Policing Power

This dissertation is about how the state influences individuals’ behavior by giving certain citizens the legal and physical means to coerce other citizens. Using field experimentation, participatory observation, and time-series analysis of two large sets of micro-data on crime to study policing in West Africa and Melanesia, the findings challenge conventional wisdom about the relationship between coercion, corruption, and the state. Empowering women by sending police officers to assist them in disputes with men may not necessarily reduce gender-based coercion, because men can preserve their privileges by drawing on alternative authorities. Conferring police officers powers to coerce other people does not necessarily induce corrupt behavior, because conferral of power may cause them to care more about their reputation than the rents they can extract. Competitive elections may not reduce petty police corruption even if they make principals accountable; instead, elections may incentivize corruption by increasing agents’ uncertainty about how principals will act in the future. These findings contribute new insights to the theory of state-building, accountability, and bureaucratic politics.