When We May Not Experiment: On the Ethics of Randomized Controlled Trials in Policy and Development

When are governments, non-governmental actors, or researchers permitted to randomly assign citizens to different policies to learn their effects? Randomized policy experiments are increasingly used in development research. Yet randomization involves treating citizens unequally. Some are denied the benefits of a policy which are extended to others. Philosophers and researchers have argued that the concern of unequal treatment can be overcome. They argue that when two policy options are on a par randomization does not treat anyone disadvantageously. In other circumstances randomization is justified as an exercise of fair supererogation. In both cases, randomization is supported by both epistemic and fairness considerations. I challenge this outlook on RCTs and provide a more restrictive perspective on the moral limits of policy experiments. I argue that, even when two policy options are on a par, unique concerns of inequality and, thereby, disadvantageous treatment can arise. I also argue that the scope of fair supererogation is limited by concerns of the costs of experimentation and conditional obligations not to experiment.