Employing Detection Controlled Models in Health and Environmental Risk Assessment: A Case in Offshore Oil Drilling

Abstract

ABSTRACT This study examines the determinants of health, safety, and environmental (HS&E) incidence and reporting behavior in Gulf of Mexico offshore drilling. One objective is to determine statistically significant variables that influence the probability of HS&E incidence and reporting. Incidence modeling employs standard qualitative response models with binary and ordered dependent variables, and a Poisson specification. A second objective is to examine the impact of imperfect reporting on inferences. Unlike previous studies, models allowing for the possibility of imperfect reporting are specified and estimated. The results of this analysis provide strong evidence to support the hypothesis that aspects of well complexity and site complexity increase the likelihood of HS&E incidents in drilling. Equally important is evidence rejecting the hypothesis that broader oil company attributes influence HS&E incidence. An analysis of reporting behavior provides mixed support for the hypothesis that well-known firms are more likely to report an incident than their more anonymous counterparts. There is weak evidence indicating differential reporting behavior between regulatory districts. Finally, the analysis provides consistent support for the conclusion that a 1996 policy change reduced HS&E incidence and increased HS&E reporting. The results of this research provide guidance to company managers and regulators on the factors driving HS&E incidence and reporting. Results also provide evidence that models of imperfect reporting can alter, in this case reverse, the interpretation of trends in reported incidents. These results can be used to support efficient resource allocation to prevention efforts, and inform regulatory policy.