Recovery site hosts bad data, GAO says

The Obama administration’s project to track all spending under the $787 billion economic stimulus law has data quality problems, according to a new report from the Government Accountability Office.

Hundreds of thousands of recipients have submitted spending reports, which have been posted online at Recovery.gov. The reports describe projects funded by, and jobs created from, the $47 billion in spending to date, said the GAO report of Nov. 19.

However, the data has not proved fully reliable, GAO said. The problems include 3,978 reports that showed no dollar amount received or expended but listed more than 50,000 jobs created or retained, and 9,247 reports that showed no jobs created, but listed expenditures of $1 billion combined.

There also were discrepancies between award amounts and the amounts reported as received which, although relatively small in number, indicate “problematic issues in the reporting,” the GAO report said.

In addition, while the Office of Management and Budget has estimated that 90 percent of the recipients required to report to date have done so, there are questions about the 10 percent who have not reported, the GAO said.

While praising the administration’s efforts at transparency, the GAO urged that the data quality problems be fixed.

“While recipients GAO contacted appear to have made good faith efforts to ensure complete and accurate reporting, GAO’s fieldwork and initial review and analysis of recipient data from Recovery.gov indicate that there are a range of significant reporting and quality issues that need to be addressed,” the report concluded.

The GAO recommended that the Office of Management and Budget clarify definitions and continue working with federal agencies to provide guidance to recipients. OMB agreed with the recommendations.

Earl Devaney, chairman of the Recovery Accountability and Transparency Board, which oversees Recovery.gov, said Nov. 19 that the board has corrected a number of errors in the data posted on Recovery.gov associated with nonexistent congressional districts.

Where incorrect congressional districts were identified in states with only a single such district, the data was changed to reflect that district. In states with multiple districts, the data was matched to zip codes used in the reports. In cases where no zip code was provided, the board has utilized a placeholder code until further investigation, Devaney said in a statement.

Devaney said the problem was fixed quickly because of its importance.

“In cases where incorrect data results in a significant risk that the public will be misled or confused, the b0oard, as it has previously done, may opt to step in and correct or omit data in lieu of the established OMB guidance that only recipients can enter or change data,” Devaney said.

Gene Dodaro, acting comptroller general, and Devaney were among the witnesses who discussed the data quality issues at a hearing held Nov. 19 by the House Committee on Oversight and Government Reform.

The recovery board recently refreshed Recovery.gov to include more features. The updated Web site got mostly good reviews, but transparency advocates were concerned about data quality and searchability.