Why do violent substate conflicts spread?

  • Spring 2011

Nathan Black is a Ph.D.candidate in the Department of Political Science at MIT and an affiliate of the Center's Security Studies Program. His dissertation seeks to explain the spread of violent coups and insurgencies across international borders. Other research interests include the security consequences of climate change and political leadership and decision-making.

By Nathan Black
May 1, 2011

The recent collapse of the authoritarian regimes in Tunisia and Egypt—and the wave of anti-authoritarian protests in numerous other Arab countries—has reacquainted scholars and policymakers with a long-known truth: Unrest in one country sometimes spreads to another. The cascade of revolution from Tunisia to Egypt shares important similarities with the cascade of Eastern European revolts against de facto Soviet control in the 1980s, reminding us that the internal politics of a state are not always entirely determined by internal forces.

Fortunately, the protests in the Arab world have been mostly nonviolent. Although hundreds of protesters died in both Tunisia and Egypt at the hands of repressive governments,1 ultimately dissidents never mobilized to resist those governments militarily. But the fruits of this wave of protest in Libya have been significantly more tragic. This also is a reminder: that there is a darker side to the international spread of civil unrest. My dissertation research is focused on a particular aspect of this dark side. I seek to explain why a substate conflict in one state—anti-government fighting involving militarized rebels that has killed at least 25 people in battle—sometimes touches off a civil war, or anti-government fighting involving militarized rebels that has exceeded 1,000 cumulative battle-related deaths, in another state. For example, the spread of substate conflict from Rwanda to what was then Zaire in 1996 has since cost millions of lives, and the spread of substate conflict in the Balkans evoked tragedy on a similar scale.

U.S. foreign policymakers seem to live in perpetual fear that the spread of unrest from one state to another will look less like Egypt and more like Zaire. In fact, nearly every U.S. military intervention since the end of World War II has been predicated, in whole or in part, on fears about the nature of the spread of substate conflict that often proved to be wide of the mark. The U.S. went to Vietnam, the far-flung outposts of the Reagan Doctrine, and to a lesser extent Korea because of fear of a "domino effect" of communist revolution; she went to the Balkans for fear of a bloody ethnonationalist conflict that, in President Bill Clinton's words, "could spread like poison throughout the entire region";2 she "stayed the course" in post-2003 Iraq to keep Islamist extremists from "toppling" neighboring governments, according to President George W. Bush;3 and a similar logic partially motivates today's mission in Afghanistan. This spring's intervention in Libya is no exception—in his address to the nation of March 28, President Obama said that a massacre in Benghazi "would have driven thousands of additional refugees across Libya's borders, putting enormous strains on the … fragile transitions in Egypt and Tunisia." If we took these presidential rationalizations at their words, we would have to conclude that substate conflicts spread constantly and uncontrollably, like a virus in a crowded room of unvaccinated toddlers, and that the result is almost always bloody civil war.

My research into the causes of "substate conflict contagion" suggests quite a different view of this mysterious and important phenomenon. I have found that the most violent cascades of substate conflict—those that result in full-fledged civil wars—can generally only occur when a sovereign state government takes actions that allow the contagion to happen. Three specific state government actions seem particularly conducive to contagion: (1) a state being taken over by rebel forces which, now at the reins of the state, try to export their revolution abroad; (2) a state government expelling rebel fighters from their borders, driving the fighters into another state which they proceed to destabilize; or (3) a state government meddling—e.g., supporting rebels—in another state's conflict, resulting in the "boomeranging" of conflict back to the meddling state.

Without these state actions, substate conflict contagion is not impossible, but it is highly unlikely. Therefore, contrary to their own beliefs about the spread of substate conflicts, American policymakers have some control over whether a given conflict will spread. By using coercive diplomacy against sovereign state governments—an imperfect tool, but one in which Washington is well practiced — policymakers can prevent the state actions that enable substate conflict contagion and thereby prevent the contagion itself in most cases.

In the remainder of this essay I will first elaborate the basic principles of "State Action Theory" in more detail and discuss the empirical support for the theory. Then I will apply the theory to the current situation in the Arab world.



State Action Theory derives its predictions from a single initial premise: Spreading large-scale substate conflict is hard. For one thing, when violent intrastate conflicts break out, they do not automatically inspire potential rebels in neighboring states to take up arms as well. In fact, substate conflicts seem just as likely, if not more so, to horrify people around the region with their brutality and futility. For instance, potential rebels in Mozambique were at one point uninspired by the bloody civil war in Angola.4 Also, once a substate conflict starts in one state in a region, the governments of the other states in that region are more alert to the possibility of such a conflict in their own state. Hence neighboring governments can take actions to "fortify" themselves against the spread of conflict from the original state.5

These natural obstacles to substate conflict contagion theoretically can be overcome through a variety of means and by a variety of actors. In general, though, nonstate actors alone cannot cause substate conflict contagion, because the obstacles to its occurrence are too great. This insight contradicts the current conflict contagion literature, which emphasizes the importance of nonstate factors such as transnational rebel networks and flows of refugees.6 Instead, I argue that the most effective means to overcoming the formidable obstacles to contagion lie within the exclusive power of state governments, which have military and economic resources far exceeding those available to nonstate actors.

Specifically, relative to nonstate actors, state governments have enhanced capability to take the following three actions, each of which facilitates substate conflict contagion:

(1) State governments which have been taken over by rebels can evangelize their revolution to other states, by providing arms and training to nascent rebel groups abroad and thereby helping them overcome the natural obstacles to full-fledged civil war. Examples include Communist China's support for the Naga rebels in India in the 1950s, and Liberia's support (under Charles Taylor) for the rebellion in Sierra Leone.

(2) State governments fighting an insurrection at home may expel rebel combatants into another state. The result is often that these rebels provide the manpower and experience necessary to start a civil war in the receiving state. Examples include Uganda's expulsion of the future Rwandan rebels back to Rwanda, and the Afghan expulsion of jihadists such as Abu Musab al-Zarqawi to Iraq.

(3) State governments also sometimes meddle with overt partiality in the internal conflicts of other states, taking the side of either the rebels or the government and providing tangible assistance, such as arms or basing rights. This meddling can backfire in one of three ways. First, the state that is the target of the meddling may retaliate, starting a civil war in the meddling state (as did Rwanda in the former Zaire). Second, the meddling state's own population may be angered by the meddling and start a civil war (as did Somalis after the failed war in Ethiopia). Third, the meddling may enable the rebels in the target state to expand their conflict into the meddler's territory (as did the Afghan Taliban into Pakistan, its most prominent state sponsor).

If these state actions are necessary though not necessarily sufficient—for most substate conflict contagion to occur, then it means that contagion is a more preventable phenomenon than the conventional wisdom suggests. The U.S. can use its significant leverage over other state governments to pressure them to refrain from evangelization, expulsion, and meddling. State-to-state coercion of this kind is certainly not easy, but it is easier than keeping dissidents from talking to each other or controlling mass population flows.

My empirical research so far lends strong support to this simple and relatively optimistic theory of substate conflict contagion. I have identified 82 cases of substate conflict contagion between 1946 and 2007—cases in which a substate conflict contributed causally to a civil war in another state, ultimately involving at least 1,000 battle-related deaths. At least one of the state actions described above was involved in 64 of these 82 cases (78 percent). In only 18 cases were nonstate factors such as rebel networks sufficient to create a cascade of conflict leading to civil war.

Furthermore, the theory seems to explain cases in which substate conflict contagion did not occur. In Central America between 1978 and 1996, a substate conflict only led to the outbreak of civil war in another country once—in El Salvador, where guerillas were aided by the Sandinista regime that had just overthrown the Nicaraguan state (a case of evangelization). Contagion did not occur elsewhere despite numerous nonstate risk factors: transnational ethnic ties between disaffected social groups, arms flows between states, and significant refugee populations spilling from conflict zones into peaceful neighbors. This pattern suggests that evangelization, expulsion, and meddling on the part of state governments are the key enablers of this dangerous phenomenon.



If we accept State Action Theory, what can we say about the recent tumult in the Arab world? First, we should note that nonviolent protest movements and violent substate conflicts are different phenomena that we should expect to spread under different conditions. Seemingly without the aid of the state actions described above, nonviolent protests spread rapidly from Tunisia to Egypt to a variety of other Arab states—yet because my theory is about the spread of militarized resistance movements, this is neither evidence for nor against the theory.

Second, however, I cannot ignore Libya. There dissidents did arm, resulting in a full-scale civil war. While State Action Theory could not have predicted the onset of this war, since it was spurred by a foreign nonviolent protest movement rather than a foreign substate conflict, it can predict the conditions under which the Libyan violence is likely to spread. Unfortunately, the prognosis is not good. One of the state actions enabling contagion is already present—the United States, France, and Great Britain have meddled in the Libyan conflict by supporting the anti-Qaddafi rebels. While this meddling is unlikely to adversely affect the internal stability of the major Western powers—advanced democracies far removed from the conflict zone—the Arab states supporting this meddling, such as Qatar,7 are at an increased risk of major violent conflict according to State Action Theory. For instance, Qaddafi could retaliate against Qatar by supporting an insurrection there. We also know little about the rebels who aspire to control the Libyan government. They could easily turn out to be expansionists or "Pan-Africanists" in the vein of Mao Tse-Tung, Charles Taylor, or Qaddafi himself, unleashing a new wave of evangelization on the region.

This brings us to a final policy implication of this research. While substate conflict contagion is significantly more preventable than many scholars and policymakers currently believe, the means of prevention are important. Coercive diplomacy—for instance, economic sanctions or the threat of force—is likely a more efficacious means than military intervention for major states trying to prevent other states from evangelizing, expelling, or meddling. Because military intervention is usually itselfmeddling with overt partiality, it carries with it the potential for significant security consequences, either for the principal meddler or for its weaker supporting allies.

So the next time a substate conflict appears in the world, U.S. policymakers should not just wring their hands. But neither, conversely, should they undertake a military intervention designed to head off contagion. Instead, they should try to use their non-kinetic leverage to keep state governments from evangelizing, expelling, or meddling. In so doing, policymakers stand a good chance of staving off the most dangerous international consequences of substate conflicts at a reasonable cost in blood, treasure, and legitimacy.



Associated Press, "Human Rights Watch: 297 Killed in Egypt Protests," February 7, 2011; Voice of America News, "UN Reports Higher Death Toll from Tunisia Unrest," February 1, 2011. 

2 Quoted in Michael E. Brown, "Introduction," in Michael E. Brown, ed., The International Dimensions of Internal Conflict (Cambridge, Mass.: MIT Press, 1996), pp. 23-24.

3 President George W. Bush's address to the nation, January 10, 2007.

4 Stephen John Stedman, "Conflict and Conciliation in Sub-Saharan Africa," in Michael E. Brown, ed., The International Dimensions of Internal Conflict (Cambridge, Mass.: MIT Press, 1996), pp. 248-249.

5 Stephen M. Walt, Revolution and War (Ithaca: Cornell University Press, 1996), p. 335.

6 See, for example, Halvard Buhaug and Kristian Skrede Gleditsch, "Contagion or Confusion? Why Conflicts Cluster in Space," International Studies Quarterly, Vol. 52, No. 2 (2008): 215-233; and Erika Forsberg, Neighbors at Risk: A Quantitative Study of Civil War Contagion (PhD Dissertation, Uppsala University, 2009).

7 Mark Landler and Steven Erlanger, "Obama Seeks to Unify Allies as More Airstrikes Rock Tripoli," The New York Times, March 22, 2011.