Regulation of Pollution in the Laboratory : Random Inspections, Ambient Inspections, and Commitment Problems

We compare experimentally a traditional random inspection policy and a variant where the agency may carry out a preliminary inspection of the level of ambient pollution before implementing any individual inspection. Since the agency may have an incentive to announce high inspection probabilities and then secretly renege on its announcement to avoid implementing costly inspections, we are also interested in the agency's commitment power. We find that overall, ambient inspections increase efficiency but the effect is weaker than expected when the agency has no commitment power; and polluters' reactions to the lack of commitment power of the agency vary depending on whether the agency uses ambient inspections or not