PT Notes
EPA RFI - NAICS Codes Based on RMP Accident History Data
PT Notes is a series of topical technical notes provided periodically by Primatech for your benefit. Please feel free to provide feedback.
On July 24, 2014, the U.S. Environmental Protection Agency (EPA) announced a Request for Information (RFI) seeking comments on potential revisions to its Risk Management Program (RMP) regulations and related programs to modernize its regulations. Multiple issues were addressed in the RFI.
One of the issues addresses the use of the RMP database of accident history reports to update and replace the existing set of NAICS codes used to determine the eligibility of processes for Program 3.
When developing the RMP rule, EPA scaled the regulatory requirements based on the potential risk posed by a source and the steps needed to address the risk, rather than imposing identical requirements on all sources. To this end, processes subject to RMP requirements were divided into three tiers: Programs 1, 2, and 3. EPA established the most stringent RMP requirements under Program 3 for those industry sectors that represented a potentially higher risk of accidental releases. Accident history associated with industry sectors was identified by EPA as a better surrogate for underlying risk than individual source accident histories because accidents are rare events.
Program 3 eligibility is based in part on the North American Industry Classification System (NAICS) codes associated with the covered process. The specific codes identified for Program 3 were based on an analysis of reported accident histories within industry areas, selecting those that evidenced a higher risk potential. Ten NAICS codes were selected.
The RMP national database now contains nearly two decades of accident history reports from covered sources, and EPA believes that these reports represent a more comprehensive picture of the relative accident risks associated with different industry sectors regulated under the RMP rule than was available to EPA prior to publication of the rule . Based on these accident reports, the set of ten NAICS codes most frequently associated with accidents in RMP-regulated processes are not the same as the set of ten NAICS codes currently used for Program 3.
Questions posed by EPA on this issue include:
- Should industry sectors represented in the RMP database as those with the most accidental releases be used to update and replace the existing set of Program 3 NAICS codes with a new set?
- How can the RMP accident history data best be used to update the current list of NAICS codes that trigger Program 3 requirements? Should the agency take into account the number of sources in each sector, or the severity of reported accidents, or other factors, in selecting updated Program 3 NAICS codes? Is the methodology used to develop the NAICS code list applicable to the RMP accident history database?
- Would limiting the data analysis or the selection of NAICS codes to only those industry sectors represented in the RMP database provide a complete and accurate picture of high risk industry sectors?
- Should an analysis of the RMP data be combined with an analysis of other current accident history databases to inform any revisions/updates? If so, what other databases should be used? How much weight should be given to the RMP data set in comparison to other sources?
- Should the original NAICS codes continue to be included? Would not including the NAICS codes historically identified under Program 3 cause increased risks to those industry sectors by having them no longer subject to the more stringent measures?
- Should an analysis of accident history data be limited to a specific time frame?
- Would it cause confusion within the regulated community to change the list of NAICS codes for which Program 3 is required?
- What would be the economic impacts of modifying the list of NAICS codes for which Program 3 is required? Are there any special circumstances involving small entities that EPA should consider with respect to modifying the list of covered NAICS codes?
