President Bush has proposed a new post of national intelligence director. Not part of the Cabinet or located in the White House, the director would be charged with "coordinating" the intelligence budget and "working with" various intelligence agencies to set priorities. Sen. John Kerry has supported a more activist role for an intelligence director recommended by the Sept. 11 commission. Both houses of Congress are holding hearings to expedite legislation.

The sense of urgency in the middle of a presidential campaign is being justified on the grounds that the country is in imminent danger; the implication is that the existing intelligence system is not capable of dealing with the immediate threats. This argument cuts both ways. Reorganization will bring with it months-or years-of adjustment throughout the executive branch, and the more sweeping the change, the more this will be true. Whatever happens, the short-term threats must be dealt with through improvements to the existing structure, which was instituted after Sept. 11. As for longer-range threats, care must be taken lest a hasty transition to a new system generate unnecessary vulnerabilities. Thoughtfulness is more important than speed.

Terrorism, forthrightly described by the Sept. 11 commission as an attack from radical fundamentalist Islam, is spearheaded by technically private groups basing themselves on the territory of sovereign states and impelled by a fanaticism transcending traditional political loyalties. Adapting the intelligence system to these new realities must start with an understanding of the problems requiring solution. The current emphasis is on centralization; the principal disagreements concern the locus and authority of the proposed director of intelligence-whether he or she should have budgetary authority and whether the role should be free-standing or in the executive office of the president. The basic premise seems to be that the cause of most intelligence failures is inadequate collection and coordination. In my observation, the breakdown usually occurs in the assessment stage. The four major intelligence failures of the past three decades illustrate the point:

First, the 1973 Middle East war, which caught both the United States and Israel by surprise; second, the Indian nuclear tests of 1998, which opened a new era of proliferation threats; third, Sept. 11; and fourth, the failure to find weapons of mass destruction in Iraq. In each of these intelligence failures-except possibly Sept. 11-the facts were at hand. The difficulties arose in interpreting what they meant. Even Sept. 11 was ascribed by the commission to a failure of imagination in connecting the dots of available knowledge.

Before the 1973 Middle East war, the U.S. and Israeli governments were aware of every detail of the Egyptian and Syrian buildup. What they misjudged was its purpose. Nobody believed the Arab armies would actually attack, because every analyst at every level was convinced they were certain to be defeated. Every event, no matter how ominous, was interpreted as confirming that premise. Even when the Soviet Union withdrew dependents from Syria and Egypt 48 hours before hostilities started, it was viewed as caused by Soviet-Arab tensions.

Similarly, with respect to the Indian nuclear tests, public evidence was ignored because the intelligence community did not believe India was capable of concealing an actual test.

On the weapons issue-as the British Butler report on intelligence demonstrates-the assessment process broke down when the analysts jumped from incontrovertible evidence-a decade of Saddam Hussein's violations of the 1991 cease-fire agreement; building of dual-purpose plants for chemical and biological agents; efforts to acquire nuclear material; elaborate measures of deception- to the assumption that the demonstrated capacity to produce had been translated into stockpiles of weapons. (As early as 1998, President Bill Clinton, in an address explaining the bombing of Iraq, gave specific quantities for chemical and biological stockpiles.) That assessment went one step too far. But what we know now would not necessarily have changed the calculus for preemption. Could the United States wait until weapons were actually produced by a country with the largest army in the region, the second-largest potential oil income, a record of having used these weapons against its own population and neighbors, and-according to the Sept. 11 commission-intelligence contact with al Qaeda?

The answer requires a primarily geopolitical, not an intelligence, judgment. This is why, in reorganizing the intelligence structure, care must be taken to keep the assessment process distinct from geopolitical and strategic advocacy. Intelligence is most reliable about events that have happened or are about to happen. It grows less definitive about the future. Intelligence agencies should be judged by their ability to collect information, to interpret it, to keep assumptions from determining conclusions and to understand underlying trends.

It is a fine line, but a crucial one for effective policymaking. Most major strategic decisions involve judgments about consequences. Intelligence should supply the facts relevant to decision; the direction of policy and the ultimate choices depend on many additional factors and must be made by political leaders. A national intelligence director in the executive office of the president would erode this distinction, give intelligence disproportionate influence in policymaking and skew intelligence away from analysis.

Similarly, the merging of foreign and domestic intelligence under a single official unchecked by any institution in the executive branch short of the chief executive gives cause for concern. This is not how most democracies handle the challenge. The frequently invoked analogy to the Joint Chiefs of Staff ignores the fact that the Joint Chiefs, while enjoying direct access to the president, must in their daily operations refine their ideas in interaction with the civilian Pentagon leadership. Until recently, the policy was to raise a wall between the foreign and domestic intelligence services to prevent emergence of a single, dominant, unchecked intelligence service. Sept. 11 showed that this effort had gone too far and impeded the coordination of evidence on terrorism. But it does not follow that eliminating the distinctions altogether is the best solution.

Reorganization needs to improve the quality of intelligence at least as much as its collection. Policy stands and falls on the ability to distill trends from information. Does a free-standing director of national intelligence, charged with coordinating (in the president's proposal) or running the entire intelligence community (as in the Sept. 11 report) solve this challenge? Or does an excessively centralized system magnify the inherent danger of intellectual conformity? What structure is most likely to achieve a sense for the intangible?

In practice, most of the proposed reorganization schemes abolish the provision in the National Security Act of 1947 that makes the head of the CIA also the director of foreign intelligence for the entire government. The CIA chief has not been able to implement his theoretical powers because of the insistence of other agencies or departments-especially the Pentagon-on autonomy for their share of the intelligence process.

Layering a new national intelligence director over the CIA director would have one of two consequences: In a world where power flows from knowledge, it would require creation of a massive new bureaucracy to redirect the flow of intelligence throughout the government and sift the intelligence input from the various components of the intelligence community. Where would the personnel for such a structure come from? Does it mean dismantling existing institutions, and which ones? Could the national intelligence director function without having the analytic branch of the CIA placed under his or her direction? If the CIA were gutted in this manner, what would become of the remnant? On the other hand, if the national director were without an agency to provide support, he or she would become little more than a conduit for the recommendations of the various agencies.

In either event, the CIA director would no longer have direct access to the president, since the national director of intelligence would be defined as the president's principal intelligence adviser. Other alternatives deserve consideration; for example, enhancing the coordinating and budgetary authorities of the CIA director on foreign intelligence, symbolized by changing his title to national intelligence director. The coordination between domestic and foreign intelligence activities could be achieved by institutions such as the "National Counterterrorism Center" proposed by the Sept. 11 commission and possibly by a presidential assistant for national intelligence, charged in addition with making certain that significant competing intelligence assessments reach the president.

There is no shortage of schemes of reorganization: the Sept. 11 commission, the Senate intelligence report, the Scowcroft commission, the Hamre proposal to centralize collection but leave the analytical functions in existing institutions. What is urgently needed is a pause for reflection to form the various proposals into a coherent concept. A small group of men and women with high-level experience in government could be assigned this task with a short deadline, say six months, based on the following principles:

• Centralization must be balanced against diversity.
• Foreign and domestic intelligence should not be merged but should be coordinated by task forces, depending on the subject.
• Special provisions must be made for the systematic enhancement of quality; it cannot be left to moving around boxes on an organizational chart.

No reorganization plan will work if attention is not paid to the morale of the men and women staffing the intelligence services. Despite the portrayal of them around the world as devious master planners dominating policy, intelligence personnel in the real world are subject to unusual psychological pressures. Separated from their compatriots by security walls, operating in a culture suspicious of even unavoidable secrecy, they are surrounded by an atmosphere of cultural ambiguity. Their unadvertised and unadvertisable successes are taken for granted, while they are blamed for policies that frequently result from strategic rather than intelligence misjudgments.

Finding themselves in a kind of political wilderness, the intelligence services have been under assault for 30 years, ever since the floodgates were opened in the 1970s by the Church and Pike committees and subsequent probes in the 1980s and 1990s, which disclosed the names of many agents and almost all clandestine operations. These attacks reflected the political debates of the period. Liberals attacked the intelligence community for being too ideological and Cold War-oriented. Conservatives were critical because they considered the intelligence community not sufficiently ideological nor conscious enough of the element of power in international affairs. Inevitably, between the terms of directors William Colby through John Deutch, the emphasis was to reduce the reliance on agents and to emphasize technical means of collection less subject to the allegations (and sometimes) the reality of abuse. This was a major contributing factor to the shortfall in human intelligence regarding the terrorist threat remarked on by all commissions dealing with recent intelligence failures.

For all these reasons, intelligence reorganization needs to bring as well some stability for intelligence personnel. Thousands of dedicated people participated, at the request of their government, in some of the most important battles of the Cold War and are even now at the front lines of the war with radical, ideological Islam. Their failures must be corrected. But they deserve recognition for their service even as the structures in which they function are being revised.

© 2004 Tribune Media Services Inc.