Duty: Memoirs of a Secretary at War

CHAPTER 10 Afghanistan: A House Divided

On a crisp, sunny day in October 1986, I stood on a ridgeline in northwestern Pakistan near the Afghan border. I was the deputy director of CIA, and I was visiting a mujahideen training camp, escorted by officials of Pakistan’s Inter-Services Intelligence directorate (ISI). There were thirty to forty fighters, all wearing brand-new parkas, and they were learning to shoot rocket-propelled grenades, using as their target whitewashed rocks on the mountainside in the outline of a Soviet T-72 tank. I assumed everyone there knew CIA was providing the funds for their war, and they were putting on a great show for the man who wrote the checks. The new parkas, the handpicked marksmen, the chilled Pepsis at lunch, and the professions of gratitude for the munitions and other supplies—it was a well-staged snow job. It would not be my last.

Behind Oz’s curtain on the Pakistani frontier that day were harsh realities. I was there principally because the ISI was stalling on providing our new Stinger antiaircraft missiles and other supplies to the Tajiks of the Panjshir Valley and other non-Pashtuns fighting the Soviets in Afghanistan. It was the Pakistanis who decided which mujahideen groups—which warlords—got our weapons. CIA could cajole and exert pressure, but President Muhammad Zia ul-Haq and the ISI were the “deciders.” While we were working closely with Zia to defeat the Soviets, he was at the same time enacting laws that would promote the Islamization of Pakistan and the strengthening of Islamic fundamentalism.

A lot of American weapons, accordingly, went to Afghan Islamic fundamentalists. Our lack of understanding of Afghanistan, its culture, its tribal and ethnic politics, its power brokers, and their relationships, was profound. After becoming secretary of defense twenty years later, I came to realize that in Afghanistan, as in Iraq, having decided to replace the regime, when it came to “with what?,” the American government had no idea what would follow. We had learned virtually nothing about the place in the twenty years since helping defeat the Soviets there.

These experiences—these ghosts—led to my strong conviction, as I stated earlier, that the idea of creating a strong, democratic (as we would define it), more or less honest and effective central government in Afghanistan, to change the culture, to build the economy and transform agriculture, was a fantasy. Our goal, I thought, should be limited to hammering the Taliban and other extremists so as to degrade their military capabilities, and to building up the Afghan army and local security forces to the point where they could keep the extremists under control and deny al Qaeda a future safe haven in Afghanistan. We needed to be thinking in terms of three to five years to accomplish those narrow goals, but we also needed to figure out how to sustain a modest civilian and military presence for many years—as necessary in Afghanistan, I believed, as in Iraq. We couldn’t just walk away again. As I thought about how to achieve those objectives, the memory of 120,000 Soviet troops and more than 15,000 dead Soviet soldiers stuck in my mind. If we had too many foreign troops in country, if there were too many civilian casualties and too little respect for Afghans and what they wanted and what they thought, the Afghans would come to see us, too, as occupiers, not partners. And we would lose, just as the Soviets had. My thinking along these lines was reflected consistently between December 2006 and late 2009 in my public statements, in my congressional testimony, and in my skepticism about adding significantly more troops there.

Prior to Obama’s inauguration, Joe Biden visited Afghanistan and Iraq, as I said. Talking to U.S. diplomats, commanders, and soldiers in Kabul, Biden found confusion at all levels about our strategy and objectives. His previous encounter with Afghan president Karzai, at a dinner in February 2008, had gone badly and ended with the then-chairman of the Senate Foreign Relations Committee throwing down his napkin and walking out on the Afghan president. At dinner with Karzai in January 2009, there was another tempestuous conversation, during which Biden went after the Afghan president on both governance and corruption. One of Biden’s messages to Karzai (and Maliki) was that Obama would not engage with them nearly as often as had Bush. There was concern among Obama’s team that Bush’s frequent videoconferences with both leaders had led to an unhealthy dependence on direct communications with the U.S. president that undercut the ability of Americans in country to do their jobs. I thought there was some validity to the concern, but I was torn. Bush had been a useful mentor for both, and when he raised issues, both leaders knew there was no higher-level appeal. Biden also met with International Security Assistance Force (ISAF) commander General David McKiernan, who made his case for 30,000 more troops, particularly to improve security before the Afghans’ August election. Biden was deeply disturbed about what he had found in Afghanistan. (Things went much better on his visit to Iraq.)

Obama had pledged during the campaign to send more troops to Afghanistan to remedy the inadequate resourcing of the war during the Bush administration, which had begun shifting its focus and priorities to Iraq within months of the fall of the Taliban. I think that all the senior national security officials of the incoming administration shared the view that we were neither winning nor losing in Afghanistan and that we needed to take a hard look at what we were doing there. I had told President Bush in my job interview in November 2006 that I thought our goals in Afghanistan were too expansive, and my concerns had only increased over the ensuing two years. In my January 26 meeting with Obama, I told him that we should have “no grandiose aspirations” in Afghanistan; we just wanted to prevent the country from again becoming a source of threats to us or our allies, as it had been under the Taliban. In a hearing before the Senate Armed Services Committee the next day, I was even more explicit: “If we set out to create [in Afghanistan] a central Asian Valhalla, we will lose. We need to keep our objectives realistic and limited, or we will set ourselves up for failure.”

The new administration’s first NSC meeting on Afghanistan was on January 23. There was much discussion on the lack of a coherent strategy. Petraeus and Mullen both pushed strongly for quick approval of McKiernan’s requested 30,000 troops. I was supportive of some more troops but ambivalent about the number, partly because of the rationale the military was advancing. The additional troops were supposed to blunt the Taliban’s summer offensive and provide security for the August elections, but many of them could not get there in time to do either. I was also still concerned about the size of our military “footprint.” Biden quite logically objected to sending more troops even before we had figured out our strategy.

The president decided to reach outside the government to ask Bruce Riedel, a Middle East expert who had advised his campaign, to lead a sixty-day review of the situation in Afghanistan and recommend changes in strategy. Holbrooke and Defense Undersecretary Flournoy would cochair the effort, with Doug Lute and his staff at the NSC in support. Riedel had been a longtime analyst at CIA and had worked for me. He was one of the best, most realistic Middle East analysts.

The immediate problem facing the president was timing: if we were to get thousands of troops, whatever the exact number, trained, equipped, and into Afghanistan in time to deal with the Taliban’s summer offensive and the elections, we needed a decision before Riedel’s report would be finished, to the chagrin of the vice president and others in the White House. On his return flight from the Munich Security Conference in early February, Biden had told the press that he wasn’t going to let the military “bully” the White House into making decisions about more troops for Afghanistan because of “artificial timelines.” The president had wanted to announce his decisions on the troop drawdowns in Iraq before announcing more troops for Afghanistan, but that wasn’t going to happen either. As the Deputies Committee, chaired by Tom Donilon, parsed the request for 30,000 troops and focused on when they could get to Afghanistan and what they would do, it became clear that the Joint Staff had not worked through how many could get there by summer. The request was eventually pared to about 17,000 additional troops.

This pressure for an early decision on a troop increase in Afghanistan had the unfortunate effect of creating suspicion in the White House that Obama was getting the “bum’s rush” from senior military officers, especially Mullen and Petraeus, to make a big decision prematurely. I believed then—and now—that this distrust was stoked by Biden, with Donilon, Emanuel, and some of Obama’s other advisers joining the chorus, including, ironically, Jim Jones and Doug Lute. The distrust may also have been attributable in part to the lack of experience with military affairs—particularly, in this case, training and logistical timelines—among the senior civilian White House officials from the vice president on down. I believe the military had no ulterior motives: failure to approve at least some troop movements quickly would, in itself, limit the president’s options, rendering him unable to blunt the Taliban summer offensive or add security before the Afghan election. Nonetheless, the suspicion would only fester and grow over time.

Incidents unrelated to Afghanistan worsened it. In late February, for example, Admiral Tim Keating, commander of all U.S. forces in the Pacific, told a press conference about U.S. capabilities to shoot down North Korea’s Tae Po Dong 2 missile and that a prospective launch would be “a stern test” of the new administration. The president was furious at what he called “freelancing” as well as the admiral’s presumption in appearing to judge the president. In his view, Keating’s remarks created serious problems for the administration: if the president ordered the missile shot down, Keating had telegraphed our punch and made non-attribution difficult to sustain; if the president decided not to act, people would wonder why. Mullen and I asked the president if he wanted Keating relieved. Obama said no, that everyone deserved a second chance, but he told me to recall Keating and reprimand him. Keating flew from Hawaii to Washington for a ten-minute meeting with me. I told him of the president’s unhappiness but that we all wanted him to stay—and to learn from the experience. Tim asked me to convey his apologies to the president and tell him this kind of thing would never happen again. And it didn’t (at least with Keating). This episode, along with the president’s problems with the outspoken director of national intelligence, Denny Blair, and increasingly Mike Mullen, showed that presidential irritation with publicity-prone admirals was another source of continuity between the Bush and Obama administrations. All too early in the administration, suspicion and distrust of senior military officers by senior White House officials—including the president and vice president—became a big problem for me as I tried to manage the relationship between the commander in chief and his military leaders.

On February 13, the president chaired an NSC meeting to consider whether to wait until after the Riedel review to decide on more troops, to send 17,000 as soon as possible, to send some troops now and the rest later, or to send the full 30,000 McKiernan had requested. Riedel and all but two of the principals—Biden and Steinberg—supported sending 17,000 at once.

On February 16, in our regular weekly meeting, the president told Mullen and me that he would have preferred to announce the Iraqi drawdown first, as we knew, but that he had decided to authorize the 17,000 troops to help stabilize the situation in Afghanistan and prevent further deterioration. Obama then said to me, “I trust you and your judgment.” The next day the White House announced the decision in a written press release. Although Obama later characterized the decision as the toughest he had made early in his term, he did not bother to announce it in person.

There would later be questions about why so many of the additional troops—Marines—were sent to Helmand province with its sparse population. Their deployment was intended primarily to prevent the security situation in the south from further worsening; that took precedence over providing election security. But an important reason the Marines deployed to Helmand was that while Marine Commandant Jim Conway was eager to get his Marines off their duffs in western Iraq and into the fight in Afghanistan, he also insisted that all the Marines deploy to a single “area of responsibility”—one battlespace—with Marine air cover and logistics. Only Helmand fit Conway’s conditions. The Marines were determined to keep operational control of their forces away from the senior U.S. commander in Kabul and in the hands of a Marine lieutenant general at Central Command in Tampa. The Marines performed with courage, brilliance, and considerable success on the ground, but their higher leadership put their own parochial service concerns above the requirements of the overall Afghan mission. Despite several failed attempts through Pace and Mullen, I did not get this and other command problems in Afghanistan fully fixed until 2010. I should have seized control of the matter well before that. It was my biggest mistake in overseeing the wars in Iraq and Afghanistan.

The other major issue discussed at the February 13 NSC meeting was the timing of the Afghan elections. The Afghan constitution required that the presidential election be held by May 22, 2009, when Karzai’s term would legally end, but the United States and our coalition partners were pressing hard to postpone the election to August 20. Holbrooke argued that a May election could undermine the opposition’s ability to compete and the ISAF’s ability to provide security. The president directed Holbrooke to tell Karzai that he, Obama, was aware of the constitutional problem of going beyond May and that we would work with him to help find a “bridge” to August elections. No one, including me, was indelicate enough to mention that the new administration, dedicated to building “the rule of law” in Afghanistan, had just decided to violate the Afghan constitution and to connive with Karzai on keeping him in power illegally for several months. In its most favorable light, the decision was intended to provide time for other presidential candidates to get organized so there would be a credible election in Afghanistan. For Holbrooke and others at the table, it provided the time necessary to identify a viable alternative to Karzai, who they thought had to go. If the Afghan constitution was an impediment to achieving this goal, the hell with it.

About the same time, Michèle Flournoy returned from her first visit to Afghanistan with some disquieting observations:

She told me that the civilian-military assistance teams—the provincial reconstruction teams—intended to help bring services and better governance to areas the military had cleared of insurgents were “woefully underresourced.” I was dismayed but not surprised by her assessment of deficiencies on the civilian side—after all, I’d been talking about this problem for two years.

I told my staff in early March that I was very disappointed in the Riedel review so far, which contained no new ideas. Among other things, his report called for significantly greater U.S. civilian advisory capacity without offering any concrete proposals as to where it could be found. Flournoy said that the draft report was all about what should be done but the how was missing. There were four options under discussion: (1) “whack-a-mole” counterterrorism—also referred to as “mowing the grass”—and walking away from any other goals; (2) counterterrorism plus some training of the Afghan security forces, cutting deals with warlords, and then getting out as soon as possible; (3) limited counterinsurgency (COIN); and (4) more ambitious COIN, going beyond McKiernan’s request in terms of troop numbers.

During a single week in mid-March, there were three Principals Committee meetings and two sessions with the president on Afghanistan. That Friday we reviewed the final Riedel report, which recommended disrupting the terrorist networks in Afghanistan and especially Pakistan, promoting a more effective government in Afghanistan, developing the Afghan security forces, ending Pakistan’s support for terrorist and insurgent groups, enhancing civilian control in Pakistan, and using U.S. diplomatic, military, and intelligence channels to reduce enmity and distrust between Pakistan and India. It was breathtaking in its ambition. Most significantly in terms of the conflicts to come between the White House and the military, the report stated, “A fully-resourced counterinsurgency campaign will enable us to regain the initiative and defend our vital interests.” All the principals except Biden concurred in the recommendations of the report and also supported full deployment of the 17,000 troops already approved and another 4,000 trainers for the Afghan security forces. Except for the focus on the need to treat Afghanistan in a regional context and, above all, the critical importance of Pakistan to the outcome of the war, the Riedel report had much in common with the review Lute had overseen at the end of the Bush administration. They also had in common the weakness pointed out by Flournoy: far too much attention was paid to what should be done and far too little to how to get it done.

Biden argued throughout the process, and would continue to argue, that the war was politically unsustainable at home. I thought he was wrong and that if the president remained steadfast and played his cards carefully, he could sustain even an unpopular war. Bush had done that with a far more unpopular war in Iraq and with both houses of Congress in the hands of the Democrats. The key was showing that we were being successful militarily, at some point announcing a drawdown of forces, and being able to show that an end was in sight. Nearly two and a half years later, when I left, we still had 100,000 U.S. troops in Afghanistan. Contrary to Biden’s gloomy forecast in early 2009, the president had been able to sustain the effort.

The president embraced most of the Riedel recommendations and announced the elements of his new “AfPak” strategy in a televised speech on March 27 with his senior advisers standing behind him. The goal, he said, would be “to disrupt, dismantle, and defeat al Qaeda in Pakistan and Afghanistan and to prevent their return to either country in the future.” He said that the 17,000 soldiers he already had approved would “take the fight to the Taliban in the south and east, and give us a greater capacity to partner with Afghan security forces and to go after insurgents along the border.” Although he added that they would also help provide security in advance of the Afghan elections, implicit in his remarks was the priority of taking the fight to the Taliban in their heartland. There would now be some 68,000 American troops in Afghanistan. Further, we would increase the training and size of the Afghan security forces.

He also called for a dramatic increase in the U.S. civilian effort—agricultural specialists, educators, engineers, and lawyers—to advance security, opportunity, and justice and to help the Afghan government serve its people and develop an economy not dominated by illicit drugs. This civilian component was central to any political strategy for denying the Taliban influence. He never used the words counterinsurgency or counterterrorism in the speech, but the strategy he announced was clearly a blend of both. Two days after the announcement, I told a television interviewer that I did not think there would be any need to ask the president to approve more troops until we saw how the troops soon to deploy were doing.

I fully supported the president’s decisions although I was deeply skeptical about two fundamental elements of the strategy. Based on our experience in Iraq, I harbored deep doubt that the required number of civilian advisers from State, the Agency for International Development, the Department of Agriculture, and other agencies could be found and deployed. My doubts would prove justified. I also doubted we could persuade the Pakistanis to change their “calculus” and go after the Afghan Taliban and other extremists on their side of the border. When a Pakistani Taliban offensive that spring reached within sixty miles of Islamabad, the Pakistani army went after them in the border provinces of Swat and South Waziristan for their own protection. Their continuing toleration of the Afghan Taliban, including harboring their leaders in Quetta, was a hedging strategy based on their lack of trust in us, given our unwillingness to stay engaged in Afghanistan in the early 1990s. The Obama administration worked hard to alleviate that mistrust, but history was working against us.

My definition of success was much narrower than Riedel’s or the president’s at that point: using military operations—a combination of selective counterinsurgency and counterterrorism—to degrade the Taliban’s capabilities to the point where larger and better trained Afghan security forces could maintain control of the country and prevent the return of al Qaeda. I would take this position for as long as I was secretary. The president’s broad new policy would help accomplish that goal. I had told Petraeus in Iraq that a key to success was recognizing the tipping point—when the Iraqis doing something barely adequately was better than us doing it excellently. I thought the same principle should apply in Afghanistan and, even in the Bush administration, I had called it “Afghan good enough.”

In June 2008, on my recommendation to the president, General Dave McKiernan became the commander of ISAF in Afghanistan, a coalition force of American troops and troops from more than forty other countries. George Casey, Army chief of staff, and Mullen thought he was the right man for the job, and I had a very high opinion of him, in no small part because he had worked so well with our allies in Europe. Nonetheless, by mid-fall, I was openly expressing concern to my immediate staff about whether I had made a mistake. To this day, it is hard for me to put a finger on what exactly it was that concerned me, but my disquiet only grew through the winter. Perhaps more than anything it was two years’ experience in watching generals like Petraeus, McChrystal, Chiarelli, Rod Rodriguez, and others innovate in blending both counterterrorism and counterinsurgency operations, and observing their flexibility in embracing new ideas, their willingness to experiment, and their ability to abandon an idea that didn’t pan out and move on to try something else. McKiernan was a very fine soldier but seemed to lack the flexibility and understanding of the battlespace required for a situation as complex as Afghanistan. Based on his recent background and experience—commanding coalition ground forces during the opening phase of the Iraq War and then leading the U.S. Army in Europe—I wondered if I had put him into a situation that did not play to his strengths.

There were some specific issues. In trying to solve the command and control problem for coalition forces in Afghanistan, Mullen and I agreed that the best alternative was to replicate the structure we had in Iraq—a four-star commander of all forces, McKiernan, with a subordinate three-star commander to manage the war on a day-to-day basis. McKiernan, like McNeill before him, spent a significant amount of time with Karzai and other Afghan officials, coalition ambassadors, and visiting government officials, and on NATO-related issues—diplomatic and political duties. That role was critically important but made apparent the need for someone else who would be totally focused on the fight. McKiernan strongly resisted such a change. I was also concerned that we were not moving fast enough or decisively enough to deal with the problem of civilian casualties. As I said before, I don’t believe any military force ever worked harder to avoid innocent victims, but it seemed like every incident was a strategic defeat, and we needed to take dramatic action. Soon after the president’s March announcement, I told Mullen, “I’ve got kids out there dying, and if I don’t have confidence I have the very best possible commander, I couldn’t live with myself.”

The issue came to a head in early April when Michèle Flournoy returned from Afghanistan and told me of her concern as to whether McKiernan was the best man for the job. The specific issues she raised paralleled my own list. Mullen and Petraeus both agreed a change was needed. Casey argued strenuously against firing Dave, calling it a “rotten” thing to do. He wrote a letter to the president expressing his views, a letter that he shared with me and I personally delivered.

I had talked on several occasions privately with the president about my misgivings and in mid-April told him I thought the time had come to make a change. Mullen, Petraeus, and I would unanimously recommend Lieutenant General Stanley McChrystal to succeed McKiernan. The president understood the potential for a political ruckus caused by firing the senior commander in the war, but he was willing to make the change.

Relieving McKiernan of command was one of the hardest decisions I ever made. He had made no egregious mistake and was deeply respected throughout the Army. Mullen had been talking with him about what was in the air for a few weeks and, in the latter part of April, flew to Afghanistan to try to persuade him to step down of his own accord. Dave made clear he wanted to remain in place until the end of his tour in the spring of 2010. I couldn’t wait that long. I flew to Kabul on May 6 and went almost immediately into a private dinner with Dave, telling him why I wanted to make a change so quickly. He acceded with extraordinary dignity and class.

I would learn only later that this was the first time a wartime commander had been relieved since Truman fired Douglas MacArthur in 1951. During World War II, Generals George Marshall and Dwight Eisenhower routinely fired commanders, many of them perfectly capable officers, including several personal friends. General Matthew Ridgway did much the same in Korea before and after taking over from MacArthur. The act was common enough not to be a career-ender or blight on the reputation of the affected general or the Army itself. But by the time of the Vietnam War, it was practically unheard of in the Army. I hope that the McKiernan episode will contribute to reestablishing accountability for senior officers for wartime performance, including the precedent that personal misconduct or serious mistakes need not be required for relief.

On May 11, I announced that McKiernan was being relieved and that I would recommend McChrystal to take his place as senior commander. My senior military assistant, Rodriguez, would become the deputy commander in charge of the day-to-day fight. A reporter asked what McKiernan had done wrong. I said absolutely nothing, that a new strategy required a new commander. When asked why McChrystal was the replacement, I said that he and Rodriguez together brought a unique skill set in both counterterrorism and counterinsurgency.

My usual practice with senior military appointments under both Bush and Obama was to take the officer in for a brief photo op with the president just before my weekly meeting in the Oval Office. My main purpose was to have the president congratulate the appointee and offer his support and confidence. The president knew of McChrystal’s extraordinary success in running counterterrorism operations in Iraq and Afghanistan. In official settings, alas, Stan was not relaxed or casual, to say the least. When he met Obama, and the president cracked a little joke, McChrystal remained rigid and unsmiling. After Stan left, Obama smiled and said, “He’s very… focused.”

Even before McChrystal was confirmed by the Senate, I was hearing from Mullen and others about the need for more troops—the president’s recent approval of 21,000 more notwithstanding. There was still an outstanding request from McKiernan for another 10,000, among other things. A little over a week after my meeting in Kabul with McKiernan, I attended a deployment orders meeting about staffing Rodriguez’s new operational headquarters. (Each week I would meet with the chairman and vice chairman, along with my staff and the Joint Staff, to approve “requests for forces” from commanders all over the world—which units, how many troops, and so on.) I was told that this new headquarters in Kabul would require several thousand more troops, perhaps going well above the presidentially approved number of 68,000. I was surprised by the request and told the group we could not go above the approved number without going back to the president. Once McChrystal was in Afghanistan, he would have to evaluate just how all those people were being used—for example, were there soldiers and Marines doing infrastructure construction or maintenance who could be reassigned to the new headquarters or to the fight?

Meanwhile there weren’t enough U.S. civilian advisers and experts, described as so critical to success in the Riedel report. On May 2, Petraeus and Holbrooke cochaired a civilian-military “coordination conference” with representatives from a number of government agencies. The embassy had asked for 421 more people, but Holbrooke asked for an “unconstrained” reanalysis of the civilian requirement down to the provincial and district levels. Holbrooke obviously shared my skepticism about State’s ability to expeditiously field a significant number of civilians to Afghanistan. I told the State Department and National Security Staff that I was prepared to provide several hundred civilian experts from Defense and from the military reserve to fill vacancies.

At a Deputies Committee meeting in late May, Tom Donilon reaffirmed the importance of the civilian component and expressed considerable impatience with the size and pace of the civilian surge. Despite the failure of State and others to deliver the needed number of civilian experts, Deputy Secretary of State Jack Lew (in charge of administration) and others did not take advantage of my offer to loan them (and pay for) Defense civilians to fill gaps. The most common response I heard was that our people weren’t an exact fit for the open positions. I felt that if a Defense civilian had half the skills or background State was looking for, that would give us 50 percent more civilian capacity than we currently had. On a related issue, I was concerned that a high percentage of the U.S. civilians in Afghanistan were stationed in Kabul, when the greatest need was in the provinces and districts where our military was attempting to clean out the Taliban. They stayed a year—with a number of weeks of vacation time—and nearly all turned over in the summer, often leaving gaps in civilian capability for months and sometimes indefinitely. The numbers and location of civilian experts would remain a source of frustration among our commanders and the rest of us at Defense. (Many of the civilians eventually sent by State did not possess the required skills either; they and too many other civilians spent their entire Afghan tour holed up in the fortified embassy compound.)

I thought there was another potential source of civilian expertise available. I knew of the international outreach programs of most U.S. land-grant colleges and universities, particularly in fields including agriculture, livestock, veterinary medicine, and water resources. Research and practitioner faculty regularly traveled to developing countries and, working in primitive and often dangerous circumstances, made a huge contribution. Repeatedly, I urged Holbrooke and AID officials to reach out to the national president of the land-grant university association to seek his assistance in enlisting help from some of these schools. Unlike many government employees, they would expect and want to be deployed to the countryside to help. The president of that association was Peter McPherson, former president of Michigan State and head of the Agency for International Development from 1981 to 1987 under President Reagan, and I was confident he would make every effort to enlist help from the universities. Like my offer of Defense civilians, though, nothing came of the idea. There wasn’t interest at State or AID.

Illustrative of another problem in getting civilian experts into Afghanistan, Secretary of Agriculture Tom Vilsack volunteered to send dozens of specialists to Afghanistan but told me he had no money to cover the cost. Could Defense do so? I had to tell him we could not because of congressional restrictions on transfers of funds among government departments. We could make such transfers only to the State Department.

On June 8, I met with McChrystal, Rodriguez, Mullen, Cartwright, and Flournoy to continue discussions on the new command structure. We needed to take this step-by-step, I cautioned, because of coalition sensitivities, so we would begin with Rodriguez solely as a deputy commander for ISAF through the fall and then see about additionally “double-hatting” him as deputy commander of U.S. forces after New Year’s. After I said we needed a better approach to dealing with civilian casualties, I told Stan I wanted him to do a sixty-day review of the situation in Afghanistan, reviewing the personnel we already had and might need. It seemed like a perfectly reasonable, indeed innocuous, request at the time. I said we needed to do the review before I approached the president about any more forces because I couldn’t nickel-and-dime him to death. Finally, I warned him that “I feel strongly that too big a footprint [too many U.S. forces] is strategically dangerous.” The president and I would rue the day I asked for that review.

The next day was my worst so far with the Obama administration. A meeting with the president began with his approval of our plans to implement the command and control changes in Afghanistan, including Rodriguez’s headquarters. I assured him we would come up with detailed plans, get them approved in the interagency process, and then take them to the allies. I then described my request to McChrystal for a sixty-day assessment, including a review of troop levels and newly identified troop needs through the end of the year—things we had not anticipated. I would then come to the president ready to justify any further increase in troop numbers and would not ask again in 2009. The room exploded. The president said testily there would be no political support for any further troop increase—the Democrats on the Hill didn’t want one, and the Republicans would just play politics. He recounted how getting approval of the FY2009 supplemental had been harder than they imagined possible. Biden and Emanuel piled on. I was aware of Biden’s conviction—and probably that of others in the room—that this request and the McChrystal assessment were part of an orchestrated squeeze play by the military to get the president to approve a lot more troops. I described my own reservations about a big increase in troop numbers but didn’t see why two to four thousand more troops should cause so much angst and hostility.

I left the meeting discouraged less about the skepticism regarding more troops than about the total focus on the politics. Biden was especially emphatic about the reaction of the Democratic base. (His remarks reminded me of Cheney’s focus on the Republican base when discussing detainee interrogations and Guantánamo.) Not a word was mentioned about doing whatever it took to achieve the goals the president had so recently set or to protect the troops. The president and his advisers all emphasized that before any more troops could be considered, we would have to show success and a change of momentum with the troops we had. I was stunned. The Democrats controlled both houses of Congress, and the White House was running scared. The skepticism I could understand; the politics I couldn’t.

McChrystal was confirmed as commander—and authorized for a fourth star—by the Senate on the same day as my ugly meeting in the White House. My earlier strategy of getting him confirmed as director of the Joint Staff and taking care of any potential Senate issues at that time paid off.

Unfortunately, by summer, the Obama foreign policy team was splintering. Biden, his staff, Emanuel, some of the National Security Staff, and probably all of the president’s White House political advisers were on a different page with respect to Afghanistan than Clinton, Mullen, Blair, and me. The same people were, to repeat, increasingly suspicious that Mullen, Petraeus, McChrystal, and other senior military officers were trying to box in—“jam”—the president and force him to approve even more troops. Donilon, Denis McDonough, and others were saying openly to people in Defense that “the White House” was not happy with Mullen’s performance as chairman “and never have been,” and they complained about his frequent interviews on television, even though they were often the ones who would ask him to go on the talk shows. McDonough bellyached to Geoff Morrell about how McChrystal’s forthcoming sixty-day assessment would be a “turd,” and he went on to say the president shouldn’t hear Mullen’s views for the first time in the papers and on TV. Rumors about Jones’s isolation in the White House and potential exit were rife. Biden, Donilon, and Lute were increasingly at odds with Holbrooke. And as mentioned earlier, the Panetta-Blair relationship had tanked. Jones told me, on his return from a trip to Afghanistan at the end of June, that he had warned Stan that any further request for troops would provoke a “Whiskey Tango Foxtrot” response from the president—military parlance for “What the fuck?” The potshotting and rumormongering by late summer created a volatile atmosphere for considering McChrystal’s report.

I had my own growing list of Whiskey Tango Foxtrot issues with the president and others in the administration by this time. After my weekly meeting (accompanied by Vice Chairman Cartwright) with the president on July 15, he asked to see me alone, an increasingly common occurrence. He then dropped a bombshell on me: he intended to meet with General Cartwright privately to ask him if he would stick around and succeed Mike Mullen as chairman of the Joint Chiefs. I was concerned that any such conversation and arrangement would leak, rendering Mike a lame duck for more than two years—unless, of course, the president’s intention was to hasten his departure. I knew nearly everyone at the White House preferred Cartwright’s briefing style, which was much crisper than Mullen’s. Cartwright could also explain highly technical matters clearly, and his analytical style meshed better with the president’s own. But Mullen’s high public profile and his independence grated as well. I pleaded with the president not to meet with Cartwright before September or October, until after he and Mullen had both been reconfirmed by the Senate in their positions for another term. When that meeting took place, I suggested the president simply ask Cartwright to remain for the full two years of his second term. (I had proposed his retirement after the first year of his second term only to stagger the chairman’s and vice chairman’s terms.) I had no problem with the idea of Cartwright succeeding Mullen as chairman at that point.

In early August, I had a long, very direct conversation with Rahm Emanuel in his White House corner office about a list of issues. I had been in that office under many previous chiefs of staff, and the décor remained essentially unchanged, with little that personalized it beyond a few family photos on a credenza behind the desk. In his shirtsleeves, Rahm greeted me cordially as always and offered me a Diet Coke. Ignoring the more formal sofa and chairs in front of his fireplace, we sat down at his conference table. The first issue I raised was a decision by Attorney General Eric Holder that the Justice Department would not defend six Navy petty officers who had been guards at Guantánamo and were being sued by a prisoner there. The sailors had done nothing wrong, but Holder did not want to have to defend the constitutionality of holding prisoners at Guantánamo. The Justice Department had told the sailors the government would pay for their defense by private lawyers, but as I told Rahm, that was not the same as having the full weight of the U.S. government on your side in a courtroom. Everyone in uniform knew that, and it upset them. I told Rahm the president’s approach to the military from the day he was elected had been pitch-perfect, but that this decision could strongly and negatively affect military morale and attitudes toward the commander in chief. I also complained that Holder had made this decision without any consultation with me. I told him in language I knew he’d understand that a decision by Justice not to defend innocent American service members was a travesty and a “huge fucking mistake.”

I also told Emanuel I was ticked off that Deputy National Security Adviser John Brennan had told the president that additional Reaper drone caps in Afghanistan should be transferred from the military to CIA, without me knowing anything about it. Those were Defense Department assets, I said, and no one in the White House had any business going to the president with such a recommendation without going through the established interagency process. This was part and parcel of an increasingly operational National Security Staff in the White House and micromanagement of military matters—a combination that had proven disastrous in the past. I told Rahm, “I’m a team player, but I’m not a patsy.”

Perhaps most important, I told Emanuel that the president needed to “take ownership of the Afghan War,” both for the troops and for our allies. My principal concern was not with his public comments on the need for an exit strategy but rather with what he wasn’t saying. He needed to acknowledge that the war could take years but that he was confident we would ultimately be successful. He needed to say publicly why the troops’ sacrifices were necessary. I told Emanuel I would likely come to the president in mid-September for additional “enablers”—more troops for counter-IED, ordnance disposal, route clearance, intelligence, surveillance, reconnaissance, and medics—but that I would try to delay any request for combat units until January. I said that the president did not want to be in the position of turning down assets that had a direct role in protecting the troops’ lives. I wasn’t trying to “jam” the president; I knew from experience that with the increase to 68,000 troops, more of these support capabilities would be needed.

Emanuel sat patiently and quietly while I vented; then he said he would see what he could do. Although I never raised my voice, I think he could tell how angry I was and chose to exercise uncharacteristic restraint. I would later hear that some of the politicos in the White House worried about my quitting.

In a June 24 videoconference, McChrystal told me for the first time that he had found the situation in Afghanistan much worse than he expected. In the south, he said, insurgents controlled five of thirteen districts in Helmand province, Kandahar was under pressure, and much of the region was “not under our control.” The Afghan forces in the south were at only about 70 percent of authorized strength, and there was a big retention problem. In the east, the Haqqani network was expanding its operational reach, “but our guys have a pretty good handle on the situation there.” Overall, he said, governance was very bad and creating a lot of problems: “There is no legitimacy.” When I asked him if he had enough ISR, his answer provoked the only smiles in the session: “Sir, I am genetically predisposed to never say I have enough.”

I first heard that McChrystal was going to ask for a lot more troops from Mike Mullen immediately upon his return from a trip to Afghanistan in mid-July. Mullen said McChrystal might ask for as many as 40,000 additional troops. I nearly fell off my chair. Questions flooded my mind: Why? What for? Did he really believe the president would approve that massive an increase so soon after agreeing to an additional 21,000? What about the size of our footprint and the impact on the Afghans? How could I personally reconcile all my public statements expressing concern about our military footprint with supporting McChrystal? Even were I to agree with McChrystal’s assessment and recommendation, I had no idea how I could get the president’s approval of even a fraction of that number. It didn’t take a clairvoyant to see a train wreck coming.

The only time as secretary of defense that I was truly alarmed was when I heard what McChrystal intended. I decided to meet him secretly in Europe on August 2 and hear firsthand what he had to say. Just before the trip, I participated in the president’s retreat with the cabinet and senior White House staff. An article in The Washington Post on the morning of the thirty-first reported that McChrystal was preparing to ask for a significant increase in troop levels in Afghanistan, a real help as I prepared to spend twenty-four hours interacting with the White House staff.

I climbed aboard my airplane late in the afternoon on the first and flew to Chievres Air Base in Belgium, where I sat down with McChrystal at eight-thirty on Sunday morning for what turned out to be a five-hour meeting. All the other key players were there as well: Mullen, Petraeus (as Central Command commander), Admiral Jim Stavridis (as supreme allied commander, McChrystal’s NATO boss), Michèle Flournoy, Rodriguez, and of course, a number of their staff. We met in a very plain, utilitarian conference room at the air base around a large U-shaped table, thus allowing everyone a view of the PowerPoint slides so essential to all military briefings. The enlisted soldiers keeping the coffee pots full and serving food seemed nervous, probably because of the array of four-star admirals and generals in the room. They seemed oblivious to the short, white-haired guy in a blue blazer with no stars.

I began the discussion by underscoring the need to keep the entire troop decision process confidential through its conclusion, which needed to be pushed beyond the Afghan elections on August 20—even though that exceeded my sixty-day deadline for McChrystal’s assessment. I said there would be four pressure points associated with any force increase: White House and congressional political opposition, the impact on Iraq, the availability of additional forces and the impact on the already stressed Army and Marine Corps, and the need for additional supplemental funding. I then asked McChrystal eight questions I had prepared on the plane:

We spent most of our time on Stan’s assessment of the situation, and he repeated to us his belief that the situation was “serious and deteriorating,” as he had told me a few weeks earlier. He spoke of Karzai’s deficiencies and those of Afghan governance more broadly throughout the country (with some exceptions), the lack of legitimacy, and massive corruption. We talked about civilian casualties and what he intended to do about that, as well as new rules for treating Afghans with respect. He made clear he intended to focus our military effort, as in the past, in the south and east, but he said he would select something like eighty districts and population centers on which to focus our efforts to provide security for the people, “inkblots” on his map where the circles of security would grow until they began to link up. Partnering with the Afghan security forces was critical, and the size and quality of those forces had to increase and improve. It was clear that counterterrorism operations would continue, as well as special operations aimed at taking Taliban commanders “off the battlefield.” We also talked about greater military engagement and cooperation with the Pakistanis in a new effort to get them to help go after Taliban safe havens on their side of the border. (I did not share my skepticism that this would work; there was no harm in trying.) These issues would frame the debate inside the administration in the months to come. “Stability in Afghanistan is an imperative,” I said. “If the Afghan government falls to the Taliban—or has insufficient capability to counter transnational terrorists—Afghanistan could again become a base for terrorism, with obvious implications for regional stability.”

McChrystal said that a new campaign strategy was needed, one that focused on protecting the population rather than on seizing terrain or destroying insurgent forces. He talked a lot about changing the operational culture to interact more closely with the population. He emphasized the urgency of the situation: “I believe the short-term fight will be decisive. Failure to gain the initiative and reverse insurgent momentum in the… next 12 months—while Afghan security capacity matures—risks an outcome where defeating the insurgency is no longer possible.” I thought to myself that he was right about the need to produce a perceptible shift of momentum in the near term; the status quo would lead to failure. I believed we could produce the shift in momentum, but expanding and improving the Afghan forces would take more time.

We then turned to the issue that everyone there knew had prompted the meeting—the question of more troops. I repeated what I had been saying for a year and a half about the history of foreign armies, the Soviet experience, and my concern about reaching a “tipping point” where the size of our presence and our conduct turned us into “occupiers.” McChrystal was ready for the subject. He knew that his chances of getting more troops were nonexistent without my support, and so my concerns had to be alleviated. He said that the size of the force (or footprint) was less important than what you did with it. In and of itself, this wasn’t an earthshaking insight. Similar debates took place over troop levels in Iraq. But I had viewed Afghanistan for so long through the lens of the Soviets’ experience that his comments had a serious impact on me. If the Afghans could see foreign and Afghan forces working together and providing sustainable protection so they could go about their daily lives and not fear the return of the Taliban, they would not resent their presence. Respect for the Afghans and their customs was critical. He spoke at considerable length on the footprint issue, and when he was finished, while not yet supportive of a big increase in troop numbers, I was at least open to considering it.

I told McChrystal that I wanted him to wait to submit the assessment until after the Afghan election so he could include at least a preliminary evaluation of its impact. He needed to ensure that the military strategy he presented in his assessment focused explicitly on implementing the broader strategy the president had announced in March: “This will be required to achieve the objectives Obama approved.” I said he should submit the assessment and troop recommendations separately because I expected the former to leak and we had to hold the latter very closely. I wanted people to focus first on the assessment, how things were going, and on strategy. Having troop options on the table at the same time would totally divert the debate to numbers, and the substance of the assessment would be ignored.

Mullen and I briefed the president on the meeting in the Oval Office on August 4. Biden, Emanuel, and Donilon were also there. We focused on Stan’s assessment and the decision-making process to come. I reminded everyone that the troop increase approved in February had preceded the president’s decisions on strategy in March. McChrystal’s assessment would describe the situation as he saw it and then describe how he would operationally implement the president’s March strategy decisions, including the resources required.

I repeated now to the president all that I had said to McChrystal. I told the president that Stan would probably need some additional capabilities for training the Afghan army and some additional enablers—ordnance disposal, counter-IED, medevac, helicopters—but “not a huge number and I’ll provide ample justification.” I continued, “I understand your priorities this fall—the heavy lift on health care, energy, the budget. I will not add to your burden.” At the end of 2009, early 2010, I said, we could evaluate where we were, and I could make further recommendations. I understood the need to justify any increase, I told him; I would not put him in the position of having the appearance or the reality of an open-ended commitment. McChrystal believed that, if properly resourced, he could have the situation in a different place in one year, I said, and the Afghan forces able to secure key population centers within three years.

The president said that he wanted a choice of real options, including not just troop-intensive counterinsurgency. “I would never do that to you,” I said. “But whatever we do, we will need more trainers for Afghan forces and more enablers—let’s do that in September, and then do a basic review in January—a return to ‘first principles.’ ” I went on to say that combat units wouldn’t be ready to deploy before spring anyway, and the real decision was whether to add more combat units or not. In January, we should have a pretty good picture of the effect of the election, whether we could accelerate the recruitment and training of the Afghan army and police, and progress with reintegration of former Taliban fighters.

Obama asked if he had to spend $100 billion a year in Afghanistan. If it was necessary for the security of the United States, he said he would do so. But was that necessary to keep al Qaeda down? Were there alternatives? What about Pakistan? Biden weighed in with his view of the level of congressional opposition to any further increase in troop levels, saying, “The Democrats hate the idea, and the Republicans will just say, ‘You’re on your own.’ ” Nothing new there. Emanuel said that the Hill had voted for the war supplemental the preceding May only as a favor to the president, and they wouldn’t do it now. The president concluded by asking for “robust options” and saying that he would look at McChrystal’s assessment and that we would look deeper at the end of the year or soon afterward into whether we were on the right path. At the end of the session, I said that we would reevaluate progress regularly, something not done in Iraq or Vietnam. The president responded that while North Vietnam had never attacked the United States, there were still points during the war when the basic approach should have been questioned: “I just don’t want McChrystal to come in determined to stay on the same path if it’s not working.” Mullen had the last word: if that was the case, he said, “I would tell you to stop.”

That discussion presaged many of the debates we would have in the months to come. I thought the president had been thoughtful and balanced, sensible in his comments and questions. He was aware of the politics but, unlike Biden and Emanuel, not driven by them. The meeting took place on the president’s forty-eighth birthday. At Leon Panetta’s suggestion, I asked Obama whether he wanted a billion-dollar presidential helicopter or an F-22 for his present. He demurred.

A week and a half later I asked Cartwright whether McChrystal could include an option that would be limited to trainers for the Afghans and enablers, in numbers up to about 7,500 troops. We could then push off a decision on combat forces until January inasmuch as we couldn’t deliver them until late spring anyway. In short, as of mid-August, I continued to be focused on a modest increase in troops in the fall and possibly more only after the first of the year, depending on a thorough evaluation of the situation.

Meanwhile Holbrooke was doing his best to bring about the defeat of Karzai in the August 20 elections. Richard had spoken for months about the need for creating a “level playing field” for all presidential candidates in Afghanistan, including ensuring that they all had security, access to independent media, and transportation to campaign around the country. What he really wanted was to have enough credible candidates running to deny Karzai a majority in the election, thus forcing a runoff in which he could be defeated. Unlike the 2004 Afghan presidential election, when the United States offered Karzai unqualified support, in the months leading up to the 2009 election our public position was one of neutrality among the candidates. But Holbrooke and U.S. ambassador to Afghanistan Karl Eikenberry were encouraging the other candidates, meeting and being photographed with them, attending their rallies, and making suggestions. Karzai might not be a great president, I figured, but he sure as hell knew what was going on in his own capital and was well aware of the American efforts to unseat him. Indeed, as Peter Lavoy, the senior intelligence officer briefing the NSC, later told us, Karzai saw the United States—the Obama administration—walking away from him and turned to the warlords and made deals to get reelected.

The election outcome was deeply marred by security problems but also large-scale fraud perpetrated by Karzai. He failed to get the magic 50 percent in the first round but still ended up with a second term. It was all ugly: our partner, the president of Afghanistan, was tainted, and our hands were dirty as well. The senior UN representative for Afghanistan, Ambassador Kai Eide, subsequently gave a report on the election to the NATO defense ministers during which he sat next to me. Before speaking publicly, he whispered to me that while he was only going to say that there was blatant foreign interference in the election, he wanted me to know he had in mind specifically the United States and Holbrooke. Our future dealings with Karzai, always hugely problematic, and his criticisms of us, are at least more understandable in the context of our clumsy and failed putsch.

For two and a half years, I had warned about the risks of a significant increase in the U.S. troop presence in Afghanistan, and during that period we had increased from about 21,000 to 68,000 troops. I was torn between my historical perspective, which screamed for caution, and what my commanders insisted was needed for accomplishing the mission they had been given by the president and by me. Three very different commanders—McNeill, McKiernan, and McChrystal—had all asked for more troops. I believed, with Mike Mullen, that the war in Afghanistan had been neglected and underresourced in the Bush administration. But how many troops were too many before reaching the tipping point in terms of Afghan attitudes and support? Embassy polling showed that in 2005 about 80 percent of Afghans saw us as allies and partners; by summer 2009, after nearly eight years of war, that was down to 60 percent.

As I thought about the tipping point, it seemed to me we had several vulnerabilities with the Afghan population. One was civilian casualties; every incident was a strategic defeat, often caused and always manipulated by the Taliban and then magnified by Karzai. Another was our thoughtless treatment of the Afghans in routine encounters, including U.S. and coalition military vehicles barreling down the roads scattering animals and scaring people. We often disrespected their culture or Islam and failed to cultivate their elders. We collaborated with Afghan officials who were ripping off ordinary citizens. In Kabul and all over the country, we and our coalition partners, as well as nongovernmental organizations, far too routinely decided what development projects to undertake without consulting the Afghans, much less working with or through them on what they wanted and needed. Was it any wonder that Karzai and others complained they had no authority in their own country? Or that even reasonably honest and competent Afghan officials got no respect from their fellow citizens? For all our hand-wringing and hectoring about corruption, we seemed oblivious to how much we were contributing to it, and on a scale that dwarfed the drug trade. Tens of billions of dollars were flooding into Afghanistan from the United States and our partners, and we turned a blind eye or simply were ignorant of how regularly some portion was going to payoffs, bribes, and bank accounts in Dubai. Our own inspectors identified how lousy—or nonexistent—U.S. government controls were. From Karzai on down, Afghans had to shake their heads at our complaints about their corruption when elements of the American government (and almost certainly a number of our closest allies) were paying off them and their relatives as agents and to secure their cooperation. Hillary Clinton and I repeatedly objected to this contradictory behavior by the United States, but to no avail.

An important way station in my “pilgrim’s progress” from skepticism to support of more troops was an essay by the historian Fred Kagan, who sent me a prepublication draft. I knew and respected Kagan. He had been a prominent proponent of the surge in Iraq, and we had talked from time to time about both wars, including one long evening conversation on the veranda of one of Saddam’s palaces in Baghdad. His essay, “We’re Not the Soviets in Afghanistan,” subsequently published in The Weekly Standard, reminded me of the brutal realities of my first Afghan war. In that conflict, an ill-trained, loutish, and often drunken Soviet army had gradually turned to an out-and-out war of terror on the Afghan people, killing at least a million and creating somewhere between three and five million refugees. (Other accounts put the number as high as seven million.) They tried to upend Afghan culture by redistributing property on a large scale and by trying to destroy “key pillars” in the social structure. As Kagan wrote, “Increasing frustration led to increased brutality, including a deliberate campaign to de-house the rural population (forcing people to concentrate in cities that the Soviets believed they could more easily secure)…. The Soviets also used chemical weapons, mines, and devices intended to cripple and maim civilians.” Kagan wasn’t telling me anything I didn’t remember about Soviet behavior in Afghanistan in the 1980s; after all, at CIA, I had watched it, reported on it, and beginning in 1986, had a direct part in countering it. What I had not done consciously as secretary of defense, Kagan’s essay made me realize, was contrast the behavior of the Soviet troops with our own. As McChrystal had said in Belgium at our meeting, the size of the footprint matters far less than what you do with it. There were reasons to be cautious about more troops, and I still was, but I now saw our experience in a light different from the Soviet one.

My thinking about more troops was further affected by President Obama’s speech on August 17 to the Veterans of Foreign Wars. He said, in reference to the war in Afghanistan, “The insurgency in Afghanistan didn’t just happen overnight, and we won’t defeat it overnight. This will not be quick, nor easy. But we must never forget: This is not a war of choice. This is a war of necessity. Those who attacked America on 9/11 are plotting to do so again. If left unchecked, the Taliban insurgency will mean an even larger safe haven from which al Qaeda would plot to kill more Americans. So this is not only a war worth fighting. This is fundamental to the defense of our people.” This was the only time I could recall him being so forthright and committed in terms of prosecuting this war to a successful conclusion. Maybe my comment to Emanuel a few days earlier about the president needing to take “ownership” of the war had penetrated.

Because the Pentagon was more accustomed to Bush’s style of decision making than Obama’s, the military’s proposed timetable for getting a decision on more troops by the end of September was naïve. As planned, McChrystal submitted his assessment to me on August 31. Only Mullen, Petraeus, and Stavridis (at NATO) got copies initially. Petraeus endorsed the assessment the next day and, contrary to later claims, specifically supported Stan’s view of the need for both reintegration of lower-level former Taliban fighters into Afghan society and reconciliation with senior Taliban commanders. Flournoy discussed the process with Donilon, and they agreed we would pass the assessment to the NSS right after Labor Day (September 7), and it would then be discussed at limited-attendance meetings of both the deputies and the principals.

Donilon didn’t want a firm deadline on resource decisions; he correctly wanted to focus discussion initially on the assessment—as I had hoped would happen—and to make sure we had the strategy right before talking about troop numbers. He said there should be no discussion of the assessment at NATO until the White House was comfortable with it. Stavridis in Brussels agreed to sit on his copy, but he and I would have to deal with a very unhappy NATO secretary general, who expected to be brought into the loop early—a reasonable position, since McChrystal was a NATO commander.

Nobody was going to keep Barack Obama in the dark for a week about what McChrystal’s assessment said. Mullen and I met with the president in the Oval Office on September 2 and, as he had insisted, gave him a copy of the report. I told him it did not represent a new strategy but focused on implementing what the president had approved in March. I indicated that the following week I would forward to him the views of Petraeus, Mullen, and the Joint Chiefs, as well as my own, on the way ahead. I promised he would get a full array of options for discussion from McChrystal, noting that there were three elements to the troop issue—combat units, trainers, and enablers (medevac, counter-IED, and the like).

Yet again I told the president I wanted to move quickly on the enablers, sending perhaps up to as many as 5,000. With increased IED attacks and casualties, the message to troops and commanders in delaying the enablers was unacceptable, I said. I requested flexibility to respond to these requests as they came in; I had been sitting on some of them for weeks, to stay under the presidentially imposed troop cap of 68,000. I asked for a decision within a week and offered to report weekly to the NSS on any additional troops sent in this category.

To my astonishment and dismay, the president reacted angrily to my request. Why do you need more enablers? he asked. Were they not anticipated as part of the 21,000? What had changed? Is this mission creep? The public and Congress don’t differentiate between combat troops and enablers, he said. Incremental increases lead to a sudden increase in commitment. Any more troops would be a heavy lift in terms of numbers and money. Biden jumped in with the familiar refrain that the Republicans would start calling it “Obama’s war.” I told them I had gotten a phone call from Senator Joe Lieberman saying that he, John McCain, and Lindsay Graham wanted to be helpful, and I had told Lieberman that they couldn’t let the Republicans take a pass on this key national security issue. I told the president I understood his concerns about an open-ended commitment and mission creep but that “war is dynamic, not static. At the end of the year, whatever the troop numbers, we’ll reevaluate and change our strategy if it’s not working.”

Just outside the Oval Office after the meeting, exasperated, I told Biden and Donilon that with respect to the 5,000 enablers, “From a moral and political standpoint, we cannot fail to take action to protect the troops.”

I was deeply disturbed by the meeting. If I couldn’t do what I thought was necessary to take care of the troops, I didn’t see how I could remain as secretary. I was in a quandary. I shared Obama’s concerns about an open-ended conflict, and while I wanted to fulfill the troop requests of the commanders, I knew they always would want more—just like all their predecessors throughout history. How did you scale the size of the commitment to the goal? How did you measure risk? But I was deeply uneasy with the Obama White House’s lack of appreciation—from the top down—of the uncertainties and inherent unpredictability of war. “They all seem to think it’s a science,” I wrote in a note to myself. I came closer to resigning that day than at any other time in my tenure, though no one knew it.

During the deliberations over Afghanistan in the weeks to come, events would routinely drag me back to the sacrifices our troops were making and to the obtuseness of many of those at home. One such event occurred two days after I received McChrystal’s assessment. That day twenty-one-year-old Lance Corporal Joshua M. Bernard’s Marine unit was ambushed, and he was mortally wounded by a rocket-propelled grenade. An Associated Press photographer took a picture of the dying Marine, being tended by two comrades. His wounds were graphically portrayed in the photo. After Bernard was buried ten days later, the AP sent a reporter to talk with his family and tell them they were going to publish the photo. Bernard’s father asked that the photograph not be circulated to the news media for publication, saying it would only hurt the family more. The AP’s intent to run the photo came to my attention on September 3, and its callousness toward the family both sickened and angered me. From early in my tenure, I had had a good relationship with the press and had spoken publicly and often to military audiences about its importance in upholding our freedom (and identifying problems that needed fixing). But publishing this photo was an outrage as far as I was concerned.

I called Tom Curley, the president and chief executive of the AP, and asked him, in consideration of the father’s wishes, not to run the picture. I said at one point in the conversation, “I am the secretary of defense, and I am begging you not to run that picture.” I had never begged anybody for anything, but the sacrifice of this young Marine and the anguish of his family had suddenly become very personal to me. Curley said he would review the decision with his editors, but he didn’t hold out much hope they would change their minds. I followed up with a letter, in which I said, “The American people understand that death is an awful and inescapable part of war,” but publishing the photo would be “an unconscionable departure from the restraint that most journalists and publications have shown covering the military since September 11.” I called the decision “appalling” and said that the issue was not one of law or constitutionality but one of “judgment and common decency.” The AP was fresh out of common decency that day and put the photo on the wire. Fortunately, most newspapers and other media had better judgment than the AP and refrained from publishing the picture. The AP’s insensitivity continues to rankle me.

I formally sent McChrystal’s assessment to the president through Jim Jones on September 10, along with a separate paper by McChrystal on why he thought a counterterrorism strategy alone would not work in Afghanistan. At that point, McChrystal was almost certainly the most lethal and successful counterterrorism practitioner in the world. The successes of the U.S. forces under his command in both Iraq and Afghanistan were legion and legendary. The paper I gave to the president was a distillation of years’ experience in hunting bad guys. McChrystal wrote that while CT (counterterrorism) operations are highly effective at disrupting terrorists, they are not the endgame to defeat a terrorist group. “CT operations are necessary to mitigate a sanctuary, but to defeat a terrorist group, host nation capacity must grow to ensure a sustainable level of security…. Without close-in access, fix and find methods become nearly impossible…. Predator [drone] strikes are effective where they complement, not replace, the capabilities of the state security apparatus, but they are not scalable in the absence of underlying infrastructure, intelligence, and physical presence.” Given McChrystal’s counterterrorism credentials, I was both astounded and amused in the weeks to come as Joe Biden; his national security adviser, Tony Blinken; Doug Lute; and others presumed to understand how to make CT work better than Stan did.

Along with the assessment, I gave the president the written endorsements and comments from Petraeus, the Joint Chiefs, and Mike Mullen. I said that “they all are essentially of one mind: that McChrystal is the right man, has the right military approach to accomplish the goals set forth in your March 27 decisions, and that he should receive proper resourcing to carry out his plans. As well, they all are, with difference only in degree, convinced that no strategy will work as long as pervasive corruption and preying upon the people continue to characterize governance in Afghanistan.” It did not dawn on me at the time that my practice of having the president hear directly from each level in the chain of command, because of the unanimity of the senior military in support of McChrystal’s recommendations, in this instance probably only reinforced Obama’s and Biden’s suspicion of a “military bloc” determined to force the commander in chief’s hand.

The same day I gave Jones McChrystal’s assessment, I also gave the president a long “eyes only” memo on my own thinking. I began by saying that, with the additional forces he had approved in February and March, I had hoped we would have until early 2010 to see if McChrystal’s approach resulted in changed momentum in Afghanistan and, if so, that we would be able to use that to justify continued and perhaps increased support. With the worsening situation the general had identified, and public statements of grave concern by U.S. officials, however, “the debate and decisions—including over resources—I had hoped could be delayed until early next year when we might be able to show some progress are, unfortunately, upon us now. In fact, circumstances have conspired to place us at a historic crossroads during the next few weeks.” I added that “as usual,” all the options were unpalatable.

The principal alternative to McChrystal’s recommendations, I felt, would be Biden’s “counterterrorism plus” strategy. I told the president that I thought that strategy had all the disadvantages of a counterterrorism strategy and not enough capability to reap any of the advantages of a counterinsurgency strategy; and “I also don’t know how to explain such a strategy to anyone.”

I wrote impertinently that any new decision that abandoned his decisions in March or the vow he had made to the VFW in August would be seen as a retreat from Afghanistan, with all the implicit messages that that would send to Afghans, Pakistan, our Arab and NATO allies, Iran, North Korea, and others about American will and staying power: “We need to give it [the March strategy] a chance.” Knowing this president, I realized that he, like me, had a number of questions that had to be answered before any decision would be made, and I laid out some of them:

The priority, I said, should be to expand the Afghan security forces as quickly as possible. Additional U.S. and allied forces should be considered a temporary “bridge” to train those Afghans while keeping the situation on the ground from deteriorating further, at least until the Afghans could protect their own territory and keep the Taliban and al Qaeda out. I also said we needed a clearer strategy for reintegration of the Taliban. “I am confident of this,” I said. “Your strategy—centered on building Afghan security capacity—gives us a chance for success; the more limited alternatives do not.” I ended on a very personal note:

During September, several events fractured what little trust remained between the senior military and the president and his staff. On September 4, The Washington Post columnist Michael Gerson published an interview with Petraeus in which the general stated flat-out that while there was no guarantee more troops would lead to success in Afghanistan, “it won’t work out if we don’t” send a lot more. He dismissed the “counterterrorism plus” strategy as insufficient, saying it had been tried before and that the way to target terrorists was with “on-the-ground intelligence,” which “takes enormous infrastructure.” Petraeus came down squarely in the interview behind McChrystal’s approach, “a fully resourced, comprehensive counterinsurgency campaign.” Virtually everybody in the Obama White House saw this as blatant lobbying designed to force the president to approve more troops. Their suspicion of Petraeus and his political ambitions was not allayed by the fact that Gerson had been a speechwriter for George W. Bush, something Petraeus denied knowing.

On September 13, the president chaired the first of nine—by my count—very long (two-to-three-hour) meetings on McChrystal’s assessment and Afghan strategy. Two days later the Senate Armed Services Committee held a confirmation hearing for Mike Mullen’s second term as chairman of the Joint Chiefs of Staff, at which time he forcefully argued for more troops in Afghanistan. He was implicitly critical of the vice president’s views, saying we could not defeat al Qaeda and prevent Afghanistan from becoming a safe haven again “from offshore…. You have to be there, where the people are when they need you there, and until they can provide for their own security.” The president—and everyone else in the White House—was livid, seeing the testimony as another effort by Mullen and the military to force the commander in chief’s hand. Rahm told me that the president “used my language” when he heard what Mullen had said. In an effort to calm things down, at a press conference soon thereafter I said the president deserved the right to absorb McChrystal’s assessment and have his questions answered, that some of the most important decisions of his presidency were involved and he should not be rushed. I suggested that “everybody should just take a deep breath.”

Then the biggest shoe of all dropped. On Monday, September 21, The Washington Post published a detailed story by Bob Woodward on McChrystal’s assessment, clearly based on a leaked copy. The four-column-wide headline read “McChrystal: More Forces or ‘Mission Failure.’ ” The Post had given us advance warning it was going to run the story, and over the weekend Cartwright, Flournoy, and Geoff Morrell negotiated with Woodward and others from the Post to remove sensitive numbers, references to intelligence gaps, Special Forces unit designations, and the like. They had some success, but they could not redact the political bombshell the story represented. The story ended with a quote from the assessment: “Failure to provide adequate resources also risks a longer conflict, greater casualties, higher overall costs, and ultimately, a critical loss of political support. Any of these risks, in turn, are likely to result in mission failure.” After I left office, I was chagrined to hear from an insider I trust that McChrystal’s staff had leaked the assessment out of impatience with both the Pentagon and the White House. If so, I’d be very surprised if Stan knew about it.

Anger and suspicion were further fueled six days later when the CBS-TV program 60 Minutes aired an interview with McChrystal in which he spelled out in detail how bad he had found the situation in Afghanistan and what needed to be done. The interview had been taped in late summer, long before the debate got under way in the administration, but the timing of its airing was awful.

McChrystal had been invited weeks earlier to give a speech on October 1 in London and asked Mullen whether he should do so, given the furor surrounding the leaked assessment. Mike encouraged him. I did not object. I should have. Stan’s speech was innocuous enough, but in response to a question afterward, he dismissed out of hand the option Biden was supporting.

An infuriated president, Mullen, and I repeatedly discussed what he regarded as military pressure on him. On September 16, Obama asked us why all this was being discussed in public. “Is it a lack of respect for me? Are they [he meant Petraeus, McChrystal, and Mullen] trying to box me in? I’ve tried to create an environment where all points of view can be expressed and have a robust debate. I’m prepared to devote any amount of time to it—however many hours or days. What is wrong? Is it the process? Are they suspicious of my politics? Do they resent that I never served in the military? Do they think because I’m young that I don’t see what they’re doing?” Mike assured him there was no lack of respect. I said we just needed to shut everyone down until the process was complete.

The president and I then talked alone. I told him Mullen had called both Petraeus and McChrystal after the incidents and thought he had the situation under control. I said Mike’s testimony had been a surprise to me, especially since we had reviewed potentially hot topics before his hearing.

Again and again I tried to persuade Obama that there was no plan, no coordinated effort by the three military men to jam him. I said that if there had been a strategy to do that, they sure as hell wouldn’t have been so obvious. I reminded him that McChrystal had never had a job before with the kind of public exposure he now had, that he was inexperienced and a bit naïve about dealing with the press and politics. I said Mullen and Petraeus were both on his team and wanted to serve him well; but particularly when testifying, or even when talking to reporters, both felt ethically compelled to say exactly what they thought, however politically awkward. I told the president that Mike’s independence had annoyed Bush as well. My assurances fell pretty much on deaf ears, which I found enormously frustrating and discouraging.

The press was reporting a campaign being mounted by the military to force acceptance of McChrystal’s recommendations, and Emanuel told me that, according to reporters, there were four different sources saying that McChrystal would quit if he didn’t get his way. A wall was going up between the military and the White House. That was bad for the country, even dangerous. I had to fix it. In a conference call with McChrystal and Petraeus on September 23, I told them that the decision the president was facing was conceivably the most significant of his presidency. The experts and politicians in Washington were divided on what to do in Afghanistan. The president was very deliberative and very analytical, and he was going to take whatever time was necessary to work through this decision. If he agreed to provide significant additional troops, he would do everything he could to make it work, though it would be a very heavy political lift at home. I directed McChrystal to provide his memorandum on force options only to me, the chairman, Petraeus, Stavridis at NATO, and the NATO secretary general. I said no copies should be made, and it should not be shared with staff or anyone else, that a leak would possibly be fatal to Stan’s case. I reassured them that the president was not questioning Stan’s assessment or recommendations for further resources, but rather whether changed circumstances on the ground required revisiting the strategy he had settled on in March. I ended the call by emphasizing that we had to actively oppose the perception in the press and embraced by some in Congress that the president and the military were pitted against each other.

Four days after McChrystal’s London gaffe, I gave a speech to the Association of the U.S. Army in which I mentioned the leaks. I said it was important to take our time to get the Afghan decision right, “and in this process it is imperative that all of us taking part in these deliberations—civilian and military alike—provide our best advice to the president candidly but privately.” Most commentators thought it was a shot at McChrystal, but my target was far broader. We heard regularly from members of the press that Biden, Jones, Donilon, McDonough, Lute, Emanuel, and Axelrod were “spilling their guts” regularly—and disparagingly—to reporters about senior military leaders, Afghanistan, and the decision-making process. I was told that The New York Times was besieged by unsolicited White House sources offering their views. I acknowledged that the Pentagon leaked. But whenever I would complain about White House leaks, there were bullshit protestations over there of innocence. Only the president would acknowledge to me he had a problem with leaks in his own shop.

As impatient and frustrated as I would get at different points, not to mention just being sick of sitting in the Situation Room hour after hour, day after day, I believe the process on Afghanistan was an important and useful one. In my entire career, I cannot think of any single issue or problem that absorbed so much of the president’s and the principals’ time and effort in such a compressed period. There was no angle or substantive point that was not thoroughly examined. If I were to fault the process, I would say that vastly more attention was focused on every aspect of the military effort than—despite Donilon’s and Holbrooke’s best efforts—on the broad challenge of getting the political and civilian part of the equation right. Too little attention was paid to the shortage of civilian advisers and experts: to determining how many people with the right skills were needed, to finding such people, and to addressing the imbalance between the number of U.S. civilians in Kabul and elsewhere in the country. Nor did we focus on the tension between our ambassadors and commanders in Afghanistan, Eikenberry and McChrystal in particular. During my tenure as secretary, there were three U.S. ambassadors to Kabul; none did well, in my opinion. None could compare to Afghan-born Zalmay Khalilzad, the U.S. ambassador in Kabul from 2003 to 2005, in coaching, counseling, and working with Karzai—or to a couple of CIA station chiefs in Afghanistan. Even Secretary Clinton would speak of Eikenberry’s insubordination, that he would not do what she directed. Though both Clinton and I wanted Eikenberry replaced—because his relationship with Karzai was beyond repair and his relationships with both Defense and State were so poor—and repeatedly told Jones so, the ambassador was protected by the White House.

From September through November, over and over again we would rehash the issues and get further into the weeds—details beyond what was needed or appropriate. Broadly, there were three substantive areas on which our many meetings focused. The first was the nature of the threat. What were the relationships between the Taliban, al Qaeda, and other extremist groups in the Afghan-Pakistani border area? Was defeat of the Taliban essential to the defeat of al Qaeda? If the Taliban regained power, would al Qaeda return to Afghanistan? Would a more stable Afghanistan change Pakistan’s strategic calculus? The second issue was which strategy for dealing with the threat would be most effective and efficient, COIN or CT-Plus. The key question with COIN was whether there was an Afghan model of governance that would be “good enough” to meet our objectives. Did the government have enough legitimacy in the eyes of its own people to permit our strategy to succeed? In the case of CT-Plus, could it work if the United States lacked the resources on the ground to protect the population and without adequate intelligence to be effective in its counterterrorism strikes? Third, if we stayed with the president’s March strategy, how would we know if and when it was time to change course?

Pakistan continued to be a critically important factor in our discussions. If Pakistan was so critical to the success of our strategy, Biden asked, why were we spending thirty dollars in Afghanistan to every one dollar in Pakistan? There was a lot of talk about more military and civilian aid to the Pakistanis. Their military was deeply suspicious of U.S. intentions in Pakistan, believing any effort to increase the number of our uniformed personnel there was part of a nefarious scheme to seize their nuclear weapons. They welcomed our cash and our equipment but not our people. And they were not particularly interested in letting us teach them how to go after targets in their own country. As for civilian assistance, their paranoia and our political ham-handedness reinforced each other. After much political effort, and the leadership of Senators John Kerry and Dick Lugar and Representative Howard Berman, Congress passed a five-year, $7.5 billion aid package for Pakistan. It was a great achievement and just what was needed, especially the multiyear aspect to demonstrate our long-term commitment. Then some idiot in the House of Representatives attached language to the bill that stipulated that the assistance was conditional on the Pakistani military not interfering with the civilian government. Not surprisingly, there was outrage in Pakistan, especially among the military. In a flash, all the actual and potential goodwill generated by the legislation was negated. I knew that nothing would change Pakistan’s hedging strategy; to think otherwise was delusional. But we needed some level of cooperation from them.

The president kept returning also to the matter of cost. He observed that the cost of the additional troops McChrystal was requesting would be about $30 billion; yet if he froze all domestic discretionary spending, he would save only $5 billion, and if he cut the same by 5 percent, that would save only $10 billion. He said that if the war continued “another eight to ten years, it would cost $800 billion,” and the nation could not afford that given needs at home. His argument was hard to disagree with. The costs of the war were staggering.

By the fifth NSC meeting, on Friday, October 9, some clarity was emerging on the key issues. Panetta set the stage with a simple observation: “We can’t leave, and we can’t accept the status quo.” The president said he thought we had reached “rough” agreement on that but also on what was achievable in terms of taking on the Taliban; that defining counterinsurgency in terms of population security as opposed to Taliban body count was sound; and that the basic “inkblot” strategy was sound—we couldn’t resource COIN throughout the country, so we had to deny the Taliban a foothold in key areas.

He then posed the next set of questions. Were the interests of the Afghan government aligned with our own? How could we ramp up training of the Afghan forces to allow us to leave in a reasonable time? How would we transition from clearing out the Taliban in an area to transferring security responsibility there to the Afghans? Did we have a strategy for reintegration of Taliban fighters? What were the timetables, and how did we sustain the effort? If we were not sending enough troops for countrywide counterinsurgency, how did we choose what to protect? How would we deal with Pakistani opposition to our adding troops? I thought these questions in themselves reflected progress in our discussions. Apparently assuming the president was leaning toward approving significantly more troops, Biden jumped in: “What if a year from now this isn’t working? What do you do then? Are you increasing the consequences of failing?”

About eight that same Friday night, as I was eating my Kentucky Fried Chicken dinner at home, the president called. “I’m really looking to you for your views on the way forward in Afghanistan. I’m counting on you,” he said. Earlier that week Biden had leaned over in the Situation Room and whispered to me, “Be very careful what you recommend to the president because he will do what you say.” I spent the weekend deciding what to say.

When I met privately with the president in the Oval Office on October 13, I told him I had thought about his call a lot and had prepared a memo for him offering my thoughts on what he should do. He grinned broadly, stuck out his hand to shake over the bowl of apples on his coffee table, and said, “You have the solution?” I wasn’t sure about that, but in the event, one of the most significant decisions of his presidency largely tracked the recommendations in my paper.

I wrote that the Afghan Taliban and al Qaeda had become symbiotic, “each benefiting from the success and mythology of the other, both inside and outside Afghanistan.” Al Qaeda clearly believed that a Taliban victory over the United States in Afghanistan would have great strategic benefit for the group.

I wrote that all three of the mission options we had been discussing were “doomed to fail, or already have.” Counterterrorism focused solely on al Qaeda could not work without a significant U.S. ground presence in Afghanistan and the opportunity to collect intelligence that this would afford us. “We tried remote-control counterterrorism in the 1990s, and it brought us 9/11.” “Counterterrorism plus,” or “counterinsurgency minus,” was what we had been doing since 2004, and “everyone seems to acknowledge that too is not working.” Fully resourced counterinsurgency “sounds a lot like nation-building at its most ambitious” and would require troop levels, time, and money that few in the United States or in the West were prepared to provide.

I wrote that the core goals and priorities Obama had decided the previous March remained valid and should be reaffirmed. However, we had to narrow the mission and better communicate what we were trying to do. We could not realistically expect to eliminate the Taliban; they were now a part of the political fabric of Afghanistan. But we could realistically work to reverse their military momentum, deny them the ability to hold or control major population centers, and pressure them along the Pakistani border. We ought to be able to reduce their level of activity and violence to that which existed in 2004 or thereabouts. I recommended focusing our military forces in the south and east and charging our allies with holding the north and west. Our military efforts should be intended to stabilize the situation in Afghanistan and buy time to expand and train the Afghan security forces, who, despite their many deficiencies, were courageous fighters; many of them were prepared to die—and had died—fighting the Taliban. We should “quietly shelve trying to develop a strong, effective central government in Afghanistan.” What we needed, I wrote, was some central government capacity in a few key ministries—defense, interior, finance, education, rural development. We should help broker some kind of “national unity” government or other means to give the Karzai government at least a modicum of legitimacy in the eyes of the Afghan people. We also had to get a handle on corruption. “Our kids must not die so that corrupt Afghan officials can line their pockets.”

All this would give us a mission that the public and the politicians could easily understand: “Deny the Taliban momentum and control, facilitate reintegration, build government capacity selectively, grow the Afghan security forces, transfer security responsibilities, and defeat al Qaeda.”

I supported McChrystal’s request for 40,000 troops, but I offered an alternative of about 30,000 troops. I urged Obama not to place a firm ceiling on the numbers because troop numbers are always estimates, and there are always unplanned needs. Because the fourth brigade combat team McChrystal had requested (bringing his number to 40,000 troops) was needed to replace the Canadians and the Dutch, who were leaving the south in 2010 and 2011, I suggested he leverage our own new commitment to get the allies to provide those replacement troops.

To assure Americans this wasn’t an open-ended commitment to a stalemate with constantly rising numbers of troops for years to come, I said I thought it was imperative to pledge that we would review progress at the end of 2010 and, if necessary, “adjust or change our approach.” I also wrote that while the deliberative process “has served you well, we cannot wait a month or two for a decision. Uncertainty about the future is beginning to impact Afghans, the Pakistanis, our allies, and our troops.”

Almost two weeks later, on October 26, the president invited Hillary and me to discuss the options. We were the only outsiders in the session, considerably outnumbered by White House insiders including Biden, Emanuel, Jim Jones, Donilon, and John Brennan. Obama said at the outset to Hillary and me, “It’s time to lay our cards on the table. Bob, what do you think?” I repeated a number of the main points I had made in my memo to him. Hillary agreed with my overall proposal but urged the president to consider approving the fourth brigade combat team if the allies wouldn’t come up with the troops.

The exchange that followed was remarkable. In strongly supporting a surge in Afghanistan, Hillary told the president that her opposition to the surge in Iraq had been political because she was facing him in the Iowa primary. She went on to say, “The Iraq surge worked.” The president conceded vaguely that opposition to the Iraq surge had been political. To hear the two of them making these admissions, and in front of me, was as surprising as it was dismaying.

Rahm charged once again that the military had waged a campaign to limit the president’s options to what McChrystal wanted. Seething inside, I ignored him and turned to a question I knew was on the president’s mind—why 40,000 more troops were needed if we were narrowing the mission. I said that the early phase of any option other than pure counterterrorism was to reverse the momentum of the Taliban and degrade their capabilities. (A counterterrorism strategy alone could not do that.) The president commented that OMB had told him 40,000 more troops would cost an additional $50 billion or more a year, putting the cost of the overall effort at maybe a trillion dollars over ten years. What were the national security implications of that for the deficit, defense investment, and so on? he asked. He then wrapped up the meeting, saying he wanted to make a decision before his Asia trip (which was to begin on November 12).

Rahm called after the meeting to apologize for the “campaign” comment but said again that the president was feeling boxed in by all the press articles, including one that same day about a Defense war game that purportedly showed the vice president’s option wouldn’t work. I told Rahm that until that story, the military had been quiet since my public warning to advise only in private, but that had not been true of people at the White House. He admitted, “I know, I know.”

Jones came to see me that same day to share his concerns over McChrystal’s plans. He had been very quiet during the meetings in the Situation Room. He said, “The idea of 100,000 American troops in RC [regional command] South and RC East blows my mind. There is something missing, the glue that holds it all together. Where is the Afghanistan-wide plan, including the NATO role?” He was also concerned about the inflexibility of the military—either 40,000 troops or none. Jim said that Mullen was seen as responsible for the contentiousness and “has real problems among some at the White House,” though not necessarily the president. I went through my familiar commentary again. I said that the notion of some kind of organized campaign was ridiculous, that McChrystal’s statement in London was an unscripted answer to a question, and that Mike had admitted that his statement in his confirmation hearing had been a mistake. I said I thought the atmosphere at the White House was getting poisonous, especially on the part of Donilon, who had characterized Mullen and the military as “insubordinate” and “in revolt.” It steamed me that someone who had never been in the military and had never even been to Afghanistan was second-guessing commanders in the field on things like why there were helicopters in certain places. Jones acknowledged he had “to get Tom back in the box.”

The next day the shit hit a new part of the fan. Roughly three weeks earlier the president had told me that he wanted to talk privately with Vice Chairman of the Joint Chiefs Cartwright to get his personal views on the path forward in Afghanistan. I told him that if it became known, Mullen would feel undercut. I advised Obama to tell no one else, and I suggested he meet Cartwright on a Saturday in the residence. The meeting took place on Tuesday, October 20, while I was in Japan. Several people at the White House were aware of it. Mullen wasn’t. A week after the fact, Jones told Mike about the meeting.

Mike came to see me that afternoon, feeling betrayed by Jones, Cartwright, and possibly me. He felt the meeting showed a lack of confidence in him on the part of the president. He asked why the president wouldn’t just tell him about wanting to meet with Cartwright (who had been sworn to secrecy). Mike said Cartwright now felt like damaged goods and was wondering how he could stay long-term. I thought Mike might resign. I described the origins of the meeting and my concerns. I admitted I had probably made a mistake in my advice to the president and should have told him to be up-front with Mike. Mullen then asked how he stood with me, and I told him I would not want to remain as secretary without him as my partner, that I had total confidence in him and felt terrible about the whole episode. I added that the president had put all three of us—Mullen, Cartwright, and me—in an awful spot. Mullen and I agreed we each had to talk privately with the president.

Mike wanted clarity on the perceived “campaign” by the military, the president’s confidence in him, and the overall White House view of the military. A few weeks earlier the president “had chewed our asses,” as Mullen put it, for public military statements. Obama had said, “On Afghanistan, my poll numbers will be stronger if I take issue with the military over Afghanistan policy.” That clearly had bothered Mike (as it did me) because it suggested we were on different teams. After our meeting, I called Rahm and asked for fifteen minutes privately with the president the next day. I said it was about a personnel matter but not mine. Rahm asked, “Mike?” I said yes.

I told Obama about my conversation with Mullen and his worry that the president had lost confidence in him. I reported also that Cartwright felt it would be difficult for him to stay on now. I acknowledged to the president that I had given him bad advice. “I should have told you to go ahead and see Cartwright but call Mike first.” The president said he could have handled it better, too, and maybe he was not sensitive enough to military protocol because he had never served. But “I feel like I should be able to talk to anyone in uniform as commander in chief,” he said. I told him I did exactly that at every post and base without the chain of command present. I told him Mike wanted to stay behind for a private talk with him after our regular meeting that afternoon. Obama said he would give Mike a full vote of confidence but would also repeat his belief that military comments had boxed him in on Afghanistan. Mike later told me they had a good conversation and cleared the air.

This episode serves as a reminder that those at the highest levels of government, tough and experienced people accustomed to the hard knocks of political life at the top in Washington, are still human beings. All of us, in varying degrees, have vulnerabilities, insecurities, and sensitivities. All hate critical press stories that question our motives, integrity, or competence. All, including hardened senior military officers and secretaries of defense, need the occasional pat on the back or gesture of support. And however independent and powerful, we need to know we have the confidence of our boss, especially when he is the president of the United States.

By early November, we were focused on three options: 20,000 additional troops (half for counterterrorism, half for training Afghan forces), the vice president’s recommendation; Option 2, McChrystal’s proposal for 40,000 more troops; and “Option 2A,” my alternative of 30,000 with a push on the allies for 5,000 to 7,000 more. Much would be written later about resentment in the Pentagon over Cartwright helping Biden and his staff craft an alternative plan from McChrystal’s. For Mullen, Petraeus, McChrystal, and others, that was probably true. I had no problem with it. My only concern was that, as was his style, Cartwright didn’t share much about what he was doing with either the senior civilians (including, at times, me) or the military in the Pentagon, which didn’t help things.

Late in our deliberations, we dealt with the important question of how fast additional troops could get into Afghanistan. The original military plan had the deployments spaced out over more than a year. The president correctly pointed out that that could hardly be characterized as a “surge” to recapture momentum. He asked Petraeus how fast the surge had arrived in Iraq. About six months, Petraeus said. Obama decided the arrival of the troops in Afghanistan had to be significantly accelerated. The military leadership ultimately agreed to get most of the troops there by the end of August 2010—a logistical nightmare, but they managed it.

How long would the surge troops stay? The military had been saying that areas cleared of the Taliban would be ready to transition to Afghan government security responsibility within two years. Because the first Marines had arrived in Helmand in the summer of 2009 to take on the Taliban, the president wanted to begin bringing out the surge troops in July 2011. I had opposed any kind of deadlines in Iraq but was supportive of the president’s timeline in Afghanistan because I felt some kind of dramatic action was required to get Karzai and the Afghan government to accept ownership of their country’s security. I also accepted the military’s two-year forecast. I knew well we were not talking about a countrywide transition in July 2011 but, rather, the beginning of a process that would go district by district or province by province. A “conditions-based” date to begin drawing down the surge forces therefore was acceptable to me. To those who said we were inviting the Taliban to just lie low until we left, I said that would only give us more opportunity to accomplish our goals.

In practical terms, a set date to begin drawing down the surge gave Obama something to work with in terms of reassuring both the public and Congress that he was not committing to an open-ended war in Afghanistan. Most Democrats and a growing number of Republicans in Congress had become increasingly skeptical about the war and its cost, both in lives and in treasure. The politics of the troop increase would be a heavy lift indeed—just as in Iraq in early 2007.

The endgame began on November 6 with a thunderbolt cable to Clinton and the White House from Ambassador Eikenberry in Kabul, which leaked almost immediately. He was dead set against a counterinsurgency strategy and a large infusion of U.S. troops. He said that adding troops was counter to “Afghanization” and “civilianization” of the mission. In his view, Karzai was not an adequate strategic partner, we overestimated the ability of the Afghan forces to take over security, and more troops would only deepen the dependency on us by the Afghan forces. He complained about the lack of a civilian counterpart to the commander of ISAF (McChrystal) and said that he ought to fill that role, not the UN or NATO senior representatives. Eikenberry recommended that we study the situation for several more months while proceeding with development projects.

I thought his recommendations were ridiculous. Analyze for four more months? How do you do development projects without security? The cable ruptured the relationship between McChrystal (and the senior military) and Eikenberry once and for all, both because of the substance but also because Eikenberry had never mentioned either his views or his cable to McChrystal.

On Veterans Day, November 11, we went around and around again on the options. The next day the president called me from Air Force One on his way to Singapore. He said he was focusing on sending two brigade combat teams and not considering a third until the summer of 2010, when we could see what Karzai had done and how we were doing. I urged him to approve all three brigades to demonstrate resolve and to ease military planning. He could then “off-ramp” the third depending on Karzai’s actions. He said he’d think about it. He then asked: Were any of the 17,000 in the first tranche of surge troops being deployed because they were needed to support a force of 40,000 more troops? Could some of the third brigade be brought forward to strengthen the second? What were the benchmarks of progress? Could we accelerate both the buildup and the drawdown? How should we treat the civilian and military components together? He said he had posed those questions to the NSS.

As he prepared to end the conversation, he asked me how to adjust his body clock for the Asia trip. I said, “The old-fashioned way—alcohol and Ambien.” He laughed and said he’d break out the Johnnie Walker.

On November 13, I invited Emanuel and Denis McDonough to my office to review the president’s questions and to make sure Defense and the NSS understood the questions in the same way. They brought Lute with them, and I was joined by Mullen. Rahm told me that my comments on sending three brigade combat teams had gotten the president’s attention. I said that sending two brigades would look like the president simply split the difference between zero and four and that, after two months of deliberation, that would be characterized as a “gutless” decision. I was sure that when senior military leaders were asked by Congress to give their personal, professional opinion, they would say two brigades were not enough.

I do not leave big issues to chance. On the fourteenth, I called Hillary in Singapore, told her of my telephone conversation with the president, explained what I’d said about three brigade combat teams, and asked whether she still supported that. She affirmed her strong support and then asked, “Where’s Jones?” I said I didn’t know, that he had been slippery on this whole matter. She agreed. I told her I’d called because the president might make a decision on his trip and she would be the only strong voice present. She laughed and said she’d do her best.

Our last meeting was on November 23, from eight to ten in the evening. The stage had been set by two opposing papers. The NSS paper recommended that the president approve two brigade combat teams (about 20,000 troops, Biden’s proposal) and reserve a decision on the third until July 2010. Mullen, in opposition, wrote a memo to the president, which he sent to Jim Jones, reasserting in the first sentence the need for 40,000 troops; McChrystal was equally adamant. Their unyielding views angered Biden, Jones, and the NSS and portended a break between the president and the military. Mullen was traveling in Europe. When I caught up with him, I said I thought that he, Petraeus, and McChrystal had agreed that my alternative of 30,000 troops plus more allied troops was workable. Mike decided to tear up his memo and redraft it. Fortunately, I had told Jones not to give the original Mullen memo to the president.

The meeting that night was straightforward and lacking in drama. Mullen, Petraeus, and McChrystal were candid in what they said but flexible, affirming their support for whatever the president decided. Hillary strongly supported McChrystal’s approach, with UN ambassador Susan Rice, Deputy Secretary of State Steinberg, Mullen, Cartwright, McChrystal, Petraeus, and me supportive of the “maximum leverage” option (my alternative). Biden, Donilon, and Brennan were all opposed. Eikenberry supported more troops but was skeptical that counterinsurgency would work because of the deficiencies of the Afghan government. Emanuel spoke mainly about the political challenge of getting the money to pay for a surge, and the impact on public opinion, health care, the deficit, and other programs. He said getting congressional approval would be difficult.

On November 27, the day after Thanksgiving, the president called me at home in the Northwest for a long talk. He was fine with the 30,000 troops with flexibility “in the range of 10 percent” for additional enablers, but he would not agree to the requests for 4,500 enablers unrelated to the new deployments that had been stacking up on my desk for over two months. He said that pushed the total number to 37,000, which would be hard to sell on the Hill, and it was too close to McChrystal’s number for the general to understand he was being given a different number and a different mission. “I’m tired of negotiating with the military,” he said. When I expressed my dismay, saying that I’d held off on this need for months pending his decision and now I would have to provide those needed enablers “out of hide,” he responded that McChrystal ought to be able to find the necessary troops: “Doug Lute tells me there’s a lot of tail to tooth [troops in support roles as opposed to fighters] in Afghanistan.” He asked me to return to Washington early for a meeting with him, Mullen, Cartwright, and Petraeus to make sure they were on board: “If they aren’t, I will revert to McChrystal’s option of 10,000 mostly trainers.” We agreed to meet at five on Sunday.

To prepare for that meeting, I held a videoconference with Mullen and Cartwright on Saturday morning and brought them up to speed. “Stan needs to grasp,” I said, “that there has been a shift in mission.” I repeated the president’s threat to go back to McChrystal’s smallest troop option. I came away from the videoconference believing everything was okay but worried nonetheless about what McChrystal might say the next day.

I complained to Jones that afternoon that the NSS decision paper for the president was trying to place precise caps on troop numbers, particularly the 10 percent flexibility the president had given me. I told him they should write it just as the president and I had agreed. I then raised the additional 4,500 enablers I had discussed with the president. Jones said he thought the president had just forgotten about them when meeting with “the acolytes” on Friday. He went on to say that “those guys—Emanuel, Axelrod, Donilon, and McDonough—were really deeply involved and stirring the pot.” He said he was isolated in the meetings.

I received word that same afternoon that the Sunday meeting with the president had been changed to nine-thirty a.m., thus requiring me to fly all night from the West Coast to make it. I saw the handiwork of the NSS in this and told my staff, “Tell them to go fuck themselves. The president and I agreed on five and that’s when I’ll be there. If they go at nine-thirty, they’ll do it without the secretary of defense.” The meeting was changed back to five.

The meeting was unlike any I ever attended in the Oval Office. Obama, Biden, Mullen, Cartwright, Petraeus, Emanuel, Jones, and I were there. Obama said he had gathered the group principally to go through his decisions one more time to determine whether Mullen and Petraeus were on board and fully committed. He said that if not, he would go back to McChrystal’s option of 10,000 troops, the option favored by most of his civilian White House advisers. He then went around the room. Mullen and Petraeus said what he wanted to hear. Emanuel—no surprise—stressed the political lift on the Hill and the danger of any daylight between the president and the military. Jones and Cartwright were supportive. I, of course, was pleased to hear my proposal being adopted.

Then there was an exchange that’s been seared into my memory. Joe Biden said he had argued for a different approach and was ready to move forward, but the military “should consider the president’s decision as an order.” “I am giving an order,” Obama quickly said. I was shocked. I had never heard a president explicitly frame a decision as a direct order. With the American military, it is completely unnecessary. As secretary of defense, I had never issued an “order” to get something done; nor had I heard any commander do so. Former chairman of the Joint Chiefs of Staff Colin Powell, in his book It Worked for Me, writes, “In my thirty-five years of service, I don’t ever recall telling anyone, ‘That’s an order.’ And now that I think about it, I don’t think I ever heard anyone else say it.” Obama’s “order,” at Biden’s urging, demonstrated, in my view, the complete unfamiliarity of both men with the American military culture. That order was unnecessary and insulting, proof positive of the depth of the Obama White House’s distrust of the nation’s military leadership.

The president announced the troop surge, at West Point on December 1. Clinton, Mullen, Jones, and I accompanied him.

In the end, I felt this major national security debate had been driven more by the White House staff and by domestic politics than any other in my entire experience. The president’s political operatives wanted to make sure that everyone knew the Pentagon wouldn’t get its way. Jones had told me David Axelrod was backgrounding the press to that effect. I thought Obama did the right things on national security, but everything came across as politically calculated.

After the president’s announcement, I wrote a note to myself: “I’m really disgusted with this process, I’m tired of politics overriding the national interest, the White House staff outweighing the national security team, and NSS (Donilon and Lute) micromanagement. May 2010 is looking a lot more likely than January 2011 [in terms of when I would leave]. I’m fed up.” When I wrote that, I was frustrated with a valuable process that had gone on way too long.

To be fair, though, national interest had trumped politics, as the president made a tough decision that was contrary to the advice of all his political advisers and almost certainly the least popular of the options before him in terms of his political constituents.

On reflection, I believe that all of us at the senior-most level did not serve the president well in this process. Our “team of rivals” let personal feelings and distrust cloud our perceptions and recommendations. I believe, for example, that my view of a geographically limited counterinsurgency, combined with aggressive counterterrorism and disruptive Special Forces attacks on Taliban leaders, emphasizing expansion and training of the Afghan security forces, was actually pretty close to what Biden had in mind. The difference between his recommendation for increased troops and mine was the difference in total force between 83,000 to 85,000 troops and 98,000 troops. His number was far above what was required for counterterrorism, and mine was far too small for a fully resourced counterinsurgency strategy. The aggressive, suspicious, and sometimes condescending and insulting questioning of our military leaders—especially by Donilon, Lute, and others at the White House—made them overly defensive, hardening their unwillingness to compromise. White House distrust and dislike of Holbrooke, and Lute’s preoccupation with the military side of the equation, contributed to inadequate attention to the civilian component of the Afghan effort. Contending teams presented alternatives to the president that were considerably more black and white than warranted. A more collegial process, one that tried to identify points of agreement rather than sharpen differences, would have had a more harmonious conclusion and done less damage to the relationship between the military and the commander in chief.

Responsibility for finding the common ground and shaping the deliberations accordingly would normally fall to the national security adviser, Jim Jones. The National Security Staff is supposed to be the “honest broker” in the policy-making process. That was not the case in the Afghan debate. Jones’s views, and the even stronger opinions of his deputy, Tom Donilon, and Lute, made the NSS an advocate rather than a neutral party, contributing to a damaging split in the government, with the White House and NSS on one side and the Defense and State Departments on the other.

My anger and frustration with the White House staff and the NSS during the process led me to become more protective of the military and a stronger advocate for its position than I should have been. In retrospect, I could have done more to bridge the differences. Fairly early in the process, after I had talked about a narrower Afghan mission in a principals meeting, Biden wrote me a note at the table saying, “What you outlined is what I’ve been trying to say.” We had breakfast together once that fall at his residence to discuss things, but I could have reached out privately to him more often to find common ground. I don’t think we would have agreed on the number of additional troops, but I believe we could have come pretty close on the strategy; that alone would have helped avoid a lot of acrimonious debate.

The rift on Afghan policy would linger for the rest of my tenure as secretary. Biden, Lute, and others in the White House who had opposed the decision would gather every negative bit of information about developments in Afghanistan and use them to try to convince the president that they had been right and the military wrong. That began before the first surge soldier set foot in Afghanistan.

In the middle of our debates over Afghanistan, a tragedy at home was a vivid reminder of the complex dangers we were facing. On November 5, Army Major Nidal Malik Hasan turned on his fellow soldiers, murdering thirteen people and wounding twenty-nine others in a shooting rampage at Fort Hood, Texas. It was the worst such attack ever on a military base in the United States. Hasan had expressed extremist Islamic views and had been in contact with Imam Anwar Al-Awlaki, an advocate of extremist violence residing in Yemen. Hasan’s attack on fellow soldiers was a wake-up call for the military to look closely at its own ranks and especially to question why Hasan’s expression of extremist views had drawn little scrutiny. The president spoke eloquently at the memorial service at Fort Hood.

Before the service, I met separately with each of the families to express my sympathy and condolences. The father of one victim, Specialist Frederick Greene, invited me to attend his son’s funeral in Mountain City, Tennessee. I had wanted to attend the funerals of fallen heroes in their hometowns since becoming secretary, but I had not done so out of concern that my presence would be a distraction and intrude on the privacy of families. I decided to accept Mr. Greene’s invitation. Mountain City, a town of about 2,400, is in the far northeastern corner of the state. The nearest airport is near Bristol, Tennessee, about an hour’s flight from Washington. I flew there on November 18 with two military assistants (and the always-present security team). I took no staff, no press. We drove across three mountain ridges to get to the remote town. Flags seemed to be hanging from every building. There were many signs acknowledging the life and sacrifice of Specialist Greene. We drove through Mountain City into the countryside to Baker’s Gap Baptist Church, a simple but picturesque country church. It was windy, cold, and rainy. The service was at the church cemetery on an adjacent hill. I met with the family privately in the church and then took my seat at graveside under the funeral tent. Fred Greene’s wife and two young daughters sat immediately in front of me. As the service proceeded, I could see in my mind’s eye other cemeteries in numberless small towns across America, where families and friends had buried local sons who had risked everything and lost everything. When the service ended, I shook hands with the members of the Army honor guard and made the long drive back to the airport.

A little over two weeks later, as I signed the deployment order sending the first 17,000 troops of the surge to Afghanistan, my thoughts returned to that bleak hillside in Mountain City.