A Field Experiment on Police Extortion with West African Truck Drivers

Research Method

Randomized Controlled Trial


West Africa




Borderless Alliance, Center for the Study of Development Strategies (funder)

Research Question

How robust are the institutions of extortion to unexpected observation by outsiders?

Research Method

Randomized Controlled Trial


Corruption is often viewed as an 'informal institution' that helps to minimize transaction costs in information-sparse environments. If extortion is an institution, how institutionalized is it? How robust are the routines and norms of extortion to shocks to actors' expectations? I present the results of a field experiment that attempted to randomly manipulate the expectations of the low-level officials who extort truck drivers for money along the main trade corridors of West Africa. Over a two-week period traveling with different truck drivers on over 1,500 miles of highway, I randomly assigned a subset of 123 interactions between drivers and officials at checkpoints in Burkina Faso, Ghana, Togo and Benin to be overtly observed and recorded by me, an obvious outsider. Prior to the experiment, I elicited expectations about how this intervention would impact such interactions from 20 experts in the field of logistics and road governance in West Africa. Using a Bayesian model to integrate both the expert and the experimental data, I discuss different kinds of updating that this framework makes possible, and how this affects posterior certainty about both treatment effects and expert bias.