What Our Intelligence Agencies Could Learn from Silicon Valley (WSJ 17/5/10)
The clamor to increase the power of the Director of National Intelligence is mistaken. We need less hierarchy and centralization.
The failure to prevent the 9/11 terrorist attacks led three years later to the Intelligence Reform Act, which decreed a reorganization of the nation’s intelligence system. Government reorganizations are a common response to government failures (we’re seeing it again, in the financial regulatory reform legislation now wending its way through Congress).
They are quick, highly visible, easily explained, and relatively cheap—and ineffective when the failures that beget them are not failures of institutional design. That’s the case with regard to financial reform: The nations of the world organize their monetary and fiscal agencies in very different ways, and there is no evidence that any organizational form was more successful than any other in anticipating or responding to the financial collapse of September 2008. The problem was not institutional design, but execution.
It’s the same with intelligence. The world’s intelligence systems are structured in very different ways, and no particular way seems to have a marked advantage in preventing intelligence failures—which are frequent because of the uncertainties inherent in trying to ferret out the secrets of a nation’s enemies.
What the government needed to do in the wake of 9/11 was to identify specific areas for improvement of our intelligence system. There were several.
The system was and is extremely complex—there are at least 20 separate U.S. intelligence agencies, not counting state and local agencies. (New York City’s police department, for example, has a formidable intelligence unit.) The nominal coordinator had been the Director of Central Intelligence, but he doubled as director of the CIA—a full-time job in itself, so that the coordination function tended to be neglected.
Further, despite the complexity of the system, there was a notable gap: We were the only major country without a domestic intelligence agency separate from the national police force (an agency such as Britain’s MI5 or the Canadian Security Intelligence Service).
The Intelligence Reform Act did nothing to fill the U.S. domestic intelligence gap. It did create a coordinator separate from the CIA director—the Director of National Intelligence (DNI). But it gave him additional powers that he did not need along with responsibilities in excess of those powers—a recipe for failure.
The new law designated him the nation’s senior intelligence adviser—a role properly played by the CIA’s director (the CIA was designed to be, and is, the nation’s central intelligence agency)—and gave the DNI commensurate staff resources. Soon he had a bureaucracy of 1,500 intelligence officers, all wanting to throw their weight around. The inevitable consequence was rivalry between the office of the DNI and the CIA. The rivalry grew toxic and now Dennis Blair, the DNI, is history, having fought and lost a bitter turf war (all turf wars in government are bitter) with the CIA’s director, Leon Panetta.
Mr. Blair was a four-star admiral, very smart, very able, but accustomed to commanding a disciplined hierarchical organization. The DNI does not command the intelligence system. Almost all its constituent intelligence agencies are part of cabinet-level departments, whose bosses outrank the DNI. The White House intelligence director, John Brennan, whose experience in intelligence exceeds Mr. Blair’s, outranked him as well, as a practical though not formal matter.
Mr. Blair’s termination is being attributed to the inherent weakness of the DNI’s position, and there is a clamor to enlarge the DNI’s powers. But increasing the power of the DNI is exactly the wrong direction for reform to take. It will exacerbate the rivalry between the DNI and the Department of Defense, which controls most of the intelligence budget; marginalize the CIA; and, worst of all, make the structure of the intelligence system more hierarchical when it should be less so.
Organization theorists distinguish usefully between “U-shaped” (unitary form) and “M-shaped” (multidivisional) organizations. The former are tightly centralized, with layers of supervisors: Information flows to the top and commands flow back down. Such an organization works well in a stable economic environment, where change is slow and production can be routinized.
M-shaped organizations are flatter and looser. The different divisions have considerable autonomy, develop somewhat different cultures, and compete with each other. Top management monitors performance, facilitates communication, encourages innovation, delegates authority, watches out for gaps and overlap, and thus coordinates a loose-knit structure that is optimal in a dynamic environment. That is the environment of an intelligence system, an information producer in a setting of bewildering uncertainty. It is no less a knowledge producer than the Silicon Valley firms that provide the most successful models of M-shaped corporate structures.
We need a Director of National Intelligence who is not the president’s senior intelligence adviser—that is the role of the CIA’s director. The DNI’s role is rightly that of a chairman focused on such urgent tasks as modernizing the intelligence system’s many computer networks (and enabling them to communicate with each other), establishing uniform standards for security clearances, and pushing for a coherent organization of domestic intelligence.
These are tasks wholly unlike briefing the president on North Korea’s belligerent intentions. They call for the skills of a top manager, ideally perhaps a former intelligence officer who had gone on to manage a knowledge-generating institution in the private sector.
Mr. Posner, a federal circuit judge and a senior lecturer at the University of Chicago Law School, is the author of “Uncertain Shield: The U.S. Intelligence System in the Throes of Reform” (Rowman &Littlefield, 2006).