general mood is populist. Populism is an ideological twin
of neo-liberalism. Both claim that social order is spontaneously
created by a single demiurge: “the market” or “the people,” the
latter always in singular, as in “le peuple,” “el pueblo,” or “lud.”
Neither sees a role for institutions: spontaneity suffices.
No wonder they appear together on the historical scene.
Many emergent parties portray themselves as “anti-
system,” “anti-establishment,” or “anti-elite.” They are “popu-
list” insofar as the image of politics they project is one of an
“elite” (“casta,” cast, in the language of the Spanish Podemos;
“swamp” in the language of Donald Trump) that betrays, abuses,
or exploits undifferentiated “people” (Mudde 2004: 543). Such
claims originate on the Left as well as the Right (Rooduijn and
Akkerman 2017). Indeed, as the French 2017 elections show, they
can also emerge from the center, even if ironically the parliament
that resulted from this election is even more elitist in social terms

87
 the present: what is happening?

than the outgoing one, just including fewer professional politi-

cians. The populist parties are not anti-democratic in the sense
that they do not advocate replacing elections by some other
method of selecting governments. Even when they express
a yearning for a strong leader, they want leaders to be elected.
Political forces that question democracy do exist but they are
completely marginal. In turn, these parties, again on the Left as
well as on the Right, claim that the traditional representative
institutions stifle the voice of “the people” and call for some new
form of democracy that would better implement “popular sover-
eignty” (Pasquino 2008) and bring governments closer to “the
people” (Canovan 2002). Popular initiative referendums are their
favorite, but otherwise their projects for constitutional reforms
are vague. Still, the populist image of politics is associated with
the rejection of representative democracy and its replacement by
a different, “direct” one. Hence, while the populist parties are not
anti-democratic, they are anti-institutional in the sense of reject-
ing the traditional model of representative democracy.
As a Mexican presidential candidate, Manuel López Obrador,
exclaimed in the aftermath of his defeat in 2006, “to hell with
your institutions” (“al diablo con vuestras instituciones”).
On economic issues, left-wing parties are resolutely
egalitarian. Those on the Right are more ambivalent: they want
to retain the support of the traditional petite bourgeoisie, which
wants lower taxes and a flexible labor market, while recruiting
industrial workers, who want more job protection and more
income redistribution (Iversflaten 2005). Both extremes are
highly protectionist (Guiso et al. 2017, Rodrik 2017). Moreover,
they oppose globalization and are strongly anti-Europe.
The result is that at least in some countries, the economic policies

88
 the signs

of the radical Left and Right do not diverge by much. For

example, comparing the electoral programs of the extreme Left
candidate in the 2017 French presidential elections, Jean-Luc
Mélenchon, and of Marine Le Pen, shows convergence on eco-
nomic, social welfare, workers’ rights, and protectionism issues.3
The similarities, however, end there. The sharp differ-
ence is with regard to immigration, immigrants, xenophobia,
and racism. Some populist parties – Podemos in Spain, Syriza in
Greece – are open to the coexistence of multiple cultures, view
immigrants as net contributors to the economy, and take
a strong stand against racism. In turn, the parties standardly
referred to as “extreme” or “radical” Right are nationalist and
xenophobic, or “nativist.” They also tend to be racist and repres-
sive. They adopt electoral strategies that emphasize the salience
of “immigration” (Arzheimer 2013). Defending “national
values” – a favorite phrase of Marine Le Pen – they advocate
excluding immigrants from publicly provided social services,
nationalistic indoctrination in education, banning Halal foods
in school cafeterias, a dress code, etc. To this extent, they are
authoritarian. With some unease, I follow Golder (2016) in using
the label of “radical Right” to denote such parties.
While one may quibble about classifying particular
parties, the trend is manifest. Figure 5.3 portrays the rise of
radical Right parties in different sets of European and Anglo-
Saxon democracies.4 This picture, however, hides important
differences among countries.

3
See www.leparisien.fr/elections/presidentielle.
4
Armingeon et al. (2016) use the label “populist right” and lump together
what Golder (2016) would distinguish as “extreme” and “radical” Right.

89
 the present: what is happening?

10
8 With Eastern Europe
Percentage
6
4

OECD-2000 only
2
0

1960 1980 2000 2020

Year of observation

Figure 5.3. Average electoral support for radical Right, by year

Source: Armigeon et al. CDPS 2016, with modifications for
Hungary and Poland.Lowess smooth

The highest current share of radical Right parties are in

Switzerland, Austria, and Denmark, where they exceed 20 per-
cent. In Austria and France, radical Right candidates won more
than 25 percent of votes in the first rounds of presidential elec-
tions. In turn, in five countries such parties either do not exist or
get no votes at the present. The trends are not homogeneous
either: radical Right parties gained strength only most recently in
Norway, Sweden, and Germany, while they peaked some time
ago in Belgium, Italy, and Japan. The open question is how to
treat the Republican Party in the United States. It now satisfies all
the criteria most scholars use to classify parties as radical Right,
even if Armingeon et al. (2016) do not classify it as such. More

They do not consider Fidesz in Hungary, Law and Justice (PiS) in Poland,
and UKIP as “Right,” which I do. The data end in 2014.

90
 the signs

generally, this classification does not take into account move-

ments of traditional right-wing parties toward the extreme,
which is perhaps why Armingeon et al. (2016) do not classify
the Hungarian Fidesz and the Polish PiS as radical Right.

Table 5.1 Share of votes of radical Right (countries that were members
of the OECD as of 2000)

Maximal Last Last

Country sharea Period parliamentary presidentialb

Austria 28.2 2008–12 26.8 35.1

Belgium 14.0 2007–9 3.7
Denmark 21.1 2015– 21.1
Finland 19.1 2011–14 17.7 9.4
France 14.9 1997–2001 14.4c 26.0c
Germany 12.6 2017– 12.6
Greece 14.4 2012–14 10.7d
Iceland 3.0 2013–16 0.0
Italy 25.8 1996–2000 4.1
Japan 14.9 2012–13 2.1
Luxembourg 2.3 1989–1999 0.0
Netherlands 17.0 2002 13.1
Norway 16.3 2013–17 15.2
Spain 2.1 1979–81 0.0
Sweden 12.9 2014– 12.9
Switzerland 28.9 2007–10 26.6
United Kingdom 3.1 2010–17 1.8

Note: As of October 15, 2017. The radical Right has never won any votes in
Australia, Canada, Ireland, New Zealand, or Portugal. (a) Maximal share of
votes before the most recent parliamentary election, to the lower house if
there is more than one. First-round results are reported for France. (b) Only
where president is directly elected. (c) Front National + Debout La France.
(d) Golden Dawn + ANEL.
Source: Armingeon et al. (2016), updated by own research.

91
 the present: what is happening?

30
SocDem
20

Conservative

Liberal
Religious
10

Communist
Right
0

1960 1980 2000 2020

Year of observation

Figure 5.4. Vote shares of parties by years in countries that were

members of the OECD before 2000
Lowess smooth. Data source: Armingeon et al. 2016 CDPS

Traditional parties lost electoral support among the

potential voters, while the support for the radical Right has
crept up. But is it because political opinions became more
polarized, with voters moving to the extremes, or because tradi-
tional parties lost touch with their supporters? The crumbling of
traditional parties need not entail an erosion of centrist, moder-
ate preferences but just disgust with the parties themselves.
When people believe that all professional politicians are the
same, self-serving, dishonest, or corrupt, they turn against
them whether they locate themselves on the left, right, or center.
Hence, the erosion of traditional parties need not signify an
erosion of the center.
The decline of traditional parties is manifest. In
Figure 5.4 the parties are, from top to bottom, the leading social
democratic, conservative, liberal, religious, and Communist

92
 the signs

85
All
80
Turnout

OECD-2000 only
75 70
65

1960 1980 2000 2020

Year

Figure 5.5. Turnout by year

Lowess smooth. Data source: Armingeon et al. 2016 CDPS

parties, as classified by Armingeon et al. (2016), while the lowest,

rising, trend is for the parties of the radical Right. Perhaps
surprisingly, this erosion of support for the traditional parties
coincided with sharply declining turnouts (Figure 5.5).
This is not just a coincidence. Guiso et al. (2017) point
out that if the decision to vote and the direction of the vote share
are common determinants, one should expect the relation
between turnout and right-wing vote to be negative. Within-
country regressions of the vote shares of the radical Right on
turnout in Figure 5.6 show that among the ten pre-2000 OECD
members in which the radical Right exists, only in Denmark is
the slope positive.5

5
In the pooled data including all the countries, fixed-effects OLS regression
generates the 95 percent confidence interval of the coefficient as [−0.168,

93
 the present: what is happening?

30
Radical Right vote share
20
10

Denmark
0

40 60 80 100
Turnout

Figure 5.6. Turnout and radical Right vote share in ten developed
democracies
Linear fit. Data source: Armingeon et al. 2016 CDPS

One cannot tell from the available data which part of the
increase of electoral shares of the radical Right is due to an
increase in the numbers of its supporters and which to the
growing abstention of centrist voters. Yet it may well be that the
increasing share of the radical Right is due more to the abstention
of centrist voters than to an increase of extreme voters.
Why would centrist voters withdraw from the electoral
process? There are two, not necessarily rival, hypotheses. One
goes like this. The stagflation crisis of the 1970s, followed by the
victories of Thatcher and Reagan, pushed traditional right-wing

−0.095; N = 1571]. Given that several countries do not have any radical
Right parties, I also estimated a random effects Tobit regression, which
gives an even more negative coefficient: [−0.487, −0.274; 453 uncensored
observations].

94
 the signs

45
40
Distance
30 25
2035

1940 1960 1980 2000 2020

Year

Figure 5.7. Ideological distance between center parties, by year

Data from the Manifestos Project, courtesy of Jose Maria
Maravall. The ideological scale ranges from −100 on the left to
+100 on the right. Countries include Western Europe plus
Australia, Israel, Japan, and New Zealand. Lowess smooth

parties to the right. For some reason, perhaps because of the

economic fiasco of the first year of Mitterrand’s government, the
Social Democrats followed by also making a “virage” to
the right, embracing the language of “trade-offs” between equal-
ity and efficiency, fiscal discipline, and flexible labor markets.
As a result, the ideological distance between the two major
center-left and center-right parties has decreased sharply during
the post-war period, perhaps with a slight upturn following the
crisis of 2008, as shown in Figure 5.7.
Yet the convergence of party platforms on the left–right
dimension is not the only plausible explanation. Already Lipset
(1960) argued that political attitudes are two-dimensional, with

95
 the present: what is happening?

the second dimension being “authoritarianism.”6 According to

Albright (2010: 714), the single left–right dimension “is steadily
diminishing in its ability to summarize party behaviour.” While
economic issues still constitute the most important dimension
along which parties compete in most countries (Huber and
Inglehart 1995, Wagner 2012), social and cultural issues have
gained in importance since the 1970s (Inglehart and Flanagan
1987). Moreover, it has been argued that in many countries the
cultural and the economic dimensions do not neatly correlate
with each other anymore, so that the political space cannot be
characterized by a single left–right axis but has to be depicted as
two-dimensional (Kitschelt 1994, Kriesi et al. 2006, 2012, Marks
et al. 2006). Brady, Ferejohn, and Papano (2017), for example,
find in a study of seven countries that traditional parties adopt
more pro-immigration policies than their supporters and attri-
bute the weakening of these parties to this distance: “immigra-
tion has driven a wedge between the major parties – those that
regularly play a role in government – and their supporters and
that this wedge opens up enormous space for new movements
either inside existing parties or outside.” For a long time, the
rhetoric of the radical Right has been that “they are taking jobs
from you,” while recently it has become more along the lines of
“you are paying for them,” that the “middle class” is paying for
the poor, particularly the immigrants, and particularly those
with a different skin color. “Illegal immigrant households receive
far more in federal welfare benefits than native American

6
On general issues concerning bi-dimensionality and party strategies in
the presence of a second dimension, see the special issue of Party Politics
(2015, vol. 21(6)), with an introduction by Elias, Szocsik, and Zuber (2015).

96
 the signs

households,” Trump wrote in a 2016 Facebook post, “I will fix

it.” As Brady, Ferejohn, and Paparo (2017: 3) put it, “immigration
puts a recognizable face on events that may well be properly
attributable to other forces.” Hence, the alternative story is that
whatever the left–right distance among them, the traditional
parties increased their distance from voter preferences on the
immigration issue, thus alienating their supporters.
An open issue is why the center parties would remain
distant from voters on the second dimension, whatever it is.
A plausible explanation is provided by Dancygier (2017).
Accommodating xenophobic preferences is costly for these par-
ties in terms of votes because it causes people she refers to as
“cosmopolitans” to move out of the electorate. Hence, center
parties face a trade-off between winning the votes of some sectors
of the potential electorate and losing them from other sectors.
They adopt xenophobic postures when it is electorally advanta-
geous and refrain from appealing to such attitudes when it would
lead to the erosion of their traditional support. Even if they
maximize their potential vote shares, in either case they face limits
to how far they can move. Hence, in equilibrium, they still remain
distant from some voters on the cultural dimension.
Before summarizing, it is instructive to look in more
detail at a particular country, namely France. First, while a large
majority, 71 percent, up from 57 percent in 2013, of French survey
respondents now agree that “the notions of the Left and the Right
are obsolete,” 94 percent are still able to locate themselves on this
dimension (Hastings 2018). So can 91 percent of Europeans
(excluding Russia) in general (Cautres 2018). During the past
forty years France experienced several partisan alternations in
office, all governments focused on reducing unemployment, and

97
 the present: what is happening?

yet unemployment never fell below 9 percent. Hence, as

Teinturier (2018: 65) reports, voters ask themselves whether
politics has any effect on their lives. Since 2013, between 75 and
83 percent of the French declared that “The democratic system
functions rather badly in France. I have an impression that my
ideas are not well represented.” Moreover, about two-thirds
agree that “Most politicians are corrupt” and between 83 and
89 percent that “They act principally in their personal interests.”
Politics evokes “disappointment” among 40 percent, “disgust”
among 20 percent, “anger” among 13 percent, and “indifference”
among 9 percent (all these numbers are from Teinturier 2018).
Electoral abstention in legislative elections has increased sharply
since the 1980s and in presidential elections since 2007. Together
these patterns indicate that while the left–right dimension
remains as salient as it was in the past, most people are just
disgusted with the traditional parties.
At the same time, there is a general perception that the
issue space is not unidimensional. Following Inglehart, Foucault
(2018) sees the second dimension as broadly cultural, but without
specifying its components or showing its independence of the
economic dimension. In a daring novel (Soumission, 2015),
Michel Houellebecq raises the specter of a confessional,
Catholic-Islam coalition opposed to a secular, republican one.
The only piece of hard evidence I could find is from Piketty
(2018), who uses exit polls to classify voters according to their
positive or negative attitudes toward redistribution and immi-
gration, and shows that in 2017 they divided almost equally
among the four cells of this two-by-two table. Hence, the evi-
dence is that immigration divides people independently of the
left–right dimension, but it is not clear what else does.

98
 the signs

Finally, the result of the 2017 presidential election was

a debacle for the traditional Left. Among their usual consti-
tuencies, the share of the Left vote among people between
eighteen and thirty-nine years old fell from 31 percent in 2012
to 7 percent; among people with more than high-school
education from 33 to 7 percent; among public employees
from 41 to 8 percent. But it seems that most of the vote the
Socialists lost was split between the extreme Left and the
center, not benefiting the Right. In turn, while the 2017 elec-
tion was the first one in which more workers voted for the
extreme Right (Front National) than the Left, the largest party
among them are non-voters (based on Foucault 2018).
In the end, even with all these data it is difficult to tell
to what extent the recent political transformations in France
are due to the general disaffection with the traditional parties –
a crisis of representation – and to what extent they are due to
an emerging salience of some second dimension that divides
people independently from the economic one. Piketty (2018:
26–7) reports that the proportion of voters who say that there
are “too many” immigrants in France has actually declined
over time, as has the salience of the religious dimension.
Hence, it is not clear whether the virtual disappearance of
the traditional center-left and center-right parties is due to
voters’ disgust with politicians or to their distance from voters
on the dimension of immigration.
More generally, we can see that support for tradi-
tional center parties has crumbled across Europe and that
some centrist voters withdrew from the electorate, while the
vote shares of radical Right parties, but not necessarily abso-
lute numbers of their supporters, have increased. To what

99
 the present: what is happening?

extent these transformations are due to a general rejection of

parties and politicians and to what extent to the rise of
some second, “cultural,” dimension is difficult to weigh.
Moreover, to repeat the caveat of Section 5.1, the rejection of
party politics may be just a transitory phenomenon: new
center parties may replace the traditional ones, mobilize cen-
trist voters, and deter a further move of the electorate toward
the radical Right, as at least for now seems to be happening in
France. Yet it is also possible that the center will continue to
erode and xenophobic, populist parties will continue to gain
strength, or that the traditionally centrist parties will success-
fully prevent the electoral rise of the radical Right only by
moving to that position themselves.

5.3 Decline of Support for Democracy in

Surveys
All kinds of surveys are cited as evidence of declining support for
democracy: “democratic backsliding” or “democratic deconso-
lidation.” In particular Foa and Mounk (2016) find it alarming
that in the six countries they examined, younger people find it
less “essential to live in a democracy.”7 Armingeon and
Guthman (2014) examined seventy-eight surveys in twenty-six
European Union countries to compare support for democracy
in 2007 and 2011. They found that this support fell in twenty
countries and increased in six, with the total mean declining by
7.2 points. Countries that were most affected by the 2008 crisis,

7
For a debate on Foa and Mounk, go to http://journalofdemocracy.org/o
nline-exchange-%E2%80%9Cdemocratic-deconsolidation%E2%80%9D.

100
 the signs

notably Greece and Spain, are where this support fell most.
Similar results about the effect of the 2008 crisis emerge from
surveys conducted by the World Values Studies that ask people
whether they have confidence in democracy, experts, the army,
or strong leaders, albeit with more heterogeneous patterns over
the longer run, and with the United States showing the sharpest
decline in the relative standing of democracy since the 1994–8
period (Weakliem 2016a). Surveys also show declining confi-
dence in other, not just representative, institutions. At least in
the United States, confidence also declined sharply for news-
papers, television, banks, big business, religion, schools, and the
medical system (Weakliem 2016b, based on data from Gallup).
Are these numbers signs of a crisis of democracy?
If a crisis is defined by these numbers, then this is just
a tautology, albeit one that is frequently made by those who
produce them. But should we take them as harbingers of
a collapse of democracy? Titles of popular articles in which
these numbers “ring bells for democracy” are ubiquitous. Yet
while such numbers are disheartening, there is not a shred of
evidence that they predict anything. Six months before the coup
in Chile only 27.5 percent of respondents thought that “a military
coup is convenient for Chile” (Navia and Osorio 2018). Whether
democracy requires democrats, whether its continued existence
depends on individual attitudes, is a controversial issue. Even if
it does, the causal relation between answers to survey questions
and the erosion of democracy must depend on the actions of
organized political groups.
Responses to survey questions are informative but not
predictive. For one, no one knows what people in different
countries and at different times understand by “democracy”

101
 the present: what is happening?

when they are asked whether “democracy” is the best form of

government or whether it is essential that their country be
governed “democratically.” While elites see democracy in insti-
tutional terms, several surveys indicate that mass publics often
conceive of it in terms of “social and economic equality.”
Moreover, even if recent surveys indicate that many people
would want to be governed by “strong leaders” and many others
by non-partisan “experts,” does it mean that they do not want to
have a voice in choosing the leaders or the experts? The taste for
selecting governments through elections is an acquired one, but
it is addictive once acquired. Wanting governments to be effec-
tive, hoping that they will be competent and effective in improv-
ing people’s lives, does not imply abdication from the right to
choose them and to replace them when they fail. Finally, with all
the variations in the support for democracy shown by surveys
conducted in different developed countries over the past thirty-
five years, democracy collapsed in none of them. We may be
worried when few people declare confidence in political parties,
parliaments, or governments, when the belief that democracy is
the best system of government declines among the mass public,
or when the yearning for strong leaders or the rule by experts
increases. But the predictive power of answers to such questions
for the outright collapse of democracy is null. One should not
draw inferences about the survival of democracy from answers
to survey questions.

102
 6

Potential Causes

Here’s an Irish joke. A couple of tourists gets lost while

trekking in Ireland. They ask a peasant cultivating his field,
“How do we get from here to Dublin?” He responds,
“First, you do not begin from here.” Where to begin the
explanations? Globalization, technological change, break-
down of class compromise, immigration, authorization of
prejudices by some insurgent politicians, or something still
else? The purpose of this chapter is just to catalogue the
potential explanations, without attempting to adjudicate
among them. Issues entailed in identifying causality are
raised in Chapter 7.

6.1 The Economy: Income Stagnation,

Inequality, and Mobility
The instinct is to start with the economy, and this is where
I begin. The economic developments of the past decades can
be grossly characterized by three transformations that gener-
ated two effects. These transformations are: (1) decline of
growth rates of the already developed countries; (2) increase
in income inequality among individuals and households, as
well as a declining labor share in manufacturing; and (3)
decline of employment in industry and the rise of the service
sector, particularly of low-paying service jobs. Here is some
evidence.

103
 the present: what is happening?

.05
.04
.03
.02
.01

1940 1960 1980 2000 2020

Year

Figure 6.1. Rate of growth of per capita income by year of

countries that were members of the OECD before 2000
The irregular line is a lowess smooth, the fitted line fractional
polynomial regression with 95% confidence interval

The rates of growth of developed democracies, which

I take as countries that were members of the OECD before
2000, declined from about 4 percent in the aftermath of
World War II to about 2 percent currently. Figure 6.1 shows
the annual averages and the trend. As shown in Figure 6.2, the
average within-country inequality increased sharply (the pic-
ture looks almost identical for countries that were members of
the OECD before 2000). Figure 6.3 shows that the average
labor share took a precipitous dip from about 1980. In turn, as
shown in Figure 6.4, the average employment in industry
declined over time in absolute terms in the developed democ-
racies, while employment in services increased.

104
 potential causes

44
42
Average Gini
40
38
36
34

1960 1980 2000 2020

Year

Figure 6.2. Average Gini coefficient of pre-fisc incomes in Europe,

Japan, Australia, New Zealand, by year
Data source: Armingeon et al. 2016 CDPS. Lowess smooth
.64
Average labor share
.62
.6
.58
.56

1940 1960 1980 2000 2020

Year

Figure 6.3. Average labor share by year among countries that were
members of the OECD before 2000

105
 the present: what is happening?

2000 4000 6000 8000 10000

services

industry

agriculture
0

1960 1980 2000 2020

Year

Figure 6.4. Average employment by sector over time, absolute

numbers
Lowess smooth. Source: Armingeon et al. 2016 CPDS

The first effect of the combination of declining growth

rates with increasing inequality is the stagnation of lower
incomes, which has been exceptionally long-lasting in the
United States, portrayed in Figure 6.5. The picture is somewhat
different in the remaining OECD-2000 countries. Figure 6.6
shows that while the distance between the income of the top
and bottom 10 percent of recipients increased sharply from the
1980s onwards, incomes below the median continued to creep
up until all incomes were hit by the crisis of 2008.
The second effect is the erosion of the belief in mate-
rial progress. According to the Pew Research Center (Spring
2015 Global Attitudes Survey), 60 percent of respondents in
the United States and 64 percent in Europe now believe that
their children will be worse off financially than they are.

106
 potential causes

Income in thousands (2011 dollars) Recession

200
180 $186,000
95th
160
140 $111,900 $143,600
90th
120
100
$88,600
80
50th (median)
60
$42,100 $50,100
40
20 10th
$12,000
$9,600
0
1967 1975 1980 1985 1990 1995 2000 2005 2011

Figure 6.5. Real household income at selected

percentiles, 1967 to 2011
Source: United States Census Bureau, public
domain.

Moreover, these are not just perceptions. Chetty et al. (2016)

calculate that in the United States 90 percent of thirty-year-
old offspring were better off than their parents at the same age
in 1970, while in 2010 only 50 percent were. This collapse of
the deeply ingrained belief in intergenerational progress is
a phenomenon at a civilizational scale. The expectation of
material progress has been an essential ingredient for Western
civilization during the past 200 years. Since about 1820 every
generation in Europe and the United States lived and
expected to live better than their parents, and yet this belief
is being shattered. This is certainly a transformation that can
have profound cultural and political consequences.
Why did these economic transformations occur? Two
hypotheses are straightforward and plausible. One is “globa-
lization,” the combination of the liberalization of commodity

107
 the present: what is happening?

20000 40000 60000 80000 100000

Top 10%

Median

Bottom 10%
0

1940 1960 1980 2000 2020

Year

Figure 6.6. Average incomes of selected groups, OECD-2000

countries excluding the United States
Calculated from PWT9.0 and WIDER, in 2011 PPP USD Lowess
smooth

and capital markets, combined with the Chinese reforms.

The second is an autogolpe (self-coup) of the bourgeoisie,
the breakdown of class compromise. Both of these events
have a definite date, more or less 1978–80, so they cannot be
distinguished by rough timing. They may or may not be
associated but I discuss them separately.
The effect of China is a subject of much controversy.
Autor et al. (2013; see also Acemoglu et al. 2016) conclude that
rising imports cause higher unemployment, lower labor-force
participation, and reduced wages in local labor markets that
house import-competing industries. They attribute one-
quarter of the contemporaneous aggregate decline in United
States manufacturing employment to import competition from
China. Yet Rothwell (2017) questions the Autor et al. estimates,

108
 potential causes

concluding that foreign competition does not appear to

elevate the risk of job loss to a greater extent than domestic
competition, and that people living in the communities most
exposed to foreign competition are no worse off on average.
Rothwell and Diego-Rosell (2016) conclude that
“Surprisingly, there appears to be no link whatsoever
between greater exposure to trade competition or competi-
tion from immigrant workers and support for nationalist
policies in America, as embodied by the Trump campaign.”
In turn, Miao (2016) points out that import competition
lowers prices, and identifies a significant welfare gain from
trade with China. Moreover, in a recent review of the litera-
ture, Helpman (2016) concludes that increased wage
inequality is due mainly to factors other than commodity
trade. Hence, economists need to sort out their disagree-
ments. What is clear is that some people lost as a result of
globalization and were not compensated by redistributive or
other policies (Rodrik 2017).
An alternative explanation is the breakdown of
class compromise. The most startling picture is for the
United States, portrayed in Figure 6.7. The same is true
for some, albeit not all, other advanced economies after
1999, shown in Figure 6.8. Other sources show the same to
have occurred in Germany as of 1997, Japan as of 2002, and
the UK as of 1988.

Until about 1978, increases in wages almost exactly

followed increases in productivity, so that the functional
distribution of income was stable. Industrial workers were
organized by unions protected by the state and, with almost

109
 the present: what is happening?

300% 1948–1973: 1973–2014:

Productivity: 96.7% Productivity: 72.2%
Hourly compensation: 91.3% Hourly compensation: 9.2%
Cumulative percent change
250
238.7%

200
Productivity
since 1948

150

109.0%
100
Hourly compensation

50

0
1960 1980 2000
Note: Data are for average hourly compensation of production/nonsupervisory workers in the private sector and
net productivity of the total economy. “Net productivity” is the growth of output of goods and services minus
depreciation per hour worked.

Figure 6.7. Disconnect between productivity and a typical

worker’s compensation, 1948–2014
Source: Economic Policy Institute.

full employment, unions had monopoly power over labor

markets. Anticipating that excessive wage demands
would cause firms to invest less, wherever they were
sufficiently centralized, unions exercised wage restraint.
Government policies were subject to the same constraint
as unions, namely that excessive income taxation would
reduce investment, and thus future consumption.
In turn, facing moderate wage and taxation demands,
firms not only invested but could also live with unions
and with democracy. As a result, a “democratic class
compromise” naturally emerged. Governments managed
this compromise by regulating markets, providing social
services, and offering incentives for investment and
innovation.

110
 potential causes

Real wage index (base year = 1999) Labour productivity index (base year = 1999)
120

115

110

105

100

95
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Note: Labour productivity is defined as GDP per employed person and used GDP in constant 2005 PPP$ for all countries.
G20 advanced economies include: Australia, Canada, France, Germany, Italy, Japan, the Republic of Korea, the United
Kingdom and the United States. Both indices are based on a weighted average of all the countries in the group that
takes into account labour productivity and the size of paid employment.

Figure 6.8. Productivity and wage index (G20 advanced

economies)
Source: ILO.

This compromise was shattered in the United

Kingdom and the United States by the respective victories of
Thatcher and Reagan, whose first targets were the unions,1
and eroded more gradually in most other countries.
As a consequence, as shown in Figure 6.9, the average union
density dropped by more than ten percentage points between
the peak in 1980 and 2010. Perhaps the most consequential
policy of the Thatcher government was the stealth opening of
capital account, which changed the trade-offs between

1
“Unions” was the most frequently used word both in the 1979
Conservative Manifesto and in Thatcher’s electoral campaign. Under the
combined pressure of unemployment and of hostile legislation, the trade
union movement was seriously weakened, losing 17 percent of its
membership in five years.

111
 the present: what is happening?

50
45
Density
40
35

1960 1980 2000 2020

Year

Figure 6.9. Union density by year in countries that were members

of the OECD before 2000
Lowess smooth. Source: Armigenon et al. 2016 CPDS

redistribution and growth and thus forced both major poli-

tical parties to reduce the extent of the redistribution they
proposed (Dunn 2000). The opening of capital account was
not an issue in the election of 1979 when Mrs. Thatcher came
to office. Yet once the decision was made, the entire spectrum
of feasible policies was moved. It bears emphasis that this
offensive by the Right was premeditated, planned, vigorously
promoted by all kinds of think tanks, and coercively spread by
the influence of the United States in the international financial
institutions, codified as the “Washington Consensus.”
Whatever the causes, plus automation, these pro-
cesses generated winners and losers. For future reference, it
makes sense to distinguish: (1) actual losers, those who lost
stable jobs with living wages in industry and either moved to
lower-paid services or were forced to retire or became long-

112
 potential causes

term unemployed; (2) prospective losers, those who fear this

fate; (3) non-winners, principally the self-employed, the tradi-
tional petite bourgeoisie whose material conditions did not
change much one way or another; and (4) winners, the reci-
pients of profit incomes, however disguised.

6.2 Divisiveness: Polarization, Racism, and

Hostility
When thinking about the intensity of political divisions, we
need to consider two distinct aspects. (1) Distributions of
preferences over some general policy dimension
(liberal–conservative in the United States, left–right in
Europe) or over specific issues, such as immigration. These
distributions can be characterized in terms of “polarization”:
a population is polarized if individual preferences divide
people into clusters that are internally homogeneous and
distant from each other (Esteban and Ray 1994). (2)
The actions that people with particular preferences are or
are not willing to engage in with regard to members of other
group(s). This is important because people with the same
ideological profile may have different postures toward those
with whom they disagree and may be willing or not to engage
in hostile acts against them.
The ideological distance of party supporters in the
United States, portrayed in Figure 6.10, has sharply increased
in the past twenty-three years. Whether the same is true
across the European countries is more difficult to diagnose
because of the prevalence of multi-party systems, in which
people sort themselves out around parties occupying several

113
 the present: what is happening?

Distribution of Democrats and Republicans on a 10-item scale of political values

1994 2004 2017

MEDIAN MEDIAN MEDIAN MEDIAN MEDIAN MEDIAN
Democrat Republican Democrat Republican Democrat Republican

Consistently Consistently Consistently Consistently Consistently Consistently

liberal conservative liberal conservative liberal conservative

Notes: Ideological consistency based on a scale of 10 political values questions (see methodology). The light grey area
in this chart represents the ideological distribution of Democrats and Democratic-leaning independents; the dark grey
area of Republicans and Republican-leaning independents. The overlap of these two distributions is shaded black.

Figure 6.10. Democrats and Republicans more ideologically

divided than in the past
Source: Pew Research Center, public domain (see www
.pewresearch.org/terms-and-conditions).

positions of the left–right spectrum. Given the availability of

several parties, one would expect that supporters of each are
more homogeneous, but the overall distance between them is
more difficult to characterize. Indeed, the evidence that voters
moved away from the center is ambiguous outside the United
States. The distribution of individual positions on the
left–right dimension, studied by Medina (2015: figure 1) in
eighteen European countries, tends to be trimodal, with
a large mode at the center and small modes left and right of
center. Between 2002–3 and 2008, the mean position shifted to
the left in six countries, to the right in six, and in six it
remained statistically indistinguishable. In terms of polariza-
tion, the size of the center mode decreased in seven countries

114
 potential causes

A larger share of Republicans are saying immigrants are “a burden” to American

society.

Republican Democrat

63%
57%
60

40

25%
20

1994 2015

Figure 6.11. Immigration wasn’t always a partisan issue

Source: Pew Research Center, public domain (see www
.pewresearch.org/terms-and-conditions).

(Belgium, Czech Republic, Denmark, Finland, Germany,

Poland, and Slovenia), increased in three, and remained the
same in eight. In turn, Moral and Best (2018) found that the
polarization of citizens increased in Australia, Denmark,
Sweden, and the United States, but decreased in Germany
and the Netherlands. Hence, even if in some countries people
moved away from the center, there is no general European
trend.
As shown in Figure 6.11, the increase of polarization is
particularly evident with regard to immigration. Immigration,
in some countries specifically the inflow of refugees, is also the
most salient and divisive issue in Europe. The distribution of

115
 the present: what is happening?

European Attitudes on Immigrants: Racial differences

Bar chart shows % of population saying “Allow NO Muslim Immigrants”
Czech Republic 18% 56% 63%
Hungary 31% 51% 62%
Estonia 11% 41% 51%
Lithuania 20% 38% 47%
Poland 13% 27% 33%
Portugal 25% 32% 42%
Ireland 12% 26% 44%
Austria 12% 22% 26%
Spain 11% 20% 27% Allow No Muslim immigrants
Slovenia 14% 20% 31%
Belgium 11% 19% 31%
Allow No Gypsies
United Kingdom 6% 17% 29%
Allow No Jewish Immigrants
Finland 5% 17% 23%
Netherlands 4% 14% 17%
Switzerland 5% 13% 19%
France 6% 13% 20%
Denmark 3% 11% 25%
Norway 2% 8% 18%
Germany 3% 8% 14%
Sweden 1% 4% 5%
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65%
% of population (citizens only) saying “Allow no Immigrants”

Figure 6.12. European attitudes to immigrants: racial differences

Source: European Social Survey.

postures toward immigration is clearly bimodal across Europe

(Spoon and Kluwer 2015). Moreover, while attitudes toward
immigration vary across European countries, it is striking that
survey respondents distinguish the potential immigrants by
ethnicity or race: as shown in Figure 6.12, Gypsies are less
desirable in almost all countries than Muslims, who are in turn
less desirable than Jews.
The language of “immigration” used by the Right
amalgamates two distinct issues. One is control over the
current flow of foreigners across borders, which is the stan-
dard-bearer of the language of “national sovereignty.” But
note that the current net flow between the United States and

116
 potential causes

Mexico is southward: according to Pew Research Center

(2015), between 2008 and 2014 the Mexican population in
the United States fell by 140,000. Hence, if President Trump
does build a wall, it will keep more Mexicans in than it
prevents from coming. Indeed, there are reasons to think
that if this border was completely open, there would be
fewer undocumented Mexicans in the United States at any
time: lowering the risk of not being able to come back would
reduce the incentives to stay illegally. The same does not hold
for countries exposed to a massive inflow of refugees, but it
does hold for France, where the net inflow is relatively low.
In fact, when Mme. Le Pen or President Trump refer to
“immigrants,” they mean the third-generation offspring of
immigrants, who happen to have a different physiognomy.
Both invoke a myth of a “national culture,” some traditional
way of life, that is being undermined by the presence of
“immigrants.” “Immigrants” is just a code word for racism.
As sacrilegious as it may seem, it is useful to delve into
the conceptual relation between racism and multiculturalism.
The obvious difference is that racism claims inequality
between groups, treating them as innately superior and infer-
ior. The second difference is that the races are defined by the
racists, and in their view one is a member of a race by virtue of
origin, independently of one’s choice, while the ideology of
multiculturalism allows individuals to choose their cultural
identity. Yet the identities we choose for ourselves are not
always those in terms of which we are perceived by others.
In a beautiful phrase of Amitav Gosh, we leave “shadow
lines”: I may not see myself as Jewish or Muslim, yet others
may still see me as one. What these ideologies have in

117
 the present: what is happening?

common is the ontology of social fragmentation that should

be acknowledged by society and the state. As Michaels (2007:
3) observed, “the goal of overcoming racism, which had some-
times been identified as the goal of creating a ‘color-blind’
society, was now reconceived as the goal of creating a diverse,
that is, a color-conscious, society.” Their commonality
becomes clear when juxtaposed against the ideology of “repub-
licanism”: the idea that as citizens we are anonymous, that when
people with different features and different self-identifications
enter the public sphere they lose all their qualities and must be
treated equally because they are indistinguishable (Rosanvallon
2004). In spite of all their differences, racism and multicultural-
ism are both ideologies that fraction society into distinct groups.
When combined with cultural relativism, postmodern
ideology implies a multiplicity of truths. The truth of a statement
is authenticated by the identity of the speaker and all identities
are equally authoritative. It creates a world that allows for differ-
ences but precludes disagreements (Michaels 2007). If I say
“As a pink male, I believe the news is . . ., ” one can claim that
this news is false for him or her. But you cannot persuade me and
I cannot persuade you: each of us has our own truth. There is
nothing to talk about: a recent study reports that in 2017
a Thanksgiving dinner with guests from electoral districts domi-
nated by different parties lasted 30–50 minutes less than with
exclusively co-partisans (the average was 257 minutes; Chen and
Rohla 2018). Our beliefs have no authority over others because
they are conditioned by our identity. In a relativist world, the
news of others are all “fake” and there is no procedure by which
they could be determined to be true or false: this is a “post-truth”
world.

118
 potential causes

In an exceptionally well-informed and incisive analy-

sis, Lewandowsky, Ecker, and Cook (2017) report some results
of research:

1. “Corrections are rarely fully effective: that is, despite being

corrected, and despite acknowledging the correction, peo-
ple by and large continue to rely at least partially on
information they know to be false . . . In some circum-
stances, when the correction challenges people’s world-
views, belief in false information may ironically even
increase” (p. 355).
2. Falsehoods induce some people to conclude that truth is
unknowable even when the false message is not credible.
3. Propagating falsehoods diverts people from recognizing
other messages as true.
4. People tend to persist with beliefs they admit to be false if
they believe they are shared by others.

They conclude, “We are now facing a situation in

which a large share of the populace is living in an epistemic
space that has abandoned conventional criteria of evidence,
internal consistency, and fact-seeking . . . An obvious hall-
mark of a post-truth world is that it empowers people to
choose their own reality, where facts and objective evidence
are trumped by existing beliefs and prejudices” (pp. 361–2).
What distinguishes people is not information but alternative
epistemologies. Powerful evidence presented by Meeuwis
et al. (2018) shows that investors who had different models
of the economy modified their portfolios differently in
response to the US election of 2016, and the differences ran
along party lines.

119
 the present: what is happening?

Moreover, even when the views of particular indi-

viduals remain fixed, their attitudes toward those with
whom they disagree can be less or more hostile. In the
United States, 86 percent of Democrats and 91 percent of
Republicans have unfavorable views of the other party, with
41 percent of Democrats and 45 percent of Republicans
seeing the other party as a “threat to the nation”
(Acherbach and Clement 2016). Poignant anecdotes about
experiences of discrimination and abuse in everyday life
abound, and many systematic data indicate that the general
level of anger and hostility is on the rise. In 2012, 33 percent
of Democrats and 43 percent of Republicans described
themselves as angry at the opposing party’s presidential
candidate “most of the time” or “just about always,” while
by 2016 the percentage of Democratic voters who said they
were this angry at Trump rose to 73 percent, and the per-
centage of Republicans with that level of hostility toward
Hillary Clinton increased to 66 percent. Where we have
more systematic evidence, albeit only for the most recent
years, is with regard to “hate crimes.” In the United States
their incidence in the nine major metropolitan areas
increased 23.3 percent from 2015 to 2016, with a total of
13,037 (NBC 2017). Another source reports that they jumped
in the aftermath of the election, with over 1,000 incidents
self-reported between November 9 and December 12, 2016
(SPLC 2016). Overall, anti-immigrant incidents (315)
remain the most reported, followed by anti-black (221),
anti-Muslim (112), and anti-LGBT (109). Anti-Trump inci-
dents numbered twenty-six. Britain saw an increase of over
40 percent in self-reported hate crime incidents between

120
 potential causes

2015 and 2016. In addition to race-based hate crimes, Britain

also saw a rise in hate crimes based on sexual orientation.
Galop, a London-based LGBT anti-violence charity,
reported that hate crimes motivated by sexual orientation
rose 147 percent during the late summer of 2016. Other
countries across Europe have also experienced an increased
rate of hate crimes over the past several years. Between 2014
and 2015, Germany reported a 77 percent increase in hate
crimes. Amnesty International reported that incidents of
race-based violence are at an all-time high since World War
II in Germany. Statistics collected by Germany’s Interior
Ministry show that asylum shelters were attacked 1,031
times in 2015, a drastic increase from 199 attacks in 2014
and sixty-nine attacks in 2013. In Spain, the Spanish
Federation of Islamic Religious Entities reported that anti-
Islam attacks increased from forty-eight in 2014 to 534 in
2015. Additionally, Spain’s Interior Ministry published sta-
tistics for 2015 reporting hundreds of hate crimes based on
disability, ideology, and sexual orientation (Human Rights
Brief 2017). France seems to be the exception, with racist
crimes (anti-Semitic, anti-Muslim, anti-Roma) having
peaked in 2015 and then falling 44.7 percent, from 2,034 to
1,125, between 2015 and 2016 (Franceinfo 2017).
These, albeit unsystematic, facts show that the divi-
sions that rip several countries apart are not merely political
but have deep roots in society. These two levels are obviously
related but which way the causality runs is hard to determine,
as social and political polarization may feed on one another
(Moral and Best 2018). What these facts do tell us is that we
should not overpoliticize our understanding of the current

121
 the present: what is happening?

situation – that we should not reduce it to the actions of

politicians. The incessant complaints about Trump’s temper
and incompetence should not obscure that fact that his elec-
tion and his continuing support reflect something deeper,
something that lurks in the everyday life of society.

122
 7

Where to Seek Explanations?

7.1 Methodological Issues

As we have already seen, explanations can begin with the
economy, culture, or strategies of traditional parties; perhaps
all of them, perhaps some combinations. Yet most of the
pictures shown in previous chapters represent averages for
groups of countries over time – central tendencies – while
particular countries are far from being the same. As of 2014,
there were no radical Right parties in Australia, Canada,
Ireland, Luxembourg, New Zealand, Portugal, or Spain.
The income shares of the top 1 percent of recipients show
a sharp increase from the 1970s in the Anglo-Saxon countries,
while they remained stable in Germany, Japan, France, Sweden,
Denmark, and the Netherlands. The proportion of survey
respondents saying that immigration should be decreased
ranges from 25 percent in Australia, through 40 percent in
the United States, to 69 percent in the United Kingdom
(Gallup surveys in 2012 and 2014). Strikingly, Figure 7.1 shows
that even Germany and France clearly differ in the relation
between wage growth and productivity growth.
It bears emphasis that the US is in many ways an
outlier. It is the only economically developed democracy
where a candidate with a radical Right program has won an
election. While a decline of median incomes occurred in most

123
 German Wages Have Not Kept Up French Wages Have Kept Up
With Productivity With Productivity

160 GERMANY 160 160 FRANCE 160

NOMINAL INDICES*: NOMINAL INDICES*:
PRODUCTIVITY (TARGET) PRODUCTIVITY (TARGET)
150 PRODUCTIVITY (ACTUAL) 150 150 PRODUCTIVITY (ACTUAL) 150
WAGES (ACTUAL) WAGES (ACTUAL)

140 140 140 140

130 130 130 130

120 120 120 120

110 110 110 110

100 100 100 100

2000 2002 2004 2006 2008 2010 2012 2014 2016 2000 2002 2004 2006 2008 2010 2012 2014 2016
*SOURCE: ECB. *SOURCE: ECB.
TARGET GROWTH IS 1.5%
REAL PRODUCTIVITY GROWTH PLUS 2% INFLATION.

Figure 7.1. Wages and productivity in Germany and France

Source: European Central Bank.
 where to seek explanations?

countries only after 2008, in the United States they have

stagnated over a longer period. Top incomes rose more in
the United States than in other countries. Political polariza-
tion is exceptionally sharp. It is a country experiencing
a diminishing international influence, as well as deteriorating
infrastructure, education, and health. Perhaps most sympto-
matic are the recent data generated by Case and Deaton (2017)
concerning mortality patterns of white males aged 45–54,
which show that these mortality rates declined sharply in
France, Germany, the United Kingdom, Australia, and
Sweden, while they have increased since the late 1990s in the
United States. Finally, it is the only presidential system with
indirect elections, generating issues of legitimacy when
a candidate who does not obtain a majority of votes is
declared as the winner according to constitutional norms.
As these cross-country differences evidence, what we
have seen thus far is only the zeitgeist, the spirit of the time. But
the devil is in the details. How then to identify causal relations?
Global causes are not sufficient: globalization cannot explain
the differences between Germany and France. At least some
interaction between global causes and national factors is neces-
sary. I am skeptical that standard social science methods will
take us far. There are many common trends and time series are
relatively short, so I am concerned that we would find causality
where there is none. Indeed, while studies at the aggregate level
tend to find relations between economic variables and support
for the radical Right, individual-level data leave a lot of doubt
(Kates and Tucker 2017: 1–2). Moreover, when different states
of the world co-evolve over time, it is hard to determine which
way the causality runs (Dancygier and Laitin 2014). Finally,

125
 the present: what is happening?

cross-country observations are not independent: there is diffu-

sion as well as dissuasion, with debates as to whether Trump’s
victory increased support for the Front National or whether it
dissuaded Dutch voters from voting for Geert Wilders in 2017.
It is not obvious how to proceed, but it is instructive
to look at the micro level.

7.2 Voting for and Supporting the Radical

Right
The obvious temptation is to rationalize these attitudes in
economic terms, as competition for employment, defense of
wages from competition, etc. Is support for the radical Right
related to individual or collective economic conditions, such as
income, experiencing economic difficulties, having suffered
a job loss, fearing a job loss, and the like? A stronger version
of this question is whether these postures are economically
rational, that is, whether policies offered by the radical Right
are in the best interest of those who support it. One cannot
assume, however, that individuals accurately perceive the eco-
nomic conditions or understand the consequences of particular
policies. Even perceptions of one’s own economic situation are
tainted by partisan loyalties or other biases. One striking exam-
ple is that following the election of Donald Trump, Democratic
supporters revised downward the perceptions of their own
economic situation over the past five years while Republican
supporters revised it upwards (Brady, Ferejohn, and Paparo
2017). Perceptions of general economic conditions are even
more vulnerable to such biases (Stokes 2001, Maravall and
Przeworski 2001). Hence, it is a different question whether

126
 where to seek explanations?

people rationalize their political postures in economic terms or

whether they are rational in the sense of being based on
accurate beliefs about the world.
The strongest support in favor of the effects of actual
job losses in the United States is provided by Autor et al. (2017),
who found a robust positive effect of rising import competition
on Republican vote share gains and attributed Trump’s victory
directly to imports from China. In turn, in a massive study of
regions within fifteen Western European countries between
1988 and 2007, Colantone and Stanig (2017: 1) report that
“At the district level, a stronger import shock leads to: (1) an
increase in support for nationalist parties; (2) a general shift to
the right in the electorate; and (3) an increase in support for
radical right parties. These results are confirmed by the analysis
of individual-level vote choices. In addition, we find evidence
that voters respond to the shock in a sociotropic way.”
Nevertheless, the effect of job loss on the vote for the radical
Right is small: the difference between regions in the 75th
percentile of job loss and those in the 25th percentile is 0.7
points of the vote share. In another study of several countries,
Guiso et al. (2017) learn that experiencing current income
difficulties makes people less inclined to vote, but if they do
vote they are more inclined to vote for the “populist” parties,
while the experience of having lost a job during the past five
years reduces the probability of voting without having an effect
on the direction of the vote. Yet Margalit (2013), as well as Ayta,
Rau, and Stokes (2017), find that the political effects of unem-
ployment are conditional on the general economic situation
and wane with time. The most detailed study of the 2016
election in the United States that I found concludes that

127
 the present: what is happening?

The results show mixed evidence that economic distress

has motivated Trump support. His supporters are less
educated and more likely to work in blue collar
occupations, but they earn relatively high household
incomes and are no less likely to be unemployed or
exposed to competition through trade or immigration.
On the other hand, living in racially isolated communities
with worse health outcomes, lower social mobility, less
social capital, greater reliance on social security income
and less reliance on capital income, predicts higher levels
of Trump support. (Rothwell and Diego-Rosell 2016)

There is also evidence that fear, rather than actual experience,

drives radical Right postures. Minkenberg (2000: 187) finds
right-wing supporters among the “second-to-last fifth of post-
modern society, a stratum which is rather secure but objec-
tively can still lose something.” Kates and Tucker (2017: 3)
conclude that “only pessimism about one’s own financial
future is positively correlated with far right ideological iden-
tification.” Fossati (2014) learns that individuals in occupa-
tions that have a high unemployment rate are more likely to
vote on the basis of economic considerations.
Particularly puzzling are the results of Andrews, Jilke,
and Van de Walle (2014), who used a survey conducted across
Europe in 2010 concerning perceptions of social tensions
along four dimensions: the poor and the rich, managers and
workers, old and young people, and different racial and ethnic
groups. What I find perplexing is that people who experience
income difficulties are more likely to perceive higher tensions
on all four dimensions. One would think that some would
find the culprit for their conditions among the rich, others

128
 where to seek explanations?

among the management, still others in the disproportionate

income of old people, and some would blame immigrants. Yet
people who perceive high tension on one dimension also
perceive it on all the other dimensions. Neither class nor racist
ideology has a clear sway on the way people analyze their
conditions. They blame everyone because they do not know
whom to blame.
In light of this, context matters. Ivarsflaten (2008)
shows that radical Right parties are successful only if they
appeal on the immigration issues, regardless of economic
changes or political corruption. The conclusions she draws,
however, are flawed because she does not consider why some
anti-establishment parties use this issue while others do not.
Dancygier (2010), in turn, argues that support for the radical
Right is larger when poor economic conditions are combined
with high immigration and when the immigrants have elec-
toral power, so that incumbents have incentives to extend
material benefits to immigrants.
Yet not everyone agrees about the relevance of economic
factors. Inglehart and Norris (2016) find support for populist
parties among people who report income difficulties and have
experienced unemployment, but they conclude – I am not clear
on what basis – that “cultural backlash” is a more convincing
explanation. Hainmueller and Hopkins (2014: 227) question the
relevance of economic factors for anti-immigration postures.
“Overall,” they report, “hypotheses grounded in self-interest
have fared poorly, meaning that there is little accumulated evi-
dence that citizens primarily form attitudes about immigration
based on its effects on their personal economic situation.”
Instead, they claim, these attitudes are driven by “sociotropic

129
 the present: what is happening?

concerns about its cultural impacts.” I have no idea what this

means, other than racism. Indeed, Lee and Roemer (2006) calcu-
lated that racism, specifically opposing the welfare services
because they would also include people with different skin
color, is economically costly to poor whites in the United States.
While the evidence that natives react negatively to
people who differ from them is abundant, the origins of such
postures remain obscure. If they do not reflect economic
conditions, where do such postures come from? What
makes people adopt xenophobic, often unabashedly racist,
postures? What makes them willing to engage in hostile acts
with regard to people who look different, speak a different
language, or eat different foods? The psychological explana-
tions favored by Hainmueller and Hopkins (2014) just relabel
what we observe in a “scientific” language without providing
enlightenment.
In turn, there is evidence that attitudes change as
a function of other transformations. McCarty, Poole, and
Rosenthal (2016) demonstrate a very strong relationship between
income inequality and polarization in the United States House of
Representatives from 1947 on. Another piece of evidence is that
the perception of immigration as an important policy issue
followed the increase of net immigration into the United
Kingdom (IPSOS/MORI 2013). Still, a rival explanation is that
xenophobia, racism, nativism, bigotry, and other prejudices were
always present – German immigrants into the United States
were “Krauts,” Italians “Dagos,” Japanese “Japs,” Poles
“Polacks,” and they were only temporarily suppressed by hypoc-
risy, which suddenly lost its “civilizing force” (Elster 1998)
because some politicians made them public. One startling

130
 where to seek explanations?

piece of evidence is a survey conducted in the United States in

1939: the respondents were asked whether the country should
receive 10,000 German, mainly Jewish, children, with 61 percent
saying “No.” Some Germans and Japanese continued to be
nationalistic even after the atrocities committed by these coun-
tries during the war, and the language of “patriotism,” recently
evoked by right-wing parties, just authorized these attitudes to
enter the public sphere. Perhaps Trump just liberated every pre-
existing prejudice from the censorship of political correctness.
All of this is second hand. I suggest a magisterial
review of this literature by Golder (2016), who advocates
combining contextual conditions and individual characteris-
tics. It may be, however, easier said than done. From what we
know as of now, it seems that the actual losers from globaliza-
tion tend to vote Right and that the prospect of job loss
frightens the potential losers. Yet while all these effects are
statistically significant, they explain only a small part of radi-
cal Right postures. The origins of these postures remain
opaque.
This is a paltry conclusion and it may disappoint
many readers. But one should not believe the flood of
accounts that have all the answers. If you read papers by
economists, you will find that the political postures are
explained by economic situations; if you read papers by psy-
chologists the secret will lie in some psychological traits. I find
neither explanation convincing: regression analyses of politi-
cal postures on economic conditions always find some con-
ditions that are statistically significant but explain very little;
psychologists tend to relabel what they seek to explain and
claim the relabeled feature is the cause. Our intellectual

131
 the present: what is happening?

deformation is to always find some sense in complex situa-

tions, to suppose that diverse phenomena that surprise us
must be somehow related, that everything must have
a cause. I have listed various factors and briefly summarized
each of several recent transformations – economic, cultural,
and purely political – that may have engendered the current
political situation, which is no more than a catalogue, and the
conclusion is paltry because it says that we cannot tell whether
they are related and which matters most. But that is the best,
I believe, we can do given what we now know.

132
 8

What May Be Unprecedented?

Before we can begin thinking about the future, it is useful to

place the current situation in the context of what we have
learned about the past. To the extent that some aspects of
present conditions repeat those of the past, we can draw on
the lessons of history suggested by the comparisons of con-
solidated democracies that collapsed and of those that sur-
vived earlier in time. Yet history does not illuminate the
future when the present conditions are unprecedented, so
there is nothing to draw lessons from. This is why we need
to place the present in the context of the past.
A simple statistical analysis indicates that, even ignor-
ing the fact that democracy in the US is 200 years old, given
the current income of the United States, the probability that
the incumbent would not hold an election or hold one making
it impossible for the opposition to win is 1 in 1.8 million
country years. If one believes in drawing lessons from history,
an outright collapse of democracy in a country with the per
capita income of the United States today is out of the realm of
the imaginable. In spite of frequent references to these tragic
events, looking back at the advent of fascism in Europe in the
1920s and 1930s is not instructive for the simple reason that the
countries where fascism came to power were miserably poor
compared to now. The per capita income (in 1996
Geary–Khamis PPP USD, from Maddison 2011) of Italy in
1922 was $2,631 while as of 2008 it was $19,909; of Germany it

133
 the present: what is happening?

was $3,362 in 1932 and $20,801 in 2008; of Austria $2,940 in

1932 and $24,131 in 2008. This was just a different world.
It was also a different world ideologically. The extreme
parties during the interwar period were anti-democratic
(Capoccia 2005): both Communists and fascists openly sought
to replace a system based on individual representation through
elections. Communists derided democracy as a mask over the
“dictatorship of the bourgeoisie” (Lenin 1919) and fought to
replace it by a “dictatorship of the proletariat” led by a single
party. Fascists disdained democracy as a system that promotes
artificial conflicts between classes and strived to replace it with
a system based on negotiated compromises among corporations
organized along functional bases (Cassese 2011). Both offered
a promise to replace “politics” with “rational administration.”
Both had a broad appeal, not only in Europe but around the
world, from Argentina to Mongolia. Yet both these ideologies
are now dead and buried. The “anti-system” parties of today are
not anti-democratic. While the label of “fascist” is carelessly
brandished to stigmatize these political forces, these parties do
not advocate replacing elections by some other way of selecting
rulers. They are ugly – most people view racism and xenophobia
as ugly – but these parties campaign under the slogan of return-
ing to “the people” the power usurped by elites, which they see as
strengthening democracy. In the words of a Trump advertise-
ment, “Our movement is about replacing a failed and corrupt
political establishment with a new government controlled by
you, the American people.”1 Marine Le Pen promised to call for
a referendum on Europe, in which “you, the people, will decide.”

1
See www.youtube.com/watch?v=vST61W4bGm8.

134
 what may be unprecedented?

Table 8.1 Economic conditions in democracies that fell or

did not fall before 2008 and the post-2008 means for
democracies that survived

Fell Didn’t fall Now

Mean Mean N Mean

a
GDP/cap 18,012 5,770 406 23,825
Growtha 0.031 0.011 406 0.020
Labor shareb 0.61 0.50 352 0.54
Gini grossc 42.6 44.6 152 42.5
Gini netc 33.8 44.6 152 33.5

Note: The columns for ‘Fell’ cover the period until 2008, those
for ‘Now’ cover the period 2008–14. Cell entries are means.
Sources: (a) In PWT9.0 PPP dollars. (b) From PWT 9.0. (c)
Gini coefficients of gross and net incomes, from SWIID
(2014).

Moreover, there is nothing anti-democratic about people want-

ing to have a “strong” or “competent and effective” govern-
ment – responses to survey questions that increased in
frequency in recent years and which some commentators inter-
pret as a symptom of declining support for democracy.
Schumpeter (1942) certainly wanted governments to be able to
govern and to govern competently, and I do not see why other
democrats would not.
Yet we are nervous. We are nervous because we sus-
pect that some of the conditions portrayed above are unprece-
dented, that extrapolations from history may be an unreliable
guide to the current prospects. Are they unprecedented?
Table 8.1 compares the economic conditions of coun-
tries in which democracies fell or survived before 2008 with the

135
 the present: what is happening?

post-2008 (and pre-2015) conditions of current democracies.

Average incomes are now higher than even those of democracies
that survived in the past. The average inequality of individual
incomes mirrors that of surviving democracies. But labor share
in manufacturing as well as growth of average incomes are
lower.
These numbers, however, do not tell the full story.
Consider first the stagnation of incomes of some percentage –
between thirty and fifty – of people receiving the lowest
incomes. Their incomes stagnated over almost forty years in
the United States but only more recently in most European
countries. In the United States the incomes of poorer people
did not increase mainly because of increased inequality; in
continental Europe the stagnation after 2008 was due princi-
pally to slower growth. This stagnation is certainly unprece-
dented during the period for which we have data on income
distribution (1950 onwards), but perhaps even if we include
the interwar period. The average rate of growth of the OECD-
2000 countries between 1978 and 2014 was 2.1 percent, while
during the interwar period it was 2.3 percent. Yet inequality
increased sharply during the recent period, while from the
fragmentary information we have it seems that it declined
quite significantly in the earlier period.2 Hence, there are
grounds to believe that lower incomes grew slower after 1978

2
In the five countries for which we can calculate Gini coefficients of gross
incomes during the interwar period, they decreased in France from forty-
five in 1919 to forty in 1939, in Germany from thirty-eight in 1926 to thirty-
five in 1932, in the Netherlands from forty-nine in 1919 to thirty-eight in
1939, in Sweden from forty-eight in 1919 to thirty-six in 1939, and in the
United States from forty-four in 1919 to thirty-eight in 1939. These

136
 what may be unprecedented?

than during the interwar period, which means that the cur-
rent income stagnation of some bottom percentage of recipi-
ents is historically unprecedented, at least over the past one
hundred years.
Particularly striking is the decline of unions, the
opening gap between increases in productivity and of wage
rates, which remain almost flat, and the decline of labor share
in manufacturing. The power of unions to control the flow of
labor to firms, conquered in several countries in the 1930s and
institutionalized in the aftermath of World War II, has been
eroding rapidly since the neo-liberal turn of the early 1980s,
and weakening the political power of their electoral allies, the
left-wing parties. This shift in the economic and political
balance between capital and labor, coupled with a relaxation
of control over capital flows and financial deregulation,
resulted in the secular stagnation of incomes of lower-strata
income recipients, punctuated by the crisis of 2008.
One should not be surprised, therefore, that beliefs in
intergenerational equality are eroding. The 60 percent of
respondents in the United States and 64 percent in Europe
who believe that their children will be financially worse off
than they are may be unduly pessimistic and still reacting to
the shock of 2008, even though Chetty et al.’s (2016) evidence
about the United States shows that these beliefs do not diverge
far from reality. We do not have quantitative evidence about
what previous generations believed, but there are good rea-
sons to think that a belief in material progress has been deeply

calculations are based on transforming the Pareto coefficient o: provided

by Atkinson, Piketty, and Saez (2011) into Gini coefficients.

137
 the present: what is happening?

ingrained in Western civilization since the industrial revolu-

tion. The average incomes among the OECD-2000 countries
increased twenty-two-fold between 1820 and 2008. In spite of
wars and economic crises, there was no thirty-year period in
the past 200 years in which average incomes declined. So if
people are now pessimistic about the future of their children,
this may be a shift at a civilizational scale.
Whether the current polarization, accompanied by
hostility with regard to people who hold different views, is
new is impossible to judge. It is certainly new in the recent
period but, as we have seen in the case studies of past crises,
there have been periods in different countries, including the
United States, when societies were deeply divided. All we
know is that such intense divisions are long-lasting and diffi-
cult to overcome. They leave scars, which are often plastered
by silence, as is the experience of the civil war in Spain or the
divisions over Allende’s period in Chile, and tend to resurge.
Finally, as confirmed in Table 8.2, perhaps the most
dramatic recent change is the erosion of traditional party
systems. The party systems that crystallized in Western
Europe in the 1920s, with few exceptions, continued to con-
stitute the main political alternatives until the end of the
twentieth century. With the exception of the period immedi-
ately following World War II, voters had a choice between
center-left and center-right. The parties representing these
options, sometimes changing labels, splitting and merging,
continued to be the two top vote-getters. New parties
appeared from time to time but they were rarely successful
and many were ephemeral. Of the two top vote-getters around
1924, 90 percent were still among the top two by the late 1990s,

138
 what may be unprecedented?

Table 8.2 Some political features of countries that were

members of the OECD as of 2000, before and after 2008

1960–2007 1960–2007 2009–14 2009–14

N Mean N Mean

Turnout 1,065 78.8 138 72.1

Effparty in the 1,065 4.05 137 4.68
electorate
Effparty in the 1,065 3.48 137 3.86
legislature
Social democratic 1,065 26.9 138 21.2
vote share
Union density 995 42.1 98 33.0

Note: Effparty is the effective number of parties.

Source: Armingeon et al. (2016).

but only about 75 percent are as of now. The number of

“effective parties” was three in 1960 and it is almost four
now. Moreover, this erosion has been accompanied by falling
rates of electoral participation, with the share of radical Right
parties increasing as turnout falls. It may be that this erosion is
due to the loss of convocational power by the traditional
parties or a shift of voters’ preferences away from the center,
as in the United States. Yet the decline of the traditional
parties is not necessarily a reflection of an erosion of the
political center. Rather, it may be a sign of their organizational
weakness as well as a general disgust with professional politi-
cians. All we can observe is that the traditional parties now
come third, fourth, or even fifth in some countries, while in
the United States a party that was once dominated by
“Rockefeller Republicans” has been captured by the radical

139
 the present: what is happening?

Right. What is new is that both types of organizations that

used to represent the working class – social democratic parties
and trade unions – have lost this capacity. Both the social
democratic vote shares and union density are distinctly lower
in the most recent period. What has also changed, or at least is
changing, is the social base of support for right-wing move-
ments (Ignazi 1992, 2003, Arzheimer 2013). Traditionally such
movements were supported by the petite bourgeoisie – self-
employed, small shop-keepers, artisans, and farmers – while
now they seek to combine this traditional base with appeals to
the working class. As the social democratic parties became
bourgeoisified, the right-wing parties became proletarianized.
The final but consequential difference between the
past and the present, one that is encouraging, is that the
military have pretty much disappeared from the political
scene. The military played the decisive role in the collapse of
nine out of fourteen democracies listed in Table 2.1. Indeed,
had this text been written some forty years ago, the political
stance of the military would have been its central preoccupa-
tion. Yet astonishingly it is no longer a political actor, even in
Latin America, and it all but disappeared from the pages of
political science as well.
Several caveats are in order to avoid unwarranted
conclusions. Most importantly, the lessons from the past
should not be treated as causal. Endogeneity is an obvious
problem. As just one example: did democracies fall because
the economies were stagnant or were the economies stagnant
because democracies were about to fall? Second, the general
patterns hide sharp differences among particular countries.
Whatever global factors affect democracies, say worldwide

140
 what may be unprecedented?

economic depression, their effects work differently depending

on the conditions specific to each country. Third, “difference”
and “similarity” is a matter of degree and we do not know how
much a particular difference matters. Finally, the list of con-
ditions that we can observe is far from complete and it may
well be that this list omits some that are crucial, alone or in
combinations. Hence, as already stated, I am not trying to
persuade the reader of anything but just provide food for
thought. Drawing lessons from history is an art, not
a science. What all of this adds up to – what the prospects
are for democracy’s survival under such conditions – one can
only speculate about.

141
Part III

The Future?
Before we begin asking questions about possible futures, we
need to understand how democracy works when it works well
and how it collapses or deteriorates. Democracy functions well
when political institutions structure, absorb, and regulate
whatever conflicts may arise in society. Elections – the
mechanism by which a collectivity decides who should govern
it and how – are the central mechanism by which conflicts are
processed in democracies. Yet this mechanism functions well
only if the stakes are not very large, if losing an election is not
a disaster, and if the defeated political forces have a reasonable
chance to win in the future. When deeply ideological parties
come to office seeking to remove institutional obstacles in
order to solidify their political advantage and gain discretion
in making policies, democracy deteriorates, or “backslides.”
This prospect is foreboding because the process need not entail
violations of constitutionality and, in turn, when backsliding
follows a constitutional path, when the government is careful
to preserve all appearances of legality, citizens have nothing
upon which to coordinate their resistance. Hence, it is reason-
able to worry about whether it could happen in the United
States or in the mature democracies of Western Europe.

143
 9

How Democracy Works

9.1 Conflicts and Institutions

The point of departure must be that at every moment in every
society, some people – individuals, groups, or organizations –
conflict over something. Often this something comprises var-
ious kinds of scarce goods, such as income, property, places at
university, replacement organs, or access to public services.
Many antagonisms, however, concern issues other than dis-
tribution. Some arise because some people have strong, often
religiously motivated views about how others should act.
Some are driven by a sheer desire for power, ambition, or
vanity. Symbolic issues also evoke passions: in Weimar one
government coalition broke over the issue of the colors of the
German flag.
Not all antagonisms become political. Some of us are
intensely divided by loyalty to different sport clubs, without
such divisions becoming politicized. One woman may want to
wear a burka on the beach and another nothing at all, but such
discords may remain private. Even if some people have views
about what others should or should not do, these are still
private opinions. Antagonisms turn into political conflicts
when they entail views about the policies governments should
pursue and the laws they should adopt, most importantly
about what governments should coerce all of us to do or not to

145
 the future?

do,1 or when some groups attempt to impose their will over

others by force, say in physically blocking access to abortion
clinics or occupying someone else’s property.
Conflicts may be easier or harder to resolve peace-
fully. They differ in several aspects:
1. How divided are people about what they want most to occur
with regard to a particular issue? The answers to this question
describe distributions of “ideal points”: the policies or laws
people see as best. One way to characterize such distributions
is to ask whether there is some outcome that is preferred by
more people than any other outcome and whether the pro-
portion of people who like all other outcomes falls as the
distance from the option preferred by most people increases.
Distributions which satisfy both conditions are “unimodal.”
For example, the current distribution of attitudes toward
abortion in the United States is unimodal, with more people
opting for it being legal in most cases than in all cases, and
more people opting for it being legal in most cases than illegal
in most or all cases (Pew Research Center 2018). Yet it may
well be that people’s peak preferences concentrate around
different outcomes. Some years ago in France, for example,
a large segment of the population opposed same-sex mar-
riage, fewer people wanted to allow it without the right to
adopt, and another large segment supported it without

1
Regulation of what women can or cannot wear on beaches has been
in fact the subject of political conflicts in France. One long-standing
conflict is whether people should be allowed to be nude on public
beaches; the new one is whether they should be allowed to be clad
from toe to head.

146
 how democracy works

restrictions on adoption. This distribution was bimodal, as is

the current United States distribution of postures on the
general liberal–conservative dimension, shown in
Figure 6.11. Interestingly, Medina (2015: figure 1) shows that
voters’ positions on the left–right dimension in twenty
European countries tend to be trimodal (as Downs (1957)
predicted), with a big mode at the center and smaller modes
to the left and right of it.
2. How much do individuals care when outcomes deviate
from their ideal preference? Clearly, people dislike more
outcomes that are more distant from what they want
most. But the intensity of their loss varies across issues
as well as among individuals. Say that someone wants the
top marginal tax rate to be 40 percent and the actual rate
is 30 or 50 percent. This person views such tax rates as
too low or too high, but this dissatisfaction is unlikely to
be very intense. Yet for people who think that abortion
should not be allowed under any circumstances, even
legalizing the “morning-after” pill is anathema to them:
their utility falls sharply when this is the law. Hence, even
when the distribution of ideal points is unimodal, con-
flicts can be intense if people experience a sharp loss of
utility when outcomes deviate even minimally from their
peak preferences.
3. How closely related are positions on different issues?
Are people who want abortion laws to be more restrictive
the same as those who oppose immigration? Are people
who oppose immigration the same as those who want
more redistribution of income? If answers to these ques-
tions are positive, cleavages are superimposed; if they are

147
 the future?

negative, cleavages are cross-cutting. For example, nega-

tive postures about immigration correlate with homopho-
bia and sexism across the OECD countries. Cleavages
tend to be superimposed when preferences are associated
with some other characteristics, say religion, income, or
education. According to the Pew survey cited above, for
example, only 25 percent of white Evangelicals agree to
abortion being legal under some circumstances, while
more than 50 percent of Catholics, about 70 percent of
traditional Protestants, and 75 percent of people unaffi-
liated with any religion do so. Because these groups differ
on other moral issues as well, cleavages are superimposed.
In turn, other cleavages may be cross-cutting: Lipset
(1960) has argued that the postures toward democracy
versus authoritarianism divide the working class, and we
have already examined the division between the SPD and
KPD in Weimar Germany.

It is reasonable to expect that conflicts are more

difficult to resolve peacefully when people’s peak preferences
differ more, when the loss of utility associated with deviations
from these ideal preferences is more intense, and when clea-
vages are superimposed, clearly separating otherwise identifi-
able groups (Coser 1964). This is not to say that governments
are passive when confronting conflicts that are difficult to
manage. A natural strategy of governments is to try to per-
suade people that whatever divides them is less important
than what unites them. “Unity,” as in “united we stand,”
“harmony,” and “cooperation” are incessantly propagated
by appeals to nationalism, evocations of common roots even

148
 how democracy works

in the face of divergent origins, celebrations of national holi-

days, anthems, and flags, expressions of pride in the national
army or in the national performance in the Olympics – the list
goes on. Even intensely divisive elections are always followed
by a “unity” speech. To the best of my knowledge, Donald
Trump was the first US president not to call for unity in his
inaugural speech. Salvador Allende’s declaration, “No soy
Presidente de todos los Chilenos” (I am not the president of
all the Chileans), was an enormous blunder.
It is hard to tell whether such exhortations have much
effect, but the fact is that conflicts often persist in spite of
them. Just for heuristic purposes, imagine that preferences
can be placed on a single (utility) line, with a mass of people at
points marked as A and B:

–– A ––––– x ––––– B ––

Point x is a potential solution to the conflict. Say that

point A represents the preference for a path to citizenship
being open to all immigrants, illegal and legal; point B is the
preference for deporting all illegal immigrants regardless of
family considerations; and point x the preference for some
intermediate solution, such as legalizing the status of parents
whose children were born in the country. If A and B are
sufficiently distant from each other on the utility scale, the
conflict may have no solution. Say point x is unacceptable for
people located at B and nothing farther from A than x is
acceptable to people at A. Then the conflict has no solution
acceptable to both groups. Think of the Chilean situation: not
being able to nationalize some large firms in one stroke was
unacceptable for the government coalition, only nationalizing

149
 the future?

firms one at a time was acceptable to the opposition.

The Chilean conflict did not have a peaceful solution.
The same extends to more than one dimension.
Remember that one large German party, the SPD, was socia-
list on the economic dimension and democratic on the poli-
tical dimension, while another party, the DNVP, was
capitalist and authoritarian. Because any majority coalition
had to include both, the set of compromises that would be
supported by a majority in the parliament was empty.
How, then, do we manage to process such conflicts in
order and peace, without curtailing political freedom, relying
on procedures and rules that indicate whose interests, values,
or ambitions should prevail at a particular moment?
Political institutions orderly manage conflicts by (1)
structuring conflicts, (2) absorbing conflicts, and (3) regulat-
ing them according to rules. An institutional order prevails if
only those political forces that have institutionally constituted
access to the representative system engage in political activ-
ities, and if these organizations have incentives to pursue their
interests through the institutions and incentives to tolerate
unfavorable outcomes. Specifically, conflicts are orderly if all
political forces expect that they may achieve something, at the
present or at least in some not too distant future, by proces-
sing their interests within this framework, while they see little
to be gained by actions outside the institutional realm.
Note that thinking in strategic terms assumes that
organizations can discipline the actions of their followers.
As Maurice Thorez famously remarked in 1936, “One has to
know how to end a strike.” Organization, Pizzorno (1964)
observed, is a capacity for strategy. Organizations can act

150
 how democracy works

strategically only if they can activate and deactivate their

followers according to strategic considerations. When they
do not have this capacity, political conflicts can assume the
form of unorganized, “spontaneous” outbursts.
1. Political institutions structure conflicts. Institutions
define the actions that particular actors can adopt, they
provide incentives associated with each course of action,
and constraints to the possible outcomes. As a result, they
structure the actions which all actors would pursue given
their interests or values and shape the collective outcomes,
resulting in equilibria. Obviously, no one competes to con-
quer the office of the president in systems that have no such
position: parliamentary monarchies. Only slightly less
obvious is that the competition for the office of the pre-
sident is more intense in systems where the president is the
chief executive than in those in which he or she is only the
ceremonial head of state. A more complicated example is
the effect of electoral systems on electoral competition.
With a single-district/single-member (SMD) system, and
two parties, both parties have incentives to move toward
the center of voters’ preferences; with a high degree of
proportionality, parties want to maximize their niche,
which may lead some of them to maintain extreme pos-
tures. Such examples are endless.
Every political system molds the ways in which social
forces organize as political actors, regulates the actions they
can undertake, and constrains the policy outcomes that are
subject to institutional competition. For example, rules
according to which votes become transformed into legislative

151
 the future?

seats – electoral systems – influence the number of parties that

participate in electoral competition and the interests they
represent: functional, regional, religious, ethnic, etc. Rules
concerning unionization affect the number of trade union
associations, their sectorial organization, and the extent of
their centralization. Rules with regard to class actions deter-
mine whether only individuals or groups sharing the same
grievance can address themselves to courts. Other rules define
the actions that can be followed within the institutional fra-
mework. Most countries, for example, have laws regulating
whether business lobbies and trade unions can financially
support political parties. Most countries have laws defining
which strikes are legal and which are not. Finally, constitu-
tional courts or equivalent bodies can invalidate those out-
comes that are inconsistent with some basic principles that
stand above pluralistic competition, principles that are often
but need not be enshrined in constitutions.
Political parties mold public opinion, compete in
elections, and occupy executive and legislative offices.
Parties became at one point the main form for organizing
interests. They were a mechanism for articulating and aggre-
gating interests, vertical organizations that integrated indivi-
duals into the representative institutions. For reasons that
remain obscure, however, they transformed over time into
organizations that function intermittently only at times of
elections. They lost their socially integrative function: no
one could say today with Michael Ostrogorskij (1981),
“Do not convince them, take them in socially.” Any kind of
a daily, permanent connection is gone. And when parties do
not have a day-to-day vertical connection with the people

152
 how democracy works

who end up supporting them at the time of elections, they

cannot discipline their political actions.
Interest groups, whether lobbies of businesses, reli-
gious groups, or voluntary associations, seek to influence
political parties as well as advance their interests by addres-
sing themselves directly to the executive, including the lower
echelons of the bureaucracy. One important difference in
structuring conflicts lies in the area of regulation of function-
ally defined interests. Unions were banned in all European
countries until the middle of the nineteenth century. Even
when they were finally legalized, in all democracies the state
tightly regulates the conditions under which they can be
formed, whether one or multiple organizations can exist
within a sector of industry or a particular workplace, whether
collective agreements have the force of law, whether agree-
ments concluded by unions apply to non-union members, etc.
Note that, as shown in Figure 6.10, average union density
declined sharply after about 1980, so that the power of union
organizations over workers eroded similarly to that of politi-
cal parties over their sympathizers. Lobbies of businesses are
not equally tightly regulated, with only a few countries requir-
ing that they register as such and make their activities trans-
parent. Voluntary associations are regulated mainly through
tax laws whenever they seek a not-for-profit status.
Civil law and its adjudication by courts individualize
conflicts. Without recourse to courts many conflicts assume
a form of spontaneous collective protests, as in China. But
when individuals can direct their claims to courts, conflicts
between them and the state become decentralized: in
Argentina, for example, individuals sue the state in courts

153
 the future?

for not delivering services guaranteed in the constitution

(Smulovitz 2003). Courts are a channel for processing con-
flicts without collective organization by the claimants.
In sum, states shape the organization of political
forces that can appear on the terrain of political institutions.
Other forms of political activity are either uneasily tolerated
or actively repressed.

2. Institutions absorb political conflicts when those political

forces which can potentially engage in other ways of promot-
ing their interests or values have incentives to direct their
actions within the institutional framework. What matters is
not only whether they win or lose, but what can they win or
lose: how much is at stake. A conflict over wages, for example,
entails lower stakes than a strike over layoffs. The stakes in
a conflict over dumping toxic waste into rivers may be low for
industry, just involving somewhat lower or higher profits, and
very high for those potentially exposed to the poison.
The stakes in a decision to go to war may be enormous for
everyone. Note that in many conflicts, the benefits of govern-
ment decisions are concentrated, while the costs are diffuse:
think of a tariff on toothpaste that significantly increases the
profits of the producers and is almost imperceptible to the
consumers. Conflicts that entail future political power entail
high stakes because their outcomes are difficult to reverse.
“Flexible labor market” policies, for example, may or may not
reduce unemployment, but they undermine the organiza-
tional power of the unions and, thus, their chances to influ-
ence policies in the future.

154
 how democracy works

Schematically, think about the fact that each orga-

nized political force expects to gain something by processing
its interests within an institutional framework and has some
idea about how reversible the outcome would be if it happens
to lose, so that it has some notion of the expected value of
participating in the institutional interplay of interests.
The alternative that each political force faces is to use its
resources outside the institutional framework, using violence
or other inefficient forms of conflict processing (see below).
This choice was starkly stated by John McGurk, chairman of
the UK Labour Party in 1919: “We are either constitutionalists
or we are not constitutionalists. If we are constitutionalists, if
we believe in the efficacy of the political weapon (and we do,
or why do we have a Labour Party?) then it is both unwise and
undemocratic because we fail to get a majority at the polls to
turn around and demand that we should substitute industrial
action” (quoted in Miliband 1975: 69). His view, however, is
not always shared: for example, the leader of a new left-wing
political party in France, Jean-Luc Mélenchon, announced in
the aftermath of his electoral defeat that he would lead his
supporters to the streets. Moreover, one should not go too far
in assuming that all such choices are dictated by strategic
considerations. Each society has a fringe of fanatics, people
who act without considering the consequences.
Both the resources that particular groups bring to the
institutional interplay of interests and those they can mobilize
for actions outside the institutional framework are group-
specific. Multinational corporations have an effective lobbying
power but no capacity to bring people to the streets. Unions
may have less political influence but a damaging power to

155
 the future?

strike. The military are not supposed to have any institu-

tional power but they are the ones who have arms. To be
effective in absorbing conflicts, the power of particular
actors within the institutional framework cannot diverge
too much from their capacity to realize their objectives out-
side of it. Institutions function under the shadow of non-
institutional power.
3. Institutions regulate conflicts if the losers accept outcomes
determined by applying institutional rules. Political actors
may use political institutions and still reject an unfavorable
outcome. One may think, and some theorists do, that such
situations are not possible. The argument is that if one group
would adopt a strategy of “I will try within the institutions and
if I fail I will go outside the institutions,” then the group(s)
with which it is in conflict would not direct their actions
within the institutions, knowing that their institutional vic-
tory would be hollow. Hence, the argument goes, “if actors
agree to some rules, they will obey them” or “if they do not
intend to obey them, actors will not agree to the rules”
(Buchanan and Tullock 1962, Calvert 1994). Yet we do witness
situations in which a conflict should have been terminated
according to some rules and still the losers do not accept the
outcome, reverting to non-institutional actions. Collective
agreements concluded by union organizations are sometimes
rejected by rank-and-file workers, who engage in wildcat
strikes. A legislature may pass a law that brings people to
the streets in protest: educational reforms in France routinely
mobilize massive opposition. Even election results are not
always accepted by the losers: among democracies that fell,

156
 how democracy works

this was the case in Honduras in 1932 and Costa Rica in 1958.
The answer lies in uncertainty: outcomes of institutional
interplay cannot be predicted exactly. Hence, a group may
calculate ex ante that it would get something by directing its
activities within the institutional framework, only to discover
that it has lost, and that the resulting status quo is worse than
what it can expect to get by going outside the institutional
channels. In turn, the other group(s) may believe ex ante that
a loss would be tolerable for their opponents, only to discover
ex post that it is not.
One important aspect of institutions is whether they
provide determinate rules according to which conflicts should
be terminated. We have seen in the Chilean case, for example,
that the legal framework contained two contradictory rules for
treating the state monopoly of arms: on the one hand, the
Congress passed the law giving jurisdiction over this monopoly
to the military, empowering them to search for arms in govern-
ment buildings; on the other hand, the law gave the president
the authority not to allow the military to enter public buildings.
Hence, the constitutional status of the search for arms became
indeterminate, which undermined the posture of those gener-
als who adhered to the principle of non-intervention as long as
the president did not violate the constitution. Perhaps the most
flagrant example of constitutional indeterminacy occurred in
Ecuador in 1977, when three persons could claim with some
justification that they were the president and the Supreme
Court refused to arbitrate the conflict (Sanchez-Cuenca 2003:
78–9). Examples are many, but the general point is that some-
times constitutions and laws do not provide clear guidance for

157
 the future?

solving particular conflicts, and then the very distinction

between institutional and non-institutional breaks down.
Given this characterization of conflicts and institu-
tions, a question that naturally arises is whether all institutions
can manage all conflicts in an orderly way. For example, some
scholars think that a less proportional electoral system would
have generated stable governments in Weimar Germany.
Others, in turn, see the institutional culprit of Weimar in
Article 48 of the constitution that allowed a president to
appoint a government without the support of parliament and
even in opposition to it (Bracher 1966: 119). Conversely, some
scholars think that had Chile had a parliamentary instead of
a presidential system, a center-right majority coalition would
have been formed and democracy would have survived. One
may also wonder what would have happened to democracy in
France had the Fourth Republic continued rather than being
replaced by a presidential system. Unfortunately, such claims
must invoke counterfactuals, so they are inevitably speculative.
We know enough about institutions to understand that, given
the structure of political cleavages, some institutions could
generate effective and stable governments while other institu-
tions could not. Whether, however, a different institutional
framework would have prevented the advent of Hitler to
power or the fall of democracy in Chile is impossible to tell:
too many contingencies are entailed.
The most important institution by which conflicts
are processed in democracies are elections. Elections,
however, are a peculiar way of processing conflicts, in
that they occur on particular dates, are fixed indepen-
dently of the current political situation in most countries,

158
 how democracy works

and are supposed to determine the relations of political

power for some definite future. Political life, however,
never stops. For one, the day an election is over, parties
already begin to campaign for the next election. But
politics between elections is not limited to electoral pol-
itics. The policies of governments elected by a majority
may meet with opposition from groups that feel intensely
about particular issues. Moreover, even if governments
are elected by a majority, not all of the policies they
propose need to enjoy majority support. Hence, we need
to examine separately what happens in elections and what
happens during the periods between elections.

9.2 Elections as a Method of Processing

Conflicts
We select our governments through elections. Parties propose
policies and present candidates, we vote, someone is declared
to be winner according to pre-established rules, the winner
moves into the government office, and the loser goes home.
Glitches do sometimes occur, but mostly the process works
smoothly. We are governed for a few years and then have
a chance to decide whether to retain the incumbents or throw
the rascals out. All of this is so routine that we take it for
granted. What makes it possible?
Here is the puzzle stripped to its bare bones. Suppose
that I want something that someone else wants as well; some-
times I want what is not mine. An application of some rule
indicates that someone else should get it. Why would I obey
this rule?

159
 the future?

The very prospect that governments may change can

result in a peaceful regulation of conflicts. To see this argu-
ment in its starkest form, imagine that governments are
selected by a toss of a, not necessarily fair, coin: “heads”
mean that the incumbents should remain in office, “tails”
that they should leave. Thus, a reading of the toss designates
“winners” and “losers.” This designation is an instruction
regarding what the winners and the losers should and should
not do: the winners should move into a White, Blue, or Pink
House or perhaps even a palace; while there they can take
everything up to the constitutional constraint for themselves
and their supporters; and they should toss the same coin again
when their term is up. The losers should not move into the
House and should accept not getting more than whatever they
are given.
When the authorization to rule is determined by
a lottery, citizens have no electoral sanction, prospective or
retrospective, and incumbents have no electoral incentives to
behave well while in office. Because electing governments
through a lottery makes their chances of survival independent
of their conduct, there are no reasons to expect that govern-
ments would act in a representative fashion because they want
to earn re-election: any link between elections and represen-
tation is severed. Yet the very prospect that governments
would alternate may induce the conflicting political forces
to comply with the rules rather than engage in violence.
Although the losers suffer temporarily by accepting the out-
come of the current round, if they have a sufficient chance to
win in future rounds they may prefer to comply with the
verdict of the coin toss rather than revert to violence in the

160
 how democracy works

quest for power. Similarly, while the winners would prefer not
to toss the coin again, they may be better off peacefully leaving
office rather than provoking violent resistance to their usur-
pation of power. Examine the situation from the point of view
of the losers in a particular election. They face the choice of
either reverting to violence in order to grab power by force or
accepting the cost of having lost and waiting to win the coin
toss the next time around. What they will do depends on their
chances of prevailing by force, on the cost of fighting, on the
loss entailed by being governed against their will, and on their
chances of winning the next time. This calculus may go either
way, but they will wait so long as the policies imposed by the
winners are not too extreme or so long as their chance to win
at the next opportunity is sufficiently high. In turn, the win-
ners know that to prevent the losers from rising in arms they
have to moderate their policies or not abuse their incumbent
advantage to deny the current losers the chance to win in the
future. Regulating conflicts by a coin toss generates a situation
in which peacefully waiting for one’s chance may be best for
each party given that the other party does the same.
Bloodshed is avoided by the mere fact that the political forces
expect to take turns.
Yet we do not use random devices; we vote. Voting is
an imposition of a will over a will. When a decision is reached
by voting, some people must submit to an opinion different
from theirs or to a decision contrary to their interest. Voting
generates winners and losers, and it authorizes the winners to
impose their will, even if within constraints, on the losers.
What difference does it make that we vote? One answer to this
question is that the right to vote imposes an obligation to

161
 the future?

respect the results of voting. In this view, losers obey because

they see it as their duty to obey outcomes resulting from
a decision process in which they voluntarily participated.
Outcomes of elections are “legitimate” in the sense that people
are ready to accept the decisions of as-yet undetermined
content so long as they can participate in the making of
those decisions. I do not find this view persuasive, yet
I think that voting does induce compliance, through
a different mechanism. Voting constitutes “flexing muscles”:
a reading of chances in the eventual conflict. If all men are
equally strong (or armed) then the distribution of votes is
a proxy for the outcome of war. Clearly, once physical force
diverges from sheer numbers, when the ability to wage war
becomes professionalized and technical, voting no longer
provides a reading of chances in a violent conflict. But voting
does reveal information about passions, values, and interests.
If elections are a peaceful substitute for rebellion, it is because
they inform everyone who would mutiny and against what.
They inform the losers – “Here is the distribution of force: if
you disobey the instructions conveyed by the results of the
election, I will be more likely to beat you than you will be able
to beat me in a violent confrontation” – and the winners –
“If you do not hold elections again or if you grab too much,
I will be able to put up a forbidding resistance.” Elections,
even those in which incumbents enjoy an overwhelming
advantage, provide some information about the chances of
conflicting political forces in an eventual violent resistance.
They reduce political violence by revealing the limits to rule.
In the end, elections induce peace because they enable
intertemporal horizons. Even if one thinks that people care

162
 how democracy works

about outcomes rather than procedures, the prospect that

parties sympathetic to their interests may gain the reins of
government induces hope and generates patience. For many,
the United States election of 2000 was a disaster, but we knew
that there would be another one in 2004. When the 2004
election ended up even worse, we still hoped for 2008. And,
as unbelievable as it still appears, the country that elected and
re-elected Bush and Cheney, voted for Obama. Those who
voted against Trump now hope he will be defeated in 2020.
Elections are the siren of democracy. They incessantly rekin-
dle our hopes. We are repeatedly eager to be lured by pro-
mises, to put our stakes on electoral bets. Hence, we obey and
wait. The miracle of democracy is that conflicting political
forces obey the results of voting. People who have guns obey
those without them. Incumbents risk their control of govern-
mental offices by holding elections. Losers wait for their
chance to win office. Conflicts are regulated, processed
according to rules, and thus limited. This is not consensus,
yet not mayhem either. Just regulated conflict; conflict with-
out killing. Ballots are “paper stones.”
Yet this mechanism does not always work. Elections
peacefully process conflicts if something is at stake in their
outcomes, but not too much (Przeworski, Rivero, and Xi
2015). If nothing is at stake, if policies remain the same
regardless of who wins, people observe that they voted in
election after election, governments changed, and their lives
remained the same. They may conclude that elections have no
consequences and lose incentives to participate. The mirror
danger occurs when too much is at stake, when having been
on the losing side is highly costly to some groups and their

163
 the future?

prospects to be on the winning side in the future are dim, so

that they see their losses as permanent or at least long-lasting.
When incumbent governments make it next to impossible for
the opposition to win elections, the opposition has no choice
but to turn away from elections.

9.3 Government and Opposition Between

Elections
An argument can be made that maintaining public order
between competitive elections should not be problematic,
precisely because the prospect of being able to win future
elections is sufficient to induce the current losers to suffer in
silence between elections. While O’Donnell (1994) diag-
nosed the reduction of politics to elections as a Latin
American pathology, “delegative democracy,” for James
Madison this was how representative government should
function: the people should elect governments but then
have no role in governing. Lippman (1956) insisted that
the duty of citizens, “is to fill the office and not to direct
the office-holder.” Schumpeter (1942) admonished voters
that they “must understand that, once they elected an indi-
vidual, political action is his business not theirs. This means
that they must refrain from instructing him what he is
to do.”
As a description, this picture is obviously inaccurate
(Manin 1997, 2017). Conflicts over policies are the bread and
butter of everyday politics. Political activities are not limited
to elections, nor even to efforts oriented toward influencing
the outcomes of future elections. Moreover, while opposition

164
 how democracy works

to government policies can be limited to the institutional

framework, under some conditions it spills outside of it.
The parliamentary opposition can stop or modify
some actions of the government. If a policy proposed by the
government is subject to legislative approval, the government
may fail in parliament. Opposition parties may persuade
government supporters to modify their views; they can exer-
cise its institutional prerogatives to block some legislation (in
Germany presidencies of parliamentary committees are dis-
tributed proportionately to party strength; in the United
Kingdom the Committee of Public Accounts is by convention
controlled by the opposition; in Argentina passing legislation
requires a supermajoritarian quorum); they can threaten with
obstructive tactics (a government proposal to privatize an
electric utility company was met with thousands of amend-
ments in France; filibustering in the United States Senate);
they can threaten non-cooperation at the lower levels of
governments they control. Note that if elections are expected
to be competitive, the opposition faces a strategic choice of
either accepting concessions from the government or going
for broke with the hope of unseating the government in the
next election. For example, in Brazil under the presidency of
Fernando Henrique Cardoso, most parties were willing to
support the government in exchange for pork barrel spend-
ing, but the Workers’ Party (PT) invariably voted against the
government and won the subsequent presidential election.
The opposition may also seek recourse to constitu-
tional courts in order to restrict the actions of the govern-
ment. Note that the logic of the role of elections in peacefully
processing conflicts extends to the courts. Conflicting sides

165
 the future?

are willing to respect the verdicts of constitutional tribunals

when they believe in their impartiality, specifically, that the
tribunal considers each case on its merits. The losing side
obeys the courts when it believes that in some future cases it
may find itself to be the winner. When courts are blatantly
partisan, this belief is eroded, and addressing conflicting
issues to constitutional tribunals becomes futile.
Opposition, however, need not be limited to legisla-
tures and courts. It can take place on the streets, in factories,
or in offices. Street demonstrations are a standard repertoire
of democratic opposition, as are strikes. As long as they are
orderly and peaceful, they are just a routine tactic by which
some groups signal their opposition to particular policies or
their general dissatisfaction with the government. But
demonstrations are not always peaceful: sometimes they are
gratuitously repressed, sometimes they deteriorate into vio-
lence by marginal groups of demonstrators (they are called
casseurs, “breakers” in France). The line between legal and
illegal is thin. Hofstadter’s (1969: 7) observation that
“The normal view of governments about organized opposi-
tion is that it is intrinsically subversive and illegitimate” con-
tinues to be haunting. The idea that opposition to government
policies does not need to signify treason or obstruction was
first recognized in Great Britain in a parliamentary speech of
1828. But what kind of opposition is loyal and what kind
subversive? Must opposition to government policies be chan-
neled through the framework of representative institutions, or
can people act in any way they please? Babasaheb Ambedkar,
the father of the Indian constitution, thought that while civil
disobedience was appropriate under colonial rule, it is

166
 how democracy works

“nothing but the Grammar of Anarchy” under democracy.

In the words of David Cameron, the former British prime
minister, demonstrations by students against raising tuition
fees “were a part of democracy but violence and law-breaking
was not” (BBC News). Actions such as blocking roads and
bridges, occupying buildings, lock-outs, civil disobedience,
rioting, and in the extreme terrorism, are intended to under-
mine the government by undermining public order.
Moreover, violence is not always directed against govern-
ments. We have seen that in many instances private groups,
sometimes organized as paramilitary organizations but some-
times just forming spontaneously, engage in violence against
each other: this was the case in Weimar Germany and Chile,
as well as in the United States during the 1960s.
Demonstrations that end in violence, violent labor
conflicts, blockages of roads and bridges, occupations of
buildings, lock-outs, civil disobedience, street fights, riots,
and terrorism are what I mean by conflicts spilling outside
the institutional boundaries. They constitute breakdowns of
public order. They are costly to the perpetrators, to the gov-
ernment, and often to third parties. They may occur as a result
of strategic decisions of some groups but they may also erupt
spontaneously.
Consider a situation in which a government has the
monopoly of legislative initiative and is assured of the support
of a majority in the legislature. All bills are initiated by the
executive and all the bills become laws. Moreover, the govern-
ment acts with full legality or the courts are partisan, so any
recourse to the judicial system would be futile. Examine this
situation from the point of view of a social group opposed to

167
 the future?

a particular policy. This group has no chance of influencing

government policy within the institutional system: the gov-
ernment wants to adopt the policy, the legislature is just
a rubber stamp and offers no recourse. The most this group
can expect of the system of representative institutions is that if
the policy turns out to be sufficiently unpopular, the govern-
ment would lose the next election and the policy would be
reversed. But suppose that in addition, the government has
a good chance of being re-elected. Then this group has noth-
ing to gain by acting within the institutional framework.
Under such conditions it may be sufficiently desperate to try
stopping the policy by acting outside the institutional
channels.
Saiegh generated interesting information relating the
rate at which bills proposed by the executive are approved as
laws by legislatures (“Box score” in Figure 9.1) and the inci-
dence of riots. Governments do not always get what they want
in the legislatures: according to Saiegh (2009), democratic
legislatures approved only 76 percent of bills proposed by
the executive during the 783 country years for which these
data are available. In turn, under democracy (Saiegh’s regime
classification is based on Alvarez et al. 1996), riots are more
frequent when the executive is either not at all effective or
when the legislature is just a rubber stamp.
I interpret these patterns as saying that institutions
are successful in regulating conflicts when the government is
sufficiently able to govern but the opposition has an impor-
tant voice in policy making. Politics spills out of institutional
bounds either when governments are too weak to be able to
pass legislation or so strong that they do not need to

168
 how democracy works

6
5
Predicted number of riots
4
3
2
1
0
•1

0 10 20 30 40 50 60 70 80 90 100
Box score

Figure 9.1. Proportion of bills passed and riots

Source: Saiegh (2009)

accommodate legislative opposition. As several French poli-

ticians commented in the aftermath of President Macron’s
massive legislative victory, “if the debate does not take place in
the parliament, it will occur on the streets.”
Breakdowns of public order tend to spiral.
Historical experience suggests that when conflicts spill
onto the streets, public support for authoritarian measures
designed to maintain public order tends to increase, even
when protests are targeted precisely against the authoritar-
ian tendencies of governments. People do expect govern-
ments to maintain order; indeed, no society can tolerate
permanent disorder. Protracted public transportation
strikes or strikes causing shortages, road blockages, or

169
 the future?

other actions that paralyze everyday life provoke

a backlash even among the people who are sympathetic
to the cause of strikers. Repeated street fights induce the
atmosphere of disorder and insecurity. Hence, govern-
ments are always tempted to portray actions against them
as illegal. Particularly dangerous are “situations in which
the authorities, the police, and the judiciary, even if dis-
approving of violent political acts, dealt leniently with
them because they felt sympathetic to the motives of
those engaging in them or hostile to their victims” (Linz
1978: 57). In turn, in such political climates the repressive
forces, whether ordinary or riot police, feel authorized to
use violence even in the face of peaceful protests: think of
the “police riot” during the 1968 Chicago Democratic Party
convention. When these forces are not well trained and
disciplined, tragic accidents are almost inevitable: think of
the massacre of students at Kent State University on May 4,
1970. And when peaceful actions are brutally repressed,
some people conclude that they are being pushed out of
the institutional framework and revert to terrorism, as in
the United States, Germany, and Italy in the 1960s–1970s.
I do not claim that these are regular patterns: we
know little that is systematic about dynamics of disorder
and repression. The only conclusion one can draw from
these examples is that breakdown of public order is something
which all governments must fear. Faced with demonstrations
that turn violent, road blockages, protracted transportation
strikes, or fights between private groups, governments have
only two choices: either to persevere with their policies while
reverting to repression or to abandon their policies in order to

170
 how democracy works

placate the opposition. Neither alternative is attractive.

The spirals of unrest and repression undermine public
order, while repeated concessions render governments unable
to implement any stable policies.

9.4 How Democracies Fail

Democracy works well when representative institutions
structure conflicts, absorb them, and regulate them
according to rules. Elections fail as a mechanism for
processing conflicts either when their outcomes have no
consequences for people’s lives or when incumbents abuse
their advantage to the point of making them non-
competitive. Once elected, governments must be able to
govern, but they cannot ignore the views of intense mino-
rities. When conflicts are intense and a society is highly
polarized, finding policies acceptable to all major political
forces is difficult and may be impossible. Miscalculations,
whether by governments or different groups opposing
them, lead to institutional breakdowns. When govern-
ments ignore all opposition to their policies, when they
interpret all opposition as subversive, when they engage
in gratuitous repression, they push the opposing groups
out of the institutional framework: opposition turns into
resistance. When some groups of the opposition refuse to
accept policies resulting from applying the institutional
rules, governments may have no choice but to engage in
repression to maintain public order. Finding the right
balance between concession and repression is a subtle
choice. Failures are inevitable.

171
 10

Subversion by Stealth

10.1 Democratic Backsliding

The dream of all politicians is to remain forever in office and
to use their tenure to do whatever they want. Most democratic
governments attempt to advance these goals by building pop-
ular support within the established institutional framework.
Some, however, seek to protect their tenure in office and to
remove obstacles to their discretion in choosing policies by
undermining institutions and disabling all opposition.
Prominent recent examples are Turkey under the government
of the Justice and Development Party (AKP), Venezuela
under Chavez and Maduro, Hungary under the second gov-
ernment of Fidesz, and Poland under the second government
of PiS.
Democratic “deconsolidation” or “backsliding” is
a process of gradual erosion of democratic institutions and
norms. Ginsburg and Huq (2018a: 17) use the term “authoritar-
ian retrogression,” distinguished from outright “reversion,” and
define it as “a process of incremental (but ultimately still sub-
stantial) decay in the three basic predicates of democracy –
competitive elections, liberal rights to speech and association,
and the rule of law.” As “backsliding” or “deconsolidation” or
“retrogression,” whatever one wants to call it, advances, the
opposition becomes unable to win elections or assume office if

172
 subversion by stealth

it wins, established institutions lose the capacity to control the

executive, and manifestations of popular protest are repressed
by force. This process is propelled by the desire of a government
to monopolize power and to remove obstacles to realizing its
ideal policies. Yet it is a process of interaction between the
government and various actors that seek to block it. Hence,
the strategy of the governments that go down this path concen-
trates on disabling potential blockers, who differ from case to
case but typically include the opposition parties, the judicial
system, and the media, as well as the streets.
Think about it as follows. A government deeply com-
mitted to a particular ideological goal – such as Islamization
in Turkey, “Bolivarianism” in Venezuela, “preserving the
purity of the nation” in Hungary, or “defending
Christianity” in Poland – wins an election.1 This government
decides whether to take steps to increase its ability to remain
in office or steps that would increase its discretion in policy
making. As Lust and Waldner (2015: 7) put it, “Backsliding
occurs through a series of discrete changes in the rules and
informal procedures that shape elections, rights and

1
The definition of “victory” is not as simple as it may appear. Electoral laws
play an important role: in Turkey, the AKP won 34.3 percent of votes to
obtain 66 percent of seats when it first assumed power in 2002, in
Hungary Fidesz won 53 percent of votes and 68 percent of seats in 2010, in
Poland PiS got 37.5 percent of votes and 51 percent of seats. The ascension
of Chavez to office in Venezuela was convoluted: the traditional parties
actually won the legislative election of 1998, Chavez won the presidential
election with 56.4 percent, the referendum for a new constitution was
passed by 71.8 percent, then Chavez won a new election with 59.8 percent
while his party obtained 44.4 percent of the votes and 55.7 of the seats in
the legislative election of 2000.

173
 the future?

accountability. These take place over time, separated by

months or even years.” Examples of the first kind of steps
include changing electoral formulae, redistricting, changing
voting qualifications (age, eligibility of citizens residing
abroad), harassing the partisan opposition, or imposing
restrictions on non-governmental organizations. Examples
of the second kind of steps include shifting power from the
legislature to the executive, reducing the independence of the
judicial system, or using referendums to overcome constitu-
tional barriers. Some measures, such as instituting constitu-
tional reforms, imposing partisan control over state
apparatuses, or controlling the media, have both effects.
Having observed such steps, citizens who value democracy
may turn against the government even if they support its
policies or enjoy outcomes which they attribute to its policies.
If the opposition rises, the government may be removed from
office or it may decide to stop taking further steps, anticipat-
ing its rise.
In principle, the opposition could prevent govern-
ments from taking the next step either by stopping it by
legal measures, such as defeating a bill in parliament, or
obtaining a presidential veto or favorable court ruling.
The history of the four cases we referred to above shows,
however, that governments regularly overcome legal obsta-
cles. In Turkey, when in 2007 the president vetoed
a constitutional amendment passed by the parliament for
direct election of the president, the government organized
a referendum and won. In Venezuela, when the opposition
won a legislative election in December 2015, Maduro
replaced the Congress with a newly elected Constitutional

174
 subversion by stealth

Assembly. In Hungary, when the Constitutional Tribunal

invalidated an electoral reform in 2013, the government
passed a constitutional amendment curbing the power of
the tribunal. In Poland, when the president vetoed two
laws concerning the courts, he was quickly persuaded to
change his mind. This is not to say that governments
always prevail: in Poland, for example, the government
withdrew from fining an opposition TV station that is
owned by US interests. Nevertheless, it seems that legal
counter-moves by the opposition at most slow down the
process but are ineffective in stopping it. This is why we
think of the opposition in terms of an effective threat to
removing the government and reversing the process of
democratic deterioration.
The obvious question is why some governments
decide to go down this path while most refrain from it.
The second is whether, once a government takes such
steps, it can ever be stopped short of the realization of
complete domination while still remaining in office.
The third is whether the potential opposition would be
able to remove the government and reverse this process.
It bears emphasis that we try to understand the “how”
rather than the “when”: how does deconsolidation evolve
when it does, rather than under what conditions it is likely
to occur. We learn from the experiences of the four countries
mentioned above – we sample on the dependent variable –
because we want to understand how democracies can be
destroyed by small steps, not why backsliding occurs in
some countries and not in others, for which see Maeda
(2010), Svolik (2015), and Graham, Miller, and Strøm (2017).

175
 the future?

Our question is not “Will it happen here?” but “Can it happen

anywhere?”

10.2 Stealth
The puzzle entailed in the destruction of democracy by “back-
sliding” is how a catastrophic state of the world can be gra-
dually brought about by small steps, against which people
who would be adversely affected do not react in time.
As Ginsburg and Huq (2018b: 91) pose it, “The key to under-
standing democratic erosion is to see how discrete measures,
which either in isolation or in the abstract might be justified as
consistent with democratic norms, can nevertheless be
deployed as mechanisms to unravel liberal constitutional
democracy.” In the parable of “the frog in the pot,” if a frog
is put suddenly into hot water, it will jump out, but if it is put
in cold water which is then heated slowly, it will not perceive
the danger and will be cooked to death. Yet the parable is not
true: recent experiments show that the frog will get uncom-
fortable when the water is heated and will try to jump out (see
“Boiling Frog” on Wikipedia). How, then, can gradual back-
sliding succeed in destroying democracy?
The first lesson we are learning from recent experi-
ences is that democracies do not contain institutional mechan-
isms that safeguard them from being subverted by duly elected
governments observing constitutional norms. When Hitler
came to power, through an “authoritarian gap in the Weimar
Constitution” (Article 48, which allowed the president to
empower the government to rule by decree; Bracher 1966:
119), the possibility of a legal path to dictatorship was seen as

176
 subversion by stealth

a flaw of this particular constitution. Yet such gaps may be

generic. The father of constitutionalism, Montesquieu (1995:
326), insisted that “For the abuse of power to be impossible, it is
necessary that by the disposition of things, the power stop the
power.” But, pace Madison (Federalist #51), checks and bal-
ances do not operate effectively when different powers of the
government are controlled by the same party: as Madison
himself was almost immediately to discover (Dunn 2004:
47–61), the constitutional separation of powers is vulnerable
to partisan interests. Courts, constitutional as well as ordinary,
can be packed, intimidated, or circumvented. Wholesale
changes of constitutions, amendments, or referendums can
constitutionally overcome extant constitutional obstacles.
Public bureaucracies, including security agencies, can be
instrumentalized for partisan purposes. Public media can be
controlled by partisan regulatory bodies, while private media
can be legally intimidated or destroyed financially. All such
measures can be taken legally. As Landau (2013: 192–3)
observes, “The set of formal rules found in constitutions is
proving to be a mere parchment barrier against authoritarian
and quasi-authoritarian regimes. There is even worse news:
existing democracy-protecting mechanisms in international
and comparative constitutional law have proven ineffective
against this new threat.”
Democratic deconsolidation need not entail viola-
tions of constitutionality. Thinking about the United States,
a constitutional lawyer writes,

If it happens here, it won’t happen all at once . . . Each step

might be objectionable but not, by itself, alarming . . . there

177
 the future?

will have been no single, cataclysmic point at which

democratic institutions were demolished . . . the steps
toward authoritarianism will not always, or even usually,
be obviously illegal . . . In fact, each step might conform to
the letter of the law. But each step, legal in itself, might
undermine liberal democracy a little bit more. (Strauss
2018: 365–6)

In a broader context, another constitutional lawyer

concludes that “it is difficult to identify a tipping point during
the events: no single new law, decision or transformation seems
sufficient to cry wolf; only ex-post do we realize that the line
dividing liberal democracy from a fake one has been crossed:
threshold moments are not seen as such when we live in them”
(Sadurski 2018: 5). This is what we mean by “stealth”: “the use of
legal mechanisms that exist in regimes with favorable demo-
cratic credentials for anti-democratic ends” (Varol 2015).
In turn, when the government takes steps that are not
flagrantly unconstitutional or undemocratic, citizens who ben-
efit from its policies but still value democracy are not certain
how to react. Some measures that backsliding governments
adopt do not even require legal acts, just a change of practices.
For example, the Polish ruling party, PiS, gradually altered the
parliamentary procedure concerning the introduction of new
bills: the parliamentary rules say that bills proposed by the
government must be subject to public hearings, while private
member bills are not, and the government shifted to offering its
proposal as private bills of its deputies (Sadurski 2018: 6).
Moreover, nothing is constitutionally wrong with legal measures
such as a parliamentary act easing restrictions on the teaching of

178
 subversion by stealth

the Koran (June 2005 in Turkey), anti-terror laws (June 2006 in

Turkey, May 2016 in Poland), or a statute that requires non-
governmental organizations to register as foreign organizations
if they receive funding from abroad (June 2017 in Hungary).
These are ordinary laws, passed according to constitutional
provisions by the legally competent bodies, a prerogative of
any democratic government. Even changes of constitutions are
valid so long as they observe constitutional provisions, as they
did in Hungary in April 2011, and in Turkey after the referen-
dums of October 2007 (direct election of the president, after the
incumbent president vetoed a parliamentary act),
September 2010 (increased civilian control over the military
and the courts), and April 2011 (introducing a presidential sys-
tem). As a result, several observers agree, “there is usually no
single event or governmental conduct which may mobilize the
resistance by sending a clear signal that democratic norms are
imperilled” (Ginsburg and Huq, 2018a), “slow slides towards
authoritarianism often lack both the bright spark that ignites
an effective call to action and the opposition and movement
leaders who can voice that clarion call” (Bermeo 2016: 14).
Protests against legal measures taken by a govern-
ment that just won an election only show that the opposition
is a sore loser, that it does not respect democratic norms. Even
more perverse are the situations in which governments suc-
ceed in packing or controlling the constitutional tribunals and
then use judicial review to legitimize their actions, as they did
in Venezuela, Turkey, and Hungary. They are perverse
because they allow governments to portray actions of the
opposition as anti-constitutional: in President Trump’s

179
 the future?

tweet the appointment of Special Counsel Robert Mueller was

“totally UNCONSTITUTIONAL” (@realDonaldTrump June
4, 2018 at 4.01 p.m.). Part of the rhetorical games entailed in
backsliding is to compete for the mantle of “democracy” and
“constitutionalism,” in which the opposition does not always
prevail. Thus a pro-Putin Russian journalist, Mikhail
Leontiev, disingenuously observes: “I do not understand
what is undemocratic in that some force enjoying overwhelm-
ing social support wins elections” (an interview with a Polish
newspaper, Dziennik, January 19, 2008).
Note that claims of “unconstitutional” in principle
require a more rigorous test than claims of “undemocratic.”
Declarations that a particular statute or an action of the gov-
ernment violates the constitution are issued by specialized
bodies designated by a constitution and are couched as inter-
pretations of its text. But courts are populated bodies (Ferejohn
and Pasquino 2003): their decisions are made by particular
people, appointed by politicians. Hence, if a government suc-
ceeds in stuffing these bodies with its partisan supporters, they
issue decisions favorable to the government. Venezuela is
a flagrant example. Moreover, constitutions can be amended
or completely replaced, still by constitutional provisions. Even
when the extant constitution is entrenched, the entrenchment
clauses can be modified to enable constitutional change.
True, constitutionality can be questioned even when
a court is constitutionally formed and when it rules that
actions of the government are constitutional. One issue is
whether an act adopted following constitutional provisions
that abrogates constitutionality altogether can ever be con-
sidered as constitutional. The prominent example is the

180
 subversion by stealth

German Enabling Act (Gesetz zur Behebung der Not von Volk
und Reich) of March 24, 1933 that gave the government the
power to act in an extra-constitutional way: “In addition to
the procedure prescribed by the constitution, laws of the
Reich may also be enacted by the government of the Reich”
(Article 1). A literal interpretation of “constitutionality” does
not provide a decisive criterion. Some broader conception
must be invoked to deem actions of a government as violating
constitutionality when the pertinent bodies declare them to be
constitutional. Thus Landau (2013: 195) considers “the use of
mechanisms of constitutional change in order to make a state
significantly less democratic than it was before” as “abusive
constitutionalism,” while the Colombian Constitutional
Court ruled that even duly adopted constitutional amend-
ments can be unconstitutional (Ginsburg and Huq 2018b:
188). Given that such conceptions are inescapably vague,
partisan-based disagreements are inevitable.
The notion of “undemocratic” is even more per-
missive. There is nothing “undemocratic” about the elec-
tion of Donald Trump: in the words of his advertisement,
“Our movement is about replacing a failed and corrupt
political establishment with a new government controlled
by you, the American people.”2 It is even more paradoxical
to claim the same about referendums, in which Marine Le
Pen, advocating a vote on “Frexit,” promised “You, the
people, will decide.” Not holding scheduled elections or
committing blatant fraud are almost universally seen as
violations of democratic norms, but short of this

2
See www.youtube.com/watch?v=vST61W4bGm8.

181
 the future?

instinctive marker the norms of what constitute undemo-

cratic practices are less crystallized and diverge even more
across partisan lines. For example, in the United States
almost everyone believes that fraud-free elections with
equal voting rights are important for democracy, but
fewer people think that it is important that districts not
be biased or that the government not interfere with the
press, with divergence between supporters and opponents
of President Trump (Bright Line Watch 2018).
Moreover, it is often controversial whether some
steps taken by governments are anti-democratic. Imagine
that a government extends voting rights to citizens residing
abroad (Erdoğan did it, but also Berlusconi), or it adopts
legislation to require additional documentation at the polling
place, or it remaps electoral districts. Are these steps anti-
democratic? The government says “We want to extend rights
to all citizens,” “We want to prevent fraud,” “We want every
vote to have equal weight.” The opposition says
“The government does not care about rights and is extending
the vote to Turks in Berlin only because they will vote for it,”
“The government does not care about fraud, it only wants to
prevent poor people, who do not have the documents, from
voting,” “The government is redistricting in its favor.” All
these steps are adopted following constitutional provisions,
so they are not “undemocratic” in the sense of violating
procedural norms. The disagreements are not about facts
but about intentions and intentions are not directly
observable.
Now, governments may take actions that are fla-
grantly unconstitutional or undemocratic. Refusing to

182
 subversion by stealth

comply with court rulings is clearly unconstitutional and in

the United States makes the perpetrator subject to contempt-
of-court charges. Banning an opposition newspaper flagrantly
violates democratic norms. Moreover, some are offensive
whether or not they are either: shooting peaceful demonstra-
tors, as at Kent State University on May 4, 1970, may provoke
widespread outrage whatever the legal or normative niceties.
Stealth is a process by which the government takes certain
steps, none of which is flagrantly unconstitutional or unde-
mocratic, and which cumulate in undermining the capacity of
the opposition to remove it or enlarge its discretion in making
policies.

10.3 Dynamics of Subversion from Above

“Parchment barriers” are not sufficient to prevent the erosion
of democracy by governments that proceed by stealth.
The question, then, is whether a government intent on back-
sliding can be dissuaded from pursuing it or be removed
because of the rise of popular opposition.
Suppose that at each moment the government deci-
des whether to take a step that increases its incumbent advan-
tage or brings the policy closer to its ideals or both,
anticipating the size of the opposition to such steps. In turn,
individual citizens decide whether to turn against the govern-
ment or to support it, perhaps waiting to see if the govern-
ment would continue taking additional steps. The probability
that the government remains in office depends on the number
of steps it takes to protect its tenure from the potential
opposition and on the size of the opposition. The more such

183
 the future?

steps the government successfully takes, the larger the oppo-

sition required to generate the same chance of removing it.
The government enters the backsliding path if taking some
steps will make it better off in terms of staying in office and
implementing its preferred policies than the institutional
status quo. It stops if the gain from taking further steps is
outweighed by the danger of increased opposition. At each
step it may be removed from office according to the probabil-
ities described above.
Individuals attach different weights to the extent to
which they care about democracy and about some policies,
such as anti-immigration measures, or outcomes, such as
income growth (Svolik 2017, Graham and Svolik 2018). People
who do not like the outcomes the government generates, say
environmentalists under the Trump government, oppose it
regardless of the value they place on democracy. In turn, people
who support government policies trade off their benefits from
policies against the damage to democracy at different rates,
where “democracy” means that the government can be
removed by elections (or some constitutional provisions, such
as impeachment or a vote of non-confidence), when a sufficient
majority opposes it. Obviously, there may be people who do
not care about democracy at all.
While this much is general, there are many possibi-
lities. The opposition may not rise at all; it may remain
dormant at first and only then suddenly increase; it may
remain at some constant level or just rise sporadically in
reaction to particular measures of the government. Given
that by making appropriate assumptions one can obtain any
result one wants, to generate conclusions by assuming them,

184
 subversion by stealth

we keep these possibilities as open as possible by exploring the

generic conditions under which a government enters on the
path of deconsolidation, as well as the conditions, if any,
under which it ever stops under the threat of opposition, or
is removed from office while deconsolidating.
The central conclusion3 is that everything depends on
whether people who are at all concerned about democracy
anticipate the effects of particular steps on the long-term
future. If individuals anticipate the cumulative effect of back-
sliding, those who value democracy would turn rapidly
against the backsliding government and, expecting this reac-
tion, a government intent on backsliding would desist from
taking or continuing far on this path. This would be true even
if people were uncertain to begin with whether the govern-
ment intends to backslide and would update their beliefs
about the type of government they face only when the govern-
ment takes some steps. In turn, if individuals react only to
their current situation, the opposition rises too slowly to
prevent a government intent on backsliding from taking
a sufficient number of steps to secure its incumbent advantage
and to remove institutional obstacles to its discretion in
making policy. The opposition rises somewhat faster if some
policy measures of a backsliding government would be also
taken by a democratic one, and it rises somewhat slower if
individuals are uncertain whether the particular steps are
intended to increase incumbent advantage, but neither possi-
bility prevents the government from backsliding. Hence,

3
The conclusions are based on a series of mathematical models developed
jointly with Zhaotian Luo. See Luo and Przeworski (2018).

185
 the future?

unless people react at the onset against actions by the govern-

ment that would have a cumulative effect of eroding democ-
racy, democracy erodes.
Defending democracy imposes a difficult, perhaps
impossible, challenge for individual citizens. To act now
against the government that may in some future destroy
democracy, individuals who currently enjoy its policies or
some outcomes which they attribute to policies4 must see
the long-term effect of current policies. Even if individuals
have consistent time preferences (Akerlof 1991), and even if
they are concerned about the future, they must be able to
calculate the cumulative effect of particular, seemingly
democratic steps – they must be able to see through the
stealth. This is a formidable task and, even if the incapacity
to anticipate the future violates the assumption of full
rationality, it should not be surprising if people cannot
perform it. Consider a sequence of events in which the
government first adopts legislation to require additional
documentation at the polling place, then has its cronies
buy an opposition newspaper, then remaps electoral dis-
tricts, and then controls the bodies that administer and
supervise elections. What people need to see is that, while
each of these measures may have little effect, their cumulative
effect is to protect the incumbent from being defeated even
by a largely majoritarian opposition. Moreover, policies have

4
Note that in Turkey per capita incomes grew at the annual rate of 4.4%
under the AKP government, Venezuela enjoyed spectacular growth
between 2004 and 2011 (except for 2009) due to oil prices, Hungary grew
at the rate of 3.5% under Fidesz, and Polish incomes continued to grow
under PiS. (Data from PWT 9.0, ending in 2014.)

186
 subversion by stealth

interactive effects that are even more difficult to calculate:

Sheppele’s (2013) example is the interaction of Article 48 of
the Weimar Constitution, that permitted the president to
declare a state of emergency, subject to the check that parlia-
ment could reject such a declaration, and Article 25 that
permitted the president to dissolve the parliament once for
any reason, the effect of which was that once parliament was
dissolved, the president could declare an emergency at will.
Even the most competent constitutionalists, who authored
this constitution, did not see the potential effects of this
combination, and it proved fatal.
One might think that people would be educated to
look toward the long-term future by leaders of the opposi-
tion. But their potential role is limited. Exhortations by
partisan opposition are not likely to be effective in influen-
cing beliefs. People know that the goal of the leaders of the
opposition is to replace the incumbents, whether they act for
good or bad reasons. If opposition leaders criticize every
action of the government, people tend to dismiss their mes-
sages: as Austen-Smith (1992) put it, speech that is predict-
able according to interests is not credible. And if only the
extreme opponents of the government turn out on the
streets, the government can claim that the opposition is
undemocratic, and others are less likely to join (Shadmehr
and Berhardt 2011).
The fact is that backsliding governments have enjoyed
continued popular support. To the best of my knowledge, the
only case in which a backsliding government lost an election and
left office is in Sri Lanka in 2015, and this outcome resulted from
massive defections from the ruling coalition, with the winner

187
 the future?

previously being a minister in the outgoing government. Other

backsliding governments have suffered temporary reversals but
were able to recover and continue: with 40.9 percent of votes, the
AKP failed to win a majority of seats in the election of June 7,
2015, but it called for a new election and won 49.5 percent of the
vote five months later. Three years later, in June 2018, Erdoğan
won the presidential election with 52.6 percent. In Poland,
a majority of survey respondents thought that the government
was “performing badly” when it began to tinker with the
Constitutional Court in 2015, but the government was positively
evaluated by a majority two years later, by October 2017 (Kantar
Public 2018). In Hungary, Fidesz and its allies won re-election
in April 2018 with 44.9 percent of the vote. In Venezuela, Chavez
won re-election in 2006 with 62.8 percent of the vote and again
in 2,102 with 55.1 percent. He enjoyed majority support in the
polls and the opposition became majoritarian only after his
death (Venezuelabarometro). And in the United States, the
popularity of President Trump hovers narrowly around 40 per-
cent regardless of anything. The implication must be that either
many people do not care at all about democracy or that they do
not see the long-term consequences for democracy when they
vote or answer survey questions.

10.4 Could It Happen Here?

All conclusions must be speculative. Intentions matter, determi-
nation to pursue them matters, resistance is effective only when it
is optimally timed and sustained, and the conditions under
which the opposition rises are difficult to satisfy. But the opti-
mism that citizens would effectively threaten governments that

188
 subversion by stealth

commit transgressions against democracy and thus prevent

them from taking this path (Montesquieu 1995: book 19, chapter
19, Weingast 1997, 2015, Fearon 2011) is sadly unfounded.
Montesquieu had hoped that if power were to be abused, “every-
thing would [be] united against it,” there would be a revolution,
“which would not change the form of government or its con-
stitution: for revolutions shaped by liberty are but a confirmation
of liberty.” Yet these views are based on the assumption that the
government commits some acts that flagrantly threaten liberty,
violate constitutional norms, or undermine democracy. But
when a government proceeds by stealth, citizens turn against it
only if they see what its actions are leading to in the long run.
Hence, resistance against a backsliding government imposes
a difficult challenge on individual citizens. The effect of stealth
is to obscure the long-term danger. And if the opposition does
not stop the government from taking some series of legal steps, it
will be too late to prevent it from taking illegal ones.
Can it happen anywhere? Could it happen in the
United States? Here is a nightmare scenario.
First: Congress passes a law prohibiting the publica-
tion of “false, scandalous, and malicious writings against the
government of the United States, the Congress, or the
President, with intent to bring them into contempt or dis-
repute” (Sedition Act of 1798; in Stone: 2018: 491) or those
using “disloyal, profane, scurrilous, or abusive language”
about the United States government, its flag, or its armed
forces or that caused others to view the American government
or its institutions with contempt. Those convicted under the
act will receive sentences of imprisonment for five to twenty
years (Sedition Act of 1918).

189
 the future?

Next: the Supreme Court grants large discretion to

state legislatures in designing districting plans.
Next: Congress passes a law, according to which mak-
ing public official documents without the authorization of the
government agency is subject to fines or imprisonment. “[T]he
Constitution imposes little constraint on the selective disclosure
(or nondisclosure) of information by the state in ways that can
shunt public debate away from questions that would embarrass
or undermine political leaders” (Ginsburg and Huq 2018b: 67).
Next: Congress passes a law against electoral fraud,
mandating states to adopt rules concerning documents
required to register to vote.
Next: the number of federal judges nominated by the
administration reaches 112 (the total number of vacancies to
be filled before 2020).
Next: the president issues an executive order decree-
ing that “all persons privileged to be employed in the depart-
ments and agencies of the Government, shall be reliable,
trustworthy, of good conduct and character,” and should
demonstrate “unswerving loyalty to the United States”
(Executive Order 10450 of President Eisenhower in 1953,
quoted in Goldsmith 2018: 106). Hundreds of civil servants
appointed before 2016 are purged.
Next: Congress passes a law withdrawing tax-exempt
status from non-governmental organizations which “obstruct
the implementation of duly adopted laws and regulations.”
Next: two Supreme Court justices are replaced by
administration nominees.
Next: Congress passes a new anti-terrorism law
according to which anyone who “threatens national security”

190
 subversion by stealth

is subject to preventive detention. “[Y]ou are kidding yourself

if you think the same thing will not happen again. Because
Inter arma enim silent leges . . . in times of war, the laws fall
silent” (Justice Scalia, referring to Korematsu vs United States,
quoted in Minow 2018: 321).
Next: the president is re-elected.
Next: the president issues a series of executive orders
on issues previously subject to legislation. Congress remains
mute. The Supreme Court remains mute.
The fat lady sings.

191
 11

What Can and Cannot Happen?

I will disclose to you what raised me to my position. Our

problems seemed complicated. The people did not know
what to do about them. In these circumstances people
preferred to leave them to the professional politicians. I, on
the other hand, have simplified the problem and reduced
them to the simplest formula. The masses recognized this
and followed me. (Adolf Hitler, quoted in Linz 1978: 53–4)
As J.K. Galbraith once remarked, “The only function of eco-
nomic forecasting is to make astrology look respectable.”
Making political predictions is even more hazardous.
Uncertainty is inherent under any circumstances and it is
exceptionally high when big issues are at stake and conflicts
are intense. Hence, the best we can perhaps achieve is to
identify the range of possibilities contained in the current crisis.
The most optimistic scenario in the economic realm is
that the crisis – the stagnation of lower incomes, job insecurity,
and the erosion of beliefs in intergenerational mobility – will just
blow over and, to the extent that the political discontent is driven
by the economy, so will the political impasse. Growth is accel-
erating in developed countries, with average incomes finally
surpassing the pre-2008 levels. Unemployment has declined
from post-2008 levels. Globalization, specifically the outflow of
jobs to low-wage countries, is slowing down. Wages in China
have increased 64 percent in five years, even if they remain low in
other countries. The rates of return to domestically located

192
 what can and cannot happen?

activities are in several sectors converging on those of foreign

investment. Internally, labor market reforms designed to
increase flexibility are expected by some to increase total employ-
ment. Accompanied by income-insurance programs, the com-
bination of flexible employment with income protection would
enhance efficiency while providing material security. Hence, in
this scenario, the economic crisis is fading, and in retrospect it
will be seen as just a temporary bump on the long-run path of
material progress.
The most dire scenario is that there is nothing on the
horizon that would reverse the stagnation of low incomes and
the insecurity resulting from the disappearance of better-paid
jobs. Even if growth does accelerate, there are no grounds to
expect that wages would increase at a similar rate. Hence,
inequality will continue to increase. Protectionism is unlikely
to protect jobs entailing traditional skills, either in industry or in
several service sectors. Even if protectionism reduces the outflow
of jobs, it will promote labor saving. The very idea of “bringing
jobs back” is an empty campaign slogan. While the historical
experience has been that replacing people by machines did not
reduce total employment much or at all, because new jobs were
created in place of those lost, a careful study by the McKinsey
Global Institute (2017) concludes that about 60 percent of all
occupations currently have at least 30 percent of constituent
components that could be automated, and that the rate of
increase in employment will slow down. There are also more
apocalyptic claims that artificial intelligence is a revolutionary
innovation, replacing brains and not just muscles, and that the
rate of replacement will be much larger (see Brynjolfsson, Rock,
and Syverson 2017). Yet even if the emergence of new jobs

193
 the future?

compensates for the disappearance of traditional ones, the occu-

pations for which there will be greater demand will be those with
low productivity and low wages. The fastest-growing sector in
the United States is personal services, the sector with the lowest
pay. The US Bureau of Labor Statistics predicts that between 2016
and 2026 non-agricultural wage and salaried jobs will grow at
a rate of 0.7 percent per year, with manufacturing jobs declining
at an annual rate of 0.6 percent and health care and social
assistance jobs increasing at a rate of 1.9 per year. The average
annual wage outside agriculture in the United States in 2016 was
$49,630, while home health aides earned $23,600 and fast-food
cooks $20,570. Among the twenty-eight members of the
European Community the fastest-growing sectors between
2008 and 2016 were also the lowest paying ones, with the excep-
tion of professional, scientific, and technical workers. Hence,
there are reasons to expect that many people will experience
the necessity to move to lower-pay occupations, with the atten-
dant loss of social status and a perception of downward mobility.
The distributive effects of this scenario would be alle-
viated by redistributive income policies. It is intuitive to expect
that when inequality increases, so do political demands for
redistribution through taxes and transfers or social services:
this belief is the cornerstone of political economy. Yet empirical
evidence in favor of this theory is at best shaky: indeed, the
question “Why don’t the poor take it away from the rich?” is
a source of unceasing puzzlement, with divergent answers.1 The,

1
For overviews of this literature, see Putterman (1996), Roemer (1998),
Harms and Zink (2003), Lind (2005), Ansell and Samuels (2010),
Acemoglu et al. (2015).

194
 what can and cannot happen?

not so new, idea on the political horizon in several countries is

a universal minimum, “citizen,” income. We do need to treat
time free from labor not in terms of “unemployment” but as
liberation from unnecessary activities, some of which are highly
unpleasant. But even if universal income at some decent level is
fiscally feasible, it may not be sufficient to overcome ghettoiza-
tion: the vicious circle of residential segregation, bad schools,
unemployment, and crime. Once the ghettoes, “villes,” or “bar-
rios” are formed, all policies seem impotent in breaking them.
The Right does not know what to do and neither does the Left.
Hence, in this scenario, not only inequality but social segrega-
tion are here to stay, and perhaps increase.
As we can see, the range of possible economic futures
is wide, from the possibility that the crisis would just con-
veniently disappear to a scenario in which it would become
even more profound, without redistribution via the political
system. As my description of these possibilities may reveal,
I am leaning toward the pessimistic scenario, but perhaps
current events in the United States, where the government is
embarking on a major program to increase inequality and
reduce income protection, unduly colors my more general
views. Certainly several European societies are more averse to
inequality and several European governments are more
attuned to its dangers. I must abdicate to the reader the task
of deciding where in this range of possibilities the future lies.
In the political realm, both the left- and the right-
wing insurgent parties are “populist.” They claim that tradi-
tional representative institutions serve the interests of elites
and do not provide sufficient voice to “the people.” Even
though the very word “populism” emerged only at the end

195
 the future?

of the nineteenth century, such claims are as old as represen-

tative institutions. Already, for Anti-Federalists “political
aristocracy” was as much of a danger as social aristocracy.
If the rulers were other than the ruled, they feared,
“Corruption and tyranny would be rampant as they have
always been when those who exercised power felt little con-
nection with the people. This would be true, moreover, for
elected representatives, as well as for kings and nobles and
bishops” (in Ketcham 1986: 18). Hence, they were preoccupied
with the duration of terms, as short as six months in New
Jersey at one time, term limits, restrictions on representatives
to determine their own salaries, and revoking and censuring
procedures – almost the same measures as those proposed by
President Macron in France.
Representative democracy, the political system ush-
ered in by the American and French revolutions and gradu-
ally adopted around the world, recurrently confronts
widespread and intense dissatisfaction. Some of this dissatis-
faction stems from the intrinsic features of any system in
which people decide as a collectivity who will govern them –
the limitations of representative democracy that are inevitable
yet disagreeable. But some originate from pathologies of the
particular systems of representative institutions.
One widespread misunderstanding of the way democ-
racy works is that elections do not offer a choice. The parties
offer “tweedledum and tweedledee,” “bonnet blanc et blanc
bonnet.” The Cohn-Bendit brothers (1968) saw elections as
a choice between “gin and tonic and tonic and gin,” the journal-
ist Friedman (2001) as between “Pepsi or Coke.” Clearly, people
can choose only between or among the proposals offered by

196
 what can and cannot happen?

parties; not every conceivable political program is represented in

electoral competition. The alternatives that are subject to elec-
toral choice are particularly restricted in systems in which form-
ing new parties is next to impossible, as in the United States, or
where the two major parties form grand coalitions. Yet the very
fact that individuals are offered little choice on election day does
not mean that the people as a collectivity does not choose. What
parties propose in elections is what they think is most likely to
make them win, and what is most likely to make them win is
what most people want. Hence, if parties know exactly what
people want, they offer the same platforms and individuals have
no choice at the polls; electoral platforms diverge only to the
extent that parties are uncertain about individual preferences.
Yet had the majority or plurality wanted something different,
parties would offer something different; platforms would still
converge to the center of public opinion but would be different.
Hence, people as a collectivity choose even if individuals have
little choice when they cast their votes.
Another source of dissatisfaction is that in elections
no single individual decides anything, creating a feeling of
inefficacy. When individuals make private choices, they cause
outcomes. Yet from an individual point of view, the outcome
of an election is independent of one’s action. No one can say
“I voted for A, therefore A will win”; the most each of us can
do is to cast our ballot, go home, and impatiently wait in front
of a television set to see how others have voted. When collec-
tive decisions are made using a simple majority rule by many
individuals endowed with an equal influence over the out-
come, no individual has a causal effect on the collective
decision. The value of elections is not that each voter has

197
 the future?

real influence on the final result, but that collective choice is

made by summing individual wills. Yet even if people appreci-
ate elections as a mechanism of collective decision making,
they still feel politically impotent as individuals.
At the root of the periodic flares of dissatisfaction
with representative institutions lies something even more
profound. Democracy is a system in which the people decide
as a collectivity who will rule them, at least over some period
of time. But even if rulers are selected through elections, we
are ruled, which means that we are sometimes forbidden to do
what some of us would want to do and ordered to do what
some of us would not want to do. The ideal that justified the
founding of modern representative institutions was “self-
government of the people.” The problem to be solved, as
posed by Rousseau (1964: 182), was to “find a form of associa-
tion which defends and protects with all the shared force the
person and the goods of each associate, and through which
each, uniting with all, still obeys but himself, remaining as free
as before.” But this problem would have a solution only if
everyone wanted the same; only then would obeying others be
the same as obeying oneself. In a society with conflicting
interests and heterogeneous values, being ruled means having
to yield to the will of others, against one’s own.
Democratically elected governments can take money away
from some and give it to others, force parents to inoculate
their children, keep people in prisons, and in some barbaric
countries even kill them. No wonder no one likes being ruled,
even if to live together in peace ruled we must be.
These sources of dissatisfaction with representative
democracy are just due to the inherent limits imposed on

198
 what can and cannot happen?

individuals by the requirements of living together in peace.

Democracy may still be, and I believe it is (see Przeworski
2010), the least bad way of organizing our lives as
a collectivity, but any political arrangements face limits as to
what they can achieve. It is only natural that this latent dis-
satisfaction flares up when democracy does not deliver what
people most care about, whether material security, public
order, or the realization of cultural values and norms.
Hence, there are reasons to expect that postures toward
democracy would be outcome-contingent, that democracy
may experience crises.
These general features of democracy do not explain,
however, the current popularity of “anti-elite,” “anti-
establishment,” “anti-system,” populist rhetoric. The slogan
of insurgent parties everywhere echoes the cry of Argentine
streets during the crisis of 2001: “Everyone Out” (“Fuera
Todos”). A cynical interpretation would be that it is just an
instrument of outsiders to elbow their way into office by
replacing all traditional parties. But is it not true that our
representative institutions, as they were designed, favor the
interests of elites? There is something inconsistent in
bemoaning persistent inequality and at the same time com-
plaining about the populist critique of representative
institutions.
Our systems of representative government were born
from a fear of participation by the broad masses of the popu-
lation, a large part of whom were poor and illiterate. One
would not err much in thinking that the strategic problem of
the “founders,” pretty much everywhere, was how to con-
struct representative government for the elites while

199
 the future?

protecting it from the poor. While governments were to be

selected by elections, their role was to ratify the superiority of
those entitled to govern by their social and economic position.
Created under a shadow of religious and economic conflicts,
representative institutions were designed to bar, or at least
minimize, the voice of the people between elections, treating
all “intermediate organizations” – clubs, associations, trade
unions, as well as political parties – as a danger to civil peace.
Intended as a bulwark against despotism, they were designed
to disable governments from doing much of anything, bad or
good, by checking and balancing powers, and by protecting
the status quo from the will of the majority. The poor were
instructed that their interests would be represented by the
wealthy, women that their interests would be guarded by men,
the “uncivilized” that they needed to be guided by their
colonizers. When the fear for property took hold, self-
government, equality, and liberty were dressed up in elaborate
intellectual constructions to make them compatible with the
rule by a few. The people cannot be trusted because it can
“err”: James Madison said it, Simón Bolivar said it, and so did
Henry Kissinger, when he declared that President Allende was
elected “due to the irresponsibility of the Chilean people.”
The particular forms of our representative institu-
tions were designed to protect the status quo – whatever it
was, but centrally the property relations against temporary
majorities. Bicameralism and presidential veto power meant
that the status quo could be altered only by super-majorities.
Restrictions of franchise, open voting, and indirect elections
protected the political influence of elites. These trenches pro-
tecting property were gradually removed: suffrage became

200
 what can and cannot happen?

universal, ballots secret, elections direct, legislatures more

frequently unicameral. Yet they were replaced by new anti-
majoritarian mechanisms: judicial review (Ginsburg and
Versteeg 2012), delegation of monetary policy to non-elected
central banks (Cukierman, Edwards, and Tabellini 1992), and
independent regulatory bodies. The 1992 Maastricht Treaty,
which restricted annual deficits to no more than 3 percent of
GDP, deprived European governments of the possibility to
pursue anti-cyclical economic policies and put bounds on
social expenditures.
Even without these institutional trenches, elections
are inherently an elitist, Manin (1997) would say “aristo-
cratic,” mechanism. Voters recognize that not everyone is
equally equipped to govern, and they elect people whom
they see as in some ways distinct. Voters are free to take any
qualities they want as signs of this capacity, but most want to
vote for people different from themselves. The result is that in
no country does the composition of the elected bodies even
vaguely resemble the composition of the electorate.
The United States Senate is a “millionaires club.” Perhaps
most ironically the 2017 French legislative elections, won by
a party campaigning on anti-elite slogans, generated
a parliament even more elitist in terms of education and
income than the outgoing one.
When the Italian political philosopher, Norberto
Bobbio (1987), analyzed the differences between democracies
and dictatorships, all he could come up with was the distinc-
tion between systems in which “elites propose themselves and
elites impose themselves.” But people have no power in
a system ruled by elites. No wonder then that calls for

201
 the future?

institutional reforms that would make “the voice of the

people” louder, and some measures toward “direct democ-
racy,” dominate the populist institutional agenda. Some of
these proposals return to the Anti-Federalist demands men-
tioned above: short terms, term limits, revocation of man-
dates, reduction of pay of legislators, and limitations on
circulating between the public and private sectors. In the
United States, the obvious measures would be direct elec-
tions of the president and the delegation of electoral district-
ing from state legislatures to independent bodies. In Europe,
such proposals range from the inane, such as the “survey
democracy” advocated by the Five Stars (Cinque Strelle)
party in Italy, to increased reliance on popular initiative
referendums, to convocations of randomly selected “para-
legislatures” (bodies of randomly selected citizens that con-
sider particular legislative proposals without having the
authority to adopt laws). Particularly interesting is
a proposal that surged during the last French elections,
according to which voters would be able to cast votes for
“none of the above” (vote en blanc), and if such votes would
win a plurality, another election would be called, with none
of the previous candidates eligible to present themselves.
One wonders what this mechanism would have produced
in the US 2016 presidential elections: most likely, neither
Trump nor Clinton.
Yet, as justified as the populist dissatisfaction with the
extant representative institutions may be, all such measures
are no more than palliatives. They may temporarily restore
some confidence in democratic institutions, but they hurl
themselves against the inescapable: the mere fact that each

202
 what can and cannot happen?

of us must be ruled by someone else and being ruled must

entail policies and laws we do not like. There are gradations –
some institutional frameworks induce better representation
than others – but in the end, as J.S. Mill (1991) observed, it is
not possible for everyone to rule at the same time. Hence, even
if some institutional reforms emerge from the current crisis,
I fear that they will not change much.
“Europe” is a topic apart. Both the European Union
and the Eurozone offer attractive targets for populists. For
one, there is no subject, no “the people” in singular, they can
possibly represent. Moreover, they are even more distant
from the people in plural than their respective governments.
The criticism that they are ruled by foreign, often read as
“German,” elites is plausible. Hence, the calls for isolation and
protectionism are appealing.
I am more sanguine about the electoral menace of the
radical Right. While several traditional parties are already
accommodating the anti-immigration sentiments, electoral
victories of the radical Right are not on the horizon in most
European countries. The radical Right seems to have a hard
core of supporters of about one-fourth of the potential voters
in most developed democracies. Trump won only because he
was able to take over a traditional party, and many people
voted for him because they hated the Clintons, not because of
his personality or program. My guess is that the genie is
already out of the bottle and that it grew as much as it possibly
could.
Yet I doubt that policies directed against “immigra-
tion,” either by radical Right parties in office or by centrist
parties responding to its menace, will appease anyone. While

203
 the future?

the Right adopts the language of “national sovereignty” and

campaigns for measures to control the current flow of people
across borders, such measures will have no effect on the
ethnic, cultural, and religious conflicts that tear societies
apart. The sources of these conflicts are not at the borders;
they are deeply embedded in the social fabric. Moreover,
polarization has now reached the basic unit of social structure:
the family. In 1960, 5 percent of Republican sympathizers and
4 percent of Democratic ones would have been displeased if
their offspring were to marry across party lines, while in 2010
these percentages were forty-nine for Republicans and thirty-
three for Democrats. Political polarization has deep roots and
it will not go away with contingent political events.
While these are just speculations, suppose that noth-
ing much will change in the foreseeable future: growth will be
slow, inequality and segregation will persist, good jobs will
continue dwindling, and traditional parties will accommodate
anti-immigration sentiments while trying to cope with
inequality and segregation using the same repertoire of poli-
cies. Does such a scenario threaten democracy?
The danger is that democracy would gradually and
surreptitiously deteriorate. This is the danger that incumbents
may intimidate the hostile media and create a propaganda
machine of their own; that they would politicize security agen-
cies, harass political opponents, use state power to reward
sympathetic private firms, selectively enforce laws, provoke
foreign conflicts to monger fear, and rig elections. The danger
in the countries where the radical Right does not accede to
office is that governments might go too far in accommodating
nativist and racist demands, and restrict civil liberties without

204
 what can and cannot happen?

improving the material conditions of the people most dissatis-

fied with the status quo.
Hence, we should not be desperate but also not san-
guine. Something profound is going on. Perhaps the best
diagnosis of the current situation in many democracies is
“intense partisanship with weak parties” (Azari 2016).
Democratic elections peacefully process conflicts only when
political parties are successful in structuring conflicts and
channeling political actions into elections. Representative
institutions absorb conflicts only if everyone has the right to
participate within these institutions, if conflicts are structured
by political parties, if parties have the capacity to control their
supporters, and if these organizations have the incentives to
pursue their interests through the representative system.
My fear is that neither the government of Trump, nor
Brexit, nor the governments that will be elected on the
European continent will improve the everyday lives of most
people, which will only strengthen “anti-establishment” or
“anti-system” sentiments. In a typical election about one in
two voters ends up on the losing side. In presidential systems
the winner rarely receives much more than 50 percent of the
vote, and in parliamentary multi-party systems the largest
share is rarely higher than 40 percent. Moreover, many people
who voted for the winners are dismayed with their perfor-
mance in office. So most of us are left disappointed, either
with the outcome or with the performance of the winner. Still,
election after election, most of us hope that our favorite
candidate will win the next time around and will not disap-
point. It is, thus, only natural that when people participate in
successive elections, see governments change, and discover

205
 the future?

that their lives remain the same, they find something wrong
with “the system” or “the establishment.”
As a Polish adage has it, “A pessimist is but an
informed optimist.” I am moderately pessimistic about the
future. I do not think that the very survival of democracy is at
stake in most countries, but I do not see what would get us out
of the current discontent. It will not be alleviated by contin-
gent political events, the results of future elections. This crisis
is not just political; it has deep roots in the economy and in
society. This is what I find ominous.

206
 references

Acemoglu, Daron, and James Robinson. 2000. “Why Did the West
Extend the Franchise? Democracy, Inequality, and Growth in
Historical Perspective.” Quarterly Journal of Economics 115:
1167–99.
Acemoglu, Daron, Suresh Naidu, Pascual Restrepo, and James
A. Robinson. 2015. “Democracy, Redistribution, and Inequality.”
Handbook of Income Distribution, Vol. 2. Amsterdam: Elsevier.
Pages 1886–966.
Acemoglu, Daron, David Autor, David Dorn, Gordon H. Hanson,
and Brendan Price. 2016. “Import Competition and the Great
United States Employment Sag of the 2000s.”Journal of Labor
Economics 34(S1): S141–S198.
Acherbach, Joel, and Scott Clement. 2016. “America Really is More
Divided Than Ever.” Washington Post, July 16.
Ageron, Charles-Robert. 1976. “L’opinion française devant la guerre
d’Algérie.” Revue française d’outre mer 231: 256–85.
Akerlof, George A. 1991. “Procrastination and obedience.” American
Economic Review 81: 1–19.
Albright, Jeremy. 2010. “The Multidimensional Nature of Party
Competition.” Party Politics 16: 699–719.
Alon, Amos. 2002. The Pity of It All: Portrait of the German-Jewish
Epoch 1743–1933. New York: Picador.
Altamirano, Carlos. 1979. Chili: les raisons d’une défaite. Paris:
Flammarion.

207
 references

Alvarez, Michael, José Antonio Cheibub, Fernando Limongi, and

Adam Przeworski. 1996. “Classifying Political Regimes.” (ACLP)
Studies in International Political Development 31: 3–36.
Andrews, Rhys, Sebastian Jilke, and Steven Van de Walle. 2014.
“Economic Strain and Perceptions of Social Cohesion in
Europe: Does Institutional Trust Matter?”European Journal of
Political Research 53: 559–79.
Ansell, Christopher, and David Samuels. 2010. “Democracy and
Redistribution, 1880–1930: Reassessing the Evidence.” Paper pre-
sented at the Annual Meeting of the American Political Science
Association, Washington, DC.
Armingeon, Klaus, and Kai Guthmann. 2014. “Democracy in Crisis?
The Declining Support for National Democracy in European
Countries, 2007–2011.” European Journal of Political Research 53:
424–42.
Armingeon, Klaus, Christian Isler, Laura Knopfel, David Weisstanner,
and Sarah Engler. 2016 Comparative Political Data Set. 1960–2014.
www.cpds-data.org.
Arzheimer, Kai. 2013. “Working Class Parties 2.0? Competition
between Centre Left and Extreme Right Parties.” In Jens Rydren
(ed.), Class Politics and the Radical Right. London, New York:
Routledge. Pages 75–90.
Atkinson, Anthony B.,Thomas Piketty, and Emmanuel Saez. 2011.
“Top Incomes in the Long Run of History.” Journal of Economic
Literature 49: 3–71.
Austen-Smith, David. 1992. “Strategic Models of Talk in Political
Decision Making.” International Political Science Review 13:
45–58.
Autor, David, David Dorn, and Gordon H. Hanson. 2013.
“The China Syndrome: Local Labor Market Effects of Import
Competition in the United States.” American Economic Review
103: 2121–68.

208
 references

Autor, David, David Dorn, Gordon H. Hanson, and Kaveh Majlesi.

2017. “A Note on the Effect of Rising Trade Exposure on the 2016
Presidential Election.” Appendix to Autor et al. “Importing
Political Polarization? The Electoral Consequences of Rising
Trade Exposure.” Working paper.
Ayta, Selim Erdem, Eli Rau, and Susan Stokes. 2017.
“Unemployment and Turnout.” Working paper, Department of
Politics, Yale University.
Azari, Julia. 2016. “Weak Parties with Strong Partisanship Are a Bad
Combination.” www.vox.com, November 3.
Banks, Arthur S. 1996. Cross-National Time-Series Data Archive.
Databanks International.
Bermeo, Nancy. 2016. “On Democratic Backsliding.” Journal of
Democracy 27(1): 5–19.
Bitar, Sergio, and Crisostomo Pizarro. No date. La Caida de Allende
y la Huelga de el Teniente. Santiago: Las Ediciones del Ornitorrinco.
Bobbio, Norberto. 1987. Democracy and Dictatorship. Minneapolis:
University of Minnesota Press.
Boix, Carles, Michael Miller, and Sebastian Rosato. 2012.
“A Complete Data Set of Political Regimes, 1800–2007.”
Comparative Political Studies 46: 1523–54.
Bon, Frederic. 1978. Les élections en France. Paris: Seuil.
Bracher, Karl Dietrich. 1966. “The Technique of the National
Socialist Seizure of Power.” In Fritz Stern (ed.), The Path to
Dictatorship 1918–1933: Ten Essays by German Scholars. Garden
City, NY: Anchor Books. Pages 113–32.
Brady, David, John Ferejohn, and Aldo Paparo. 2017. “Immigration
and Politics: A Seven Nation Study.” Working paper.
Bright Line Watch. 2018. “Wave 5.” April. http://brightlinewatch
.org/wave5/.
Bruno, Michael, and Jeffrey Sachs. 1985. Economics of Worldwide
Stagflation. Cambridge, MA: Harvard University Press.

209
 references

Brynjolfsson, Erik, Daniel Rock, and Chad Syverson. 2017.

“Artificial Intelligence and the Modern Productivity Paradox:
A Clash of Expectations and Statistics.” Working paper.
Buchanan, James M., and Gordon Tullock. 1962. The Calculus of
Consent: Logical Foundations of Constitutional Democracy. Ann
Arbor: University of Michigan Press.
Calvert, Randall. 1994. “Rational Actors, Equilibrium, and Social
Institutions.” In J. Knight and I. Sened (eds.), Explaining
Social Institutions. Ann Arbor: University of Michigan Press.
Canovan, Margaret. 2002 “Taking Politics to the People: Populism
as the Ideology of Democracy.” In Yves Meny and Yves Surel
(eds.), Democracies and the Populist Challenge. New York:
Palgrave. Pages 25–44.
Capoccia, Giovanni. 2005. Defending Democracy: Reactions of
Extremism in Interwar Europe. Baltimore: Johns Hopkins
University Press.
Carr, William. 1969. The History of Germany 1815–1945. New York:
St. Martin’s Press.
Case, Anne, and Angus Deaton. 2017. “Mortality and Morbidity in
the 21st Century.” Brookings Papers on Economic Activity. BPEA
Conference Drafts, March 23–4.
Cassese, Sabino. 2011. Lo Stato fascista. Milano: il Mulino.
Cautres, Bruno. 2018. “Le clivage gauche-droite dans les démocraties
modernes.” Cahiers français 404 (May–June): 52–61.
Cheibub, Jose Antonio. 2007. Presidentialism, Parliamentarism, and
Democracy. New York: Cambridge University Press.
Cheibub, Jose Antonio, Jennifer Gandhi, and James
Raymond Vreeland. 2010. “Democracy and Dictatorship
Revisited.” Public Choice 143: 67–101.
Chen, M. Keith, and Ryne Rohla. 2018. “The Effect of Partisanship
and Political Advertising on Close Family Ties.” Science 360
(June 1): 1020–4.

210
 references

Chetty, Raj, David Grusky, Maximilian Hell, Nathaniel Hendren,

Robert Manduca, and Jimmy Narang. 2016. “The Fading
American Dream: Trends in Absolute Income Mobility Since
1940.” Working Paper 22910. www.nber.org/papers/w22910.
Chiaramonte, Alessandro, and Vincenzo Emanuele. 2017. “Party
System Volatility, Regeneration and De-institutionalization in
Western Europe (1945–2015).” Party Politics 23: 376–88.
Cohn-Bendit, Daniel, and Gabriel Cohn-Bendit, 1968. Obsolete
Communism: The Left-Wing Alternative. New York: McGraw-
Hill.
Colantone, Italo, and Piero Stanig. 2017. “The Trade Origins of
Economic Nationalism: Import Competition and Voting
Behavior in Western Europe.” Balfi Carefin Centre Research
Paper No. 2017–49. Università Commerciale Luigi Bocconi,
Milan.
Cornell, Agnes, Jørgen Møller, and Svend-Erik Skaaning. 2017.
“The Real Lessons of the Interwar Years.” Journal of Democracy
28: 14–28.
Corvalan, Luis L. 2003. El Gobierno de Salvador Allende. Santiago:
LOM Ediciones.
Coser, Lewis A. 1964. The Functions of Social Conflict:
An Examination of the Concept of Social Conflict and Its Use in
Empirical Sociological Research. New York: Free Press.
Cukierman, Alex, Sebastian Edwards, and Guido Tabellini. 1992.
“Seigniorage and Political Instability.” American Economic
Review 82: 537–55.
Dahl, Robert A. 1971. Polyarchy: Participation and Opposition. New
Haven: Yale University Press.
Dancygier, Rafaela M. 2010. Immigration and Conflict in Europe.
New York: Cambridge University Press.
Dancygier, Rafaela M. 2017. Dilemmas of Inclusion: Muslims in
European Politics. Princeton: Princeton University Press.

211
 references

Dancygier, Rafaela M., and David D. Laitin. 2014. “Immigration into

Europe: Economic Discrimination, Violence, and Public Policy.”
Annual Review of Political Science 17: 43–64.
Delmer, Sefton. 1972. Weimar Germany: Democracy on Trial.
London: Macdonald.
Denquin, Jean-Marie. 1988. 1958: La Genèse de la Ve République.
Paris: Presses Universtaires de France.
De Vylder, Stefan. 1974. Allende’s Chile: The Political Economy of the
Rise and Fall of the Unidad Popular. Cambridge: Cambridge
University Press.
Diamond, Larry. 2002. “Thinking about Hybrid Regimes.” Journal
of Democracy 13(2): 21–35.
Dimsdale, Nicholas, Nicholas Horsewood, and Allard Van Riel.
2004. “Unemployment and Real Wages in Weimar Germany.”
University of Oxford Discussion Papers in Economic and Social
History, Number 56.
Dixit, Avinash, Gene M. Grossman, and Faruk Gul. 2000.
“The Dynamics of Political Compromise.” Journal of Political
Economy 108: 531–68.
Downs, Anthony. 1957. An Economic Theory of Democracy.
New York: Harper and Row.
Droz, Bernard, and Evelyne Lever. 1991. Histoire de la Guerre
d’Algérie. Paris: Éditions du Seuil.
Dunn, John. 2000. The Cunning of Unreason. Cambridge:
Cambridge University Press.
Dunn, Susan. 2004. Jefferson’s Second Revolution: the Election Crisis
of 1800 and the Triumph of Republicanism. Boston: Houghton
Mifflin.
Elias, Anwen, Edina Szocsik, and Christina Isabel Zuber. 2015.
“Position, Selective Emphasis and Framing: How Parties Deal
with a Second Dimension in Competition.” Party Politics 21:
839–50.

212
 references

Elster, Jon. 1998. “Deliberation and Constitution Making.”

In Jon Elster (ed.), Deliberative Democracy. Cambridge:
Cambridge University Press. Pages 97–122.
Ermakoff, Ivan. 2008. Ruling Onself Out: A Theory of Collective
Abdications. Durham, NC: Duke University Press.
Esteban, Joan-Maria, and Debraj Ray. 1994. “On the Measurement
of Polarization.” Econometrica 62: 819–51.
Evans, Richard J. 2003. The Coming of the Third Reich. Penguin
Publishing Group. Kindle Edition.
Fearon, James. 2011. “Self-enforcing Democracy.” Quarterly Journal
of Economics 126: 1661–708.
Ferejohn, John, and Pasquale Pasquino. 2003. “Rule of Democracy
and Rule of Law.” In Jose Maria Maravall and Adam Przeworski
(eds.), Democracy and the Rule of Law. New York: Cambridge
University Press. Pages 242–60.
Flechtheim, Ossip K. 1966. “The Role of the Communist Party.”
In Fritz Stern (ed.), The Path to Dictatorship 1918–1933: Ten Essays
by German Scholars. Garden City, NY: Anchor Books. Pages 89–112.
Foa, Roberto Stefan, and Yascha Mounk. 2016. “The Democratic
Disconnect.” Journal of Democracy 27(6): 5–17.
Fossati, Diego. 2014. “Economic Vulnerability and Economic
Voting in 14 OECD Countries.” European Journal of Political
Research 53: 116–35.
Foucault, Martial. 2018. “Les transformations de la sociologie du
vote.” Cahiers français 404 (May–June): 42–51.
Franceinfo. 2017. “En 2016, les actes racistes, antisémites et antimusul-
mans ont baissé en France, mais pas les actes antichrétiens.”
February 2. www.francetvinfo.fr/societe/religion/en-2016-les-actes-
racistes-antisemites-et-antimusulmans-ont-baisse-en-france-mais-
pas-les-actes-antichretiens_2044983.html.
Friedmann, T. 2001. The Lexus and the Olive Tree: Understanding
Globalization. New York: Anchor Books.

213
 references

Gargarella, Roberto. 2003. “The Majoritarian Reading of the Rule of

Law.” In Jose Maria Maravall and Adam Przeworski (eds.),
Democracy and the Rule of Law. New York: Cambridge
University Press. Pages 147–67.
Ginsburg, Tom, and Mila Versteeg. 2012. “The Global Spread of
Constitutional Review: An Empirical Analysis.” Working paper,
University of Chicago Law School.
Ginsburg, Tom, and Aziz Z. Huq. 2018a. “How to Lose
a Constitutional Democracy.”UCLA Law Review 65(1): 78–169.
Ginsburg, Tom and Aziz Z. Huq. 2018b. How to Save a Constitutional
Democracy. Chicago: University of Chicago Press.
Golder, Matt. 2016. “Far Right Parties in Europe.” Annual Review of
Political Science 19: 477–97.
Goldsmith, Jack. 2018. “Paradoxes of the Deep State.” In Cass
R. Sunstein (ed.), Can It Happen Here? New York:
HarperCollins. Pages 105–34.
Graham, Benjamin A.T., Michael K. Miller, and Kaare Strøm. 2017.
“Safeguarding Democracy: Powersharing and Democratic
Survival.” American Political Science Review 111: 686–704.
Graf, Rudiger, and Konrad H. Jarausch. 2017. “‘Crisis’ in Contemporary
History and Historiography.” www.docupedia.de/zg/
Graf_jarausch_crisis_en_2017.
Graham, Matthew, and Milan W. Svolik. 2018. “Democracy in
America? Partisanship, Polarization, and the Robustness of
Support for Democracy in the United States.” Unpublished
manuscript, Yale University.
Gramsci, Antonio. 1971 (c.1930). Prison Notebooks. New York:
International Publishers.
Guiso, Luigi, Helios Herrera, Massimo Morelli, and
Tommaso Sonno. 2017. “Demand and Supply of Populism.”
Discussion Paper 11871. Center for Economic Policy Research,
London.

214
 references

Habermas, Jurgen. 1973. Legitimation Crisis. Boston: Beacon Press.

Haffner, Sebastian. 2002. Histoire d’un Allemand. Paris: Babel.
Hainmueller, Jens, and Daniel J. Hopkins. 2014. “Public Attitudes
Toward Immigration.” Annual Review of Political Science 17:
225–49.
Harms, Philipp, and Stefan Zink. 2003. “Limits to Redistribution in
a Democracy: A Survey.” European Journal of Political Economy
19: 651–68.
Hastings, Michel. 2018. “Le clivage gauche-droite: disparition ou
renouvellement?” Cahiers français 404 (May–June): 34–41.
Helpman, Elhanan. 2016. “Globalization and Wage Inequality.”
Working paper, Department of Economics, Harvard
University.
Hofstadter, Richard. 1969. The Idea of a Party System: The Rise of
Legitimate Opposition in the United States, 1780–1840. Berkeley:
University of California Press.
Huber, John, and Ronald Inglehart. 1995. “Expert Interpretations of
Party Space and Party Locations in 42 Societies.” Party Politics 1:
73–111.
Human Rights Brief. 2017. “Xenophobic and Racist Hate
Crimes Surge in the European Union.” February 28. http://
hrbrief.org/2017/02/xenophobic-racist-hate-crimes-surge-eur
opean-union/.
Hutchison, Elizabeth Quay, Thomas Miller Klubock, and Nara
B. Milanich (eds.). 2013. “The Chilean Road to Socialism:
Reform and Revolution.” Chile Reader: History, Culture,
Politics. Durham, NC: Duke University Press. Pages 343–428.
Ignazi, Piero. 1992. “The Silent Counter-Revolution: Hypotheses on
the Emergence of Extreme Right-Wing Parties in Europe.”
European Journal of Political Research 22: 3–34.
Ignazi, Piero. 2003. Extreme Right Parties in Western Europe.
Oxford: Oxford University Press.

215
 references

Inglehart Ronald F., and Flanagan S.C. 1987. “Value Change in

Industrial Societies.” American Political Science Review 81:
1289–319.
Inglehart, Ronald F., and Pippa Norris. 2016. “Trump, Brexit, and
the Rise of Populism: Economic Have-Nots and Cultural
Backlash.” Faculty Research Working Paper Series, Harvard
Kennedy School.
Ivarsflaten, Elisabeth. 2008. “What Unites Right-Wing Populists in
Western Europe? Re-examining Grievance Mobilization Models
in Seven Successful Cases.” Comparative Political Studies 41: 3–23.
Jung, Florian. 2011. “Income Inequality, Economic Development,
and Political Institutions.” Doctoral dissertation no. 3196 pre-
sented at the University of St. Gallen.
Kalecki, Michal. 1972 [1932]. “Czy mozliwe jest ‘kapitalistyczne’
wyjście z kryzysu?” In Kapitalizm, koniunktura i zatrudnienie.
Warsaw: Państwowe Wydawnictwo Ekonomiczne. Pages 75–81.
Kantar Public. 2018. “Ocena dzialalnosci rzadu, premiera i prezydenta.”
Online K002/18, January.
Karl, Terry Lynn. 1995. “The Hybrid Regimes of Central
America.”Journal of Democracy 6(3): 72–87.
Kates, Sean, and Joshua Tucker. 2017. “We Never Change, Do We?
Economic Anxiety and the Far Right in a Post Crisis Europe.”
Working paper, Department of Politics, New York University.
Ketcham, Ralph (ed.). 1986. The Anti-Federalist Papers and the
Constitutional Convention Debates. New York: Mentor Books.
King, Gary, and Langche Zheng. 2007. “When Can History Be Our
Guide? The Pitfalls of Counterfactual Inference.” International
Studies Quarterly 51: 183–210.
King, Gary, Ori Rosen, Martin Tanner, and Alexander F. Wagner.
2008. “Ordinary Economic Voting Behavior in the
Extraordinary Election of Adolf Hitler.” Journal of Economic
History 68: 951–96.

216
 references

Kitschelt, Herbert. 1994. The Transformation of European Social

Democracy. Cambridge: Cambridge University Press.
Kriesi, Hanspeter, Edgar Grande, Martin Dolezal, Simon Bornschier,
and Timotheos Frey. 2006. “Globalization and the Transformation
of the National Political Space: Six European Countries Compared.”
European Journal of Political Research 45: 921–56.
Kriesi, Hanspeter et al. 2012. Political Conflict in Western Europe.
Cambridge: Cambridge University Press.
Lambrecht Plaza, Karen. 2011. “La Distribución del Ingreso en Chile:
1960–2000.” Facultad de Economia y Negocios, Universidad de
Chile, Santiago.
Landau, David. 2013. “Abusive Constitutionalism.” University of
California at Davis Law Review 47: 189–260.
Landsberger, Henry A., and Tim McDaniel. 1976. “Hypermobilization
in Chile, 1970–1973.” World Politics 28: 502–41.
Lee, Woojin, and John E. Roemer. 2006. “Race and
Redistribution in the United States: A Solution to the
Problem of American Exceptionalism.” Journal of Public
Economics 90: 1027–52.
Le Gac Julie, Anne-Laure Olivier, and Raphael Spina. 2015. La
France en chiffres, de 1870 à nos jours: Sous la direction d'Olivier
Wieviorka. Paris: Perrin.
Lenin, Vladimir I. 1959 [1919]. “Letter to the Workers of Europe and
America.” In Against Revisionism. Moscow: Foreign Languages
Publishing House. Pages 479–86.
Lepsius, Rainer M. 1978. “From Fragmented Party Democracy to
Government by Emergency Decree and Nationalist Socialist
Takeover: Germany.” In Juan J. Linz and Alfred Stepan (eds.),
The Breakdown of Democratic Regimes: Europe. Baltimore: Johns
Hopkins University Press. Pages 34–79.

217
 references

Levitsky, Steve, and Lucan A. Way. 2010. Competitive Authoritarianism:

Hybrid Regimes after the Cold War. New York: Cambridge
University Press.
Lewandowsky, Stephan, Ullrich K.H. Ecker, and John Cook. 2017.
“Beyond Misinformation: Understanding and Coping with the
‘Post-Truth’ Era.” Journal of Applied Research in Memory and
Cognition 6: 353–69.
Lind, Jo T. 2005. “Why Is there So Little Redistribution?” Nordic
Journal of Political Economy 31: 111–25.
Lindvall, Johannes. 2014. “The Electoral Consequences of Two Great
Crises.” European Journal of Political Research 53: 747–65.
Linz, Juan J. 1978. The Breakdown of Democratic Regimes: Crisis,
Breakdown, and Reequilibration. Baltimore: Johns Hopkins
University Press.
Lippmann, Walter. 1956. The Public Philosophy. New York: Mentor
Books.
Lipset, Seymour Martin. 1960. Political Man. Garden City, NY:
Doubleday.
Los mil dias de Allende. 1997. www.cepchile.cl.
Luo, Zhaotian, and Adam Przeworski. 2018. “Subversion by Stealth:
Dynamics of Democratic Backsliding.” Working paper,
Department of Politics, New York University.
Lust, Ellen, and David Waldner. 2015. “Unwelcome Change:
Understanding, Evaluating and Extending Theories of Democratic
Backsliding.” USAID. pdf.usaid.gov/pdf_docs/PBAAD635.pdf.
McCarty, Nolan, Keith Poole, and Howard Rosenthal. 2016.
Polarized America: The Dance of Ideology and Unequal Riches.
Cambridge, MA: MIT Press.
McGann, Anthony. 2006. The Logic of Democracy: Reconciling
Equality, Deliberation, and Minority Protection. Ann Arbor:
University of Michigan Press.

218
 references

McKinsey Global Institute. 2017. “A Future That Works:

Automation, Employment, and Productivity.”
Macaulay, Thomas B. 1900. Complete Writings, Vol. 17. Boston and
New York: Houghton-Mifflin.
Maddison, Agnus. 2011. Statistics on World Population, GDP and Per
Capita GDP, 1–2008 AD. www.ggdc.net/maddison/oriindex.htm.
Maeda, Ko. 2010. “Two Modes of Democratic Breakdown:
A Competing Risks Analysis of Democratic Durability.”Journal
of Politics 72(4): 1129–43.
Magaloni, Beatriz. 2017. “Thought Piece on How Democracies Fail.”
Paper presented at the How Do Democracies Fall Apart (And
Could it Happen Here)? conference, Yale University, October 6.
Manin, Bernard. 1997. The Principles of Representative Government.
Cambridge: Cambridge University Press.
Manin, Bernard. 2017. “Les habits neufs de la représentation.” Esprit
437: 71–85.
Maravall, Jose Maria. 2016. Demands on Democracy. Oxford: Oxford
University Press.
Maravall, Jose Maria, and Adam Przeworski. 2001. “Political
Reactions to the Economy: The Spanish Experience.” In Susan
C. Stokes (ed.), Public Support for Economic Reforms in New
Democracies. New York: Cambridge University Press.
Margalit, Yotam M. 2013. “Explaining Social Policy Preferences:
Evidence from the Great Recession.” The American Political
Science Review 107: 80–103.
Marks G, L. Hooghe, M. Nelson, et al. 2006. “Party Competition and
European Integration in the East and West: Different Structure,
Same Causality.” Comparative Political Studies 39: 158–9.
Martner, Gonzalo. 1988. El gobierno del presidente Salvador Allende
1970–1973. Santiago: Editorial LAR.
Marx, Karl. 1934 [1852]. The Eighteenth Brumaire of Louis Bonaparte.
Moscow: Progress Publishers.

219
 references

Marx, Karl. 1952 [1851]. Class Struggles in France, 1848 to 1850.

Moscow: Progress Publishers.
Marx, Karl. 1971. Writings on the Paris Commune. Edited by
H. Draper. New York: International Publishers.
Marx, Karl. 1979 [1859]. A Contribution to the Critique of Political
Economy. New York: International Publishers.
Medina, Lucia. 2015. “Partisan Supply and Voters’ Positioning on
the Left–Right Scale in Europe.” Party Politics 21: 775–90.
Meeuwis, Maarten, Jonathan A. Parker, Antoinette Schoar, and Duncan
I. Simester. 2018. “Belief Disagreement and Portfolio Choice.” NBER
Working Paper 25108. www.nber.org/papers/w25108.
Miao, Ouyang. 2016. “The Pro-Competitive Effect of Chinese
Imports: Amplification through the Input–Output Network.”
Job Market Paper, Department of Economics, Brandeis
University.
Michaels, Walter Benn. 2007. The Trouble with Diversity: How
We Learned to Love Identity and Ignore Inequality. New York:
Henry Holt and Co.
Miliband, Ralph. 1975. Parliamentary Socialism: A Study in the
Politics of Labour. 2nd edn. London: Merlin Press.
Mill, John Stuart. 1977 [1859]. The Collected Works of John Stuart
Mill, Volume XVIII – Essays on Politics and Society Part I, edited
by J.M. Robson. Toronto: University of Toronto Press.
Mill, John Stuart. 1991 [1857]. Considerations on Representative
Government. Cambridge: Cambridge University Press.
Minkenberg, Michael. 2000. “The Renewal of the Radical Right:
Between Modernity and Anti-modernity.” Government and
Opposition 35: 170–88.
Minow, Martha. 2018. “Could Mass Detentions Without Process
Happen Here?” In Cass R. Sunstein (ed.), Can It Happen Here?
New York: HarperCollins. Pages 313–28.
Montesquieu. 1995 [1748]. De l’esprit des lois. Paris: Gallimard.

220
 references

Moral, Mert, and Roben E. Best. 2018. “On the Reciprocal

Relationship between Party Polarization and Citizen
Polarization.” Paper presented at the 2018 Annual Meeting of
the Midwest Political Science Association, Chicago.
Mudde, Cas. 2004. “The Populist Zeitgeist.” Government and
Opposition 39: 541–63.
Navia, Patricio, and Rodrigo Osorio. 2017. “‘Make the Economy
Scream’? Economic, Ideological and Social Determinants of
Support for Salvador Allende in Chile, 1970–3.” Journal of Latin
American Studies 49: 771–97.
Navia, Patricio, and Rodrigo Osorio. 2018. “Attitudes Toward
Democracy and Authoritarianism Before, During and After
Military Rule: The Case of Chile, 1972–2013.” Contemporary
Politics: 1–23.
NBC. 2017. “U.S. Hate Crimes Up 20 Percent in 2016, Fueled by Election
Campaign: Report.” March 14. www.nbcnews.com/news/us-news/u
-s-hate-crimes-20-percent-2016- fueled-election-campaign-n733306.
ODEPLAN. 1971. Plan de la Economia Nacional. Santiago: Oficina
de Planificacion Nacional.
O’Donnell, Guillermo. 1994. “Delegative Democracy.” Journal of
Democracy 5(1): 55–69.
Ostrogorskij, Michael. 1981 [1927]. Democracy and the Organization
of Political Parties. Piscataway, NJ: Transaction Publishers.
Pasquino, Gianfranco. 2008. “Populism and Democracy.” In
Daniele M. Albertazzi and Duncan McDonnell (eds.), Twenty-First
Century Populism. New York: Palgrave Macmillan. Pages 15–29.
Pew Research Center. 2015. “More Mexicans Leaving Than Coming
to the U.S.” November 19. www.pewhispanic.org/2015/11/19/more-
mexicans-leaving-than-coming-to-the-u-s/.
Pew Research Center. 2018. “Public Opinion on Abortion.”
October 15. www.pewforum.org/fact-sheet/public-opinion-on-
abortion/.

221
 references

Piketty, Thomas. 2018. “Brahmin Left vs Merchant Right: Rising

Inequality and the Changing Structure of Political Conflict
(Evidence from France, Britain and the US, 1948–2017).” WID.
World Working Paper Series No. 2018/7.
Pizzorno, Alessandro. 1964. “The Individualistic Mobilization of
Europe.” Daedalus (December): 199–224.
Popper, Karl. 1962. The Open Society and Its Enemies. London:
Routledge and Kegan Paul.
Prothro, James W., and Patricio E. Chaparro. 1976. “Public Opinion
and the Movement of the Chilean Government to the Left,
1952–73.” In Arturo Valenzuela and J. Samuel Valenzuela (eds.),
Chile: Politics and Society. New Brunswick, NJ: Transaction
Books. Pages 67–114.
Przeworski, Adam. 1986. Capitalism and Social Democracy.
New York: Cambridge University Press.
Przeworski, Adam. 2010. Democracy and the Limits of Self-
Government. New York: Cambridge University Press.
Przeworski, Adam. 2015. “Acquiring the Habit of Changing
Governments through Elections.” Comparative Political Studies
48: 1–29.
Przeworski, Adam, and Fernando Limongi. 1997. “Modernization:
Theories and Facts.” World Politics 49: 155–83.
Przeworski, Adam, Gonzalo Rivero, and Tianyang Xi. 2015.
“Elections as a Method of Processing Conflicts.” European
Journal of Political Economy 39: 235–48.
Putterman, Louis. 1996. “Why Have the Rabble not Redistributed
the Wealth? On the Stability of Democracy and Unequal Wealth.”
In John E. Roemer (ed.), Property Relations, Incentives, and
Welfare. London: McMillan. Pages 359–89.
Rodrik, Dani. 2017. “Populism and the Economics of Globalization.”
Discussion Paper 12119, Centre for Economic Policy Research,
London.

222
 references

Roemer, John E. 1998. “Why the Poor Do Not Expropriate the Rich:
An Old Argument in New Garb.” Journal of Public Economics 70:
399–424.
Rooduijn, Matthijs, and Tjitske Akkerman. 2017. “Flank Attacks:
Populism and Left–Right Radicalism in Western Europe.” Party
Politics 23: 193–204.
Rosanvallon, Pierre. 2004. Le Modèle Politique Français: La société
civile contre le jacobinisme de 1789 à nos jours. Paris: Seuil.
Rosanvallon, Pierre. 2009. “Réinventer la démocratie.” Le Monde,
April 30.
Rothwell, Jonathan. 2017. “Cutting the Losses: Reassessing the Costs
of Import Competition to Workers and Communities.” Working
paper, Gallup, Inc.
Rothwell, Jonathan, and Pablo Diego-Rosell. 2016. “Explaining
Nationalist Political Views: The Case of Donald Trump.”
Working paper, Gallup. Last revised November 2, 2016.
Rousseau, Jean-Jacques. 1964 [1762]. Du contrat social. Edited by
Robert Derathe. Paris: Gallimard.
Sadurski, Wojciech. 2018. “How Democracy Dies (in Poland):
A Case Study of Anti-Constitutional Populist Backsliding.”
Legal Studies Research Paper No. 18/01, Sydney Law School.
Saiegh, Sebastian. 2009. “Political Prowess or Lady Luck? Evaluating
Chief Executives’ Legislative Success Rates.” Journal of Politics 71:
1342–56.
Sanchez-Cuenca, Ignacio. 2003. “Power, Rules, and Compliance.”
In Jose Maria Maravall and Adam Przeworski (eds.), Democracy
and the Rule of Law. New York: Cambridge University Press.
Pages 62–93.
Schedler, Andreas. 2006. Electoral Authoritarianism: The Dynamics
of Unfree Competition. Boulder, CO: Lynn Rienner.
Schorske, Carl E. 1955. German Social Democracy 1905–1917:
The Development of the Great Schism. New York: Harper & Row.

223
 references

Schumann, Dirk. 2009. Political Violence in the Weimar Republic,

1918–1933: Fight for the Streets and the Fear of Civil War.
New York: Berghahn Books.
Schumpeter, Joseph A. 1942. Capitalism, Socialism, and Democracy.
New York: Harper & Brothers.
Shadmehr, Mehdi, and Dan Berhardt. 2011. “Collective Action with
Uncertain Payoffs: Coordination, Public Signals and Punishment
Dilemmas.” American Political Science Review 105: 829–51.
Sheppele, Kim Lane. 2013. “The Rule of Law and the Frankenstate:
Why Governance Checklists Do Not Work.” Governance 26: 559–62.
Skinner, Quentin. 1973. “The Empirical Theorists of Democracy and
Their Critics: A Plague on Both Houses.” Political Theory 1:
287–306.
Smulovitz, Catalina. 2003. “How Can the Rule of Law Rule? Cost
Imposition through Decentralized Mechanisms.” In Jose
Maria Maravall and Adam Przeworski (eds.), Democracy and the
Rule of Law. New York: Cambridge University Press. Pages 168–87.
Sontheimer, Kurt. 1966. “Anti-Democratic Thought in the Weimar
Republic.” In Fritz Stern (ed.), The Path to Dictatorship 1918–1933:
Ten Essays by German Scholars. Garden City, NY: Anchor Books.
Pages 32–49.
SPLC. 2016. “Update: 1,094 Bias-Related Incidents in the Month
Following the Election.” December 16. www.splcenter.org/hate
watch/2016/12/16/update-1094-bias-related-incidents-month-fol
lowing-election.
Spoon, Jae-Jae, and Heike Kluwer. 2015. “Voter Polarisation and
Party Responsiveness: Why Parties Emphasise Divided Issues,
But Remain Silent on Unified Issues.” European Journal of
Political Research 54: 343–62.
Stern, Fritz. 1966. “Introduction,” In The Path to Dictatorship
1918–1933: Ten Essays by German Scholars. Garden City, NY:
Anchor Books.

224
 references

Stokes, Susan C. (ed.). 2001. Public Support for Economic

Reforms in New Democracies. New York: Cambridge
University Press.
Stone, Geoffrey R. 2018. Perilous Times: Free Speech in Wartime,
from the Sedition Act of 1798 to the War on Terrorism. New York:
W.W. Norton.
Strauss, David A. 2018. “Law and the Slow-Motion Emergency.”
In Cass R. Sunstein (ed.), Can It Happen Here? New York:
HarperCollins. Pages 365–86.
Svolik, Milan W. 2015. “Which Democracies Will Last? Coups,
Incumbent Takeovers, and the Dynamic of Democratic
Consolidation.” British Journal of Political Science 45(4):
715–38.
Svolik, Milan W. 2017. “When Polarization Trumps Civic Virtue:
Partisan Conflict and the Subversion of Democracy by
Incumbents.” Unpublished manuscript, Yale University.
SWIID. 2014. “The Standardized World Income Inequality Data
Base.” Version 4.0. https://dataverse.harvard.edu/dataset.xhtml?
persistentId=hdl:1902.1/11992.
Teinturier, Brice. 2018. “Perceptions de la politique et vote: ce qui
a changé.” Cahiers français 404 (May–June): 62–71.
Tingsten, Herbert. 1973. The Swedish Social Democrats. Totowa:
Bedminster Press.
Turner, Henry Ashby Jr. 1985. German Big Business and the Rise of
Hitler. New York: Oxford University Press.
UNU-WIDER. 2014. “World Income Inequality Database
(WIID3.0A).” June. www.wider.unu.edu/research/WIID-3a/en_
GB/database/
Varol, Ozan O. 2015. “Stealth Authoritarianism.” Iowa Law Review
100: 1673–742.
Wagner, M. 2012. “Defining and Measuring Niche Parties.” Party
Politics 18: 854–64.

225
 references

Weakliem, David. 2016a. “Declining Support for Democracy?”

December 7. https://justthesocialfacts.blogspot.com/2016/12/decl
ining-support-for-democracy.html?m=0.
Weakliem, David. 2016b. “Going Downhill.” December 2. https://
justthesocialfacts.blogspot.com/2016/12/going-downhill.html?
m=0.
Weingast, Barry R. 1997. “Political Foundations of Democracy
and the Rule of Law.” American Political Science Review 91:
245–63.
Weingast, Barry R. 2015. “Capitalism, Democracy, and
Countermajoritarian Institutions.” Supreme Court Economic
Review 23: 255–77.
Wilson, Kenneth A. 2017. Cross-National Time-Series Data Archive.
Yocelevzky, Ricardo A. 2002. Chile: partidos politicos, democracia
y dictadura, 1970–1990. Mexico: Fondo de Cultura Economica.
Zakaria, Fareed. 1997. “The Rise of Illiberal Democracy.” Foreign
Affairs 76(6): 22–43.

226
 index

abuse of power, 77, 177 authoritarianism, 40, 43, 49, 51, 89,
abusive constitutionalism, 181 96, 148–50, 176–8, 179
actors, political, 13, 140, 151, 156 competitive, 15
agriculture, 53, 59, 194 electoral, 15, 26
Alessandri Autor, D., 108, 127
Arturo, 52–4 average incomes, 41, 106, 136,
Jorge, 52 138, 192
Algeria, 65, 67–72 Aylwin, Patricio, 57, 62
Allende, President Salvador, 9, 13,
52–5, 57–64, 79, 138, 149, 200 backsliding, 2, 15, 26, 143, 172–6, 180,
alternations, 21–5, 30, 37–8, 41, 52, 97 183–5
anger, 98, 120 governments, 185, 187–9
antagonisms, 7–8, 145 ballots, 13, 64, 69, 163, 197, 201
anti-government demonstrations, Bangladesh, 30
37–8, 47 banks, 1, 53, 56, 101
anti-immigration sentiments, 203–4 central, 5, 123, 201
anti-system parties, 87, 134 Bavarian People’s Party, see BVP.
anti-system rhetoric, 199 Belgium, 32, 84, 90–1, 115
anti-system sentiments, 1, 205 beliefs, 22, 102, 106–7, 118–19, 127,
Argentina, 32–3, 37, 134, 153, 165 137, 185–7, 192, 194
armed forces, 58–9, 62, 189 ideological, 52
Armigeon, K., 90–3, 139 ingrained, 107
artificial intelligence, 193 biases, 38, 126
assassinations, 38, 47, 60, 70 bicameralism, 5, 200
associations, 152–3, 200 bills, 56, 167–8, 174, 178
attitudes, 80–3, 97, 116, 120, 126, Blum, Leon, 65
129–31, 146 Bobbio, Norberto, 5, 201
political, 61, 95 bourgeoisie, 2, 9, 18, 55, 108
Australia, 91, 104, 115, 123–5 dictatorship of the, 134
Austria, 41, 90, 134 petty, 18, 88, 113

227
 index

breakdown of class compromise, choice, 63, 117, 138, 155, 161, 164,
103, 108–9 170–1, 196–7
breakdown of public order, 12, 14, collective, 198
21, 167, 169–70 and elections, 196–7
Brüning, Chancellor Heinrich, 44, Christian Democrats, 54–6, 62, 79
49–51 citizens, 115, 118, 160, 164, 174, 178,
bureaucracies 182, 188, 195
non-partisan, 20 individual, 183, 186, 189
public, 8, 20, 177 polarization, 115
BVP (Bavarian People’s Party), civil disobedience, 166–7
43–4, 79 civil law, 153
civil peace, 8, 200
Canada, 12, 31, 91, 123 civil war, 16, 62, 74, 138
candidates, presidential, 55, 88, 120 class compromise
capitalism, 11, 16–19, 39, 42–3, 52, breakdown, 103, 108–9
64, 150 democratic, 110
crises, 11 cleavages, 147–8
and democracy, 16–19, 42, 64 political, 42, 158
Carr, W., 46–8, 51 rural–urban, 53
causality, 29, 80, 103, 121–5 Clintons, 120, 202–3
center parties, 97, 99–100, 203 coalitions, 43–6, 54–6, 63–4, 66, 68,
ideological distance between, 95 87, 98, 150, 158
Center Right, 95 electoral, 66
center-right parties, 85, 99 government, 41, 43, 49, 55, 145,
central banks, 5, 123, 201 149, 187
Chetty, R., 107, 137 coexistence, 17, 19, 89
children, 106, 131, 137–8, 149, coin toss, 160–1
198 collectivity, 143, 196–9
Chile, 9, 22, 26–30, 38–9, 52–3, 61–3, Colombia, Constitutional Court, 181
79, 101, 158 color-blind society, 118
democracy, 52 Communists; see also KPD, 11, 41–2,
economy, 53, 56 45, 47, 52, 54–5, 63, 66, 134
Ley de las Areas, 56–7, 79 comparisons, 22, 27, 34, 38–9, 133
outcome of problems, 63–4 competition, 109, 126–8, 151
problem signs, 58–63 electoral, 151–2, 197
threats, 52–8 import, 108–9, 127
China, 16, 108–9, 127, 153, 192 political, 8, 19

228
 index

competitive authoritarianism, 15 political, 129, 134, 181

compromise, 18–19, 55, 110–11, Costa Rica, 31, 37, 157
134, 150 counterfactual questions, 73, 77
concessions, 14, 165, 171 coups, 16, 26–30, 39, 50, 58, 62, 65,
confidence, 1–3, 5, 12–14, 49, 68–70, 101
101–2, 202 courts, 6–8, 32, 76–7, 152, 153–4,
conflicts, 7–9, 21, 32, 39, 58–60, 70–1, 165–6, 167, 175, 177–80
162–4, 171, 204–5 constitutional, 5, 152, 165–6, 175,
differences between, 146–9 179–81
elections as method of supreme, 77, 157, 190–1
processing, 159–64 crises, 1–5, 9–16, 19–21, 27–33, 35–7,
and institutions, 145–59 39–40, 73–8, 80–7, 99–101
political, 8, 13, 21, 145, 151, 154 economic, 11–12, 14, 30–2, 34,
processing, 7, 154–5, 158–9, 163, 41–2, 48–9, 138, 193
165, 171, 205 governmental, 35–6, 38
regulating, 161, 168 political, see political crises
structuring, 150, 153, 205 stagflation, 11, 94
Congress, 54, 56–8, 64, 74, 77, 157, cultural transformations, 22, 80–3
174, 189–91 cultures, national, 117
consolidated democracies, 29–35, Czech Republic, 115
37, 133
constitutional courts, 5, 152, 165–6, Dancygier, R.M., 97, 125, 129
175, 179–81 data, systematic, 79, 120
constitutionalism, 177, 180 DDP (German Democratic Party),
abusive, 181 40, 43–5
constitutions, 18, 50–1, 58, 65, 69, de Gaulle, General Charles, 65,
71–2, 157–8, 180–1, 189–90 67–9, 71–2
changes, 177–9 Dean, John, 76
Weimar, 49–50, 176, 187 deconsolidation, 15, 172, 175, 185
constraints, 20, 110, 151, 161, 190 democratic, 100, 177
constitutional, 160 defeats, electoral, 16, 155
contramajoritarian, 4 demands, 8, 18, 63, 71
electoral, 21 democracy
control, 20–1, 25, 55, 63, 64, 137, 163, breakdowns, 26–7
165, 204–5 and capitalism, 16–19, 42, 64
corrections, 119 consolidated, 29–35, 37, 133
corruption, 12, 32, 196 crises of, 2–16, 25–80, 101

229
 index

democracy (cont.) dissatisfaction, 62, 147, 166,

decline of support for democracy 196–9, 202
in surveys, 100–2 districts, electoral, 65, 118, 182,
erosion, 23–6, 101, 183 186, 202
liberal, 5, 176–8 divisiveness, 39, 113–22, 149
mature, 1, 143 DNVP (German National People’s
past experience of, 37 Party), 43–5
presidential, 35, 38 Dominican Republic, 32
and quest for power, 19–21 DStP, 43–5
representative, 5, 88, 196, 198 DVP (German People’s Party), 43–5
Weimar, 27, 39
democratic breakdown, economic conditions, 78, 126,
probability, 36 129–30, 131, 138
democratic credentials, favorable, pre- and post-2008, 135
15, 178 economic crises, 11–12, 14, 30–2, 34,
democratic deconsolidation, 100, 177 41–2, 48–9, 138, 193
democratic institutions, 10, 12–15, economic forecasting, 192
21, 76, 78, 172, 178, 202 economic growth, see growth
democratic norms, 87, 176, 179, 181–3 economic interests, 8, 16
democratic opposition, 73, 166 economic issues, 44–5, 88, 96
Democrats, 77, 113, 120, 204 economic policies, 42, 88, 201
demonstrations, 37, 57, 70, economic transformations, 64, 107
166–7, 170 economy, 33–5, 55–6, 80, 89, 103–13,
anti-government, 37–8, 47 119, 123, 140
peaceful, 37 Ecuador, 30–1, 157
demonstrators, 75, 166, 183 education, 29, 60, 89, 99, 125, 148,
Denmark, 90–1, 93, 115, 123 156, 201
destabilization, 22, 84–5 elected governments, 23, 176, 198
dictatorship, 14, 25, 35, 72, 176, 201 elections, 3–9, 21–7, 40–3, 50–4,
of the bourgeoisie, 134 119–27, 158–64, 171–3, 196–202,
of the proletariat, 134 205–6
direct elections, 72, 174, 179, 202 and choice, 196–7
disasters, 10–12, 14–15, 25, 30, 41, direct, 72, 174, 179, 202
71–2, 78, 143, 163 France, 89, 201–2
economic, 14, 41 government and opposition
discretion, 143, 172–3, 183, 185 between, 164–71
disorder, 169–70 indirect, 125, 200

230
 index

as method of processing conflicts, failure of democracies, 171

159–64 falsehoods, 119
parliamentary, 58, 65, 69, 73 Fehrenbach, Konstantin, 41
presidential, 54, 72, 79, 89–90, Finland, 31, 84, 91, 115
98–9, 165, 188, 202 flexible labor markets, 88, 95, 154
electoral authoritarianism, 15, 26 force, 16, 25, 40–1, 55, 62, 146, 161–2,
electoral competition, 151–2, 197 173, 198
electoral defeats, 16, 155 forecasting, economic, 192
electoral districts, 65, 118, 182, 186, 202 Fourth Republic, 9, 65–9, 71–2, 158
electoral incentives, 6, 160 frameworks, institutional, 7–9, 25,
electoral participation, see turnout 29, 71, 152, 154–8, 165, 168,
electoral strategies, 42, 89 170–2
electoral systems, 9, 20, 42, 151–2, 158 France, 37–9, 64–73, 84–7, 90–1,
electoral victories, 21, 203 97–8, 99–100, 121–5, 155–6
electorates, 86–7, 97, 99–100, 127, democracy, 64
139, 201 Fourth Republic, 9, 65–9, 71–2, 158
elites, 87, 102, 134, 195, 199–203 outcome of problems, 71–3
elitism, 87, 201 problem signs, 66–70
emergency threats, 65–6
powers, 49, 51, 64 unrest by year, 73
state of, 60, 67, 187 freedoms, 5, 17, 21, 25
states of, 60, 67, 187 political, 17, 150
employment, 19, 41, 103–6, 109, Frei, Eduardo, 52–4, 56, 59, 62
126, 193 Front National, 91, 99, 126
equality, 16–17, 19, 95, 102, 200
inter-generational, 137 Gaullists, 66–8, 72
Estonia, 27, 30 GDP, 34, 135, 201
European Union, 100, 203 general patterns, 29–38, 140
Evans, R.J., 46–7, 50 general strikes, 37–8, 47
everyday life, 2, 13, 37, 120, 122 generals, 58, 68–70, 72, 157
evidence, 52, 98, 100–1, 103, 114, 119, German Democratic Party,
127–31, 137 see DDP
executive orders, 190–1 German National People’s Party, see
experience, 25, 37, 127–8, 138, 194 DNVP
historical, 21, 169, 193 Germany, 8–9, 30–1, 39–51, 65–6,
experts, 101–2 78–9, 90–1, 115, 121–5
extreme Left, 40, 89, 99 democracy, 40–1

231
 index

Germany (cont.) historical experience, 21, 169, 193

location of parties on left-right Hitler, Adolf, 43, 49–51, 79, 158,
and democratic-authoritarian 176, 192
space, 45 Honduras, 30, 157
Nazi, 73 hostility, 78, 113, 120, 138
outcome of problems, 49–51 Hungary, 21, 172–5, 179, 188
problem signs, 46–9 Huq, A.Z., 3, 172, 176, 179–81, 190
Reichstag, 27, 43, 47–50
stab in the back, 42 Iceland, 91
threats, 41–6 identities, 117–18
Weimar Constitution, 49–50, ideologies, 1, 52, 85, 95, 113, 117–18,
176, 187 121, 134
Gini coefficients, 34, 41, 104, 135 illegal acts, 35, 77
Ginsburg, T., 3, 172, 176, 179–81, illegal immigrants, 96, 149
190, 201 immigrants, 89, 96, 116–17, 129, 149
globalization, 22, 80, 88, 103, 107–9, illegal, 96, 149
125, 131, 192 immigration, 89, 96–7, 98–9, 103,
government coalitions, 41, 43, 49, 55, 113–16, 123, 128, 129–30, 147–8
145, 149, 187 anti-immigration sentiments,
government policies, 37, 110, 159, 203–4
165, 166 impeachment, 32, 58, 62–4, 77, 184
governmental crises, 35–6, 38 import competition, 108–9, 127
governments, 4–6, 8–9, 12–13, 16–17, impotence, 195, 198
40–2, 44–6, 56–73, 158–60, incentives, 7, 110, 117, 129, 150–1, 154,
163–90 163, 205
backsliding, 185, 187–9 electoral, 6, 160
elected, 23, 176, 198 income inequality, 19, 29, 41, 53, 103,
incumbent, 5, 164 130, 136
and oppositions between income protection, 193, 195
elections, 164–71 income stagnation, 103, 137
Greece, 30, 89–91, 101 incomes, 11, 13, 17, 19, 33–5, 106, 109,
growth, 19, 34, 136–7, 145
Guatemala, 32 average, 41, 106, 136, 138, 192
Guyana, 32 low, 22, 62, 106, 136, 192–3
redistribution, 19, 41, 88, 147,
hate crime, 120–1 194
Hindenburg, 49, 51, 79 incumbent advantage, 161, 183, 185

232
 index

incumbents, 5, 13, 16, 20, 26, 31, Italy, 12, 90–1, 133, 170, 202
159–60, 162–4, 186–7 Ivarsflaten, E., 88, 129
indirect elections, 125, 200
industrial revolution, 138 Jamaica, 31–2
industrial workers, 53, 88, 109 Japan, 84, 90–1, 104, 109, 123
industry, 53, 59, 103–6, 112, 153–4, 193 job losses, 109, 126–7
inequality, 5, 19, 34, 38, 103, 136, 193, jobs, 80, 127, 193, 204
194–5, 204 outflow, 192–3
income, 19, 29, 41, 53, 103, 130, 136 judicial review, 6, 179, 201
influence judiciary, 4–5, 170
equal, 197
political, 155, 200 Kent State University, 75, 170, 183
institutional frameworks, 7–9, 25, KPD (German Communist Party),
29, 71, 152, 154–8, 165, 168, 170–2 41, 43–5, 79, 148
institutional reforms, 14, 39, 202–3
institutional rules, 25, 156, 171 labels, 15, 26, 83, 85, 89, 134, 138
institutional status quo, 10, 15, labor, 80, 137, 195
32, 184 labor, markets, 108–9, 193
institutional systems, 36, 42, 77, 168 labor markets
institutions, 1, 6, 7–8, 10–11, 13–14, flexible, 88, 95, 154
72, 77–9, 168, 172 labor share, 34, 53, 136–7
and conflicts, 145–59 language, 87, 95, 116, 130–1, 204
democratic, 10, 12–15, 21, 76, 78, Latin America; see also individual
172, 178, 202 countries, 35, 140, 164
political, 3, 7, 143, 150–1, 154, 156 laws, 3–7, 56–7, 58, 145–7, 152–3,
representative, 10, 14, 15, 74, 77, 156–7, 167–8, 189–91, 202–4
166–8, 195–6, 198–200, 202 Le Pen, Marine, 89, 117
insurgent parties, 195, 199 leaders, 64, 68, 73, 79, 84, 88, 102,
interest groups, 153 155, 187
interests, 7–9, 13, 25, 150–5, 161–3, strong, 14, 88, 101–2
175, 198, 200, 205 Left, 31, 45, 54–7, 59, 83, 87–9, 95,
economic, 8, 16 97–9, 137
of elites, 195, 199 extreme, 40, 89, 99
institutional interplay of, 13, 155 left–right dimension, 95–7, 98,
partisan, 77, 177 114, 147
intergenerational equality, 137 legislation, 56, 64, 165, 182, 186,
interwar period, 37, 83, 134, 136–7 191

233
 index

lessons from history, 22, 78–80, multiculturalism, 117–18

133, 141 multi-party systems, 113, 205
Ley de las Areas, 56–7, 79
liberal democracy, 5, 176–8 national culture, 117
liberal rights, 3, 5, 172 National Socialist German Workers
liberty, 8, 13, 189, 200 Party, see NSDAP
losers, 8, 112, 131, 156, 159–66, national sovereignty, 116, 204
179, 205 nationalization, 56–7
nativists, 87, 89, 204
Maddison, A., 31, 33, 133 Nazis, 42–3, 47–8, 51
Magaloni, B., 26 Netherlands, 91, 115, 123
Manin, B., 164, 201 New Zealand, 91, 104, 123
manufacturing, 103, 108, 136–7, 194 NGOs (non-governmental
material conditions, 10, 19, 113, 205 organizations), 174, 179, 190
material progress, 22, 106–7, 137, 193 Nixon, President Richard, 39, 75–7
Mauritius, 30, 32 non-governmental organizations,
media, 1, 21, 76, 173–4, 177, 204 see NGOs
median incomes, 53, 123 non-voters, 48, 99
Medina, L., 114, 147 norms, democratic, 87, 176, 179,
Mélenchon, Jean-Luc, 89, 155 181–3
Mexicans, 88, 117 Norway, 37, 90–1
Michaels, W.B., 118 NSDAP (National Socialist German
military, 32, 35, 60, 62–4, 67–8, 71, 73, Workers Party), 44–5, 48, 50, 84
140, 156–7
coups, see coups. occupations, 53, 59, 73, 128, 167,
ministers, 46, 55–7, 60, 63, 66, 68, 188 193–4
MIR (Movimiento de Izquierda OECD, 30, 86, 91–3, 104–6, 112,
Revolucionaria), 55, 60 136–8, 139, 148
mobility, 19, 103, 128, 192–4 opponents, 20–1, 47, 50, 75, 157, 182
mobilization, 36, 42, 100, 155–6, 179 opposition, 13, 20–1, 56–8, 61–2,
monarchists, 40, 44 72–6, 164–6, 168–75, 179–80,
monopoly power, 19, 109, 173 182–9
Montesquieu, 177, 189 democratic, 73, 166
Movimiento de Izquierda and government between
Revolucionaria (MIR), 55, 60 elections, 164–71
MRP, 66–7 media, 175, 183, 186
Muller, Hermann, 41 organized, 26, 166

234
 index

parties, 165, 173 right-wing, 91, 94, 131, 140

partisan, 174, 187 traditional, 48, 62, 80, 91–3, 96–7,
potential, 175, 183 98–9, 139, 199, 203–4
organizations, 7, 18–19, 71, 140, 145, weak, 205
150, 152–4, 205 working-class, 18
paramilitary, 46, 167 party politics, 48, 62, 100
terrorist, 70 party systems, 22, 86, 138
union, 153, 156 erosion, 83–7
organized political forces, 71, 155 past experience of democracy, 37
outcomes, 8–11, 13, 15, 38–40, 146–7, Patria y Libertad, 59–60
154–7, 160–3, 184–6, 197–9 patterns, 85, 98, 168
Chile, 63–4 general, 29–38, 140
final, 49, 51 peace, 150, 162, 198–9
France, 71–3 civil, 8, 200
Germany, 49–51 social, 47
unfavorable, 7, 150, 156 peak preferences, 146–8
United States, 77 peasants, 56, 59–60, 63, 103
people, the, 87–8, 134, 195, 198, 202–3
Pakistan, 30 per capita income, 30–1, 33, 42,
paramilitary organizations, 46, 167 104, 133
parliamentary elections, 58, 65, perceptions, 51, 98, 107, 126, 128,
69, 73 130, 194
parliamentary systems, 32, 36, 79 Peru, 30–1
parliaments, 66–9, 71, 80, 87, 158, petite bourgeoisie, 18, 88, 113, 140
165, 169, 174, 187 Pfilmin, Pierre, 67–8
parties, 42–6, 54–5, 66, 80–93, 96–7, Philippines, 30
118–20, 151–3, 159–61, 196–7 Piketty, T., 41, 87, 98–9
center, 95, 97, 99–100, 203 plurality, 43, 52, 64, 197, 202
center-right, 85, 99 Podemos, 87–9
competing, 20, 25, 43 Poland, 20, 115, 172–5, 179, 188
effective number of, 43, 86, 139 polarization, 9, 42, 78, 92, 113–16,
left-wing, 137 121–5, 130, 138, 204
major, 83, 96 police, 60, 68, 70, 73–5, 170
new, 1, 84, 138, 155, 197 policies, 109, 112, 126, 159–61, 163–6,
populist, 88–9, 100, 127, 129 168–74, 178, 183–6, 203–4
radical Right, 89–90, 99, 123, 127, economic, 42, 88, 201
129, 139, 203 government, 37, 110, 159, 165, 166

235
 index

political actions, 153, 164, 205 monopoly, 19, 109, 173

political actors, 13, 140, 151, 156 political, 10, 16–18, 137, 154, 159
political attitudes, 61, 95 to rule by decrees, 68, 71
political cleavages, 42, 158 usurpation of, 26, 39, 68, 78, 161
political competition, 8, 19 Prats, General, 58–60
political conflicts, 8, 13, 21, 145, preferences, 113, 139, 148–9, 151
151, 154 ideal, 147–8
political corruption, 129, 134, 181 individual, 113, 197
political crises, 31–3, 39 peak, 146–8
political forces, 7, 10, 14, 25, 72, 88, presidential candidates, 55, 88, 120
150, 154, 160–2 presidential democracies, 35, 38
major, 9, 171 presidential elections, 54, 72, 79,
organized, 71, 155 89–90, 98–9, 165, 188, 202
political freedom, 17, 150 presidential systems, 32, 35–6, 63,
political influence, 155, 200 125, 158, 179, 205
political institutions, 3, 7, 143, 150–1, presidential veto, 5, 174, 200
154, 156 presidents, 57–9, 62–4, 71–2, 75–7,
political parties, see parties 151, 157–8, 174–6, 187, 189–91
political polarization, 78, 121–5, 204 elected, 36, 69, 72
political power, 10, 16–18, 137, pressures, 32, 40, 57
154, 159 productive resources, 11, 18
political systems, 66, 151, 195–6 productivity, 109, 123, 137
political violence, 47, 162 professional politicians, 88, 92,
politicians, 1, 6–7, 12, 19, 26, 39, 46, 139, 192
50, 99–100 progress, material, 22, 106–7, 137,
professional, 88, 92, 139, 192 193
politics, 62, 87–8, 98, 134, 159, 164 property, 10, 17–18, 145–6, 200
popular initiative referendums, protectionism, 89, 193, 203
88, 202 public bureaucracies, 8, 20, 177
populism, 1, 195, 199, 202–3 public order, 3, 5, 7, 12–13, 68, 164,
rise, 87–100 167, 169–71, 199
populist parties, 88–9, 100, 127, 129 breakdown, 12, 14, 21, 167, 169–70
post-truth world, 118–19
power, 5–6, 18–19, 26–30, 49–50, quasi-authoritarian regimes, 177
63–4, 68, 71–2, 133–7, 175–7 questions
abuse of, 77, 177 open, 90
emergency, 49, 51, 64 survey, 101–2, 135, 188

236
 index

race-based violence, 121 respondents, 52, 58, 69, 106, 116, 123,
racism, 83, 89, 113, 116–18, 130, 134 131, 137
Radical Party, 52, 54 revolution, 14, 18, 42, 64, 189, 196
radical Right, 14, 89–93, 128, 131 industrial, 138
parties, 89–90, 99, 123, 127, 129, Right, 40–2, 45, 58–9, 80–3, 87–100,
139, 203 123–7, 129–31, 139–40, 203–4
voting and supporting, 126–32 center, 95
Reagan, 94, 111 radical, 14, 89–93, 123–6, 128, 131
redistribution, 98, 112, 194–5 rights
redistribution of incomes, 19, 41, 88, liberal, 3, 5, 172
147, 194 voting, 182
reelection, 76, 160, 188 right-wing parties, 91, 94, 131, 140
referendums, 69–70, 134, 174, right-wing populism, see populism
177–9, 181 riots, 21, 34, 37–8, 47, 167–8
popular initiative, 88, 202 Romania, 32
reforms, 57, 67 Rosanvallon, P., 3–4, 118
educational, 156 Rothwell, J., 108–9, 128
institutional, 14, 39, 202–3 rule of law, 3–7, 172
partial, 14–15
refugees, 115–17 Santiago, 52, 58, 59–62
regimes, 7, 15, 25–6, 34, 178 slums, 53–5, 60
Reichstag, 27, 43, 47–50 Schneider, General, 59–60, 62
religion, 29, 80, 99–101, 121, 148, Schumpeter, J.A., 5, 9, 135, 164
153, 204 segregation, 195, 204
reparations, 41–2 separation of powers, 6, 9
representative government, 164, 199 services, 53, 103–6, 154, 193
representative institutions, 10, 14, 15, personal, 194
74, 77, 166–8, 195–6, social, 89, 110, 194
198–200, 202 sexual orientation, 121
representative system, 7, 150, 205 SFIO, 66–7
repression, 5, 12–13, 20–2, 74–5, shadow lines, 117
170–1 Slovenia, 115
Republicans, 77, 98, 113, 120, 204 slums, Santiago, 53–5, 60
Rockefeller, 139 Social Democrats, 40, 49–50, 55, 64,
resistance, 13, 59, 143, 161, 171, 179, 83, 95, 140
188–9 social peace, 47
violent, 161–2 social services, 89, 110, 194

237
 index

socialism, 42–3, 52, 55, 62, 66, respondents, 52, 58, 69, 106, 116,
83–4, 150 123, 131, 137
Solomon Islands, 30 Sweden, 6, 90–1, 115, 123–5
sovereignty Switzerland, 90–1
national, 116, 204
popular, 88 taxes, 80, 88, 194
Spain, 89, 91, 101, 121–3, 138 Teinturier, B., 98
Podemos, 87–9 terrorism, 65, 73, 167, 170
SPD (Social Democratic Party), Thailand, 30, 33
40–1, 43–5, 64, 148–50 Thatcher, 94, 111–12
Sri Lanka, 30, 187 threats, 5, 10, 14–17, 32, 185
stability, 83–5 Chile, 52–8
stagflation crisis, 11, 94 France, 65–6
stagnation, 22, 106, 136–7, 192–3 Germany, 41–6
stakes, 8–9, 20–1, 143, 154, 163, United States, 74
192, 206 trade unions, see unions
state of emergency, 60, 67, 187 traditional parties, 48, 62, 80, 91–3,
status quo, 9–10, 14, 69, 157, 200, 96–7, 98–9, 139, 199, 203–4
205 transformations, 31, 100, 103–7,
institutional, 10, 15, 32, 184 130–2, 178
stealth, 15, 172–91 cultural, 22, 80–3
street fights, 167, 170 economic, 64, 107
Stresemann, 43–5 Trump, President Donald, 87, 97,
strikes, 34, 37–8, 47, 59, 65, 152, 156, 109, 117, 120, 179–82, 188,
166, 169–70 202–3, 205
general, 37–8, 47 Turkey, 172–4, 179
strong leaders, 14, 88, 101–2 turnout, 1, 48, 62, 93, 139
subversion, 15, 23, 172–91
from above, 183–8 Ukraine, 32
suffrage, 14, 200 unemployment, 11, 41, 97–8, 127–9,
individual, 16 154, 192, 195
universal, 17–18 unfavorable outcomes, 7, 150, 156
supreme courts, 77, 157, 190–1 Unidad Popular, 54–5
surveys unions, 18–19, 109–11, 137, 140, 152–4,
decline of support for democracy 155, 200
in, 100–2 decline, 137
questions, 101–2, 135, 188 density, 112–39, 140

238
 index

United Kingdom, 6, 31, 84, 91, 111, vote shares, 62, 93, 97, 99, 127
123–5, 130, 165 voters, 40, 92, 97–9, 138–9, 147, 151,
United States, 38–9, 73–7, 106–7, 197, 201–2, 205
109–17, 120–5, 130–1, 136–9, potential, 92, 203
188–90, 194–5 votes, 54–6, 90–3, 97, 125, 127–8,
democracy, 73 161–2, 181–2, 188, 201–2
Kent State University, 75, 170, 183 voting, 69, 126–7, 161–2, 182
outcomes, 77 voting, results, 162–3
problem signs, 75–7 voting rights, 182
Senate, 165, 201
and subversion by stealth, 189–91 wages, 3, 11, 109–12, 123–6, 137, 154,
threats, 74 192–4
universal suffrage, 17–18 war, 37, 41–2, 46, 65, 70–2, 75, 83–4,
unrest, 38, 47, 60–1, 73–7, 171 131, 138
Uruguay, 12, 31 Watergate scandal, 76–7
usurpation of power, 26, 39, 68, Weimar, see Germany
78, 161 winners, 52, 85, 112–13, 125, 159–62,
166, 187, 205
Venezuela, 26, 31, 172–4, 179–80, 188 women, 59, 200
vetos, presidential, 5, 174, 200 workers, 52–3, 57, 59–61, 89, 99, 109,
victories, 52, 94 128, 153
electoral, 21, 203 industrial, 53, 88, 109
institutional, 156 working class, 55, 80, 140, 148
violence, 13, 46, 60–1, 75, 155, 160–1,
166–7, 170 xenophobia, 83, 87, 89, 97, 100,
physical, 37 130, 134
political, 47, 162
race-based, 121 Z+BVP, 44
violent resistance, 161–2 Zentrum, 40–1, 43–5, 50, 79

239