The three pillars of stability: Legitimation, repression, and co-optation in autocratic regimes

Abstract

Why do some autocracies remain stable while others collapse? This article presents a theoretical framework that seeks to explain the longevity of autocracies by referring to three pillars of stability: legitimation, repression, and co-optation. These three causal factors are derived by distilling and synthesizing the main arguments of classic and more recent research efforts. Particular emphasis is paid to re-incorporate legitimation in the explanation of stable autocracies. The article conceptionalizes the three pillars and discusses methods of concrete measurement. It then moves on to explain the stabilization process. How do these pillars develop their stabilizing effect? It is argued that reinforcement processes take place both within and between the pillars. They take the form of exogenous reinforcement, self-reinforcement, and reciprocal reinforcement. To illustrate the inner logic of these processes, I draw on empirical examples. I also state what we would need to observe empirically and how we can approach the three pillars methodically. A theoretical framework of this nature has two advantages: it is able to take the complexity of autocratic regimes into account while remaining parsimonious enough to be applicable to all autocratic regimes, irrespective of their subtype; and it integrates a static view to explain stability, with the emphasis on the underlying stabilization mechanisms and facilitates within-case and cross-case comparisons.

Full text

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 13.
The three pillars of stability: legitimation, repression, and co-optation in
autocratic regimes
Johannes Gerschewski*
Research Unit “Democracy: Structures, Performance, Challenges”, Social Science Research Center Berlin (WZB),
Berlin, Germany
Why do some autocracies remain stable while others collapse? This article presents a theoretical
framework that seeks to explain the longevity of autocracies by referring to three pillars of stability:
legitimation, repression, and co-optation. These three causal factors are derived by distilling and
synthesizing the main arguments of classic and more recent research efforts. Particular emphasis is
paid to re-incorporate legitimation in the explanation of stable autocracies. The article
conceptionalizes the three pillars and discusses methods of concrete measurement. It then moves on
to explain the stabilization process. How do these pillars develop their stabilizing effect? It is argued
that reinforcement processes take place both within and between the pillars. They take the form of
exogenous reinforcement, self-reinforcement, and reciprocal reinforcement. To illustrate the inner
logic of these processes, I draw on empirical examples. I also state what we would need to observe
empirically and how we can approach the three pillars methodically. A theoretical framework of this
nature has two advantages: it is able to take the complexity of autocratic regimes into account while
remaining parsimonious enough to be applicable to all autocratic regimes, irrespective of their
subtype; and it integrates a static view to explain stability, with the emphasis on the underlying
stabilization mechanisms and facilitates within-case and cross-case comparisons.
Keywords: autocracy · stability · legitimation · repression · co-optation
What makes autocracies endure?
In 1996, in a widely cited article, Adam Przeworski and his colleagues asked “what makes
democracies endure?” They came to the conclusion that it was mainly economic performance and
the institutional choice of parliamentarism that contributes decisively to the longevity of
democracies.1 Their article stands in a long tradition, and while the research on democracy and
democratization has produced an abundance of theories and empirical results, the research on
autocratic political regimes as the “counterpart” to democracies has not moved beyond the classics
of the 1960s and 1970s.2 This changed as the “third wave of democratization” ebbed away and
* Email: johannes.gerschewski@wzb.eu
This article was published by Taylor and Francis in
Democratization, 2013, Vol. 20 (2013), Iss. 1, pp. 13–38 (2013/01/29):
https://doi.org/10.1080/13510347.2013.738860.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 14.
the widespread optimism both in politics and political science was replaced by a more sobering
assessment. Stable autocracies came more and more to the forefront of scholarly attention. New and
innovative answers to long-standing problems have been recently proposed to shed light on the
logics of autocracies.3 This article is located within this lively ongoing debate and asks what, in turn,
makes autocracies endure?
The main aim is to propose a theoretical framework for the analysis of stable autocratic regimes.4 It
tries to synthesize the main arguments that have been brought forward in both the classics in
autocracy research and in more recent studies regarding why many autocracies have remained
stable. It will be argued that the stability of all autocracies – irrespective of their subtype – can be
explained with reference to what might be aptly called the three pillars of stability: legitimation,
repression, and co-optation.5
These three pillars are not there from the outset, but need to develop over time. How do they get
built, that is how can we make sense theoretically of the stabilization process? I argue that
reinforcement processes take place both within and between the pillars. These processes can take
different forms: an exogenous reinforcement process that is propelled by the available power and
material resources of the ruling regime; an endogenous self-reinforcement process that triggers
path-dependency; and, lastly, a reciprocal reinforcement process that leads to a complementarity
advantage between the pillars. I suggest that these three processes should be studied closely for
explaining the stabilization of autocratic regimes.
A theoretical framework of this kind has, in my view, two advantages. First, it is capable of taking the
complexity of autocracies into account while remaining parsimonious enough to be applicable to all
autocratic regimes. Second, it integrates a static view to explain stability with a more dynamic
perspective to look closely at the stabilization processes and, as a result, facilitates both within-case
and cross-case comparisons.
The remainder of the article is structured as follows: the second section conducts a review of the
literature to identify the three main pillars. The third section engages in concept-building and
operationalizes their specifications in more detail. Against this backdrop, I theorize in the fourth
section how these three pillars come into being, before drawing a conclusion in the final section.
What does the literature say?
Why are legitimation, repression, and co-optation key to the stability of autocracies? To answer this,
a review of the literature will lay open different strands of explanatory schemes that can be
integrated within one comprehensive and coherent framework. Broadly speaking, three research
waves can be identified: the totalitarianism paradigm until the mid-1960s that highlighted ideology
and terror; the rise of authoritarianism until the 1980s that placed more emphasis on socio-economic
factors; and, starting with Geddes' seminal article in 1999, a renaissance of autocracy research that
centred mostly on strategic repression and co-optation.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 15.
First wave, 1930s–1960s: the totalitarianism paradigm
Starting with early works in the 1930s,6 the use of the concept of totalitarianism to characterize a
new social phenomenon became widespread. The benchmark study – at least within political science
– is still the study conducted by Friedrich and Brzezinski.7 In the 1950s, they proposed their famous
“six-point catalogue” for identifying totalitarian regimes.8 The creation of a “new man” was
stipulated by an all-encompassing ideology with strong chiliastic and utopian fervour and was
implemented by the use of terror, by propaganda measures and a strong party.9 While Friedrich and
Brzezinski's argument is structuralist in character, Hannah Arendt's work was a socio-philosophical
attempt to understand the essence of totalitarianism.10 In her work, she highlighted what she
identified as the two main features of totalitarian Herrschaft: ideology and terror. She saw
totalitarianism, with its attempt to atomize society, as the radical negation of what she understood
as the political. In a similar vein, a strong emphasis on ideology and terror was also underlined in a
third approach. Totalitarian movements have been compared to political religion in order to explain
their “tremendum et fascinosum”.11
However, in the 1960s, the totalitarianism paradigm faced mounting critique. In addition to criticizing
its inherent statism, Barber also complained about the fuzzy conceptual foundations of the term.12
Given the scarcity of empirical cases, this led to a situation in which totalitarianism was increasingly
edged out by the concept of authoritarianism.
Second wave, 1960s–1980s: the rise of authoritarianism
The concept of authoritarianism had more modest beginnings. Juan Linz claimed in 1964 that
Francoist Spain should not be studied through the lenses of totalitarianism, but that authoritarianism
constituted a new and distinct phenomenon.13 The rise of authoritarianism as a distinct regime type
began, with O'Donnell's study on “bureaucratic authoritarianism” as the most prominent one.14 In
brief, O'Donnell argued that, due to the fact that Argentina and Brazil reached limits in their import
substitution strategy, bureaucratic elites from the military and business that became increasingly
frustrated with the political and economic circumstances emerged. In response to the perceived
crisis, they formed a coup coalition that finally established an authoritarian regime and sought
economic progress and state order. Although his explanations encountered difficulties in travelling to
other Latin American countries, he emphasized a point that had previously been overshadowed: the
socio-economic dimension functioned both as a driver for the emergence and the maintenance of
authoritarian regimes.15 I will take up this explanatory factor later in the discussion about “specific
support”.16
Two other trends are worth considering within this wave: a proliferation of subtypes and a
“regionalization” of explanations.17 Especially in Latin America and Southeast Asia, military regimes
emerged in the 1960s and 1970s that were characterized by the persistent political role of the
military and that acted behind

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 16.
a quasi-civilian façade. Simultaneously, sub-Saharan Africa experienced an increase in one-party
regimes, while the autocracies in many Arab countries began to build their stability on neo-
patrimonial rule and a “social contract” between the rulers and the ruled.18
What all these explanations have in common is that they varied greatly from the totalitarianism
paradigm and searched for more nuanced and tailored explanations for new (regional) phenomena.
The quintessence was that autocrats cannot rely solely on ideology and terror, but also need to
deliver improved socio-economic performances, sometimes focused in informal reward policies.
Third wave: a renaissance in studies of autocracies
The third wave of autocracy research started after a time lag. It was not until the seminal article by
Barbara Geddes in 1999 that scholarly attention returned to questions surrounding autocracies.19
Most prominently, neo-institutionalist approaches have recently entered the research on autocracies
and have highlighted the stabilizing effect of institutions.20 The role of political parties, legislatures,
and elections as co-optation mechanisms has been of particular importance. In older research,
elections and legislatures were seen as mere window dressing and parties were mainly reduced to
entities that “provide a following for the dictator”.21 Gandhi and Przeworski go beyond this reasoning
and argue that autocratic rulers do make systematic use of these seemingly democratic institutions
to prolong their rule.22 In particular, it has been demonstrated that one-party regimes are more
robust than other regimes. Parties seem to have a stabilizing effect on autocratic rule, as they can
settle and mediate intra-elite conflicts.23 Also, elections serve as a tool for co-optation. They provide
information for the national leaders regarding the loyalty and competence of regional and local
incumbents.24
In contrast to the repressive abuse in totalitarian regimes, Acemoglu and Robinson compellingly
point out that strategic and, in game-theoretic language, “optimal” degrees of repression play a
particularly decisive role in the longevity of autocratic regimes.25 In their rational choice approach, a
second topic is of special interest: intra-elite cohesion. The underlying premise here is that all
autocratic regimes have to share power in an insecure environment, in which any defection must be
avoided by the (threat of the) use of force. The latent danger of intra-elite splits must be tackled. Co-
opting institutions are crucial here, as they alleviate moral hazard problems.26
To conclude, while the totalitarianism paradigm especially emphasized the use of terror and the role
of political ideologies, subsequent studies have differed in their research focus. Socio-economic
conditions, as well as more informal characteristics of authoritarian rule, have come to the forefront
and have largely dominated the discourse. In the last decade, the stabilizing effect that strategic
repression and formal institutions like elections, legislatures, and parties have in shoring up an intra-
elite cohesion have been analysed more rigorously. Table 1 provides an overview.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 17.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 18.
In order to synthesize these different strands in one coherent analytical framework, I propose the
three pillars of legitimation, repression, and co-optation. I argue that they are able to thwart the
danger of regime breakdown that could stem from three sources: from the ordinary citizens whose
non-compliance usually takes the form of popular uprisings and rebellions; from oppositional actors
that organize resistance; and lastly from intra-elite splits in which strategically important elites
deviate from the ruling elite's course.27 The conceptualization, operationalization, and methods of
concrete measurement are dealt with in the following section.
The concepts: legitimation, repression, and co-optation
Reincorporating legitimation in the study of autocracies
Recent research efforts have gradually lost sight of the legitimation dimension. As shown above,
legitimation was at the core of classic studies. Closely linked to the demise of the totalitarianism
paradigm, it currently plays only a secondary role. Contemporary critiques come from three sides:
normative, substantive, and methodical. Normatively, it is argued that a “legitimate autocracy”
constitutes nothing more than an oxymoron. Substantively, it is claimed that legitimation simply does
not matter for the stability of autocracies, as such regimes do not need to rely on people's support.28
And lastly, from a more pragmatic viewpoint, the methodical challenges in measuring legitimation in
autocratic settings remain unsolved.
These are good reasons. However, I argue that we miss an important causal factor when we bracket
out legitimation. Instead, I make the case for re-incorporating it into the explanation of stable
autocracies. Taking up the aforementioned critique, legitimation will be defined here as the process
of gaining support which is based on an empirical, Weberian tradition of “legitimacy belief”.29 It is
then free from normative connotations and therefore does not run counter to the oxymoron
criticism. This was exactly Weber's aim. He attempted to classify political rule without reference to
normative judgments that are supposed to be the “right” rule. Legitimation seeks to guarantee active
consent, compliance with the rules, passive obedience, or mere toleration within the population.
The substantial critique puts the relevance of legitimation for maintaining stability into question.
However, Rousseau's famous dictum that even the strongest needs to transform strength into right,
as he would be never strong enough to be always the master, applies to the autocratic context as
well. I explicitly start from the assumption that behind every political order there must be a
“legitimacy idea”.30 Today's autocracies cannot rely (at least in the long term) entirely on their abuse
of power in a strictly hierarchical, pyramid-shaped political order as the unconstrained tyrants of the
past – from whom all power was derived – might have done. Wittfogel's old idea of an “oriental
despotism”, in which the elite can regulate the access to water and can so rule unboundedly in a
“hydraulic society”, does not seem to be suitable for today's autocracies.31 Instead, they are
characterized by many more interdependencies between the ruler and the ruled.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 19.
Even now, only anecdotal evidence is available as to why legitimation matters. A more solid
theoretical basis is needed. For illustrative purposes it might suffice here to refer to the basic logic in
different empirical cases. Regime maintenance in the former German Democratic Republic (GDR)
has, for example, been explained by a mixture of charismatic and rational legitimacy beliefs. Despite
all mobilization efforts, this legitimacy belief was ironically held by an emergent apolitical middle
class, whose reference points were the family or colleagues from work, instead of focusing on “high
politics”. Yet when the regime came under mounting pressure in the 1970s to give up its utopian
promises and focus more closely on performance criteria, it failed to “deliver”. The GDR's weak spot
was the citizens' disenchantment and perception of a growing discrepancy between the official
ideological claims and the social reality that finally led to its breakdown.32 A similar pattern can be
seen contemporaneously in the so-called “Arab Spring”. For a long time, these populist authoritarian
states were said to rely on a “social contract”: a reciprocal relationship that guaranteed political
acquiescence in return for relatively acceptable economic performances.33 Due to growing
disillusionment, the ruled people in Tunisia and Egypt withdrew from this “social contract” and
protested. From Mexico we know, following Magaloni's study, that the monopolist Partido
Revolucionario Institucional (PRI) gained public support in times of economic growth and that
Mexicans “voted for autocracy” for decades; it was only the younger generation, who had not
experienced the times of economic stability, which finally dumped the PRI from office in 2000.34
Alternatively, take the case of the People's Republic of China, in which the ruling Communist Party
(CP) was very successful in (re)gaining public legitimacy after the Tiananmen protests in 1989 by
using a mixture of economic performance, nationalism, and ideology.35 These exemplary empirical
cases do not, however, provide systematic evidence; a broader comparative study of legitimation in
autocracies is still pending. But these cases illustrate the crucial importance for autocratic regimes to
build a strong legitimizing basis.
The examples also suggest that autocratic regimes are more performance-dependent than is often
assumed. For a long time, the legitimation dimension has been equated with “ideocracies” like Cuba
or North Korea and “theocracies” like Iran.36 The underlying idea was that massive indoctrination
turns citizens into “true believers”. However, it seems that autocratic regimes cannot maintain a
utopian ideology and shield the people from external influences over a longer period. The
indoctrination mechanism reaches its limits. Even North Korea, globally the most reclusive state, has
growing problems maintaining its informational monopoly and needs to legitimate its rule more
strongly with reference to its nuclear ambitions.37 Such as in the case of the GDR, ideocracies arrive
sooner or later at a state in which they need to deliver. This can turn out to be the Trojan horse for
the autocratic elite and makes them more vulnerable to people's assessments.38 Therefore, we
should go beyond the reasoning of the intensity of ideological indoctrination and include
performance and output legitimation as a different legitimation source in our studies.39 I propose
that the old Eastonian

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 20.
distinction between “diffuse” and “specific support” captures the concept of legitimation most
appropriately.
“Specific support” can be defined as the “quid pro quo for the fulfilment of demands”,40 and
particularly includes the performance orientation. As has been emphasized in the second wave of
autocracy research, autocracies have to address popular demands for socio-economic development
and physical security. Besides economic conditions, specific support can also stem from the state's
ability to maintain internal order and social security. “Diffuse support” refers to what the regime
“actually is or represents”.41 In contrast to specific support, it is more general and long-term-
oriented. Diffuse support of this nature can stem from both the political ideologies that have been
the main focus in classic totalitarian research, and also from religious, nationalistic, or traditional
claims, from the charisma of autocratic leaders as well as from external threats that lead to domestic
rally-around-the-flag effects.42
One challenge remains. Even if we agree to have good reasons to reincorporate legitimation in the
study of autocracies, how can we measure it? For specific support, we can assume, in line with the
conceptual discussion that the regime needs to deliver. The better it performs economically, socially,
and in terms of public order, the more legitimate it is in the eyes of the ruled. The rulers fulfil the
social contract. Adequate performance indicators are existent and can also be aggregated to specific
indices.43
The different World Bank Development Indicators, such as Gross Domestic Product (GDP) growth
rates, inflation rates, growth of energy consumption/capita, but also vehicles per capita or
telephones and televisions per capita, are good indicators for the economic performance of the
regime. The growth or decline in social, health, and educational expenditures, the GINI index for
social inequality, the school enrolment per capita and the literacy rate, the physicians per capita, or
the overall Human Development Index (HDI) are adequate indicators that can be used for measuring
specific support.44 To what extent the promise to guarantee internal security and public order has
been kept can be measured by proxy-indicators like the number of riots, strikes, guerrilla warfare, or
the country's crime rate.45
Diffuse legitimation is without doubt a harder nut to crack.46 Measuring legitimation in democracies
routinely relies on survey data about people's attitudes. For autocracies, this kind of data is either
unreliable, as it faces the problem of preference falsification, or simply does not exist. How can we
evaluate how successful an ideology has been? Three routes to come closer to the degree of diffuse
support can be proposed. First, the number and intensity of public protests can be taken as a proxy-
indicator for societal discontent. This route would not derive its informative value from people's
attitude, but more from its empirically observable behaviour.47 The higher the number of protests,
the less a regime is legitimized. However, two problems arise: on the one hand, protest data just
measures the absence of legitimation and is therefore “one-sided”; on the other hand, public
protests depend on the population's ability to protest in the first place, that is the weakening of
repression. Therefore, repression needs to be discounted. A second route for approximating a

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 21.
legitimation measure can be via qualitative assessments of country experts or assessments in
secondary literature, which are both labour-intensive tasks. Third, official legitimacy claims by the
ruling elite can be taken more seriously and can be classified by using content analysis techniques.
Under the premise that the ruling elite must also keep its ideological promises in autocratic settings,
a perceived gap between the promises and the social reality erodes the legitimation basis for the
autocratic elite.
Repression as the backbone of autocracies
Repression is undoubtedly one of the backbones of autocracies and is sometimes even referred to as
a defining feature of autocracies. Yet repression alone cannot account for the longevity of
autocracies. Repression is too costly a way to maintain stability in the long run; as the pointed saying
attributed to Talleyrand goes, you can do anything with bayonets except sit on them. Following
Davenport, repression can be defined as the “actual or threatened use of physical sanctions against
an individual or organization, within the territorial jurisdiction of the state, for the purpose of
imposing a cost on the target as well as deterring specific activities”.48 Its main function lies in
channelling public demands vis-à-vis the political system in a way that these demands do not
endanger the autocratic regime.
In operationalizing the concept, I will make use of the instructive distinction by Levitsky and Way
between “high” and “low intensity” repression.49 In their work they distinguish repression according
to the targeted people or institution and the form of violence used. Against this backdrop, high
intensity coercion can be defined as visible acts that are targeted either at well-known individuals like
opposition leaders, at a larger number of people, or at major oppositional organizations. Concrete
measures include the (violent) repression of mass demonstrations, (violent) campaigns against
parties, and the attempted assassination or imprisonment of opposition leaders. Lower intensity
coercion would then aim at groups of minor importance, is less visible, and often takes more subtle
forms. Concrete measures can be the use of (formal and informal) surveillance apparatus, low
intensity physical harassment and intimidation, and also non-physical forms such as the denial of
certain job and education opportunities as well as the curtailment of political rights like the freedom
of assembly. The measurement of repression is straightforward as data is available. The distinction
by Freedom House (FH) between “political rights” and “civil liberties” mirror to some extent the
distinction between harder and softer forms of repression. The Cingranelli-Richards Human Rights
Dataset (CIRI) is, however, arguably more suitable, as it makes the composition of its indices
transparent and so allows for the construction of separate indices. CIRI's distinction between its
“New Empowerment Index” and the “Physical Rights Index” captures low and high intensity
repression correspondingly. A third alternative database would be the Political Terror Scale (PTS)
project, which bases its assessments on the yearly reports by Amnesty International and the US State
Department Country Reports on Human Rights Practices.50

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 22.
The third pillar: co-optation
I define co-optation as the capacity to tie strategically-relevant actors (or a group of actors) to the
regime elite. In Bueno de Mesquita's terms, members of the “selectorate” need to be bound to the
“winning coalition”.51 Co-optation needs to be exerted so that the actor is “persuaded not to exercise
his power to obstruct” and instead to use the resources in line with the ruling elite's demands.52 The
“players” in this “autocratic subgame” are on the intra-elite level. Military and business elites of
strategic importance need to be co-opted.53 The function of co-optation can be characterized as
inclusionary. It works as a transmission belt to ensure both the intra-elite cohesion and the steering
capacity of the political elite. The ability of the political elite to maintain the balance between
competing subordinate actors and to avoid a situation in which one actor grows too strong by
simultaneously tying all relevant actors to the regime is crucial for the stability of autocracies.
As discussed above, Gandhi and Przeworski make a strong case for co-optation via formal channels.54
Prima facie democratic institutions like parliaments, parties, or elections have vital functions for the
co-optation of strategic elites from business or military ranks. There are, however, also informal ways
of binding actors to the regime. In this regard, (neo) patrimonialism was already seen in the 1970s as
the most widespread type of autocratic rule.55 This often implies that the autocratic elite rules by and
through a close network of direct and indirect ties to subordinate actors. Patronage, clientelism, and
corruption are the most commonly used instruments.56
Concrete measurement again poses some challenges. There is no indicator for co-optation that can
be taken off the shelf. Gandhi and Przeworski have introduced a “degree of institutionalization”. The
basic idea behind this is intriguing. They argue that autocrats need to respond to their corresponding
threat level. They use proxy-indicators like the (military or civilian) origin of the leader, the number of
changes of the leader, the inherited parties, the percentage of other democracies in the world, and
the mineral resource endowment as predictors for institutionalization (meaning: no parties, one
party, or more than one party in the legislature). They then compare the number of predicted parties
to the actual number and capture the effect of over- and under-institutionalization on dictators'
survival.57
To further elaborate on Gandhi and Przeworski's idea, the inclusiveness of parties might also be
measured by the years in office of the leader's party, the government fractionalization, and the
number of cabinet changes. When it comes to binding the military to the ruling regime, the military
expenditures, the size of the military, and militaries in cabinets might indicate the intensity of co-
optation efforts.58
Applying the same reasoning for informal methods, the “demand-side” of co-optation can be
captured by the linguistic, ethnic, religious, and cultural heterogeneity measurements by Alesina and
colleagues as well as by Fearon.59 Another powerful source is the Ethnic Power Relations Dataset.60
However, to what extent this demand is compensated by the use of informal instruments like

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 23.
clientelism and patronage still depends on the qualitative assessment of country experts. To my
knowledge, there is no suitable indicator for larger studies at hand. Given the importance of informal
co-optation, to systematically collect data on informal co-optation mechanisms might be a rewarding
task for future research.61 Figure 1 illustrates the three pillars of stability.
The stabilization process
Having set out three crucial concepts – legitimation, repression, and co-optation – that, I argue, are
causal for the stability of autocracies, we need to go one step further. Stability refers usually to a
status quo, while stabilization refers to a process. Autocracies rest on the three pillars, but how do
these pillars come about? The pillars are not in existence from the very beginning of the regime, but
need to be built over time. As depicted in Table 2, I suggest that we need to focus on three processes
that take place within and between these pillars to explain the stabilization processes:
1) exogenous reinforcement processes that are propelled by the availability of outside power
and material resources of the regime, but take place within the pillars;
2) self-reinforcing processes within the pillars that lead to path-dependency; and
Figure 1. The three pillars of stability (adapted from Gerschewski et al., “Warum überleben
Diktaturen”).

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 24.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 25.
3) reciprocal reinforcement processes and the emergence of complementarity advantages
between the pillars.
In the following section, I will theorize on these processes, explicitly spell out what we would need to
observe if the theory holds, and draw on empirical cases for illustrative purposes.
Reinforcing autocratic rule within the pillars
In order to adequately theorize the time-dimensional character of the process within the pillars, I
borrow insights from the rich neo-institutional literature.62 In order to uncover their stabilizing effect,
I argue that the three pillars need to be institutionalized over time.63 The criss-crossing of multiple
interactions that take place within the “arena” of legitimation, repression, and co-optation needs to
go beyond a mere situational and ad hoc basis and needs to develop into stable institutions. In this
sense, institutions are here understood as behavioural patterns. They are “compliance procedures
and standard operating practices that structure the relationship between individuals”.64
Institutionalized legitimation means that citizens have internalized the legitimating norm;
institutionalized repression would mean that oppositional actors are structurally prevented from
revolting; and institutionalized co-optation represents an interaction between political elite on the
one hand and business and military elites on the other hand, in which cooperation benefits outweigh
the costs. But how does the institutionalization process take place?
In answering this question, I identify two different processes within the pillars. The first one is the
more widespread and “normal” case of an exogenously reinforced institutionalization. The second
one is theoretically very appealing, but empirically rarer: a self-reinforcing institutionalization process
in the case of increasing returns. The distinction is drawn for analytical clarity, but should not rule out
the possibility that these logics change, and sometimes parallel.
The most common and intuitive form of institutionalization refers to a process that is not self-
sustaining, but needs to be propelled externally, that is in our case from outside the pillars.
Legitimation, repression, and co-optation have no in-built self-reinforcing mechanism, but they are in
need of constant external drive. Therefore they are more vulnerable and dependent on the
availability of external resources. The power and material resources determine the limits of the
institutionalization process. If the political regime elites are no longer able, due to hard budget
constraints and/or declining power resources, to foster legitimation within the population, to uphold
a functioning repression apparatus, or to distribute enough material benefits to co-opt strategic
actors in a sufficient way, the institutionalization process within the pillars comes to a halt. To use a
metaphor, the ruling regime acts like a plate spinner who needs to hold the plates (the institutions)
in motion constantly, otherwise the plates fall, and the institutionalization process ends.65
The second form of institutionalization process refers to path-dependent explanations. Once an
institution is set on track, the path is difficult to alter and

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 26.
reinforces itself. Mahoney has proposed different mechanisms of self-reinforcement: a legitimation
mechanism that reproduces itself “because actors believe it is morally just or appropriate”;66 a
mechanism based on power asymmetries; and a utilitarian mechanism that reproduces itself due to
corresponding cost-benefit calculations.67 These three mechanisms can be adopted. Legitimation can
reinforce itself due to the nurturing of people's supporting attitudes vis-à-vis the regime. Repression
functions by reproducing power asymmetries between the ruler and the opposition. Co-optation can
be seen as more strategic action, in which both the ruling elite and the elite to be co-opted weigh
their individual costs and benefits.
In line with Mahoney, I argue that these mechanisms are self-reinforcing, which leads to path-
dependency, and also concur with him that, for analytical added value, we need to go beyond a
“history-matters” argument. I part company, however, when it comes to show this phenomenon
empirically. I suggest narrowing down path-dependency to demonstrate that these mechanisms
exhibit increasing returns.68 Stemming originally from economics, increasing returns occur when a
change of one unit on the explanatory side is followed by a change of more than one unit on the
explanandum. In other words, the marginal costs that occur when “producing” one more unit
decrease with growing numbers of the product. The marginal cost curve is convex and monotonically
decreasing.
Although path-dependency is often posited in social science explanations, it is much rarer than the
aforementioned exogenously-propelled institutionalizations. But what would we need to observe
empirically? How can we know if a process is self-reinforcing, instead of being reproduced by
exogenous sources? This indeed poses a challenge, but increasing returns and therefore self-
reinforcement processes can occur when we observe the following:
1) High fixed or set-up costs that are amortized with growing numbers and that result in a
situation in which subsequent investments in alternative projects become costlier,
sometimes even ruling them out.
2) Learning and coordination effects due to additional knowledge or experience make the
production process more efficient, so that costs decrease and higher returns are to be
expected.
3) The creation of adaptive expectations refers to the effect of anticipation. Actors form their
future expectations on the base of current information and adapt their behaviour to it, with
self-fulfilling effects. If a person projects, for example, that the use of one technology will
supersede an alternative, then s/he acts in a way that means that this comes true.
4) The network effect refers to the benefits of a network good that increases disproportionally
with the number of users of this good, for example a telephone’s benefit increases from four
and five users by 12 to 20 possible connections.
Let me illustrate this with some empirical examples. The diffusion of a “legitimacy belief” can be
understood as a network good that is mostly connected to high

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 27.
initial start-up costs that trigger subsequent investments and make alternatives costlier, sometimes
even crowding them out. Take the case of North Korea. While in the 1950s, a limited pluralism of
Weltanschauungen was still observable, the rapid spread of the Juche ideology after its first public
mentioning in 1955 is a good case in point. While initial propaganda efforts sidelined and catalyzed
the dissemination of the Juche idea, its spread was at least to some extent self-reinforcing. Adding
new believers as nodes to the network led to a disproportional increase in the network's
countrywide coverage. The possible connections within this belief network increased sharply, leading
to a reduction of marginal “persuasion costs”.69
Repression can also be interpreted as a network good whose dynamic institutionalization process is
self-reinforcing. Analogous to telephones or railways, the regime elite tries to infiltrate and penetrate
society. Following the famous dictum of Mussolini that the party should be the capillary organization
of the regime,70 adding one new regime party office or one security police facility in the country can
increase its benefit for the regime beyond the one unit that is invested. To stick to the North Korean
example, the institutionalization of the repression and surveillance apparatus at the end of the 1950s
can be interpreted as following this network logic of repression. Within a relatively short period of
time, a dense network of prison camps of various types and societal controlling mechanisms were
installed all over the country, which made the repression mechanism – at least temporarily – self-
reinforcing. In addition, coordination effects not just facilitated, but also reinforced, the construction
and maintenance of the North Korean repression apparatus. Secret Service, Secret Police, the
Ministry for Public Security, the Ministry of State Security, and the Ministry of Defence began to
coordinate their work in the 1950s, which led to increasing returns.71
Learning and coordination effects might be detected in autocracies' co-optation efforts. Hlaing
argues persuasively for the Burmese case that the Tatmadaw, the military organization, has
undergone a stabilizing process after its electoral defeat in 1988 and the withdrawal of Ne Win. The
way the co-optation mechanisms have been balanced out and fine-tuned over the last two decades
suggests that a case for path-dependency and organizational learning can be made here as well. The
Burmese junta institutionalized a system of collecting incriminating evidence against military officers
in the event that they were to be disciplined. Moreover, they perfected the rule of “reserving”
discrete domains of operation, which gave officers enough benefit to remain loyal to the regime
while none of them got too powerful.72 In this light, co-optation, repression, and legitimation were
not just exogenously propelled by the regime elite via huge material investments, but their
institutionalization also had internal and structurally in-built reasons.
Reciprocal reinforcement and complementarity between the pillars
Institutions rarely work in isolation, but are part of an institutional ensemble. The relationship
between them can, in general, be neutral, substitutive, conflictive, or

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 28.
complementary. I suggest complementarity between the three pillars. But what does
complementarity mean? The most advanced conceptualizations have been offered by economic
sociologists and comparative political economists.73 Höpner provides an excellent overview of the
research fields that have adopted the concept. He concludes from the various applications that
complementarity refers to the concurrence and co-functioning of institutions that constitute
together a “whole”.74 In addition, I suggest a reciprocal reinforcement between the three pillars.
This means that the existence of one pillar provokes that of another, which in turn strengthens the
first again, etc.75 Reciprocal reinforcement stresses not just the functional interdependence, but also
the mutual strengthening.
In the concepts part of this article, I have touched upon the different functions of the pillars. Gaining
diffuse and specific support from the citizens is the function of legitimation; channelling the demands
from the opposition towards the ruling elite is secured by harder and softer forms of repression; and
lastly, maintaining cohesion and steering capability within the elite is fulfilled through the use of
formal and informal methods of co-optation. How can these functions be complementary and how
can they reinforce reciprocally?
Maintaining elite cohesion is complementary to the function of channelling oppositional demands.
Co-opting potentially deviant elites by using formal and informal instruments reduces the danger of
the emergence of oppositional leaders, while soft and hard repression raises the mobilization costs
for such oppositional figures. We can assume a strong linkage between forms of repression and
forms of co-optation. This also holds true for the relationship between co-optation and legitimation.
Co-optation binds intra-elite actors to the ruling regime and therefore reduces the danger of an
upcoming personal (charismatic) alternative that takes the lead in promoting an alternative
Weltanschauung. Performance as well as ideational legitimation reduces in turn the persuasion
costs, which make co-optation easier.
The Achilles heel can be seen between forms of legitimation and repression. Gaining support
decreases the potential for opposition that would need to be repressed. Reciprocally, repression
seeks to undermine all attempts of oppositional movements for alternative legitimation sources.
However, the linkage between these two pillars is weaker and ambivalent, as repression is a double-
edged sword. Davenport has highlighted the “punishment puzzle”, in which the impact of repression
on public dissent has been shown as being positive, negative, or non-existent.76 Repression comes
with unintended consequences and can weaken the legitimation function, which again brings
repression mounting problems. This can lead to a destabilizing downward spiral. Moreover, a regime
based on strong ideologies might be able to use both forms of repression. The ideology then serves
as the hermeneutic frame and even justifies the use of hard repression. However, hard repression
might prove to be incompatible with performance-based legitimation. Figure 2 displays the
complementarity triangle.
What might we empirically observe if the theoretical claim of reciprocal reinforcement and
complementarity is true? Deeg is right in his assessment that

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 29.
Figure 2. Reciprocal reinforcement and complementarity.
complementarity presumes that the number of successful combinations of elements is limited.77
Certain combinations of pillar specifications “go together”. They should not be randomly distributed,
but should display systematic patterns. To test the complementarity hypothesis empirically, we
would expect that the regimes cluster.
Two stable configurations can be hypothesized to exist. It has been argued above that diffuse
support via ideologies and hard repression have been the two most dominant characteristics in
classic studies. By adding co-optation, particularly via party structures, we would arrive at one
hypothesized stable configuration that leans toward the old totalitarian reasoning: high diffuse
support, high soft and hard repression, and formal co-optation. This configuration can aptly be called
an “over-politicization configuration”. We might, however, be able to overcome the idiosyncratic
tendencies of the totalitarianism paradigm and adequately capture more empirical cases by
simultaneously preserving some of the most valuable insights in understanding the essence of such
an extreme and ideal-typical regime form.
In the spirit of the work by Linz,78 a second path that can be identified is a “de-politicization
configuration”. This configuration would exhibit a high degree of performance orientation that would
demand more subtle and softer forms of repression as well as co-optation by informal channels,
ranging from patronage to clientelistic networks. The result of this stabilizing combination is what
O'Donnell has pointedly described as “low-intensity citizenship”.79 In particular, the rentier state
approach and the literature on developmental authoritarianism

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 30.
that have been reviewed above have highlighted the idea of an unspoken “social contract” between
the performing rulers and those rendered passive.
Empirically showing that these two configurations yield complementarity advantages and therefore
lead to stable configurations is still on the agenda for future research. Two routes for causal
inference can be identified. Firstly, we can rely on counterfactual theories of causation and engage in
deep case studies. We would then need to show that if there is no complementarity, then, ceteris
paribus, the institutions would perform worse and would suffer from lower efficiency gains.80
Secondly, we can increase causal leverage by relying on the regularity theory of causation and could
make use of Cluster Analysis or systematic Qualitative Comparative Analysis (QCA). If we find these
two clusters over a large number of cases, the regularity with which a certain combination is stable
suggests a causal relationship. The “over-politicization” and “under-politicization configuration”
would then constitute the “two worlds of autocracies”.
Concluding remarks
When in 1975 Juan Linz wrote his classic handbook article, he was given so much space as there was
almost no other entry or reference to non-democracies in the whole six-volume project.81 With the
recent renaissance of autocracy studies, this has changed considerably. New insights have been
gained; and while there are still blind spots in understanding autocratic regimes' inner logic, the
academic debate is vital and ongoing. To place this theoretical framework within this discussion was
the aim of the article.
I have argued that autocratic regimes rest on three pillars. Legitimation, repression, and co-optation
have been derived as key by extracting and synthesizing the main arguments of classic and more
recent research efforts. I have paid special attention to re-incorporating legitimation into the studies
of autocratic regimes, a dimension that has recently been considered less relevant. In order to
account for the time-dependent stabilization process, I have suggested closely looking at three
processes: exogenous reinforcement and self-reinforcement within the pillars, and reciprocal
reinforcement between them. In this sense, the proposed theoretical framework aims at integrating
a static view to explain stability by three causal factors, with a dynamic perspective to uncover the
underlying reinforcement mechanisms. To think further along these lines will hopefully prove useful
for future cross-case and within-case comparative work.
Acknowledgements
I would like to thank the three anonymous reviewers, as well as Patrick Köllner, Steffen Kailitz, Jeff
Haynes, and Andreas Schedler for helpful comments and advice. The article has been written within a
research project on “Critical Junctures and the Survival of Dictatorships”, which is funded by the
German Research Foundation. I benefited greatly from

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 31.
discussions with Wolfgang Merkel, Christoph Stefes, and Dag Tanneberg. I am particularly grateful to
Alex Schmotz.
Notes
1. Przeworski et al., “What Makes Democracies Endure”.
2. For example, Linz, “Authoritarian Regime”; Linz, “Totalitarian and Authoritarian Regime”; O'Donnell,
Bureaucratic Authoritarianism.
3. See recently, among other contributions, Wintrobe, Political Economy of Dictatorship; Brownlee,
Authoritarianism in Age of Globalization; Gandhi and Przeworski, “Authoritarian Institutions”; Gandhi,
Political Institutions under Dictatorship; Magaloni, Voting for Autocracy; Svolik, “Power Sharing”.
4. I refer to “autocracy” as the opposite of democracy. It is therefore used as an overarching umbrella term
that encompasses the classic distinction by Linz and others into authoritarian, totalitarian, and post-
totalitarian regimes, as well as more recent classification efforts into personalist, one-party, or military
regimes. See Tullock, Autocracy, 1–17; Merkel, Systemtransformation, 40–54. A “regime” is defined here
with Fishman as the “formal and informal organization of the center of political power, and of its relations
to the broader society”, Fishman, “Rethinking State and Regime”, 428.
5. I use the term “pillar” here. Others might prefer “arena” or “field” or narrow it down to concrete variables
or conditions. In the second part of the article, where I discuss the stabilization process, I argue for an
understanding of the three pillars as institutions.
6. The history of non-democratic regimes is of course as long as the one for democracies, with altering
notions of, for example, tyranny, despotism, absolutism, feudalism, or dictatorships. A short etymological
discussion can be found in Sartori, Theory of Democracy.
7. Friedrich and Brzezinski, Totalitarian Dictatorship
8. The six features are: ideology, mass party, terror system, monopoly of mass communication, monopoly of
force, state-run economy.
9. For a further development see Schapiro, Totalitarianism, 20.
10. Arendt, Origins of Totalitarianism.
11. Maier, “Concepts for the Comparison of Dictatorships”, 205.
12. Barber, “Conceptual Foundations”.
13. Linz, “Authoritarian Regime”; also Linz, “Totalitarian and Authoritarian Regimes”.
14. O'Donnell, Bureaucratic Authoritarianism. A very helpful and critical review is provided by Remmer and
Merckx, “Bureaucratic Authoritarianism Revisited”.
15. See, among numerous others, Collier, The New Authoritarianism. It has been applied to other cases; see
Im, “Bureaucratic Authoritarianism in South Korea”.
16. Easton, A Systems Analysis.
17. See also Perlmutter's attempt to differentiate between the Corporate State Model and the Praetorian
State Model, Perlmutter, Modern Authoritarianism, 38–44.
18. See, for example, Finer, Man on Horseback; Nordlinger, Soldiers in Politics; Eisenstadt, Tradition, Change,
and Modernity; Beblawi and Luciani, The Rentier State.
19. Geddes, “Democratization after Twenty Years”; Merkel, “Are Dictatorships Returning”.
20. For an overview, also see two recent German reviews of the relevant literature: Köllner, “Autoritäre
Regime”; Kailitz, “Autokratieforschung”.
21. Friedrich and Brzezinski, Totalitarian Dictatorship, 47.
22. Gandhi and Przeworski, “Authoritarian Institutions”, 1293; also see Gandhi, Political Institution under
Dictatorship.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 32.
23. Brownlee, Authoritarianism in Age of Democratization; Magaloni, Voting for Autocracy.
24. Schedler, “Elections without Democracy”; Schedler, “Electoral Authoritarianism”; Gandhi and Lust-Okar,
“Elections under Authoritarianism”.
25. Acemoglu and Robinson, Economic Origins of Dictatorship.
26. Gandhi and Przeworski, “Cooperation, Cooptation and Rebellion”; Magaloni, “Credible Power-Sharing”;
Svolik, “Power Sharing”.
27. I have proposed a complementary argument as to why the three pillars are crucial for the stability of
autocracies by referring to actors' motivation. While in the legitimation pillar, the main motive for the
people to support the regime is a sense of righteousness, or at least acquiescence; repression builds on
fear as the motive for potential opposition. The motive for elites to be co-opted is a utilitarian cost-benefit
calculus. See Gerschewski et al., “Warum überleben Diktaturen?”.
28. The most prominent and influential critique stems from Przeworski, Democracy and the Market.
29. Weber, Wirtschaft und Gesellschaft, 122.
30. Kielmannsegg, “Legitimität”, 389.
31. Wittfogel, Oriental Despotism.
32. See Meuschel, Legitimation und Parteiherrschaft, 10–28.
33. See, for example, Heydemann, Authoritarianism in Syria, 30–54; Albrecht and Schlumberger, “Waiting for
Godot”, 376–378.
34. Magaloni, Voting for Autocracy.
35. Holbig and Gilley, “Legitimacy in China”.
36. On the concept of “ideocracies”, see Piekalkiewicz and Penn, Politics of Ideocracy; on the concept of
“theocracies”, please refer to, for example, Ferrero and Wintrobe, Political Economy of Theocracy.
37. See, for example, Lankov, “Death of North Korean Stalinism”.
38. See also the work by Linz and Stepan and their excellent description of the erosion of totalitarianism to
post-totalitarian forms of rule; Linz and Stepan, Democratic Transition, 42–50.
39. For the economic performance, see White, “Economic Performance and Communist Legitimacy”; and
Scharpf, Governing in Europe for performance orientation.
40. Easton, A Systems Analysis, 268.
41. Easton, A Systems Analysis, 444.
42. For a discussion of these sources of legitimation, please refer to Burnell, “Autocratic Opening”; Beetham,
Legitimation of Power.
43. There is one caveat in using these indicators: while modernization theory would argue that absolute levels
of socio-economic indicators measure an increase in the likelihood of democratization, relative indicators
that measure growth or decline might be more apt for measuring specific support in autocracies.
44. See The World Bank, World Development Indicators.
45. Banks and Wilson, Cross-National Times-Series Data Archive.
46. Pioneering work is currently done by Kailitz, who typologizes political systems along their legitimation
strategy; see Kailitz, “Classification by Legitimation”.
47. Protest movements can be framed as a collective action problem; see Chwe, “Collective Action”.
48. Davenport, “State Repression”, 2.
49. Levitsky and Way, “Competitive Authoritarianism”.
50. See Freedom House, “Freedom in the World”; Cingranelli and Richards, “CIRI Human Rights Dataset”;
Gibney, Cornett, and Wood, “Political Terror Scale”.
51. Bueno de Mesquita et al., The Logic of Political Survival.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 33.
52. Shleifer and Treisman, Without a Map, 8–9; Bertocchi and Spagat, “Politics of Co-optation”.
53. The concept of co-optation is limited here to the elite level. Mass co-optation that is directed towards
broader parts of the population is subsumed here as a form of generating specific support.
54. Gandhi and Przeworski, “Cooperation, Cooptation, and Rebellion”; Gandhi and Przeworski, “Authoritarian
Institutions”.
55. Eisenstadt, Tradition, Change, and Modernity.
56. Stefes, Understanding Post-Soviet Transitions.
57. Gandhi and Przeworski, “Authoritarian Institutions”.
58. These indicators can be found in Banks and Wilson, Cross-National Times-Series Data Archive; Beck, Keefer,
and Clarke, “Database of Political Institutions”.
59. Beck, Keefer, and Clarke, “Database of Political Institutions”; Banks and Wilson, Cross-National Times-
Series Data Archive; Alesina et al., “Fractionalization”; Fearon, “Ethnic and Cultural Diversity”.
60. Wimmer, Cedermann, and Min, “Ethnic Politics”.
61. Alexander Schmotz is working on a more complex measurement of cooptation. See Schmotz, “Co-
optation”.
62. Still the best overview of the three forms of neo-institutionalism is given by Hall and Taylor, “Three New
Institutionalism”; for the historical institutionalism, please refer to Thelen, “Historical Institutionalism”.
63. Huntington, Political Order, 12.
64. Hall, Governing the Economy, 19. See, for a similar definition, North, Institutions, 3.
65. I am grateful to Kathleen Thelen who has suggested this metaphor to me.
66. Mahoney, “Path Dependence”, 517.
67. Mahoney identifies a fourth mechanism based on functionalism and systems theory. This mechanism is
more to be seen as an overarching mechanism for the stability of the autocratic systems as a whole and
does not correspond to the institutionalization process within the pillar. Furthermore, Mahoney sees
“reactive sequences” as path-dependent, in which chains of events can be traced back. To what extent
these sequences exhibit increasing returns is unclear. Increasing returns are to be seen, however, as a
necessary condition for path-dependency, so that reactive sequences will be left out here, being aware
that they might feature increasing returns.
68. David, “Economics of QWERTY”; Arthur, Increasing Returns; North, Institutions. Paul Pierson has made this
argument more popular for social sciences; see Pierson, “Increasing Returns”. For excellent discussions on
path-dependency, please refer to Beyer, “Mechanisms of Path Dependence”; Rixen and Viola, “Taxonomy
of Institutional Change”.
69. See Park, “Nature and Evolution of Juche Ideology”; Suh, Leadership and Political Culture in North Korea,
31–53.
70. Gentile, “Party in Italian Fascism”, 263.
71. See Szalontai, Kim Il Sung in the Khrushchev Era. For today's repression apparatus, please refer to Haggard
and Noland, Witness to Transformation, 81–100.
72. Hlaing, “Rules for Survival”.
73. The debate has been intensified by Hall and Soskice, Varieties of Capitalism.
74. Höpner, “Complementarity”, 383–384.
75. Crouch has pointed to such a relationship; see Crouch, “Three Meanings”, 362.
76. Davenport, “State Repression”, 8.
77. Deeg, “Complementarity and Institutional Change”, 612.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 34.
78. See Linz, “Authoritarian Regime”, 304. Linz argued then that the Spanish regime elite tried to depoliticize
public life to maintain stability.
79. O'Donnell, Bureaucratic Authoritarianism, 143.
80. Deeg, “Complementarity”, 614.
81. Linz, “Further Reflections”.
Notes on contributor
Johannes Gerschewski is a Research Fellow at the Research Unit “Democracy: Structures,
Performance, Challenges” at the Social Science Research Center Berlin (WZB), a PhD candidate at the
Berlin Graduate School of Social Sciences (BGSS), Humboldt-Universität zu Berlin, and an Associated
Research Fellow at the German Institute of Global and Area Studies, Hamburg. His research interest
lies in the comparative study of autocracies and democratization processes, with a special focus on
the East Asian region.
Bibliography
Acemoglu, Daron, and James A. Robinson. Economic Origins of Dictatorship and Democracy. Cambridge:
Cambridge University Press, 2005.
Albrecht, Holger, and Oliver Schlumberger. “Waiting for Godot. Regime Change Without Democratization in the
Middle East.” International Political Science Review 25 (2004): 371–392.
Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat, and Romain Wacziarg.
“Fractionalization.” Journal of Economic Growth 8 (2003): 155–194.
Arendt, Hannah. The Origins of Totalitarianism. New York: Schocken, 1951.
Arthur, Brian W. Increasing Returns and Path Dependence in Economics. Ann Arbor: University of Michigan
Press, 1994.
Banks, Arthur S., and Kenneth A. Wilson. “Cross-National Time-Series Data Archive.” (2012).
http://www.databankinternational.com.
Barber, Benjamin. “Conceptual Foundations of Totalitarianism.” In Totalitarianism in Perspective: Three Views,
edited by Carl Joachim Friedrich, Michael Curtis, and Benjamin Barber, 3–52. London: Pall Mall, 1969.
Beblawi, Hazem, and Giacomo Luciani, eds. The Rentier State. Kent: Istituto Affari Internazionali, 1987.
Beck, Thorsten, Philip E. Keefer, and George R. Clarke. “New Tools in Comparative Political Economy: The
Database of Political Institutions.” World Bank Economic Review 15 (2001): 165–176.
Beetham, David. The Legitimation of Power. Houndmills: Palgrave, 1991.
Bertocchi, Graziella, and Michal Spagat. “Politics of Co-optation.” Journal of Comparative Economics 29 (2001):
591–607.
Beyer, Jürgen. “The Same or Not the Same – On the Variety of Mechanisms of Path Dependence.” International
Journal of Social Science 5 (2010): 1–11.
Brownlee, Jason. Authoritarianism in an Age of Democratization. Cambridge: Cambridge University Press, 2007.
Bueno De Mesquita, Bruce, Alastair Smith, Randolph M. Siverson, and James D. Morrow. The Logic of Political
Survival. Cambridge, MA: The MIT Press, 2003.
Burnell, Peter. “Autocratic Opening to Democracy. Why Legitimacy Matters.” Third World Quarterly 27 (2006):
545–562.
Chwe, Michael Suk-Young. “Structure and Strategy in Collective Action.” American Journal of Sociology 105
(1999): 128–156.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 35.
Cingranelli, David L., and David L. Richards. “The Cingranelli-Richards Human Rights Dataset Version
2011.12.09.” (2011). http://www.humanrightsdata.org.
Collier, David, ed. The New Authoritarianism in Latin America. Princeton, NJ: Princeton University Press, 1979.
Crouch, Colin. “Three Meanings of Complementarity.” Socio-Economic Review 3 (2005): 359–363.
Davenport, Christian. “State Repression and Political Order.” Annual Review of Political Science 10 (2007): 1–23.
David, Paul A. “Clio and the Economics of QWERTY.” American Economic Review 75 (1985): 332–337.
Deeg, Richard. “Complementarity and Institutional Change in Capitalist Systems.” Journal of European Public
Policy 14 (2007): 611–630.
Easton, David. A Systems Analysis of Political Life. Chicago: Chicago University Press, 1979.
Eisenstadt, Shmuel Noah. Tradition, Change, and Modernity. New York: Wiley, 1973.
Fearon, James D. “Ethnic and Cultural Diversity by Country.” Journal of Economic Growth 8 (2003): 195–222.
Ferrero, Mario, and Ronald Wintrobe, eds. The Political Economy of Theocracy. New York: Palgrave Macmillan,
2009.
Finer, S. E. The Man on Horseback. The Role of Military in Politics. London: Pall Mall, 1962.
Fishman, Robert M. “Rethinking State and Regime: Southern Europe's Transition to Democracy.” World Politics
42 (1990): 422–440.
Freedom House. “Freedom in the World 2012.” (2012). http://www.freedomhouse.org/report/freedom-
world/freedom-world-2012 .
Friedrich, Carl Joachim, and Zbigniew Brzezinski. Totalitarian Dictatorship and Autocracy. New York: Praeger,
1968.
Gandhi, Jennifer. Political Institutions under Dictatorship. Cambridge: Cambridge University Press, 2008.
Gandhi, Jennifer, and Ellen Lust-Okar. “Elections under Authoritarianism.” Annual Review of Political Science 14
(2009): 403–422.
Gandhi, Jennifer, and Adam Przeworski. “Authoritarian Institutions and the Survival of Autocrats.” Comparative
Political Studies 40 (2007): 1279–1301.
Gandhi, Jennifer, and Adam Przeworski. “Cooperation, Cooptation, and Rebellion under Dictatorships.”
Economics & Politics 18 (2006): 1–26.
Geddes, Barbara. “What Do We Know About Democratization after Twenty Years?” Annual Review of Political
Science 2 (1999): 115–144.
Gentile, Emilio. “The Problem of the Party in Italian Fascism.” Journal of Contemporary History 19 (1984): 251–
274.
Gerschewski, Johannes, Wolfgang Merkel, Alexander Schmotz, Christoph Stefes, and Dag Tanneberg. “Warum
überleben Diktaturen?” Politische Vierteljahresschrift, forthcoming
Gibney, Mark, Linda Cornett, and Reed Wood. “Political Terror Scale 1976–2006.” (2012)
http://www.politicalterrorscale.org
Hall, Peter A. Governing the Economy. The Politics of State Intervention in Britain and France. New York: Oxford
University Press, 1986.
Hall, Peter A, and Rosemary C. Taylor. “Political Science and the Three New Institutionalism.” Political Studies
44 (1996): 936–957.
Hall, Peter A., and David Soskice, eds. Varieties of Capitalism: The Institutional Foundations of Comparative
Advantage. Oxford: Oxford University Press, 2001.
Haggard, Stephen, and Marcus Noland. Witness to Transformation. Refugee Insights into North Korea.
Washington, DC: Peterson Institute for International Economics, 2001.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 36.
Heydemann, Steven. Authoritarianism in Syria. Institutions and Social Conflict, 1946–1970. Ithaca, NY: Cornell
University Press, 1999.
Hlaing, Kyaw Yin. “Setting the Rules for Survival. Why the Burmese Military Regime Survives in an Age of
Democratization.” The Pacific Review 22 (2009): 271–291.
Höpner, Martin. “Epilogue to ‘Explaining Institutional Complementarity.’” Socio-Economic Review 3 (2005):
383–388.
Holbig, Heike, and Bruce Gilley. “Reclaiming Legitimacy in China.” Politics and Policy 38 (2010): 395–422.
Huntington, Samuel. Political Order in Changing Societies. London: Yale University Press, 1968.
Im, Hyug Baeg. “The Rise of Bureaucratic Authoritarianism in South Korea.” World Politics 39 (1987): 231–257.
Kailitz, Steffen. “Classification by Legitimation.” Paper presented at the 6th ECPR General Conference,
Reykjavik, August 26, 2011.
Kailitz, Steffen. “Stand und Perspektiven der Autokratieforschung.” Zeitschrift für Politikwissenschaft 19 (2009):
437–488.
Kielmannsegg, Peter Graf. “Legitimität als Analytische Kategorie.” Politische Vierteljahresschrift 12 (1971): 360–
401.
Köllner, Patrick. “Autoritäre Regime – Ein Überblick über die jüngere Literatur.” Zeitschrift für Vergleichende
Politikwissenschaft 2 (2008): 351–366.
Lankov, Andrei. “The Natural Death of North Korean Stalinism.” Asia Policy (2006): 95–121.
Levitsky, Steven, and Lucan A. Way. “Elections without Democracy: The Rise of Competitive Authoritarianism.”
Journal of Democracy 13 (2002): 51–65.
Linz, Juan J. “An Authoritarian Regime. The Case of Spain.” In Cleavages, Ideologies, and Party Systems.
Contributions to Comparative Political Sociology, edited by Erik Allardt and Irgo Littunen, 291–342. Helsinki:
The Academic Bookstore, 1964.
Linz, Juan J. “Further Reflections on Totalitarian and Authoritarian Regimes.” In Totalitarian and Authoritarian
Regimes, edited by Juan J. Linz, 1–48. Boulder, CO: Lynne Rienner, 2000.
Linz, Juan J. “Totalitarian and Authoritarian Regimes.” In Handbook on Political Science, edited by Fred I.
Greenstein and Nelson W. Polsby, 175–411. Reading, MA: Addison Wesley, 1975.
Linz, Juan J., and Alfred Stepan. Problems of Democratic Transition and Consolidation. Southern Europe, South
America, and Post-Communist Europe. Baltimore, MD: The Johns Hopkins University Press, 1996.
Magaloni, Beatriz. “Credible Power-Sharing and the Longevity of Authoritarian Rule.” Comparative Political
Studies 41 (2008): 715–741.
Magaloni, Beatriz. Voting for Autocracy. Hegemonic Party Survival and Its Demise in Mexico. Cambridge:
Cambridge University Press, 2006.
Mahoney, James. “Path Dependence in Historical Sociology.” Theory and Society 29 (2000): 507–548.
Maier, Hans. “Concepts for the Comparison of Dictatorships. ‘Totalitarianism’ and ‘Political Religion’”. In
Totalitarianism and Political Religions. Concepts for the Comparison of Dictatorships, vol. 1, edited by Hans
Maier, 199–215. New York: Routledge, 2004.
Merkel, Wolfgang. “Are Dictatorships Returning? Revisiting the ‘Democratic Rollback’ Hypothesis.”
Contemporary Politics 16 (2010): 17–31.
Merkel, Wolfgang. Systemtransformation. Eine Einführung in die Theorie und Empirie der
Transformationsforschung. Wiesbaden: VS Verlag für Sozialwissenschaften, 2010.
Meuschel, Sigrid. Legitimation und Parteiherrschaft in der DDR. Frankfurt a.M.: Suhrkamp, 1991.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 37.
Nordlinger, Eric A. Soldiers in Politics. Military Coups and Governments. Englewood Cliffs, NJ: Prentice Hall,
1977.
North, Douglass C. Institutions, Institutional Change and Economic Performance. New York: Cambridge
University Press, 1990.
O'Donnell, Guillermo. Modernization and Bureaucratic Authoritarianism. Studies in South American Politics.
Berkeley: University of California Press, 1979.
Park, Han-Shik. “The Nature and Evolution of Juche Ideology.” In North Korea. Ideology, Politics, Economy,
edited by Han-Shik Park, 9–18. Englewood Cliffs, NJ: Prentice Hall, 1996.
Perlmutter, Amos. Modern Authoritarianism: A Comparative Institutional Analysis. New Haven, CT: Yale
University Press, 1981.
Piekalkiewicz, Jaroslaw, and Alfred Wayne Penn. Politics of Ideocracy. Albany: State University of New York
Press, 1995.
Pierson, Paul. “Increasing Returns, Path Dependence, and the Study of Politics.” The American Political Science
Review 94 (2000): 251–267.
Przeworski, Adam. Democracy and the Market. Political and Economic Reforms in Eastern Europe and Latin
America. Cambridge: Cambridge University Press, 1991.
Przeworski, Adam, Michael Alvarez, José Antonio Cheibub, and Fernando Limongi. “What Makes Democracies
Endure?” Journal of Democracy 7 (1996): 39–55.
Remmer, Karen L., and Gilbert W. Merckx. “Bureaucratic Authoritarianism Revisited.” Latin American Research
Review 17 (1982): 3–40.
Rixen, Thomas, and Lora Viola. “Toward a Taxonomy of Institutional Change.” Unpublished manuscript, 2012.
Sartori, Giovanni. A Theory of Democracy Revisited. Chatham: Chatham House Publisher, 1987.
Schapiro, Leonard. Totalitarianism. London: Praeger, 1972.
Scharpf, Fritz W. Governing in Europe. Oxford: Oxford University Press, 1999.
Schedler, Andreas. “Elections without Democracy. The Menu of Manipulation.” Journal of Democracy 13
(2002): 36–50.
Schedler, Andreas. “The Logic of Electoral Authoritarianism.” In Electoral Authoritarianism. The Dynamics of
Unfree Competition, edited by Andreas Schedler, 1–23. Boulder, CO: Lynne Rienner, 2006.
Schmotz, Alexander. “Measuring Co-optation.” Unpublished manuscript, 2012.
Shleifer, Andrei, and Daniel Treisman. Without a Map. Political Tactics and Economic Reform in Russia.
Cambridge: Cambridge University Press, 2000.
Stefes, Christoph H. Understanding Post-Soviet Transitions. Corruption, Collusion and Clientelism. New York:
Palgrave Macmillan, 2006.
Suh, Dae-Sook. Leadership and Political Culture in North Korea. Seoul: Yonsei University Press, 2000.
Svolik, Milan W. “Power Sharing and Leadership Dynamics in Authoritarian Regimes.” American Journal of
Political Science 53 (2009): 477–494.
Szalontai, Balácz. Kim Il Sung in the Khrushchev Era. Soviet-DPRK Relations and the Roots of North Korean
Depotism 1953-1964. Washington, DC: Woodrow Wilson Center Press, 2005.
Thelen, Kathleen A. “Historical Institutionalism in Comparative Politics.” Annual Review of Political Science 2
(1999): 369–404.
Tullock, Gordon. Autocracy. Lancaster: Kluwer Academic Publisher, 1987.
Weber, Max. Wirtschaft und Gesellschaft. Grundriß der verstehenden Soziologie. Tübingen: Mohr, 1956.
White, Stephen. “Economic Performance and Communist Legitimacy.” World Politics 38 (1986): 462–482.

Originally published in:
Democratization, Vol. 20 (2013) Iss. 1, pp. 38.
Wimmer, Andreas, Lars-Erik Cederman, and Brian Min. “Ethnic Politics and Armed Conflict. A Configurational
Analysis of a New Global Dataset.” American Sociological Review 74 (2009): 316–337.
Wintrobe, Ronald. The Political Economy of Dictatorships. Cambridge: Cambridge University Press, 1998.
Wittfogel, Karl. Oriental Despotism. A Comparative Study of Total Power. New York: Random House, 1957.
World Bank. World Development Indicators 2011. Washington, DC: World Bank, 2011.