United States intelligence community

The U.S. Intelligence Community is the decentralized network of American agencies charged with protecting the country from attack. Its history exposes a central organizational dilemma: specialization gives analysts distinct expertise and local knowledge, but those advantages are weakened when information and judgments cannot travel across institutional boundaries. The book argues that the community’s failures before September 11, 2001, did not prove that decentralization was inherently mistaken; they revealed a system that distributed intelligence work without providing an adequate way to aggregate what its separate parts knew.

From Centralization to Fragmentation

In 1946, as the possibility of conflict with the Soviet Union and the arrival of technologies such as rockets, atomic weapons, and bacteriological warfare made national vulnerability seem newly urgent, “Wild Bill” Donovan called for a centralized, impartial, independent agency to direct American intelligence operations. The National Security Act of 1947 created the CIA, with the stated aim of achieving unity of command and unity of intelligence. The intended answer to fragmentation was a single coordinating authority.

That centralization never fully materialized. The CIA began as the leading postwar intelligence agency, but the community developed into an “alphabet soup” of organizations with overlapping responsibilities: the CIA, the National Security Agency, the National Imagery and Mapping Agency, the National Reconnaissance Office, the Defense Intelligence Agency, and the intelligence arms of the three major military services. Although the CIA director was theoretically in charge of the community as a whole, in practice he had little supervisory power over the other agencies. Most intelligence funding came from the Department of Defense, while the FBI, responsible for domestic law enforcement, operated almost entirely outside the community’s orbit—even though information about foreign terrorists inside the United States could matter to the CIA.

The resulting system consisted of virtually autonomous groups that shared the broad goal of keeping the United States safe but pursued it in different ways. Its fragmentation was not simply a matter of having many agencies. Each held particular information and developed specialized ways of interpreting it, but no central repository or dependable coordinating process made their separate views into a common picture. The structure preserved organizational independence without fulfilling Donovan’s hope for unified intelligence.

Missed Warnings and the Case for Reform

Before September 11, the system’s weaknesses were easy 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 bombing of the USS Cole in Yemen. After September 11, the failures appeared unmistakable. The Congressional Joint Inquiry concluded that agencies had failed to recognize the individual and collective significance of information relevant to the attacks, missed opportunities to disrupt the plot, and let potentially important clues pass unnoticed. The book likens the experience to Pearl Harbor all over again.

The retrospective judgment requires caution. A post-attack inquiry can identify items that seem relevant once the outcome is known, without proving that analysts could reasonably have recognized their importance beforehand. Roberta Wohlstetter’s account of Pearl Harbor describes how indications of an impending attack were lost amid what she called “the buzzing and blooming confusion.” If a vast Japanese naval attack was difficult to foresee, the book observes, a terrorist operation carried out by nineteen men might have been harder still. Predicting the precise attacks on the World Trade Center and the Pentagon may have been impossible.

But uncertainty about prediction does not dispose of the organizational question. From 1993 through 2001, the community failed to anticipate four major terrorist attacks. The book asks whether a different structure could at least have made it more likely that analysts would recognize the combined importance of information already available and produce a reasonable estimate of the danger. Congress’s answer was that better processes could have improved the result, especially by addressing the lack of information sharing among agencies.

Senator Richard Shelby’s criticism focused on the FBI’s decentralized structure, which he said left information holdings divided among largely independent fiefdoms. The agencies generated localized snapshots rather than a coherent account of the threats facing the country; information was not reliably placed in the hands of people who could use it. Shelby therefore called for a body above and independent of the disputatious bureaucracies, reviving the centralizing idea Donovan had advanced decades earlier. The book acknowledges the force of the diagnosis—information needed to be shared—while challenging the proposed cure: putting one agency in charge would make central control easier, but might sacrifice the diversity, independence, and local knowledge that a decentralized system can provide.

Decentralization Needs Aggregation

The book distinguishes decentralization from the particular form it took in the intelligence community. Dividing work among agencies can make sense: specialization supports more detailed analysis, while distinct perspectives can reveal relationships that one group would miss. Acting Defense Intelligence Agency director Lowell Jacoby made this point in testimony to Congress: information one set of analysts considered irrelevant noise might provide critical clues when another set scrutinized it. Different agencies’ perspectives were therefore a potential strength, not merely an obstacle to coordination.

The missing element was a mechanism for aggregating not just information but judgments. The system could draw on the knowledge of National Security Agency analysts, CIA officers, and FBI agents without making their collective interpretive capacity usable as a whole. It had decentralization without aggregation, and so lacked the organization needed to turn separate contributions into shared intelligence. Shelby’s proposal could have created a central authority to sift information, but, the book argues, no small group—however intelligent—could be expected to be smarter than the larger body of analysts whose knowledge it would oversee. The more promising resource was the collective judgment of the intelligence community itself. In this formulation, centralization is not the answer; aggregation is.

One practical improvement would have been to link the agencies’ computer databases while allowing them to retain their autonomy. The book notes that even two years after September 11, the government still lacked one unified watch list drawing on data from all parts of the intelligence community. A basic technical connection could have helped information move without requiring every agency to surrender its separate structure. The example shows the distinction between sharing and central command: information can be made accessible across organizations without concentrating every decision in one office.

The Value and Limits of Local Knowledge

Decentralization’s strength is that it encourages independence and specialization while allowing people to coordinate their work. The book connects this principle to tacit knowledge: knowledge grounded in a particular place, job, or experience that cannot easily be summarized and passed along, yet can be essential to solving a problem. People close to a problem may be better positioned to find a workable answer because they possess details that are difficult to transmit upward through a hierarchy.

That advantage comes with a corresponding weakness. Valuable knowledge discovered in one part of a system may never reach the rest of it; information that remains local cannot contribute fully to a collective judgment. A capable organization therefore needs to preserve the specificity that makes local knowledge useful while finding ways to make it globally available. The book uses Google as an analogy: the company draws on the local knowledge of millions of Web-page operators to improve its searches. The intelligence community, like a market or corporation, faces the same design problem—how to connect knowledge without flattening or replacing the independent sources from which it comes.

This is also why the book resists treating decentralization as a universal cure. In its account of the intelligence agencies, the problem was not that the work was distributed, but that the agencies lacked ways to share information and combine analysis. The relevant contrast is not simply between a scattered system and a centralized one. It is between a system that leaves specialized knowledge isolated and one that allows distinct contributions to inform a collective view. Effective coordination requires both local independence and channels through which the results of that independence can travel.

A Broader Organizational Principle

The intelligence case belongs to the book’s wider examination of how groups turn dispersed knowledge into intelligent action. Across different kinds of organization, local information and specialized judgment can increase the amount and diversity of knowledge available, while weak communication can prevent those advantages from affecting decisions. The U.S. Intelligence Community stands as a particularly consequential example: agencies could gather information and develop expertise in parallel, yet the absence of effective aggregation left the community unable to make full use of its own resources.

The book therefore treats the community’s pre-September 11 failures neither as conclusive proof that no attack could have been foreseen nor as a simple argument for one all-powerful intelligence agency. Its more exacting conclusion is structural. Central command may make responsibility and information control appear clearer, but the system also needs to preserve competing perspectives and knowledge close to the work. The task is to connect independent agencies well enough that separate clues and judgments can become collectively useful—without assuming that a small center can replace the intelligence of the whole.