Tough Questions We Were Right to Ask

By Douglas J. Feith
Wednesday, February 14, 2007

Promoters of the "Bush Lied, People Died" line claim that the recent Pentagon inspector general's report concerning my former office's work on Iraq intelligence supports their cause. What the IG actually said is a different story.

The IG, Thomas Gimble, focused on a single Pentagon briefing from 2002 -- a critique of the CIA's work on the Iraq-al-Qaeda relationship. His report concluded that the work my office generated was entirely lawful and authorized, and that Sen. Carl Levin was wrong to allege that we misled Congress.

Gimble made Levin happy, however, by calling the Pentagon briefing "inappropriate," a word the senator has whipped into a political lather. At issue is a simple but critical question: whether policy officials should be free to raise questions about CIA work. In Gimble's opinion, apparently, the answer is no. I disagree.

The CIA has a hard job. Some of its work has been good; some has been famously and disastrously bad, as everyone familiar with the Iraqi weapons of mass destruction fiasco knows. Intelligence is inherently sketchy and speculative -- and historically often wrong. It is improved when policy officials freely probe and challenge it.

In evaluating our policy toward Iraq after Sept. 11, 2001, my office realized that CIA analysts were suppressing some of their information. They excluded reports conflicting with their favored theory: that the secular Iraqi Baathist regime would not cooperate with al-Qaeda jihadists. (We now face a strategic alliance of jihadists and former Baathists in Iraq.) Pentagon officials did not buy that theory, and in 2002 they gave a briefing that reflected their skepticism. Their aim was not to enthrone a different theory, but to urge the CIA not to exclude any relevant information from what it provided to policymakers. Only four top-level government officials received the briefing: Donald Rumsfeld, George Tenet, and (together) Stephen Hadley and I. Lewis "Scooter" Libby.

A 2004 Senate intelligence committee report praised the quality of the Pentagon's Iraq-al-Qaeda work -- the critical briefing and the related Pentagon-CIA dialogue. The policy officials "played by [intelligence community] rules" and asked questions that "actually improved the Central Intelligence Agency's products," it said. Levin and Sen. Jay Rockefeller both endorsed that judgment.

The recent inspector general's report argues that policy officials "undercut" the CIA by pointing out " 'fundamental problems' with the way the Intelligence Community was assessing information" on the issue of Iraq-al-Qaeda relations -- even though Gimble last week said at a Senate hearing: "Again, I need to just remind everyone, we didn't make an assessment on the validity of either side of this issue." He labeled the Pentagon briefing "inappropriate" not because of any errors in it but because he viewed it as an "intelligence activity" that "varied" from "the intelligence community consensus." The Pentagon officials told the IG, however, that the briefing was a policy activity -- a critique of an intelligence product.

If this report hadn't become part of a political battle, Gimble's position would be scoffed at across the political spectrum. Sensible people recognize the importance of vigorous questioning of intelligence by the CIA's "customers." In bipartisan, unanimous reports on Iraq intelligence, both the Senate intelligence committee and the Silberman-Robb WMD commission called for more such questioning.

Specifically on the Pentagon's criticism of Iraq-al-Qaeda intelligence, the 2004 Senate report noted that our challenges were helpful: Intelligence analysts "stated that the questions had forced them to go back and review the intelligence reporting, and that during this exercise they came across information they had overlooked in initial readings."

In his Senate testimony, Gimble said his report -- and therefore all related claims that my office "manipulated intelligence" -- concerned only this single briefing. His whole argument rests on the claim that the briefing was "disseminated" as "an intelligence product" rather than a policy product. But he acknowledged that neither Rumsfeld nor Tenet could have mistaken the briefers for intelligence community spokesmen. His objections applied solely to the briefing that Hadley and Libby received in September 2002.

Astonishingly, the IG acknowledged that his office had not interviewed either of these officials to ask whether they thought the briefing was an intelligence product. Knowing well that the briefers worked for me, neither could have believed that Pentagon policy officials were speaking for the intelligence community. One of the briefing slides was "Fundamental Problems With How the Intelligence Community Is Assessing Information." Gimble had no basis to say the briefing was seen as an intelligence activity.

In his report, Gimble wrote that the Pentagon briefing was not the "most accurate analysis of intelligence." This has been taken to suggest it was false or deceptive. But the IG said he meant only that the briefing was at "variance with the consensus of the Intelligence Community." Of course it was at variance! It was a critique. That's why it was prepared in the first place.

Gimble's characterization is absurdly circular. Cheered on by the chairmen of the Senate intelligence and armed services committees, he is giving bad advice based on incomplete fact-finding and poor logic. He is discouraging tough questioning of intelligence. Our government needs more such questioning, not less.

The author was undersecretary of defense for policy from July 2001 until August 2005.

© 2007 The Washington Post Company