The vast majority of civil wars occur in a small number of countries. What causes conflicts to geographically cluster in this way?
Studies of intrastate armed conflicts show that the majority of civil wars cluster in a small number of states. According to the widely-used Uppsala Conflict Data Program’s Armed Conflict Database, 30 states experienced more than 60 percent of all new armed conflict onsets between 1946 and 2013. In this period, Burma, Ethiopia, India, Indonesia, and Sudan alone account for about 30% of the world’s new ethnic conflicts.
The conflict trap
Conflict researchers and development economists such as Paul Collier attribute the clustering of internal war to state failure and conflict traps: weak states cannot deter rebellion. Civil war, in turn, impoverishes individuals, destroys institutions, and plants feelings of revenge. All of these factors increase the risk of conflict recurrence.
Yet neither India and Burma nor Ethiopia and Indonesia qualify as failed states. Moreover, their political regimes cannot explain the frequency of rebellion either. Burma and Sudan have been repressive autocracies for most of the period but India has been democratic for the vast majority of its existence. Existing explanations, then, do not fully account for why armed conflict clusters in these countries.
Civil war diffusion within states
In a recent study in International Studies Quarterly, my co-author Jesse Hammond and I highlight an alternative explanation for the concentration of so many conflicts in these multi-ethnic states. We explore the diffusion of ethnic civil wars within one country. Unlike earlier research on the diffusion of armed conflict across international borders, we study how government’s decision to fight one rebel group can trigger additional rebellions by rebels from other ethnic groups.
To separate diffusion from recurrence dynamics, we move from country-level to ethnic-group-level analysis. Our study includes all states between 1946 and 2009 that (1) experienced at least one civil war and (2) contain at least three distinct ethnic groups – two in conflict, and one potential challenger. This selection leaves us with 49 states, 415 ethnic groups, and 127 ethnic armed conflicts.
On the basis of this data, we model the yearly probability of a new ethnic conflict breaking out. According to our theory of diffusion, the location of ongoing conflicts as well as the duration and number of armed challengers are the main factors that affect the probability of new conflicts. Nearby conflicts should increase the motivation for additional rebellions; longer conflicts and more rebels should increase the opportunity for fighting.
To construct those measures of motivation and opportunity, we combined data on the geographic location of ethnic groups’ settlement areas from the Geographic Research of War – Unified Platform at ETH Zurich with data on conflict zones from the Peace Research Institute Oslo. Whereas the left panel in Figure 1 shows the settlement areas of ethnic groups in Chad, the right panel shows the extent of an active armed conflict between 1999 and 2002. For these years, we compute the distance between peaceful ethnic group and the conflicts zones and note whether some groups are directly affected by fighting. We repeat this for all ethnic groups in all states in our sample.
Equipped with these measures our study argues that there are four pathways of conflict diffusion within states– two that affect the motivation of potential challengers, and two more that increase their opportunity to rebel.
How armed conflict increases the motivation for additional rebellions
On the motivation side, ongoing fighting may harm members of nearby but previously neutral ethnic groups. Even if fighting does not directly affect other ethnic groups, increased state repression that results from fear of additional uprisings might. In turn, members of previously peaceful ethnic groups become aggrieved about state violence and decide to take up arms to defend themselves. India’s repressive policy in its Northeastern states may have had exactly this effect.
Our second motivational mechanism states that an ongoing civil war encourages already disaffected groups to take up rebellion as a strategy. Witnessing nearby groups’ rebellions provides a blueprint on how to potentially overcome political and economic inequalities such as exclusion from state power.
On its own, political discrimination does not frequently trigger rebellion; disadvantaged groups exist for long periods of time without mobilizing. However, seeing nearby groups with similar political disadvantages rise up against repressive political regimes can provide the spark for additional rebellions.
Patterns of armed uprising against the Burmese and Indonesian states soon after decolonization exemplify these patterns at the domestic level. Although it goes beyond the scope of our study, we argue that similar mechanisms operate at the international level. Although the states in North Africa and the Middle East have been among the most repressive and ethnically discriminatory regimes in the world for decades, Arab citizens only rose up their rulers in 2011 after witnessing the Tunisian revolution.
How armed conflict increases the opportunity for additional rebellions
Turning to our opportunity mechanisms, we argue that ongoing internal armed conflicts can provide important signals about the government’s repressive capacity. If the government is strong, it will crush any rebellion quickly. If it fails to quickly and decisively defeat one rebel organization, other ethnic groups may perceive the government as weak and rebel to gain concessions from the state.
While the 2003 rebellion in Sudan’s Darfur region has various causes, our opportunity logic offers a good explanation for its timing. For two decades, the Sudanese government was unable to decisively defeat the Sudan People’s Liberation Army and its various offshoots. As the southern rebellion endured, aggrieved groups in the Darfur region realized that Khartoum might be vulnerable to extending concessions to them when facing additional violence.
A similar dynamic is at play when the government fights multiple challengers at the same time. The economic and military costs of armed conflict drain governments’ resources. This makes it possible for additional ethnic challengers that were too weak to confront the government alone to join the fray. The increasing number of ethnic challengers in Burma exemplifies this last pathway to domestic conflict diffusion.
To summarize, governments that violently confront rebel groups rather than negotiate enter a slippery slope that may lead to even more civil wars. Armed conflicts with one ethnic rebel group have inspired members of other ethnic groups to rebel in Northeast India, Burma, Indonesia, Sudan, or the Democratic Republic of Congo.
Why then do governments fight rebels rather than accommodate them? One answer may be that government leaders prefer monopolizing power rather than sharing it to extract more resources from the state or to reduce the risk of coups. Where the cost of conflict is not borne by elites but by citizens, such a strategy may pay off.
Other research shows that giving in to rebel demands makes governments appear weak and potentially triggers additional challenges. Future research will have to uncover the exact conditions under which governments prefer one risk over the other.
Our study adds to our understanding of countries caught in conflict traps. We believe that our study’s findings are particularly relevant for counterinsurgency and peacekeeping strategies. In addition to ending one civil war and keeping it peaceful, governments and international institutions need to contain armed conflicts in space. Otherwise, they are very likely to infect other ethnic groups in the same country.
Nils-Christian Bormann is lecturer and Humanities and Social Science Fellow in the Politics Department at the University of Exeter.
Jesse Hammond is assistant professor in the Department of Defense Analysis at the Naval Postgraduate School in Monterey, California.