The logic is that corrupt police are less likely to enforce counterinsurgency, implying a positive correlation between insurgency and corruption in locations with active insurgents. This research tests that prediction in Afghanistan, using survey data about police corruption, and two available measures of insurgency: number of attacks against the police and government forces, and number of violent incidents with civilian casualties. Actors of this research form three separate groups: police, civilians and insurgents. The hearts-and-minds approach is the theoretical argument used for analysis of actor interaction in the local context. The approach states that the reduced willingness of non-combatants (civilians) to provide intelligence to counterinsurgents (police) constrains insurgent activities and survival. It emphasizes population attitudes as a predictor of violence. According to this, public support for active insurgencies has a negative relationship with the number of civilian casualties resulting from insurgent violence. This implies a seesaw effect between support for insurgents and support for counterinsurgents among the civilian population. Is public attitudes towards police negatively affected by police corruption? If so, police involvement in illicit activities may cause more violence - directly and indirectly - and reduce public willingness to share information with them about insurgents in order to combat insurgent violence. Police corruption may alter counterinsurgency effectiveness in the short run, and implicates government control. The research problem pertains to geographic areas of Afghanistan with active insurgencies and the relationship between violence and police corruption. First, it is hypothesized that corrupt police officers directly undermine security by diverting labor and capital away from crime fighting and counterinsurgency. Second, they indirectly undermine security by incentivizing civilians not to share information about active insurgencies. Instead of a top-down approach to security, this research follows a bottom-up approach to understanding violence in the local context and supporting anti-corruption measures.
