A virtual fence along the U.S.-Mexico border at a test facility in Playas, N.M AP Photo/U.S. Customs and Border Protection

The federal government’s investigative arm issued yet another scathing report about Department of Homeland Security’s virtual fence program Thursday, questioning the cost-effectiveness of an initiative that has plagued by glitches and delays over the past five years.

The Government Accountability Office’s report said Homeland Security has “yet to demonstrate the cost-effectiveness of the proposed SBInet solution, and thus whether the considerable time and money being invested represented a prudent use of limited resources.”

Despite allocations of $1.6 billion, the Boeing Co.-led SBInet program’s promised technology is still not available to the Border Patrol, a May report from the Government Accountability Office found. Two grids along 53 miles of border in Southwestern Arizona are up, but not yet working.

The problems prompted Homeland Security Secretary Janet Napolitano to order a reassessment of the program in January and freeze future funds in March beyond work on the two Arizona systems. When she announced the freezing of funds, she also reallocated $50 million from stimulus funds to buy commercially available, stand-alone technology. More than half of that - $31.7 million - is scheduled to buy mobile surveillance systems that Border Patrol agents say are very effective.

The report says officials have not reliably estimated the life cycle cost of the two systems up in Arizona, called “Block 1.”

“To do so requires DHS to ensure that the estimate meets key practices that relevant guidance states are important to having an estimate that is comprehensive, well-documented, accurate, and credible,” the report says. “However, DHS’s cost estimate for Block 1, which is about $1.3 billion, does not sufficiently possess any of these characteristics.

The SBInet program that once was scheduled to cover three Border Patrol Sectors covering 655 miles has been scaled back to cover two sectors covering 387 miles, the report says. And performance measures have been laxed, too.

“The stringency of the performance measures was relaxed, to the point that system performance is now deemed acceptable if it identifies less than 50 percent of items of interest that cross the border,” the report says. “According to program officials, the decreases are due to poorly defined requirements and limitations in the capabilities of commercially available system components. The result will be a deployed and operational system that does not live up to user expectations and provides less mission support than was envisioned.”