Why Intelligence Fails: Lessons from the Iranian Revolution and the Iraq War (review)

As the Obama administration considers its options for dealing with Iran's nuclear weapons program—policy determinations that will hinge on assessments of the Iranian regime's intentions, capabilities, and political durability—the publication of Robert Jervis's trenchant new book on intelligence failures could not be timelier. Jervis, a distinguished political scientist perhaps best known for his path-breaking work on perceptions and misperceptions in foreign policy decision-making, examines two of the most consequential U.S. intelligence failures in recent history: the flawed assessment by the Central Intelligence Agency (CIA) in 1978 of the Shah of Iran's vulnerability to revolutionary overthrow, and the CIA's mistaken judgment that Saddam Hussein's regime in Iraq had retained active programs of weapons of mass destruction (WMD) after 1991 in contravention of United Nations Security Council resolutions (a judgment often cited in late 2002 and early 2003 by George W. Bush's administration in justifying the decision to go to war against Iraq in 2003). Jervis extends his analysis of these two detailed case studies to draw broader, highly policy-relevant conclusions about the politics and psychology of intelligence analysis and intelligence reform.

The Iran case is based on the recently declassified study the CIA commissioned Jervis to write when he was a scholar-in-residence at the agency in the late 1970s. An insightful and occasionally wry memoir of Jervis's experience as an academic in what was then the CIA's National Foreign Assessment Center (NFAC) precedes his postmortem. When preparing his report, Jervis was surprised by the paucity of CIA resources dedicated to Iran relative to the agency's primary Cold War intelligence target, the Soviet Union. That problem was compounded by the weakness of the CIA's collection effort—a glaring example being the dearth of information about the revolution's leader, Ayatollah Ruhollah Khomeini, beyond what was available to the U.S. embassy and CIA station through newspapers. Reporting in general on the domestic politics of a key U.S. ally was given short shrift. Jervis's report focused on the CIA analytical division's performance in responding to Iran's unfolding domestic crisis from its onset in mid-1977 to its decisive turn in November 1978. An analytical paper prepared by the CIA in August 1978 on the prospects for an orderly succession to the Shah claimed that "Iran is not in a revolutionary or even a 'prerevolutionary' situation" (p. 45).

Jervis notes that U.S. intelligence analysts missed and misunderstood the roles that religion and nationalism (as manifested in anti-Americanism) would play as drivers of revolutionary fervor. His report concluded that "even had observers grasped the depth of popular discontent in Iran,… [t]he idea that one of the world's most powerful monarchs could be overthrown by an unarmed mob of religiously-inspired fanatics was simply incredible" (p. 38). At the heart of this CIA assessment was a flawed inference that verged on tautology. The NFAC analysts expected that the Shah would "crack down" if domestic instability escalated to the point of threatening regime survival; in turn, as events unfolded in the autumn of 1978, the Shah's decision not to employ massive force to quell demonstrations was read by U.S. analysts as an indicator that the Shah had the domestic situation under control. The persisting riddle of the Iranian revolution is why the Shah never resorted to full-scale repression. A key factor unbeknownst to the CIA, Jervis notes, was that the Shah was undergoing treatment during this period for a terminal illness.

Jervis's examination of the Iraq WMD intelligence failure draws on the official British, Australian, and American retrospectives, as well as the myriad journalistic accounts. His assessment challenges the conventional wisdom that intelligence was politicized to make the case for a preventive war to topple Saddam Hussein's regime. Jervis concludes that correctable errors were made and that analysis could have been...