CIA’s command relationship with the U.S. Intelligence Community

The CIA’s command relationship with the U.S. Intelligence Community emerged from a postwar attempt to replace fragmented intelligence gathering with a unified national system, but formal leadership never translated into effective command. The resulting arrangement preserved specialized agencies while leaving them with overlapping responsibilities and little capacity to combine their information and judgments—a weakness exposed by the September 11 attacks. The book’s central distinction is that intelligence needed aggregation, not simply centralization.

The Case for a Central Agency

In April 1946, at a forum organized by the New York Herald-Tribune, General “Wild Bill” Donovan delivered a speech entitled “Our Foreign Policy Needs a Central Intelligence Agency.” During World War II, Donovan had headed the Office of Strategic Services, the United States’ chief wartime intelligence organization. After the war, he became a forceful public advocate for a more powerful peacetime successor.

Before the war, intelligence-gathering responsibilities were divided among the military services. Their failure to anticipate the attack on Pearl Harbor, despite evidence that a major Japanese strike might be coming, made the limitations of that arrangement apparent. The possibility of conflict with the Soviet Union and the arrival of technologies Donovan identified as the rocket, atomic bomb, and bacteriological warfare added to the case for a more comprehensive system. He called for “a centralized, impartial, independent agency” to take charge of the nation’s intelligence operations.

The public campaign damaged Donovan’s prospects of returning to government: his sharp criticism alienated members of the intelligence community and probably doomed his chances of government service. But in 1947 Congress passed the National Security Act and created the CIA. The law aimed to establish unity of command and unity of intelligence, on the premise that fragmentation had left the country vulnerable to surprise attack.

Formal Leadership, Fragmented Practice

The intended centralization did not take hold. Although the CIA was initially the key player in the postwar system, the intelligence community grew more fragmented over time. The CIA operated alongside the National Security Agency, the National Imagery and Mapping Agency, the National Reconnaissance Office, the Defense Intelligence Agency, and intelligence branches of the three major military services. Their missions and responsibilities overlapped rather than converging under a single effective authority.

In theory, the CIA director led the intelligence community as a whole. In practice, the director exercised little supervision over the other agencies, while most intelligence-operation funding came from the Department of Defense. The FBI, responsible for domestic law enforcement, stood almost entirely outside the intelligence community’s orbit, even though information about foreign terrorists operating inside the United States was plainly relevant to the CIA. The result was not Donovan’s centralized repository of information and analysis, but largely autonomous groups pursuing the broad common aim of protecting the country by different means.

This arrangement illustrates the double edge of decentralization. Separate agencies could specialize, develop expertise, and examine information from different perspectives. But information uncovered in one part of the system was not guaranteed to reach the rest. The organizational structure therefore preserved local knowledge without reliably turning it into a shared picture.

September 11 and the Case for Aggregation

Before September 11, the flaws were easier to overlook. The intelligence community had not anticipated the 1993 bombing of the World Trade Center, the 1998 bombings of the U.S. embassy in Kenya, or the attack on the USS Cole in Yemen. The September 11 attacks made the failure of intelligence gathering seem undeniable. A congressional inquiry found that agencies had failed to recognize the individual and collective significance of information relevant to the plot, and had missed opportunities to disrupt it. In that sense, the failure recalled Pearl Harbor.

The book also cautions against treating the inquiry’s hindsight as proof that the attacks should have been predictable. Intelligence agencies handle enormous quantities of information, and a retrospective review can make previously scattered clues look more obviously connected than they were beforehand. In her account of Pearl Harbor, Warning and Decision, Roberta Wohlstetter described the difficulty of selecting meaningful signals from the confusion surrounding them. If a major naval attack involving hundreds of aircraft and ships was hard to foresee, the book observes, a terrorist operation by nineteen men could be harder still. Strategic surprise may be intrinsically difficult to prevent.

Yet difficulty does not settle whether a different organization could have improved the odds. The community had failed to anticipate four major terrorist attacks between 1993 and 2001. Congress concluded that better processes could have produced better results, particularly by correcting the lack of information sharing. Agencies had generated localized snapshots rather than a coherent account of threats. Senator Richard Shelby sharply criticized the FBI’s decentralized structure, arguing that it left information fragmented among largely independent fiefdoms; he called for an agency above the disputatious bureaucracies, restoring Donovan’s centralizing ambition.

The book accepts the importance of sharing but rejects the assumption that a central authority or a small group of officials should be responsible for sifting all the evidence. Specialization allows a finer-grained reading of information, while diverse perspectives can expose relationships that one group of analysts might miss. Acting Defense Intelligence Agency director Lowell Jacoby made that point in congressional testimony: information that one set of analysts regarded as irrelevant noise might reveal critical clues when examined by another. The missing element was a means to aggregate not only information but judgments—to draw on the collective intelligence of analysts across the agencies. Centralization might put someone in charge of the information; aggregation could make the whole community more capable of interpreting it.

Designing Connections Without Erasing Autonomy

One relatively simple remedy was to link the agencies’ computer databases, improving the flow of information while allowing each agency to retain its autonomy. Two years after September 11, the government still lacked a unified watch list drawing on data from all parts of the intelligence community. The book presents this gap as evidence that relatively mechanical changes could have made the system significantly smarter without first replacing its structure with a single commanding authority.

More ambitious was FutureMAP, an abortive program funded by the Defense Advanced Research Projects Agency. It proposed decision markets in which analysts from different agencies and bureaucracies could buy and sell futures contracts reflecting their expectations about developments in the Middle East and elsewhere. One proposed element was a small internal market—perhaps twenty or thirty participants—open to intelligence analysts and possibly a limited number of outside experts. The proposal represented a different response to fragmentation: rather than assume that all useful judgments had to pass through a hierarchy, it sought a mechanism for pooling dispersed expectations.

The broader principle is that decentralization works only when local knowledge can become collectively useful. Knowledge tied to a particular job, experience, or place may be hard to convey, yet can be valuable precisely because it is specific. Specialization increases the information available to a system; without channels for sharing and combining it, that added knowledge can remain trapped in separate parts. The CIA’s nominal leadership of the U.S. Intelligence Community never resolved this tension. Its history instead illustrates why command, information sharing, and collective judgment are related but distinct organizational problems.