Ethnicity, nonviolent protest, and lethal repression in Africa
| Published date | 01 July 2019 |
| Author | Cullen S Hendrix,Idean Salehyan |
| Date | 01 July 2019 |
| DOI | 10.1177/0022343318820088 |
| Subject Matter | Regular Articles |
Regular Articles
Ethnicity, nonviolent protest, and lethal
repression in Africa
Cullen S Hendrix
Josef Korbel School of International Studies, University of Denver & Peterson
Institute for International Economics
Idean Salehyan
Department of Political Science, University of North Texas
Abstract
Why do governments use deadly force against unarmed protesters? The government’s threat perception may be a
function of the mobilization potential of the opposition and/or the size of the ruling elite’s support coalition. Given
the high salience of ethnicity in African politics, governments that depend on small ethnic coalitions will see peaceful
protests as more threatening, as the opposition may be able to draw on larger numbers of potential dissidents and
excluded groups. Alternately, governments with larger, more homogeneous ethnic coalitions will find nonviolent
mobilization less threatening and will be less likely to respond with deadly force. Using the Social Conflict Analysis
Database, we demonstrate that as the size – and to a lesser extent homogeneity – of the ethnic ruling coalition grows,
governments are significantly less likely to use deadly force against nonviolent protesters. This finding is robust to
several operationalizations of the size of the government’s support coalition, the inclusion of other measures of ethnic
demographics, and estimators that account for the hierarchical nature of the data. Threat perception hinges not only
on dissident tactics but on their demands, their mobilization potential, and their capacity to impose costs on the
government. This article demonstrates that the size and composition of the government’s ethnic support base matters
as well.
Keywords
Africa, ethnicity, protest, repression
Introduction
The decision by governments to kill nonviolent protest-
ers represents one of the most egregious violations of
international human rights norms and law. Articles 3
and 20 of the Universal Declaration of Human Rights
guarantee to all people the rights of ‘life, liberty and
security of person’ and ‘peaceful assembly and associa-
tion’ (United Nations, 1948). While governments can
be expected to respond to violent challenges with force,
the use of such force against unarmed protesters is puz-
zling. Given that the opposition uses peaceful means,
why then would governments respond with violence?
The rapacity and brutality of the regime might offer
partial explanations. However, governments are more
rational and strategic than this argument would suggest
(Sutton, Butcher & Svensson, 2014). Threat perceptions
by the government are multifaceted and take into
account a variety of factors related to protest movement
success. We argue the size and cohesion of the ruling
coalition affect its ability to endure popular protest and
overcome collective action problems in formulating a
response to social movements.
While it is uncommon, governments do use deadly
force against peaceful demonstrators. Between 1990 and
2011, roughly3% of nonviolent protests,demonstrations,
Corresponding author:
cullen.hendrix@du.edu
Journal of Peace Research
2019, Vol. 56(4) 469–484
ªThe Author(s) 2019
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/0022343318820088
journals.sagepub.com/home/jpr
and strikes across Africa were met withlethal force, result-
ing in at least 1,200 deaths.
1
If we include those protests,
demonstrations, and strikes th at began peacefully but
escalated to violence – in many cases, as a result of gov-
ernment repression – lethal force is used 7% of the time
and the death toll jumps to over 7,000. Given its contra-
vention of international norms and potential to inflame
further protests, why do governments use deadly force
against peaceful protests?
The ‘law of coercive responsiveness’ holds that more
severe threats to government authority will elicit repres-
sion (Davenport, 2007). Yet, perceived threat to the
regime is not limited to the use of violence by dissidents.
One example occurred in Guinea in 2009, after Moussa
Dadis Camara took power in a coup d’e
´tat. After pro-
mising to act as a caretaker and step down following a
democratic transition, Camara reneged, choosing to
stand for election. Defying a ban on pre-election pro-
tests, thousands marched peacefully against his decision,
eventually congregating in a football stadium in Con-
akry. Chaos erupted as members of the Guinean Presi-
dential Guard opened fire, resulting in 157 deaths and
over 1,000 people injured.
Camara had been an unpopular leader, in part because
he is a member of a minority ethnic group. Camara is
Kpelle, a group which makes up 5% of Guinea’s popu-
lation, and which ruled in coalition with other minority
ethnic groups in a fractious coalition. Thus, Camara’s
perceptions of threat may have centered on the rising
popularity of opposition leader and former prime min-
ister Cellou Dalein Diallo, a member of the Peul (Fulani)
ethnic group that comprises 40% of the population.
2
We
find the Guinean case to be emblematic of broader
trends: governments supported by small, diverse ethnic
coalitions are much more likely to respond to nonviolent
protests with deadly force. While elections in Guinea
jeopardized incumbent rule most vividly, protests in gen-
eral can signal discontent against governments and pose a
threat to their longevity.
We argue that the government’s perception of threat
is a function of the mass mobilization potential of the
opposition as well as the size and concentration of the
ruling coalition. Given the salience of ethnicity in Afri-
can politics, governments that depend on relatively small
ethnic constituencies will likely see peaceful protests as
more threatening, given that sizeable ethnic groups are
excluded from power and can potentially be mobilized
by the opposition. Governments that are based on rule
by larger, more ethnically homogeneous coalitions are
less likely to be threatened by nonviolent protest and
will tolerate such dissent.
However, the size of the ruling coalition relative to the
opposition is not the only dimension to consider. The
difficulty of collective action across ethnic lines also mat-
ters. For ruling coalitions, more diverse membership
undermines unity and loyalty to the regime; for opposi-
tion activists, forming a mu lti-ethnic alliance is mo re
difficult than organizing within groups.
We demonstrate that as the size and, to a lesser
extent, cohesion of the ethnic ruling coalition grow, the
use of deadly force against nonviolent protesters
declines. Our estimates suggest a government with a
majority ethnic support coalition (51% of the popula-
tion) would be roughly 47% less likely to use deadly
force against nonviolent protesters than one such as
Camara’s. This finding is robust to several operationa-
lizations of size and diversity, various estimators, and
the inclusion of controls.
Our empirical approach is based on event data across
countries in Africa (Salehyan et al., 2012). Much of the
literature on repression employs annual, cross-national
data to look at aggregate levels of abuse, but this
approach tells us little about the actors and events that
are targeted for repression. Other approaches look at
dynamic sequencing and government responses to dis-
sent over time, but typically focus on a small set of
countries (Moore, 1998). Using the Social Conflict
Analysis Database, we can account for structural factors
affecting the probability of repression and attributes of
the contentious events themselves.
These findings are part of a body of research on eth-
nicity and ethnic exclusion in contentious politics. Fol-
lowing Fearon & Laitin’s (2003) study, which found no
general relationship between ethnic diversity and civil
war onset, researchers have sought to identify the condi-
tions under which ethnicity forms a basis for mobiliza-
tion – either peaceful or violent – and how ethnicity
conditions responses to mobilization (Buhaug, Ceder-
man & Rød, 2008; Cederman & Girardin, 2007; Ceder-
man, Wimmer & Min, 2010; Cederman, Weidmann &
Gleditsch, 2011; Cederman, Gleditsch & Wucherpfen-
nig, 2017; Fearon, Kasara & Laitin, 2007; Heger &
Salehyan, 2007; Østby, 2008). Our contribution moves
1
Authors’ calculations based on the Social Conflict Analysis
Database, version 3.0 (http://www.scaddata.org), henceforth
SCAD. These data were previously referred to as the Social
Conflict in Africa Database; the name change reflects recently
expanded geographic coverage. In this article we focus exclusively
on sub-Saharan and North Africa.
2
Camara was eventually deposed by an assassination attempt in
December 2009.
470 journal of PEACE RESEARCH 56(4)
To continue reading
Request your trial