This report employs frequentist statistical analysis in order to model the effects of various factors, including the type of actors (state, violent non-state actor (VNSA) or civilian) involved and the prevalence of kinetic activity, on (de-)escalatory trends in Gray Zone conflicts. This is coupled with the development of a Bayesian Belief Network for predictive analysis of White, Gray and Black Zone behavior within Gray Zone conflicts.
Both sets of analyses utilize a version of the event-level data from the Worldwide Integrated Crisis Early Warning System (ICEWS). However, we heavily modified this data prior to running the analyses. Specifically, we recoded new variables of particular interest to the study of Gray Zone conflict, addressed erroneous and duplicate entries, and restructured the data in order to model temporal changes. This was accomplished using a hybrid process involving both automated recoding procedures and expert human coders.
This report's procedure was applied to three diverse gray zone conflicts: Colombia (01 January 2002 to 19 September 2016), Libya (01 January 2011 to 12 September 2016) and Ukraine (01 January 2014 to 12 September 2016). These conflicts all share two commonalities: they all entail a large amount of Gray Zone activity and myriad VNSAs. Nevertheless, the three cases vary in a number of important respects: the level of foreign involvement, the belligerents’ motives, as well as their guiding ideologies, and their geographic location. Consequently, the results are highly likely to be generalizable to a diverse array of other Gray Zone conflicts.
Three principal findings hold across both methodological approaches and are apparent in multiple cases. First, contrary to popular belief, kinetic military operations are a key aspect of Gray Zone conflicts. While it is true that these events are relatively sparse (around 20% of all events depending on the case), they have substantial influence in shaping non-kinetic events. Second, while VNSAs are less proficient than states at identifying their adversaries (de-)escalation trends, the closer VNSAs are linked to states, the less that this is a problem. Finally, legitimacy matters. For this reason, both VNSA and state forces will moderate their behavior in order to avoid being perceived as the aggressor or engaging in more (easily visible) civilian victimization than their opponents.
Koven, Barnett S., Varun Piplani, Steve Sin, and Marcus Boyd. “Quantifying Gray Zone Conflict: (De-)escalatory Trends on Gray Zone Conflicts in Colombia, Libya and Ukraine,” Report to DHS S&T Office of University Programs and DoD Strategic Multilayer Assessment Branch. College Park, MD: START, 2017. https://www.start.umd.edu/pubs/START_QuantifyingGrayZoneConflict_June2017.pdf