Twice this past week, on January 23 and 25, the Senate Select Committee on Intelligence held hearings on intelligence reform. Topics included the remaining 9/11 Commission recommendations and efforts both to facilitate information-sharing across the U.S. government's 16 intelligence agencies and to increase the number of operatives and linguists.

The committee's schedule suggests Sen. Jay Rockefeller will use his new majority status and chairmanship to increase oversight and press the Bush administration on matters ranging from CIA rendition programs to the National Security Agency's warrantless surveillance programs. Oversight should be welcome, but neither it nor the 9/11 Commission's recommendations will be enough to rectify the quality of U.S. intelligence analysis.

In a seminal article in the Economist in 1955, historian C. Northcote Parkinson described the behavior of bureaucracies. First, he observed, any "official wants to multiply subordinates, not rivals; and [second,] officials make work for each other." He used the British admiralty to illustrate his case. Between 1914 and 1928, its commissioned ships declined two-thirds. Over the same period of time, the number of officials managing them almost doubled.

As John Negroponte prepares to move from the directorship of National Intelligence to Foggy Bottom, it is clear that his legacy falls far short of real reform. He hired 1,500 employees for his new office, but missed recruitment targets for both operatives and analysts.

This is failure. As both the Iranian nuclear drive and al Qaeda's declared war on the United States continue, the nation needs spies to peer where satellites cannot and men on the ground to hear conversations that take place in caves rather than on cell phones. The failure to recruit and retain quality linguists is also a scandal. While Rockefeller criticizes wiretap procedures, the true outrage is the failure of the intelligence and law enforcement communities to put the products of such surveillance to use. On July 27, 2005, Glenn Fine, inspector general for the U.S. Department of Justice, testified before the Senate Judiciary Committee that the Federal Bureau of Investigation backlog of counterterrorism and counterintelligence audio surveillance awaiting translation had grown from almost 25,000 hours on December 31, 2003, to more than 38,000 hours on March 31, 2005. Department of Justice sources say the problem has not diminished.

While the number of spies and linguists may be a critical metric for gauging U.S. capabilities, access to raw material does not itself correlate with quality analysis. Here, the intelligence community falls short. Take Larry Johnson, a former CIA and State Department analyst to whom the CIA awarded two Exceptional Performance commendations. On July 10, 2001, Johnson penned a New York Times op-ed entitled "The Declining Terrorist Threat." As Mohamed Atta and the other 9/11 hijackers conducted dry runs for their attack, and despite Osama bin Laden's 1998 declaration of war on the United States, Johnson argued that Americans were not primary targets of terrorism. He blamed concern about Islamist terrorism on "24-hour broadcast news operations too eager to find a dramatic story line."

While Johnson is just one public example, the poor quality of the CIA's analytical products is an open secret among intelligence consumers. Reports circulated to the State Department, Pentagon, Treasury Department, National Security Council, and the White House are seldom more analytical or detailed than published newspaper accounts.

The reasons for poor analysis are multifold. The initial premise of a closed analytical shop segregated from policy was to maintain a bank of first-rate social scientists to prevent surprise and predict events. But social science has never lived up to its promise. The Soviet Union's collapse and 9/11 are just two prominent instances of the CIA's failure to predict. While it has become fashionable to scapegoat Iraqi National Congress head Ahmad Chalabi for faulty intelligence about Iraqi weapons, and thereby exculpate the CIA's perhaps $30 billion intelligence operation, the preponderance of Langley's analysis suggested Iraq was permeated with chemical and biological munitions.

Nor did Langley ever attract top academics. Many specialists shy away from government careers. In practice--unbelievable as it may seem--travel and regional experience disqualify applicants during the security clearance process. Those with native fluency in languages like Arabic, Persian, or Pashtun seldom pass CIA vetting. While Langley recruits Mormons returning from missions with linguistic ability, most intelligence hires are book smart but experience poor.

As a result, the products of the intelligence community lack both cultural nuance and a feel for personalities. Too many analysts assume that Iranian officials approach diplomacy with the sincerity of their U.S. counterparts; they cannot imagine the prospect that seminary-trained clerics practice religiously sanctioned dissimulation. Hence, many intelligence professionals at the time believed that Iranian president Mohammad Khatami was sincere in his calls for a dialogue of civilizations; now it is apparent that he pursued Iran's covert nuclear program with the same energy as his successor. When European leaders and Secretary of State Madeleine Albright relaxed sanctions and offered an olive branch to Tehran, the Islamic Republic used the resulting hard currency influx to upgrade Iran's military and fuel its covert nuclear program.

Cubicle isolation is also apparent to anyone who knows the people about whom dossiers are compiled. When writing biographies of Iraqi politicians, CIA analysts commonly erred on such basic information as the languages they spoke, let alone their predilections or personalities.

Extreme compartmentalization also reduces the chances for sound comparative analysis. As the CIA has grown, its analysts' areas of responsibility have narrowed. Expertise in arcane subjects should be welcome, but if it comes at the expense of comparative analysis, much can be lost. Analysis of Iranian nuclear capabilities, for example, should not be separated from study of the North Korean ballistic missile program or Pakistani weapons design. Nor should Iran area specialists be segregated from al Qaeda analysts. Rogue regimes and terrorists do not always compartmentalize relationships as neatly as does the U.S. bureaucracy.

The CIA does have many good analysts, but the organizational prioritization of group-think and seniority strangles them. Bureaucratic interests dominate. As reports filter up through multiple levels, officials insert trap-door statements to assert the opposite of any conclusion so that if the report's thrust is wrong, the agency can be absolved of responsibility. A single sentence questioning Saddam's weapons programs, for example, might be buried on page 17 of a report otherwise declaring their existence.

As former CIA operative Reuel Marc Gerecht points out, the formulaic assumption that any watershed event is five-to-ten years away is both the product of caution and a way to avoid acknowledging ignorance. Repeated statements that Iran is five-to-ten years away from autonomous nuclear capability, for example, have become the 21st-century equivalent of the Ten-Year Rule that left Great Britain scrambling to meet the challenge of a resurgent Germany prior to World War II.

Secrecy protects shoddy analysis. Langley may oust analysts for security reasons, but, like any government body, it seldom purges mediocrity. While intelligence analysts conflate questioning with politicization, the desire to avoid inquiry is often a sign of lack of confidence. Analysts who publish openly and under their own names must, for the sake of their reputations, produce solid work or else they will hemorrhage credibility and jeopardize their employment. Too often, though, intelligence briefers cannot answer basic questions. When queries are followed by requests to see raw intelligence, the source material does not always support the proffered conclusions. Intelligence professionals should be able and willing to defend their products.

It is this phenomenon that was at the root of tension between the CIA's Directorate of Intelligence and the Defense Department's Iraq policy shop. The forthcoming report of an investigation by the Pentagon's Office of Inspector General into the Office of Special Plans will absolve the unit of charges that it produced its own intelligence--it did not--but the report may criticize the office for questioning too much the products it received from Langley. But to move toward a standard of blind acceptance of intelligence would be both dangerous and wrong.

The traditional value of intelligence products was to provide a baseline of neutral expertise, but the era of an apolitical Langley is over. In November 2005, W. Patrick Lang, former defense intelligence officer for the Middle East, South Asia, and counterterrorism, told the American Prospect of CIA analysts' attempts to hurt the White House prior to the 2004 election. "Of course they were leaking," he said. "They told me about it at the time. They thought it was funny. They'd say things like, 'This last thing that came out, surely people will pay attention to that. They won't reelect this man [President Bush].'" Intelligence analysts should not participate in policymaking. Their frequent and as-yet-unplugged leaks may win some short-term policy battles for Langley, but such illegalities have badly damaged trust. To suggest the Directorate of Intelligence is policy-neutral is risible.

So what is the solution? Washington's inclination is always to expand hiring. But that will constrain rather than improve analysis. Today, the CIA's analytical wing is the ultimate expression of Parkinson's Law, rather than a generator of accurate explanation or prediction. Rather than expand, the government should privatize much of its analysis.

Privatization works. Already, Beltway firms like SAIC and Booz Allen Hamilton operate streamlined intelligence shops. Their analysts hold the highest security clearances. So do many think-tank scholars and some university academics. Many private-sector analysts have language abilities and experience their government counterparts lack.

Freeing analysts from some government rules and regulations could improve their products. Not only would it enable outside-of-the-box thinking, but it could also improve access. U.S. government personnel visiting Beirut, let alone Baghdad, must adhere to embassy regulations stipulating intrusive security for travel outside compound walls. Nongovernment employees roam free--or at least set their own rules for security.

Privatization would improve productivity. It can take the CIA hierarchy weeks to sign off on an analyst's report and release it to intelligence consumers across the U.S. government. Private companies react faster. Competition might also expedite exploitation of several million pages of documents seized in Afghanistan and Iraq.

A decade ago, the CIA curtailed its subscriber-based circulation of foreign newspapers and media broadcasts in translation, partly for financial reasons and partly out of misdirected hand-wringing that such products might violate even North Korean and Iranian intellectual property rights. Today, the Open Source Center, the office within the CIA that translates published material, still withholds much of its product from the public. Getting this into the hands of a wider pool of analysts would be in the national interest, even if the analysts offered differing interpretations.

Expanding the pool of professionals who hold security clearances would have auxiliary benefit. Not only would it enable more opinion and debate without the costs of salary and pension; but, in the long term, it would also erode the clearance lag. Even with "expedite" orders, whoever wins the presidency in 2008 will have to wait 15 months to staff the National Security Council with new faces unless they already hold clearances. At present, the CIA spends hundreds of thousands of dollars to screen and train analysts who may leave government service after only a couple of years. Making it easier for the U.S. government to employ such people would increase return on investment.

There would be drawbacks to more privatization--security and counterintelligence problems would expand--but the risks need not be excessive. Even State Department student interns receive top-secret clearances. Access to government products should still require background checks, security clearance, and the incumbent oaths to protect the material. The FBI and other relevant agencies should nevertheless expand counterintelligence checks. Dissemination of sensitive compartmentalized information like signals and communications intercepts should, of course, remain subject to the presence of adequate facilities to handle and protect the information.

Some outside scholars might also cherry-pick data. But then government intelligence analysts do so now. While data are open to interpretation, competition exposes bad methodology, and ultimately quality shines through. Individual authorship promotes accountability.

Would the CIA's analytical wing disappear? No. But it should shrink, as the pool of outside experts expands. Much of the money allocated for the analytical wing would be better applied to the Directorates of Operations and Science & Technology. Langley and its consumers might maintain yellow pages of analysts by expertise and repositories of finished products. Congressmen could call on individuals to explain their reports or even have multiple specialists debate interpretations. It is not uncommon in, say, the Pentagon for senior leaders to host closed debates among academics and analysts. One thing is certain, though. With threats multiplying, bloat and a culture of job security over performance will neither protect the United States nor promote the serious thinking needed to help it face new challenges.

Michael Rubin, a resident scholar at the American Enterprise Institute, was a staff adviser for Iran and Iraq in the office of the secretary of defense between 2002 and 2004.

Next Page