One of the other sources of the mobile lab information was a German with the apt code name of “Curveball.” Only after the war had started and stockpiles of WMD had failed to materialize did I hear rumblings that the Germans did not trust their own source. What I did not know during the drafting of the speech was that operatives from the Agency’s European Division wanted McLaughlin to know that the Germans had concerns with Curveball. McLaughlin’s executive assistant chaired a meeting with the operatives and the analysts. The latter insisted they were confident in the source—in large part because there were three other sources backing up his story (usually not a reason to be confident in someone’s credibility)—and that an independent review from the National Labs had stated that the reporting was plausible. (The National Labs are institutions that conduct research for the US government on nuclear and other advanced weapons.) In any event, the analysts apparently made such a persuasive case that the EA did not see a reason to highlight the concerns of the Germans. I have seen nothing to suggest to me that either McLaughlin or Tenet was made aware of the dispute at the time. As it turned out, Curveball was a fabricator and later admitted that he had lied because of a personal agenda to get rid of Saddam.
In retrospect, this could have been a turning point. Had the analysts said to Powell, Tenet, and McLaughlin that two of the four streams of information on the mobile labs lacked credibility, it would have raised eyebrows. It would have at minimum led us to review the mobile lab issue, if not the broader issue. One of the three could well have said, “So, what else lacks credibility here?”
While the secretary’s work on the WMD portion of the presentation was methodical, the discussion on the terrorism portion was short, but dramatic. The first draft of this part of the speech—which was written by Scooter Libby or one of his aides in the vice president’s office, and which did not go through my vetting process—did not say directly but implied that Saddam was complicit in 9/11. It was wrong, and Secretary Powell saw the inaccuracy immediately. Several senior White House officials, including Rice, as well as Hadley, sat in on some of the sessions with the secretary, and with Libby in the room, Powell said to Tenet, “George, you don’t believe this crap about Iraq and al Qa‘ida, do you?” Tenet said, “No, Mr. Secretary, we do not. I do not know where that came from, but we will fix it.” Tenet then turned to the deputy director of CIA’s Office of Terrorism Analysis, Philip Mudd, and said, “Rewrite this.” Libby did not say a word. It had been yet another attempt to politicize the terrorism analysis. What the secretary ultimately said about Iraq and terrorism at the United Nations matched very closely what Mudd wrote in the hours after the discussion between Powell and Tenet, and was fully consistent with the second CIA paper on the topic, Iraq Support to Terrorism.
* * *
On a number of occasions in recent years, Secretary Powell has expressed chagrin that no one from the intelligence community has publicly come forward and apologized to him for putting his well-deserved reputation for probity at risk by arming him with bad intelligence to use as the basis of his UN speech. I am absolutely confident that no one at CIA intentionally misled him, politicized analysis, or tried to provide anything but the best information—but CIA and the broader intelligence community clearly failed him and the American public. So, as someone in the chain of command at the time the Iraq WMD analysis was provided, I would like to use this opportunity to publicly apologize to Secretary Powell.
* * *
As we approached war, the views among many Middle Eastern specialists at CIA were decidedly antiwar. Many at the Agency were concerned that bringing down Saddam would open a Pandora’s box. While CIA should not put policy advice on the table, I believe we did have a responsibility prewar to produce a detailed analysis of the likely postwar scenarios—with a clear outline of the key factors that would have determined whether we ended up with stability or instability. That paper was never written (given the importance of the president’s decision, it should have been a National Intelligence Estimate). No analyst initiated such a paper and no one in the chain of command, from a first-line supervisor at CIA all the way to the president, requested it.
My own feelings about the war were mixed. On the one hand, based on the views of the analysts, I believed Saddam had chemical and biological weapons, was restarting his nuclear weapons program, and was working on missiles of various ranges to deliver all these weapons. I worried less about Saddam’s giving these weapons to al Qa‘ida than I did about his eventually using them himself. And I worried that his acquisition of a nuclear weapon would give him the confidence to pursue a very aggressive foreign policy in the region. To top it off, Saddam had ignored nearly a dozen UN Security Council resolutions for years, and either the UN matters or it does not. On the other hand, Saddam was not an imminent threat, and I was uncomfortable with US military action—putting the lives of young Americans at risk—in the absence of imminence. And I worried about the unintended consequences of military action, which are always numerous and significant.
In retrospect it is easy to criticize the decision to go to war. But I understand why the president felt it necessary. And it is hard to say that anyone presented with the same facts and burdens would have come to a different conclusion. After all, most of Congress saw the war as necessary for the same reasons that the president did.
* * *
I was well into another assignment by the time we learned that we had been wrong in our judgments. This awareness did not happen overnight. It came on slowly, as coalition forces on the ground in Iraq could find nothing at all related to an Iraqi WMD program and as Tenet set up a survey group to thoroughly investigate the matter. When it finally became clear, we knew that this had been one of the largest intelligence failures in the history of the Agency.
Much has been written and said about what went wrong and my intent here is not to offer excuses or to be defensive. But after spending much time analyzing why we were so wrong, I think I have some explanations of value. This is certainly not a thorough or complete analysis of this massive intelligence failure, but it highlights what I think was most important.
The first thing that should be said is that CIA’s judgment about Saddam and WMD was nothing new, nor was it unique. Neither of these facts makes it OK that CIA was wrong, but they are important to consider. This analytic conclusion went back to the Clinton administration, and it was not just the US intelligence community that believed Saddam had a weapons program. All foreign intelligence services that looked at the issue thought the same, as did think tanks, distinguished university professors, and even the United Nations. If you had polled Iraqi military and intelligence officers in 2002, I am convinced you would have found a large majority saying their boss had active WMD programs.
The perception that the Bush administration pushed the intelligence community toward believing that Saddam had WMD is just wrong. No one pushed us—we were already there. The notion that we were telling the White House what it wanted to hear can easily be debunked. Look at the question of Saddam’s connections to al Qa‘ida. We held our ground and refused to go where the intelligence did not take us. On WMD, if we’d believed it was likely that Saddam had none, it would have been an act of madness to take the position we did. Following an invasion, a stockpile would either turn up or not. To go to war knowing you are soon going to be proven wrong would be insane.
So how could we get it so wrong? The answer is major mistakes on both the analytic and collection sides of the intelligence community—the latter being something that has not often been discussed.
First let me address the analytic question. The problems with the analysis were numerous, although a lack of resources was not one of them (Iraq was a priority focus for analysts at CIA and throughout the intelligence community). One problem was what I call “analytic creep.” What had begun as assumptions about Saddam and his program became firm judgments. It took a team of our smartest analysts, looking back with twenty-twenty hindsight, to identify this an
alytic creep.
In addition, biases were rampant. One of these was the “hindsight bias” on the part of the intelligence community analysts who had missed Saddam’s nearly acquiring a nuclear weapon in the early 1990s and did not want to make that mistake twice. Another was the “historical bias”—we knew that Saddam had chemical weapons and had used them on Iraqi Kurds and on the Iranians; and if he’d possessed this capability once, he must still have had it. The third was “review bias.” In the late 1990s, Donald Rumsfeld, then a private citizen, had led a commission that assessed the intelligence community’s judgments over time regarding the foreign ballistic missile threat to the United States. His critique of the analysts was blistering. He chastised them—in both his team’s interviews with them and the final report—for focusing on the most likely outcome rather than on what could be the worst-case outcome. The truth is that analysts must do both, but there is no doubt in my mind that the Rumsfeld Commission created its own bias with regard to Saddam’s weapons program.
Another problem was confirmation bias—the tendency to accept facts as true if they support one’s view and to reject them otherwise. Some of the sources on which we relied were clearly sent our way by outside groups with an agenda, like the dissident group the Iraqi National Congress (INC), which never met a “Saddam has WMD” story it didn’t like, and we accepted much of the reporting of these sources as fact. After the invasion, INC leader Ahmed Chalabi bragged about the information provided by members of his group, calling them “heroes in error.” Well, he got that half right.
But by far the biggest mistake made by the analysts—and one that encompasses all the above issues—was not that they came to the wrong conclusion about Iraq’s WMD program, but rather that they did not rigorously ask themselves how confident they were in their judgments. If you took all the intelligence available to the analysts when they drafted the NIE and looked at it today, you would come to the conclusion that Saddam had chemical and biological weapons programs and was on the verge of restarting his nuclear weapons program. But had the analysts at the time thoughtfully and rigorously asked themselves how confident they were in those judgments, they would most likely have said, “Not very.” That would have been a very different message to the president and other policy-makers and potentially could have affected their policy decision.
Anyone who has seen the now-declassified Key Judgments from the NIE knows that a text box at the end briefly discusses confidence levels and notes that analysts had high confidence in their judgments that Saddam had chemical and biological weapons and was reconstituting his nuclear program. But the analysts did not really think about that statement before making it. It was a reflection of their gut view. It did not reflect a thorough assessment of the question of confidence levels. Such a rigorous assessment was missing. It was simply not part of an analyst’s tool kit in those days.
Why do I think we would have reported low confidence if we had carefully considered the question? Because the vast majority of the information we were working with was at least four years old. It dated back to 1998, before UN weapons inspectors were kicked out of Iraq; once they’d left the country, our intelligence became more limited. And to top it off, the case in support of the notion that Saddam had active WMD programs was largely circumstantial. As noted earlier, we also relied too heavily on the knowledge that he’d once had chemical weapons, once used chemical weapons, and, at the start of the first Gulf War, been much closer to having a nuclear weapon than we had previously thought. The fact that he’d once had chemical weapons and had once pursued a nuclear weapon was actually irrelevant to whether he was doing so in 2003.
In addition to the brief sentences on confidence levels in the Key Judgments of the NIE, we created an impression that the analysts had high confidence in other ways. The text of the NIE was nuanced, but the Key Judgments were not. When we wrote pieces for the president, the analysts wrote with authority on the issue. This is why I personally never found fault with George Tenet’s alleged “slam dunk” comment. The way the analysts talked and wrote about their judgments would have led anyone to think it was a slam dunk—that is, that Saddam definitely had active WMD programs. No one ever said to me, Miscik, McLaughlin, Tenet, Rice, or the president, “You know, there is a chance he might not have them.” Such a statement would have gotten everyone’s attention.
In the aftermath of the Iraq debacle, those of us on the analytic side of CIA spent a lot of time thinking about how we had failed. Miscik established a lessons-learned task force called the Iraq Review Group. This was an important step, as it showed the DI was willing to take ownership for the failure. The Review Group did its job with thoroughness and rigor. It identified a number of failures, and it made a number of important recommendations. Stealing an idea from the military, which periodically holds “safety stand-downs” following aviation mishaps, Miscik set aside a couple of days for the Review Group to present its findings to every analyst at the Agency. Miscik is to be commended for how she handled the aftermath. She showed leadership, integrity, and professionalism, and the Agency today is much better off for the effort.
The key recommendation made by the Review Group—which was adopted—was that in all future major intelligence products, analysts be required to include a thorough assessment and explicit statement regarding their level of confidence in the judgments expressed. This practice is now ingrained in both analysts and policy-makers. Confidence levels are part of the normal conversation that happens every day in the Agency, in the intelligence community, and in the Situation Room. They were part of the discussion about whether Syrian president Bashar al-Assad was using chemical weapons against his own people in 2012 and 2013. At first the analysts were at low confidence, but over a period of months they moved to high confidence. Policy-makers were fully aware of the analysts’ changing view on this, as we discussed it regularly in Deputies Committee meetings during this time.
There was a second major factor in the intelligence community’s failure on Iraq WMD intelligence. Unlike the analytic failure, this one has rarely been discussed in public or within government. It was not studied in the aftermath for lessons learned. This was a failure of intelligence collectors—CIA and the NSA—to penetrate Saddam’s inner circle, where they might have been able to learn the truth. The leadership of the operational sides of CIA and NSA should have requested a lessons-learned assessment, but they did not.
Of course, gaining access to the handful of people close to Saddam, who truly understood the status of Iraq’s WMD programs, was hard, in large part because Iraq was a “denied area,” a place where there is minimal or no official US government presence. We face the same problem in other opaque regimes like Iran and North Korea that go out of their way to make sure their actions are shrouded in darkness. But hard can’t be an excuse at the end of the day, because the intelligence community is paid to do hard. And, at a minimum, we should have told the president and his national security team that we did not have good access to Saddam and his inner circle.
I believe that one of the reasons CIA failed on the collection front—which should be a lesson learned moving forward—was our focus on covert action in Iraq. During the 1990s the United States had been intent on regime change in Iraq—in 1998, Congress made it the stated policy of the United States—and CIA had been in the lead. The day-to-day aim of our operations officers at that time had been to build ties to the Kurds in northern Iraq who might play a role in the overthrow of Saddam, and who were providing us with locations from which to operate against him. Our collection focus was on finding Sunnis in Saddam’s military who might be willing and able to overthrow him and take control of the country (and develop a new and much different relationship with the United States in the process). With all this, collection on other issues related to Iraq—including WMD—suffered.
It is important to remember that CIA doesn’t authorize covert action. That’s a policy decision that requires the direction and signature of the president o
f the United States. Covert action has a number of unseen costs. One is that it diverts you from traditional foreign intelligence collection. When an administration gives CIA the mission of conducting a covert action, it doesn’t assign additional people to perform the mission. The Congressional oversight committees are briefed on covert actions, and they sometimes provide additional funding, but they do not raise the Agency’s personnel ceiling just because there is a new plan. So the folks who could have been trying to figure out how to collect intelligence from Saddam’s inner circle to discover Saddam’s plans, intentions, and capabilities with regard to weapons of mass destruction were diverted to trying to find generals willing to overthrow him.
We had the perfect storm of imperfect intelligence. We were not collecting the kind of information that would have saved us from inaccurate analysis, and we were not rigorously asking ourselves how confident we were in the collected information. Had we done either, the intelligence outcome would have been different—and possibly the policy outcome as well.
Charles Duelfer, who led the US WMD hunt in Iraq after the invasion, concluded that Saddam had wanted to maintain the appearance of having weapons of mass destruction in order to deter his number one enemy, Iran. But Duelfer found that Saddam had thought that US intelligence was good enough to figure out the real story and, therefore, that the United States would eventually lower the sanctions and, more important, not attack him. Even Saddam turned out to be overconfident in US intelligence capabilities.
The Great War of Our Time: The CIA's Fight Against Terrorism--From Al Qa'ida to ISIS Page 11