by Steve Coll
There was plenty that looked truly dangerous. The CIA pushed European security services, Pakistan, Egypt, Saudi Arabia, and other governments to crack down that autumn on known associates of bin Laden. Cooperation was mixed, but several dozen militants were arrested, including bin Laden’s longtime spokesman in London. Computers and telephone records seized in these cases made plain that al Qaeda’s global cells were metastasizing. The CIA saw a level of lethality, professionalism, and imagination among some of these detained Islamists—particularly among the well-educated Arabs who had settled in Europe—that was on a par with the more sophisticated secular Palestinian terrorist groups of the 1970s. Their connections with one another and with bin Laden often seemed loose. Yet increasingly the Islamist cells were united in determination to carry out the anti-American fatwas that had been issued by bin Laden and al-Zawahiri from Afghanistan.
Within the morass of intelligence lay ominous patterns. One was an interest by bin Laden’s operatives in the use of aircraft. A classified September 1998 threat report warned that in bin Laden’s next strike his operatives might fly an explosive-laden airplane into an American airport and blow it up. Another report that fall, unavailable to the public, highlighted a plot involving aircraft in New York and Washington. In a third case, in November, Turkish authorities broke up a plan by an Islamic extremist group to fly a plane loaded with explosives into the tomb of modern Turkey’s founder, Ataturk, during a ceremony marking the anniversary of his death. Some of these threats against aviation targets were included in classified databases about bin Laden and his followers maintained by the FBI and the CIA.6 There these strands joined the evidence about suicide airplane attacks and aircraft bombings dating back to the 1995 arrest of Ramzi Yousef. Yet at the Counterterrorism Security Group meetings and at the CIA’s Counterterrorist Center there was no special emphasis placed on bin Laden’s threat to civil aviation or on the several exposed plots where his followers had considered turning hijacked airplanes into cruise missiles. Aviation had been a terrorist target for three decades; hijacking threats and even suicide airplane plots had for years been part of the analytical landscape. The threat reports and the pattern of past bin Laden attacks emphasized other target categories more prominently, such as embassies and military bases. If they had any analytical bias, Clinton, Tenet, and Clarke tended to be most worried about weapons of mass destruction because of the casualties and economic damage such an attack might produce. Several classified reports that fall warned that bin Laden was considering a new attack using poisons in food, water, or the air shafts of American embassies. Aviation was an issue but not a priority.7
A second pattern in the threats that fall did galvanize attention: It seemed increasingly obvious that bin Laden planned to attack inside the United States. In September the CIA and the FBI prepared a classified memo for Clinton’s national security cabinet outlining al Qaeda’s American infrastructure, including charities and other groups that sometimes operated as fronts for terrorist activity. In October the intelligence community picked up reports that bin Laden sought to establish an operations cell in the United States by recruiting American Islamists or Arab expatriates. In November came another classified report that a bin Laden cell was seeking to recruit a group of five to seven young men from the United States to travel to the Middle East for training. When the FBI announced a $5 million reward for bin Laden’s capture, the CIA picked up reports that bin Laden had authorized $9 million bounties for the assassination of each of four top CIA officers. Many of the intelligence reports were vague. Still, the pattern was unmistakable. “The intelligence community has strong indications,” declared a December classified memo endorsed by the CIA and circulated at the highest levels of the U.S. government, “that bin Laden intends to conduct or sponsor attacks inside the United States.”8
AFTER THE AFRICA EMBASSY BOMBINGS, National Security Adviser Sandy Berger ordered the Pentagon to station navy ships and two cruise-missile-bearing submarines beneath the Arabian Sea, off the coast of Pakistan. The secret deployment order was so closely held that even some senior directors at the CIA did not know the submarines were in place. Cruise missiles could twist and turn across hundreds of miles as they flew preprogrammed paths to their targets. Their software guided them to coordinates marked by satellites in fixed, stationary orbits. The White House hoped that CIA agents on the ground in Afghanistan—at this time mainly the tribal agents operating near Kandahar—would track bin Laden and relay the coordinates of one of his meeting places or overnight guest houses. Some of the coordinates of bin Laden’s known camps, such as the buildings in the Tarnak Farm compound, had already been loaded into the submarines’ missile computers. Other places could be marked by laser targeting equipment carried by the mobile tracking team, then quickly relayed as GPS coordinates to the submarines. Clinton made it clear to his senior White House aides that if they could produce strong intelligence about bin Laden’s location, he would give the order to strike. Clarke’s counterterrorism office initiated classified exercises with the Pentagon and discovered they could reduce the time from a presidential order to missile impact in Afghanistan to as little as four hours. Still, as they considered a launch, a vexing question remained: How certain did they need to be that bin Laden was really at the target? In a political-military plan he called “Delenda,” from the Latin “to destroy,” Clarke argued that they should move beyond trying to decapitate al Qaeda’s leadership and should instead strike broadly at bin Laden’s infrastructure. The Delenda Plan recommended diplomatic approaches, financial disruption, covert action inside Afghanistan, and sustained military strikes against Taliban and al Qaeda targets. Some of Clarke’s ideas rolled into continuing discussions about how to pressure bin Laden, but none of Clinton’s national security cabinet agreed with his approach to military targeting. Broad strikes in Afghanistan would provide “little benefit, lots of blowback against [a] bomb-happy U.S.,” recalled Deputy National Security Adviser James Steinberg, who like Berger opposed Clarke’s proposal for attacks against al Qaeda camps or Taliban infrastructure, such as it might be. Still, Clinton and his senior aides said they remained ready to fire missiles directly at bin Laden or his most senior leaders if they could be located precisely.9
Late in 1998 the CIA relayed a report to the White House from one of its agents that bin Laden had been tracked to Kandahar. The report was that bin Laden would sleep in the Haji Habash house, part of the governor’s residence complex. The CIA had reported that there were fifty-two Stinger missiles hidden on the grounds. A bomb or cruise missile might kill bin Laden and Omar or other senior Taliban, plus destroy the missiles—a counterterrorism trifecta. “Hit him tonight,” Gary Schroen cabled from Islamabad. “We may not get another chance.” The Cabinet principals on terrorism issues, including Tenet and Richard Clarke, discussed the report.10 Target maps showed that the building where they expected bin Laden to be was near a small mosque. Clinton knew from painful experience that for all the amazing accuracy of cruise missiles, they were far from perfect. When the president had launched missiles against the Iraqi intelligence headquarters in 1993, one missile had fallen a few hundred yards short and had killed one of the most prominent female artists in the Arab world. Clinton had never forgotten that. Now he and others around the table worried that if one of the missiles fell short again, it would destroy the mosque and whoever happened to be inside. Civilian casualties had not been an issue for Clinton during discussions about the August cruise missile strikes, he told a colleague years later, because Clinton felt they had a serious chance in that attack to get bin Laden. Now the prospect of success seemed less certain, Clinton believed. The president said that he would not allow minimizing civilian casualties to become a higher priority than killing or capturing bin Laden, but he wanted to achieve both objectives if possible. In a memo written at the time, Clarke said Clinton had been forced to weigh 50 percent confidence in the CIA’s intelligence against the possibility of as many as three hundred casualties. The t
wo issues—the likelihood of innocent deaths and the uncertainty about bin Laden’s exact whereabouts—often were discussed together in the Small Group after mid-1998.
Tenet said the intelligence he had about bin Laden’s location this time was “single-threaded,” meaning that he lacked a second, independent source. The CIA was searching for confirmation of bin Laden’s presence but didn’t yet have it. As his Delenda memo reflected, Clarke believed that they should fire the missiles anyway. He felt that if they missed bin Laden, Clinton could just declare to the public that he had been targeting Taliban and al Qaeda “infrastructure” and “terrorist training camps” because of continuing threats. Clinton, however, was not enthusiastic about bombing Taliban and al Qaeda camps that Hugh Shelton derided as little more than “jungle gyms” if there was scant expectation that bin Laden or his top lieutenants would be killed. To strike at bin Laden and miss would hurt the United States, Clinton believed.11
Berger told his colleagues that the costs of failure might be very high. Every time the United States shot off one of its expensive missiles at bin Laden and failed to get him, it looked feckless, Berger argued, reinforcing Clinton’s view. As Berger later recalled it: “The judgment was [that] to hit a camp and not get top bin Laden people would have made the United States look weak and bin Laden look strong.”12 Berger did not demand absolute certainty from Tenet or the CIA about bin Laden’s location. The standard he laid down for a decision to strike was a “significant” or “substantial” probability of success. But could the CIA promise even that much?13
Tenet reported back to the group: He did not have a second source. He would not recommend a missile launch. In this judgment he was supported by several of his senior aides at the CIA and the Pentagon’s commanders. The submarines returned to their patrols off Pakistan, still on alert. “I’m sure we’ll regret not acting last night,” wrote Mike Scheuer, the bin Laden unit chief, to Gary Schroen. “We should have done it last night,” Schroen replied. Increasingly, the CIA was chasing a roving spectre.
IN ADDITION TO the submarine order Clinton signed a Top Secret “Memorandum of Notification” within days of the embassy bombings to authorize the CIA or its agents to use lethal force if necessary in an attempt to capture bin Laden, Ayman al-Zawahiri, and several other top lieutenants. Clinton had a specific understanding of bin Laden’s leadership group. He understood al-Zawahiri as someone who was “as smart as bin Laden, not quite as charismatic, but equally ruthless.” The squat Egyptian doctor remained fixed in Clinton’s mind as a participant in the conspiracy that assassinated Anwar Sadat, whom Clinton saw as a rare progressive in the Middle East. His memo provided legal authority for CIA covert operations aimed at taking a specific list of al Qaeda leaders into custody for purposes of returning them to the United States for trial on federal charges of terrorism and murder.14
The MON, as it was called, added new specificity to a previously approved CIA covert action program. The agency already had legal authority to disrupt and arrest terrorists under the 1986 presidential finding that established its Counterterrorist Center. A new finding would trigger all sorts of complex bureaucratic, budgetary, and legal steps. It seemed wiser to use a MON to amend the legal authority the center already possessed, to make it more specific.
By 1998 government lawyers had been intimately woven into the American system of spying and covert action. After the Iran-Contra scandal the White House established a new position of chief legal counsel to the president’s national security adviser. This office, headed at the time of the Africa embassy bombings by Jamie Baker, occupied a suite on the third floor of the Old Executive Office Building, next to the chief White House adviser on intelligence policy. Baker ran a highly secret interagency committee of lawyers that drafted, debated, and approved presidential findings and MONs. They spent long hours on subtle legal issues that arose in America’s lethal covert action programs:When is a targeted killing not an assassination? When is it permissible to shoot a suspect overseas in the course of an attempted arrest?15
Those and similar questions swirled around the CIA’s secret program to track and capture Osama bin Laden in Afghanistan. From Tenet on down, the CIA’s senior managers wanted the White House lawyers to be crystal clear about what was permissible and what was not. They wanted the rules of engagement spelled out in writing and signed by the president so that every CIA officer in the field who ever handed a gun or a map to an Afghan agent could be assured that he was operating legally.16
This was the role of the MON. It was typically about seven or eight pages long, written in the form of a presidential decision memo drafted for Clinton’s signature. The August 1998 memo began with what the lawyers called a “predicate,” or a statement about how bin Laden and his aides had attacked the United States. It also outlined and analyzed possible repercussions of the covert action being planned to arrest them. The MON made clear that the president was aware of the risks he was assuming as he sent the CIA into action. Any covert arrest operations in Afghanistan might go sour, and agents or civilians might be killed. Difficulties might be created for American diplomacy if the operations failed or were exposed. There was also language to address the issue of civilian casualties. Typically this was a boilerplate phrase which in effect urged that “every effort must be taken” by the CIA to avoid such casualties where possible.17
Some of the most sensitive language in the MON concerned the specific authorization to use deadly force. The lawyers had to make clear to the CIA in writing that it was okay to shoot and kill bin Laden’s bodyguards or bin Laden himself as long as the force was employed in self-defense and in the course of a legitimate attempt to make an arrest. “We wanted to make clear to the people in the field that we preferred arrest, but we recognized that that probably wasn’t going to be possible,” Richard Clarke said later. After the Africa bombings the intent of the White House, Clinton’s national security aides insisted later, was to encourage the CIA to carry out an operation, not to riddle the agency with constraints or doubts. Yet Clinton’s aides did not want to write the authorization so that it could be interpreted as an unrestricted license to kill. For one thing, the Justice Department signaled that it would oppose such language if it was brought to Clinton for a signature. Their compromise language, in a succession of bin Laden–focused MONs, always expressed some ambiguity. Typical language might instruct the CIA to “apprehend with lethal force as authorized.”18 Those sorts of abstract phrases had wiggle room in them. Some CIA officers and supervisors read their MONs and worried that if an operation in Afghanistan went bad, they would be accused of having acted outside the memo’s scope.
As time passed, private recriminations grew between the CIA and the White House. It was common among senior National Security Council aides to see the CIA as much too cautious, paralyzed by fears of legal and political risk. They were not alone in this view. Porter Goss, a former CIA officer who had entered Congress and now chaired the House intelligence committee, declared just six weeks after the Africa bombings that the Directorate of Operations had become too “gun-shy.” The CIA’s outgoing inspector general, Fred Hitz, wrote at the same time that the CIA “needs to recapture the esprit de corps it manifested during the height of the Cold War.”19 At Langley this criticism rankled. Midlevel officers noted that they were the ones who had developed the Tarnak snatch operation even before the Africa attacks, only to have it turned down. The CIA’s senior managers felt that Clinton’s White House aides, in particular, wanted to have it both ways. They liked to blame the CIA for its supposed lack of aggression, yet the White House lawyers wrote covert action authorities full of wiggle words. CIA managers had been conditioned by history to read their written findings and MONs literally.Where the words were not clear, they recommended caution to their officers in the field.
The classified legal memos reflected a wider ambiguity in Clinton’s covert policy toward bin Laden that autumn. There was little question at either the National Security Council or the CIA
that under American law it was entirely permissible to kill Osama bin Laden and his top aides, at least after evidence showed they were responsible for the Africa attacks. The ban on assassinations contained in Executive Order 12333 did not apply to military targets, the Office of Legal Counsel in Clinton’s Justice Department had previously ruled in classified opinions.20 Tarnak Farm or other terrorist encampments in Afghanistan were legitimate military targets under this definition, the White House lawyers agreed. In addition, the assassination ban did not apply to attacks carried out in preemptive self-defense where it seemed likely that the target was planning to strike the United States. Clearly bin Laden qualified under this standard as well. Under American law, then, Clinton might have signed MONs that made no reference to seeking bin Laden’s arrest, capture, or rendition for trial. He might have legally authorized the agency to carry out covert action for the sole purpose of killing bin Laden, al-Zawahiri, and other al Qaeda leaders.
But Clinton did not choose this path. Janet Reno, the attorney general, from whom Clinton was somewhat estranged, opposed MONs that would approve pure lethal operations against bin Laden by the CIA. Reno’s position, expressed in Jamie Baker’s top-secret council of lawyers and in other communications with Richard Clarke’s counterterrorism group, was nuanced and complex, according to officials who interacted with the attorney general and her aides. She told the White House that she would approve lethal strikes against bin Laden if the Saudi threatened an imminent attack against the United States. But what was the definition of “imminent”? Clarke argued that the threat reporting about bin Laden made clear that al Qaeda had attacks in motion, but it was impossible to be sure about the timing or location of specific bin Laden operations. Reno accepted that they could not predict specific attacks, but when the strikes that Clarke warned about did not occur right away, Reno sometimes renewed her private objections to broad lethal authority for the CIA.
