Ghost Wars

by Steve Coll

  There was no way to be certain how Taliban troops would react to a U.S. Special Forces raid; any sensible plan had to assume the Taliban would be hostile. A raid in an urban area, therefore, looked highly dangerous. The CIA’s clandestine effort to track bin Laden outside of “KKJ” and snare him in less heavily defended border areas made more sense in theory, but there was no joint planning with the CIA about this possibility. In any event the Pentagon saw huge tactical and political problems if the United States tried to operate on its own anywhere near Pakistan.20

  Clinton, Berger, the National Security Council staff, and Pickering at the State Department all saw Shelton as too cautious, too mired in conventional Pentagon doctrine about logistics and force protection. Pickering saw Shelton’s slide shows about how many thousands of troops would be required to snatch bin Laden as “a standard military position—give us forty-eight months and five divisions. These were gold-plated arguments… . They thought, perhaps with some justification, that the NSC and State wanted to correct every problem with them as cannon fodder.” Clinton pleaded with Shelton after a Cabinet meeting for even a symbolic raid: “You know,” the president told the general, “it would scare the shit out of al Qaeda if suddenly a bunch of black ninjas rappelled out of helicopters into the middle of their camp. It would get us enormous deterrence and show these guys we’re not afraid.” But when Shelton returned with an options briefing, his plans all outlined large deployments and cautioned that there would be scant probability of success.21

  Shelton felt the pressure from Richard Clarke especially. Clarke pressed the Pentagon relentlessly for smaller, stealthier plans to attack bin Laden. Shelton saw the White House counterterrorism chief as “a rabid dog.” He conceded that “you need that in government—you need somebody who won’t take no for an answer.” Still, Shelton and the generals felt Clarke and other White House civilians had “some dumb-ass ideas, not militarily feasible. They read something in a Tom Clancy novel and thought you can ignore distances, you can ignore the time-distance factors.”22

  In Special Forces doctrine the quality of intelligence determines the size of the force required to conduct a raid. The more uncertain the intelligence, the larger the required force. The calculation is as much art as science, but it rests on common sense. If an American Delta Force commando, for instance, is able to watch a target with his own eyes and communicate by secure radio to attacking forces, then a commander can be highly certain about when to launch, and he might feel confident about sending a relatively small force. But if the tactical intelligence is being relayed by non-Americans of uncertain competence or loyalty, and if their intelligence is fragmentary or subject to sudden change—as was the case with the CIA’s reporting about bin Laden in Afghanistan—then a commander should size the attacking force to cope with unpredictable resistance. Shelton felt he had a very hard time convincing the civilians in Clinton’s White House of these plain ideas.23

  Any raid by American forces into Afghanistan would have to launch from the sea and cross either Iranian or Pakistani airspace. The Pentagon had no land-basing arrangements close enough to Afghanistan for a helicopter to make a round-trip. Special Forces helicopters and some specially equipped C-130 support transports could evade Iranian or Pakistani radar, but seaborne helicopter carriers would have to circle in waters off the coast and could not hide. Pakistan and Iran both kept close watch on ships moving in international waters near their shores. Pentagon intelligence had monitored Pakistani communications well enough to know that Pakistan tracked American warships and reported on their positions when they neared Pakistan’s shores. Only submarines could reliably evade such detection. The Pentagon had permanently stationed cruise missile–equipped subs rather than surface ships off Pakistan’s coast in case the president ordered another missile strike against bin Laden. The Pentagon assumed that Pakistan maintained spy networks in Oman and the Persian Gulf to watch American armadas come and go. Shelton also assumed that if Pakistan detected a U.S. raiding mission, it would alert the Taliban; the Taliban would then alert bin Laden, allowing him to escape or prepare an ambush for American forces. The list of catastrophic precedents rang in Shelton’s ears: Desert One, the failed U.S. Special Forces raid in 1980 to rescue American hostages in Tehran; the 1993 disaster in Mogadishu, Somalia (which al Qaeda operatives had helped to carry out); the ambush losses suffered by Soviet special forces in Afghanistan during the late 1980s. Shelton repeatedly cited Desert One to Clinton’s White House aides as a cautionary example. He made an impression. Some of Clinton’s senior aides believed that that failed raid had effectively ended the presidency of the last Democrat in the White House, Jimmy Carter.24

  A generation earlier the CIA had possessed its own sizable covert paramilitary forces—sea, land, and air—which it used to attack problems like this one. The CIA had run a small war in Guatemala, a failed raid at Cuba’s Bay of Pigs, and a secret air war in Laos. The agency’s Special Activities Division retained some paramilitary assets, but the unit was a fraction of its previous size. Its strengths were intelligence collection missions, covert operations with local militia forces, and very small strikes. Some American officials believed it did not possess the airplanes or support facilities to pull off a mission in Afghanistan without help from the Pentagon. Still, it bothered some of Clinton’s aides that the CIA never even suggested using its own forces to go after bin Laden.25

  For their part, officers at the Pentagon and the CIA believed that Clinton, as commander in chief, had failed to make—or to force his Cabinet to make—a firm tactical decision about how best to capture or kill bin Laden and his lieutenants. There were no good options, they all admitted. But the White House fostered dispersed, highly compartmented, isolated operations and planning at the CIA and the Pentagon. Clinton’s policy seemed to involve the pursuit of many policies at once. He did not make clear, for instance, whether his priority was to kill bin Laden with cruise missiles or to mount a lethal capture operation. Clarke’s Counterterrorism Security Group tried to fuse and share intelligence reporting and to seize opportunities for sudden strikes against al Qaeda, but Clinton himself hung back. He goaded Clarke’s efforts with “need to do more”–style notations on the margins of National Security Council memos, but he never insisted on final plans or attack decisions. As a result the CIA Counterterrorist Center attempted to develop both the cruise missile track and a snatch operation using proxy forces, but its officers never collaborated with the Pentagon in a concentrated fashion on either one. Staggering through impeachment, it would have taken an exceptional act of will for the president to push through a decision to attack, given the difficulties of the target and the divisions in his Cabinet. At the White House, Clinton’s National Security Council aides firmly believed that they were the aggressive ones on the al Qaeda case, pursuing every possible avenue to get at bin Laden over calcified resistance or incompetence within the CIA and Pentagon bureaucracies. From the other side of the Potomac, Clinton’s White House often looked undisciplined, unfocused, and uncertain—and the bin Laden planning was no exception.26

  Politics entwined these debates with more threads of doubt. In the context of impeachment and Clinton’s uncomfortable relations with the military, some White House aides suspected that Shelton’s reluctance to attack bin Laden was partially political, that neither he nor other generals were prepared to take risks for a weakened president they did not trust. For his part, Clinton worried about his “personal responsibility to the soldiers and their families,” recalled one of his senior aides. “People underestimate what that’s like.” The worst case would be “a failed mission in which you insert a few hundred Special Forces and they get routed.”27

  They all kept returning to the same issues. Among the most important was the status of the Taliban. By 2000 there were still a few analysts at the State Department’s intelligence bureau who argued for patient engagement with the Taliban. But most of Clinton’s Cabinet now accepted that al Qaeda had hijacked Mullah Omar.
Clinton squeezed the Taliban with economic sanctions, but he also continued to endorse negotiations with them, declared a policy of neutrality in Afghanistan’s war, and resisted entreaties to aid Massoud.

  This divided policy affected internal debates about the cruise missile option. Clarke said to Berger that if the White House openly recognized the Taliban as the enemy, it could take a more flexible approach to cruise missile strikes. Clinton then would no longer require precise, two-source intelligence about bin Laden’s location. Clinton could pursue a bomb-and-pause approach against the Taliban, choosing his targets carefully based on the best available CIA intelligence about bin Laden, but defending the strikes in public as an attack against the Taliban and terrorist infrastructure. The strikes could be tied to the long-standing American demand that the Taliban turn bin Laden and his lieutenants over for trial. If the Taliban refused, the United States could just bomb again, especially when it had strong intelligence about bin Laden, al-Zawahiri, Atef, or other leaders. Shelton recalled that the idea of hitting Taliban infrastructure and leadership targets developed to the point where he was asked to examine the residences of Taliban leaders and places where they worked, and to develop target data “in the event that we wanted to make that decision,” to bomb the Taliban directly.28

  Sandy Berger rejected this proposal for a wider war. The August 1998 cruise missile strikes against al Qaeda had been a political disaster at home and abroad. The repeated firing of cruise missiles at impoverished, long-suffering Afghanistan—without strong intelligence about who would be killed and with the near-certainty of civilian deaths—would only raise bin Laden’s standing in the Islamic world, foster new al Qaeda recruitments, and draw worldwide condemnation of the United States. Pickering agreed with Berger. “We had force in the region and were prepared to use it,” Pickering recalled, if they had a precise fix on bin Laden’s location. “But we were not prepared to fire Tomahawks on a daily basis or to try to use bombing aircraft, crossing Pakistani airspace when, in fact, we didn’t have even the right intelligence or the right predicate to do it.” Sixty-seven Americans had been killed by terrorists during the Clinton presidency, Berger noted pointedly. There was no political context for an American war in Afghanistan. Instead Berger worked on the issues he felt were realistic. After the Millennium near-miss Clarke wrote that it seemed clear that the U.S. campaign against al Qaeda had “not put too much of a dent” in bin Laden’s organization and that “sleeper cells” had formed on American soil. Berger pulled the national security cabinet together on March 10 to endorse new efforts: More support for CIA operations abroad; more attention to foreign terrorist groups at home; and tighter border security. It was a campaign of budget allocations, law tightening, and foreign liaison programs—practical but limited.29

  IT WOULD BE SO MUCH EASIER if Massoud or his allies would just take care of bin Laden themselves. But would he do so even if he had the chance? The White House and the CIA debated Massoud’s motivations. The officers who met the commander in the Panjshir or who had known him previously understood that Massoud was a pious Muslim who saw himself as a global Islamic leader. If he struck out against bin Laden and killed him—or worse, if he bundled him off to the Americans—he would pay a heavy price in the Muslim world. Massoud might be able to defend a decision to kill bin Laden in battle—he was in a war—but to kidnap an Islamic sheikh on behalf of the CIA and to deliver him to a humiliating trial in an American courtroom? That would not do much to burnish Massoud’s reputation as an independent-minded guerrilla legend. What was his incentive to take that kind of chance even if such an operation were possible? The CIA could not offer him the prospect of American military or even political support against the Taliban. Shelton and others at the Pentagon were skeptical about even a covert military partnership with Massoud. The Northern Alliance “had its own baggage,” Shelton recalled, “and when you attach the U.S. flag to their formation, and you become a partner with them, then you also become one who can be held accountable for their actions.” Massoud was not a partner that Shelton wanted to embrace.30

  Still, the CIA deepened its intelligence partnership with Massoud’s men during 2000. Some of Clinton’s White House aides figured Massoud would just tell the agency what it wanted to hear, pocket the relatively small amounts of money and equipment on offer, and go about his business as before. But even so, why not give it a try? They had few other plausible options.

  There was an argument about which section of the agency would make additional secret trips into northern Afghanistan. Was Massoud now a Counterterrorist Center account, or did he belong to the Afghanistan section of Near East? The discussions produced a Solomonic decision: Future missions to the Panjshir would be alternated between the Counterterrorist Center’s JAWBREAKER teams and the NALT teams drawn from the Near East Division.31

  Cofer Black flew out to Tajikistan with a team from the bin Laden unit in the early summer of 2000. In tattered Dushanbe, Massoud’s men picked him up in an old Mercedes-Benz which they proudly claimed had belonged once to Najibullah, Afghanistan’s communistera secret police chief and doomed president. They drove the American team to one of Massoud’s safehouses. Inside, with aides and translators, Massoud laid out a battle map and reviewed the Taliban’s positions. As always when he had an American audience, he talked about the broader threat that the Taliban posed to the Islamic world and to the West. He talked about the sufferings endured by the Afghan people under Taliban oppression.32

  Black wanted to solidify their partnership and advance their efforts at shared intelligence collection. He asked if Massoud had any Arab prisoners who could be interrogated. Massoud said he had only a few, none of any value. They had trouble taking prisoners when they fought Brigade 55, bin Laden’s Arab mercenary force. When Massoud’s forces closed in, the Arab soldiers hurriedly gathered in a circle, pulled the pins on their grenades, and committed collective suicide. Massoud became very specific about “problems of the resistance,” as one of his intelligence aides in the meeting recalled, including “the problems of purchasing weapons from Russia,” and the kind of military equipment that the Americans could supply if they wanted to make a difference in the war.

  Massoud “made it very clear to the American side that it was a good time, if they wanted, to somehow punish the Taliban,” his aide remembered. The Taliban were weakening politically, but Massoud’s forces were struggling. The CIA team reported to their colleagues that Massoud portrayed himself as the only anchor, the only force challenging the Taliban. Massoud had asked them for substantial support, they reported to Langley.

  The CIA team said they were arguing on his behalf in interagency councils in Washington. “They were trying to show Mr. Massoud that he had succeeded in finding an audience in the United States,” recalled Massoud’s intelligence aide, “and that his mission and his cause was on the U.S. agenda… . They wanted to tell him that maybe in the future they will assist him.”

  Massoud’s men knew it would be hard for the CIA to keep that promise. The agency’s intelligence aid was helpful, but as a means to change American policy in Afghanistan, the CIA seemed like a limited partner. “Things were going well but very slowly—very slowly,” recalled Abdullah, Massoud’s foreign policy adviser. “I was never of the opinion that we could get big changes” even with the CIA’s help in policy debates. “The system in the United States—it takes dramatic events for things to move.”33

  28

  “Is There Any Policy?”

  PERVEZ MUSHARRAF HOPED to position himself as a modern, even progressive military usurper. He called himself Pakistan’s “chief executive,” appeared publicly in business suits, and issued extravagant promises about reform and democratic restoration. He hired a Washington lobbyist, Lanny J. Davis, who had been Clinton’s mouthpiece during impeachment, to convince the White House of his liberal outlook. But in Islamabad, within the councils of his own army, Musharraf had to establish a new order, and he could not pay a lobbyist to help. He was especia
lly beholden to one general, Mahmoud Ahmed, who had been the frontline commander of the raid into Kargil, reporting directly to Musharraf, and who at the time of the coup had been the commander of the Tenth Corps, the army unit barracked in Rawalpindi and responsible for security in the capital. On that perilous October evening in 1999, as his superior circled uncertainly on a plane above Karachi, Mahmoud (as he was called by his colleagues) rolled a brigade into Islamabad to detain Nawaz Sharif and secure the government for the army. Then, honoring the chain of command, Mahmoud stood aside. All of the Pakistani political elite understood that Musharraf owed his power to Mahmoud’s conduct, and they watched in the first weeks after the coup to see how this debt would be repaid. They did not have to wait long: Musharraf quickly announced that Mahmoud would become the new director-general of ISI. Mahmoud would clean up the mess left by Ziauddin, Sharif’s lackey who now was under house arrest in Lahore.1