Development Assistance and the Diffusion of Insurgent Violence Open Access
Downloadable ContentDownload PDF
Development assistance has become a cornerstone of counterinsurgency (COIN) strategy. Despite extensive reliance on this approach to COIN, a fundamental question remains unanswered: is development assistance an effective COIN tool? Existing scholarship has arrived at very different and conflicting conclusions. This is partly the result of the fact that much of the existing research utilized traditional statistical techniques, which only accounted for the effect of developmental interventions in the immediate areas receiving assistance. By ignoring how these endeavors affect surrounding areas, previous research is affected by spatial autocorrelation (SAC) that systematically biases the results. On the other hand, this dissertation advances and tests a theory explaining the effects of assistance on violence both where it is implemented and in surrounding areas. Specifically, it posits that the effect of aid is a product of two, under-studied factors: whether development interventions are corrupted and if they are undertaken by a belligerent or a neutral party. In doing so, this investigation combines the literatures on conflict diffusion and development assistance. It leverages extensive archival- and interview-based evidence as well as original datasets and quantitative methods that incorporate the effects of aid on violence in surrounding areas, thereby avoiding bias caused by SAC. This dissertation covers three distinct cases: Iraq, Peru and Colombia. Unfortunately all four possible combinations of (un)corrupted projects and actor types are not present in any of the three cases. Consequently, a piecemeal approach to examining these different scenarios across conflicts is employed. While not ideal, this is not only necessary given the realities of how COIN aid has been used in practice, it also represents an advance over existing research, which has tended to examine single conflicts in isolation.