An "Intelligence Failure"

How It Happened

The Intelligence Community's errors were not the result of simple bad luck, or a once-in-a-lifetime "perfect storm," as some would have it. Rather, they were the product of poor intelligence collection, an analytical process that was driven by assumptions and inferences rather than data, inadequate validation and vetting of dubious intelligence sources, and numerous other breakdowns in the various processes that Intelligence Community professionals collectively describe as intelligence "tradecraft." In many ways, the Intelligence Community simply did not do the job that it exists to do.

Our review revealed failings at each stage of the intelligence process. Many past discussions of the Iraq intelligence failure have focused on intelligence analysis, and we indeed will have much to say about how analysts tackled the Iraq WMD question. But they could not analyze data that they did not have, so we begin by addressing the failure of the Intelligence Community to collect more useful intelligence in Iraq.

There is no question that collecting intelligence on Iraq's weapons programs was difficult. Saddam Hussein's regime had a robust and ruthless security system and engaged in sophisticated efforts to conceal or disguise its activities from outside intelligence services--efforts referred to within the Intelligence Community as "denial and deception." The United States had no Iraq embassy or official in-country presence; human intelligence operations were often conducted at a distance. And much of what we wanted to know was concealed in compartmented corners of the Iraqi regime to which few even at high levels in the Iraqi government had access.

Still, Iraq was a high-priority target for years, and the Intelligence Community should have done better. It collected precious little information about Iraq's weapons programs in the years before the Iraq war. And not only did the Community collect too little, but much of what it managed to collect had grave defects that should have been clear to analysts and policymakers at the time. Indeed, one of the most serious failures by the Intelligence Community was its failure to apply sufficiently rigorous tests to the evidence it collected. This failure touched all the most salient pieces of evidence relied on by our intelligence agencies, including the aluminum tubes, reporting on mobile BW, uranium from Niger, and assertions about UAVs.

One of the most painful errors, however, concerned Iraq's biological weapons programs. Virtually all of the Intelligence Community's information on Iraq's alleged mobile biological weapons facilities was supplied by a source, codenamed "Curveball," who was a fabricator. We discuss at length how Curveball came to play so prominent a role in the Intelligence Community's biological weapons assessments. It is, at bottom, a story of Defense Department collectors who abdicated their responsibility to vet a critical source; of Central Intelligence Agency (CIA) analysts who placed undue emphasis on the source's reporting because the tales he told were consistent with what they already believed; and, ultimately, of Intelligence Community leaders who failed to tell policymakers about Curveball's flaws in the weeks before war.

Curveball was not the only bad source the Intelligence Community used. Even more indefensibly, information from a source who was already known to be a fabricator found its way into finished pre-war intelligence products, including the October 2002 NIE. This intelligence was also allowed into Secretary of State Colin Powell's speech to the United Nations Security Council, despite the source having been officially discredited almost a year earlier. This communications breakdown could have been avoided if the Intelligence Community had a uniform requirement to reissue or recall reporting from a source whose information turns out to be fabricated, so that analysts do not continue to rely on an unreliable report. In the absence of such a system, however, the Defense Intelligence Agency (DIA), which disseminated the report in the first place, had a responsibility to make sure that its bad source did not continue to pollute policy judgments; DIA did not fulfill this obligation.

Lacking reliable data about Iraq's programs, analysts' starting point was Iraq's history--its past use of chemical weapons, its successful concealment of WMD programs both before and after the Gulf War, and its failure to account for previously declared stockpiles. The analysts' operating hypothesis, therefore, was that Iraq probably still possessed hidden chemical and biological weapons, was still seeking to rebuild its nuclear weapons program, and was seeking to increase its capability to produce and deliver chemical and biological weapons. This hypothesis was not unreasonable; the problem was that, over time, it hardened into a presumption. This hard and fast presumption then contributed to analysts' readiness to accept pieces of evidence that, even at the time, they should have seen as seriously flawed.

In essence, analysts shifted the burden of proof, requiring evidence that Iraq did not have WMD. More troubling, some analysts started to disregard evidence that did not support their premise. Chastened by the effectiveness of Iraq's deceptions before the Gulf War, they viewed contradictory information not as evidence that their premise might be mistaken, but as evidence that Iraq was continuing to conceal its weapons programs.

The Intelligence Community's analysis of the high-strength aluminum tubes offers an illustration of these problems. Most agencies in the Intelligence Community assessed--incorrectly--that these were intended for use in a uranium enrichment program. The reasoning that supported this position was, first, that the tubes could be used in centrifuges and, second, that Iraq was good at hiding its nuclear program.

By focusing on whether the tubes could be used for centrifuges, analysts effectively set aside evidence that the tubes were better suited for use in rockets, such as the fact that the tubes had precisely the same dimensions and were made of the same material as tubes used in the conventional rockets that Iraq had declared to international inspectors in 1996. And Iraq's denial and deception capabilities allowed analysts to find support for their view even from information that seemed to contradict it. Thus, Iraqi claims that the tubes were for rockets were described as an Iraqi "cover story" designed to conceal the nuclear end-use for the tubes. In short, analysts erected a theory that almost could not be disproved--both confirming and contradictory facts were construed as support for the theory that the tubes were destined for use in centrifuges.

In the absence of direct evidence, premises and inferences must do. Analysts cannot be faulted for failures of collection. But they can be faulted for not telling policymakers just how little evidence they had to back up their inferences and how uncertain even that evidence itself was. The October 2002 NIE and other pre-war intelligence assessments failed to articulate the thinness of the intelligence upon which critical judgments about Iraq's weapons programs hinged.

Our study also revealed deficiencies in particular intelligence products that are used to convey intelligence information to senior policymakers. As noted above, during the course of its investigation the Commission reviewed a number of articles from the President's Daily Brief (PDB) relating to Iraq's WMD programs. Not surprisingly, many of the flaws in other intelligence products can also be found in the PDBs. But we found some flaws that were inherent in the format of the PDBs--a series of short "articles" often based on current intelligence reporting that are presented to the President each morning. Their brevity leaves little room for doubts or nuance--and their "headlines" designed to grab the reader's attention leave no room at all. Also, a daily drumbeat of reports on the same topic gives an impression of confirming evidence, even when the reports all come from the same source.

The Commission also learned that, on the eve of war, the Intelligence Community failed to convey important information to policymakers. After the October 2002 NIE was published, but before Secretary of State Powell made his address about Iraq's WMD programs to the United Nations, serious doubts became known within the Intelligence Community about Curveball, the aforementioned human intelligence source whose reporting was so critical to the Intelligence Community's pre-war biological warfare assessments. These doubts never found their way to Secretary Powell, who was at that time attempting to strip questionable information from his speech.

These are errors--serious errors. But these errors stem from poor tradecraft and poor management. The Commission found no evidence of political pressure to influence the Intelligence Community's pre-war assessments of Iraq's weapons programs. As we discuss in detail in the body of our report, analysts universally asserted that in no instance did political pressure cause them to skew or alter any of their analytical judgments. We conclude that it was the paucity of intelligence and poor analytical tradecraft, rather than political pressure, that produced the inaccurate pre-war intelligence assessments.

The Iraq Study

This case study proceeds in two parts. The study first details the stream of pre-war intelligence assessments, from the Gulf War to Operation Iraqi Freedom, and compares those to the post-war findings of the Iraq Survey Group. That comparison is provided for each weapons type--nuclear, biological, chemical, and their delivery systems--and also for the political context in Iraq during this time period. For each of these sections, the report also offers the Commission's findings, which often identify specific flaws that led to the inaccuracies in the assessments. The study then identifies the overarching conclusions about the collection, analysis, and dissemination of intelligence that we drew from our examination of the Intelligence Community's performance on the Iraq WMD question.

The pre-war estimate of Iraq's nuclear program, as reflected in the October 2002 NIE Iraq's Continuing Programs for Weapons of Mass Destruction, was that, in the view of most agencies, Baghdad was "reconstituting its nuclear weapons program" and "if left unchecked, [would] probably...have a nuclear weapon during this decade," although it would be unlikely before 2007 to 2009. ³ The NIE explained that, in the view of most agencies, "compelling evidence" of reconstitution was provided by Iraq's "aggressive pursuit of high-strength aluminum tubes." ⁴ The NIE also pointed to additional indicators, such as other dual-use procurement activity, supporting reconstitution. The assessment that Iraq was reconstituting its nuclear program and could therefore have a weapon by the end of the decade was made with "moderate confidence." ⁵

Based on its post-war investigations, the Iraq Survey Group (ISG) concluded--contrary to the Intelligence Community's pre-war assessments--that Iraq had not tried to reconstitute a capability to produce nuclear weapons after 1991. ⁶ Moreover, the ISG judged that Iraq's work on uranium enrichment, including development of gas centrifuges, essentially ended in 1991, and that its ability to reconstitute its enrichment program progressively decayed after that time. ⁷ With respect to the aluminum tubes, the ISG concluded that Iraq's effort to procure the tubes is "best explained by its efforts to produce 81-mm rockets," and the ISG uncovered no evidence that the tubes were intended for use in a gas centrifuge. ⁸

The Community was, in brief, decidedly wrong on what many would view as the single most important judgment it made. The reasons why the Community was so wrong are not particularly glamorous--failures of analysts to question assumptions and apply their tradecraft correctly, errors in technical and factual analysis, a paucity of collection, and failure by the Community to authenticate relevant documents. But these seemingly workaday shortcomings collectively led to a major mis-estimation of a critical intelligence question.

This chapter details our review of the Intelligence Community's performance on the nuclear issue. Like the chapters that follow on the Community's assessments of other aspects of Iraq's weapons programs, this chapter is divided into three sections. First, we review the Intelligence Community's pre-war assessments of Iraq's nuclear program. We then summarize the findings of the ISG regarding Iraq's nuclear efforts and how those findings compare to the Intelligence Community's assessments. The final section contains our findings concerning the causes of the Intelligence Community's failures on the aluminum tubes issue and the now-infamous Niger story.

The Intelligence Community's Pre-War Assessments

The Intelligence Community's assessments of Iraq's pre-war nuclear program were not made in a vacuum. Rather, as the Intelligence Community later explained, its assessments were informed by its analysis of Iraq's nuclear ambitions and capabilities spanning the preceding fifteen years, as well as by "lessons learned from over a decade of dealing with Iraqi intransigence on this issue." ⁹ Thus the proper starting point for an evaluation of the Intelligence Community's assessments lies at the conclusion of the first Gulf War--when the Intelligence Community reviewed the state of Saddam Hussein's nuclear programs and was surprised by what it found.

Post-Gulf War. Following the Gulf War, based on a variety of sources of intelligence including reporting from defectors, the Intelligence Community learned that Iraq's nuclear weapons program went "far beyond what had been assessed by any intelligence organization" in 1990-1991. ¹⁰ Before the Gulf War, in November 1990, the Community had assessed that, because analysts had not detected a formal, coordinated nuclear weapons program, Iraq likely would not have a nuclear weapon until the late 1990s. ¹¹ Thus after the war the Intelligence Community was surprised to discover the breadth of Iraq's nuclear weapons program, including the wide range of technologies Iraq had been pursuing for uranium enrichment, which in turn indicated that Iraq "had been much closer to a weapon than virtually anyone expected." ¹² This humbling discovery that Iraq had successfully concealed a sophisticated nuclear program from the U.S. Intelligence Community exercised a major influence on the Intelligence Community's assessments throughout the early 1990s and afterwards.

Iraq's subsequent and continuing attempts to evade and deceive international inspectors heightened analysts' concerns. ¹³ In a 1994 Joint Atomic Energy Intelligence Committee (JAEIC) assessment, Iraq's Nuclear Weapons Program: Elements of Reconstitution, the Intelligence Community agreed that the "Iraqi government is determined to covertly reconstitute its nuclear weapons program," and that, although Iraq had not yet begun reconstitution, it "would most likely choose the gas centrifuge route" and would "invest a great deal of time and effort" to "conceal its efforts from long-term monitoring." ¹⁴

Mid-1990s. Still, through the mid-1990s, analysts continued to assess that Iraq had not yet reconstituted its nuclear program. Most agencies judged in a 1993 NIE that "if sanctions are lifted and especially if inspections cease, Baghdad will rapidly accelerate its effort" to produce nuclear weapons. ¹⁵ And all agencies agreed in a September 1994 JAEIC assessment that Iraq "still seems to be pursuing" its former program. ¹⁶ The Intelligence Community believed that if Iraq were able to mount a dedicated centrifuge program, it would probably take the Iraqis five to seven years to produce enough fissile material for a nuclear weapon. ¹⁷ This consensus was best reflected by an October 1997 assessment by the JAEIC, which reaffirmed its previous judgments that Iraq would need five to seven years to produce fissile material indigenously, assuming some availability of foreign technical assistance and supplies. ¹⁸ Whether that five to seven year clock had started to run, however, was unclear: this assessment noted that although there was "no firm evidence that reconstitution had begun, six years had passed since the Gulf War and the Community could not be certain whether the starting point for the five to seven year timeline was in the past or future." ¹⁹

During this period, the lack of specific intelligence on the subject continued to complicate analysts' abilities to assess Iraq's ability to reconstitute its nuclear program. The Intelligence Community noted in a 1998 assessment, for instance, that there was limited and often contradictory human intelligence reporting on Iraqi nuclear efforts, with some human intelligence sources indicating that Iraq was continuing "low-level theoretical research for a weapons program" while other sources reported that "all nuclear-related activity [had been] halted." ²⁰ The Intelligence Community acknowledged that it had an "incomplete picture of the Iraqi nuclear program." ²¹

Post-1998. The end of international inspections in 1998, prompted by Saddam Hussein's preventing the inspectors from doing their work, increased concern among analysts that Iraq would use that opportunity to reconstitute its nuclear program. Accordingly, in 1999, the JAEIC noted that although it still had no specific evidence that reconstitution had begun, the absence of inspectors gave Iraq greater opportunity to conduct covert research and development. ²² As of December 2000, however, an Intelligence Community Assessment noted that Iraq still did not appear to have taken major steps toward reconstitution. ²³ Thus, after the departure of inspectors, the Intelligence Community assumed that Iraq had the opportunity and the desire to jumpstart its covert nuclear weapons program; by the end of 2000, however, the Community had seen no firm evidence that this was actually happening.

This judgment began to shift in early 2001 as a result of a discovery that, in hindsight, was the critical moment in the development of the Intelligence Community's assessment of Iraq's nuclear program. In March 2001, intelligence reporting indicated that Iraq was seeking high-strength tubes made of 7075 T6 aluminum alloy. ²⁴ The Intelligence Community obtained samples of the tubes when a shipment bound for Iraq was seized overseas. ²⁵

At this point, a debate began within the Intelligence Community about the reason why Iraq had procured the tubes. The CIA assessed that the tubes were most likely for gas centrifuges for enriching uranium and believed that the tubes provided compelling evidence that Iraq had renewed its gas centrifuge uranium enrichment program. ²⁶ CIA subsequently identified possible non-nuclear applications for the tubes, ²⁷ but continued to judge that the tubes were destined for use in Iraqi gas centrifuges ²⁸ --even while acknowledging that the Intelligence Community had very little information on Iraq's WMD programs to corroborate this assessment. ²⁹

This judgment concerning the tubes' likely intended use was echoed by another expert technical entity within the Intelligence Community. Analysts from the National Ground Intelligence Center (NGIC), a component of the U.S. Army recognized as the national experts on conventional military systems, judged that while it could "not totally rule out the possibility" that the tubes could be used for rockets and thus were not destined for a nuclear-related use, the tubes were, technically speaking, poor choices for rocket bodies. NGIC's expert judgment was therefore that there was a very low probability the tubes were designed for conventional use in rockets. ³⁰ Because of NGIC's expertise on conventional weapons systems such as rockets, NGIC's view that the tubes were poor choices for rocket bodies gave CIA analysts greater confidence in their own judgment that the tubes were likely for use in centrifuges. ³¹

Other entities took a different view, however. The Department of Energy (DOE), the U.S. government's primary repository of expertise on nuclear matters, assessed that the tubes--although they "could be used to manufacture centrifuge rotors"--were "not well-suited for a centrifuge application" and were more likely intended for use in Iraq's Nasser 81 millimeter Multiple Rocket Launcher (MRL) program. ³² The International Atomic Energy Agency (IAEA) agreed with DOE's assessment, concluding that the tubes were usable in a gas centrifuge application but that they were not directly suited to that use. ³³

Despite this disagreement, the CIA informed senior policymakers that it believed the tubes were destined for use in Iraqi gas centrifuges. ³⁴ While noting that there was disagreement within the Intelligence Community concerning the most likely use for the tubes, the CIA pointed out that there was also interagency consensus that the tubes could be used for centrifuge enrichment. ³⁵ This consensus on capability led many analysts at both CIA and DIA to think that the tubes supplied the evidence that Iraq was starting to "reconstitute" its nuclear program. ³⁶

Other streams of evidence also raised flags. At about the same time, analysts began to see indications that Iraq was seeking procurement of other dual-use items that would be consistent with a possible renewed effort at developing centrifuges. ³⁷ This activity concerned even DOE, which had expressed skepticism that the intercepted tubes had centrifuge applications. ³⁸ These concerns were affected by the Intelligence Community's history of underestimating Iraq's nuclear program; as the National Intelligence Council (NIC) would later observe, analysts became concerned during 2002 that "they may again be facing a surprise similar to the one in 1991." ³⁹

In the months before the October 2002 NIE, the CIA continued to assess that the tubes were intended for use in gas centrifuges, albeit with slight variations in the strength of that formulation, pointing out that Iraq's interest in the tubes was "key" to the assessment that Iraq was "reconstituting its centrifuge program." ⁴⁰ CIA presented this view in an Intelligence Assessment, entitled Iraq's Hunt for Aluminum Tubes: Evidence of a Renewed Uranium Enrichment Program, in which CIA concluded that the aluminum tubes "are most likely for gas centrifuges for enriching uranium" and that Iraq's pursuit of such tubes provided "compelling evidence that Iraq has renewed its gas centrifuge uranium enrichment program." ⁴¹ The assessment noted that "some" in the Intelligence Community believed conventional armament applications, such as multiple rocket launchers, were "more likely end-uses," but the assessment noted that NGIC, the "national experts on conventional military systems," had found such uses "highly unlikely." ⁴² At the same time, DOE disseminated a separate assessment arguing that, while the tubes could be modified for use as centrifuge rotors, "other conventional military uses [we]re more plausible." ⁴³ The Department of State's Bureau of Intelligence and Research (INR) agreed with DOE's assessment. ⁴⁴

October 2002 NIE. The Intelligence Community judged in the NIE with moderate confidence that "Baghdad ha[d] reconstituted its nuclear weapons program." ⁴⁵ Only INR dissented from this assessment, although INR judged in the President's Summary of the NIE that the overall evidence "indicates, at most, a limited Iraqi nuclear reconstitution effort." ⁴⁶ By reconstitution, the Intelligence Community meant that Iraq was in the "process of restoring [its] uranium enrichment capability." ⁴⁷ To the relevant CIA and DIA analysts, the pursuit of aluminum tubes provided "compelling evidence" of reconstitution. ⁴⁸ In particular, the composition, dimensions, cost, and tight manufacturing tolerances for the tubes were assessed by CIA and DIA to exceed by far those needed for non-nuclear purposes, thus demonstrating that the tubes were intended for a nuclear-related use. ⁴⁹ At the interagency coordination meeting for the NIE, both NSA and the National Geospatial-Intelligence Agency (NGA) agreed with the CIA/DIA position on the tubes. ⁵⁰ DOE and INR dissented from the tubes judgment, assessing that the tubes were more likely for use in tactical rockets. ⁵¹

The NIE stated that the conclusion that the tubes indicated reconstitution was bolstered by additional evidence that suggested Iraq could be rebuilding its nuclear program:

1. Other Dual-Use Procurements. Reporting indicated that Iraq was attempting to procure other dual-use items that would be required to build a gas centrifuge plant, such as magnets, "high-speed balancing machines," and machine tools. ⁵² These items are all dual-use materials, however, and the reporting provided no direct indication that the materials were intended for use in a nuclear program, as indicated in the NIE. ⁵³

2. Nuclear Cadre. The NIE also pointed to evidence that Iraq was making efforts to preserve, and in some cases re-establish and enhance, its cadre of weapons personnel. ⁵⁴ Reporting indicated that some scientists had been reassigned to the Iraqi Atomic Energy Commission (IAEC) and that Iraq had "reassembled" many scientists, engineers, and managers from Iraq's previous nuclear program. ⁵⁵

3. Activity at Suspect Sites. Sources indicated that Iraq was trying to procure a magnet production line in 1999-2001 and one report indicated the plant would be located at Al-Tahadi, where analysis suggested construction of buildings in late 2000 that could have housed a magnet production line. ⁵⁶ Both sources indicated, however, that magnet procurements were likely affiliated with Iraq's missile program, rather than with nuclear applications, though some reporting noted that the cadre of scientists and technicians at the site formerly worked in the nuclear program. ⁵⁷

Uranium from Niger. Although the NIE did not include uranium acquisition in the list of elements bolstering its conclusion about reconstitution, it did note that Iraq was "vigorously trying to procure uranium ore and yellowcake" from Africa. ⁵⁸ This statement was based largely on reporting from a foreign government intelligence service that Niger planned to send up to 500 tons of yellowcake uranium to Iraq. ⁵⁹ The status of the arrangement was unclear, however, at the time of the coordination of the Estimate and the NIE therefore noted that the Intelligence Community could not confirm whether Iraq succeeded in acquiring the uranium. ⁶⁰ Iraq's alleged pursuit of uranium from Africa was thus not included among the NIE's Key Judgments. ⁶¹ For reasons discussed at length below, several months after the NIE, the reporting that Iraq was seeking uranium from Niger was judged to be based on forged documents and was recalled. ⁶²

In short, all of the coordinating agencies, with the exception of INR, agreed that Iraq was reconstituting its nuclear program. ⁶³ Of those agencies that agreed on reconstitution, all but DOE agreed that the tubes provided "compelling evidence" for that conclusion. DOE reaffirmed its previous assessments that, while the tubes could be modified for use in a gas centrifuge, they were poorly suited for such a function and were most likely designed for use in conventional rockets. ⁶⁴ On the question of reconstitution, DOE believed that the other factors--the attempted procurement of magnets and balancing machines, efforts to reconstitute the nuclear cadre, activity at suspect sites, and evidence of Iraqi efforts to obtain uranium from Africa--justified the conclusion that Iraq was reconstituting its nuclear program. ⁶⁵ None of the other agencies placed significant weight on reporting about attempts to procure uranium from Africa to support their conclusion of reconstitution. ⁶⁶

Post-NIE. The publication of the NIE did not settle the dispute about the aluminum tubes and so, in the period between the NIE and the invasion of Iraq, debate within the Intelligence Community over their significance continued. INR, for its part, continued to see "no compelling reason to judge that Iraq ha[d] entered" the timeframe of at least five to seven years that the Intelligence Community agreed Baghdad would need to produce sufficient fissile material for a nuclear weapon. ⁶⁷ DOE, meanwhile, continued to believe that reconstitution was underway but that the "tubes probably were not part of the program," ⁶⁸ assessing instead that the tubes were intended for use in conventional rockets. ⁶⁹ On the other side of the dispute, NGIC and CIA continued to assess that the tubes were destined for use in gas centrifuges. ⁷⁰ Outside the Intelligence Community, the IAEA, after inspections resumed in fall 2002, also weighed in on the dispute, concluding with DOE and INR that the tubes were likely intended for use in Iraq's 81 millimeter rocket program. ⁷¹

During this time the CIA continued to explain to senior policymakers that the Intelligence Community was not of one view on the most likely use for the tubes, ⁷² but CIA offered its own view that the "alternative explanation" for the tubes' intended use--that they would be used for rockets--was likely an Iraqi "cover story." ⁷³ The CIA also noted the overall paucity of information on Iraq's programs, but suggested that the lack of information was due in part to Iraq's successful efforts to hide its illicit activity. ⁷⁴

Other countries' intelligence agencies views of the tubes were, on balance, somewhat more circumspect than that of the majority in the NIE. For its part, the British Joint Intelligence Committee assessed, as did the NIE, that the aluminum tubes, with some modifications, would be suitable for use in a centrifuge, but noted that there was no definitive intelligence that the tubes were destined for the nuclear program. ⁷⁵ The views of the Australian Office of National Assessments on the relevance of the tubes to Iraq's nuclear program were "inconsistent and changeable." ⁷⁶

Post-War Findings of the Iraq Survey Group

The Iraq Survey Group concluded that Iraq had not tried to reconstitute a capability to produce nuclear weapons after 1991. ⁷⁷ It concluded that Iraq's efforts to develop gas centrifuges for uranium enrichment ended in 1991, as did Iraq's work on other uranium enrichment programs, which Iraq had explored prior to the Gulf War. ⁷⁸ The ISG also found no evidence that Iraq had taken steps to advance its pre-1991 work in nuclear weapons design and development. ⁷⁹ Although the ISG did find indications that Saddam remained interested in reconstitution of the nuclear program after sanctions were lifted, it concluded that Iraq's ability to reconstitute its program progressively decayed after 1991. ⁸⁰

Not long after the start of the Iran-Iraq war in 1980, Iraq started to pursue formally a uranium enrichment program using a variety of uranium enrichment techniques. ⁸¹ By 1990, Iraq had built two magnetic-bearing centrifuges (with foreign assistance) using imported carbon fiber rotors and two oil-bearing centrifuges. ⁸² During the first Gulf War, however, nearly all of the key nuclear facilities in Iraq--those involved in the processing of nuclear material or weapons research--were bombed and many of the facilities were largely destroyed. ⁸³

After the Gulf War, Iraq initially chose not to disclose the extent of its nuclear program and instead sought to hide any evidence of it. Accordingly, the director of Iraq's Military Industrialization Commission, Hussein Kamil, ordered the collection of all inculpatory documents and equipment. The equipment and documentation were then moved to a variety of locations to hide them from the IAEA. Hussein Kamil ordered at least one set of all nuclear-related documents and some equipment to be retained by a senior scientist. ⁸⁴

Despite Iraqi efforts, in early summer 1991 the IAEA confronted Baghdad with evidence of uranium enrichment components during the course of its inspections. At that point Baghdad admitted to its large pre-war enrichment programs, but still did not fully declare the extent of its centrifuge program. ⁸⁵

Indeed, Iraq continued to resist more comprehensive disclosure of its pre-1991 nuclear program until after the defection of Hussein Kamil in 1995, when a large number of documents and equipment fell into the hands of UNSCOM and the IAEA. From this point forward, according to the ISG, the Iraqis appear to have been more cooperative and provided more complete information. For example, the Iraqis largely declared their pre-1991 centrifuge program, although a full set of documents obtained by Iraq from German engineers in the 1980s was not supplied to IAEA inspectors. ⁸⁶

Although the Iraqis did not make more comprehensive disclosures about their nuclear program until 1995, the Iraq Survey Group concluded that Iraq had actually ended its nuclear program in 1991. More specifically, the ISG assessed that Iraq's development of gas centrifuges essentially ended in 1991 and that Iraq did not continue work on any of the other pre-1991 enrichment methods it had explored, including electromagnetic isotope separation (EMIS). ⁸⁷ The ISG did point out, however, that many of the former EMIS engineers and scientists continued to work for either the Iraqi Atomic Energy Commission or the Military Industrialization Commission in roles that could preserve their technical skills. ⁸⁸

Despite these efforts to preserve the skills and talent of the nuclear cadre, the intellectual capital underlying Iraq's nuclear program decayed in the years after 1991. ⁸⁹ For example, starting around 1992, the Director of Iraq's Military Industrialization Commission transferred personnel from the former nuclear program to various military research and production facilities. Some of the work performed by these former nuclear scientists by its nature preserved for Iraq capabilities that would be needed for a reconstituted nuclear program. Still, the ISG noted that the overall decline of the Iraqi economy made it very difficult to retain scientists, many of whom departed for better prospects abroad. ⁹⁰

With the influx of funds from the Oil-for-Food program and later the suspension of cooperation with UNSCOM, Saddam began to pay renewed attention to former members of the Iraq nuclear program. In the late 1990s, for instance, he raised salaries for those in the Military Industrialization and Iraqi Atomic Energy Commissions, and new programs, such as joint programs with universities, were initiated to employ the talent of former nuclear program employees. ⁹¹ In the year before Operation Iraqi Freedom, Iraq's Military Industrialization Commission also took steps to improve capabilities that could have been applied to a renewed centrifuge program for uranium enrichment. But the ISG did not uncover information indicating that the technologies being pursued were intended to support such a program. ⁹²

With respect to Iraq's interest in procuring high-strength aluminum tubes, the ISG concluded that the Iraqi attempt to procure the tubes is best explained by Iraq's efforts to produce effective 81 millimeter rockets; the ISG uncovered no evidence that the tubes were intended for use in a gas centrifuge. ⁹³ The ISG arrived at this conclusion only after investigating the key indicators that suggested a possible centrifuge end-use for the tubes--for example, the tubes' dimensions and tight manufacturing tolerances--and found no evidence of a program to design or develop an 81 millimeter aluminum rotor centrifuge. ⁹⁴

What the ISG found instead was that, with respect to the dimensions of the tubes, Iraqi nuclear scientists thought it was at best impractical for Iraq to have made a centrifuge with 81 millimeter rotors. For example, Ja'far Diya Ja'far, the head of Iraq's pre-1991 uranium enrichment program, stated in post-war debriefings that, while it was possible to make a rotor from the tubes, he thought it would be impractical to do so. ⁹⁵ He also said that using 81 millimeter rockets as a "cover story" for a centrifuge project would not have been very useful, because Iraq had difficulty importing any goods. ⁹⁶ Ja'far similarly did not consider it reasonable that Iraq could have pursued a centrifuge program based on 81 millimeter aluminum tubes, judging the technical challenges to doing so were too great. ⁹⁷

Conversely, the Iraq Survey Group investigation did uncover what it judged to be plausible accounts that linked the tubes to 81 millimeter rockets, and which answered questions about why the Iraqis had sought such tight manufacturing specifications for the tubes. For example, some sources indicated to the ISG that the tight tolerance requests were driven by a desire to improve the accuracy of the rockets. Inconsistencies among rockets had resulted in past variations in range and accuracy, according to these sources, and the Iraqis chose to address this problem by tightening specifications. ⁹⁸ Another explanation was that the engineering drawings for the Iraqi 81 millimeter rocket, which was originally reverse-engineered from an Italian air-to-ground rocket (the Medusa), had undergone many ad hoc revisions over the years because the Iraqis were using their 81 millimeter rockets as ground-to-ground rockets. An Iraqi military committee was convened to return the design to the original Italian-based design, according to the ISG report, and that military committee then set new, and more strict, specifications. ⁹⁹ The ISG also learned that misfires sometimes resulted from pitting in the tubes caused by improper storage and corrosion, a problem that could explain the requirement that the tubes be anodized and shipped carefully. ¹⁰⁰

Though ultimately concluding that the evidence did not show that the Iraqis intended a nuclear end-use for the tubes, the Iraq Survey Group did note some inconsistencies in the explanation that the tubes were intended for use in tactical rockets. ¹⁰¹ For example, the ISG found technical drawings that showed that Iraq's 81 millimeter rocket program had a history of using tubes that fell short of the strict manufacturing standards demanded in the procurement attempts before the war. ¹⁰² Also, the ISG found evidence that, in the months just before the war, the Iraqis accepted lower-quality, indigenously produced aluminum tubes for its 81 millimeter rockets, despite the continuing efforts to procure high-specification tubes from abroad. ¹⁰³ Iraq also explored the possibility (about a year before the war) of using steel for the rocket bodies. This approach was rejected, however, because it would have required significant design modifications for the existing 81 millimeter rocket design. ¹⁰⁴ The ISG noted that these efforts raise questions about whether high-specification tubes were really needed for rockets. ¹⁰⁵

The ISG reconciled this evidence by judging that Iraq's continued efforts to obtain tubes from abroad, even while simultaneously accepting some indigenously produced tubes for use in rockets, could be explained in large measure by bureaucratic inefficiencies and fear of senior officials in the ranks of the Iraqi government. ¹⁰⁶ For example, Dr. Huwaysh, the head of the Military Industrialization Commission, "exhibited a rigid managerial style" and frequently made unreasonable production demands. The fear of being held responsible for rejected tubes or components affected the lead production engineer and he therefore decided to tighten specifications for the rocket program. Similarly, a report from the rocket program noted that some engineers requested tight specifications in order to appear effective in addressing problems. Also, because Huwaysh demanded results quickly, the engineers did not have time to attempt a detailed analysis of the causes for rocket scatter and inaccuracy; instead, the engineers simply tightened some specifications in the hope that that would improve accuracy. ¹⁰⁷ Other factors influencing the continuing efforts to procure tubes from abroad included the "lack of sufficient indigenous manufacturing capabilities"--an effort that Iraq only began in 2002--the high costs of production, and the "pressure of the impending war." ¹⁰⁸

The ISG noted that one other factor that the Intelligence Community had cited as evidence that the tubes were intended for use in a centrifuge was that the potential supplier was asked to provide 84 millimeter tubes--a change that would have meant the tubes could not be used in an 81 millimeter rocket. ¹⁰⁹ But the ISG found no clear indication that it was Iraq (or an Iraqi entity) that was making these inquiries about 84 millimeter tubes. ¹¹⁰ In any event, the ISG concluded that, although a larger diameter tube would be better for use in a centrifuge, Iraq already had 500 tons of 120 millimeter diameter aluminum shafts which it had imported before sanctions were imposed in 1990. And, furthermore, Iraq was using those shafts in the months before Operation Iraqi Freedom to support the flow-forming operations related to the 81 millimeter rocket program. ¹¹¹

With respect to alleged "high-level interest" in tubes by Iraqi leaders, the ISG concluded that such interest in the tubes appears to have focused on efforts to produce 81 millimeter rockets rather than on any element of a nuclear program. ¹¹²

The Iraq Survey Group also found no evidence that Iraq sought uranium from abroad after 1991. ¹¹³ With respect to the reports that Iraq sought uranium from Niger, ISG interviews with Ja'far Diya Ja'far, the head of Iraq's pre-1991 enrichment programs, indicated that Iraq had only two contacts with the Nigerien government after 1998--neither of which was related to uranium. ¹¹⁴ One such contact was a visit to Niger by the Iraqi Ambassador to the Vatican Wissam Zahawie, the purpose of which Ja'far said was to invite the Nigerien President to visit Iraq (a story told publicly by Zahawie). ¹¹⁵ The second contact was a visit to Iraq by a Nigerien minister to discuss Nigerien purchases of oil from Iraq--with no mention of "any kind of payment, quid pro quo, or offer to provide Iraq with uranium ore, other than cash in exchange for petroleum." ¹¹⁶ The use of the last method of payment is supported by a crude oil contract, dated June 26, 2001, recovered by the ISG. ¹¹⁷

The ISG found only one offer of uranium to Baghdad since 1991--an offer that Iraq appears to have turned down. ¹¹⁸ The ISG found a document in the headquarters of the Iraqi Intelligence Service that reveals that a Ugandan businessman had approached the Iraqi Embassy in Nairobi with an offer to sell uranium, reportedly from the Congo. The Iraqi Embassy in Nairobi, reporting back to Baghdad on the matter on May 20, 2001, indicated that the Embassy told the Ugandan that Iraq did not deal with "these materials" because of the sanctions. ¹¹⁹

Finally, and on a broader plane, even if an order to reconstitute had been given, Iraq Survey Group interviews with former senior officials indicated that Iraq would not have been able to do so given the conditions inside the country in 2002. ¹²⁰ Unsurprisingly, therefore, the ISG found no indication that Iraq had resumed fissile material or nuclear weapon research and development activities after 1991. ¹²¹

Analysis of the Intelligence Community's Pre-War Assessments

This marked disjuncture between the Intelligence Community's assessments and the findings of the Iraq Survey Group about Iraq's purported nuclear weapons program was not solely the product of bad luck or the inherent difficulties of making intelligence judgments. It arose out of fundamental flaws in the way the Intelligence Community approached its business.

Above all, the Intelligence Community's failure on the nuclear issue was a failure of analysis. To be sure, the paucity of intelligence contributed to that failure. Although signals intelligence played a key role in some respects that we cannot discuss in an unclassified format, on the whole it was not useful. Similarly, though imagery intelligence showed some construction at a possible suspect nuclear site in or around 2000, imagery provided little helpful insight into the purpose of that activity and nothing beyond that. And, other than information on the alleged uranium deal that was later determined to be unreliable, very little human intelligence was available to provide insight into Iraq's intentions. The time pressures of the October 2002 NIE also may have hampered the normal thorough review before dissemination. ¹²²

But on the crucial question of whether the aluminum tubes were for use in a gas centrifuge or in tactical rockets--an analytical question--the Intelligence Community got it wrong. ¹²³ And, notably, it was not one of the difficult and inherently speculative questions intelligence analysts often confront; it was not a question that required the Intelligence Community to make a prediction about future events or to draw conclusions about the state of the world based upon limited information. Rather, the critical question was, at bottom, largely a technical one, where the critical facts were known or knowable: namely, how well-suited were the aluminum tubes for tactical rockets and centrifuges, respectively? An even-handed assessment of the evidence should have led the Intelligence Community to conclude that the tubes were more likely destined for tactical rockets. This section examines this analytic failure and other issues uncovered by our review of the Intelligence Community's performance.

The Intelligence Community's Pre-War Assessments

The Intelligence Community's assessment of Iraq's BW program--like its judgments about Iraq's other WMD programs--evolved over time. The October 2002 NIE reflected a shift, however, in the Community's judgments about the state of Iraq's BW program. Previous Community estimates had assessed that Iraq could have biological weapons; the October 2002 estimate, in contrast, assessed with "high confidence" that Iraq "has" biological weapons. This shift in view, which began in 2000 and culminated in the October 2002 NIE, was based largely on information from a single source--Curveball--who indicated that Iraq had mobile facilities for producing BW agents.

Background. In the early 1990s, the Intelligence Community knew little about Iraq's BW program. ²²⁴ Prior to the Gulf War, the Intelligence Community judged that Iraq was developing several BW agents, including anthrax and botulinum toxin, at a number of facilities. ²²⁵ The Intelligence Community further assessed that Iraq might have produced up to 1,000 liters of BW agent, and that Iraq had used some of it to fill aerial bombs and artillery shells. At that time, however, the Community judged that it had insufficient information to make assessments about BW agent testing and deployment of filled munitions. ²²⁶ Between 1991 and 1995, the Intelligence Community learned little more about Iraq's BW program. However, there was some additional human intelligence reporting indicating that pre-Gulf War assessments of Iraq's BW program had substantially underestimated the quantities of biological weapons that Iraq had produced. Moreover, this reporting suggested that the Intelligence Community was unaware of some Iraqi BW facilities. ²²⁷

It was not until 1995--when UNSCOM presented the Iraqis with evidence of continuing BW-related imports and Saddam Hussein's son-in-law, Hussein Kamil, defected--that Iraq made substantial declarations to the United Nations about its activities prior to the Gulf War, admitting that it had produced and weaponized BW agents. ²²⁸ These declarations confirmed that the Intelligence Community had substantially underestimated the scale and maturity of Iraq's pre-Desert Storm BW program. Iraq had, before the Gulf War, weaponized several agents, including anthrax, botulinum toxin, and aflatoxin; produced 30,165 liters of BW agent; and deployed some of its 157 bombs and 25 missile warheads armed with BW agents to locations throughout Iraq. ²²⁹ Following these declarations, the Intelligence Community estimated in 1997 that Iraq was still concealing elements of its BW program, and it assessed that Iraq would likely wait until either sanctions were lifted or the UNSCOM presence was reduced before restarting agent production. ²³⁰

After 1998, the Intelligence Community found it difficult to determine whether activity at known dual-use facilities was related to WMD production. The departed inspectors had never been able to confirm what might be happening at Iraq's suspect facilities. Accordingly, the Intelligence Community noted that it had no reliable intelligence to indicate resumed production of biological weapons, but assessed that in the absence of inspectors Iraq probably would expand its BW activities. ²³¹ These assessments were colored by the Community's earlier underestimation of Iraq's programs, its lack of reliable intelligence, and its realization that previous underestimates were due in part to effective deception by the Iraqis. ²³² By 1999, the CIA assessed that there was some Iraqi research and development on BW and that Iraq could restart production of biological weapons within a short period of time. The 1999 NIE on Worldwide BW Programs judged that Iraq was "revitalizing its BW program" and was "probably continuing work to develop and produce BW agents." ²³³

Growing concern. The Intelligence Community's concern about Iraq's BW program increased in early 2000, and the Community began to adjust upward its estimates of the Iraq BW threat, based on a "substantial volume" of "new information" regarding mobile BW facilities in Iraq. ²³⁴ This information came from an Iraqi chemical engineer, subsequently codenamed Curveball, who came to the attention of the Intelligence Community through a foreign liaison service. That liaison service debriefed Curveball and then shared the debriefing results with the United States. The foreign liaison service would not, however, provide the United States with direct access to Curveball. Instead, information about Curveball was passed from the liaison service to DIA's Defense HUMINT Service, which in turn disseminated information about Curveball throughout the Intelligence Community.

Between January 2000 and September 2001, DIA's Defense HUMINT Service disseminated almost 100 reports from Curveball regarding mobile BW facilities in Iraq. ²³⁵ These reports claimed that Iraq had several mobile production units and that one of those units had begun production of BW agents as early as 1997. ²³⁶

Shortly after Curveball started reporting, in the spring of 2000, his information was provided to senior policymakers. ²³⁷ It was also incorporated into an update to a 1999 NIE on Worldwide BW Programs. The update reported that "new intelligence acquired in 2000...causes [the IC] to adjust our assessment upward of the BW threat posed by Iraq...The new information suggests that Baghdad has expanded its offensive BW program by establishing a large-scale, redundant, and concealed BW agent production capability." ²³⁸ In December 2000, the Intelligence Community produced a Special Intelligence Report that was based on reporting from Curveball, noting that "credible reporting from a single source suggests" that Iraq has produced biological agents, but cautioned that "[w]e cannot confirm whether Iraq has produced...biological agents." ²³⁹

By 2001, however, the assessments became more assertive. A WINPAC report in October 2001, also based on Curveball's reporting about mobile facilities, judged "that Iraq continues to produce at least...three BW agents" and possibly two others. This assessment also concluded that "the establishment of mobile BW agent production plants and continued delivery system development provide Baghdad with BW capabilities surpassing the pre-Gulf War era." ²⁴⁰ Similar assessments were provided to senior policymakers. ²⁴¹ In late September 2002, DCI Tenet told the Senate's Intelligence and Armed Services Committees (and subsequently the Senate Foreign Relations Committee) that "we know Iraq has developed a redundant capability to produce biological warfare agents using mobile production units." ²⁴²

October 2002 NIE. The October 2002 NIE reflected this upward assessment of the Iraqi BW threat that had developed since Curveball began reporting in January 2000. The October 2002 NIE reflected the shift from the late-1990s assessments that Iraq could have biological weapons to the definitive conclusion that Iraq "has" biological weapons, and that its BW program was larger and more advanced than before the Gulf War. ²⁴³ Information about Iraq's dual-use facilities and its failure to account fully for previously declared stockpiles contributed to this shift in assessments. ²⁴⁴ The information that Iraq had mobile BW production units, however, was instrumental in adjusting upward the assessment of Iraq's BW threat. ²⁴⁵ And for this conclusion, the NIE relied primarily on reporting from Curveball, who, as noted, provided a large volume of reporting through Defense HUMINT channels regarding mobile BW production facilities in Iraq. ²⁴⁶ Only in May 2004, more than a year after the commencement of Operation Iraqi Freedom, did CIA formally deem Curveball's reporting fabricated and recall it. ²⁴⁷ At the time of the NIE, however, reporting from three other human sources--who provided one report each on mobile BW facilities--was thought to have corroborated Curveball's information about the mobile facilities. ²⁴⁸ These three sources also proved problematic, however, as discussed below.

Another asylum seeker (hereinafter "the second source") reporting through Defense HUMINT channels provided one report in June 2001 that Iraq had transportable facilities for the production of BW. ²⁴⁹ This second source recanted in October 2003, however, and the recantation was reflected in a Defense HUMINT report in which the source flatly contradicted his June 2001 statements about transportable facilities. ²⁵⁰ Though CIA analysts told Commission staff that they had requested that Defense HUMINT follow-up with this second source to ascertain the reasons for his recantation, DIA's Defense HUMINT Service has provided no further information on this issue. ²⁵¹ Nor, for that matter, was the report ever recalled or corrected. ²⁵²

Another source, associated with the Iraqi National Congress (INC) (hereinafter "the INC source"), was brought to the attention of DIA by Washington-based representatives of the INC. Like Curveball, his reporting was handled by Defense HUMINT. He provided one report that Iraq had decided in 1996 to establish mobile laboratories for BW agents to evade inspectors. ²⁵³ Shortly after Defense HUMINT's initial debriefing of the INC source in February 2002, however, a foreign liaison service and the CIA's Directorate of Operations (DO) judged him to be a fabricator and recommended that Defense HUMINT issue a notice to that effect, which Defense HUMINT did in May 2002. Senior policymakers were informed that the INC source and his reporting were unreliable. The INC source's information, however, began to be used again in finished intelligence in July 2002, including the October 2002 NIE, because, although a fabrication notice had been issued several months earlier, Defense HUMINT had failed to recall the reporting. ²⁵⁴

The classified report here discusses a fourth source (hereinafter "the fourth source") who provided a single report that Iraq had mobile fermentation units mounted on trucks and railway cars.

Post-NIE. After publication of the NIE in October 2002, the Intelligence Community continued to assert that Baghdad's biological weapons program was active and posed a threat, relying on the same set of sources upon which the NIE's judgments were based. ²⁵⁵ For example, a November 2002 paper produced by CIA's Directorate of Intelligence (DI) reiterated the NIE's assessment that Iraq had a "broad range of lethal and incapacitating agents" and that the "BW program is more robust than it was prior to the Gulf War." ²⁵⁶ The piece contended that Iraq was capable of producing an array of agents and probably retained strains of the smallpox virus. It further argued that technological advances increased the potential Iraqi BW threat to U.S. interests. And a February 2003 CIA Intelligence Assessment anticipated Iraqi options for BW (and CW) use against the United States and other members of the Coalition; the report stated that Iraq "maintains a wide range of...biological agents and delivery systems" and enumerated 21 BW agents which it judged Iraq could employ. ²⁵⁷

Statements about biological weapons also appeared in Administration statements about Iraq in the months preceding the war. Secretary of State Colin Powell's speech to the United Nations Security Council on February 5, 2003, relied on the same human sources relied upon in the NIE. ²⁵⁸ Secretary Powell was not informed that one of these sources-- the INC source --had been judged a fabricator almost a year earlier. And as will be discussed at length below, serious doubts about Curveball had also surfaced within CIA's Directorate of Operations at the time of the speech--but these doubts also were not communicated to Secretary Powell before his United Nations address.

Reliance on Curveball's reporting also affected post-war assessments of Iraq's BW program. A May 2003 CIA Intelligence Assessment pointed to the post-invasion discovery of "two probable mobile BW agent productions plants" by Coalition forces in Iraq as evidence that "Iraq was hiding a biological warfare program." ²⁵⁹ Curveball, when shown photos of the trailers, identified components that he said were similar to those on the mobile BW production facilities that he had described in his earlier reporting. ²⁶⁰

Post-War Findings of the Iraq Survey Group

The Iraq Survey Group found that the Intelligence Community's pre-war assessments about Iraq's BW program were almost entirely wrong. The ISG concluded that "Iraq appears to have destroyed its undeclared stocks of BW weapons and probably destroyed remaining holdings of bulk BW agent" shortly after the Gulf War. ²⁶¹ According to the ISG, Iraq initially intended to retain elements of its biological weapons program after the Gulf War. UNSCOM inspections proved unexpectedly intrusive, however, and to avoid detection, Saddam Hussein ordered his son-in-law and Minister of the Military Industrial Commission Hussein Kamil to destroy, unilaterally, Iraq's stocks of BW agents. ²⁶² This took place in either the late spring or summer of 1991. ²⁶³ But Iraq retained a physical plant at Al-Hakam and the intellectual capital necessary to resuscitate the BW program. ²⁶⁴ Simultaneously, Iraq embarked on an effort to hide this remaining infrastructure and to conceal its pre-war BW-related activities. ²⁶⁵

In early 1995, however, UNSCOM inspectors confronted Iraqi officials with evidence of 1988 imports of bacterial growth media in quantities that had no civilian use within Iraq's limited biotechnology industry. ²⁶⁶ This confrontation, followed by the defection of Hussein Kamil in August 1995, prompted Iraq to admit that it had produced large quantities of bulk BW agent before the Gulf War. ²⁶⁷ Iraq also released a large cache of documents and issued the first of several "Full, Final and Complete Declaration[s]" on June 22, 1996, further detailing its BW program. UNSCOM subsequently supervised the destruction of BW-related facilities at Al-Hakam in 1996. ²⁶⁸

The Iraq Survey Group found that the destruction of the Al-Hakam facility effectively marked the end of Iraq's large-scale BW ambitions. ²⁶⁹ The ISG did judge that after 1996 Iraq "continued small-scale BW-related efforts" under the auspices of the Iraqi Intelligence Service, and also retained a trained cadre of scientists who could work on BW programs and some dual-use facilities capable of conversion to small-scale BW agent production. ²⁷⁰ Nevertheless, the ISG "found no direct evidence that Iraq, after 1996, had plans for a new BW program or was conducting BW-specific work for military purposes." ²⁷¹

With respect to mobile BW production facilities, the "ISG found no evidence that Iraq possessed or was developing production systems on road vehicles or railway wagons." ²⁷² The ISG's "exhaustive investigation" of the two trailers captured by Coalition forces in spring 2003 revealed that the trailers were "almost certainly designed and built exclusively for the generation of hydrogen." The ISG judged that the trailers "cannot ... be part of any BW program." ²⁷³

Analysis of the Intelligence Community's Pre-War Assessments

The Intelligence Community fundamentally misjudged the status of Iraq's BW programs. As the above discussion demonstrates, the central basis for the Intelligence Community's pre-war assessments about Iraq's BW program was the reporting of a single human source, Curveball. This single source, whose reporting came into question in late 2002, later proved to be a fabricator.

Our intelligence agencies get burned by human sources sometimes--it is a fact of life in the murky world of espionage. If our investigation revealed merely that our Intelligence Community had a source who later turned out to be lying, despite the best tradecraft practices designed to ferret out such liars, that would be one thing. But Curveball's reporting became a central part of the Intelligence Community's pre-war assessments through a serious breakdown in several aspects of the intelligence process. The Curveball story is at the same time one of poor asset validation by our human collection agencies; of a tendency of analysts to believe that which fits their theories; of inadequate communication between the Intelligence Community and the policymakers it serves; and, ultimately, of poor leadership and management. This section thus focuses primarily on our investigation of the Curveball episode, and the findings we drew from it.

The problems with the Intelligence Community's performance on Curveball began almost immediately after the source first became known to the U.S. government in early 2000. As noted above, Curveball was not a source who worked directly with the United States; rather, the Intelligence Community obtained information about Curveball through a foreign service. The foreign service would not provide the United States with direct access to Curveball, claiming that Curveball would refuse to speak to Americans. ²⁷⁴ Instead, the foreign intelligence service debriefed Curveball and passed the debriefing information to DIA's Defense HUMINT Service, the human intelligence collection agency of the Department of Defense.

The lack of direct access to Curveball made it more difficult to assess his veracity. But such lack of access does not preclude the Intelligence Community from attempting to assess the source's bona fides and the credibility of the source's reporting. Indeed, it is incumbent upon professional intelligence officers to attempt to do so, through a process referred to within the Intelligence Community as "vetting" or "asset validation."

Defense HUMINT, however, did not even attempt to determine Curveball's veracity. A Defense HUMINT official explained to Commission staff that Defense HUMINT believed that it was just a "conduit" for Curveball's reporting--that it had no responsibility for vetting Curveball or validating his information. ²⁷⁵ In Defense HUMINT's view, asset validation is solely the responsibility of analysts--in their judgment if the analysts believe the information is credible, then the source is validated. ²⁷⁶ This line echoes what Defense HUMINT officials responsible for disseminating Curveball's reporting told the Senate Select Committee on Intelligence; they told the Committee that it was not their responsibility to assess the source's credibility, but that it instead was up to the analysts who read the reports to judge the accuracy of the contents. ²⁷⁷

The Senate Select Committee on Intelligence concluded that this view represents a "serious lapse" in tradecraft, and we agree. ²⁷⁸ Analysts obviously play a crucial role in validating sources by evaluating the credibility of their reporting, corroborating that reporting, and reviewing the body of reporting to ensure that it is consistent with the source's access. But analysts' validation can only extend to whether what a source says is internally consistent, technically plausible, and credible given the source's claimed access. The process of validation also must include efforts by the operational elements to confirm the source's bona fides (i.e., authenticating that the source has the access he claims), to test the source's reliability and motivations, and to ensure that the source is free from hostile control. ²⁷⁹ To be sure, these steps are particularly difficult for a source such as Curveball, to whom the collection agency has no direct access. But human intelligence collectors can often obtain valuable information weighing on even a liaison source's credibility, and the CIA's DO routinely attempts to determine the credibility even of sources to whom it has no direct access. In light of this, we are surprised by the Defense HUMINT's apparent position that it had no responsibility even to attempt to validate Curveball.

As a footnote to this episode, while DIA's Defense HUMINT Service felt no obligation to vet Curveball or validate his veracity, it would later appear affronted that another agency--CIA--would try to do so. On February 11, 2003, after questions about Curveball's credibility had begun to emerge, an element of the DO sent a message to Defense HUMINT officials expressing concern that Curveball had not been vetted. The next day the Defense HUMINT division chief who received that message forwarded it by electronic mail to a subordinate, requesting input to answer CIA's query. In that electronic mail message, the Defense HUMINT division chief said he was "shocked" by CIA's suggestion that Curveball might be unreliable. The reply--which the Defense HUMINT official intended for Defense HUMINT recipients only but which was inadvertently sent to CIA as well--observed that "CIA is up to their old tricks" and that CIA did not "have a clue" about the process by which Curveball's information was passed from the foreign service. ²⁸⁰

Conclusion

This section has revealed that Intelligence Community management was remiss in not taking action based on expressed concerns about Curveball's reliability. In retrospect, we conclude that the Intelligence Community's leadership should have more aggressively investigated Curveball's bona fides, rather than seeing the confidence of the analysts and the responsible liaison service as sufficient reason to dismiss the rival concerns of the operators and other liaison services. These leaders also should have pushed harder for access to Curveball--even at the cost of significant inter-liaison capital--given that the source's reporting was so critical to the judgment that Iraq was developing a mobile BW capability. After the NIE, CIA leadership should have paid closer heed to mounting concerns from the DO and, at the very least, informed senior policymakers about these concerns.

This said, the Community's failure to get the Iraq BW question right was not at its core the result of these managerial shortcomings. We need more and better human intelligence, but all such sources are inherently uncertain. Even if there had not been--as there was--affirmative reason to doubt Curveball's reporting, it is questionable whether such a broad conclusion (that Iraq had an active biological weapons production capability) should have been based almost entirely on the evidence of a single source to whom the U.S. Intelligence Community had never gained access. The Intelligence Community's failure to get the BW question right stemmed, first and foremost, from the strong prevailing assumptions about Iraq's intentions and behavior that led the Intelligence Community to conclude that Curveball's reporting was sufficient evidence to judge with "high confidence" that Iraq's offensive BW program was active and more advanced than it had been before the first Gulf War. The Intelligence Community placed too much weight on one source to whom the Community lacked direct access--and did so without making clear to policymakers the extent of the judgment's reliance on this single, unvetted source.

The Intelligence Community's Pre-War Assessments

The Intelligence Community's assessment of Iraq's CW programs and capabilities remained relatively stable during the 1990s, judging that Iraq retained a modest capability to restart a chemical warfare program. The October 2002 NIE therefore marked a shift from previous assessments in that it concluded that Iraq had actually begun renewed production of chemical agents on a sizable scale. ⁴²³ This shift was based primarily on imagery, although analysts also saw support for their assessment in a small stream of human and signals intelligence on Iraq's CW capabilities. ⁴²⁴

Background. For more than ten years, the Intelligence Community believed that Iraq retained the capability to jumpstart its CW program. After Operation Desert Storm in 1991, the Community judged that Iraq retained CW munitions and CW-related materials; the Community based these judgments primarily on accounting discrepancies between Iraq's declarations about its chemical weapons program and what UNSCOM had actually discovered. ⁴²⁵ As with assessments of Iraq's nuclear and biological weapons programs, the conclusion that Iraq still had CW munitions was "reinforced by Iraq's continuing efforts to frustrate" United Nations inspectors. ⁴²⁶ Encapsulating this line of reasoning, in 1995 the CIA judged that Iraq could "begin producing [chemical] agent in a matter of weeks after a decision to do so," based on the assessment that Iraq had "sequestered ...at least some tens of metric tons" of CW precursors. ⁴²⁷ This assessment cautioned, however, that building Iraq's "CW program to its previous levels" would require two to three years. ⁴²⁸

Mid-1990s: Growing concern. The Intelligence Community's understanding of Iraq's CW program was altered with the defection in August 1995 of Hussein Kamil, the head of Iraq's Military Industrialization Committee and, as such, the head of Iraq's WMD programs. Among a host of damning revelations, Kamil released details previously unknown to the U.S. Intelligence Community about Iraq's pre-1991 production and use of VX nerve gas. More specifically, Iraq subsequently admitted that it had worked on in-flight mixing of binary CW weapons before the Gulf War, produced larger amounts of VX agent than previously admitted, and perfected long-term storage of a VX precursor. These admissions about Iraqi work on VX--a potent nerve agent and an advanced chemical weapon--all played an important role in shaping subsequent Intelligence Community assessments about Iraq's CW program. ⁴²⁹

Two further revelations about the extent of Iraq's pre-1991 CW efforts also markedly influenced the Community's view of Iraq's CW programs. First, in June 1998, U.S. tests of warhead fragments from an Iraqi al-Hussein missile yielded traces of degraded VX. ⁴³⁰ This finding was noteworthy to Community analysts because it established beyond any doubt (in analysts' eyes) that Iraq, before 1991, had successfully weaponized VX--a technical advance that Iraq refused to admit in its United Nations declarations both before and after the United States became aware of the test results. ⁴³¹

Second, in July 1998, weapons inspectors found documents--now commonly known as the "Air Force Documents"--that detailed Iraqi CW use in the Iran-Iraq War. ⁴³² This finding was significant because the documents indicated Iraq had expended far fewer CW munitions in the Iran-Iraq War than previously thought, thus suggesting that Iraq possessed more unexpended CW munitions than analysts believed. Analysts lent additional credence to the information because Iraqi officials refused to let inspectors actually keep the relevant document, which suggested to analysts that the documents were incriminating and important. ⁴³³ Though both of these revelations concerned Iraq's pre-1991 CW effort, analysts saw them as lending support to the assessment that Iraq was continuing its deliberate efforts to obscure elements of its CW capabilities.

By 1998, the Intelligence Community was continuing to assess that Baghdad retained "key elements of its CW program including personnel, production data, and hidden stocks of production equipment and precursor chemicals" and that "Iraq could begin limited CW agent production within weeks after United Nations sanctions are lifted and intrusive inspections cease." ⁴³⁴ The Community noted, however, that it lacked "reporting to confirm whether [CW] production [was] taking place." ⁴³⁵

2001-2002: Little change. The Community continued through 2001 to note that there was no evidence that Iraq had started large-scale production of CW. ⁴³⁶ Though analysts continued to believe that Iraq's capability to produce CW was increasing, primarily through the development of an indigenous chemical industry, and that Iraq might have engaged in small-scale production, ⁴³⁷ the Community continued to assess that Iraq had not restarted large-scale production. ⁴³⁸ Even after the terrorist attacks of September 11, 2001--when the Intelligence Community detected what it determined to be the dispersal of Iraqi military units in anticipation of U.S. military strikes ⁴³⁹ --the CIA found no evidence that the munitions Iraq was moving were CW-related. ⁴⁴⁰ And additional reporting during this time did not reveal whether certain suspect sites were actively engaged in CW weapons production--although it remained impossible to determine whether dual-use precursor chemicals were being produced for illicit purposes. ⁴⁴¹

With respect to possible CW stockpiles, as of 2002 the Community assessed that Iraq possessed between 10 and 100 metric tons of CW agent and that it might have had sufficient precursors to produce an additional 200 metric tons. ⁴⁴² This estimated stockpile was smaller than the stockpiles Iraq possessed before the Gulf War, as an early 2002 Senior Executive Memorandum noted. ⁴⁴³ But according to a CIA analyst's mid-2002 briefing to senior officials, Iraq could restart CW production in a matter of days by using dual-use facilities and hidden precursors. ⁴⁴⁴ These assessments, however, did not go so far as to conclude that Iraq had restarted production or, relatedly, had sizable CW stockpiles.

The October 2002 NIE. The October 2002 NIE reflected a shift in the Intelligence Community's judgment about Iraq's CW program in two ways: (1) the NIE assessed that Iraq had large stockpiles of CW; and (2) the NIE unequivocally stated that Iraq had restarted CW production. ⁴⁴⁵

Regarding stockpiles, the NIE stated that "[a]lthough we have little specific information on Iraq's CW stockpile, Saddam probably has stocked at least 100 metric tons and possibly as much as 500 metric tons of CW agents--much of it added in the last year." ⁴⁴⁶ This judgment represented a significant increase in the Intelligence Community's estimate of the size of Iraq's CW stockpile.

This stockpile estimate rested primarily on Iraqi accounting discrepancies, Iraq's CW production capacity, estimates of Iraqi precursor stocks, and--at the upper limit (500 metric tons)--on practical considerations such as the size of pre-Gulf War stockpiles and Iraq's limited delivery options. ⁴⁴⁷ This calculation was also informed by the Intelligence Community's assessments of Iraqi military requirements, ammunition demand, and possible changes in Iraqi use doctrine. ⁴⁴⁸

The lower end of this stockpile range (100 metric tons) was premised on the aforementioned 1999 estimate that Iraq possessed between 10 and 100 metric tons of CW agents and that Iraq "could" produce an additional 200 tons of agents "using unaccounted-for precursor chemicals." ⁴⁴⁹ This 1999 estimate was itself premised on previous Iraqi CW accounting irregularities. ⁴⁵⁰ The Community assessments of the range of Iraq's CW stockpile thus rested largely on what analysts estimated Iraq could do with unaccounted-for precursors and production capabilities.

In addition to assessing the size of the Iraqi CW stockpile, the NIE judged that "much" of the CW stockpile had been "added in the last year." ⁴⁵¹ This latter assessment, in turn, rested on the NIE's second major CW conclusion: that Baghdad had "begun renewed production of mustard, sarin, GF (cyclosarin), and VX."

The NIE's judgment that Iraq had restarted CW production was based primarily on imagery intelligence. ⁴⁵² As analysts subsequently explained, this imagery showed trucks transshipping materials to and from ammunition depots, including suspect CW sites, in Iraq. These transshipments began in March 2002 and continued until early 2003. ⁴⁵³ At approximately 11 sites, imagery analysts saw a number of "indicators" in the imagery that suggested to them that some of the trucks were possibly moving CW munitions; then, because imagery analysts observed evidence of numerous such shipments, CW analysts in turn assessed that Iraq was moving significant volumes of CW munitions and therefore that Iraq had restarted CW production. ⁴⁵⁴ These indicators included the presence of "Samarra-type" trucks--a distinctive type of tanker truck--which were regularly associated with CW shipments in the late 1980s and during the Gulf War; atypical security patterns "associated with" the Special Republican Guard, which was believed to be responsible for protecting parts of Iraq's WMD programs; at least at one site, the grading of the topsoil, which likewise suggested to analysts deliberate concealment of suspect activity; and other indicators. ⁴⁵⁵

Although the NIE's judgment that Iraq had restarted CW production was based primarily on imagery, that judgment was also supported by small streams of human and signals intelligence. The NIC subsequently explained in its Statement for the Record that this human intelligence reporting consisted of "a number of specific reports alleging that Iraq had resumed large-scale production of CW agents." ⁴⁵⁶ None of these reports was considered "highly reliable," however, and only six were deemed "moderately reliable." ⁴⁵⁷

Of these reports, Community analysts identified to us several as having been most significant, although subsequent analysis of the reports revealed--in some cases--serious flaws in the reporting. The key reports were: one involving a foreign source in 1999 who reported that two Iraqi companies were involved in the production of nerve gas; ⁴⁵⁸ reporting concerning a factory for the production of castor oil that could be used to make "sarin"; ⁴⁵⁹ information from an Iraqi defector, who claimed to be an expert in VX production, describing the production of "tons" of nerve agents in mobile labs; ⁴⁶⁰ reporting from a source with "good but historical access" asserting that, as of 1998, mustard and binary chemical agents were being produced in Iraq; ⁴⁶¹ a source who reported that Iraq was producing a binary compound and mustard as of fall 2001; ⁴⁶² and reporting on the production of CW at dual-use facilities. ⁴⁶³

Finally, a liaison service reported in September 2002 that a senior Iraqi official had indicated that Iraq was producing and stockpiling chemical weapons. ⁴⁶⁴ Although this report was distributed to a very small group of senior officials prior to the publication of the NIE--including to the NIE's principal author--it was not made available to most analysts. ⁴⁶⁵ In any event, as described below, the senior Iraqi official later denied having made such statements.

In addition to these imagery indicators of transshipment activity and human intelligence, the NIE also drew upon a handful of additional pieces of information--based largely on other Intelligence Community reporting --to support the assessment that Baghdad had restarted CW production. This information suggested suspect activity at dual-use sites and included: indications that Iraq was expanding its indigenous chemical industry in ways that were deemed unlikely to be for civilian purposes, specifically by increasing the indigenous production capacity for chlorine--despite the fact that Iraq's civilian chlorine needs were met through United Nations-permitted imports; ⁴⁶⁶ the "management" of key chemical facilities by "previously identified CW personnel"; ⁴⁶⁷ attempted procurement of nuclear, biological, and chemical weapons defensive materials; and the attempted procurement of dual-use materials associated with CW. ⁴⁶⁸ Although the NIE noted that the Intelligence Community could not "link definitively Iraq's procurement of CW precursors, technology, and specialized equipment from foreign sources directly" to its CW program, ⁴⁶⁹ it nevertheless assessed that "Iraq's procurements have contributed to the rebuilding of dual-use facilities that probably are adding to Iraq's overall CW agent capability." ⁴⁷⁰ In drawing this conclusion, the NIE drew particular attention to Iraq's attempts to obtain necessary precursors for nerve agents. ⁴⁷¹

Finally, reporting on other aspects of Iraq's unconventional weapons programs also influenced some analysts' CW-related conclusions. Specifically, reporting on the existence of Iraqi mobile BW production facilities--namely, reports from Curveball--buttressed some analysts' certainty in their CW judgments. As one CIA analyst put it, "much of the CW confidence [in the pre-war assessments] was built on the BW confidence." ⁴⁷² In other words, although some CW analysts at times questioned the existence of significant Iraqi CW stockpiles, the reports that Iraq had a hidden, mobile BW program pushed the analysts "in the other direction" and helped convince them of their ultimate conclusion: that Iraq was hiding a CW program. ⁴⁷³

Post-October 2002 NIE reports. In November 2002, the NIC published a Memorandum to Holders of the October NIE entitled Iraq's Chemical Warfare Capabilities: Potential for Dusty and Fourth-Generation Agents. ⁴⁷⁴ The Memorandum warned that Iraq might possess dusty agent ⁴⁷⁵ and that it had the technical expertise to develop fourth-generation agents ⁴⁷⁶ that could be extremely lethal. Identifying the "Key Intelligence Gaps" on Iraq's CW program, the Memorandum observed that although the Intelligence Community "assess[ed]" that Iraq was producing blister and nerve agents, the Intelligence Community had not "identified key production facilities" and did "not know the extent of indigenous production or procurement of CW precursors." ⁴⁷⁷ But just as the NIE had cautioned that the Intelligence Community had "little specific information on Iraq's CW stockpile," the Memorandum stated that the Intelligence Community had "almost no information on the size, composition, or location of Iraq's CW stockpile." ⁴⁷⁸ In a separate NIE published in January 2003, however, the Community reiterated its estimate that Iraq "ha[d] 100 to 500 metric tons of weaponized bulk agent." ⁴⁷⁹

In December 2002, CIA's WINPAC published a coordinated Intelligence Community paper that reiterated its belief that "Iraq retain[ed] an offensive CW program," but it did not specifically describe the extent of any CW stockpiles. ⁴⁸⁰ In addition, the CIA reported the Intelligence Community had "low confidence" in its ability to monitor the Iraqi CW program due to "stringent operational security" and "successful denial and deception practices." ⁴⁸¹

Post-War Findings of the Iraq Survey Group

The Iraq Survey Group's findings undermined both the Intelligence Community's assessments about Iraq's pre-war CW program and, indeed, the very fundamental assumptions upon which those assessments were based. The ISG concluded--contrary to the Intelligence Community's pre-war assessments--that Iraq had actually unilaterally destroyed its undeclared CW stockpile in 1991 and that there were no credible indications that Baghdad resumed production of CW thereafter. ⁴⁸² Iraq had not regained its pre-1991 CW technical sophistication or production capabilities prior to the war. Further, pre-war concerns of Iraqi plans to use CW if Coalition forces crossed certain defensive "red lines" were groundless; the "red lines" referred to conventional military planning only. ⁴⁸³ Finally, the only CW the Iraq Survey Group recovered were weapons manufactured before the first Gulf War; the ISG concluded that, after 1991, Iraq maintained only small, covert labs to research chemicals and poisons, primarily for intelligence operations. ⁴⁸⁴ However, "Saddam never abandoned his intentions to resume a CW effort when sanctions were lifted and conditions were judged favorable," and Iraq's post-1996 infrastructure improvements "would have enhanced Iraq's ability to produce CW" if it had chosen to do so. ⁴⁸⁵

Despite having "expended considerable time and expertise searching for extant CW munitions,"--the vaunted stockpiles--the ISG concluded with "high confidence that there are no CW present in the Iraqi inventory." ⁴⁸⁶ The ISG specifically investigated 11 sites that were associated with suspected CW transshipment activity, conducting an in-depth inspection of two of the sites, which were "assessed prior to war to have the strongest indicators of CW movement." ⁴⁸⁷ Neither of these sites revealed any CW munitions. ⁴⁸⁸ Further, the ISG's "review of documents, interviews, intelligence reporting, and site exploitations revealed alternate, plausible explanations" for pre-war transshipment activity that the Intelligence Community judged to have been CW-related. ⁴⁸⁹

Regarding Iraq's dual-use chemical infrastructure and personnel, the Iraq Survey Group found no direct link to a CW program. Instead, investigators found that, though Iraq's chemical industry began expanding after 1996, in part due to the influx of funds and resources from the Oil-for-Food program, the country's CW capabilities remained less than those which existed prior to the Gulf War. ⁴⁹⁰ The ISG also interviewed 30 of the approximately 60 "key" Iraqi CW scientists, all of whom denied having been involved in any CW activity since 1990 and the vast majority of whom denied having any knowledge of any CW activity occurring. ⁴⁹¹

The ISG also cited a number of reasons why Iraq's expansion of its chlorine capacity was not, contrary to the NIE's assessment, capable of being diverted to CW production. ⁴⁹² Specifically, Iraq experienced a "country-wide chlorine shortage," and Iraq's chlorine plants "suffered from corroded condensers and were only able to produce aqueous chlorine." ⁴⁹³ Further, "[t]echnical problems and poor maintenance of aging equipment throughout the 1990s resulted in many chemical plants, including ethylene and chlorine production plants, operating at less than half capacity despite the improvements to the chemical industry." ⁴⁹⁴

In sum, the Iraq Survey Group found no direct link between Iraq's dual-use infrastructure and its CW program. However, "concerns" about some aspects of the infrastructure ⁴⁹⁵ arising out of "an extensive, yet fragmentary and circumstantial body of evidence" suggested Saddam intended to maintain his CW capabilities by preserving CW-related assets and expertise. ⁴⁹⁶

Regarding Iraqi decisionmaking about its CW program after 1991, the ISG concluded that, in the aftermath of the Gulf War, "Iraq initially chose not to fully declare its CW" in anticipation that inspections would be short-lived and ineffective. This position changed after a particularly invasive search in late June 1991, after which Iraq destroyed its hidden CW and precursors while retaining some documents and dual-use equipment. Iraq kept these latter items for the next five years, but did not renew its CW efforts out of fear that such a move would imperil its effort to have sanctions lifted. In August 1995, however, after the defection of Hussein Kamil, Saddam relented and revealed to inspectors extensive VX research and other, more advanced, technologies. ⁴⁹⁷

Overall, although the vast majority of CW munitions had been destroyed, the Iraq Survey Group recognized that questions remained relating to the disposition of hundreds of pre-1991 CW munitions. ⁴⁹⁸ Still, given that, of the dozens of CW munitions that the ISG discovered, all had been manufactured before 1991, the Intelligence Community's 2002 assessments that Iraq had restarted its CW program turned out to have been seriously off the mark. ⁴⁹⁹

Finally, on two ancillary issues the ISG found little or no evidence to support indications of Iraqi CW efforts. First, with respect to a "red line" defense of Baghdad, the ISG found no information that such a defense--which amounted to a multi-ring conventional defense of the city--called for the use of CW. ⁵⁰⁰ According to a senior Iraqi military officer, the "red line" was simply the line at which Iraqi military units would no longer retreat. ⁵⁰¹ At the same time, both generals and high-level defense officials believed that a plan for CW use existed, even though they themselves knew nothing about it. ⁵⁰²

Second, with respect to CW work by the Iraqi Intelligence Service, there was "no evidence" of CW production in clandestine labs, other than the Service's laboratory effort to develop substances to kill or incapacitate targeted individuals. ⁵⁰³

Analysis of the Intelligence Community's Pre-War Assessments

As the foregoing comparison illustrates, the Intelligence Community's pre-war assessments of Iraq's CW program were well off the mark. Iraq did not have CW stockpiles; it was not producing CW agent; and its chemical infrastructure was in far worse shape than the Intelligence Community believed. It is a daunting task in any circumstance to distinguish a normal chemical infrastructure and conventional military establishment on the one hand from a chemical warfare program on the other. But the Community made more difficult the challenges of identifying a CW program in Iraq by latching on to ambiguous imagery indicators and by failing to collect enough good intelligence to keep analytic judgments tethered to reality.

There are several reasons for the significant gap between the Intelligence Community's pre-war assessment of Iraq's CW program and the Iraq Survey Group's findings. Chief among these was the over-reliance on a single, ambiguous source (the Samarra-type tanker trucks) to support multiple judgments. Less central, although still significant, were the failure of analysts to understand fully the limitations of technical collection; the lack of quality human intelligence sources; the lack of quality signals intelligence; and, on a broader plane, the universal difficulty of establishing the existence of a CW program in light of the prevalence of dual-use technology.

As noted, the pre-war assessment that Iraq had restarted CW production relied primarily on CW analysts' assessments of imagery intelligence. ⁵⁰⁴ This imagery showed trucks transshipping materials to and from ammunition depots, including suspect CW-sites, in Iraq. ⁵⁰⁵ In the late spring of 2002, analysts started to believe that these shipments involved CW munitions. ⁵⁰⁶ This belief was based on the aforementioned "indicators" seen on the imagery--that is, activity and circumstances surrounding the shipments that were thought to be indicative of CW activity. The most important of these indicators was the presence of "Samarra-type" trucks--a distinctive type of tanker truck--which had been regularly associated with Iraqi CW shipments in the late 1980s and during the Gulf War. ⁵⁰⁷ Based on the assessment that the presence of these Samarra-type trucks (in combination with the other indicators) suggested CW shipments, CW analysts then judged that the frequency of such transshipments pointed to the assessment that "CW was already deployed with the military logistics chain," which, in turn, indicated to these analysts that Iraq had added to its CW stockpile in the last year. That assessment, in turn, indicated to analysts that Iraq had restarted CW production.

In short, the key pre-war assessments about Iraq's CW program--that Iraq was actively producing CW and had increased its stockpile of CW--rested on the following evidence and associated reasoning:

·                     Imagery revealed the presence of Samarra-type trucks at suspect weapons sites;

·                     The presence of Samarra-type trucks indicated CW activity;

·                     The scale of the Samarra-type trucks' involvement demonstrated Iraq had already deployed CW with their forces; and

·                     For CW to be deployed with Iraqi forces, Iraq had to have restarted CW production within the past year--the period during which analysts had seen Samarra-type trucks.

As this logic train illustrates, the final conclusion regarding restarted CW production was, therefore, fundamentally grounded on the single assessment that the Samarra-type trucks seen on imagery were in fact CW-related. ⁵⁰⁸ This assessment, however, proved to be incorrect--thereby eliminating the crucial pillar on which the Community's judgment about Iraq's CW program rested.

Post-war investigation revealed how the Intelligence Community ran astray. After the war, NGA "reassessed" the imagery from one of the sites thought to bear the strongest indications of CW activity--the Al Musayyib Barracks--by incorporating information from ISG inspections and debriefings of key personnel. ⁵⁰⁹ Contrary to pre-war assessments, NGA concluded that the activity represented "conventional maintenance and logistical activity rather than chemical weapons." ⁵¹⁰ NGA analysts drew this conclusion in part after reexamining imagery and in part on ISG debriefs of former commanders of the Al Musayyib site. ⁵¹¹

More detailed analysis of other imagery intelligence--in particular, surface grading-- also revealed the absence of a clear link to CW. ⁵¹² NGA assessed that grading could be associated with innocuous, routine activities. ⁵¹³ The rationales behind that assessment are discussed in the classified report.

The story is much the same with respect to pre-war assessments of other imagery evidence regarding certain security patterns. ⁵¹⁴ Post-war analysis by NGA could not confirm pre-war assessments that these security patterns were indicative of Special Republican Guard activity associated with security at CW-related sites. Indeed, at least one human source debriefed after the war said the security activity in question was not related to the Special Republican Guard and that it was actually related to the performance of miscellaneous jobs associated with the ammunition depot. ⁵¹⁵

Finally, post-war debriefings suggested that other CW-related imagery evidence was also innocuous, although this suggestion was neither definitively confirmed nor refuted by the imagery reassessment. ⁵¹⁶ And NGA notes that it is generally not possible to determine from imagery whether some activities, such as certain safety measures, are intended to support the training of offensive or defensive chemical warfare troops. And NGA has noted that imagery, when used alone, may not definitely determine the intended purpose of an adversary's activity.

The Community's over-reliance on ambiguous imagery indicators thus played a pivotal role in its ultimate misjudgment that Iraq had restarted CW production and had increased its CW stockpiles. In our view analysts relied too heavily on the presence of Samarra tanker trucks--backed by other, even more ambiguous imagery indicators--to support multiple, interdependent, and wide-ranging judgments about Iraq's chemical warfare program. And the Community did so despite the truism about which NGA itself has cautioned: imagery alone can neither prove nor disprove a CW association. ⁵¹⁷

Building one assessment upon another in this fashion--without carrying forward the uncertainty of each "layer" of assessment--results in a false impression of certainty for analysts' ultimate judgment. We believe, therefore, that at a minimum analysts must communicate the uncertainty of their judgments, and the degree to which they rely on narrow assessments about specific indicators. Moreover, avoiding the pitfalls of such layering requires careful consideration of alternative hypotheses, such as, in this case, the possibility that the shipments involved conventional weapons and that the trucks were for water supply or fire suppression.

We do not discount the fact that analysts must sometimes focus on seemingly mundane indicators. But at the same time analysts must always recognize, and communicate to decisionmakers, the tenuous quality of their reasoning.

Analytical flaws in assessing the significance of the imagery indicators were not the only factors leading to the misassessment of the imagery intelligence. In addition, analysts may have misperceived the significance of the imagery on Iraq's supposed CW program because they did not fully understand--and the collectors did not fully explain--the scope and nature of imagery collection against the target. Indeed, we cannot rule out the possibility that the analytic judgment that Iraq had added to its CW stockpile in the preceding year rested, at least in part, on a simple increase in collection and reporting rather than any rise in Iraqi activity.

Pre-war, analysts relied upon imagery to detect transshipment activity at suspected CW sites, and beginning in March 2002, analysts believed that they were seeing an "increase" in such activity. ⁵¹⁸ In reality, however, the "increase" in transshipment activity that analysts saw starting in March 2002 may have been due, at least in part, to an increased volume of imagery collected by U.S. satellites rather than to any increased activity by the Iraqis. To only somewhat oversimplify the matter, it wasn't that the Iraqis were using Samarra trucks more often in 2002--it was that in 2002 the United States was taking more pictures of places where the Samarra trucks were being used. And this failure to distinguish between actual increased activity at suspect CW sites and the appearance of increased activity due to increased imaging likely contributed to the mistaken assessment that Iraq was ramping up CW production in 2002.

This error sprung from the fact that not all Community analysts were fully cognizant of a major change in NGA collection that occurred in the spring of 2002. ⁵¹⁹ Until 2000, imagery collection on Iraq had been oriented primarily toward supporting military operations associated with the no-fly zones. ⁵²⁰ But in 2001 and 2002, imagery collection against Iraq WMD more than doubled, prompted by recommendations that more attention be given to the target. ⁵²¹ Most significantly, the United States began "expanded imagery collection over Baghdad [and] suspect WMD sites" in March 2002--not coincidentally the same time that analysts began to "see" new activity they associated with CW transshipments. ⁵²²

Thus, in drawing their conclusions about the state of Iraq's CW production based on increased transshipment activity, analysts did not realize the necessity of distinguishing between the "new" activity they saw, on the one hand, at sites that had been previously imaged on a regular basis (e.g., suspect WMD sites) and, on the other, at sites that had not been previously imaged on a regular basis (e.g., ammunition depots that had not been previously associated with WMD). ⁵²³ Whereas increased activity at the former could be attributed to changes in Iraqi behavior (since the United States had been photographing the sites prior to March 2002), the same could not be said for the latter category (since there was no "baseline" of activity with which to compare levels of activity seen from March 2002 on). ⁵²⁴

This problem extended to one of the sites that was key to analysts' conclusions about Iraqi CW production--the Al Musayyib Barracks. According to NGA, Al Musayyib had not been regularly imaged prior to the March 2002 imaging blitz because it had not been previously associated with Iraq's chemical or biological weapons programs. ⁵²⁵ Unaware of this important fact, analysts confidently assessed that the Iraqis had expanded transshipment activity at Al Musayyib, as well as other sites, when they began to see more images of Samarra-type truck activity. In short, analysts attributed what they saw to nefarious Iraqi activity when it could just as easily have been attributed to changes in U.S. collection priorities. In our view, this failure is the direct result of poor communication between analysts and collectors about a crucial change in the scope and nature of collection against a vital target.

Conclusion

Similar to its assessments about Iraq's nuclear and biological efforts, the Intelligence Community's mistaken assessments about Iraq's chemical weapons program can be traced in large part to a single point of failure--the Community's over-reliance on ambiguous imagery indicators. But the Community's bottom line on Iraq's chemical weapons capabilities was further influenced by a breakdown in communication between imagery collectors and analysts; a basic paucity of quality intelligence, particularly quality signals intelligence; and the fact that much of the human and signals intelligence that was collected was bad.

It is, however, understandable that analysts assessed--as they did throughout the 1990s--that Iraq retained a chemical warfare capability. Iraq's pre-Gulf War chemical weapons stockpile was large and relatively sophisticated. Nor did Saddam's uncooperative and secretive behavior after the war encourage confidence that he had converted from the CW path. The Community's failure on CW was therefore not in thinking that Iraq had such a capability--that was, in many ways, the only sensible conclusion, given the evidence. Rather, analysts erred in their assessment--based largely on ambiguous imagery indicators that could not logically support the judgment--that Iraq had in fact resumed producing and stockpiling significant quantities of CW.

The Intelligence Community assessed in the October 2002 NIE that Iraq was developing small Unmanned Aerial Vehicles (UAVs) capable of autonomous flight, which most agencies assessed were "probably" intended to deliver biological warfare agents. ⁵⁴⁸ The Intelligence Community also judged that these UAVs could threaten the U.S. homeland. ⁵⁴⁹ This latter assessment was based on an Iraqi attempt to procure commercially available civilian U.S. mapping software for its UAVs. That attempted procurement, the Intelligence Community assessed, "strongly suggest[ed] that Iraq [was] investigating the use of these UAVs for missions targeting the United States." ⁵⁵⁰

By January 2003, however, the Intelligence Community had pulled back from its view that Iraq intended to target the United States. ⁵⁵¹ This re-assessment reflected a belief among CIA analysts that the Iraqi attempt to procure U.S. mapping software may have been inadvertent. ⁵⁵² As a result, the Intelligence Community assessed in January 2003 that while the mapping software could provide the capability to target the United States, the purchasing attempt did not necessarily indicate an intent to do so. ⁵⁵³ By early March 2003, CIA had further retreated from the view that the purchase of the mapping software evidenced an intent to target the United States and, in early March 2003, on the eve of the invasion of Iraq, CIA advised senior policymakers that it was an open question whether the attempted software procurement evinced the intent to target the United States at all. ⁵⁵⁴

Following its exhaustive investigation in Iraq, the Iraq Survey Group concluded that Iraq had indeed been developing small UAVs, but found no evidence that the UAVs had been designed to deliver biological agent. ⁵⁵⁵ Instead, the ISG concluded that Iraq had been developing and had flight tested a small, autonomous UAV intended for use as a reconnaissance platform, ⁵⁵⁶ and had developed a prototype for another small UAV for use in electronic warfare missions. ⁵⁵⁷ Although both UAVs had the range, payload, guidance, and autonomy necessary to deliver a biological agent, the ISG found no evidence that Iraq intended to use them in such a way. ⁵⁵⁸ With respect to the mapping software, Iraqi officials told ISG investigators that the software in question had been included as part of a package deal with autopilots they had purchased for the UAVs; the Iraqis, the ISG judged, had not actually intended to buy the mapping software. ⁵⁵⁹

The October 2002 NIE had also examined whether Iraq was deploying missiles capable of reaching beyond the 150 kilometer limit imposed by the United Nations. The NIE assessed that Iraq was deploying two types of short-range ballistic missiles capable of flying beyond the United Nations-authorized range limit. ⁵⁶⁰ The NIE also assessed, based largely on Iraqi accounting discrepancies and incomplete records and record keeping, that Iraq retained a covert force of up to a few dozen Scud-variant missiles in defiance of United Nations resolutions. ⁵⁶¹ The ISG concluded--consistent with this assessment--that Iraq had been developing and deploying ballistic missiles that exceeded United Nations restrictions, although the ISG also found, contrary to pre-war assessments, that Iraq had not retained Scud or Scud-variant missiles after 1991. ⁵⁶²

The Intelligence Community's assessments of Iraq's delivery systems developments offered both a bright and a dark spot on its Iraq record. While far from perfect (which can never be reasonably expected in intelligence work), the Community's judgments about the progress of Iraq's ballistic missile programs were substantively accurate. As the ISG discovered, the Iraqis were indeed violating United Nations strictures by working on missiles that exceeded the 150 kilometer range limit. But on the issue of whether Iraq was developing UAVs to deliver biological agent against U.S. targets--including the U.S. homeland--the Community erred, once again attributing more to spotty intelligence than that information could bear.

This section describes the Community's analysis of Iraq's work on delivery systems between the first Gulf War and Operation Iraqi Freedom, as well as the ISG's findings concerning the same. The Commission then offers its findings based on a thorough investigation into the Community's efforts on Iraqi delivery systems, concentrating particularly on the analytical flaws apparent from the Community's products on the uses of Iraqi UAVs.

The Intelligence Community's Pre-War Assessments

As with other aspects of Iraq's WMD programs, the Intelligence Community's assessment of Iraq's delivery systems evolved over the course of many years and was heavily influenced by Iraq's past actions and intransigence.

Background. Before the Gulf War Iraq had been in the early stages of a project to convert the MiG-21 jet aircraft into UAVs for BW delivery. ⁵⁶³ In addition, Iraq had experimented in 1990 on a BW spray system, designed to be used with the MiG-21 UAV. ⁵⁶⁴ Iraq admitted to this program in 1995, after the defection of Hussein Kamil. ⁵⁶⁵ Subsequent UNSCOM inspections discovered video showing the spray-system experiments. ⁵⁶⁶ Also, analysts in the early 1990s had observed continued activity at Salman Pak--Iraq's primary BW research and development facility prior to the Gulf War--where, UNSCOM reported, work continued on modified commercial crop sprayers for BW delivery and the presence of UAV program personnel. ⁵⁶⁷ Iraq claimed that, because of the war, it had abandoned the MiG-21 UAV project after conducting only one experiment in 1991, but UNSCOM inspections could not confirm this claim. ⁵⁶⁸ In the mid-1990s Iraq also began testing another modified jet aircraft, the L-29, as a UAV, that analysts believed was a follow-on to the converted MiG-21 program. ⁵⁶⁹

These discoveries also cast new light, in analysts' minds, on UNSCOM's earlier discovery of 11 small-to-medium sized UAV drones at the Salman Pak compound in 1991. ⁵⁷⁰ Although Iraq denied having developed these UAVs for BW delivery, Iraq's later admission--after an initial denial--that the MiG-21 program was for the purpose of delivering biological agents led analysts to believe, given Iraqi deception, that Iraq's small UAVs had a similar purpose. ⁵⁷¹ Analysts also focused on Iraqi admissions--in their 1996 declaration to the United Nations--that, in the late 1980s, senior Iraqi officials had met to discuss the feasibility of using small UAVs as BW delivery vehicles. ⁵⁷²

This history, along with evidence that Iraq had flight-tested small and medium-sized UAVs, led most Intelligence Community analysts to conclude consistently from the late 1990s through 2002 that Iraq was maintaining its UAV program for BW and CW delivery. ⁵⁷³ Briefings and written products to senior policymakers in mid-2002 reflected this assessment. ⁵⁷⁴ As with the other elements of Iraq's purported weapons programs, however, intelligence on UAVs in the years preceding 2002 was partial and ambiguous. While it was clear that Iraq did have a UAV program, the key question--whether that program was meant to be a delivery system--remained unanswered. Therefore, analysts' judgments again depended heavily upon assumptions based on Iraq's earlier behavior and Community views about Iraq's sophisticated denial and deception activities. ⁵⁷⁵

With respect to ballistic missiles, the Intelligence Community judged in 1992 that Iraq's ballistic missile programs were more advanced than the Community had assessed before the Gulf War. ⁵⁷⁶ Iraq was further along in its production capability for Scud and Scud-derivative missiles and had produced more components indigenously than the Intelligence Community had assessed before the Gulf War. ⁵⁷⁷ By 1995, the Intelligence Community judged that Iraq was developing liquid-propellant missiles with an expected range of about 150 kilometers. ⁵⁷⁸ In 1998, the Community assessed that these missiles, named the al-Samoud, were capable of flying farther than the 150 kilometer limit imposed by the United Nations and that Iraq was also developing solid-propellant missiles. ⁵⁷⁹ By early 2002, the Intelligence Community judged that Iraq probably still retained a small force of Scud missiles and that both its liquid-propellant and solid-propellant missiles were capable of flying over 150 kilometers. ⁵⁸⁰

October 2002 NIE. The October 2002 NIE judged, with a dissent from the Director of Air Force Intelligence, that Iraq was developing small UAVs "probably" for BW delivery which could be used against U.S. forces and allies in the region. ⁵⁸¹ In addition, the NIE mentioned the concern of most agencies about the possible intent to use UAVs as delivery systems against the U.S. homeland. ⁵⁸² This possible use was based on the attempted procurement of U.S. mapping software by an Iraqi procurement agent. ⁵⁸³

As noted, the Director of Air Force Intelligence dissented from the majority view. In contrast to other organizations, the Air Force judged that Iraq was developing UAVs "primarily for reconnaissance rather than [as] delivery platforms for [CW or BW] agents." ⁵⁸⁴ The Air Force further noted that CW or BW delivery is "an inherent capability of UAVs but probably is not the impetus for Iraq's recent UAV programs." ⁵⁸⁵

Analysts' judgments that Iraq's small UAVs were intended for BW delivery were based on the following logic: the Iraqis had admitted that the MiG-21 program was intended for BW delivery, and analysts judged that the L-29 program, for which there was some evidence of a BW-delivery mission, was the successor to the MiG-21 program. Because the L-29 program had suffered set-backs in late 2000 after a crash, analysts then deduced that Iraq's new, small UAVs may have been designed to replace the L-29 effort, and that they were therefore also intended to deliver BW agents. ⁵⁸⁶

There was very little reporting, however, to support the conclusion that the small UAVs were "probably" intended for BW delivery. Only one human intelligence report indicated that small UAVs were intended for CW or BW delivery. ⁵⁸⁷ Given the dearth of reporting on the purpose for the small UAVs, analysts instead deduced their intended purpose from Iraq's previous admissions and from what was assessed about the characteristics of Iraq's other UAV programs.

For example, analysts pointed to several human intelligence reports that suggested that Iraq's L-29 UAV program could be used to deliver CW or BW agents. ⁵⁸⁸ Only one of those reports, however, stated explicitly that the L-29 UAV was intended for biological or chemical weapon delivery, and that early 1998 report was based on a report of unknown reliability. ⁵⁸⁹ Analysts believed, though, that this conclusion was reinforced by separate reporting indicating that Iraq was prepared to use modified L-29 UAVs against U.S. forces in the Persian Gulf area; these UAVs, the reasoning went, would have been useless for delivery of conventional weapons and BW was therefore a likelier function. ⁵⁹⁰

But there were other indications that the UAVs were not intended for BW delivery. Iraq's 1996 declaration to the United Nations indicated that the drones discovered in 1991 were actually intended for reconnaissance and aerial targeting--not BW delivery. ⁵⁹¹ Intelligence reporting supported this view; Iraq was attempting to procure equipment for its small UAVs, which suggested the UAVs' purpose was reconnaissance. ⁵⁹² Finally, as noted in the Air Force dissent, the small UAVs were not ideally suited for BW or CW delivery; the Air Force assessed instead that "the small size of Iraq's new UAV strongly suggests a primary role of reconnaissance, although chemical/biological weapons (CBW) delivery is an inherent capability." ⁵⁹³ Although CIA's WINPAC had published an Intelligence Assessment in 2001 that discussed these possible non-BW delivery missions for Iraq's UAVs, such alternative missions were not emphasized in the October 2002 NIE because WINPAC's "focus [in] the NIE was WMD delivery systems and not the Iraqi UAV program as a whole." ⁵⁹⁴

In sum, the evidentiary basis for the pre-war assessment that Iraq was developing UAVs "probably intended" for BW delivery was based largely on the BW focus of Iraq's pre-1991 UAV programs and a thin stream of (primarily human intelligence) reporting that hinted at such a function for post-1991 UAVs. ⁵⁹⁵

As noted above, the NIE also judged that Iraq's UAVs "could threaten...the U.S. Homeland." ⁵⁹⁶ This assessment was based on two streams of reporting: first, intelligence reporting indicating that the UAVs had a range of over 500 kilometers and could be launched from a truck; and, second, reporting that an Iraqi procurement agent was attempting to buy U.S. mapping software for its small UAVs. ⁵⁹⁷ The latter piece of information was, however, the only evidence that supported Iraq's intent to target the United States. Based on this stream of reporting, the NIE reasoned that, because the mapping software would be useless outside the United States, its procurement "strongly suggest[ed]" Iraq was interested in using the UAVs to target the United States. ⁵⁹⁸

The procurement effort revealed by the reporting was spearheaded by an Iraqi procurement agent who had been involved in the pre-Gulf War Iraqi UAV program ("the procurement agent"). The procurement agent had subsequently emigrated to another country where he ran an illicit procurement network for Iraq. ⁵⁹⁹ In late 2000 or early 2001, the procurement agent received a "shopping list" from an Iraqi general associated with the UAV program that included autopilots and gyroscopes. To fill this request, the procurement agent researched potential suppliers for these items, and in May 2001 he submitted requests for price quotes to a manufacturer and a distributor for the requested items, which included autopilots and gyroscopes but also included "Map Source" mapping software. The distributor responded with a price quote for the autopilot package, which included "Garmin 50 State" topographic mapping software, also sold as "Map Source." After consulting with Baghdad and soliciting a final price quote, in early 2002 the procurement agent submitted a final procurement list, which included the Garmin 50 State mapping software, to the distributor. ⁶⁰⁰

Although the distributor had been assured by the procurement agent that the end-user was "legitimate," the distributor remained concerned about the procurement agent's interest in these items and contacted its own country's authorities in March 2002. The distributor also removed the mapping software from its website. ⁶⁰¹

Following the attempted procurement, several analytical assessments were published regarding the attempted procurement of the mapping software. An Intelligence Community Assessment titled Current and Future Air Threats to the US Homeland, published July 29, 2002, noted that Iraq was seeking route planning software and an associated topographic database "likely intended to use with its UAVs" and "almost certainly relate[d] to the United States." ⁶⁰² CIA's Office of Near Eastern and South Asian Analysis also disseminated an intelligence assessment on August 1, 2002, observing that the mapping software would "provide precise guidance, tracking, and targeting in the United States." ⁶⁰³

A liaison intelligence service subsequently approached the procurement agent to question him about the attempted procurement. ⁶⁰⁴ In these discussions, the procurement agent claimed that he had not intended to purchase mapping software of the United States. Although he admitted that the software he had ordered had not been "bundled" with other items he ordered, he explained that he had not well understood all of the elements of the package and had not wanted to miss out on an important piece of software. He said he had been concerned that the other system pieces might not work if he did not purchase the mapping software; it was cheap; and he had thought the system would allow the user to scan maps and program them into a GPS. Asked by the liaison service to submit to a thorough examination, the procurement agent refused. ⁶⁰⁵ Thus, by fall 2002, the CIA was still uncertain whether the procurement agent was lying.

While the October 2002 NIE was being coordinated, a CIA analyst interviewed the procurement agent in an effort to determine if his attempted procurement of the U.S. mapping software had in fact been inadvertent, as he claimed. The analyst initially concluded that the procurement agent was lying because a review of the website showed that, contrary to the procurement agent's claims, the option to purchase the mapping software was not on the page with the autopilots and gyroscopes. After further research, however, the analyst determined that the version of the website that the procurement agent had accessed in early 2001 had in fact contained the configuration and software option that the procurement agent described. This discovery led the analyst to believe that the purchase order may have indeed been inadvertent. ⁶⁰⁶

Although the CIA was now beginning to obtain indications that the procurement agent's attempted purchase of the U.S. mapping software may in fact have been inadvertent as the procurement agent claimed, CIA remained uncertain whether the procurement agent was lying. ⁶⁰⁷ As the National Foreign Intelligence Board was convening to review and approve the NIE, several CIA analysts expressed concern about its use of the words "strongly suggests" and recommended that the language be toned down. But these concerns did not reach the DCI himself until the Board process had concluded. ⁶⁰⁸ With the lengthy Board meeting finished, the DCI concluded that the word "strongly" would remain in the NIE because the coordination process was complete at that point and the new information had not been confirmed. ⁶⁰⁹

As noted, the NIE also stated that gaps in accounting suggested that Iraq retained a small covert Scud force, and the NIE assessed that Iraq was deploying missiles capable of flying farther than the United Nations limit of 150 kilometers. ⁶¹⁰

Post-NIE. The Intelligence Community's assessment that the UAVs were "probably" for BW delivery remained unchanged in the run-up to the war. ⁶¹¹ In a paper sent to the National Security Council in January 2003, the CIA noted that an Iraqi Ministry of Defense official had indicated that Iraq considered its UAVs to be an important strategic weapon. ⁶¹² And in testimony before the Senate Select Committee on Intelligence in early February 2003, DCI Tenet stated that "[w]e are concerned that Iraq's UAVs can dispense chemical and biological weapons." ⁶¹³

The Intelligence Community did, however, begin to retreat from its assessment that Iraq intended to target the U.S. homeland, though not quickly enough to prevent the charge's inclusion in the President's speech in Cincinnati in October 2002. In the immediate aftermath of the publication of the October 2002 NIE, CIA increasingly believed that the attempted purchase of the mapping software--on which this judgment was based--may have been inadvertent. ⁶¹⁴ Accordingly, at least one CIA analyst recommended that a reference to the UAVs targeting the United States be deleted from a draft Presidential speech. Because of persistent uncertainty within the analytical ranks about the significance of the mapping software, however, CIA and the Intelligence Community's official position remained unchanged from the NIE. The President's speech, which was delivered on October 7, 2002 in Cincinnati, therefore expressed concern "that Iraq is exploring ways of using these UAVs for missions targeting the United States." ⁶¹⁵

Subsequent analytical products did begin to reflect the uncertainty over the significance of the mapping software, though. An NIE addressing the UAV question, entitled Nontraditional Threats to the US Homeland Through 2007, which was approved by the National Foreign Intelligence Board in November 2002, was not published for two months because of disagreement over whether the order for the U.S. mapping software indicated Iraqi intent to target the U.S. homeland. ⁶¹⁶ The Nontraditional Threats NIE ultimately addressed the UAV issue in terms of capabilities rather than intent. That is, that NIE phrased the first judgment like the October 2002 Iraq NIE, noting that Iraqi UAVs "could strike the US Homeland if transported to within a few hundred kilometers," but phrased the software judgment only in terms of capability, noting that this "[route planning] software...could support [the] programming of a UAV autopilot for operation in the United States." For their parts, the Air Force, DIA, and the Army assessed that the purpose of the acquisition was to obtain generic mapping capability and that that goal was "not necessarily indicative of an intent to target the US homeland." ⁶¹⁷

By early March 2003, days before the March 19 invasion of Iraq, the CIA had further pulled back from its October NIE view, concluding in a memorandum to the Chairman of the House Permanent Select Committee on Intelligence that it was an open question whether the attempted procurement of the mapping software had been the result of a specific request from Baghdad or had been inadvertent. ⁶¹⁸ CIA also advised senior policymakers of this change in view. In the memorandum, the CIA stated that it "[had] no definite indications that Baghdad [was] planning to use WMD-armed UAVs against the U.S. mainland....[Although] we cannot exclude the possibility that th[e] purchase [of mapping software] was directed by Baghdad, information acquired in October suggests that it may have been inadvertent." ⁶¹⁹

With respect to ballistic missiles, CIA's position remained unchanged after the NIE. ⁶²⁰ Subsequent to the NIE, the Intelligence Community confirmed from Iraq's December 2002 declaration to the United Nations that Iraq had two versions of the al-Samoud missile, as described in the NIE. The longer-range version was inefficiently designed and did not go as far as the NIE had postulated, but it did have a range in excess of 150 kilometers. ⁶²¹

Post-War Findings of the Iraq Survey Group

The Iraq Survey Group concluded that, although Iraq had pursued UAVs as BW delivery systems in the past, Iraq's pre-Operation Iraqi Freedom program to develop small, autonomous-flight UAVs had actually been intended to fulfill reconnaissance and airborne electronic warfare missions. The ISG found no evidence suggesting that Iraq had, at the time of the war, any intent to use UAVs as BW or CW delivery systems. ⁶²²

The ISG concluded that Iraq's purpose in converting a MiG-21 into a Remotely Piloted Vehicle (RPV) in early 1991 had been to create a CBW delivery system. ⁶²³ After the MiG-21 RPV program failed, Iraq in 1995 resumed efforts to convert manned aircraft into RPVs, this time with an L-29 jet trainer. ⁶²⁴ The ISG, however, was unable to establish whether the L-29 had an intended CBW role, although the ISG did obtain some indirect evidence that the L-29 RPV may have been intended for CBW delivery. ⁶²⁵ The ISG also concluded that Iraq had the capability to develop chemical or biological spray systems for the L-29, but found no evidence of any work along these lines. ⁶²⁶ The L-29 program ended in 2001. ⁶²⁷

After several crashes of the L-29s, Iraq began to pursue long-range UAV options, probably at some point in 2000. ⁶²⁸ The ISG assessed, however, that these small UAVs had not been intended for use as chemical or biological delivery systems. ⁶²⁹ Specifically, although these small UAVs had the range, payload, guidance, and autonomy necessary to be used as BW delivery platforms, the ISG found no evidence that Iraq had intended to use them for such a purpose, had a suitable dispenser available, or had conducted research and development activity associated with use as a BW delivery system. ⁶³⁰

The more advanced of Iraq's two UAV programs, the Al-Musayara-20, had actually been developed for use as a reconnaissance platform, according to a senior Iraqi official. ⁶³¹ An interview with an Iraqi military official after Operation Iraqi Freedom revealed that many general officers had been shot down on helicopter reconnaissance missions during the Iran-Iraq war and therefore the military was interested in developing a UAV to perform such missions. ⁶³² According to another official, although the Al-Musayara-20 was developed for a reconnaissance role, other roles, such as for the delivery of high explosives, were also considered. ⁶³³

A competing program to the Al-Musayara, the Al Quds UAV program, had been less advanced but had included prototypes of varying sizes and weights. ⁶³⁴ The ISG concluded that the Al Quds program had been intended as an airborne electronic warfare platform. ⁶³⁵ Like the Al-Musayara, the Al Quds UAV had the range, autonomous guidance, and payload to enable it to deliver CBW. ⁶³⁶ The ISG uncovered no evidence, however, that Iraq had been developing a dispenser or had the intent to use the UAV as a BW delivery system. The Al Quds UAV was still in development when the war started. ⁶³⁷

According to the Iraq Survey Group, Iraqi officials denied deliberately seeking to acquire mapping software for the United States, but did say they received mapping software that came as part of the package with the autopilots they purchased. ⁶³⁸ An official claimed to have received several autopilots for UAVs through the procurement agent, but asserted that these autopilots were never installed because they arrived on the eve of the war. The official was unaware of the current location of the autopilots. ⁶³⁹

Regarding missile systems, the Iraq Survey Group concluded that Iraq had been developing and deploying ballistic missiles that exceeded United Nations restrictions. ⁶⁴⁰ The ISG concluded that Iraq had not possessed Scud or Scud-variant missiles after 1991, having by then either expended or unilaterally destroyed its stockpile. ⁶⁴¹

Analysis of the Intelligence Community's Pre-War Assessments

The Iraq Survey Group's uncovering of ballistic missile work that violated United Nations' restrictions affords a bright spot for the Intelligence Community's record of assessments on Iraq's unconventional weapons programs. The NIE accurately assessed that Iraq was deploying ballistic missiles with ranges exceeding United Nations restrictions. ⁶⁴² And although the NIE did not assess accurately the status of Iraq's Scud missile force, we are not especially troubled by this inaccuracy in light of the NIE's clear statement that this assessment was based merely on accounting discrepancies. ⁶⁴³

The record of the Intelligence Community's performance on the UAVs is more mixed (in part because the Intelligence Community's assessments themselves shifted during the pre-war period). While these assessments accurately described the Iraqi UAVs technical capability to deliver BW, the Intelligence Community's assessments that the UAVs were intended for this purpose--or that Iraq intended to strike the United States--were not borne out by the ISG's findings.

It is worth considering why the Intelligence Community's assessments were more correct in this area than they were with respect to other aspects of Iraq's arsenal. One possible answer is that--unlike the status of Iraq's BW and CW stockpiles--certain questions about Iraq's delivery systems--especially missiles--could be answered through technical means that operate from outside of the denied area, and which are generally less subject to questions about reliability. The intentions of a closed regime, however, are difficult to penetrate, and the reliability of any such information is difficult to determine. In areas of analysis that turn largely on intent, therefore, such as whether a regime is producing BW or intends to use its UAVs for BW delivery, the quality of the analysis will be largely dependent on the quality of the available human intelligence and on the ability of signals intelligence to penetrate communications. This highlights the imperative for analysts to explain the premise of their judgments, particularly when the ultimate judgment may rest on a very thin stream of information or on a chain of assumptions about intent.

With that said, the pre-war assessments on Iraq's delivery systems reflect significant shortcomings in analysis.

The NIE went beyond what one could reasonably conclude from the intelligence by judging that Iraq's UAVs were "probably intended to deliver biological warfare agents." Although past Iraqi interest in UAVs as BW vehicles was a reasonable indicator that the interest may have continued, the paucity of subsequent evidence should have led to a more nuanced statement in the NIE--such as that BW delivery was a possible use, but not necessarily an intended one. That the NIE did not discuss in any detail other possible missions for the UAVs only compounded this problem. ⁶⁴⁴ Moreover, most analysts discounted specific reporting indicating that Iraq was seeking equipment suited to a reconnaissance mission for its UAVs.

The Intelligence Community's assessments about the purpose of Iraq's UAV programs rested largely on inferences drawn from the inherent capabilities of such UAVs and knowledge about Iraq's past UAV programs, as discussed above. The conclusion that the UAVs were probably intended for BW delivery, however, reached beyond what the intelligence would reasonably bear.

Similarly, the single stream of reporting that the Iraqi procurement agent was attempting to purchase U.S. mapping software was insufficient to justify the NIE's statement that this interest "strongly suggest[ed]" that Iraq was investigating ways to target the U.S. homeland with UAVs. While certain analysts took the proper steps to push the Intelligence Community back from this judgment after doubts about this reporting emerged, the Intelligence Community as a whole was slow to assimilate this new information--particularly given its critical importance.

Whether or not any statement about attacking the U.S. homeland merited inclusion in the NIE, it is clear that the rather thin foundation for these assessments was not clearly communicated to policymakers. And the NIE's assessment that the UAVs were "probably intended" for BW delivery did not make clear that this conclusion rested largely on analytical assumptions about Iraqi intent based on the history of Iraq's UAV programs and on the UAVs' inherent capabilities. Nor did the NIE explain why it focused only on a possible weapons-related role for UAVs. A WINPAC analyst subsequently explained that the NIE's purpose was to discuss Iraq's WMD programs, and that accordingly the UAV section addressed the UAVs' use as a BW delivery platform and not their other possible uses. The failure to explain that reasoning in the NIE, however, leaves the impression that other possible uses for the UAV had been rejected rather than simply not discussed. ⁶⁴⁵

Finally, once again, the UAV episode reflects the tendency of Intelligence Community analysts to view data through the lens of its overall assumptions about Saddam Hussein's behavior. As noted, the NIE itself did not discuss other possible purposes for the UAVs or explain why the Estimate focused only on a weapons-related purpose. In addition, however, the Intelligence Community was too quick to characterize evidence that contradicted the theory that UAVs were intended for BW delivery as an Iraqi "deception" or "cover story." And a Senior Executive Memorandum warned that Iraq "probably will assert that UAVs are intended as target drones or reconnaissance platforms" to counter the claim in the British and U.S. "white papers" that the UAVs have a BW delivery role. ⁶⁴⁶

We commend the Intelligence Community for correctly assessing that Iraq was working on ballistic missile programs that violated United Nations strictures. As the ISG's findings demonstrate, however, many of the Community's specific estimates were off the mark. The Community judged, for instance, that Iraq retained a force of "up to a few dozen Scud-variant SRBMs [short-range ballistic missiles]." ⁶⁴⁷ The ISG concluded, however, that Iraq did not have any Scud missiles after 1991. ⁶⁴⁸ Similarly, the Community stated in the NIE that "in January 2002, Iraq flight-tested an extended-range version of the al-Samoud that flew beyond the 150-km range limit." The Community subsequently learned that it had misidentified the missile and had incorrectly deduced the missile's range; in actuality, the missile, while it had a range that exceeded 150 kilometers, did not exceed that limit by as much as analysts initially thought because the engine was less effective than they estimated. ⁶⁴⁹

In short, while the Community was technically correct that Iraq's missile systems violated United Nations strictures, it erred significantly in degree.

Conclusion

As has proven the case with other pre-war Intelligence Community judgments about Iraq's unconventional weapons programs, the assumptions held by Iraq analysts about Saddam Hussein's behavior were not unreasonable ones. These assumptions, however, drove the Intelligence Community to make overly inferential leaps about Iraq's UAV program based on thin evidence, and to fail to communicate this thin evidentiary basis to policymakers. While we fully understand that, in the wake of September 11, the Community felt obliged to report even relatively unlikely threats against the United States, the Community should have at a minimum explained more fully the uncertainties underlying its assessments.

The Intelligence Community failed to examine seriously the possibility that domestic or regional political pressures or some other factors might have prompted Saddam Hussein to destroy his stockpiles and to forswear active development of weapons of mass destruction after the first Gulf War. ⁶⁵⁰ The Community was certainly aware of the overall political dynamics that underpinned Saddam Hussein's regime--that he was a brutal dictator who ruled Iraq through a combination of violence, secrecy, mendacity, and fear--but the Community did not seriously consider the range of possible decisions that Saddam might make regarding his weapons programs given his idiosyncratic decisionmaking processes.

Though the likelihood that one of those possible decisions was to destroy his weapons seemed very remote to almost all outside observers, it was one that Community analysts at least should have seriously considered. In truth, any assessment of the effect of Saddam's political situation on his decisions about WMD in the years from 1991 to 2003 would more likely than not have resulted--and, in point of fact, did result--in the conclusion that Saddam retained his WMD programs. ⁶⁵¹ But whether or not it was extraordinarily difficult (if not effectively impossible) for the Intelligence Community to have discerned Saddam Hussein's true intentions, the Community's lack of imagination about the range of strategies and tactics Saddam might adopt left the Community with an incomplete analytical picture.

Having gained access to Iraq and its leaders, the Iraq Survey Group concluded that the unlikely course of voluntary abandonment by Saddam Hussein of his weapons of mass destruction was, in fact, the reality. According to the ISG, Saddam's regime, under severe pressure from United Nations sanctions, reacted by unilaterally destroying its WMD stockpiles and halting work on its WMD programs. ⁶⁵² Saddam decided to abandon his weapons programs because the economy and infrastructure of Iraq were collapsing under the weight of the sanctions. Saddam therefore ordered the unilateral destruction of biological and chemical weapons stockpiles in 1991 and chose to focus on securing sanctions relief before resuming WMD development. ⁶⁵³ At the same time, in an attempt to project power--both domestically as well as against perceived regional threats such as Iran and Israel--Iraq chose to obfuscate whether it actually possessed WMD. ⁶⁵⁴ As a result, the U.S. Intelligence Community--and many other intelligence services around the world--believed that Iraq continued to possess unconventional weapons in large part because Iraqis were acting as if they did have them.

Like previous chapters, this section begins with a brief description of how the Intelligence Community assessed Baghdad's decisionmaking before the war and then compares that with the ISG's findings. We then describe the Community's lack of creative thinking about Saddam's motives that led to the failure even to consider the possibility that Saddam Hussein had decided to abandon his banned weapons programs.

The Intelligence Community's Pre-War Assessments

The Intelligence Community's assessments of Saddam's thought processes in the decade before Operation Iraqi Freedom are reflected in two broad lines of analysis: the threats to Saddam's regime and his threat to regional security. Throughout both these areas, one aspect remained relatively constant--the Intelligence Community emphasized repeatedly that it lacked "solid information about the activities and intentions of major players in Iraq" and was, in the words of one senior intelligence official, "flying blind" on the subject. ⁶⁵⁵

Regime stability and decisionmaking. The Intelligence Community early on identified sanctions as a significant threat to Saddam's regime, but never assessed whether Saddam might address that threat by destroying his WMD. Immediately after the Gulf War, for example, the Intelligence Community prepared a Special National Intelligence Estimate assessing Saddam's prospects for survival in power. ⁶⁵⁶ That assessment noted that economic vulnerabilities presented a threat to Saddam's regime and that the "lifting of sanctions...would provide relief to the regime and would strengthen Saddam's prospects for survival." ⁶⁵⁷ The Special Estimate therefore assessed that Saddam would concentrate on getting sanctions eased or removed. ⁶⁵⁸

Through the mid-1990s, the Intelligence Community continued to judge that the sanctions were a threat to the regime, but that Saddam "probably believe[d]" he could "outlast" them. ⁶⁵⁹ For example, in December 1993, the Intelligence Community produced another NIE on Saddam's prospects for survival, judging that the United Nations sanctions were "Saddam's Achilles' heel" because of their debilitating effect on the Iraqi economy. ⁶⁶⁰ The NIE did not consider the possibility that Iraq would actually comply with United Nations resolutions. In fact, the Estimate identified as one of the assumptions underlying the analysis that "Saddam Husayn would not fully comply with U.N. Resolutions." ⁶⁶¹

By June 1995, as living conditions and the economy continued to decline, the Intelligence Community assessed that Saddam's overall strategy was to seek a lifting of sanctions with the lowest possible level of compliance with UNSCOM's demands for a full accounting of Iraq's WMD programs. ⁶⁶² Laying out Saddam's options, a June 1995 Special Estimate judged that in the short term Saddam was "likely to make a gesture to UNSCOM...by providing limited additional information on Iraq's BW program." ⁶⁶³ If that gesture failed to achieve relief from sanctions within three months, however, Saddam "probably [would] return to a confrontational mode." ⁶⁶⁴ Such a "confrontational mode" included suspending cooperation with UNSCOM, sabotaging or obstructing UNSCOM monitoring, and expelling or taking hostage United Nations personnel. ⁶⁶⁵ In short, the Intelligence Community judged that Saddam would choose confrontation over greater cooperation with the United Nations as a way to end sanctions. ⁶⁶⁶

Throughout the remainder of the 1990s, the Intelligence Community continued to assess that sanctions threatened Saddam's regime, but also that "Saddam [was] determined to maintain elements of his WMD programs and probably calculate[d] he [could] stonewall UNSCOM while wearing down the Security Council's will to maintain sanctions." ⁶⁶⁷ Saddam's success in undermining international support for the sanctions and in repressing internal dissent also gave him greater confidence and resolve. ⁶⁶⁸ But more importantly, the commerce allowed under the Oil-for-Food program fueled international perceptions that sanctions had weakened. ⁶⁶⁹ This weakening, combined with the failure of UNSCOM to "uncover tangible proof of Iraqi concealment of weapons of mass destruction," bolstered domestic and international perceptions of the regime's strength. ⁶⁷⁰

At the same time, by the end of the decade the Community assessed that Saddam "appear[ed] to have made a strategic decision that confrontation would be necessary to gain an end to the sanctions." ⁶⁷¹ Saddam felt "that putting pressure on UNSCOM and the Security Council [was] the only way to achieve his goal of ending sanctions," according to the Intelligence Community, because Saddam did "not intend to fully comply with relevant Security Council resolutions." ⁶⁷²

The Intelligence Community viewed Iraq's behavior vis-à-vis the United Nations inspections during this time against the backdrop of these assessments and of Iraq's history of concealing its WMD programs. ⁶⁷³ Accordingly, the Community judged that Iraq would continue to obstruct inspections "to the degree they believe[d] the inspections [would] undermine the security apparatus or uncover proscribed materials." ⁶⁷⁴ Thus, when Iraq agreed to the resumption of inspections in 2002, the Intelligence Community judged that Iraq did so in part because of confidence in its ability to hide its weapons-related activities. ⁶⁷⁵ The Community also assessed that Saddam was motivated to reengage with the United Nations in order to avoid U.S. military intervention. ⁶⁷⁶ If such delaying tactics failed to divert an attack, Iraq "could make a tactical retreat by acceding to some United Nations and U.S. demands and then reneg[ing] on them at the earliest opportunity." ⁶⁷⁷ Although Iraq had tried to open several back channels to the United States seeking improved relations, the Community viewed these moves as public relations efforts and did not consider as an option the possibility that Iraq would actually comply with United Nations resolutions. ⁶⁷⁸

Still, analysis of Saddam's thinking and motivations remained largely speculative. In addition to the simple lack of information on Saddam's plans and intentions, the nature of Saddam's decisionmaking process, which the Intelligence Community assessed as highly centralized and therefore difficult to penetrate, compounded analysts' difficulties. ⁶⁷⁹ Saddam made "all key policy decisions" with little input from the bureaucracy, and he usually acted quickly and decisively. ⁶⁸⁰ He could also be "impulsive and deceptive" about his decisions. ⁶⁸¹ Moreover, the Intelligence Community judged that Saddam "rule[d] primarily by fear," using his control over the military, security, and intelligence services to "impose his absolute authority and crush resistance." ⁶⁸² Saddam reinforce[d] this control through "prominent members of his Tikriti clan who occup[ied] key leadership positions." ⁶⁸³ As a result, "all major decisions [were] made by Saddam and a few close relatives and associates." ⁶⁸⁴ The Intelligence Community noted that these characteristics of Saddam's leadership style made it very difficult to read his intentions. ⁶⁸⁵

Regional security and decisionmaking. The Intelligence Community assessed that regional supremacy for Iraq remained Saddam Hussein's fundamental goal from 1991 through 2003. ⁶⁸⁶ The Community judged, though, that to achieve that goal Saddam would need to rebuild Iraq's military might--including weapons of mass destruction. ⁶⁸⁷

But, according to the Intelligence Community, Iraq's conventional military capabilities had deteriorated significantly during this time. By 1999, after four more years of sanctions and damage inflicted by U.S. military operations, Saddam's military was "smaller and much less well-equipped than it was on the eve of his 1990 invasion of Kuwait." ⁶⁸⁸ By 2002, the Community assessed that "Iraqi military morale and battlefield cohesion [were] more fragile today than in 1991." ⁶⁸⁹

With respect to WMD capabilities, on the other hand, the Community's assessments that Iraq "retain[ed] residual chemical and biological weapons of mass destruction" remained constant. ⁶⁹⁰ Although cautioning that reading Saddam's intentions was difficult and that "critical factors important in shaping his behavior [we]re largely hidden from us," the Community nonetheless assessed that Saddam was "determined to retain elements of his WMD programs so that he [would] be able to intimidate his neighbors and deter potential adversaries such as Iran, Israel, and the United States." ⁶⁹¹ Given Iraq's history with WMD, its desire for regional dominance, and the weaknesses in its conventional military forces, the Community did not consider the possibility that Saddam would try to achieve such intimidation and deterrence while bluffing about his possession of WMD. ⁶⁹²

Post-War Findings of the Iraq Survey Group

The Iraq Survey Group concluded that Saddam Hussein unilaterally destroyed his WMD stocks in 1991. Saddam apparently concluded that economic sanctions posed such a threat to his regime that, although he valued the possession of WMD, he concluded that he had to focus on sanctions relief before resuming WMD development.

Background. Iraq's successful use of CW to repel human-wave attacks in the Iran-Iraq war had convinced Saddam Hussein of the importance of WMD and it became an "article of faith" for Saddam that WMD and theater ballistic missiles were necessary to secure Iraqi national security. ⁶⁹³ Saddam also believed that Iraq's possession of WMD and Iraq's willingness to use it "contributed substantially to deterring the United States from going to Baghdad in 1991." ⁶⁹⁴

The destruction of WMD. After the Gulf War, however, the United Nations passed resolutions explicitly linking the removal of economic sanctions with Iraq's WMD disarmament. ⁶⁹⁵ Saddam Hussein initially judged that the sanctions would be short-lived, that Iraq could weather them by making a few limited concessions, and that Iraq could successfully hide much of its pre-existing weaponry and documentation. ⁶⁹⁶ Accordingly, Iraq declared to the United Nations part of its ballistic missile and chemical warfare programs, but not its biological or nuclear weapons programs. ⁶⁹⁷ But after initial inspections proved much more thorough and intrusive than Baghdad had expected, Saddam became concerned. In order to prevent discovery of his still-hidden pre-1991 WMD programs, Saddam ordered Hussein Kamil to destroy large numbers of undeclared weapons and related materials in July 1991. ⁶⁹⁸

According to the Iraq Survey Group, Saddam's decision to destroy Iraq's WMD stockpiles in 1991 was likely shared with only a handful of senior Iraqi officials, a decision that would have important and lasting consequences. ⁶⁹⁹ Saddam so dominated the political structure of the Iraqi regime that his strategic policy and intent were synonymous with the regime's strategic policy and intent. ⁷⁰⁰ Moreover, in addition to dominating the regime's decisionmaking, Saddam also maintained secrecy and compartmentalization in his decisions, relying on a few close advisors and family members. ⁷⁰¹ And Saddam's penchant for using violence to ensure loyalty and suppress dissent encouraged a "culture of lying" and discouraged administrative transparency. ⁷⁰² As a result, the ISG concluded that instructions to subordinates were rarely documented and often shrouded in uncertainty. ⁷⁰³ The decision to destroy the WMD stockpiles was therefore confined to a very small group of people at the top of the Ba'ath pyramid.

The sanctions bind. By the mid-1990s, United Nations sanctions were taking a serious toll; removing them therefore became Saddam's first priority, according to the ISG. ⁷⁰⁴ Iraq's failure to document its unilateral destruction of WMD, however, complicated this effort. Also complicating Saddam's goal of sanctions removal was his continuing concern with regional threats to his security. Although he had destroyed his militarily significant WMD stocks, his "perceived requirement to bluff about WMD capabilities made it too dangerous to clearly reveal" Iraq's lack of WMD to the international community, especially Iran. ⁷⁰⁵ Saddam was therefore in a bind, on the one hand wanting to avoid being caught in a violation of United Nations sanctions but, on the other, not wanting his rivals to know of his weakness.

Saddam decided to strike the balance between these competing objectives, according to the ISG, by preserving Iraq's ability to reconstitute his WMD while simultaneously seeking sanctions relief through the appearance of cooperation with the IAEA, UNSCOM, and, later, the United Nations Monitoring Verification and Inspection Commission (UNMOVIC). ⁷⁰⁶ Iraq's behavior under the sanctions reflects that the Iraqis "never got the balance right." ⁷⁰⁷ Though Saddam repeatedly told his ministers not to participate in WMD-related activity, he at the same time was working to preserve the capability eventually to reconstitute his unconventional weapons programs. ⁷⁰⁸ And the Iraqis continued to conceal proscribed materials from United Nations inspectors. ⁷⁰⁹ Moreover, even when there was nothing incriminating to hide, the Iraqis did not fully cooperate with the inspectors, judging that an effective United Nations inspection process would expose Iraq's lack of WMD and therefore expose its vulnerability, especially vis-à-vis Iran. ⁷¹⁰

The regime's decision to disclose long-concealed WMD documents in the wake of Hussein Kamil's defection in 1995 further eroded confidence in the credibility of Iraqi declarations. The ISG concluded that the release of these documents served only to validate UNSCOM concerns that Iraq was still concealing its WMD programs. ⁷¹¹

Suspending cooperation with the United Nations. Angered by the continuing sanctions, inspections, and military attacks such as Operation Desert Fox, Saddam Hussein in a secret meeting in 1998 unilaterally abrogated Iraqi compliance with all United Nations resolutions, though, according to the ISG, it is unclear if anything concrete followed from this decision. ⁷¹² Meanwhile, Iraq continued to take advantage of the Oil-for-Food Program to augment regime revenue streams. Saddam Hussein used much of Iraq's growing reserves of hard currency to invest in Iraq's military-industrial complex, to procure dual-use materials, and to initiate military research and development projects. Sanctions remained in place, however. ⁷¹³

With international scrutiny bearing down on Iraq in late 2002, Saddam Hussein finally revealed to his senior military officials that Iraq had no weapons of mass destruction. ⁷¹⁴ His generals were "surprised" to learn this fact, because Saddam's "boasting" had led many to believe Iraq had some hidden WMD capacity and because Saddam's secretive decisionmaking style fostered uncertainty. ⁷¹⁵ In fact, senior officials were still convinced that Iraq had WMD in March 2003 because Saddam had assured them that if the United States invades, they need only "resist one week" and then Saddam would "take over." ⁷¹⁶

Analysis of the Intelligence Community's Pre-War Assessments

Saddam Hussein's decisionmaking process was, as the Intelligence Community assessed before the war and the Iraq Survey Group confirmed, secretive and highly centralized. And in this sense, the Intelligence Community cannot be faulted for failing to penetrate this process. But we believe the Community is open to criticism for failing to appreciate the full range of Saddam's strategic and tactical decisionmaking options regarding his weapons programs. At the very least, the Community should have considered the possibility that Saddam had halted active pursuit of his WMD programs after 1991.

Saddam and his regime repeatedly insisted that all of Iraq's banned weapons had been destroyed and that there were no active programs to reconstitute the capability. The United Nations inspectors, after 1996, found no conclusive evidence that these claims were wrong. In retrospect, as found by the ISG, it is clear that the stockpiles and programs were not there to be found. The question therefore arises of why the Intelligence Community did not discover that fact before the war, or at least consider the possibility that, however improbably, Saddam was telling the truth.

As discussed above, the Intelligence Community made multiple--and avoidable--errors in concluding "with high confidence" that Saddam retained WMD stockpiles and programs. It is a separate question why the Community failed to conclude affirmatively that he did not have them.

In large part the explanation lies in Saddam's own behavior. He had concealed crucial facts about his WMD efforts. He did repeatedly and continually obstruct the inspectors, to the point, in 1998, of completely terminating cooperation and forcing the inspectors to conclude that they could no longer do their work. When someone acts like he is hiding something, it is hard to entertain the conclusion that he really has nothing to hide.

The failure to conclude that Saddam had abandoned his weapons programs was therefore an understandable one. ⁷¹⁷ And even a human source in Saddam's inner circle, or intercepts of conversations between senior Iraqi leaders, may not have been sufficient for analysts to have concluded that Saddam ordered the destruction of his WMD stockpiles in 1991--and this kind of intelligence is extremely difficult to get. According to Charles Duelfer, the Special Advisor to the Director of Central Intelligence for Iraq's Weapons of Mass Destruction and head of the Iraq Survey Group, only six or seven senior officials were likely privy to Saddam's decision to halt his WMD programs. ⁷¹⁸ Moreover, because of Saddam's secretive and highly centralized decisionmaking process, as well as the "culture of lies" within the Iraqi bureaucracy, even after Saddam informed his senior military leaders in December 2002 that Iraq had no WMD, there was uncertainty among these officers as to the truth, and many senior commanders evidently believed that there were chemical weapons retained for use if conventional defenses failed. ⁷¹⁹

That it would have been very difficult to get such evidence is, however, not the end of the story. Failing to conclude that Saddam had ended his banned weapons programs is one thing--not even considering it as a possibility is another. The Intelligence Community did not even evaluate the possibility that Saddam would destroy his stockpiles and halt work on his nuclear program. The absence of such a discussion within the Intelligence Community is, in our view, indicative of the rut that the Community found itself in throughout the 1990s. Rather than thinking imaginatively, and considering seemingly unlikely and unpopular possibilities, the Intelligence Community instead found itself wedded to a set of assumptions about Iraq, focusing on intelligence reporting that appeared to confirm those assumptions.

Over the course of 12 years the Intelligence Community did not produce a single analytical product that examined the possibility that Saddam Hussein's desire to escape sanctions, fear of being "caught" decisively, or anything else would cause him to destroy his WMD. ⁷²⁰ The National Intelligence Officer for Near East and South Asia noted that such a hypothesis was so far removed from analysts' understanding of Iraq that it would have been very difficult to get such an idea published even as a "red-team" exercise. ⁷²¹ An intellectual culture or atmosphere in which certain ideas were simply too "unrespectable" and out of synch with prevailing policy and analytic perspectives pervaded the Intelligence Community. But much of the conventional wisdom that led analysts to reject even the consideration of this alternative hypothesis was itself based largely on assumptions rather than derived from analysis of hard data. ⁷²² In our view, rather than relying on inherited assumptions, analysts need to test favored hypotheses even more rigorously when the paucity of intelligence forces analysts to rely, not on specific intelligence, but on a country's history, politics, and observed behavior. ⁷²³

Conclusion

Iraq's decision to abandon its unconventional weapons programs while simultaneously hiding this decision was, at the very least, a counterintuitive one. And given the nature of the regime, the Intelligence Community can hardly be blamed for not penetrating Saddam's decisionmaking process. In this light, it is worth noting that Saddam's fellow Arabs (including, evidently, his senior military leadership as well as many of the rest of the world's intelligence agencies and most inspectors) also thought he had retained his weapons programs, thus responding to charges that the Community was projecting Western thinking onto a product of a foreign culture.

What the Intelligence Community can be blamed for, however, is not considering whether Saddam might have taken this counterintuitive route. Community analysts should have been more imaginative in contemplating the range of options from which Saddam might select. While such imaginative analysis would not necessarily or even likely have ultimately led analysts to the right conclusion, serious discussion of it in finished intelligence would have at least warned policymakers of the range of possibilities, a function that is critically important in the inherently uncertain arena of political analysis.

Collection

The task of collecting meaningful intelligence on Iraq's weapons programs was extraordinarily difficult. Iraq's highly effective denial and deception program (which was employed against all methods of U.S. collection), the absence of United Nations inspectors after 1998, and the lack of a U.S. diplomatic presence in-country all contributed to difficulties in gathering data on the Iraqi regime's purported nuclear, biological, and chemical programs. And these difficulties were compounded by the challenge of discerning regime intentions.

Nonetheless, we believe the Intelligence Community could have done better. We had precious little human intelligence, and virtually no useful signals intelligence, on a target that was one of the United States' top intelligence priorities. The preceding sections, which have focused on the Intelligence Community's assessments on particular aspects of Iraq's weapons programs, have tended to reflect shortcomings in what is commonly referred to as "tradecraft"; the focus has been on questions such as whether a critical human source was properly validated, or whether analysts drew unduly sweeping inferences from limited or dubious intelligence. But it should not be forgotten why these tradecraft failures took on such extraordinary importance. They were important because of how little additional information our collection agencies managed to provide on Iraq's weapons programs.

This was a problem the Intelligence Community saw coming. As early as September 1998, the Community recognized its limited collection on Iraq. ⁷²⁵ The National Intelligence Council noted these limits in 1998, the specifics of which cannot be discussed in an unclassified forum. ⁷²⁶ Yet the Intelligence Community was still unwilling--or unable--to take steps necessary to improve its capabilities after late 1998. In short, as one senior policymaker described it, the Intelligence Community after 1998 "was running on fumes," depending on "inference and assumptions rather than hard data." ⁷²⁷

This section examines and assesses the performance of each of the collection disciplines on Iraq's weapons programs.

Analysis

Intelligence analysis is a tricky business. Analysts are often forced to make predictions in the absence of clear evidence--and then are pilloried after twenty-twenty hindsight reveals that they failed to paint a full picture from conflicting and scattered pieces of evidence. As we have seen, assessing the scope of an adversary's nuclear, biological, and chemical weapons programs poses an especially formidable challenge in this regard; extrapolations from past experience and thin streams of reporting are usually necessary.

Even the best analytical practices, therefore, will sometimes result in assessments that later prove inaccurate. But given the difficulties inherent in analyzing WMD programs--and the serious consequences for judging the capabilities and intentions of such programs incorrectly--it is imperative that the analysis on which such judgments are based be as rigorous, thorough, and candid as possible. In the case of Iraq, the analytical community fell short of this standard.

Analysts have indicated that their starting point for evaluating Iraq's WMD programs was Iraq's past. Analysts' assumptions were formed based on Iraq's history of producing CW and BW, its use of CW, its history of effectively concealing its nuclear program before the Gulf War, and the regime's failure to account for its previously declared stockpiles. ⁷⁶⁴ Thus, the analysts operated from the premise that Iraq very likely still possessed CW and BW, was still hiding it from inspectors, and was still seeking to rebuild its nuclear weapons program. The analytical flaw was not that this premise was unreasonable (for it was not); rather, it was that this premise hardened into a presumption, and analysts began to fit the facts to the theory rather than the other way around.

One consequence of this tendency was that analysts effectively shifted the burden of proof, requiring proof that Iraq did not have active WMD programs rather than requiring affirmative proof of their existence. Though the U.S. policy position was that Iraq bore the responsibility to prove that it did not have banned weapons programs, the Intelligence Community's burden of proof should have been more objective. CIA's WINPAC nuclear analysts explained that, given Iraq's history of successful deception regarding the state of its nuclear program and evidence that Iraq was attempting to procure components that could be used in a uranium enrichment program, they could not envision having reached the conclusion that Iraq was not reconstituting its nuclear program. The analysts noted that they could have reached such a conclusion only if they had specific information from a very well-placed, reliable human source. ⁷⁶⁵ By raising the evidentiary burden so high, analysts artificially skewed the analytical process toward confirmation of their original hypothesis--that Iraq had active WMD programs.

Indeed, it appears that in some instances analysts' presumptions were so firm that they simply disregarded evidence that did not support their hypotheses. As we saw in several instances, when confronted with evidence that indicated Iraq did not have WMD, analysts tended to discount such information. Rather than weighing the evidence independently, analysts accepted information that fit the prevailing theory and rejected information that contradicted it. ⁷⁶⁶ While analysts must adopt some frame of reference to interpret the flood of data they see, their baseline assumptions must be flexible enough to permit revision by discordant information. The analysts' frame of reference on Iraq's WMD programs--formed as it was by Iraq's previous use of such weapons, Iraq's continued efforts to conceal its activities, and Iraq's past success at hiding such programs--was so strong, however, that contradictory data was often discounted as likely false.

Analysts' discounting of contradictory information reflected, in part, an awareness of Iraq's sophisticated denial and deception efforts and of Iraq's past success in hiding the extent of its WMD programs. Reacting to that lesson, analysts understandably (if not wholly defensibly) began to view the absence of evidence of WMD as evidence of Iraq's ability to deceive the United States about its existence. For example, both CIA and the National Ground Intelligence Center simply assumed that Iraq's claims that the aluminum tubes were for rockets was a "cover story" designed to deflect attention from Iraq's nuclear program. Similarly, analysts had imagery intelligence from 2001 that contradicted Curveball's information about mobile BW facilities, but analysts believed that this discrepancy was attributable to Iraq's denial and deception capabilities. ⁷⁶⁷

The disciplined use of alternative hypotheses could have helped counter the natural cognitive tendency to force new information into existing paradigms. Alternative hypotheses are particularly important for assessing WMD programs, which can be easily concealed under the guise of dual-use activity. With the aluminum tubes, the "transshipment" activity at ammunition depots, and the development of small UAVs, analysts did not fully consider the alternative (and non-WMD related) explanations. Analysts set aside evidence indicating a reconnaissance mission for the UAVs, and did not fully explore the possibility that the transshipment activity involved only conventional munitions. And with respect to the aluminum tubes, CIA and DIA analysts concluded that the tubes were destined for use in a gas centrifuge largely because they could be used for such a purpose, in the process discounting evidence that the tubes were in many respects better suited for use in rockets. ⁷⁶⁸

The widely recognized need for alternative analysis drives many to propose organizational solutions, such as "red teams" and other formal mechanisms. Indeed, the Intelligence Reform and Terrorism Prevention Act mandates the establishment of such mechanisms to ensure that analysts conduct alternative analysis. Any such organs, the creation of which we encourage, must do more than just "alternative analysis," though. The Community should institute a formal system for competitive--and even explicitly contrarian--analysis. Such groups must be licensed to be troublesome. Further, they must take contrarian positions, not just ones that take a harder line (a flaw with the Team B exercise of the 1970s). ⁷⁶⁹

The Iraq case shows, however, that alternative analysis mechanisms offer, at best, an incomplete solution to the problem. In addition to testing fully-developed judgments with formal red team exercises, analysts must incorporate the discipline of alternative hypotheses into the foundation of their analytical tradecraft, testing and weighing each piece of evidence. It would be unrealistic to "zero-base" every assessment, or to ignore history when forming analytical judgments. But the conventional wisdom must be tested throughout the analytical process to ensure that a position is not adopted without rigorous questioning. We offer a variety of approaches to this problem in Chapter Eight (Analysis) of our report.

Competitive analysis must also take place at the institutional level. In other words, the need for individual analysts to question their hypotheses and challenge the conventional wisdom also applies to the Intelligence Community as a whole, and suggests the need to strengthen competitive analysis among agencies in the Intelligence Community. ⁷⁷⁰

After September 11, the Intelligence Community was criticized for its failure to communicate and share information across agency lines. That failure prevented analysts from "connecting the dots" because information known to one agency was not put together with information known to another. With each agency holding one or two pieces of the puzzle, none could see the whole picture. The logical response, therefore, was to recommend the formation of centers to bring all the relevant information together. The Iraq story, however, presents a different set of problems. As discussed, the strength of the prevailing assumptions about Iraq presented a distinct picture to analysts and pieces of the puzzle that did not fit that picture were either made to fit awkwardly or discarded. The problem, therefore, was not that analysts lacked awareness of what other analysts were thinking; rather, the problem was that most analysts were thinking the same thing.

Strengthening competitive analysis among components of the Intelligence Community could help alleviate that problem. There was of course some competitive analysis on Iraq--the NIE contained dissenting positions from State's Bureau of Intelligence and Research (INR), DOE, and the Air Force. ⁷⁷¹ And those dissenting positions were at least somewhat closer to the truth than the majority position. Although reasonable minds can differ as to how significant the dissents were (at least in the cases of INR and DOE), ⁷⁷² such competitive analyses in general encourage the consideration of alternative views and ensure that those independent views reach policymakers.

The problem of discounting contrary evidence was compounded by inexcusable analytical lapses. One reason that CIA analysts were confident in their conclusion that the aluminum tubes were for use in centrifuges and not rockets was that the "rocket experts" in the Intelligence Community, the National Ground Intelligence Center (NGIC), assessed that the tolerances of the tubes Iraq was seeking were "excessive" for rockets. But NGIC rocket analysts told Commission staff that at the time they made that assessment they were not aware of the tolerances required for the Iraqi Nasser 81 rockets, for the Italian Medusa rocket on which the Nasser 81 was based, or for comparable U.S. rockets. ⁷⁷³ NGIC should have been aware of these facts.

The reasons for this failure of technical analysis were not particularly grand. Rather, analysts in NGIC, used to focusing almost exclusively on Soviet weapons systems, simply did not do their homework in tracking down information about Iraqi and U.S. weapons that would have shed light on the question whether the aluminum tubes could be used in conventional rockets. CIA analysts, for their part, were too quick to see confirmation of their hypothesis--that Iraq would seek to reconstitute its nuclear program at the first opportunity--based on somewhat dubious technical evidence.

A related concern is the problem of layering of analysis: the building of one judgment upon another without carrying forward the uncertainties of the earlier judgments. ⁷⁷⁴ The judgment in the October 2002 NIE that Iraq was reconstituting its weapons programs was built on previous assessments about Iraq's weapons programs. These earlier assessments, however, were based on relatively thin streams of reporting, yet the cumulative level of uncertainty was not reflected in the Key Judgments nor in some of the NIE's discussions. In brief, previous assessments based on uncertain information formed, through repetition, a relatively unquestioned baseline for the analysis in the pre-war assessments.

The NIE's CW assessments offer an example of the phenomenon. The NIE's estimates that Iraq had up to 500 metric tons of chemical weapons were based largely on accounting discrepancies and Iraq's CW production capacity rather than positive evidence. ⁷⁷⁵ Although the NIE conceded that "we have little specific information on Iraq's CW stockpile," it did not make clear that the baseline assumption rested largely on Iraqi accounting discrepancies. Because that baseline assumption was not made clear, the NIE gave the impression of greater certainty about the actual existence and size of stockpiles than was warranted. Similarly, the assessment that "much" of that stockpile was "added in the last year" was based largely on imagery evidence of "transshipment" in the spring of 2002. ⁷⁷⁶ Analysts assessed that Iraq had added to its CW stockpile in the previous year because the level of transshipment activity seen on imagery indicated that "CW is already deployed with the military logistics chain." ⁷⁷⁷ But that assessment in turn rested on whether the activity seen on imagery was CW-related. As the post-war reassessment by NGA concluded, it was not. By building one assessment on top of another without carrying forward the uncertainty from the first layer, the NIE gave the impression of greater certainty about its judgments than was warranted. ⁷⁷⁸

This "layering" phenomenon occurred not only with respect to one line of analysis over time, but it also occurred across analytical lines. For example, a senior CW analyst related that he and other CW analysts had been "drifting" in the direction of concluding that Iraq did not have much of a CW program. The appearance of Curveball's reporting on BW, however, "pushed [CW analysts] the other way." The analyst explained that if Iraq was producing and hiding BW, then it was probably also producing and hiding CW. In other words, "much of the CW confidence was built on the BW confidence." ⁷⁷⁹

Another shortcoming of the pre-war assessments of Iraq's WMD programs was the failure to analyze the state of these programs within the context of Iraq's overall political, social, cultural, and economic situation. ⁷⁸⁰ In short, the Intelligence Community did not sufficiently understand the political dynamics of Saddam Hussein's Iraq, and as a consequence did not understand the political and economic pressures that led to his decision to destroy his WMD stockpiles while continuing to obfuscate about Iraq's possession of WMD.

As the Iraq Survey Group found, Saddam was facing two opposing pressures--the need to get relief from sanctions and the need to project strength at home and abroad. Saddam reacted to these pressures, according to the ISG, by destroying his WMD stockpiles after the Gulf War and focusing on sanctions relief before resuming WMD development. At the same time, Saddam continued to hinder the inspectors and sow confusion about Iraq's WMD programs. ⁷⁸¹

Yet the weapons analysts did not consider how the political situation might have affected Baghdad's decisions regarding its weapons programs. To be sure, it is doubtful that such consideration would have changed the analytical outcome--the regional analysts were also operating under certain assumptions about Saddam's regime, and those assumptions did not allow for the possibility that Saddam would destroy his CW and BW stocks and halt work on his nuclear programs, as the ISG found. But the failure even to consider how the political dynamics in Iraq might have affected Saddam's decisions about his WMD programs was a serious shortcoming that resulted in an incomplete analytical picture. ⁷⁸² The failure by the Intelligence Community to entertain the possibility that Saddam was actually telling the truth also inclined analysts to accept deeply problematic evidence that might have been more rigorously questioned if the Community had actually considered the possibility that Saddam had abandoned his banned programs.

Several related problems contribute to the lack of context in analytical products. One, there is not yet an institutionalized, effective method to exploit open source resources that would have allowed a better understanding of developments in Iraq. Two, analysts are rarely assigned to one substantive account for any length of time (with the exception of INR analysts) and cannot therefore develop the requisite expertise to evaluate contextual influences. (Of course, longevity on one account can exacerbate the problem of over-reliance on past judgments.) And three, the pressure to respond to current intelligence needs as opposed to long-term research efforts degrades the overall level of expertise on all accounts. Given limited analytical resources, the demand for current intelligence suffocates long-term research and therefore largely precludes development of the kind of in-depth knowledge that such research fosters. ⁷⁸³ A related aspect of this problem is the current system of incentives for analysts, which rewards analysts for the quantity of finished intelligence pieces produced, and therefore encourages analysts to focus on current intelligence. CIA's Directorate of Intelligence is exploring ways to provide incentives for long-term research. Also, the Directorate's creation of a Senior Analytical Service to enable analysts to continue at the working-level (instead of moving into management) and still be promoted should help build expertise. We address these and other related issues in Chapter Eight (Analysis).

More generally, the pre-war assessments highlight the importance of correct presentation of material to consumers, particularly regarding the uncertainties of given judgments and how these judgments were made. While finished intelligence needs to offer a bottom line to be useful to the policymaker, it should also clearly spell out how and from what its conclusions were derived. In the case of WMD programs in hard target nations like Iraq, this means that policymakers must be made aware when--as will often necessarily be the case--many of the Community's estimates rely largely on inherently ambiguous indicators such as capabilities assessments, indirect reports of intentions, deductions based on denial and deception efforts associated with suspect WMD sites, and on ambiguous or thin pieces of "confirmatory" evidence. For example, the fact that the evidence for Iraq's biological weapons program relied largely on reporting from a single source, and that the evidence for Iraq's chemical weapons program derived largely from limited signature-based evidence of "transshipment" activity, should have been more transparent.

Such context is largely absent from the daily products provided to senior policymakers, however, and the daily dose of such products may provide a cumulative level of "certainty" that is unwarranted. Moreover, with respect to NIEs, the "confidence measures" used to describe the level of certainty in the judgments are not well-explained or understood. A more detailed description, explanation, and/or display of what those confidence measures mean should be incorporated. And those measurements should be rigorously and consistently applied.

Ironically, the NIE did contain numerous caveats, but their impact was diminished by their presentation. For example, as noted, the NIE stated that "[t]oday we have less direct access and know even less about the current status of Iraq's nuclear program than we did before the Gulf War." ⁷⁸⁴ Yet that caveat came on page 13 of the NIE, after it had twice stated that Iraq was reconstituting its program and could have enough fissile material for a nuclear weapon in the next several years.

The fundamental assumptions and logical premises on which analytical judgments are based should be clearly explained. Analysts noted that the "impending war" influenced their approach to the pre-war assessments of Iraq's WMD programs, particularly the October 2002 NIE. That is, with the knowledge that U.S. troops would soon have to face whatever WMD capabilities Iraq had, analysts adopted more of a worst-case-analysis approach. ⁷⁸⁵ Yet that approach was not identified or explained to the reader of the NIE. By contrast, when the CIA's Counterterrorism Center prepared a paper on possible links between Iraq and al-Qa'ida, it clearly identified the analysis underlying that paper as of the aggressive, "dot-connecting" sort. ⁷⁸⁶

Although too many qualifications can lead to equivocal analysis, when the evidence is equivocal, the conclusion must be as well. This must especially be the case when the results of debate about intelligence data or analysis will influence important policy decisions. Flagging the logical premises and baseline assumptions for the ultimate judgment would produce a better understanding by policymakers of the possible logical weaknesses in the assessment. It also would likely improve the analytic process as well, by forcing analysts themselves to articulate clearly their operative assumptions. Similarly, analysis that relies heavily on a single source, such as on Curveball's reporting and on the presence of Samarra-type trucks to support the conclusions that Iraq had BW and CW, respectively, should be highlighted.

Information Sharing

In addition to illuminating shortcomings in intelligence collection and analysis, our study of Iraq also highlighted a familiar challenge: that of ensuring effective sharing of information. In the Iraq case, the information sharing problem manifested itself in three specific ways: intelligence was not passed (1) from the collectors to the analysts; (2) from the analysts to the collectors; and (3) from foreign liaison services to the Intelligence Community.

The lack of an effective system for information sharing between collectors and analysts is a well-known systemic problem, but one that has proven highly resistant to resolution. Intelligence Community collectors retain a strong institutional bias against sharing operational information with analysts--CIA's Directorate of Operations is often reluctant to share relevant operational information with CIA's Directorate of Intelligence, let alone with the rest of the Community or with policymakers. Similarly, NSA is reluctant to share raw data with anyone outside of NSA. ⁷⁸⁷ Both NSA and the DO have legitimate concerns for the protection of sources and methods, and this concern must be weighed carefully when determining whether, and in what form, to share information across the Community or even across directorates.

Our review of the Intelligence Community's performance on Iraq identified several specific shortcomings in the way that collectors share intelligence with analysts. First, the source descriptions on raw human source reporting often provided insufficient detail and clarity to allow analysts adequate insight into the source's reliability. For example, the CIA report on the alleged uranium deal that was sourced to Ambassador Wilson described him (unhelpfully) as "a contact with excellent access who does not have an established reporting record." ⁷⁸⁸ Source descriptions that provide more explicit information on the context in which the information was obtained can significantly improve analysts' ability to gauge the credibility of that information. In September 2004, the CIA's DO implemented new source descriptions that are designed to provide additional such contextual detail. ⁷⁸⁹ This is an important step in the right direction, but more needs to be done.

Second, with CIA reporting, analysts were often unable to determine whether a series of raw human intelligence reporting came from the same source. For most reporting, there is currently no way to determine from the face of the CIA report whether a series of reports represents one source reporting similar information several times or several different sources independently providing the same information. For obvious reasons, it is important to distinguish corroboration from repetition. The improved source descriptions should help alleviate this problem, as will increased dialogue between collectors and analysts.

Finally, analysts often obtain insufficient insights into the operational details bearing on the reliability of sources. ⁷⁹⁰ Such information sharing is not an end in itself, of course. In the case of Curveball, for example, the DO did share operational information with DI analysts--including information that indicated possible problems with the source's reliability--but analysts' belief in Curveball's information remained unshaken. Increased dialogue, rather than simply sharing traffic, may help bridge these gaps.

It must be acknowledged that sharing operational details presents a great threat to the protection of sources and methods. Accordingly, any information sharing protocol must therefore be carefully tailored. The CIA recently conducted a DI-DO information sharing pilot program, which addressed the operational as well as technical barriers to effective information sharing within CIA. ⁷⁹¹ Such pilot programs, however, are of little use if the recommended protocols are not implemented across the board.

A separate, but related problem is the lack of a mechanism to ensure that information calling into question a prior piece of intelligence is swiftly communicated to those analysts (and policymakers) who received the intelligence. This problem was most acutely demonstrated in the case of the Iraqi National Congress source , in which Defense HUMINT failed to reissue the reporting (either with the fabrication notice or recall notice attached)--a failure that led analysts and senior policymakers to accept the reporting months after it was known to be worthless. Defense HUMINT has taken steps to ensure that fabricated reporting is recalled, and the Director of the CIA is currently working to establish Community-wide standards to ensure that the original reporting, the fabrication notice, and the recalled reporting are electronically linked. It remains to be seen, however, whether the information-technology hurdles involved in linking related reporting can be overcome.

Dissemination

The collection, analysis, and dissemination of finished intelligence is a cycle, and many of the issues related to collection and analysis also affect dissemination of the product. But at least one issue merits separate discussion. The interface between the Intelligence Community and the policymaker--the way that intelligence analysis is conveyed to the consumer--needs reexamination.

As part of its investigation, this Commission was provided access, on a limited basis, to a number of articles from the President's Daily Brief (PDB) relating to Iraq's WMD programs. Although we saw only a limited cross-section of this product, we can make several observations about the art form. In short, many of the same problems that occurred with other intelligence products occurred with the PDBs, only in a magnified manner. For instance, the PDBs often failed to explain, or even signal, the uncertainties underlying their judgments. Information from a known fabricator was used in PDBs, despite the publication of a fabrication notice on that source months earlier. PDB articles discounted information that appeared to contradict the prevailing analytical view by characterizing, without justifications, such information as a "cover story" or purposeful deception. The PDBs attributed information to multiple sources without making clear that the information rested very heavily on only one of those sources. And the titles of PDB articles were sometimes more alarmist than the text would support.

In addition to the problems it shares with other intelligence products, the PDB format presents some unique problems as well. As discussed above, the emphasis on current intelligence can adversely affect the distribution of analytical resources and can reduce the level of expertise needed for contextual analysis. But the focus on current intelligence may also adversely affect the consumers of intelligence. In particular, the daily exposure to current intelligence products such as the PDB may create, over time, a greater perception of certainty about their judgments than is warranted. And the way these products are generated and disseminated may actually skew the way their content is perceived. For example, when senior policymakers are briefed with the President's Daily Brief or a similar product, they often levy follow-up questions on the briefer. The response to those questions is then typically disseminated in the same format. Therefore, if one policymaker has an intense interest in one area and actively seeks follow-up, that questioning can itself generate numerous PDBs or Senior Executive Memoranda. A large volume of reporting on one topic can result, and that large volume may skew the sense among other policymakers as to the topic's importance.

Politicization

Many observers of the Intelligence Community have expressed concern that Intelligence Community judgments concerning Iraq's purported WMD programs may have been warped by inappropriate political pressure. ⁸²⁹ To discuss whether those judgments were "politicized," that term must first be defined.

ACCOUNTABILITY

Numerous failures within the Intelligence Community contributed to the flawed estimates on Iraq. Many of these failures are systemic--flaws in the way the Intelligence Community is managed, organized, and structured. Part Two of this report contains dozens of recommendations for systemic reform based on the lessons learned from Iraq and other case studies. But reform requires more than changing the Community's systems; it also requires accountability.

Individuals. There are unfortunately a number of examples in the Iraq assessments of individuals whose conduct fell short of what the Intelligence Community has a right to expect. Among these is the handling of Curveball's reporting on mobile BW. In late January of 2003, the Secretary of State was engaged in an intense personal effort to explore every flaw in the intelligence he was about to present to the United Nations Security Council. By then, a division in the CIA's Directorate of Operations had spent months pointing out Curveball's flaws with some persistence. Yet the Secretary of State never learned of those doubts.

A number of individuals stood between the two and could have made the connection. Some acknowledge knowing about Curveball's problems but did not understand that he was the key to the entire BW assessment. Others knew how central Curveball was to the BW case but deny knowing about Curveball's problems. Still others--particularly in CIA's WINPAC--were aware of both sides of the issue and did not present the doubts to the Secretary or other policymakers. Finally, the most senior officials of the Agency insist the serious concerns expressed about Curveball's reliability were never conveyed to them--despite assertions to the contrary.

This Commission was not established to adjudicate personal responsibility for the intelligence errors on Iraq. We are not an adjudicatory body, nor did we take testimony under oath. We were not authorized or equipped to assign blame to specific individuals, particularly when there are disputes about critical facts. We are, however, equipped to address the question of organizational accountability.

Organizations. Almost every organization in the Intelligence Community--collectors, analysts, and management--performed poorly on Iraq. But there are differences among the agencies, both in their initial performance and in how they responded when their mistakes became clear. The National Intelligence Council, for example, faltered badly in producing the flawed NIE on Iraq's WMD programs. But it also learned from its errors. It now brings the collection agencies into the NIE process to evaluate their sources, and its recent estimates are more candid about intelligence gaps, weak sources, and divergent viewpoints.

For some organizations, however, problems run deeper. Three agencies made such serious errors, or resisted admitting their errors so stubbornly, that questions may fairly be raised about the fundamental culture or capabilities of the organizations themselves.

1. The performance of the National Ground Intelligence Center (NGIC) in assessing the aluminum tubes was a gross failure. NGIC got completely wrong the question of the tubes' suitability for conventional rockets--a question that is at the core of NGIC's assigned area of expertise. And NGIC was not aware of, and did not pursue, basic information that was critical to its assessments. ⁸⁶¹

2. The Defense HUMINT Service inexcusably failed to recall reporting from a known fabricator, and compounded that error by failing to notice when its discredited reporting crept into Secretary Powell's speech. Defense HUMINT also bears heavy responsibility for the Curveball episode. Defense HUMINT disseminated Curveball's reporting while taking little or no responsibility for checking the accuracy of his reports. In fact, Defense HUMINT still calls itself merely a "conduit" for Curveball's information and resists the idea that it had any real responsibility to vet his veracity. ⁸⁶²

3. CIA's Weapons Intelligence, Nonproliferation, and Arms Control Center (WINPAC) is the Intelligence Community's center for all-source analysis on weapons of mass destruction. As such, it was at the heart of many of the errors discussed earlier, from the mobile BW case to the aluminum tubes. Just as bad, some WINPAC analysts--and WINPAC as an institution--showed great reluctance to correct these errors, even long after they had become obvious. ⁸⁶³ Creating an intelligence center always carries some risk that alternative views will be sacrificed in pursuit of consensus, and we fear that a culture of enforced consensus has infected WINPAC as an organization.

In short, we have doubts that the broad reforms described in Part Two will be enough to change the organizational culture of NGIC, Defense HUMINT, and WINPAC. Yet the cultures of each contributed crucially to the Iraq WMD debacle. We therefore recommend that the Director of National Intelligence give serious consideration to whether each of these organizations should be reconstituted, substantially reorganized, or made subject to detailed oversight.