Guest Post by Michael Ward

I asked Michael Ward to write about his article appearing in the October issue of AJPS.  His piece “Antigovernment Networks in Civil Conflicts: How Network Structures Affect Conflictual Behavior” is co-authored with Nils W. Metternich , Cassy Dorff , Max Gallop and Simon Weschle.  He writes:

• Rarely is political conflicts between two parties. In the case of the civil war in Iraq, for example, as many as 19 different factions engaged in violent interactions, including the Islamic Army in Iraq, Al Qaida in Iraq, the Jihadist Leagues, and the Just Punishment Brigades. In Syria, we see a similar picture, including the Free Syrian Army, the Syrian Liberation Front, the Syrian Islamic Front, and Jabhat al-Nusra, plus many others. The usual approach to understanding these kinds of situations treats all the varied rebel forces as a single entity, united against the government. Yet neither rebels nor even the government forces are unified and monolithic. We explore a theory of the interactions among factions in order to better understand what is likely to happen in conflict situations. Our analysis combines both strategic calculation and the role of networks in predicting conflict. The basic insight is the old saw, often attributed to the 6th century (BCE) Chinese general, Sun Tzu: hold your friends close, and your enemies closer.
•  One approach to understanding multilateral conflicts is that if there are lots of different groups, each group can gain by uniting with the others. This might be called the all-for-one and one-for-all approach.  By contrast our network model, assumes that if there are two groups that are close to each other in terms of their goals and ideologies, each is likely to assume the other will undertake any desired, yet costly action. This dilemma is often described as “free riding,” wherein my group is willing to let your group pay all of the costs of conflict, especially if we share in the benefits of winning without having to be directly engaged. The implication is that unity may bring inaction, and isolation may bring greater engagement in the conflict. In short, groups are more likely to be engaged when they are isolated from other groups, not when they are closely aligned with them.
•  Our research examines this idea in the case of Thailand (2001-2010), which witnessed two major ongoing conflicts, with changing alliances involving multiple groups. We collect textual descriptions of the ebb and flow of interactions among the various (four dozen) government and anti-government groups in Thailand. These are analyzed to see whether the periods characterized by free-riding networks witnessed less conflict. Our empirical analysis shows that this is the case from 2001 to 2009. Against conventional wisdom, conflict is lower when group unity and the potential of free riding was greatest. We then take this exact model and test whether it stands up against new data, collected for the first ten months of 2010. These predictions show that the network model of free riding in multilateral conflicts is an improvement over models that only have structural aspects such as economic growth, characteristics of the regime, and the existence of conflict in the region.
•  The data used in this research is based on a re-emerging methodology which utilizes textual information from news sources to capture data on events in different countries all over the world. In addition, it takes advantage of the network effects in conflict situations, rather than treating each actor and action as independent. Further, it uses current advances in strategic thinking (i.e., game theory) to specify what should be expected in the network. We use the case of Thailand to show that free riding in networks does affect conflict dynamics. In fact, the model is sufficiently robust that it allows better predictions for new data, than do standard approaches. This work was undertaken as part of a research project aimed at creating a set of models that help to predict and understand political events around the world. It was partially funded by the Defense Advanced Research Projects Agency (DARPA) and the Office of Naval Research (ONR), under a program widely known as ICEWS. The results and findings are however, solely, those of the authors, not the sponsors. Additional projects of our group can be found at https:\\www.mdwardlab.com.