Postwar period

The Postwar period in this account is not simply a backdrop to organizational change. It is a testing ground for how institutions handle dispersed knowledge: central authority can unite effort, but it can also suppress local expertise; decentralization can preserve diverse judgment, but without aggregation, valuable information remains isolated. The period’s cases—from American intelligence and corporate hierarchy to a submarine search and international disease research—show that the challenge is not choosing one pole, but building channels that make separate knowledge collectively useful.

Centralization and the Intelligence Problem

The end of World War II sharpened the United States’ concern that divided intelligence responsibilities could leave the country vulnerable. Before the war, the military services gathered intelligence separately; the failure to anticipate the attack on Pearl Harbor, despite evidence that a major strike was being prepared, exposed the limits of that arrangement. The possible conflict with the Soviet Union and technologies Donovan identified as the rocket, the atomic bomb, and bacteriological warfare added urgency to the question of how intelligence should be organized.

In April 1946, “Wild Bill” Donovan publicly called for a centralized, impartial, independent agency to take charge of national intelligence. His advocacy drew on his wartime leadership of the Office of Strategic Services, but his criticism of the existing intelligence community alienated its members and likely damaged his prospects of returning to government service. Congress nevertheless passed the National Security Act in 1947 and created the Central Intelligence Agency, seeking unity of command and intelligence in place of fragmentation.

The later assessment in the book distinguishes centralization from the information-sharing the intelligence system needed. Before September 11, agencies’ division of labor could support specialization and diverse analysis, but the community lacked a genuine way to aggregate judgments as well as information. A centralized authority might make information easier to collect, yet risk losing the independence, local knowledge, and diversity that specialized agencies contributed. The proposed alternative was aggregation without simply placing a small group at the center: linking agency databases, for example, could improve information flow while preserving autonomy. Even two years after September 11, the government still lacked a unified watch list drawing on all parts of the intelligence community.

Local Knowledge and Organizational Design

The book frames decentralization as a way to preserve knowledge that cannot readily be summarized or transferred: expertise tied to a particular place, job, or experience. People close to a problem may be best positioned to solve it, and specialization can enlarge the knowledge available across a system. But decentralization also creates a problem of transmission. Information discovered in one part of an organization may never reach the others, so local knowledge requires mechanisms that can assemble it without erasing its specificity.

That tension shaped the postwar American corporation. The dominant model combined vertical integration, layers of management, and centralized final authority. Even General Motors, whose divisions operated day to day much like independent businesses, reserved major strategic and organizational decisions for headquarters. In this structure, difficult problems moved up the hierarchy, while decisions and instructions returned downward. The result could be coordination at scale, but the many layers also obstructed information and made it difficult for senior managers to know whether their picture of the company reflected conditions on the ground.

The costs of that hierarchy appeared in both routine work and strategic failure. At a General Motors plant in Van Nuys, a worker had been welding a bracket for two years without knowing it held a car’s sunshade; she learned its purpose only after a supervisor showed workers the costly repair required when it was missed. In another company, a chart reproduced in In Search of Excellence showed that a new product idea had to pass through 223 possible bureaucratic paths before acceptance. Meanwhile, executives anxious to advance often communicated problem-related information less accurately, and successful managers could avoid reporting fights or budget problems. When Japanese and West German competitors began producing better products faster, American firms’ insulated hierarchies proved poorly suited to the changes required. Ford’s Pinto, introduced in 1971, epitomized the trouble; by the decade’s end, corporate profits, market share, and productivity growth had fallen sharply, while Chrysler and Lockheed required government bailouts.

Coordination Across Different Scales

The period’s organizational challenge appears in cases that could hardly be more different in scale or subject. In May 1968, the U.S. submarine Scorpion disappeared on its return to Newport News after duty in the North Atlantic. The Navy knew its last reported position but not what had happened or how far it had traveled after radio contact. The initial search area was a circle twenty miles wide and thousands of feet deep. Rather than rely on a handful of leading experts to choose a search location, John Craven assembled people with varied knowledge, including mathematicians, submarine specialists, and salvage men. He asked each to wager on the likelihood of alternative scenarios, including the submarine’s speed and descent. He then combined their estimates using Bayes’s theorem to infer its final location. No individual estimate supplied the answer; the composite made the scattered judgments useful together.

A different kind of coordination began in early 2003, when China notified the World Health Organization that a severe respiratory disease had affected 305 people in Guangdong Province since November 2002 and killed five. Reports from Hanoi and Hong Kong followed, and by early March the illness, called SARS, appeared to be a new disease rather than a new strain of flu. The WHO issued a global warning and activated a surveillance system, while also organizing an international effort to identify the disease’s cause. On March 15 and 16, it contacted eleven laboratories across countries and regions including France, Germany, Japan, the United States, Hong Kong, Singapore, and China. The laboratories agreed to collaborate, beginning a multicenter research project on March 17. The episode extends the period’s central organizational question into science: independent centers can contribute specialized work, but a shared effort must connect their results.

The Period’s Institutional Lesson

The cases do not support a simple preference for central command or for dispersed authority. Donovan’s campaign shows why fragmentation can be dangerous when agencies cannot unite intelligence. The failures of hierarchical corporations show why a central chain of command can also impede the information it needs. The submarine search demonstrates how judgments from distinct specialties can be combined without making them identical; the SARS project shows how research centers can collaborate across institutional and national boundaries.

Across these examples, the decisive distinction is between decentralization alone and decentralization paired with aggregation. Local experience and specialized analysis increase the range of knowledge available, but only channels for sharing and combining that knowledge can make it effective at the scale of a whole organization. The Postwar period therefore serves as a historical frame for a continuing design problem: how to coordinate collective action without making the center the sole source of judgment.