Support for Counterinsurgencies: Experimental Evidence from a Survey in Iraq

• In Event: Aiding and Abetting: The Role of Non-Combatants in Conflict Zones

Abstract

This study presents unique evidence on popular support for ISIS and the counterinsurgency from an original national survey fielded in Iraq at the height of the insurgency in 2015. As direct question techniques would likely yield biased responses to sensitive questions about views on warring factions in conflict theaters, we administered the first survey experiments in Iraq in order to attain reliable measures of civilian attitudes toward ISIS, the Iraqi government, and the international coalition airstrikes. This research design and data enables us to test an original argument on the effect of civilian victimization in multi-party counterinsurgencies. Our findings run counter to existing public opinion research in rebellious environments.

A near-consensus exists that the outcome of counterinsurgencies depends on the success of each side to win the support of civilians. Recent studies contend that violence against civilians shifts victims’ allegiance away from the perpetrator and toward its opponent (Condra and Shapiro, 2012; Dell and Querubin, 2018), and populations with in-group bias respond less negatively to violence by local rebels than to victimization at the hands of foreign counterinsurgents (Lyall, Blair and Imai, 2013). However, in many recent civil wars (e.g., in Afghanistan, Iraq, Mali, Somalia, and Syria), external interveners fight a rebellion alongside a local government. When the fight against a rebellion is conducted by two actors - the local government and a foreign intervener - the extant theory is less clear regarding how popular support would change in response to violence against civilians by one of these counterinsurgent actors.

In this paper, we argue that an external military intervention that fights an unpopular insurgency changes the way civilians respond to violence inflicted upon them. While the recent literature suggests that violence against civilians committed by the government should lead to a shift in attitudes away from the counterinsurgency and towards the rebellion, we argue civilians do not engage in such a binary choice when a third external actor intervenes militarily – as long as civilians perceive the local government and the foreign intervener as independent actors rather than as close allies. Specifically, we expect that harm against civilians that is inflicted by the government does not shift support away from the counterinsurgency and toward a loathed insurgency if an external intervener also fights this rebellion. Instead, we expect that government-inflicted harm increases support for the external intervention, as long as that intervention is viewed as independent from the government that harms its citizens. Thus, civilians’ perception of the ties between local and foreign counterinsurgents shapes their reaction to harm inflicted upon them by one of them.

Experimental evidence from an original national survey that we conducted in Iraq in 2015 is consistent with this argument. List experiments reveal that popular support for ISIS was extremely low (at 1.4%) while support for U.S.-led international coalition airstrikes was substantially higher at 31.7%. We find that Iraqi civilians do not engage in a binary choice between supporting the insurgency or the counterinsurgency. Our survey experiments reveal that victimization at the hands of government forces - one counterinsurgent actor - does not shift support away from the broad counterinsurgency. Instead, it is associated with stronger support for U.S.-led coalition airstrikes against ISIS - the other counterinsurgent actor - but only among respondents who view the coalition airstrikes as independent of the Iraqi government. In contrast, victimization by government forces does not systematically affect support for U.S.-led coalition airstrikes among those who are primed to view the strikes as tied to the Iraqi government. Victimization by domestic or foreign counterinsurgents also does not increase support for ISIS, contrary to what the existing literature would expect. The survey experimentally varies the cue about the Iraqi government’s endorsement of the airstrikes in order to manipulate the perceived distance between the two counterinsurgents.

This study makes three contributions. First, we show that the relationship between counterinsurgent actors determines how civilians respond to harm inflicted upon them by one of these actors. Second, we test our argument with list experiments in an original national survey in Iraq to causally estimate the effect of randomized cues about the ties between local and foreign counterinsurgents. Third, our findings have important policy implications for counterinsurgents and their external supporters: Being viewed by the domestic public as two independent alternatives to the insurgency can help maintain popular support for the broad counterinsurgency even if one counterinsurgent inflicts harm on the civilian population.

Authors

• Saurabh Pant, University of Essex

• Christoph Mikulaschek, Harvard University