Technical Difficulties: The CIA, Iraq, and the Bomb

Michael A. Levi

Long before the Senate Select Committee on Intelligence released its Iraq report last week, observers had already adopted two modes of critiquing America’s prewar intelligence. Some focused on the CIA’s faulty means of dealing with uncertainty, by which mere suspicions of Iraqi perfidy were transformed into definitive conclusions that Saddam was stockpiling banned arms. Others have emphasized the Bush administration’s alleged exaggeration of intelligence in making its case for war. Last week’s Senate report thoroughly addressed the former issue; another Senate report slated for release in January will address the latter.

Yet lost in this way of thinking about intelligence mistakes is a third sort of failure, albeit a much more mundane one: sheer technical incompetence. And last week’s Senate report revealed that in our assessments of Iraq’s nuclear program, sheer technical incompetence played a rather large role. Analysts made naïve judgments about Iraqi procurement and technology, and compounded their errors with simple yet critical numerical mistakes. Our elite intelligence analysts have some explaining to do; some remedial courses on logic and math wouldn’t hurt either.

The buffoonery revolves around the now-infamous hollow aluminum tubes that the Bush administration and parts of the intelligence community wielded as proof that Saddam was developing nuclear weapons. Iraq had been caught trying to procure 60,000 aluminum tubes that were banned under U.N. sanctions. At the time, intelligence analysts argued that those tubes were well suited to building gas centrifuges for uranium enrichment, and were too sophisticated to be plausibly meant for any alternative application, including, most prominently, artillery rockets. In short, the tubes were clear evidence of a nuclear weapons program.

In the first salvo of the internal intelligence debate, a June 14, 2001 CIA document declared, “The dimensions of the [Iraqi] tubes match those of a publicly available gas centrifuge design from the 1950s, known as the Zippe centrifuge.” That was transparent nonsense. According to the Senate report, “The dimensions of the tubes seized … do not ‘match’ the dimensions of any of Zippe’s centrifuge designs.” A myth, however, had been born. On August 2, 2001, the Defense Intelligence Agency (DIA) compared Iraq’s tubes to Zippe’s, noting that the latter had “a wall thickness of 2.8 mm,” similar to those Iraq had procured. There was only one problem: Zippe’s design, as was publicly known at the time and as the inventor himself would later confirm, used tubes with walls that were roughly 1 mm thick. Yet these incorrect numbers crept into the intelligence picture, showing up in the pivotal October 2002 National Intelligence Estimate (NIE) as evidence that Iraq was seeking to acquire nuclear arms.

In addition to the Zippe design, analysts focused on the Beams design, a 1940s-era centrifuge scheme that no state had ever been able to implement on a significant scale, but which still couldn’t responsibly be ignored. Critics of the Zippe theory had noted that while the diameter of Iraq’s tubes was close to the Zippe design, their length and thickness was not. No problem: With respect to length and thickness, the October NIE asserted, Iraq’s tubes resembled the Beams design instead. This, of course, was a logically useless argument. Imagine you found a living object that weighed 50 pounds, just like a young human being, and was 6-feet tall, just like an older human being. Would you then conclude that you had a (50 pound, 6-foot tall) human being on your hands?

Having (not) made the case that the tubes were well-suited for centrifuges, analysts turned to debunking possible other applications for the tubes. The chief candidate was artillery rockets, but the CIA and DIA were skeptical. To show how ill-suited for this task the tubes would be, the National Ground Intelligence Center (NGIC) produced a chart that compared the dimensions and quality of the tubes to those used in the American Mark-40 and Mark-66 rockets. The dimensions were wildly different, and the quality was superior—evidence, the analysts asserted, that Iraq had in mind something different and more sophisticated than mere rockets.

There was only one problem: While the tubes were different from those used in American rockets, they were identical to those used in rockets Iraq had employed for over a decade, which the United States knew well. They were also similar to Italian “Medusa” rockets, which Iraqi designers had used as inspiration for their own. Meanwhile, engineers outside the CIA addressed the quality argument, noting that it was not unreasonable for a rocket program employing unskilled engineers to demand high-quality tubes. These analysts, according to the Senate report, argued that unskilled engineers, lacking confidence in their own skills, would have tried hard to mimic foreign rocket designs as precisely as possible—which, in turn, would have required high-quality aluminum tubes.

To top things off, the NIE, backed by the CIA and DIA, asserted that Iraqi agents had agreed to pay premium prices for the tubes, suggesting that the Iraqi regime accorded them high value. Yet the Department of Energy (DOE) had earlier contacted a U.S. manufacturer for a quote on the same tubes, and had gotten a price higher than the one Iraq had negotiated.

Indeed, the DOE was the only consistent bright spot in this mess, producing detailed technical analyses that, in retrospect, were right. But their analysts were effectively shut out of the process. (The State Department’s analysts didn’t contribute much new thought, although they at least had the good sense to defer to the DOE.) Asked by the Senate Intelligence Committee why the DOE was not brought in on one particular analysis of the tubes, the CIA’s centrifuge expert replied: “Because we funded it. It was our testing. We were trying to prove some things that we wanted to prove with the testing.” Alas, they succeeded with flying colors.