High-Tech Border Surveillance: A Costly, Buggy, Blundering Flop

The Obama administration's recent surprise decision to suspend new work on a multibillion-dollar high-tech border control system -- the third attempted since 1997 -- raises further questions about the government's use of computer networks and sensors in an effort to seal the border with Mexico.

As part of a broad illegal immigration crackdown called the Secure Border Initiative, a seamless "virtual fence" launched in 2005 was supposed to be up and running by last year. Among other things, the project known as SBInet called for a vast surveillance system along the 2,000-mile Southwestern border capable of detecting immigrants, drug traffickers, and potential terrorists as they attempted to cross into the United States.

But at the current slow pace of construction, SBInet would take decades to complete. Only a 28-mile-long prototype is fully up and running in Arizona, to mixed reviews.

In March, shortly before the House Homeland Security Committee met to discuss SBInet, Homeland Security Secretary Janet Napolitano announced that $50 million in planned stimulus spending on the project had been diverted to other border protection initiatives. That marked a sharp turnabout for the Obama administration, which had professed strong support for the effort.

Napolitano, a former Arizona governor who backed using such technological solutions on the border during her nomination process, said that SBInet "has been plagued with cost overruns and missed deadlines." It also has been marked by technical problems and poor government planning, according to congressional testimony and independent investigations by watchdogs like the Government Accountability Office.

The lead contractor on the project, Boeing Co., has earned $615 million so far. Company officials declined requests from the Center for Investigative Reporting to comment on SBInet. Officials from U.S. Customs and Border Protection also would not answer questions about the project.

The problems with SBInet represent what critics call the federal government's history of overreliance on speculative technology sold at great cost to taxpayers. Washington has spent some $800 million on SBInet. When earlier failed border technology projects are included, the price tag jumps to $1.1 billion, according to figures released during congressional hearings.

Some lawmakers now are questioning whether SBInet should be abandoned.

"At least $800 million so far has been wasted," Sen. John McCain, R-Ariz., fumed in an April hearing. "Think of how that money could have been spent to try to improve our border security. There's been a lack of oversight; there's been a lack of accountability, and by most reports, this virtual fence has been a complete failure."

Backers say that bugs in SBInet technology are being ironed out and the prototype is beginning to produce results, including the apprehension of intruders. Still, the continued march of illegal immigrants into the United States, despite the Secure Border Initiative, has undoubtedly set back efforts at comprehensive immigration reform and perhaps will lead other states to follow the example of Arizona, where a tough new immigration law has triggered protests across the country.

Years in the Making

High-tech border surveillance took on a new sense of urgency in the wake of the terror attacks of Sept. 11, 2001. Concerns about immigration and national security merged in a campaign for an array of cameras, sensors, and computer technology that would alert border protection officials when intruders were detected.

It wasn't the first time such an approach was tried. In 1997, immigration officials from the Clinton administration attempted a network of thousands of infrared, seismic, and magnetic sensors combined with surveillance cameras. The idea wasn't far from SBInet: leverage technology to extend the reach of border patrol agents and save the government money on personnel costs.

The network was known as the Integrated Surveillance Intelligence System, or ISIS. Government auditors reported widespread problems in 2003 and later, including numerous cost overruns; an operations center where contractors and government employees "did little or no work for over a year;" 60-foot poles that were paid for but never installed, and gear that was put in place but did not work. Government officials paid "chronic inattention to contract administration and project management," according to one report. ISIS covered just 4 percent of the border by 2005.

Shortly thereafter, the system received a public-image overhaul with a new but short-lived name, America's Shield Initiative. The government again pledged 24-hour surveillance with state-of-the-art technology that could deter illegal immigrants and drug traffickers. Washington spent more than $340 million on ISIS and the shield initiative combined as of 2005, the year the Secure Border Initiative took shape.

A House Homeland Security subcommittee met that summer to assess years of border surveillance attempts, and the hearing came with denunciations that sounded remarkably similar to what was said about SBInet on Capitol Hill in recent weeks. The subcommittee's chairman, Rep. Mike Rogers (R-Ala.), lashed out at ISIS, saying he was "struck by how much money has been abused" and arguing that problems with it "seriously weakened" border security.

"What we have here, plain and simple, is a case of gross mismanagement of a multimillion-dollar contract," Rogers said of ISIS at the time. "This agreement has violated federal contracting rules and it's wasted taxpayers' dollars."

By then, protests over the torrent of illegal immigration were growing; the Bush administration needed to do something significant and visible in response. During a November 2005 speech in Houston, then-Homeland Security chief Michael Chertoff unveiled the Secure Border Initiative. Under his multilayered plan, the federal government would deploy a surge of border patrol agents and investigators, bolster work site enforcement of immigration laws, and expand the number of beds available for immigrant detainees to a total of 20,000.

Republicans in Congress followed with a bill directing Chertoff to construct hundreds of miles of actual physical fencing or "tactical infrastructure" that, according to plans, would work alongside SBInet's sophisticated devices to help keep intruders out.

The SBInet blueprint called for private contractors to install hundreds of towers along the southwestern United States topped by surveillance cameras and coupled with ground sensors lining the desert floor. Together they would become a wired wall of surveillance that expanded how far authorities could "see," notifying them of border crossers by transmitting signals to command centers and patrol cars.

Mixed Reviews

Like past digital border attempts, however, the promise of new technology wouldn't come cheap. The first detailed budget documents submitted to Congress in late 2006 estimated the total cost between $7 billion and $8 billion in the coming years. A report from the Department of Homeland Security's inspector general during that time concluded the real amount could climb as high as $30 billion, although no explanation was given for the larger figure. The inspector general added that authorities didn't have the capacity to control costs and effectively plan for SBInet.

An SBInet prototype called P-28 is now in place in Arizona. Border patrol officials, however, concede that it faced bugs and setbacks, and agents on the ground have said it's not optimal for what they need. Numerous unfavorable reports from the inspector general and the GAO subsequently identified technological problems and other issues. The P-28 prototype fell eight months behind schedule, according to one such report, "because the contractor-delivered system did not perform as intended," adding, "contractor oversight was limited."

Another GAO report found that software selected by Boeing to gather data coming in from the border was built to accommodate a law enforcement dispatch system and "not designed to process and distribute the type of information being collected by the cameras, radars, and sensors." Rainstorms accidentally set off the radars, and surveillance cameras couldn't see even half as far as hoped.

"The bottom line is we've been waiting for a while to get something that works," says Richard Stana, a homeland security director at the Government Accountability Office, the investigative arm of Congress.

The challenges SBInet must surmount are considerable. The project's success depends on the integration of many different types of technology along a dry, mountainous border. Government officials admit that, due to tight deadlines, tests were not performed during the prototype's construction to ensure cameras and radars would operate together effectively before being deployed in the desert.

Boeing executives have told lawmakers that the prototype is now a key tool in the seizure of illegal border crossers and the interception of narcotics. A second phase under way in Arizona is performing "reliably and effectively seven days a week," Roger Krone, Boeing's president of network and space systems, said in a congressional hearing in March.

Krone said that border agents are quickly adapting to SBInet and the technology has increased their tactical advantage. "In terms of performance on the program, progress is evident," Krone testified. "We are not seeing any system-wide issues."

But Randolph Hite, an information technology director for the GAO, said during the same hearing that hundreds of smaller defects have cause additional delays. Those included blurry camera images and false radar detections due to bad weather. High winds, typical in the sparse desert landscape, caused surveillance towers to sway, and radars sometimes shut down when rotating.

Further, reports from testing of the system were repeatedly changed, raising questions about their reliability, according to Hite. He told Congress at the hearing that evidence suggested some tests were designed to ensure the technology received a passing grade.

The Center for Investigative Reporting requested Department of Homeland Security evaluations of Boeing's performance under the Freedom of Information Act, but Customs and Border Protection denied the request, citing provisions of open-government laws that exempt proprietary business information.

The project recently received praise from Rep. Ann Kirkpatrick (D-Ariz.), who visited a Tucson-based SBInet command center in March. Though she supports the administration's decision to freeze financing for the project until a review is completed, Kirkpatrick reportedly said she was impressed with the video surveillance capabilities. She called SBInet a safer way for agents to track smugglers than traveling alone in patrol vehicles, according to press accounts.

Janice Kephart, director of national security policy at the Center for Immigration Studies, believes the security situation in the Southwest is dire. She said it's unfair to judge Boeing and the company's work on SBInet when the system hasn't been deployed along a much larger area of the border, where major weaknesses exist, in order to truly determine its capabilities. In addition, Kephart said, SBInet is not a silver bullet. Among other things, virtual eyes and ears must be combined with the physical barrier. Plus, she added, the government needs to hire more border patrol agents and they need greater permission to confront intruders. Agents now are afraid of excessive force accusations, she said.

"If you don't have the resources on the ground, then they're just going to go around the towers," Kephart said of border crossers.

All of that means spending endless sums of taxpayer money, counters Marc Rosenblum, a senior analyst at the Migration Policy Institute, which argues immigrants contribute more to the American way of life than they take when given the opportunity. The return on investment in border infrastructure is limited because undocumented immigrants and traffickers will find new ways into the United States. Though politically controversial, he says, the real answer is comprehensive immigration reform.

"On the one hand, there's some evidence these investments have made the border more secure," Roenblum said. "But nobody who's studied the issue thinks you can totally solve the problem [with technology]. . . . You have to have realistic expectations about what can be accomplished."