Share This

NEI: GAO Report Mischaracterizes Data on NRC Violations

Oct. 17, 2013—A new study by the Government Accountability Office (GAO) found that inspectors in the U.S. Nuclear Regulatory Commission’s four regions differ on how they characterize nuclear power plant inspection findings, how they divide them into sub-parts, and how they grade them. While only small differences emerged for the more safety-significant violations, GAO said the NRC should find out why the regions vary so much in the numbers of low-level violations reported per inspection hour.

“The report is a good first step, but it presents overly simplified data,” said James Slider, senior project manager for safety-focused regulation at the Nuclear Energy Institute. “GAO presents summary tables of raw numbers and ratios, and then uses those ratios to speculate on factors that might explain some of the regional differences,” he said. Slider said it is difficult to draw any conclusions from tables that combine data from 2000 to 2012.

He added that the raw numbers should not be viewed as the focus of the report. “The counts themselves are not an absolute measure of a licensee’s performance, and comparing counts between licensees is not appropriate for ranking their performance.”

GAO said the NRC is “aware of differences across its regions in the identification and resolution of findings that result from its oversight processes and has taken some steps to address them,” particularly in the case of findings that have low safety significance. “However … NRC has not conducted a comprehensive analysis of these differences,” the report said. Without such an analysis, “NRC cannot ensure that oversight efforts are objective and consistent.”

GAO also found fault with the NRC’s efforts to promote “transparency and openness for all stakeholders.” While the NRC makes public an enormous number of records, GAO said, “we found this system difficult to navigate, and external users cannot use it to independently track findings, all documents related to the findings, and the findings’ resolution.” The complexity of the NRC’s record keeping also can hinder its staff in ensuring consistency, leading some to seek information from colleagues because they cannot readily access the pertinent records, the report said.

Among its recommendations, GAO suggested that the NRC commissioners “direct agency officials to evaluate the challenges inspectors face in retrieving all relevant information on plant performance and previous oversight activities and improve its systems accordingly to address these challenges.”

In its formal response to the report, the NRC said it generally agrees with GAO’s recommendations. The NRC said it believes its internal controls are adequate to ensure alignment between the regions. Even so, “the agency agrees to seek enhancements, particularly as they related to less significant findings and violations,” said Mark Satorius, executive director for operation, in a Sept. 12 letter.

However, the NRC said it does not share GAO’s perspective on the use of “professional judgment” in the reactor oversight process (ROP) and traditional enforcement. “The repeated reference [to the use of professional judgment] implies there is a high degree of subjectivity … and that professional judgment is used in an excessive or inconsistent manner,” Satorius said. “The agency believes that the use of professional judgment is limited and controlled through detailed guidance in the enforcement policy, enforcement manual and inspection manual chapters.”

NEI’s Slider said the nuclear energy industry agrees with GAO’s recommendation that the NRC study regional differences in detail to better understand their causes and develop solutions. However, he said GAO’s emphasis on differences in the counts of low-level findings misses a key point of the reactor oversight process. “The inspection framework was reformulated under the ROP to direct inspection resources to areas showing weaker performance—in other words, to direct attention to areas producing higher-level findings,” Slider said.

“Differences in counts of low-level findings are less important in the ROP framework than ensuring that higher-level findings are properly characterized and addressed,” Slider said. “The evidence indicates the ROP is achieving that very result.”

In a statement on the report, Dale Klein, a former chairman of the NRC, characterized it as a “mixed bag.”