Kidnappings by violent political groups: Explaining between-group and within-group variations

Kidnapping is one of the most common tactics employed by violent political groups. It yields money, political concession, publicity, intimidation effects, among many others. However, notable violent political groups differ significantly in how frequently they engage in kidnappings. Meanwhile, even avid kidnapping groups show episodes of particularly high numbers of kidnappings committed in particular years. This PhD thesis focuses on this specific phenomenon and aims to probe an explanation for the differences and changes in kidnapping activities among/within violent political groups. Violent political groups use kidnapping for a broad range of purposes. Kidnapping is also known to be particularly costly for its logistic complexity and intensive need for human resources. As such, the decision by violent political groups to engage in kidnappings is likely influenced by multiple considerations and causal mechanisms. Existing literature and policymakers have mainly considered kidnappings as a strategy for coercive bargaining to pursue ransom income or political concessions. This thesis, however, focuses on the role of governance in an extra-legal context. Where the state’s monopoly of violence is contested by the presence of violent political groups, efforts to establish one’s political power and governance may necessitate the use of kidnappings to impose and enforce rules for social and political control. For example, one may use kidnappings to enforce “taxation” and “protection” rackets, smuggling of goods, to expel or punish individuals disrupting social orders or working against the groups’ interests. To test this proposed extra-legal governance mechanism that may influence violent political groups’ engagement in kidnappings, this thesis examined the associations between a set of organisational factors – measuring different facets of extra-legal governance – and the numbers of kidnappings committed by violent political groups. Analyses were conducted both for the variations among different violent political groups and the temporal changes within-groups, using data from the Global Terrorism Database (GTD) and the Big Allied and Dangerous Version 2. Block-wise multiple linear regressions and path analyses were conducted to analyse the between-group variations, and fixed effect regressions were performed to examine the within-group changes. Findings from both the between-group and within-group analyses support the notion that extra-legal governance is a relevant factor for explaining kidnappings by violent political groups. In particular, organised forms of illicit financing (e.g., extortion, smuggling, drug-trafficking) – theorised as an essential aspect of governance in an extra-legal context – is significantly associated with both the between-group and within-group variations in kidnappings. Provision of public social services (e.g., education, health, transportation, security) is a significant predictor for whether kidnappings are committed in the within-group analyses but not in the between-group analyses. Moreover, findings from this PhD research also echoed what was found in prior studies – highlighting the importance of mechanisms other than extra-legal governance on kidnappings, namely, group capacity and social learning from ones’ allies. In addition to findings on the main research question discussed above, this thesis also examined patterns of missingness in perpetrators’ identity among the GTD kidnappings. This is a necessary step to assess the potential sampling bias caused by the systematic exclusion of GTD kidnappings without information on perpetrators’ identity in the analyses of the main research question. Based on descriptive analyses and logistic regression models, this study found that perpetrators’ identity information was not missing at random among the GTD kidnappings. Instead, they followed certain temporal and geographic trends. Incident-level characteristics also significantly predicted the missingness in perpetrator identity, such as whether ransom was requested, the types of weapons used, event fatalities. The potential sampling bias caused by missingness in perpetrators’ identity is a common issue present in all studies using datasets developed from open-source information. Analyses on missingness in this thesis provide valuable insights into possible ways research findings may be influenced by the selection process of publishing and collecting open-source information by datasets like the GTD. Despite the limitation identified in the missing data analyses, this PhD research provides important preliminary evidence suggesting that kidnappings may be related to the general exertion of political power and social control by violent political groups in extra-legal contexts. The apparent monetary gains and political concessions may not provide a complete picture in explaining the root cause of kidnappings in violent political campaigns. Future research and policymakers should consider the possible role of extra-legal governance in the theoretical explanation of kidnappings and the respective prevention strategies