All U.S. intelligence efforts on proliferation continue to be seen through the prism of the October 2002 NIE on Iraq WMD. The absence of WMD in Iraq was a major factor in the impetus behind the 2004 intelligence legislation, which ostensibly addresses the issue of combating terrorism. Of the two issues—September 11 and Iraq WMD—the Iraq issue is far more serious in terms of the future of the intelligence community. For all of the pre-September 11 warnings about al Qaeda hostility, including the possibility of the use of aircraft, insufficient intelligence existed to act upon and disrupt the plot. Nor, in the pre-attack atmosphere, would it have been possible to implement the types of security steps in place now. The Iraq WMD issue, however, raised serious questions about analytic tradecraft, not only in WMD issues but also across the board. The Senate Intelligence Committee focused on the problem of groupthink, but more serious issues may have been at play:
• The effect of not allowing analysts better insight into the nature of HUMINT sources
• The proper way to pose alternative analytic questions that yield true alternative hypotheses instead of supporting or simply refuting the current one
• The need to rethink the prevalence of denial and deception (see chap. 6)
• The larger estimative process (see chap. 6)
The proliferation issue was then made even more contentious politically by the release in December 2007 of the unclassified KJs of a new NIE on Iran’s nuclear program, which concluded that Iran had halted its nuclear weapons program in 2003, a reversal of the judgments made in a 2005 estimate.
Iraq WMD, like the Cuban Missile Crisis and a few other intelligence experiences, will probably be a touchstone for years to come in debates over intelligence analysis. (Iraq may also have an ironic and dangerous effect on other would-be proliferators. The lesson they may take away from Iraq’s fate could be: Get a nuclear weapon. Iraq, without a weapon, was overrun with impunity, whereas North Korea, which claims to have tested a nuclear weapon, is going to receive aid in exchange for ceasing its nuclear weapons program.)
The role of intelligence in the WMD policy area is fairly obvious: Identify proliferation programs early enough to stop them before they are completed. As former DCI Tenet noted in his memoirs, for proliferation policy to be successful, intelligence must identify and discern the nature of a program before a test occurs, not record the fact of a test, as was the usual case in tracking Soviet weapons developments. Intelligence also targets the clandestine international commerce in some of the specialty items required to manufacture WMD. However, proliferation programs are, by their very nature, covert. Thus, the types of collection that the United States must undertake tend to come from the clandestine side of the intelligence community. The evidence of nascent programs—as well as mature programs—that U.S. intelligence might obtain may be ambiguous. Fuzzy information complicates the ability of policy makers to confront potential proliferators with confidence or to convince other nations that a problem exists. As the exposure of Pakistani A. Q. Khan’s nuclear proliferation network shows, however, doing so is not an impossible task. But it is time-consuming (the effort against Khan went on for years) and sensitive diplomatically. In the case of the Khan network, the sensitivities of Pakistan had to be taken into account, given its support for the war on terrorism. Khan’s activities also confirmed the international nature of nuclear proliferation. His enterprises spanned three continents and may have been involved in more than just the Pakistan and Libyan programs. This points up another intelligence challenge: determining how vast the interconnections are between would-be proliferators and would-be providers. Although the disruption of the Khan network was a major intelligence success, parts of the program could continue to operate without Khan’s guidance.
STOPPING PROLIFERATION. Beyond the problem of amassing convincing intelligence lies the policy question: How can a would-be proliferator be stopped? The preferred means is diplomacy, but the track record in this area is unimpressive. No nation has been talked out of developing nuclear weapons by diplomacy alone. The United States has used its influence, and its leverage as the guarantor of a state’s national security, to pressure a state into desisting from nuclear weapons development. Press accounts allege that the United States used this method with Taiwan in the 1980s. Some other nations—for reasons of their own—decided to abandon nuclear programs. Japan and Sweden chose not to develop programs. Argentina and Brazil agreed bilaterally to abandon their fledging efforts. The white South African government gave up its nuclear weapons and its capabilities on the eve of the black majority’s advent to power. Libya’s admission in 2003 that it had a range of covert WMD programs that it had formerly denied was largely a result of two factors: successful HUMINT that caught shipments going to Libya and Libya’s concerns about potential U.S. actions after the invasion of Iraq. The Libya case was an intelligence and policy success but not a result of diplomacy. Many other states—Iran, Israel, and North Korea—remain unconvinced by U.S. diplomacy. Given the minimal success of moral suasion, some people have argued that the only workable solution is an active nonproliferation policy—intervening to destroy the capability, as both Israel and the 1991 Persian GulfWar allies did with Iraq.
(See box, “Iraq’s Nuclear Program: A Cautionary Tale.”)
The September 2007 Israeli air strike against a presumed nuclear site in Syria underscores these concerns as well as the inherent ambiguities involved. After the raid, Syria denied that it had occurred, although subsequent commercial imagery revealed considerable Syrian efforts to both clean up and mask the site by extensive bulldozing. In April 2008, the United States released its conclusion that North Korea had been assisting Syria in building a plutonium processing plant, and not a peaceful nuclear use plant, at the site. There are several issues at play in this incident. First, if it was a nuclear site, then once again there is the circumstance of unilateral military action being taken as a means of ensuring that the program will be stopped. Second, if there was North Korean assistance to Syria, does this indicate a possible violation of North Korea’s agreement with the United States (and China, Russia, and Japan) to cease nuclear weapons activity or, at a minimum, an effort to circumvent that agreement by exporting part of its program? Third, it raises the specter of yet another clandestine nuclear relationship to be tracked.
IRAQ’S NUCLEAR PROGRAM: A CAUTIONARY TALE
During the 1980s, Iraq was one of the nations whose nuclear weapons program was closely watched by U.S. experts. The existence of a program was not in question; its status was.
On the eve of the 1991 Persian Gulf War, the considered analytical judgment, according to subsequent accounts, was that Iraq was at least five years away from a nuclear capability. After Iraq’s defeat in the war, analysts learned that Iraq had been much closer to success, even though Israel had attacked and destroyed some of its facilities some years earlier
What had gone wrong with U.S. estimates?
Iraq was a closed target, one of the most repressive and heavily policed states in the world. The state’s nature makes collection more difficult, but that is not the answer to the question.
The answer lies in an analytical flaw, namely, mirror imaging To manufacture the fissionable material it required, Iraq chose a method abandoned by the United States in the early days of its own nuclear program after World War II The method works, but it is a very slow and tedious way to produce fissionable material.
For Iraq, however, it was the perfect method, not because it was slow, but because foreign analysts disregarded it. The method allowed Iraq to procure materials that were more difficult to associate with a nuclear weapons program, to mask its status. A program of this sort was also more difficult for Western analysts to spot because they largely dismissed the approach out of hand, assuming that Iraq would want—just as the United States and others had—to find the fastest way to produce fissionable material.
In the course of U.S. military action in Iraq that commenced in 2003, expected Iraqi weapons of mass destruction programs were not found Some wondered if analysts had compensated for their earlier error by overinterpreting evidence of a possible program without considering alternative interpretations. The analysts themselves denied this assessment, and none of the postwar investigations of the intelligence community’s performance found overinterpretation to have been a factor.
Even without the possible Syria connection, the 2007 nuclear agreement with North Korea poses other intelligence issues. Under the terms of this agreement, North Korea will seal and eventually dismantle its Yongbyon nuclear facility and account for its nuclear activities. Although the Yongbyon facility can be verified easily by imagery, there will be no definitive way to know if North Korea has accounted for all of its nuclear activity. As with arms control agreements with the Soviet Union, an assessment must be made between the gains made by the parts of the agreement that can be monitored and verified with high confidence and the uncertainties faced by those activities about which the intelligence monitoring confidence will be less certain. It is important to note that in arms control intelligence parlance, “high confidence” means a certainty of around 90 percent. This is high but it still leaves open a 10 percent chance of some activities going unnoticed.
Pakistan’s nuclear weaponry has increased concerns about the stability of the Pakistani government. Two factors are at play: the fractious internal politics of Pakistan and the internal political effects of Pakistan’s cooperation with the United States against Muslim terrorists, including the presumed presence of bin Ladin and other al Qaeda leaders in western Pakistan, where the government has virtually no authority. According to press accounts, the United States has given Pakistan technical equipment and assistance designed to help safeguard the security of Pakistan’s nuclear arsenal, although this effort has been made more difficult by Pakistan’s reluctance to provide details about the nature and location of its weapons.
The loose nukes aspect of the issue adds a new and more difficult complication. The Soviet Union agreed with the goal of nuclear nonproliferation, recognizing that it could be a target of would-be proliferators. But far more daunting is the prospect of tracking unknown quantities of weapons-grade material (which even Russian and other authorities have been unable to account for with accuracy) and the international movement of experts from former Soviet states. The collapse of the post-Soviet economy and the end of the privileged status that scientists once enjoyed are incentives to would-be proliferators.
CW and BW proliferation require much less expertise and technical capability than nuclear proliferation does. CW and BW weapons are far less accurate than nuclear weapons, but the random terror they portend is part of their appeal to nations and terrorists. Such programs are more difficult than nuclear programs to identify and track. The anthrax scare of late 2001 underscores these points and also indicates how difficult it is to detect this type of attack in advance or to stop it once under way.
The intelligence experience in WMD is mixed. In Iraq, the analysis did not bear out. The exposure of A.Q. Khan’s network points out the importance of years of determined analysis and highly successful operations to penetrate the network until enough intelligence had been established to make the case incontrovertible. The Libyan surrender also owes much to years of collection, analysis, and some highly successful operations. Reactions to the 2007 Iran nuclear NIE indicate the continuing controversial nature of proliferation intelligence. As noted, the 2007 NIE reversed the 2005 judgment that Iran was determined to build a nuclear weapon. According to the new NIE, this reversal in Iranian policy came in 2003, within the timeframe of the earlier estimate. Thus, a logical first question would be why this earlier view was held if it postdated the Iranian decision? According to press accounts and background briefings by intelligence officials, the change hinges on new intelligence that called into question earlier judgments. Analysts apparently subjected the new intelligence to intense review to be certain that it was not part of an Iranian deception plan and came away satisfied with the reliability of the new intelligence. Reactions to the new Iran estimate were largely political in nature, some seeing it as a case of the intelligence community “learning the lessons” of Iraq WMD, or expiating themselves for the earlier mistaken estimate, or attempting to prevent President George W. Bush from taking military steps against Iran. Although the first explanation (learning the “lessons” of Iraq) may be true, the other two certainly are not. These seem to be more cases of individuals projecting their own views on to the estimative process. It is worth noting that the sharp reactions to the new Iran estimate prompted the principal deputy DNI, Donald Kerr (2007- ), to issue a statement defending the analytic tradecraft and judgments in the new NIE.
Because of the political controversy surrounding the new judgments, other issues about the Iran nuclear program and the NIE were lost in the noise. First, the new estimate did not appreciably change the estimated timelines by which Iran could achieve a nuclear weapon if it wanted to. Second, there was little discussion about why Iran would have ceased its weapons program in 2003, although a logical conclusion might be Iran’s concerns after the allied invasion of Iraq because of ostensible WMD programs. Third, as Secretary of Defense Robert Gates and Joint Chiefs Chairman Admiral Mike Mullen both noted, there was nothing preventing Iran from reversing its 2003 decision and resuming weapons development. Finally, there is also the possibility that the new estimate is in error. The key estimative judgment in the 2007 NIE was made with “high confidence,” meaning high quality intelligence. But, as the “Estimative Language” text box that accompanies each NIE notes: “A ‘high confidence’ judgment is not a fact or a certainty... and such judgments still carry a risk of being wrong.”