How migration intentions change during periods of political instability and violence: Panel survey evidence from Kenya

Abstract Forced migration research suggests that violence plays a central role in individuals’ decision to flee. However, research has not yet examined how individuals form migration decisions in violent contexts over time. I argue that forced migration decision making consists of anticipatory and reactive processes. Distinguishing these stages improves our understanding of seemingly contradictory evidence on the violence–displacement link. I posit that some individuals anticipate security risks based on personal characteristics such as affiliations with specific identity groups and based on previous experiences. Consequently, they consider leaving due to security concerns relatively early and are likely to flee in anticipation of violence. Others will only react to direct consequences of political conflict, but leave quickly when they experience violence regardless of previous intentions. To evaluate this argument empirically, I analyze an original panel dataset among the adult population of Nairobi and Mombasa which tracks individual migration considerations and actual displacement during the violent 2017 Kenyan elections over time. The longitudinal design disaggregates migration decision making and studies when and for which reasons people consider leaving as well as when individuals implement their plans to leave their homes. The results underscore the theoretical argument: Some individuals react quickly to changing events and start to consider leaving their homes. Moreover, these considerations translate into action: individuals who report security-related migration intentions are much more likely to flee. Regardless of the previous migration plans, however, a second group of individuals flees once they personally experience violence. The decision logic mirrors two-step models of non-conflict migration decision making.