A Sept. 27, 2009 satellite photograph shows a suspected uranium-enrichment facility near Qom, Iran 156 km (97 miles) southwest of Tehran. (DigitalGlobe/Handout/Reuters)

This post is part of the “Iran and the Nuclear Deal” symposium.

In 2002, the intelligence community produced a flawed estimate of Iraq’s nuclear, biological and chemical weapons capabilities. Intelligence analysts had very little reliable information at their disposal, especially because weapons inspectors had been out of the country for several years. Making matters worse, the George W. Bush administration began to lean on the community to exaggerate the Iraqi threat, and it used intelligence to sell the war to Congress and the public. Despite the patchy and unreliable underlying information, intelligence reports became increasingly assertive about the growing danger posed by Saddam Hussein’s illusory arsenal.

In 2007, the intelligence community produced another controversial National Intelligence Estimate (NIE). This time the topic was Iran’s nuclear program. Unlike the first case, this estimate was prepared under the assumption that it would remain classified, and analysts were surely surprised when then-President Bush ordered its publication. The estimate became the target of intense criticism, especially from Republicans who accused intelligence agencies of undermining the administration’s aggressive posture toward Iran. Former secretary of state Henry Kissinger called it “policy conjecture” masquerading as objective intelligence. Peter Hoekstra, the former House intelligence committee chairman, called it a “piece of trash.”

In reality, the NIE was accurate and prescient. It concluded that Iran had disbanded its organized nuclear weapons research program in 2003. At the same time, it noted that Iran was continuing enrichment work apace and that Iran would have sufficient material for a bomb by 2015 if it chose to enrich its uranium stockpile to weapons grade. This prediction, which was supported in later threat assessments, has been borne out in International Atomic Energy Agency (IAEA) reports and open source analyses. Meanwhile there is no evidence to suggest that Iran had resuscitated its weaponization effort at any point between 2003 and 2007. If the estimate was so naive, as critics would have it, they are at a loss to find proof that it was substantively wrong.

Despite all the criticism, intelligence on Iran’s nuclear program has been a success story. The 2007 NIE made the useful distinction between Iran’s suspended weapons effort and its ongoing enrichment program. It served as the baseline for subsequent analyses, which refined what was known and unknown about Iranian capabilities and intentions. The Director of National Intelligence’s annual threat assessments give a flavor of the evolving view of Iran’s capabilities and intentions. The conclusions incorporate new information about Iran’s nuclear program but do not contradict the bottom line in the original NIE. Iran was building the scientific and technical infrastructure to master the nuclear fuel cycle, but it had not restarted the weapons program.

Nor was the intelligence community surprised by Iran’s revelation of a second enrichment plant called Fordow. On the contrary, it had been surveilling the site for months and perhaps years before Iran started installing equipment for centrifuges in 2009. Intelligence officials have publicly and privately asserted that they were watching closely but were reluctant to come forward until they could make a convincing argument that the facility, buried under a mountain near the city of Qom, was designed to house uranium centrifuges. No subsequent reporting appears to challenge these claims.

President Obama appears impressed by this record. The White House has expressed confidence in the intelligence community’s ability to keep track of Iran, and Obama has a particularly close relationship with CIA Director John Brennan, whom he has backed despite calls for his resignation. All of this suggests that policymakers are using intelligence to help inform their judgment about the nuclear deal and to monitor Iranian compliance in the aftermath. So far, so good.

The problem is that policymakers are also using intelligence for political purposes. Rather than simply letting secret intelligence inform its private discussions, the administration is enlisting it to help sell the nascent nuclear deal with Iran. Last week, for instance, Brennan spoke about the ongoing negotiations at Harvard University. Beyond discussing general issues related to intelligence, he included praise for U.S. policy, arguing that sanctions had badly hurt Iran’s economy and caused Tehran to give away far more than expected. The deal, he said, was “as solid as you can get.” Brennan also took aim at critics, some of whom are “wholly disingenuous” for their claims that the deal provides Iran with a pathway to the bomb.

It is easy to understand the temptation to use intelligence as a public relations vehicle. Individuals tend to believe that private documents are more reliable than public statements, and they associate information quality with secrecy. Thus when leaders use secret intelligence to justify their policy choices, they remind skeptics that they are privy to unique sources and thus deserve the benefit of the doubt. Selectively releasing intelligence also implies that more valuable information remains classified.

But using intelligence in public is dangerous. My research shows that it often pushes the community toward firm conclusions even when the underlying information is open to multiple interpretations. Leaders involved in policy disputes do not benefit from intelligence that betrays uncertainty or doubt. If a gap appears between intelligence conclusions and policy statements, policymakers may pressure intelligence officials to alter the tone and substance of their conclusions. Examples abound. In 1967, Johnson administration officials pressured the CIA to provide optimistic assessments of progress in Vietnam in order to overcome growing opposition to the war. Two years later, the Nixon administration leaned on intelligence to hype the Soviet strategic threat in order to help sell a controversial missile defense program in Congress. In both cases the underlying information was ambiguous and contested inside and outside the intelligence community, but the demands of the public debate meant that policymakers could not tolerate signs of doubt or disagreement. So they removed them.

In addition, using intelligence to win public debates discourages reassessment – even if new information appears that contradicts previous beliefs. Intelligence leaders are reluctant to review their findings after making bold public pronouncements, because doing so would amount to an embarrassing admission of failure. In the months leading up to the 2003 Iraq War, for example, the intelligence community benefited from new information from inspectors as well as new secret sources. Officials were loathe to reassess their earlier findings, however, despite the fact that it was increasingly hard to justify the earlier estimates. The United Nations and IAEA conducted several hundred inspections, but they found no evidence of active unconventional programs or stockpiles of old weapons. Some mid-level CIA officers were desperate to reconsider the NIE and follow new leads, but they were stymied. “It’s time you learn it’s not about intelligence anymore,” one was told. “It’s about regime change.”

Finally, the decision to use intelligence in public may poison intelligence-policy relations over the long-term. Right now the Obama administration and the intelligence community seem to share a common view of Iran’s nuclear program. But their views may diverge, and intelligence leaders may become unwilling to make the kind of unequivocal statements that political leaders crave. If this occurs there may be a falling out that outlasts the current administration. Past intelligence-policy breakdowns have created mutual mistrust and hostility that lingered for years after the fact.

As the administration pushes to complete the Iran deal it should keep these dangers in mind. The expectation that intelligence will be part of the foreign policy debate has already led to surprisingly specific revelations about issues including Syria’s use of chemical weapons and U.S.-Saudi intelligence sharing in Yemen. U.S. policy is somewhat ambivalent on these issues, however, meaning that the risk of politicization is low. In the case of the Iranian nuclear deal there is no ambivalence: the administration is clearly staking itself to a nuclear deal in the face of substantial Senate opposition, and it is using intelligence to help make the case. This is a recipe for politicization. If intelligence conclusions start to drift from policy beliefs the White House will be strongly tempted to bring it back into line.

The administration should also reflect on the reasons that intelligence on Iraq was a disaster while intelligence on Iran was a triumph. Before the war in Iraq, intelligence was buffeted by the demands of an administration that needed to use it to justify the invasion. In 2007, however, there was no expectation among analysts that their work would be aired in public. The result was an estimate that has stood the test of time and subsequent intelligence built on the NIE to form a wide-ranging picture of Iran’s nuclear activities. If the White House continues to use intelligence to sell the Iran deal, it risks sacrificing that record.

Joshua Rovner is the John G. Tower Distinguished Chair in International Politics and National Security at SMU. He is the author of “Fixing the Facts: National Security and the Politics of Intelligence.”