This paper examines differences in compliance with air pollution regulation for U.S. pulp and paper mills. Our analysis is based on confidential, plant-level Census data from the Longitudinal Research Database for 116 pulp and paper mills, covering the 1979-1990 period. The LRD provides us with data on shipments, investment, productivity, age, and production technology. We also have plant-level pollution abatement expenditures from the Pollution Abatement Costs and Expenditures (PACE) survey. Using ownership data, we link in firm-level financial data taken from Compustat, identifying firm size and profitability. Finally, we use several regulatory data sets. From EPA, the Compliance Data System provides measures of air pollution enforcement activity and compliance status during the period, while the Permit Compliance System and the Toxic Release Inventory provide information on other pollution media. OSHA's Integrated Management Information System provides data on OSHA enforcement and compliance. Anticipating our results we find significant effects of some plant characteristics on compliance rates: plants which include a pulping process, which are older, and which are larger are all less likely to be in compliance. Compliance also seems to be correlated across media: plants violating water pollution or OSHA regulations are more likely to violate air pollution regulations. Firm-level characteristics are not significant determinants of compliance rates. Furthermore, once we control for the endogeneity of regulatory enforcement, we find the expected positive relationship between enforcement and compliance. We also find some differences across plants and firms in their responsiveness to enforcement. Pulp mills, already less likely to be in compliance, are also less sensitive to inspections. Some firm characteristics also matter here: plants owned by larger firms, whether measured in terms of their employment or by the number of other paper mills they own, are less sensitive to inspections and more sensitive to other enforcement actions, consistent with our expectations and with other researcher's results.
