PreprintPDF Available

Durability not Diversity: Uncovering the Organizational Roots of Democratization

Authors:
Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

We refine and test theoretical mechanisms linking mass mobilization to democratization by focusing on variation in the organizations that participate in collective dissent. Specifically, we investigate the effects of organizational diversity and durability on the likelihood of democratization. Using new data on maximalist claims-making organizations that engaged in resistance events in Africa from 1990-2015, we find little evidence that organizational diversity on its own improves democratization, which we link to what we call the "diversity dilemma." While greater movement diversity may increase democratic preferences, it undermines movement capacity to realize these preferences by increasing collective action problems and reducing a movement's ability to make credible commitments. In contrast, the participation of durable organizations such as trade unions and religious organizations significantly increases longer-term democratization prospects, which we argue reflects their enduring pro-democratic preferences and ability to credibly threaten re-mobilization during a transition. Our findings have important implications for our understanding of democratization and mass mobilization. Word Count: 9841
Content may be subject to copyright.
Durability not Diversity
Durability Not Diversity: Uncovering the Organizational Roots of
Democratization
JONATHAN PINCKNEY NTNU
CHARLES BUTCHER NTNU
JESSICA MAVES BRAITHWAITE University of Arizona
We
refine and test theoretical mechanisms linking mass mobilization to democra-
tization by focusing on variation in the organizations that participate in collective
dissent. Specifically, we investigate the effects of organizational diversity and
durability on the likelihood of democratization. Using new data on maximalist claims-making
organizations that engaged in resistance events in Africa from 1990-2015, we find little
evidence that organizational diversity on its own improves democratization, which we link
to what we call the “diversity dilemma.” While greater movement diversity may increase
democratic preferences, it undermines movement capacity to realize these preferences by
increasing collective action problems and reducing a movement’s ability to make credible
commitments. In contrast, the participation of durable organizations such as trade unions
and religious organizations significantly increases longer-term democratization prospects,
which we argue reflects their enduring pro-democratic preferences and ability to credibly
threaten re-mobilization during a transition. Our findings have important implications for
our understanding of democratization and mass mobilization.
Word Count: 9841
The authors would like to thank Felix Bethke, Kristian Skrede Gleditsch, Nikolay Marinov, and all those
who attended our panel at the 2018 American Political Science Association meeting, the 2019 International
Studies Association meeting and our online presentation as part of the "Securing the Victory" project for helpful
comments on earlier versions of this paper. The authors are varying their name order as part of a series of
projects. Equal authorship implied. This research was supported in part by a grant from the United States
Institute of Peace.
This is a working paper
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
INTRODUCTION
Re
cent studies have found that popular mobilization in autocracies, particularly nonviolent
mobilization, correlates with democratic transitions.
1
In contrast to an earlier literature,
which focused on the elite origins of democracy (O’Donnell and Schmitter 1986; Higley
and Burton 1989), this research indicates a path to democracy through mass mobilization, carried
forward by resistance movements. Yet many of the problematic transitions of the last several decades
show that this path is indirect and rocky. For one thing, a resistance movement’s opposition to autocracy
is no guarantee of its democratic preferences. In the post-World War II era, almost every major
political transformation has been undertaken in the name of democracy. Yet this claim frequently
falls by the wayside once the claimants are in power. Even if a resistance movement is led by genuine
democrats, these leaders may lack the capacity to carry the democratizing impulse through a political
transition, or the political strength to resist attempts at re-autocratization by anti-democratic forces.
To know when mass mobilization is likely to lead to democracy, we must therefore have insight into
two questions: when will ostensibly-democratic resistance movements be characterized by genuine
democratic preferences? And when will those with genuine democratic preferences have the capacity
to see those preferences realized?
Our novel theoretical entry point to answer these questions is to derive democratic preferences and
capacity from the organizations that comprise resistance movements. While many resistance movements
claim to represent an underlying popular will, this will is rarely manifested without organizational
mediation. Parsing resistance movements into their participating organizations and comparing their
patterns of participation provides crucial new insight into their behaviors and consequences. The
particular constellation of participating organizations in mobilization significantly shapes the incentive
structure that shifts mass mobilization towards or away from democratic preferences and offers the
capacity to carry out those preferences.
Our first argument deals with the organizational character of resistance movements as a whole.
1
There is an expansive and growing literature that examines this subject. See, for example Bayer et al. (2016);
Bethke and Pinckney (2016); Brancati (2016); Chenoweth and Stephan (2011); Haggard and Kaufman (2016);
Teorell (2010)
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
Existing literature on mass mobilization has emphasized movement diversity as a key factor in
encouraging democratic preferences. We find these arguments plausible but incomplete, since diverse
movements (those with many different organizations) also likely face difficulty in making credible
commitments and may fragment after an initial breakthrough, leading to an increased risk of democratic
backsliding. We thus identify a “diversity dilemma:” Mass mobilization campaigns with many
participating organizations may have more democratic preferences on average, but will have less
capacity to ensure that those preferences are actualized.
Our second argument focuses on the type of participating organizations in resistance movements. We
find strong theoretical rationales that the participation of organizations rooted in quotidian relationships
to a mass constituency and not designed themselves to compete for political power, such as labor unions
or religious organizations, will lead to greater democracy (what we call the “durability advantage”).
These organizations should prefer democracy on average because it typically provides greater autonomy
for their actions and a relatively low-cost mechanism of holding decision-makers accountable through
the ballot box. Their strong links to identifiable constituencies should also give them greater capacity
to maintain mobilization and resist attempts at re-autocratization through political transitions, giving
them greater capacity to ensure that democratization takes place.
We test these arguments using a novel dataset of organizations participating in violent and nonviolent
resistance movements in Africa from 1990-2015. Our data is one of the first sources to perform this
disaggregation and comparison on a broad temporal and geographic scale. We find support for both
the ambiguous impact of diversity and the strong positive influence of durability. The overall structure
of movements, as measured by their number of organizational participants, has little impact of future
levels of democracy. The participation of durable organizations in resistance, however, has strong and
consistently positive impacts on democratization in both the short and medium term.
The paper proceeds as follows. First, we outline the theoretical underpinnings of our focus on
organizational participation in resistance movements. Second, we present the arguments for our two
central hypotheses: the diversity dilemma and the durability advantage. Third, we introduce our
research design, particularly focusing on our novel data on organizations in resistance movements.
Fourth, we present findings showing that the participation of durable organizations in resistance
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
movements correlates positively with democracy up to 5 years into the future, while there is little
evidence that organizational diversity improves democracy. The final section concludes.
THEORY
Organizations in Resistance - Our Theoretical Priors
There is growing consensus that resistance “from below” plays a key role in democratization (Haggard
and Kaufman 2016; Teorell 2010). Protests that are large, enduring (Teorell 2010; Brancati 2016;
Kadivar 2018), and nonviolent (Bayer et al. 2016; Celestino and Gleditsch 2013; Chenoweth and
Stephan 2011; Karatnycky and Ackerman 2005; Schock 2005) appear to enhance the prospects of
democratization compared to armed conflict or no mobilization. Other studies emphasize that the
mobilization of specific social organizations or groups are critical to democratization. Collier (1999)
identified the importance of labor unions or labor-connected parties in driving democratization in
South America, Europe and Southeast Asia. Labor organizations also play a foundational role in
Rueschemeyer et al. (1992), Bermeo (2003) and Bratton and van de Walle (1997) and working-class
insurgency is central to the account of democratization in Wood (2000). Recently, Dahlum et al. (2017)
have identified a correlation between the mobilization of industrial workers and democracy, and Butcher
et al. (2018) show that the participation of organized labor improved the prospects for democratization,
on average, across 123 nonviolent campaigns. Peasant-middle class alliances also feature prominently
as explanations in Moore (1993) and more recently, Haggard and Kaufman (2016) argued that labor
unions have been a key player in so-called “distributive” democratic transitions. Many structural
accounts of democratization also include causal mechanisms that hinge on the threat or manifestation
on mass mobilization (Acemoglu and Robinson 2005; Boix 2003; Lipset 1959; Przeworski 1991).
Yet resistance movements making demands for regime change in autocracies also do not inherently
have enduring democratic preferences. Individuals, organizations, and communities seek regime
change in autocracies for diverse and contingent reasons which may be only tangentially related to an
affinity for democracy (Brancati and Snyder 2013). In many cases, ostensibly democratic transitions
simply involve elite circulation, with only the “semblance” of democratic transformation (Beissinger
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
2013). While opposition figures may articulate democratic preferences when out of power, flawed
transitions of the last several decades feature many of these figures engaging in the same or even
worse anti-democratic behavior once in power. For example, while in opposition, Zambian trade union
leader Frederick Chiluba articulated a strong belief in the value of democracy and was imprisoned
for his work fighting against Zambia’s single party authoritarian regime (Larmer 2016). However,
after his leadership of a nonviolent resistance movement led to his election as president of Zambia,
Chiluba orchestrated a significant closing of civic space and sought to indefinitely extend his tenure
through changing the Zambian constitution (Van Donge 2002). Thus, we cannot assume the democratic
preferences of ostensibly pro-democratic movements. Instead, it is necessary to theoretically motivate
when these preferences are more likely to emerge and when movements are likely to have the capacity
to compel a regime to respond to them (Dahlum et al. 2017).
Theoretically motivating democratic preferences in resistance movements raises the more basic
question of clearly defining democracy, since we must have a clear sense of what movements may or
may not prefer. This is difficult, as democracy is multidimensional, with some dimensions orthogonal
to or even in conflict with one another (Coppedge et al. 2018). A “thick” conception of democracy can
include institutions as diverse as free and fair elections, protections for minority groups and human
rights generally, proportional representation, an independent judiciary, and many more. To theorize
about democratic preferences and capacity in resistance movements, we simplify the definition of
democracy to the commonly-used threshold for a democratic system: one in which leaders achieve
power through free and fair elections (Schumpeter 2010), and in which there is relative certainty in the
institutions for alternation in power but uncertainty as to the outcome of those alternation procedures
(Przeworski 1991). This is the standard threshold definition of democracy in the democratization
literature, and informs much of the quantitative work on democratization.2
Based on this definition, we expect mass mobilization to lead to democracy in circumstances
where major political actors disagree over specific policy agendas but have relative consensus around
establishing a set of institutional rules to regulate their competition, specifically through free and
2
For example, with slight variations, this is the standard underlying the major cross-national datasets that define
democracy as a binary variable. See Boix et al. (2013); Cheibub et al. (2010); Geddes et al. (2014); Przeworski
et al. (2000)
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
fair elections. In addition, resistance movements should have the capabilities to impose sufficient
political pressure to bring about this arrangement. How might we disaggregate “bottom up” resistance
movements to gain insight into when these preferences and capabilities are likely to emerge? Previous
studies have focused on the distinction between violent movements and nonviolent movements (Bayer
et al. 2016; Chenoweth and Stephan 2011; Pinckney 2018). Yet these two categories have indistinct
boundaries, and contain significant variation.
We thus seek to deduce democratic preferences and capabilities from a different characteristic
of resistance movements: their organizational makeup. Our theoretical starting point is that violent
and nonviolent resistance movements consist of a network of organizations and social groups (Diani
2003; Dahlum et al. 2017; Staniland 2014; Weinstein 2006; Metternich et al. 2013; Leventoğlu and
Metternich 2018). No resistance movement against autocracy is monolithic. Neither are resistance
movements typically spontaneous, leaderless uprisings. While some uprisings appear spontaneous
they are usually mobilized through pre-existing organizations or quickly develop an organizational
infrastructure (Haggard and Kaufman 2016; McAdam 2010; Butcher et al. 2018). Organizations
formulate and coordinate resistance strategies, bargain with authoritarian regimes, assume power in
transitional periods and re-mobilize in the face of threats of re-autocratizaton. For example, the Tunisian
General Labor Union (UGTT) and the Tunisian Bar Association won a Nobel prize for their role in the
2011 revolution and democratization process in Tunisia. Organizations are also the primary way in
which individuals with shared preferences pool resources - a powerful reason why public preferences
are often expressed through these formalized groups.
Conceptualizing resistance movements as interacting sets of organizations suggests two avenues
through which to examine democratic preferences and capacity: a movement’s overall organizational
structure and the characteristics of its individual constituent organizations. The overall organizational
structure of a movement considers the number of organizations and their relative positions to one another.
Examining individual organizations means looking at characteristics of individual organizations that
should shape the movement as a whole when these organizations participate and to a greater degree
when they become a larger proportion of the movement. We first look at movement structure, and then
turn to individual group characteristics.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
The Diversity Dilemma
The number of organizations that make up a resistance movement is the most basic characteristic of
its organizational structure. Movements can range from being comprised of very few organizations
(though rarely if ever a single organization) to a large number of organizations. We describe this
characteristic of a movement as the movement’s “diversity.”
Higher diversity may be one mechanism linking mass mobilization to democratization, particularly
by encouraging democratic preferences. First, diverse movements often develop consensus-based
forms of decision-making that develop into enduring democratic norms. In resistance movements
composed of many organizations, coordinating and accommodating various interests becomes common
and familiar practice. For example, during the Velvet Revolution in Czechoslovakia - in which a wide
variety of groups participated - dissidents held daily “plenums” where tactical decisions were made
only after receiving buy-in by students, workers, intellectuals, and representatives of both the Czech
and Slovak components of the movement (Ash 2014). These types of power-sharing arrangements
prefigure the cooperative arrangements necessary for successful democratic politics and governance.
Diverse movements should also have aggregate preferences for democracy because any single
constituent organization has a small chance of capturing state power, yet is vulnerable to state repression.
This increases the attractiveness of institutions with checks on executive power and the ability to
regularly revise who has access to such power (Bueno de Mesquita et al. 2005; Acemoglu and Robinson
2005; Gleditsch and Ward 2006; Celestino and Gleditsch 2013; Przeworski 1991). Important tenets
of democracy such as rule of law, separation of powers, and regular elections limit the short-term
advantages of capturing the state. They thus can help prevent the emergence of a new concentration of
power around a single actor and make it more likely that those temporarily out of power will be able to
achieve it at some point in the future.
In contrast, when campaigns consist of only one type of organization, or feature a single dominant
organization, their preferences are likely to be dominated by the ideological position of that organization
type. Dedication to this vision may trump any particular commitment to the norms and institutions of
democracy, which require compromise. For example, in some civil wars rebel groups will have both
practical and ideological incentives to close off competition to other opponents if they ever capture
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
state power. Paul Kagame of Rwanda provides a prime example of this dynamic, consolidating power
around himself in an authoritarian manner following a victorious campaign spearheaded solely by his
Rwandan Patriotic Front.
Organizational diversity can therefore create latent preferences for democracy. Furthermore, if
diverse movements are also big and have greater opportunities for tactical innovation and adaptation
(Schock 2005; Chenoweth and Stephan 2011), then this may translate into greater levels of future
democracy. However, other mechanisms in the literature on civil war suggest reasons to be skeptical of
a general diversity-democracy link. We think of diversity as a “dilemma” that confronts resistance
movements. On one hand there are strong incentives to build big, diverse coalitions capable of
generating mass participation, imposing significant costs on the regime and inducing elite defections
(Schock 2005; Leventoğlu and Metternich 2018; DeNardo 2014; Chenoweth and Stephan 2011). On the
other hand, diversity usually creates bargaining problems. As more groups with heterogeneous interests
become involved in resistance movements, it becomes difficult for the regime and opposition to locate
satisfactory bargains for all constituencies. Bargaining breaks down more often in these circumstances
making it more likely that repression becomes an attractive strategy for the regime. Studies of civil war
have shown that rebel group fragmentation lengthens wars and reduces the likelihood of negotiated
settlements (Cunningham 2006, 2013b). Diversity also makes it more difficult for the opposition to
credibly commit to upholding agreements that might be made with the regime as internal factional
splintering may lead to continued resistance (Cunningham 2013a).
Moreover, all else being equal, diversity and its attendant democratic preferences do not necessarily
provide a movement with the ability to force autocratic leaders to democratize, and may actually have
the opposite effect. Although Acemoglu and Robinson (2005) see democracy as a solution to the
commitment problem implied by economic redistribution, promises of democratic reform entail their
own commitment problems because they are implemented over a period of time. New constitutions can
take years to be written and elections are not typically held until one or two years after initial democratic
breakthroughs. A promise to liberalize today, when the opposition is strong, may not be credible in two
years time if the opposition weakens. Contemporary examples of attempted re-autocratization include
Egypt (a successful coup in 2013) and Burkina Faso (a failed coup in 2015). Diverse coalitions may
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
have special problems navigating this transitional period and maintaining the ability to force and sustain
democratic reforms. While diverse coalitions may be able to maintain unity over a shared short-term
goal such as regime change, bargaining over the procedural issues that must be negotiated during any
transitional period can cause splits (Beissinger 2013). Numerous revolutionary youth groups, leftist
political parties and Islamist political parties were able to unite over the ousting of Hosni Mubarak in
2011, but in 2012 they divided over whether legislative elections should be held before the constituent
assembly would sit in order to write a new constitution. This division opened the door for the military
to re-intervene in 2013. Such tendencies to fragment over issues that arise during transitions in diverse
coalitions means their power may, on average, decline faster after a transition, making opportunistic
re-autocratization more likely.
We (and existing studies) are agnostic ex ante on the relative strength of these effects. It may
be the case that the pro-democratic preference effect of greater diversity is so high that the negative
impact of diversity on bargaining is not sufficient to counteract it. However, it is also possible that the
negative effects of diversity on bargaining are so strong that the ultimate impact of diversity on future
democracy through changed preferences is completely eliminated. This is the “diversity dilemma.
Thus, although we expect that organizational diversity creates latent preferences within a movement for
democracy, we do not necessarily expect diversity to independently impact the likelihood of subsequent
democratization due to the countervailing bargaining problems it entails. Thus our first hypothesis is:
H
1
: Increasing organizational diversity in resistance campaigns has no significant effect on future
levels of democracy.
The Durability Advantage
The second characteristic of resistance movements’ organizational makeup that likely impacts democratic
preferences and capabilities concerns the types of organizations that participate in these movements.
Our theory draws upon existing scholarship that offers insights about which organizations in societies
are particularly predisposed to promoting democratization. Case-based studies by Collier (1999),
Rueschemeyer et al. (1992), and Bratton and van de Walle (1997) have argued and shown that
democratization processes have often been forced by trade unions or labor based parties. Recently,
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
Dahlum et al. (2017), Haggard and Kaufman (2016), and Butcher et al. (2018) have presented
quantitative evidence showing that mobilized industrial workers and national trade unions are strongly
correlated with democratization after resistance campaigns. Huang (2016) and Kadivar (2018) also
argue that insurgent groups can create durable mobilization infrastructures over years of struggle
or through creating institutions to tax civilians. What unites these approaches is the insight that
particular types of organizations, when mobilized against the regime, have features that make dissident
mobilization especially likely to result in democratization.
What makes organizations like labor unions so closely associate with improved prospects for
democratization when mobilized against autocratic regimes? We argue that labor unions are part
of a larger class of the “durable” organizations that comprise civil society. Durable organizations
are embedded in social networks and practices that are stable over time and they are not formally
designed to compete for or capture state power. Rather, they exist to advocate for the interests of
a clearly-identifiable constituency. In addition to labor unions, religious organizations and cultural
advocacy groups are also durable organizations performing similar functions. We believe these durable
organizations are particularly inclined to create pro-democratic preferences and enable resistance
movements to sustain mobilization and credibly threaten punishment for attempts at re-autocratization
during democratic transitions.
First, we argue that durable organizations have good reasons to prefer democracy over autocracy,
all else being equal. Durable organizations are not designed to compete for and hold formal political
offices. Yet, they rely on toleration by the state and often seek some degree of state preference.
Autocracies typically attempt to repress durable organizations or to co-opt them by offering some
degree of preferential treatment, for instance through state sponsorship of religious groups or corporatist
arrangements with labor unions. Yet in an autocracy, where political violence is often the ultimate
deciding factor and there are fewer limitations on the exercising of state power (Svolik 2012), durable
organizations have reasons to mistrust the long-term capacity of cooperative arrangements with the
state as a way to ensure their specific preferences.
In contrast, in a democracy, greater limitations on state power provide durable organizations with
greater flexibility to pursue their goals. They have less reason to fear that, if preferential arrangements
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
with the state break down, their activities will be met with state repression. Furthermore, if they face
potential conflict with the state then they have a less costly means of advocating for their constituency
through the ballot box.
Durable organizations will not all have uniformly democratic preferences. Structural factors such
as the degree of economic inequality, the social fractionalization across religious or ethnic lines, or
the particular dynamics of autocratic regime co-optation, will doubtless affect their preferences. Yet,
because of their core goals (group advocacy rather than holding political office) and the preference for
some limitations on state power to ensure their operation without state repression, we expect that on
average when durable organizations participate in resistance movements they should be more likely to
promote democratization following the campaign. In particular, they should be more likely to hold
genuinely democratic preferences than other groups that participate in resistance movements that are
explicitly designed to seek partial or complete state power, such as political parties or rebel groups.
Second, durable organizations should have greater capacity to ensure that these democratic
preferences are realized; this is what we call the “durability advantage.” As we highlighted in the
previous section, there is an important temporal dimension to democratization (Linz and Stepan 1996;
O’Donnell and Schmitter 1986; Przeworski 1991) that can create commitment problems and open the
door to re-autocratization attempts. A regime may promise democratization today, but that promise is
only credible when the opposition believes it can punish attempts at defection by the regime in the
future. Resistance movements may not accept offers of democratization if they do not believe they will
have the capability to compel post-conflict regimes to follow through on such promises. Similarly,
resistance movements that do not have the capacity to punish attempts at defection may miscalculate
and take the bait of short-term assurances of democratization only to discover that the regime is able to
renege on these promises (Bueno de Mesquita et al. 2005; Haggard and Kaufman 2016; Butcher et al.
2018; Huang 2016).
Durable organizations help resistance movements overcome these problems. First, embeddedness
in quotidian practices such as work or worship mean these organizations maintain their mobilization
infrastructures over time (Parkinson 2013).
3
These concerns motivate them to build more enduring
3
Kadivar (2018) and Huang (2016) argue that insurgent groups can create durable mobilization infrastructures
over years of struggle or through creating institutions to tax civilians. Here we emphasise the literature that
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
mobilization infrastructures that can be called upon to continue to advocate for their constituencies.
These durable infrastructures in turn enable movements to retain their capabilities over time and
re-mobilize, or threaten to re-mobilize, if democratic backsliding occurs. The Tunisian General Labor
Union (UGTT), for example, organized general strikes after left-wing politicians were assassinated in
2013. Trade unions were also central to overturning the coup by the Presidential Guard in Burkina
Faso in 2015, now one of only a handful of countries in Africa to score over 0.7 on the Polyarchy index
(discussed in more detail below).
Other types of organizations have difficulty sustaining their mobilization infrastructures through
post-campaign transition periods. Rebel groups, for example, face high transaction costs for converting
their wartime mobilization apparatus to electoral context and have generally done poorly in post-conflict
elections (Dunning 2011). Spontaneous protest movements such as the so called “Twitter” revolutions
may be effective at generating turnout in the short term but the proliferation of weak ties makes these
movements vulnerable to repression and prone to fragmentation in a transition (McAdam et al. 2001;
McAdam 1986; Tufekci 2017).
Second, periods of democratic transition are often turbulent because a member of the resistance
movement assumes state power if the campaign is successful. This transforms one organization’s
preferences from anti-regime to pro-regime. This organization is often the largest and best able to
mobilize, leaving the residual opposition groups weaker and potentially unable to deter the return
to anti-democratic practices. Political parties, for example, will seek to capture state power in a
transitional period, activating this dilemma. Rebel groups that achieve military victories or concessions
in power-sharing arrangements also exit the opposition coalition and become (part of) the government.
However, civil society organizations are not directly designed to hold positions in government and
therefore often remain a potential part of the opposition during a transition. This slows the decay rate
of the opposition’s mobilization capacity and means these organizations retain strong incentives to
compel transitional regimes to build checks and balances into the new constitution (such as freedom of
speech and assembly and term limits) as they do not expect to take state power but are vulnerable to
argues movements commonly draw upon pre-existing organizations rather than creating them ex-nihilo (Staniland
2014). However, the logic of these arguments is similar to that which we highlight here.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
repression by the regime.
If resistance movements consist entirely of groups with weak ties to inconsistent social bases
and no natural reason for continuing activities once the initial concessions of regime change have
occurred, they will have little capacity to re-initiate resistance during the transition. Movements that
feature strong and durable organizational roots will, therefore, be better able to overcome commitment
problems in the short term and accept regime concessions, initiating the democratization process. They
should also be more likely to successfully deter attempts at democratic backsliding by a transitional
regime or to fight back against this backsliding should it happen. This leads to our second hypothesis:
H
2
: Episodes of collective dissent with the participation of a greater proportion of durable
organizations will lead to higher levels of democracy in the future.
In the preceding theoretical discussion we identified and refined two mechanisms that are thought to
link mass mobilization to democratization. One emphasizes organizational diversity within resistance
campaigns while the other emphasizes organization type. The next section describe this study’s research
design, including how we operationalize diversity and durability using new data on the organizational
composition of resistance movements in Africa from 1990-2015.
RESEARCH DESIGN
We test our theoretical expectations quantitatively, using original data on the organizational composition
of violent and nonviolent resistance campaigns in post-Cold War African states. All country-years
in North and Sub-Saharan Africa from 1990-2015 are included.
4
The data come from two primary
4
This means that our sample contains cases with widely heterogeneous regional contexts and socio-cultural
backgrounds. We consider this a strength of the data. The temporal and geographic range is based on limitations
in available data, rather than scope conditions of the theory; however, we are cognizant of the challenges of
extrapolating general trends from regionally limited data. The resistance movements we capture in our data
diverge significantly in terms of major contextual and internal characteristics, which gives us some degree of
confidence in our findings’ external validity. However, we can, of course, only say with complete confidence that
our findings apply to cases within Africa.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
sources, the Varieties of Democracy dataset (V-Dem) (Coppedge et al. 2018a) from the V-Dem
Project at the University of Gothenberg, and original data collection on organizational participation in
resistance movements. We discuss the data collection process for the relevant variables in detail below.
Dependent variable
Our dependent variable is the future level of democracy in a given country. We measure this with a
series of leads on the “Polyarchy” score from the Varieties of Democracy (V-Dem) project (Coppedge
et al. 2018a), measured from time t+1 to t+5. The Polyarchy score is an index of various measures of
democratic institutions intended to capture the extent to which a country approximates an ideal electoral
democracy in terms of the ability of a widely enfranchised proportion of the country’s population
to participate in free elections that determine the composition of executive power. Included in the
Polyarchy score’s index are components that capture freedom of expression and association as these
affect the population’s ability to make informed choices between candidates. We also ran tests on two
additional indices from V-Dem that capture different dimensions of democracy: Liberal Democracy
and Participatory Democracy (Coppedge et al. 2018a). The Liberal Democracy measure captures the
degree to which government is limited and the rights of individuals and minority groups are protected
against both the state and the majority. The Participatory Democracy measure captures the degree
to which a citizens are actively engaged in all aspects of the political process, both electoral and
non-electoral (Coppedge et al. 2018b, 40-41).
Independent Variables: Organizational Participation in Resistance
The main independent variables are measures of the organizational diversity of nonviolent and violent
resistance movements and the extent of participation by durable organizations. These measures
come from an original data collection effort, described in detail here. The data record organizational
participation in events of collective dissent in North and Sub-Saharan Africa from 1990 to 2015. The
events we examine are violent and nonviolent events of contentious politics such as protests, strikes,
terrorist attacks, and conventional as well as irregular civil conflict.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
To focus our attention on incidents of dissent most likely to lead to major political changes, we limit
our data collection to dissent with “maximalist” claims. Maximalist claims are demands for changes in
the political structure which, if implemented, would significantly alter the executive branch’s immediate
access to state power, the rules through which executives are selected, or the policy or geographic
areas over which the executive has the right to exercise authority. This builds on and refines the
definition of maximalism in (Chenoweth and Stephan 2011). Maximalist demands are not the same as
grievances. A protest against corruption, for example, may generate calls for an independent watchdog
group (not maximalist) or for the head of state to step down (maximalist). Pro-democracy demands in
autocratic settings are a subset of all maximalist demands, but our broader definition also allows us to
also examine episodes of contention that are aimed at the preservation of democracy rather than the
creation of democracy (for example, against changes to term limits) or establishing new states which
may or may not be democratic. By implication this also includes other claims for enfranchisement
on the basis of gender or collective identity such as during the civil rights movement in the United
States, the anti-apartheid movement in South Africa, or the female suffrage movement in the United
Kingdom. Our definition would not include human rights movements such as an LGBTQ movement
since it does not alter the political rules around who gets to choose the executive or those areas over
which the executive can exercise control (unless LGBTQ people were also disenfranchised).
We identified events of maximalist dissent and organizational participation by drawing on four
well-known events datasets for instances of maximalist claims-making and organizational participation.
These datasets are the Uppsala Conflict Data Program Geo-referenced Event Dataset (UCDP-GED:
Sundberg and Melander (2013)), the Armed Conflict Location and Event Dataset (ACLED: Raleigh
et al. (2010)), The Mass Mobilization Data (MM: Clark and Linzer (2015)) and the Social Conflict
Analysis Data (SCAD: Salehyan et al. (2012)). The first step was creating a set of candidate events
of maximalist dissent with a combination of subsetting on existing variables in these datasets and
text-matching.5
5
All SCAD events with goals of elections, democracy and ethnic/religious issues or events where the event
notes matched a customized series of text strings were included in the candidate list, specifically any events that
included any of the following strings: resign, election, autonomy, independen, step down, democra, stand down,
coup, dissolve, constitution, pluralism, president, multiparty, elector, transitional, bring, down, junta, military
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
We investigated each candidate event to determine whether participants in the event made maximalist
demands and whether any named organizations participated in or organized the event. We consulted
both secondary sources and newswires in the FACTIVA and LexisNexis databases for this process,
6
.
For every maximalist event, we recorded the organization(s) that participated in executing the event, if
there were any that were formally identifiable, typically with an official name. Multiple organizations
can be associated with events of maximalist dissent, but only organizations that directly mobilized and
participated were included in our dataset. For instance, in a protest planned by a labor union and a
student group calling for the overthrow of a non-democratic regime we would include both the labor
union and the student group but not, say, a political party that only issued a statement praising the
protest. We coded organizational participation when sources stated that a specific group organized the
event, members of the leadership were at the event, supporters of an organization were associated with
the event, or organizations called for the event (often in the case of protests/strikes) before they occurred.
We included all rebel groups included in the UCDP as their demands are by definition maximalist. We
also included all organizations associated with interaction “12” events in ACLED as by definition they
make maximalist claims in the ACLED codebook (Raleigh et al. 2010). We examined 10,437 events
for maximalist claims and organizational participation across the SCAD, MM and ACLED datasets. Of
rule, dictato, autocra, fraud, regime, new, government, corruption, anti government, anti-govern, islamic state,
religious state, liberation, sharia, remov, koran, bible, or secession. For the MM dataset, any events that matched
the text string or where protesters made demands that were “political,” related to “police brutality,” or called for
a politician to step down were included. All ACLED events where the event notes matched the text string and
had interaction codes 12, 13, 15, 16, 50 or 60 were also included in the candidate list. These codes correspond to
interactions of “military versus rebels,” “military versus political militia,” “military versus communal militia,
“military versus rioters,” “military versus protesters,” “sole rioter action,” and “sole protester action. All GED
were included as demands over issues related to government or territory as defined by the UCDP are maximalist
by our criteria (Gleditsch et al. 2002).
6
Specifically the Associated Press, Agence France Presse, BBC World News Monitoring, Reuters Newswires,
and Africa News. Our major secondary sources were the Historical Dictionary Series, Political Handbook of the
World,Political Parties of the World and Trade Unions of the World, occasionally supplemented by additional
scholarly sources specific to the country or resistance movement.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
these events, we coded maximalist claims and participation by at least one organization in 6,722 events.
This process produced a list of organizations linked to events in the GED, SCAD, MM and
ACLED datasets. We then aggregated this list up to the organization-country-year to produce a list
of organizations involved in events of maximalist dissent, both violent and nonviolent, in Africa
from 1990-2015. The data includes information on 1,199 distinct organizations active across 2,980
country-organization-years. Just 141 (11.8%) of these organizations are associated with events in
the GED. Thus, the vast majority of organizations that engaged in maximalist dissent in Africa from
1990-2015 were not rebel groups involved in civil wars, but organizations involved in nonviolent
resistance or low-level violence that did not produce 25 or more battle deaths in a year. These data are
possibly biased toward larger, more prominent organizations and those with experience or incentives to
interact with the news media, especially political parties but probably also trade unions and insurgent
groups. In particular, we suspect that the number of civil society organizations (CSOs) that participate
in maximalist dissent are underestimated, especially in non-democracies. We may record a greater
number of CSOs in more democratic states where press freedom entails better reporting or where civil
society is allowed more leeway to organize and act. This raises the possibility that any correlation
between the percentage of CSOs, which we use as one independent variable below, and levels of future
democracy may be spuriously correlated with levels of democracy at time t. A lagged dependent
variable to address this problem below along with a measure of how open civil society is and country
fixed effects that should capture the likelihood of organizations or certain types of organizations making
into the data. In addition, for our main analyses, only organizations that were associated with events
that had more than 1,000 participants (as recorded by either the SCAD or MM dataset), were in the
GED, or associated with ACLED events where more than 25 people were killed were included. These
more “severe” events should receive better media coverage and limit the extent to which media biases
affect the results (Weidmann 2016).
Diversity
We used two measures to capture the organizational diversity of a movement: (1) a count
of the number of active organizations engaged in maximalist dissent, and (2) a measure of the
fractionalization of participation across different types of organizations.
Our first measure is a raw count of the number of organizations engaged in maximalist dissent in
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
the country-year. If organizations form at least in part around shared individual preferences, then the
number of organizations is a plausible proxy for the diversity of these preferences across the movement.
A movement with five political parties, for example, likely has more diverse political preferences than
a movement with just one party. We normalized this measure by transforming it with the natural
logarithm. Our data collection procedure should be effective at capturing the “main” organizations
in a movement, but the ability to capture smaller organizations (and therefore a greater number of
organizations) might vary from country to country. The log transformation should reduce the impact of
this potential measurement error.
Our second measure of movement diversity is a fractionalization index of the organization types
active in a country-year. All organizations were classified into one of nine categories based on their
primary goals. These categories are shown in Table 1. The first category, “revolutionary organizations,”
captures organizations that form with the primary goal of prosecuting maximalist claims such as regime
change through non-institutional means. This category includes nearly all armed rebel organizations
along with some protest groups with explicitly maximalist goals such as Tamarod in Egypt. All
organizations in the data make maximalist demands during periods of dissent but not all have maximalist
goals as their primary objective. For instance, trade unions participate in maximalist movements but
their primary goal is to further the interests of their workers. The second category is political parties,
that is to say organizations designed to contest and win elections. Our subsequent categories are groups
with the primary goal of representing the interests of specific social groups starting with workers
(trade unions) professional groups, women, ethnic groups, religious groups and youth/student groups.
Importantly, organizations that are categorized as religious or ethnic are not primarily religious political
parties (such as the Muslim Brotherhood) or ethnically based insurgent groups. These groups would be
coded as political parties or revolutionary organizations, respectively. Our categories of religious and
ethnic organizations are civil society based religious or ethnic organizations without a primary aim
of contesting elections or pursuing maximalist goals through contentious dissent. The final category,
“other CSOS,” captures organizations such as human rights advocacy groups with non-maximalist goals
but no clearly identifiable social group that they claim to represent. Table 1 illustrates the categorization
of organization types with examples from several of the countries included in our data.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
TABLE 1. Organization Types
Type Org. Goals Example
Revolutionary Maximalist goals
AFDL: Alliance of Democratic Forces
for the Liberation of the Congo
Political Party Win elections
UFDG: Union of Democratic Forces
of Guinea
Trade Union
Worker’s rights/interests in a
trade or industry
CTNG: The National Confereration
of Guinean Workers
Professional
Professionals’s rights/interests
ONAT: Tunisian Bar Association
Women’s Women’s rights/interests
AFTD: Tunisian Association of Demo-
cratic Women
Ethnic Ethnic rights/interests
MAK: Movement for the Autonomy of
Kabylie (Algeria)
Religious Coordinate religious activities
FNAI: Tunisian Front of Islamic Asso-
ciations
Student or Youth Youth/student interests
GUTS: General Union of Tunisian
Students
Other CSOs
Broad, universal non-
maximalist goals (i.e human
rights)
CNOSCG: National Council of Civil
Society Organizations
We argue that these different social groups are likely to have different policy preferences and
must answer to different constituencies (labor unions will push for labor rights while religious groups
typically advocate for greater religious freedoms, for example). As such, the degree to which a
movement is spread across these different social groups is an indicator of the movement’s diversity. To
measure this we created a variable (Measure 2) to capture the extent of movement fractionalization by
type, that is to say the degree to which the organizations active in a given country year were either
dominated by one organization type or dispersed across a number of types. Fractionalization was
measured using the inverse Herfindahl index, commonly used to measure domination of a market by a
firm, and given in equation 1:
Fr actional i zation =1
n
Õ
i=1
s2
i(1)
In this equation sis the proportion of all organizations engaged in maximalist resistance accounted
for by organization type i, and nis the total number of organization types engaged in maximalist
resistance. As above, if different types of organizations form around different social groups and develop
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
distinct political preferences then movements that are more evenly dispersed across organization types
should have more diverse preferences.
Durability
We measured the presence of durable organizations (H
2
) with the percentage of organi-
zations in a resistance movement-year that were trade unions, religious organizations, professional
organizations, women’s organizations or ethnic organizations.
7
These organizations are characterized
by enduring social relations centred on everyday interactions such as work, worship or collective identity
and do not have the primary aim of competing for state power. They are thus more likely to endure
after short term changes in the regime and will by and large retain their mobilization infrastructures.
Some of these groups may even grow in transitional periods. This is in contrast to political parties
which may exit anti-government resistance networks if they are able to capture state power, or rebel
groups that must weather high costs transitioning from structures designed for the pursuit of political
goals with violent means to nonviolent and institutional settings.
In some contexts student/youth organizations and other CSOs may be built on quotidian ties
(although we think it is more likely that trade unions, religious organizations, professional organizations
and ethnic organizations are rooted in everyday practices). As an alternate test of the percentage
of durable organizations we also created a variable measuring the percentage of organizations in a
resistance campaign-year that were not political parties or rebel groups (% CSOs). This percentage
includes our primary measure of durable organizations as well as other kinds of civil society groups
such as student groups and human rights advocacy organizations.
Control variables
The diversity of resistance movements and the presence of durable organizations are not randomly
distributed. The types of organizations that could join resistance movements, as well as the presence
and strength of social groups that could spawn organizations and their abilities to coordinate, are likely
affected by past levels of civil society openness and regime type, past levels of contention, structural
economic factors and economic growth. These factors also very likely impact democratization, our
7
Other potential measures of durability would be organizational age or size, but these data do not currently exist.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
outcome of interest. Our modelling strategy is designed to minimize the effects of omitted variable
bias to the greatest extent possible.
In particular, we seek to include all relevant control variables that may affect our independent
and dependent variables. We include GDP growth because economic crises may make regimes more
vulnerable to defections and democratic transitions (Brancati 2016). Declining economic conditions
may also lower the opportunity costs for organizations to engage in resistance against the regime
(Haggard and Kaufman 2016). We also control for economic development using GDP per capita, as
development is associated with increasing organizational density in civil society and may independently
cause regimes to liberalize (Boix 2003; Lipset 1959; Haggard and Kaufman 2016; Diamond 2008).
These two variables come from the World Bank (2018). The participation of CSOs in resistance
campaigns is probably more likely in countries where civil society is denser or less repressed. Thus we
control for underlying civil society potential with the Core Civil Society Index from V-Dem (Coppedge
et al. 2018). The widespread use of violence may deter dense or diverse organizational participation and,
again, make democratization less likely by increasing elite cohesion (Chenoweth and Stephan 2011).
We thus include a control for the percentage of organizations associated with violent events.
8
We also
include a linear year trend to capture continent-wide factors that may have encouraged democratization
and the strong likelihood that more organizations are captured with better newswire coverage over the
25 years of our study.
9
Finally, we include the lagged dependent variable (measured at t- 1) because
the level of democracy at time t + nis related to past levels of democracy and is also related to freedom
of the press and our ability to identify organizational participation. We measured all controls at t- 1 to
avoid contaminating the control variables with the effects of organizational participation at time tand
introducing post-treatment bias.10
All models include either mixed effects with random country intercepts or country fixed effects
unless specified otherwise. The country fixed effects models capture country-specific features (such as
8GED events, ACLED interaction 12 events and “violent riots” or “anti-government violence” in SCAD.
9
Although more recent events likely have poorer coverage in secondary/historical sources. In robustness tests we
also use year fixed effects to control for continent-wide shocks that may be driving the results with very similar
results to those reported below.
10We ran robustness checks with several additional control variables (detailed in the appendix).
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
the nature of authoritarianism prior to 1990 and colonial heritage) that affect the number and types
of organizations that participate in resistance campaigns and country-specific factors that may have
deterred or encouraged democratization. OLS regression models were employed because the dependent
variable is continuous. We performed all the analysis in the R statistical computing platform.
Table 3.3 reports summary statistics on the main variables used in our analysis. There is wide
variation in the number of organizations participating in resistance movements. Out of a total of 1,253
country-years in the data there was maximalist resistance with at least one organization participating
in 481 years, or roughly 38% of all country-years in Africa between 1990 and 2015. Of these 481
observations, 68 country-years have between 5 and 10 organizations participating in maximalist
resistance, and 24 country-years have more than 10 organizations participating.
TABLE 2. Summary Statistics on Main Variables
Statistic N Mean St. Dev. Min Max
No. of Orgs 1,253 1.314 3.324 0 37
Fractionalization 1,253 0.048 0.160 0.000 0.960
% Durable 1,253 0.018 0.102 0.000 1.000
% CSOs 1,253 0.033 0.145 0.000 1.000
% Violent Orgs. 1,253 0.317 0.457 0.000 1.000
Polyarchy (t-1) 1,252 0.353 0.185 0.074 0.803
GDP per cap (log) 1,148 7.722 0.976 4.898 10.750
GDP Growth 1,148 0.020 0.113 0.590 1.309
Robust CS Index 1,252 0.590 0.250 0.021 0.976
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
RESULTS
Descriptive statistics
Figure 1 shows the distribution of organizations we record in active maximalist dissent over the
1990-2015 period by country. Grey crosses are the better-studied violent rebel groups (those associated
with GED events), while open circles are political parties, triangles are trade unions and other CSOs,
and other revolutionary organizations are squares.
11
Organizational participation in dissent has grown
over the period of observation with peaks in the early 1990s, mid 2000s and the largest peak during
the Arab Spring from 2011 onward. The types of organizations that have participated in dissent have
remained fairly stable over time with political parties making up 49% of active organizations in 1991
and 38% in 2015. Rebel groups constituted 35% of active organizations in 1991 and 44% in 2015.
Other civil society organizations made up 15% in 1991 and roughly 21% in 2015.
Figure 2 shows how democratization rates vary across three categories of observations in our data
as an initial test of H
1
and H
2
. The observations labelled “diverse” (triangles) consist of country-years
with 10 or more organizations but where less than 25 % of those organizations were classified as
“durable.” The observations labelled “durable” (squares) are country-years where 25% or more of
the active organizations were “durable” organizations in our schema. The “control” group includes
observations without 10 or more participating organizations or 25% durable organizations and that
were not exposed to either of these treatments 1-5 years after the observation year (grey dots). Along
the x-axis is length of the lead from 1-5 years and the y-axis measures the change in the Polyarchy
score from t- 1 to t+ n as indicated on the x-axis.
Figure 2 offers some initial support for our hypotheses. Years with the participation of 25% or more
“durable” organizations show positive changes in their Polyarchy scores. By t+ 2 this is roughly 0.1
points higher than the Polyarchy score at t- 1. Up until t+ 4 the mean Polyarchy score for country years
with diverse (but not durable) opposition coalitions is very close to the control mean, but rises slightly
at t+ 5. While this offers initial support for H
1
and H
2
, the percentage of durable organizations and the
11
These are both rebel groups with fewer than 25 casualties in any year, such as the Niger Delta Freedom Fighters
(NDFF) in Nigeria, and nonviolent groups formed with an explicitly maximalist goal, such as Tamarod in Egypt.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
FIGURE 1. Organizational Participation in Resistance in Africa, 1990-2015
FIGURE 2. Democratization by Diversity and Durability, Africa, 1990-2015
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
diversity of organizational participation are not randomly assigned and are likely correlated with factors
that also affect democratization. Thus, it is crucial to examine these questions through regression
models that can take the systematic effects of other factors that affect organizational participation and
democratization into account.
Regression Models
In this section, we present the results of our primary regression models, as well as a number of robustness
checks.
12
Figure 3 displays results from the country fixed effects models of our two measures of
organizational diversity.
13
The points are the average marginal effect in models including our full set
of controls, while the vertical lines are 95% confidence intervals. As the plot shows, organizational
diversity – measured either by the logged number of organizations or the fractionalization across
organization types – has little impact on future levels of democracy. The expected Polyarchy scores 2-5
years into the future are not significantly different for a country that experiences dissent with high levels
of organizational diversity when compared to a country-year with no dissent at all. Polyarchy scores are
significantly lower at t+ 1 for episodes of dissent involving many different organization types and with
a large number of organizations overall, indicating that such movements may have particular problems
negotiating transition periods or difficulty attracting short term government concessions. These results
provide evidence supporting H
1
: high organizational diversity does not, on average, lead to greater
democratization, indicating that the countervailing forces of the “diversity dilemma” may be at work.
The picture is quite different when examining the influence of durable organizations, as shown
in Figure 4. Our primary measure of the “durability advantage,” the percentage of participants in
resistance that are durable organizations, exhibits a significant positive impact on the Polyarchy score
at each point from t+ 1 to t+ 5. The impact is smaller for CSOs, a measure that includes human rights
or other activist organizations that may have lower mobilization capacity and thus be less able to shape
post-conflict institutions. For the percentage of durable organizations the size of the effect is relatively
12
The full regression tables including results for both mixed and fixed effects models can be found in Appendix
tables 2-5
13Results are substantively identical in the mixed-effects models with random intercepts.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
FIGURE 3. Fixed Effects Models of Diversity and Future Democracy t+1–t+ 5
large - a country that experiences an episode of dissent characterized by 100% durable organizations
is expected to score more than 0.1 points higher on the Polyarchy index 2-3 years after. The models
also include a significant and positively signed lagged dependent variable, meaning the substantive
effects are larger because higher past democracy scores make it more likely that democracy scores in
the future will be higher.
The control variables behave mostly as expected, with the Polyarchy score at t- 1 proving to be a
consistent and significant predictor of future Polyarchy at t+ 2 but becoming insignificant at t+ 4.
Higher scores on the core civil society index at t- 1 are also strongly and positively correlated with
future levels of democracy up to 5 years into the future. Neither GDP per capita nor GDP growth
are ever significant in the main models. The proportion of organizations associated with at least one
violent event is negatively related to democratization at t+1 but not after. There is also a strong linear
trend towards democracy, as indicated by the positive sign and high degree of significance on the year
variable.
Robustness Checks
These results provide further evidence strengthening the ambiguous impact of
organizational diversity and the strong positive impact of durable organizations. Yet there are a number
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
FIGURE 4. Fixed Effects Models of Durability and Future Democracy t+1–t+ 5
of potential weaknesses of these tests that necessitate a number of robustness checks. First, we ran
several tests changing the specific variables included in the model. It is plausible that our results may
be attributable to some systematic omitted variable bias. Thus, we re-ran our models with several
additional control variables to account for other “usual suspects” in the mobilization-democracy story.
14
Far from eliminating the correlations identified in the main models, our results become more highly
significant when including these additional controls. It is also possible that institutional changes toward
or away from democracy may not manifest as dissent immediately. Institutional openings in the more
distant past may still provide opportunities for mobilization even if the regime becomes less democratic
in the short term. To address this, we re-ran our models including longer lags of the dependent variable
and the civil society index. Our measures of the participation of durable organizations remains positive
and significant. Next, while the Polyarchy score – our primary dependent variable – captures a very
broad conception of democracy, it leaves out important dimensions of democracy relevant for our
theoretical story, such as the protection of minority rights or participatory institutions. Thus, we
replicated our main models but replaced the Polyarchy score with V-Dem’s Liberal Democracy and
14
Specifically, we include controls for oil wealth, regional democratic diffusion, government repression, and
campaign size. Details on our sources and coding are in the Appendix.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
Participatory Democracy indexes. The results are substantively similar.
Second, our results may be driven by a systematically different subset of the diverse set of country-
years on which we perform our tests. To account for this, we re-ran our main models on several subsets
of the full dataset. Specifically, we split the sample based on whether the regime at t- 1 was autocratic
or democratic;
15
and into “peace years” where no organizations were associated with GED events and
“war years” where there was an ongoing UCDP civil war. Our results are robust in autocratic and peace
years. The percentage of durable organization is negatively and significantly correlated with Polyarchy
scores at t+ 1 in the sample of democracies and not significantly associated from t+ 2 to t+ 5. We
suspect that the participation of these organizations in maximalist resistance in democracies signals a
period of democratic crisis.
Third, in our main tests we follow Weidmann (2016)’s suggestion of avoiding reporting bias by
including only organizations that participate in high-intensity events (those with 1,000 participants or
25 casualties). It is possible that this exclusion biases our results towards those resistance movements
most likely to achieve democratic progress. We re-tested the results measuring diversity and durability
indicators for all organizations regardless of the size of the events they were associated with (i.e.
including organizations associated with smaller events). As Weidmann’s argument on reporting bias
would lead us to expect, there is some additional noise in the data when these smaller events are
included, but the size and significance of the coefficients remains substantively similar.
Finally, we conducted two tests on two small samples of country-years highly likely to be
“transitional:” years where there was at least one protest of 10,000 people or more in the country-year
(56 cases) and years where there was an irregular regime change (43 cases). Although the small number
of cases here precluded the use of country fixed effects, the models do still involve the full range of
control variables including the lagged dependent variable and lagged civil society index which should
capture differences in key institutions, capacities for collective action and media-biases across countries
and time. Our results retain the direction and significance of the main tests in the years of irregular
regime change, but lose significance in some models limited to years of large protests.
In none of the robustness tests does the number of organizations have a significant positive effect
15With the distinction between these two defined by a Polyarchy score threshold of 0.5.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
on future democratization. Fractionalizaton continues to have a significant negative coefficient at t
+ 1 in a number of the robustness tests. Specifically, the fractionalization coefficient is negative and
significant in civil war years, democracies, after large protests, with extra controls and with small
protests included. It is very close to statistical significance in the remaining robustness tests with the
exception of periods following irregular leadership changes. That we continue to find few significant
correlations between diversity and future democracy following big protests and in peace years suggests
that the diversity dilemma applies to periods of nonviolent contention and violent contention.
Instrumental Variable Analysis
Causal identification in any observational study is challenging. Our regression models aim to address
concerns by conditioning on many of the extensive number of variables identified as important in the
democratization literature, as well as by including country fixed effects and the lagged dependent variable.
The conservative nature of this modeling strategy makes it unlikely that there is some systematic
variable sufficiently highly correlated with our independent and dependent variables to explain the
observed effects. It is possible, however, that organizations anticipate medium-term democratization
and participate in dissent in response to this expectation and/or that other unobserved features cause the
participation of durable organizations or diverse coalitions and also cause democratization. However,
these unobserved factors must be independent of the features controlled for in the regression models.
To address this potential concern, we ran two instrumental variable analyses (shown in the Appendix).
The starting point comes from Smith (2014) who claims that food price shocks, instrumented with
trade-adjusted changes in international grain prices, are causally related to mobilization, along with
6 and 9 month deviations in rainfall. Food price shocks predict diverse organizational participation
well, and we argue in the Appendix that there are good arguments that this instrument satisfies the
exogeneity and exclusivity assumptions in relation to democratization.
Food price shocks and rainfall deviations do not predict the percentage of durable organizations or
CSOs that participate in resistance movements. Youth bulges do, however, and there is a significant
literature arguing that this is a plausible instrument for mass mobilization, specifically in relation to
democratization (Fuller and Pitts 1990; Urdal 2006). Youth bulges measured as the proportion of the
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
population between 0 and 14 do (negatively) predict CSO and durable organizational participation and
we argue in the appendix that this variable likely satisfies the exogeneity assumption. However, there is
more uncertainty over the exclusivity assumption in our case. The instrumental variable models reported
in the appendix have additional controls from the V-DEM data to capture the link between youth bulges
and preemptive repression
16
that may create institutional “lock ins” that make democratization difficult
or empower the military (Nordås and Davenport 2013; Urdal 2006). Nonetheless these results should
be treated with caution, although we emphasize that variation in organizational participation predicted
by youth bulges is unlikely to be driven by anticipated democratization and goes some way to easing
concerns that the findings are driven by reverse causality. Omitted variable bias remains a potential
problem.
The results of two-stage least squares regression models using food price shocks as an instrument
for diverse participation and youth bulges as an instrument for the % of durable organizations and
CSOs can be found in the Appendix. The models retain country fixed effects and the main control
variables. The results are consistent with the fixed effects models in the main text.
In summary, we find consistent evidence in support of both H
1
and H
2
. There is little evidence
for a direct effect of organizational diversity on future democracy, measured either as the number of
organizations or dispersion across organization types, indicating the existence of a diversity dilemma.
In contrast, there is strong and consistent evidence of a durability advantage. Durable organizations
participation in resistance is positively associated with levels of democracy 2-5 years into the future. This
result holds with extra control variables and in autocracies, which matches the theoretical notion that
durable organizations help overcome commitment problems and punish attempts at re-autocratization
which are likely to be especially severe in more autocratic settings. We also note that our overall
modelling strategy is very conservative. The main results are from OLS models with country fixed
effects, a lagged dependent variable and a time trend. They are also robust to the inclusion of year
dummies and additional control variables that capture yearly variations in media coverage within
countries. Thus, it is unlikely that country- or regime-specific features are driving the results or that
16
Specifically, one and two year lags of the CSO repression variable, along with an indicator of whether the regime
was a military regime on December 31 of the previous year (Hadenius and Teorell 2007; Wahman et al. 2013).
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
continent-wide shocks (such as the Arab Spring) account for the associations. The inclusion of an
indicator of the strength of civil society in combination with country fixed effects means our results are
also probably not driven by media biases particular to a particular country or within countries over
time.
CONCLUSION
In this manuscript we have proposed and tested new mechanisms linking mass mobilization to
democratization through the organizational characteristics of resistance movements. We first identified
a “diversity dilemma. Some studies argue that diverse coalitions are most likely to democratize
because they mobilize large numbers of people and have aggregate preferences for democracy. On
the other hand, diversity can introduce bargaining problems that may make these movements less
likely to succeed in the short term and vulnerable to fragmentation and re-autocratization in the longer
term. We also identified a “durability advantage,” arguing that democratization is more likely when
organizations embedded in quotidian social networks with no direct designs on achieving power
durable organizations – join resistance movements.
We tested these hypotheses using a new dataset of organizational participation in maximalist dissent
in Africa from 1990-2015. Adopting a conservative modelling strategy employing either random
country intercepts or country fixed effects, a lagged dependent variable and time trends, we found little
evidence that movement diversity is associated with future levels of democracy, measured either as the
number of organizations or dispersion across organization types. Durable organizations on the other
hand, were positively correlated with future levels of democracy for up to 5 years. This result was
robust across a number of modelling strategies and subsamples, as well as in an instrumental variable
analysis.
These results suggest that the link between mass mobilization and democracy is largely attributable
to the participation of durable organizations in resistance movements. These organizations have deep
roots in stable social networks that provide a consistent mobilization infrastructure. Furthermore,
because they are not designed to compete for or capture state power, these organizations are also ideally
placed to act as a bulwark against attempts to re-establish autocracy during a transition. Our research
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
also suggests that the shift towards democratic preferences typically assumed to be a part of diverse
movements alone is not enough to bring about democracy. Indeed, the consistent non-effect we find
between measures of movement diversity and democratization suggests that the dilemmas associated
with building big-diverse movements are a significant challenge that can undermine democratization.
That adherence to nonviolent tactics does not overcome this dilemma suggests that this problem is
especially difficult to navigate. In addition, this research has implications for the relationship between
structural factors and democratization. Numerous studies propose causal mechanisms that run through
the mobilization of collective actors (Ross 2001; Bueno de Mesquita et al. 2005), or even specific
collective actors such as trade unions (Butcher and Svensson 2016; Haggard and Kaufman 2016).
However, these mechanisms are rarely tested. Our results suggest that a useful step forwards would be to
examine the extent to which structural factors and democratization are related through the mobilization
of durable organizations.
The strategy of examining variation in resistance movements through their organizational make-up
is a research agenda in its early stages, with many potential expansions for future research. For instance,
the effect of diversity in resistance movements is an avenue open for much more quantitative research.
Our measures of diversity are proxies – we are not yet able to directly measure the diversity among
stated goals or preferences of participating organizations. More direct measurements on this dimension
may produce different or more nuanced results than those we have presented here. In addition, diversity
is closely related to the dispersion of power among opposition actors. As such, it may not be how
diverse the actors are that drives democracy but rather a relatively even balance of power between
them (Bueno de Mesquita et al. 2005). Our measure of fractionalization across types is closer to
dispersion than the raw count of organizations, but this too could be more directly measured with
the fractionalization of the movement by organization size. While our measure of the raw counts of
numbers of organizations participating in resistance captures organizational diversity to some degree,
we are likely conflating many resistance movements featuring many genuine centers of power and
competition among themselves with campaigns that may have several organizations participating, but in
which these organizations are all more or less satellites of a single central organization that overwhelms
them in size.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
Our analysis is also limited in time and scope. Organizational complexity has played an important
role in resistance movements in Southeast Asia (Boudreau 2009), Latin America (Stepan 1989),
and Eastern Europe (Bunce and Wolchik 2011). While there is significant diversity across African
countries and the campaigns therein, and we have no reason to believe that the effects of organizational
participation would be systematically different in other parts of the world, neither can we confidently
say that these patterns will hold true for resistance campaigns in other regions. This is a crucial
empirical question for future research.
These findings offer some preliminary lessons for pro-democracy activists and other practitioners.
Namely, those engaged in an anti-government resistance who are committed to pursuing democratic
reforms should focus on encouraging the participation of durable organizations, such as labor unions,
religious groups, or other context-specific organizations. Including a wide variety of groups in a
resistance campaign can also be beneficial for democratization prospects when a large proportion of
these groups are more durable. Furthermore, for third parties seeking to promote democracy following
anti-government movements, they should be ready to provide more assistance in situations where the
participant organizations themselves may struggle to enforce such reforms, for example in places where
the campaign was comprised of nascent groups or lacked engagement from civil society organizations.
REFERENCES
Acemoglu, Daron and James A. Robinson (2005, December). Economic Origins of Dictatorship and Democracy.
Cambridge University Press.
Ash, Timothy Garton (2014). The magic lantern: the revolution of’89 witnessed in Warsaw, Budapest, Berlin
and Prague. London, UK: Atlantic Books Ltd.
Bayer, Markus , Felix S Bethke, and Daniel Lambach (2016, November). The democratic dividend of nonviolent
resistance. Journal of Peace Research 53(6), 758–771.
Beissinger, Mark R. (2013, August). The Semblance of Democratic Revolution: Coalitions in Ukraine’s Orange
Revolution. American Political Science Review 107(3), 574–592.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
Bermeo, Nancy Gina (2003). Ordinary People in Extraordinary Times: The Citizenry and the Breakdown of
Democracy. Princeton University Press.
Bethke, Felix S and Jonathan C Pinckney (2016). Nonviolent resistance and the quality of democracy. V-Dem
Users Working Papers. Available at https://www.v-dem.net/files/45/Users%20Working%20Paper%203.pdf.
Boix, Carles (2003, July). Democracy and Redistribution. Cambridge University Press.
Boix, Carles , Michael Miller, and Sebastian Rosato (2013). A complete data set of political regimes, 1800–2007.
Comparative Political Studies 46(12), 1523–1554.
Boudreau, Vincent (2009). Resisting Dictatorship: Repression and Protest in Southeast Asia. New York, NY:
Cambridge University Press.
Brancati, Dawn (2016, September). Democracy Protests. Cambridge University Press.
Brancati, Dawn and Jack L. Snyder (2013, October). Time to Kill: The Impact of Election Timing on Postconflict
Stability. Journal of Conflict Resolution 57(5), 822–853.
Bratton, Michael and Nicholas van de Walle (1997, August). Democratic Experiments in Africa: Regime
Transitions in Comparative Perspective. Cambridge University Press.
Bueno de Mesquita, Bruce , Alastair Smith, James D. Morrow, and Randolph M. Siverson (2005). The Logic of
Political Survival. MIT Press.
Bunce, Valerie J and Sharon L Wolchik (2011). Defeating Authoritarian Leaders in Postcommunist Countries.
New York, NY: Cambridge University Press.
Butcher, Charles , John Laidlaw Gray, and Liesel Mitchell (2018, July). Striking It Free? Organized Labor and
the Outcomes of Civil Resistance. Journal of Global Security Studies 3(3), 302–321.
Butcher, Charles and Isak Svensson (2016, March). Manufacturing Dissent: Modernization and the Onset of
Major Nonviolent Resistance Campaigns. Journal of Conflict Resolution 60(2), 311–339.
Celestino, Mauricio Rivera and Kristian Skrede Gleditsch (2013, May). Fresh carnations or all thorn, no rose?
Nonviolent campaigns and transitions in autocracies. Journal of Peace Research 50(3), 385–400.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
Cheibub, José Antonio , Jennifer Gandhi, and James Raymond Vreeland (2010). Democracy and dictatorship
revisited. Public choice 143(1-2), 67–101.
Chenoweth, Erica and Maria J. Stephan (2011). Why Civil Resistance Works: The Strategic Logic of Nonviolent
Conflict. Columbia University Press.
Clark, Tom S. and Drew A. Linzer (2015, May). Should I Use Fixed or Random Effects? Political Science
Research and Methods 3(2), 399–408.
Collier, Ruth Berins (1999, September). Paths toward Democracy: The Working Class and Elites in Western
Europe and South America. Cambridge University Press.
Coppedge, Michael , John Gerring, Carl Henrik Knutsen, Staffan I. Lindberg, Svend-Erik Skaaning, Jan Teorell,
David Altman, Michael Bernhard, M. Steven Fish, Agnes Cornell, Sirianne Dahlum, Haakon Gjerløw, Adam
Glynn, Allen Hicken, Joshua Krusell, Anna Lührmann, Kyle L. Marquardt, Kelly McMann, Valeriya Mechkova,
Juraj Medzihorsky, Moa Olin, Pamela Paxton, Daniel Pemstein, Josefine Pernes, Johannes Von Römer, Brigitte
Seim, Rachel Sigman, Jeffrey Staton, Natalia Stepanova, Aksel Sundström, Eitan Tzelgov, Yi-Ting Wang,
Tore Wig, Steven Wilson, and Daniel Ziblatt (2018). V-Dem Country-Year Dataset 2018.
Cunningham, David E. (2006, October). Veto Players and Civil War Duration. American Journal of Political
Science 50(4), 875–892.
Cunningham, Kathleen Gallagher (2013a, July). Actor Fragmentation and Civil War Bargaining: How Internal
Divisions Generate Civil Conflict. American Journal of Political Science 57(3), 659–672.
Cunningham, Kathleen Gallagher (2013b, May). Understanding strategic choice: The determinants of civil war
and nonviolent campaign in self-determination disputes. Journal of Peace Research 50(3), 291–304.
Dahlum, Sirianne , Tore Wig, and Carl Henrik Knutsen (2017). Who Revolts? Empirically Revisiting the
Social Origins of Democracy. Working Paper, 62. Available at https://www.researchgate.net/publication/
319502382_Who_Revolts_Empirically_Revisiting_the_Social_Origins_of_Democracy.
DeNardo, James (2014, July). Power in Numbers: The Political Strategy of Protest and Rebellion. Princeton
University Press.
Diamond, Larry (2008). The Spirit of Democracy: The Struggle to Build Free Societies throughout the World.
Macmillan.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
Diani, Mario (2003). Social Movements and Networks: A Research Program. In M. Diani (Ed.), Social Movement
and Networks: Relational Approaches to Collective Action. Hoboken, NJ: John Wiley & Sons.
Dunning, Thad (2011, June). Fighting and Voting: Violent Conflict and Electoral Politics. Journal of Conflict
Resolution 55(3), 327–339.
Fuller, Gary and Forrest R Pitts (1990). Youth cohorts and political unrest in south korea. Political Geography
Quarterly 9(1), 9–22.
Geddes, Barbara , Joseph Wright, and Erica Frantz (2014). Autocratic breakdown and regime transitions: A new
data set. Perspectives on Politics 12(2), 313–331.
Gleditsch, Kristian Skrede and Michael D. Ward (2006, October). Diffusion and the International Context of
Democratization. International Organization 60(4), 911–933.
Gleditsch, Nils Petter , Peter Wallensteen, Mikael Eriksson, Margareta Sollenberg, and Håvard Strand (2002,
September). Armed Conflict 1946-2001: A New Dataset. Journal of Peace Research 39(5), 615–637.
Hadenius, Axel and Jan Teorell (2007, January). Pathways from Authoritarianism. Journal of Democracy 18(1),
143–157.
Haggard, Stephan and Robert R. Kaufman (2016, August). Dictators and Democrats: Masses, Elites, and
Regime Change. Princeton University Press.
Higley, John and Michael G Burton (1989). The elite variable in democratic transitions and breakdowns.
American sociological review 54, 17–32.
Huang, Reyko (2016, October). The Wartime Origins of Democratization: Civil War, Rebel Governance, and
Political Regimes. Cambridge University Press.
Kadivar, Mohammad Ali (2018, April). Mass Mobilization and the Durability of New Democracies. American
Sociological Review 83(2), 390–417.
Karatnycky, Adrian and Peter Ackerman (2005). How Freedom Is Won: From Civic Resistance to Durable
Democracy. Washington, DC: Freedom House.
Larmer, Miles (2016). Rethinking African Politics: A History of Opposition in Zambia. Abingdon-on-Thames,
UK: Routledge.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
Leventoğlu, Bahar and Nils W. Metternich (2018, July). Born Weak, Growing Strong: Anti-Government Protests
as a Signal of Rebel Strength in the Context of Civil Wars. American Journal of Political Science 62(3),
581–596.
Linz, Juan J and Alfred Stepan (1996). Problems of Democratic Transition and Consolidation: Southern Europe,
South America, and Post-Communist Europe. Baltimore, MD: JHU Press.
Lipset, Seymour Martin (1959). Some Social Requisites of Democracy: Economic Development and Political
Legitimacy. The American Political Science Review 53(1), 69–105.
McAdam, Doug (1986, July). Recruitment to High-Risk Activism: The Case of Freedom Summer. American
Journal of Sociology 92(1), 64–90.
McAdam, Doug (2010, May). Political Process and the Development of Black Insurgency, 1930-1970. University
of Chicago Press.
McAdam, Doug , Sidney G. Tarrow, and Charles Tilly (2001). Dynamics of Contention. Cambridge studies in
contentious politics. Cambridge ; New York: Cambridge University Press.
Metternich, Nils W. , Cassy Dorff, Max Gallop, Simon Weschle, and Michael D. Ward (2013, October).
Antigovernment Networks in Civil Conflicts: How Network Structures Affect Conflictual Behavior. American
Journal of Political Science 57(4), 892–911.
Moore, Barrington (1993). Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making of
the Modern World, Volume 268. Beacon Press.
Nordås, Ragnhild and Christian Davenport (2013). Fight the Youth: Youth Bulges and State Repression.
American Journal of Political Science 57(4), 926–940.
O’Donnell, Guillermo and Philippe C Schmitter (1986). Transitions from Authoritarian Rule: Tentative
Conclusions about Uncertain Democracies. Baltimore, MD: JHU Press.
Parkinson, Sarah Elizabeth (2013, August). Organizing Rebellion: Rethinking High-Risk Mobilization and
Social Networks in War. American Political Science Review 107(3), 418–432.
Pinckney, Jonathan (2018). When Civil Resistance Succeeds: Building Democracy after Popular Nonviolent
Uprisings. Washington, DC: ICNC Press.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Pinckney, Butcher, and Braithwaite
Przeworski, Adam (1991, July). Democracy and the Market: Political and Economic Reforms in Eastern Europe
and Latin America. Cambridge University Press.
Przeworski, Adam , Michael E Alvarez, Jose Antonio Cheibub, and Fernando Limongi (2000). Democracy and
development: political institutions and well-being in the world, 1950-1990, Volume 3. Cambridge University
Press.
Raleigh, Clionadh , Andrew Linke, Håvard Hegre, and Joakim Karlsen (2010, September). Introducing ACLED:
An Armed Conflict Location and Event Dataset: Special Data Feature. Journal of Peace Research 47(5),
651–660.
Ross, Michael L. (2001). Does Oil Hinder Democracy? World Politics 53(3), 325–361.
Rueschemeyer, Dietrich , Evelyne Huber Stephens, Evelyne Huber, and John D. Stephens (1992). Capitalist
Development and Democracy. University of Chicago Press.
Salehyan, Idean , Cullen S. Hendrix, Jesse Hamner, Christina Case, Christopher Linebarger, Emily Stull,
and Jennifer Williams (2012, September). Social Conflict in Africa: A New Database. International
Interactions 38(4), 503–511.
Schock, Kurt (2005). Unarmed Insurrections: People Power Movements in Nondemocracies. U of Minnesota
Press.
Schumpeter, Joseph A (2010). Capitalism, Socialism and Democracy. Abingdon-on-Thames: Routledge.
Smith, Todd Graham (2014, November). Feeding unrest: Disentangling the causal relationship between food
price shocks and sociopolitical conflict in urban Africa. Journal of Peace Research 51(6), 679–695.
Staniland, Paul (2014). Networks of Rebellion: Explaining Insurgent Cohesion and Collapse. Cornell University
Press.
Stepan, Alfred C (ed.) (1989). Democratizing Brazil: Problems of Transition and Consolidation. New York, NY:
Oxford University Press.
Sundberg, Ralph and Erik Melander (2013, July). Introducing the UCDP Georeferenced Event Dataset. Journal
of Peace Research 50(4), 523–532.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
Durability not Diversity
Svolik, Milan W. (2012, September). The Politics of Authoritarian Rule. Cambridge University Press.
Teorell, Jan (2010, October). Determinants of Democratization: Explaining Regime Change in the World,
1972–2006. Cambridge University Press.
Tufekci, Zeynep (2017). Twitter and Tear Gas: The Power and Fragility of Networked Protest. New Haven, CT:
Yale University Press.
Urdal, Henrik (2006, September). A Clash of Generations? Youth Bulges and Political Violence. International
Studies Quarterly 50(3), 607–629.
Van Donge, Jan Kees (2002). Kaunda and chiluba: Enduring patterns of political culture. In J. A. Wiseman (Ed.),
Democracy and political change in Sub-Saharan Africa, pp. 205–231. Abingdon-on-Thames, UK: Routledge.
Wahman, Michael , Jan Teorell, and Axel Hadenius (2013, March). Authoritarian regime types revisited: Updated
data in comparative perspective. Contemporary Politics 19(1), 19–34.
Weidmann, Nils B (2016). A closer look at reporting bias in conflict event data. American Journal of Political
Science 60(1), 206–218.
Weinstein, Jeremy M. (2006, October). Inside Rebellion: The Politics of Insurgent Violence. Cambridge
University Press.
Wood, Elisabeth Jean (2000, October). Forging Democracy from Below: Insurgent Transitions in South Africa
and El Salvador. Cambridge University Press.
Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper Working Paper
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Different prominent theories suggest that regime-challenging collective action is especially conducive to democratization when it is organized by particular urban groups. Yet, systematic large-n evidence that directly tests this proposition, using data capturing the involvement of the supposedly relevant actors, is missing. In this paper, we further develop the argument that both industrial worker and urban middle class movements are conducive to democratization, albeit for somewhat different reasons. Thereafter, we present new data on the social composition of anti-regime opposition movements across the world from 1900–2006. These data enable us to conduct the first large-n empirical study of its kind on this much-debated topic. The results show that movements dominated by one of the two urban groups correspond with increased democratization prospects, both when compared to other movements and to situations without any organized mass opposition. When further differentiating the urban groups and accounting for plausible alternative explanations, the relationship between industrial worker campaigns and democratization turns out to be very robust, whereas the evidence is mixed on the relationship between middle class campaigns and democratization. Further tests suggest that urban opposition campaigns, as expected, relate more strongly to democratization in urban than in rural societies.
Article
Full-text available
Research suggests that nonviolent resistance (NVR) campaigns are more successful in deposing dictators than armed rebellions. However, ousting dictators is only the first step in the process of democratization. After deposing an autocratic regime, societies enter a transition phase where they must learn to consolidate the gains of democracy and bargain about the new rules of the democratic regime. But even if free, fair, and competitive elections are held, indicating a successful transition to democratic rule, uncertainty about its stability remains salient. In the period that follows, either democracy survives and proves to be resilient, or an autocratic backslide occurs. In this article, we analyze the effect of NVR campaigns on the survival of democratic regimes. Building on the literature on modes of transitions and nonviolent resistance, we argue that those democratic regimes that come into being as a result of a NVR campaign are less prone to democratic breakdown. The main mechanism which produces this effect is that the organizational culture of NVR campaigns spills over to the subsequent democratic regime fostering conditions favorable for democratic survival. We test the effect of NVR campaigns on democratic regime survival using survival analysis and propensity score matching. The results show that democratic regimes that experience NVR during the transition phase survive substantially longer than regimes without NVR.
Article
Full-text available
This study takes an organizational approach to understanding the outcomes of civil resistance campaigns. We argue that organizations embedded in extensive interpersonal networks upon which the state partially depends bring ‘leverage’ to civil resistance campaigns and increase the likelihood of major government concessions in the short term. Organizations with strong local ties and a confederal ‘national’ structure bring resilience and reduce the probability of short term failure. Campaigns can shape institutions in the post-conflict period when this leverage and resilience is durable. We argue that, in general, ‘National Trade Unions’ approximate these organizational features. Using new data on the participation of NTUs in civil resistance campaigns, our results suggest that NTU participation increases the likelihood of short term success, decreases the chances of short term failure and improves the prospects of post-conflict democratization. In contrast, campaign size is robustly associated with short term success but big campaigns do not significantly reduce the probability of failure or significantly improve the prospects of post-conflict democratization. Our research suggests that an ‘organizational’ turn is a productive step towards understanding the short and long-term outcomes of nonviolent campaigns.
Article
Full-text available
A growing research field examines the conditions under which major nonviolent resistance campaigns—that is, popular nonviolent uprisings for regime or territorial change—are successful. Why these campaigns emerge in the first place is less well understood. We argue that extensive social networks that are economically interdependent with the state make strategic nonviolence more feasible. These networks are larger and more powerful in states whose economies rely upon organized labor. Global quantitative analysis of the onset of violent and nonviolent campaigns from 1960 to 2006 (NAVCO), and major protest events in Africa from 1990 to 2009 (SCAD) shows that the likelihood of nonviolent conflict onset increases with the proportion of manufacturing to gross domestic product. This study points to a link between modernization and social conflict, a link that has been often hypothesized, but, hitherto, unsupported by empirical studies.
Article
From 1998 to 2005, six elections took place in postcommunist Europe that had the surprising outcome of empowering the opposition and defeating authoritarian incumbents or their designated successors. Velerie J. Bunce and Sharon L. Wolchik compare these unexpected electoral breakthroughs. They draw three conclusions. First, the opposition was victorious because of the hard and creative work of a transnational network composed of local opposition and civil society groups, members of the international democracy assistance community and graduates of successful electoral challenges to authoritarian rule in other countries. Second, the remarkable run of these upset elections reflected the ability of this network to diffuse an ensemble of innovative electoral strategies across state boundaries. Finally, elections can serve as a powerful mechanism for democratic change. This is especially the case when civil society is strong, the transfer of political power is through constitutional means, and opposition leaders win with small mandates.
Article
Vince Boudreau compares strategies of repression and protest in post-war Burma, Indonesia and the Philippines because these alternative strategies shaped the social bases and opposition cultures available to dissidents and, in turn, influenced their effectiveness. He includes first-hand research as well as the the social movements' literature to consider the interactions between the regimes in the wake of repression, and the subsequent emergence of democracy. Boudreau offers a genuinely comparative study of dictatorship and resistance in South East Asia.
Article
Empirical analyses in social science frequently confront quantitative data that are clustered or grouped. To account for group-level variation and improve model fit, researchers will commonly specify either a fixed- or random-effects model. But current advice on which approach should be preferred, and under what conditions, remains vague and sometimes contradictory. This study performs a series of Monte Carlo simulations to evaluate the total error due to bias and variance in the inferences of each model, for typical sizes and types of datasets encountered in applied research. The results offer a typology of dataset characteristics to help researchers choose a preferred model.
Article
This article updates and describes a widely used data set on democracy. Covering 1800-2007 and 219 countries, it represents the most comprehensive dichotomous measure of democracy currently available. We argue that our measure's distinguishing features-a concrete, dichotomous coding and a long time span-are of critical value to empirical work on democracy. Inspired by Robert Dahl, we define a country as democratic if it satisfies conditions for both contestation and participation. Specifically, democracies feature political leaders chosen through free and fair elections and satisfy a threshold value of suffrage. After comparing our coding to that of other popular measures, we illustrate how democracy's predictive factors have evolved since 1800. In particular, we show that economic modernization variables have steadily declined in their correlation with democracy over time.
Article
Wars within states have become much more common than wars between them. A dominant approach to understanding civil war assumes that opposition movements are unitary, when empirically, most of them are not. I develop a theory for how internal divisions within opposition movements affect their ability to bargain with the state and avoid conflict. I argue that more divided movements generate greater commitment and information problems, thus making civil war more likely. I test this expectation using new annual data on the internal structure of opposition movements seeking self-determination. I find that more divided movements are much more likely to experience civil war onset and incidence. This analysis suggests that the assumption that these movements are unitary has severely limited our understanding of when these disputes degenerate into civil wars.