[NousDefionsDoc]

The War

Al Qaeda was the classic failure for the CIA. Al Qaeda was not a national government but a small, apparently eccentric, collection of Islamists. This was already outside of the CIA's sweet spot. The Sept. 11 attacks were completely outside the paradigm that the CIA -- and others, including Stratfor -- was working with. The model of terrorism they had studied for a generation did not include an attack of this order. Therefore, since the CIA was dealing with a non-state group and with a historical discontinuity, the agency continued its record for getting it wrong.

The problem the CIA has is that it also failed in what was supposed to be its sweet spot -- covert gathering of intelligence from senior state officials in Iraq concerning a war that had been going on, in effect, since 1990. There were no surprises here, no discontinuities, no funky, off-the-wall groups. This was mainline intelligence-gathering.

It was here that the CIA made the core mistakes:

1. It did not tell either Presidents Bill Clinton or George W. Bush that Iraq had no weapons of mass destruction. It told both of them that it did.
2. It did not understand Saddam Hussein's war plan and did not warn Bush that the fall of Baghdad would trigger an organized guerrilla war. Warning of unrest is absolutely not the same as warning of a war plan.
3. It did not provide clear intelligence on the status of the Shia in Iraq and the degree of organization that had been achieved by Iranian intelligence.

The failure to predict the Sept. 11 attacks was ultimately a systematic failure hardwired into the CIA. The ways in which it collects intelligence and the way its analytic process works have consistently generated failures on this level. When an institution fails to do a certain type of work well for 50 years, it is hardly fair to condemn it when it repeats the failure. The failures it can be condemned for, however, are the mainstream collection and analytic failures that shaped the Iraq campaign.

This is what the debate has raged over. The Bush administration gave the CIA a pass over Sept. 11; they are not giving it a pass over Iraq. The CIA is responding by arguing (a) they were forced to skew data and (b) they did provide accurate analysis but were ignored. The administration is arguing (a) no one forced them to skew data and (b) the claim that they did provide accurate intelligence undermines the claim in (a). They say that the CIA was just dead wrong in its intelligence and then tried to cover it up by savaging the administration.

There are two conclusions here. First, the fact that the agency is being given a pass on Sept. 11 is the most serious problem. The consistent inability of the CIA to capture hard-to-source discontinuities is not a charming foible, but an unacceptable shortcoming. Being good in the small things doesn't matter if you can't do the big things. On that basis alone, the CIA should be rebuilt. But it is not on that basis that the administration is going after them.

Instead, they are going after the failure of the agency to do the small things right. We tend to agree that the CIA's failures in Iraq are too numerous to be explained by political pressure. The consistent inability to generate radical analysis is caused by the inherent conservatism of the complex process that Tenet put into place. Where committees rule, the product will be the lowest common denominator. On this, we side with the administration.

However, the problem is not simply to streamline a process that works well on small things. The administration doesn't appreciate the fact that the enormous failures of the CIA on the big things are the real problem. It is what gave us everything from the Chosen Reservoir to bafflement at the sight of the Berlin Wall coming down. The administration is missing its chance to rebuild American intelligence in fundamental ways -- and there is no better time to do it than during a war, as "Wild Bill" Donovan and the OSS showed.