Bits, Bytes and Bureaucrats: New Federal Stance Is Needed
The American government gave birth to the modern computer in World War II by funding the University of Pennsylvania’s ENIAC program, an attempt to create a machine capable of analyzing elements of anti-submarine warfare and artillery bombardment. But Washington is an innovator no more. As a Times series concluded this week, “much of the government is operating with computers designed and built when Studebaker was making cars.”
Myriad legal and regulatory obstacles have done everything from delaying delivery of computers for years to dictating that the government buy the cheapest models, regardless of performance. However, Congress and the Clinton administration have adopted a series of reasonable reforms aimed at clearing these obstacles. Most of the changes focus on decentralizing the federal decision-making process, permitting regional officials to assess their own computer needs and procure the equipment.
This decentralization and the authority that goes with it are appropriate, but the Harvard University professor who has spearheaded much of the effort underestimated the difficulty of getting the federal government to behave like the commercial world. The laws that Steven Kelman helped draft basically just exhort bureaucrats to do better. For this approach to work, the government must confront two significant problems:
% The sense in Congress that American voters are unwilling to support the investments required to help Washington cross that celebrated bridge to the 21st century. The White House and Congress need to convince Americans that modernization is essential to address such intense public concerns as, say, airline safety, which demands replacement of the 1950s-era computers on which the nation’s air traffic control system still depends. Government should rally public support further by underscoring that the IRS loses $170 billion annually because its antiquated computers are unable to track down tax cheats.
% The need to educate and motivate federal workers. The government command structure tends to insulate many employees from the responsibility of decision making, so when bureaucrats rise to the point of making computer purchasing decisions they lack the experience and training essential to effectively use that authority.
The best way to ensure that the government buys computers that enhance productivity is to employ the system of “benchmarking,” common in private industry, which compares how well different companies complete given tasks. The Government Performance and Results Act, which Congress passed in 1993, called on federal agencies to come up with common ways of assessing the effectiveness of information technologies. Computer industry analysts, however, say that the legislation has yet to produce satisfactory results. What’s needed is for Washington to complete the process and assess the capabilities of its newly decentralized agencies.
Control from Washington--centralized planning--has a bad name. One tends to associate it with repressive and sluggish Soviet-style bureaucracies or with the impenetrable regulatory manuals for which Washington is famous. But an enlightened central control ultimately should test the performance of the regional agencies and decide which of them uses technology most efficiently. That then becomes the standard.
