Collaborative Research: Simulating the Dynamics of Insurgency

Historical evidence suggests that once insurgencies become well-established, resolving them can take several years, is often quite bloody, and may not be definitive (many insurgencies recur). In many cases, the cost of conducting a prolonged counter-insurgency campaign is very costly for governments while producing only marginal success. On the other hand, while we often focus on long and bloody insurgencies, history also demonstrates that most nascent insurgencies actually fail quickly. This project develops a series of agent-based computer simulation models to aid in understand some of the complexities associated with defeating insurgencies and determining which are likely to drag on, and which will be defeated quickly. In particular, these models will help us address three questions. First, what leads some insurgencies to become well established and resistant to collapse, while others fail quickly? Second, when do insurgencies reach the point where they are likely to survive, despite government repression? Finally, how can governments effectively adapt their counter-insurgency strategies to stem the growth of insurgency?

The simulation models the early stages of insurgency where the interactions of civilians, insurgents, and soldiers (government forces) are critical. At early stages of insurgency, civilians constitute an audience for insurgent-soldier interactions, and can join the insurgency or be deterred from joining it. Actions by insurgents who can attack government targets, and in turn be attacked and captured or killed, set the stage for civilians to change loyalties. Initial simulations have revealed the critical importance of accuracy on the part of governments/soldiers who seek to capture insurgents, since government military actions against insurgents sometimes backfire. A series of extensions to the initial simulation model are developed, focusing on insurgent and government recruitment, learning, and the interactions of multiple insurgent groups.

The development of the simulation models is coupled with statistical analysis to determine whether the patterns observed in the simulations are also present in international politics. The project examines a population of insurgencies and sees whether factors that the model suggests will predict patterns of insurgencies in fact do so. In particular, insurgency duration, the speed at which insurgency grows, and the scope of penetration of the insurgency into the local population are examined. Along with construction of various proxy measures of the models? concepts, the project incorporates the collection of new data about insurgencies and counter-insurgency efforts to be able to better test the hypotheses that emerge from the simulations.

The broader impacts of the proposal include the revelation of important patterns of duration and termination of insurgency, both from the simulation and empirical analysis, and confirming the presence of these patterns and relationships between variables in the international system. In particular, understanding the circumstances under which different types of government actions may be useful or counterproductive may offer specific policy advice, with empirical backing. Understanding these patterns and relationships may help in the design of appropriate strategies for confronting existing insurgencies and avoiding their emergence in the first place.