by Steve Coll
—
Three months later, on October 9, more than eight million Afghans poured from their homes to cast ballots for president and dip their fingers in a pot of ink, to confirm their act of civic participation. The ink was supposed to be indelible, to prevent fraud, but it washed off easily. That glitch was not enough to call the election’s legitimacy into serious question or to undermine the festival of national restoration the day of voting seemed to create. Seventy percent of registered voters turned out, more than in American presidential elections. Hamid Karzai prevailed against seventeen competitors, winning 55 percent of the reported vote. It was the first direct presidential election in Afghan history.
When the final tally was confirmed early in November, Karzai appeared in an illuminated garden at the presidential palace. “These votes are for stability,” he said. “We hope with great love and friendship to help” the Afghan people. His magnanimity extended to the United States. Without the backing of the Bush administration and without Khalilzad’s intense partnership during the previous two years the outcome might have been different. When Americans visited from Washington now, Karzai “made frequent reference to his fondness for the U.S. and Afghanistan’s reliability as a partner in the war on terror,” one note taker recorded. He spoke warmly about visiting his brother in Maryland “and his pleasure driving himself (without an entourage and security) and enjoying coffee at Starbucks, as well as his enjoyment of country music in Nashville.” As it turned out, this would be the high point of mutual regard between Karzai and America.26
Karzai and Khalilzad looked for factions of the Taliban and other armed opposition that might be persuaded to reconcile with the government. The election triumph created a new opportunity. Sections of Hizb-i-Islami, Gulbuddin Hekmatyar’s party, had decided to enter politics. (Hekmatyar, who had taken refuge in Iran and then returned to Afghanistan and Pakistan, remained at large, operating an armed wing of his movement.) Hekmatyar’s allies worked out a deal and entered parliament in Kabul.27
Zalmay Khalilzad sought new instructions from the Bush White House about terms for talking to the Taliban. The Afghan interagency group met to write an updated policy. American diplomats had talked with Taliban officials throughout the 1990s. The N.S.C. reviewed that history and now emphasized that any Taliban defector who wanted to avoid being sent to Guantánamo had to renounce Al Qaeda and be cleared of past involvement in terrorism. A few former Taliban did move to Kabul. In May 2005, however, Zalmay Khalilzad left to become ambassador to Iraq. Khalilzad was uniquely suited to defection talks with fellow Afghans. The momentum halted. Barno departed, too, a month later. He felt “there was a lost opportunity” after the election “to bring in larger numbers of the Taliban.” Yet it did not seem a decisive failure at the time. When they left, he and Khalilzad both thought, “The Taliban are on the ropes, they were politically crushed by the election, this whole effort is on a success glide slope.” They both “felt really, really good.”28
The C.I.A., too, underestimated the Taliban’s potential to regenerate. Partly this was because agency analysis concentrated on Al Qaeda. The C.I.A.’s analysts were also guilty in the first years after the Taliban’s fall of a “kind of culture of self censorship,” according to Paul Miller, an analyst at the Directorate of Intelligence, because they were reluctant to deliver bad news to the Bush White House. They feared that if their reporting on the early signs of the Taliban’s revival was interpreted as criticism of Bush administration policy—its skepticism about Afghan nation building, for example—then the White House might stop listening to the C.I.A. Even when the evidence of trouble became harder to ignore, “there was a culture of optimism” about Afghanistan across the administration. Whenever somebody would say, “Things are getting worse,” someone else would point out that the economy was growing, a new constitution was in place, they had just held a successful election, and, in any event, “Iraq is going much, much worse.”29
TWELVE
Digging a Hole in the Ocean
Early in 2006, Ashfaq Kayani flew to Bagram Airfield on an unannounced visit. Kayani’s place as Musharraf’s most trusted and powerful lieutenant was by now ratified. He had been promoted to director-general of Inter-Services Intelligence. The sergeant’s son and muddy-boots career officer now wore civilian suits at the whitewashed, manicured I.S.I. compound in Islamabad. His responsibilities encompassed the tribal areas, domestic politics, Baluch separatism, and I.S.I. platforms for anti-Indian operations such as Nepal, Sri Lanka, Bangladesh, and Thailand. Hardly a week passed without the Americans pressing onto his agenda. As a young officer, Kayani had smoked cigarettes, but he had given them up. The stress of running I.S.I. brought him back to the habit. He now chain-smoked, placing his Dunhills in a stemmed holder, which made him look like an eccentric British actor from the 1950s. Between his mumble and the clouds of smoke enshrouding him, he seemed well suited to the role of spymaster.
The C.I.A. sought to promote greater cooperation between I.S.I. and Afghanistan’s National Directorate of Security. Rich Blee was one of the architects of the effort. He had a great fondness for both countries. He was among the Bush administration’s South Asia hands who retained a hope that I.S.I. and N.D.S. had common interests in containing extremism and promoting stability. One problem was that Pakistan’s generals regarded N.D.S. as a hostile force with a director, Amrullah Saleh, whom they judged to be an ally if not an agent of India. (Saleh regarded himself as an ardent Afghan nationalist and certainly not an agent of any foreign power; he worked with India, sometimes closely, among many other allies of an independent Afghanistan, including the United States.) Blee and others at C.I.A. respected Saleh and felt they could talk with him logically about anything—except I.S.I. Saleh would present a dossier of evidence that asserted something like “Colonel Mohammed is the Quetta Shura’s contact at I.S.I.” The C.I.A. would run it by their I.S.I. counterparts, who would reply, “We have a thousand Colonel Mohammeds.” It was an objective fact that the Pakistani service had arrested and handed over hundreds of foreign Al Qaeda—many more than N.D.S. had collared. Wasn’t that evidence of some good faith?
Saleh, however, had been accumulating files and addresses of Taliban and Al Qaeda leaders hiding in Pakistan. He placed Pashto-speaking agents into the Taliban’s recruiting stream and chronicled changes in their salaries and training. He paid “watchers” to patrol the streets of Quetta and Peshawar, to map the homes, mosques, businesses, and families of exiled Taliban commanders. When N.D.S. caught would-be suicide attackers and assassins inside Afghanistan, the service exploited their cell phone records to trace support networks in Pakistan. Occasionally, Saleh would share with I.S.I. suspect cell phone numbers. Invariably, within a day or two, the numbers went dead, he told colleagues.
Saleh found the C.I.A.’s deference to I.S.I. and Musharraf highly frustrating. At one point during 2006, the United States had decided to build a new undeclared airstrip in Paktia Province, Afghanistan, near the Pakistan border, to facilitate reconnaissance operations. Crews hauled in bulldozers, metal sheeting, and other materials, but when Pakistan discovered the project, its generals demanded that it be shut down. The Americans complied. Because it was too much trouble to haul the equipment out, they bombed their own machinery so the Taliban could not steal it.1
At the Bagram conference, the idea was that Saleh and Kayani would exchange details about Al Qaeda and its allies. Detainees in N.D.S. custody had reported taking instructions in Mansehra, a mountain valley town in western Pakistan. Some even suggested Bin Laden might be hiding there. Saleh briefed Kayani on his intelligence.
“Which house?” the Pakistani spy chief asked.
“You’ll have to do the last one hundred yards yourself,” Saleh answered.
“This is unbelievable,” Kayani said, meaning the N.D.S. reporting was not credible. Saleh said he would offer access to his source if Kayani agreed to work with the C.I.A. on the matter.
&nb
sp; “Are you telling me you are spying in my country?”
“Yes.”
Kayani was furious. “I don’t need to be taught intelligence by someone the age of my son,” he said, in Saleh’s account.
The general went outside for a smoke. (The I.S.I. chief now struggled to last an hour in any meeting if smoking was not allowed.) He returned after he had cooled down. He said he hadn’t meant his comment as an insult. He asked Saleh to invite him to his home in Panjshir. Saleh demurred. He was tired of hearing the Pakistanis always name him in reference to his home province—“that Panjshiri” seemed intended as a slur. He knew how hostile many Afghans were to I.S.I. “Mr. Director,” he told Kayani, “you can’t imagine how sensitive people are in regards to Pakistan.”2
The Taliban were clearly on the march in 2006. The number of security incidents in Afghanistan documented by the United Nations grew tenfold between 2003 and the end of 2006. As the head of N.D.S., Saleh was obligated to investigate the Taliban’s bases in Pakistan. Don’t get caught, Saleh’s colleagues at the C.I.A. urged him.
Saleh decided that spring to conduct a more formal study of the Taliban’s resurgence, to inform Karzai, his cabinet, and allies of Afghanistan, including the Bush administration. He wanted to interview active Taliban field commanders personally. There were few impermeable lines in Afghanistan’s internal conflicts. Saleh traveled to Zabul, Uruzgan, Helmand, Kandahar, and other provincial capitals. His colleagues in regional N.D.S. offices negotiated safe passage agreements with Taliban commanders, who came in to talk with him. This sometimes involved paying the Taliban for their time and insights. Saleh sat with enemy commanders for long hours. His classified paper, completed in May 2006, was titled “Strategy of the Taliban.”3
Saleh regarded Pakistan as an “India-centric country,” one that had never been “Afghanistan-centric.” He concluded, based on the limited circumstantial and hard evidence available, that I.S.I. had made a decision in 2005 to support the Taliban more actively, with cash and other aid, backed by covert subsidies from Saudi Arabia. It was the 1980s and 1990s all over again. The consolidation of Karzai’s government between 2003 and 2005 explained the timing of this Pakistani turn, Saleh judged.
“What made them switch?” he asked. “Parliamentary elections, presidential elections, Afghan consensus [that] we will make the new order work, and the growing, positive relationship of Afghanistan with India.” In essence, Pakistan’s generals feared that Karzai’s legitimacy would steer Afghanistan toward a durable role as an Indian ally, with international backing, Saleh concluded. In a sense, both Pakistan and Afghanistan shared a dilemma: If they assumed the United States would not maintain a strong military commitment in the region for more than a few years, they had to maneuver now to construct alliances for a post–American scenario, recognizing that the region would almost certainly remain riven by the bitter conflict between India and Pakistan.4
Saleh’s study predicted that the Taliban mobilization would intensify, and that by 2009, the guerrillas would be advancing from rural strongholds to threaten major cities like Kandahar. The paper forecast that the Taliban would mount a full-fledged insurgency that would bog down Afghan and international troops. This would turn out to be largely accurate, except that the Taliban drive on southern cities occurred even faster than that. “The pyramid of [the] Afghanistan government’s legitimacy should not be brought down due to our inefficiency in knowing the enemy, knowing ourselves and applying resources efficiently,” Saleh warned.5
Karzai was “extremely, extremely angry” about his findings. He ridiculed the predictions and asked him never again to call the Taliban “an insurgency.”
Saleh told Karzai, “I hope time will prove me wrong. But this is a product of your intelligence service.” It should be understood as an honest forecast based on independent field research.6
—
Condoleezza Rice, then secretary of state, was the first Bush cabinet member to grasp the seriousness of Afghanistan’s deterioration. Although Iraq still overwhelmingly dominated the Bush administration’s national security agenda, in the spring of 2006 Rice commissioned a study of the Afghan war similar to the one Saleh had initiated. She selected David Kilcullen, a former Australian Army officer who had earned a doctoral degree studying guerrilla warfare. He was a stocky, sandy-haired man then in his late thirties, with a bounce in his step and a gift for loose, entertaining expression. Kilcullen was well aware of Amrullah Saleh’s view that “the Pakistanis are on the other side, and they’re running the war” on the Taliban’s behalf. Kilcullen dismissed this at the time as “a convenient excuse for” the N.D.S. chief and Karzai to evade responsibility for their government’s corruption and inability to consolidate authority. His view was not an isolated one at the White House or the State Department. Any endorsement of Amrullah Saleh’s assessment that officers within the I.S.I. commanded and controlled the Taliban’s revival as part of official strategy endorsed by Musharraf remained an unpopular point of view in the Bush administration at the time.7
Kilcullen saw the Pakistani position as mainly one of weakness. He observed the awkward position the Pakistan Army had been forced into in Waziristan. Tribal uprisings had taught the British empire to maintain a light footprint there, to maintain control by providing cash subsidies from the relative safety of Peshawar. Independent Pakistan’s generals were mostly ethnic Punjabis—effectively foreigners when they toured Waziristan. They had internalized Britain’s lessons. Through a system of local political agents, and through I.S.I.’s construction of forward operating bases during the anti-Soviet Afghan war, Pakistan had developed its own Islamism-influenced system of light presence and heavy subsidies, with an implied guarantee of autonomy for local tribes. After 2002, however, the United States had pressured Musharraf to invade Waziristan with conventional, Punjabi-manned army forces. Musharraf had done so partly to assuage the Bush administration and partly to root out particular radical networks that had attempted to assassinate him. When local militants hammered the invading Pakistani troops, they forced the army into lockdown on scattered bases. Punjabi officers had to either fight or negotiate just to drive supplies down local roads. The Taliban and Waziri tribesmen had exposed the pretense of the Pakistan Army’s invincibility.8
This left Musharraf in a complex mood by 2006: emotional about the tactical defeats and Pakistani casualties incurred in Waziristan, under American pressure; resentful of American imperiousness; cautious about further ground fighting in the tribal areas yet willing to at least consider more military action if the United States would equip the Pakistan Army for combat success.
Kilcullen traveled to Pakistan to study the war. His hosts provided him a “wish list” of upgraded defense and intelligence equipment they felt they needed. Some of the equipment Pakistan wanted the United States did not have—attack and transport helicopters, for example, which were in very short supply because of supply chain pressures created by the wars in Iraq and Afghanistan. Other equipment—night-vision goggles and infrared surveillance systems, for example—were judged too sensitive to hand over to Pakistan because they could leak to enemy guerrillas or be used against India.9
Kilcullen landed in Pakistan in May. He took briefings at the C.I.A. station. It remained a bastion of Counterterrorism Center personnel and funding. The station had evolved into an unusual forward interagency fusion center focused substantially on terrorism. There was a large room filled with cubicles occupied by officers from the National Security Agency, who ran signals intercept operations, and the National Geospatial Agency, which controlled satellite photography. The station still staffed traditional Near East/South Asia case officers who reported on local politics and counterproliferation specialists who followed Pakistan’s nuclear program, but the Counterterrorism Center had a great many positions.10
With Musharraf’s support, the C.I.A. had embedded officers on about half a dozen Pakistani military bases in the triba
l areas. Case officers who rotated there found themselves trapped inside Pakistani facilities. Art Keller, a young case officer, rotated by helicopter to a Pakistan Army base in Waziristan during this period. An I.S.I. colonel and two I.S.I. majors were his liaison officers. They told him, “You can’t come off this base because this is a secret base.” Keller learned this was a polite fiction. The local Taliban knew the C.I.A. and American Special Forces were present; the real reason for their confinement was to prevent the Americans from trying to operate unilaterally and because of the genuine safety risks outside the base, where the Pakistan Army was vulnerable on the roads.
Running human agents outside base perimeters to collect intelligence about militants in Waziristan remained treacherous and difficult. Every few weeks another dead body turned up with “American spy” pinned to his chest. Most of the victims had nothing to do with the C.I.A., but case officers did lose agents. The atmosphere was deeply hostile to outsiders. Even Punjabi interpreters warned Keller that, because of their accents, they would themselves be killed if they tried to interview locals in markets or villages about sensitive subjects. That meant an effective reporting agent for the C.I.A. had to have the right vernacular, a plausible reason to be moving around a thinly populated region, and professional competence. They were not easy to find. The best a case officer could hope to do in many cases was to communicate by computer with agents who had access to militants. Recruiting new agents or meeting in person was difficult, and in the confines of Waziristan bases, all but impossible.11
During his study tour, Kilcullen heard the generalized doubts about Pakistan expressed by C.I.A. officers in the field. Yet Bush administration policy remained firmly rooted in partnership with Musharraf. Ryan Crocker, then the U.S. ambassador to Pakistan, feared that ramping up counterterrorism operations in Waziristan would blow back on Musharraf and destabilize Pakistan. And with reason: In March 2006, when a Pakistani Special Operations Task Force raided some Haqqani compounds in North Waziristan, killing Arabs and Chechens, local militants seized Pakistani government buildings in a furious and violent reaction. The operation “stirred up a hornet’s nest that the military was unprepared for,” the U.S. embassy in Islamabad reported. Now Kilcullen was studying how to equip and support the Pakistan Army to carry out more such operations. “Dave, I’m sitting on a powder keg here, and you’re lighting matches,” Crocker told Kilcullen.12