On June 5, the Senate intelligence committee released its long-awaited report on whether President George W. Bush and his top officials knowingly exaggerated or falsified intelligence when making the case for invading Iraq.
But more significant are the report's implications (buried between its lines) for how the next president should change the way that National Intelligence Estimates are written—and, even more, the way that they are read.
The committee examined dozens of public statements about Iraq made by Bush, Vice President Dick Cheney, Secretary of Defense Donald Rumsfeld, and so forth—then compared their claims with what classified intelligence analyses were reporting at the time. All the statements turned out to be wrong. But were the officials simply, and perhaps unwittingly, reciting mistaken intelligence reports? Or were they exaggerating, twisting, or even falsifying what those reports were really saying?
Some of the officials' claims, the committee concludes, were "substantiated by available intelligence information." This was the case for allegations about Iraq's biological weapons facilities, its ballistic-missile programs, and its support for terrorist groups other than al-Qaida.
In several instances, the claims were backed by the intelligence estimate's majority view but were disputed by some of the agencies. (An NIE is a consensus product, put together by the nation's 16 intelligence agencies; if some of the agencies disagree on some point, they often file a dissenting footnote.) This was the case for the claims that Iraq was reconstituting its nuclear-weapons program and that it was building unmanned aerial vehicles for the purpose of dropping biological weapons on Americans and our allies. The Senate report chides the officials for failing, at times willfully, to take these dissents into account.
On a few other issues, officials made claims with great confidence, whereas the intelligence reports expressed considerable uncertainty. This was the case for claims about chemical-weapons production and the prospects of postwar stability.
Finally, several claims had no basis in, or were even contradicted by, the official intelligence reports. These include the claim that Saddam Hussein intended to give weapons of mass destruction to terrorist groups, that he had a partnership with al-Qaida, that he had WMD facilities in deep underground bunkers, and that 9/11 hijacker Mohamed Atta met with Iraqi intelligence officers in Prague in 2001.
It is worth noting that the claims that reflected U.S. intelligence—on biological weapons, ballistic missiles, and support for non-al-Qaida terrorist groups—were, while serious, not the sorts of threats that would rally a nation to war. Meanwhile, the claims that did galvanize support for the invasion—on nuclear weapons and alliances with al-Qaida—either exaggerated or falsified the intelligence of the day.
Another intriguing point, made fleetingly in the Senate report's preface, is that the committee reviewed "only finished analytic intelligence documents"—not "less formal communications between intelligence agencies and other parts of the executive branch." In other words (though the authors don't put it in these terms), the committee once again evaded the key question of whether the White House pressured the Central Intelligence Agency into hardening its October 2002 NIE on Iraq.
Unless this question is addressed, the report is beside the point. Its full, ungainly title is "Report on Whether Public Statements Regarding Iraq by U.S. Government Officials Were Substantiated by Intelligence Information." If those same government officials politicized the intelligence information, then the report only perpetuates the sham. (I am not saying this is the case, only that the committee should have investigated whether it is—should have reviewed those "less formal communications.")
But let's, as promised, be "forward-looking." What lessons can an Obama or McCain administration learn from this report?
On the first and last categories that the report describes—where Bush officials simply recited mistaken intelligence analyses and where they simply lied about them—there isn't much to say beyond "Mistakes were made" and "Don't lie."
But the middle categories—where officials either ignored dissents or hyped guesses into solid fact—are genuinely hard to handle. Almost all NIEs express "high confidence" in some of their judgments, "low confidence" in others. Several are cluttered with footnotes in which one or more of the 16 intelligence agencies dissents from the consensus.
Even with the most objective disposition (and the Bush people didn't have that on Iraq), how should a president handle uncertainties or disputes on intelligence, especially when questions of war and peace are at stake?
First, the intelligence process should be refined so that officials at the top are aware of dissenting footnotes. NIEs are explicitly written for the president, but by the time they're boiled down to "executive summaries"—which, in Bush's White House, sometimes amounted to a page or two—the footnotes have often been removed. The footnotes should be put back in, and if the disputes have implications for policy, the agencies should debate the issue in front of the National Security Council, including the president.
Second, when reading a dissent, pay close attention to which agency wrote it. For instance, the October 2002 NIE stated that Iraq was probably reconstituting its nuclear-weapons program. One of the two dissents to that view was written by the Energy Department's intelligence branch. Most of the other intelligence agencies concluded that Iraq's aluminum * tubes were designed for nuclear-weapons production. The Energy Department's analysts insisted that they weren't at all suitable for nukes, that they were probably artillery tubes. The Energy Department runs the U.S. nuclear-weapons program. If the president or one of his aides had known about this dissent, he might have thought, "Hmmm, maybe these guys know something about aluminum tubes."
Similarly, where the NIE asserted that Iraq's unmanned aerial vehicles could deliver biological weapons, U.S. Air Force Intelligence wrote the dissent, concluding that the vehicles were more likely designed for reconnaissance. If the president had known this, he might have thought, "Maybe the Air Force knows more than these other agencies about aerial vehicles." (Click here for an exception to this rule.)
Third, the next president or his director of national intelligence should commission a study of NIEs in the past decade—specifically of patterns in which agencies have been most consistently right and wrong on what topics. Those that have been most correct should be made the lead agency on all the NIEs on that subject. Those that have been most mistaken should be cleaned out.
A model for this exercise might be a Pentagon memo, written during the Kennedy administration, titled "But Where Did the Missile Gap Go?" John F. Kennedy had run for president charging that the Russians were ahead in missile development. He believed in this "missile gap," as did most experts. Soon after he became president, he realized there was no missile gap. He wanted to know why the NIEs of the late 1950s were wrong. (The last time I looked, the declassified memo was in the archives of the John F. Kennedy Library, specifically Box 298 of the National Security File, in the folder marked "Missile Gap, Feb.-May 1963.")
Intelligence, as they say, is as much art as science. Evidence is sifted through assumptions. People usually find what they're looking for and often ignore or misread what they're not. This is, to some degree, inevitable. The important thing is to be aware of the tendency, to keep questioning premises, and to let informed dissenters have their full say.