In response to:

How They Got Their Bloody Way from the May 27, 2010 issue


Larry Downing/Reuters

George W. Bush awarding the Presidential Medal of Freedom to former CIA director George Tenet, Washington, D.C., December 2004

To the Editors:

Thomas Powers’s conclusion that the CIA’s vast overestimate of Iraq’s WMD capability was largely the result of political pressure [“How They Got Their Bloody Way,” NYR, May 27] is the one with which I began my study Why Intelligence Fails, but logic and evidence convinced me that I was wrong. Although definitive proof is impossible, Powers ignores rather than refutes my evidence. In brief, I was struck by the fact that as far as we can tell all intelligence services, even those in countries that opposed the war, came to conclusions similar to those held by American intelligence, the parts of the American intelligence community that would have been most susceptible to political pressures dissented from parts of the WMD judgments, and most of the conclusions were reached before political pressures mounted.

The two points Powers makes at the end of his review that he considers so damning actually are central to my argument. He claims that I ignored the fact that American intelligence did not tell the UN inspectors where to find Saddam’s WMD facilities. I touch on this on page 136, and while it remains unclear how much the US did tell the inspectors, the decisions were made by the White House, not the CIA. It is also true that the CIA failed to modify its judgment on the basis of the reports received by the UN inspectors when they were permitted to reenter the country. There are mysteries here, but my judgment (pages 135–136) is that there was “politicization late in the day” as analysts saw that war was inevitable and simply got out of the way, behavior that was unfortunate but that contrasted with their earlier level of engagement.

Powers’s other point is that the CIA’s deputy director for intelligence, Jami Miscik, threatened to resign unless the White House stopped pressuring her. But her complaints were about the CIA’s refusal to affirm links between Saddam and terrorism, not about its WMD findings, which was the topic of my analysis. This is a key point. If politicization explained intelligence assessments, we would find them converging with administration preferences. But on Iraq and terrorism, they never did.

On WMDs, the administration did not need to exert pressure because it found intelligence assessments to be generally acceptable, although as all observers (including Powers and myself) note, top officials still felt the need to issue public statements that went beyond the intelligence. The “evidence” Powers cites is then one of the reasons I changed my mind—if the CIA did not bow to pressures to link Saddam with terrorism (and also wrote papers predicting that the postwar occupation would be difficult), why should we believe that its WMD assessments were the result of political pressures rather than honest error?

I would very much like to blame Bush and Cheney for the intelligence failure, as they can be blamed for the war. This fits with my views that this administration did enormous harm to the US and the world. It also would mean that we could expect much better from intelligence when it was not impeded by political pressures. Unfortunately, my research found that such an optimistic conclusion lets intelligence off the hook too easily and leads to false expectations about the accuracy of future estimates.

Robert Jervis
Adlai E . Stevenson Professor of International Politics
Columbia University
New York City

Thomas Powers replies:

There is a circumspect spirit in Robert Jervis’s letter that also struck me in his book. In my view the CIA was not merely wrong in claiming that Saddam Hussein was “actively seeking” nuclear weapons. The evidence was sparse, fragmentary, thinly supported, inconclusive, and in crucial instances flatly contradicted; but Jervis avoids the implications of what is most obvious about this error—its nature and its magnitude. The administration needed a scary intelligence estimate that would stand up long enough to persuade Congress to authorize war. This one did the job for three reasons: because the administration had created an atmosphere of crisis that Congress was too timid to challenge or resist; because the CIA put its authority on the line by claiming it had “high confidence” in its findings; and above all because the evidence was secret. To read the CIA’s estimate now is shocking; nothing in the grab bag called for the haste to act insisted on by the President at the time.

But Jervis treats the CIA’s “intelligence failure” as an honest mistake of the garden variety, an accumulation of shaky assumptions, small errors, and misreadings that led the agency astray. I cannot see why Jervis fails to connect the dots. Parsing the case does not require bales of subpoenaed evidence. The nature of the error (fluffing up the casus belli demanded by the White House) as well as the magnitude of the error (wrong in every particular) tell us what we need to know.


A reply to a letter is not the place to reargue all the particulars of the case; Jervis’s version is in his book and mine can be found in a review of George Tenet’s memoir, At the Center of the Storm [NYR, July 19, 2007]. Jervis is wrong when he says that countries like Russia, France, Canada, and Germany, all opposed to the war, shared the American view of Iraqi WMD programs. They did not. Germany in particular went out of its way to warn the CIA against trusting old reports about Iraqi biological programs from the agent known as “Curveball”—warnings that Tenet not only ignored but said he had never received.1 If the CIA bowed to political pressure, then Tenet was the man who bent first and most. But not, as Jervis fairly points out, in every case on every particular. In the now notorious instance of the allegations that Saddam Hussein was trying to buy Nigerian yellowcake, Tenet blocked the administration from citing the report so often that President Bush, in the end, cited British claims as justification instead.

It is worth remembering, however, that the CIA during its long argument over the yellowcake documents failed to note that they had been fabricated—so transparently that UN weapons inspectors pronounced them bogus within a single day. It was not just the CIA that showed a breathtaking lack of curiosity about the fabrication of these documents; Congress has also shown no interest in who cobbled them together. The broad answer is that it was someone trying to build the case for Iraq’s WMDs as a cause for war—not the CIA (which declined to make use of them), and probably not the British, but more likely an activist, freelance cabal with access to the Italian intelligence circles from which the documents emerged.

Why has no official American body wanted to know who went to such lengths to make the case for war? There are two possible answers to this question that, in their way, mirror the differing views of the “intelligence failure” on Iraq held by Jervis and myself. One answer might be that the case is too hard—who could hope to unravel a skein so knotted with shadowy operatives and stale clues? My answer would be the opposite—that such an investigation would lead too easily, too directly in troubling directions. Cui bono? Those pushing for war could be found in the White House, the Pentagon, and the neoconservative cabal that had seized on September 11 as a heaven-sent opportunity to overthrow Saddam Hussein at last.

Jervis charges the CIA only with “politicization late in the day,” but that is when politicization counts. The bottom line is what the estimate says. The CIA writes estimates. What the State Department and the Department of Energy think goes into footnotes and does not matter. The purpose of Vice President Cheney’s eight visits to the agency was to ensure that the estimate included what the White House needed. I understand Jervis’s reluctance to see dark motives in this case; arguments about motives can turn ugly. But in my view understanding intelligence disputes often requires taking a harsher view.

Let me try to spell this out briefly. When Jervis fails to identify the White House–CIA two-step that preceded the war in Iraq he is overlooking a fundamental fact about intelligence in America—on issues the president takes seriously, there is no daylight between the White House and the agency. The one does what the other wants. Moments of friction do not last long and are (almost) always resolved in favor of the White House. Asking whether the CIA’s WMD estimate was the result of White House pressure is like asking whether President Reagan knew about Iran-contra, or whether President Kennedy knew about the plots to kill Castro.

Answering such questions, common in American history since the founding of the CIA, requires only that you remember who the agency is working for. Some, perhaps including Jervis, would think this too cynical a view, but I would argue that it offers the single most useful rule for understanding what is going on. The rule works in both directions: when you know what the president wants, you know what the CIA is doing, and when you know what the CIA is doing, you know what the president wants. In the fall of 2002 the President wanted war, and the CIA did its bit to give him what he wanted.