The Value of Discretion in the Enforcement of Regulation: Experimental Evidence and Structural Estimates from Environmental Inspections in India∗

Regulatory design balances rules that deter corruption with discretion to allow use of local information for enforcement. To evaluate this trade-off we combine a field experiment in India which enforced inspection at the mandated frequency for a random set of industrial plants with structural estimation of regulatory costs. Relative to control plants, the inspection frequency, citations for high pollution and closure warnings for treatment plants more than doubled. However, the likelihood of more costly punishments (utility disconnections and plant closure) was unchanged. Compliance to the (strict) regulatory standard is weakly higher in the treatment group, and is driven by plants with pollution close to the threshold and that face relatively low compliance costs. Average pollution levels do not significantly decline. We model a dynamic game where a plant chooses to abate if it expects future sanctions to exceed present cost of compliance. The game structure and rich administrative data on regulatory and firm actions allows us to estimate the cost of regulatory penalties to plants. For the average treatment plant, abatement is relatively costly and therefore not worthwhile, as the risk of being penalized for pollution only moderately above the standard is small. Since control inspections target the heaviest polluters, the average inspection for a control plant is twice as costly in expectation for the plant compared to the rule-based treatment inspections. We interpret this difference in cost as the social value of discretion in inspections of control plants. ∗We thank Sanjiv Tyagi, R. G. Shah, and Hardik Shah for advice and support over the course of this project. We thank Pankaj Verma, Eric Dodge, Vipin Awatramani, Logan Clark, Yuanjian Li, Sam Norris, Nick Hagerty and Susanna Berkouwer for excellent research assistance and numerous seminar participants for comments. We thank the Sustainability Science Program (SSP), the Harvard Environmental Economics Program, the Center for Energy and Environmental Policy Research (CEEPR), the International Initiative for Impact Evaluation (3ie), the International Growth Centre (IGC) and the National Science Foundation (NSF Award #1066006) for financial support. All views and errors are solely ours. †MIT, eduflo@mit.edu ‡MIT, mgreenst@mit.edu §Harvard, rohini_pande@harvard.edu ¶Corresponding author. Harvard Weatherhead Center, nickryan@fas.harvard.edu