Relative Capability and Rebel Objective in Civil War
| Author | Halvard Buhaug |
| Published date | 01 November 2006 |
| Date | 01 November 2006 |
| DOI | http://doi.org/10.1177/0022343306069255 |
| Subject Matter | Articles |
691
The more restricted the strength, the more
restricted its goals must be.
Carl von Clausewitz (1832/1984: 283)
Introduction
Armed civil conflicts come in two types.
Some conflicts involve groups aiming at
overthrowing the government or in other
ways modifying the political system. Military
coups d’état and popular revolutions belong
to this category. Other conflicts concern the
control and righteous authority of a limited
territory within the state. These conflicts are
usually labeled territorial or separatist rebel-
lions. While some conflicts may seem to
include elements of both, rebel groups rarely
fail to issue some form of political demand
that corresponds to one, and only one, of
these two archetypes. For example, the
supreme, explicit goal of the Free Aceh
Movement (also known as Acheh Sumatra
National Liberation Front, ASNLF) in
Indonesia was to ensure ‘the survival of the
people of Acheh Sumatra as a nation’
© 2006 Journal of Peace Research,
vol. 43, no. 6, 2006, pp. 691–708
Sage Publications (London, Thousand Oaks, CA
and New Delhi) http://jpr.sagepub.com
DOI 10.1177/0022343306069255
Relative Capability and Rebel Objective in Civil War*
HALVARD BUHAUG
Centre for the Study of Civil War (CSCW), International Peace Research
Institute, Oslo (PRIO)
When all else fails, aggrieved groups of society often resort to violence to redress their grievance – either
by seeking to overthrow the ruling government or by attempting to secede. The strength of the rebel
group relative to the state determines what direction the conflict will take. In institutionally and econ-
omically capable countries, any opposition group is likely to be inferior to the government. These
groups will see secession as the most viable strategy to improve living conditions. Inconsistent, poor,
and resource-dependent regimes are typically quite unstable and should therefore be more likely to
attract coups and revolutions. In addition, large and ethnically diverse countries contain a higher
number of peripheral and possibly marginalized groups, as well as remote and inaccessible terrain, both
of which are expected to favor secessionist insurgency. Smaller countries, in contrast, offer few oppor-
tunities for separatist claims but, in such countries, capturing the state might also be a more realistic
objective. This article provides a first test of these presumptions by estimating the effect of several
popular explanatory factors separately on the risk of territorial and governmental conflict, 1946–99.
The analysis offers considerable support and demonstrates that territorial and governmental conflicts
are shaped, in large part, by different causal mechanisms. The reputed parabolic relationship between
democracy and risk of civil war only pertains to state-centered conflicts, whereas democracy has a pos-
itive and near-linear effect on the risk of territorial rebellion. Moreover, the analysis strongly suggests
that the puzzling no-finding of ethnicity in several prominent studies is affected by their inability to
account for rebel objective in civil war.
* E-mail address: halvardb@prio.no. Replication data are
available from http://www.prio.no/jpr/datasets. I thank
Sabine Carey, Indra de Soysa, Nils Petter Gleditsch,
Håvard Hegre, Philip A. Schrodt, and colleagues at the
CSCW and the Norwegian University of Science and Tech-
nology (NTNU) for valuable comments on previous
drafts. I am grateful to the Research Council of Norway
and the CSCW for financial support.
through separation from the ‘Javanese colo-
nialists’.1Similarly, the Revolutionary
Armed Forces of Colombia – People’s Army
(FARC-EP), the dominant guerrilla group in
the country, has issued numerous documents
and pamphlets opposing privatization of
natural resources, denouncing US inter-
ference in domestic affairs, and generally
calling for ‘a Colombia for Colombians, with
equality of opportunities and equitable
distribution of wealth’.2And even the Revol-
utionary United Front (RUF) in Sierra
Leone, which in many regards seemed more
like a loose gang of criminals preoccupied
with looting of diamonds for personal profit,
released a political manifesto stating that ‘it
is our right and duty to change the present
political system in the name of national sal-
vation and liberation’.3
Previous research on civil war has gener-
ally failed to acknowledge the government–
territory distinction. Nonetheless, there are
ample reasons to expect various state- and
group-level characteristics to be associated
with these types of conflicts in different
ways. Therefore, an aggregated research
design is likely to diminish or even conceal
important causal relationships that apply
only to conflicts of one kind. In the follow-
ing section, I elaborate on the distinction
between governmental and territorial con-
flicts and discuss why some groups seek state
control while others claim self-determina-
tion. I argue that a significant determinant of
rebel objective is the capability of the rebels
relative to the government. Objective is here
understood as the ultimate stated goal of the
rebellion. I then develop five testable
hypotheses on how characteristics of the state
are linked to type of civil conflict. These are
evaluated through multinomial logit regres-
sions of outbreak of governmental and sepa-
ratist conflicts, 1946–99. The results support
the notion that institutionally consistent
states are better able to avoid state-centered
unrest than territorial conflict, relative to the
inconsistent regimes. Economic capacity,
too, shows a relatively stronger negative
effect on governmental conflict, although
only for major civil wars. Moreover, the
analysis presents strong and robust evidence
that country size and ethnic fractionalization
only affect the propensity for separatist rebel-
lion, thus offering important insight into the
puzzling no-finding of ethnicity in studies of
civil war in general.
Governmental and Territorial
Conflicts
According to the Uppsala/PRIO Armed
Conflicts project (Gleditsch et al., 2002),
roughly two-thirds of all intrastate conflicts
since 1946 have been challenges to the
central government, the remaining being
classified as territorial disputes. Governmen-
tal conflicts are here understood as concern-
ing ‘type of political system, the replacement
of the central government, or the change of
its composition’. Territorial conflicts are
defined as involving ‘demands for secession
or autonomy’ (2002: 619). Figure 1 further
shows that while both the rate of new wars
and share of territorial conflicts remained
fairly stable throughout the Cold War, the
1990s saw a dramatic increase in both the
absolute and relative frequencies of separatist
rebellion. A similar pattern can be found in
Fearon & Laitin’s (2003) data on major civil
wars. This trend now seems reversed. The
annual average number of new conflicts in
the 21st century is not much higher than it
was twenty or thirty years ago, and the share
of conflicts over territory is again at about
one-third.
The distinction between territorial
and policy-related issues has a long and well-
journal of PEACE RESEARCH volume 43 / number 6 / november 2006
692
1See http://www.asnlf.net/topint.htm (accessed 7 March
2006).
2See http://www.farcep.org/pagina_ingles/ (accessed 7
March 2006).
3See http://www.sierra-leone.org/footpaths.html (accessed
7 March 2006).
To continue reading
Request your trial