Loyalty Strategies and Military Defection in Rebellion

This article examines military defection (whether government soldiers, instead of fight ing for the regime, desert or fight for the opposition) during rebellions in Syria, Jordan, and Iran. It distinguishes between two strategies to maintain military loyalty in prae torian regimes?individual incentives (reward and punishment) and a policy of ethnic preference in the armed forces. These strategies produce very different outcomes for defection when a rebellion arises outside the military. An individual strategy is vulner able to a cascade of defection sweeping across the whole army when a rebellion breaks out. A group-based strategy, however, may generate out-group defection, but in-group defection is much less likely. This argument focuses on the availability of information about soldiers' preferences and about the likelihood that the regime will survive. A focus on information, in turn, highlights two different self-fulfilling prophecies. In an individualized incentive system, soldiers' true preferences are hidden. They will support the regime if they believe it will survive, and will defect if they do not; but the only way of making this judgment is based on others' behavior. Thus the belief that the regime will collapse provokes a cas cade of defection, bringing about the very collapse it predicts. In a group-based system, the belief that in-group members have a proregime preference and that out-groups are opposed to the regime helps generate precisely those preferences. Thus preferences can become public by matching public prejudice. The result is a durable cleavage between in-group and out-group, where out-group soldiers are likely to defect but in-group sol diers are likely to remain loyal. These self-fulfilling prophecies may help explain why observers sometimes over estimate the resilience of some regimes and underestimate the resilience of others.1 In essence, we have lacked a theory that takes into account the way political judgments change due to rebellion itself. Some regimes seem generally successful in quelling dis sent (for example, the Shah's Iran prior to 1978) whereas others seem prone to out group anger (for example, Jordan in the 1960s and Syria in the 1960s and 1970s), and observers believe that the former can stand up to rebellion more easily than the latter. However, in the latter cases, the visible antiregime preferences of out-groups imme diately suggest regime weakness but also solidify in-group loyalty, providing a bulwark for the regime against rebellion. In the former case, people's preferences are not visible, and the viability of a regime hinges upon the judgment that the regime will remain in power?a judgment that is vulnerable in the face of a rebellion. 333