Badr to Worse

There haven't been many days during the two-year U.S. occupation of Iraq that the country appeared to be stabilizing. But, on June 7, 2004, a Coalition Provisional Authority (CPA) official named Terry Kelly thought he was watching one unfold. That day, Iraq's newly appointed prime minister, Iyad Allawi, announced that nine prominent militia groups--commanding 100,000 fighters between them--had agreed to disband. The militiamen--mostly from the large and professional Kurdish peshmerga and the Shia mini-army known as the Badr Corps, which is loyal to the Supreme Council for the Islamic Revolution in Iraq (sciri), would either enter the new Iraqi security forces, receive retraining for new civilian jobs, or draw pensions befitting retired soldiers. The deal, Allawi said, was "a watershed in establishing the rule of law."For Kelly, it was something more. The West Point graduate had arrived in Iraq that February with the unenviable assignment of dismantling the militias-- a portfolio that no previous official had held despite nearly a year of U.S. control. He understood the importance of his task: A closely held strategy paper he had written shortly after he arrived stated bluntly that "[n]o state can exist and prosper if it contains armed organizations outside of legitimate government control." But making the strategy a reality took a Herculean effort. Kelly met constantly with representatives of Iraq's militias to convince them to relinquish their power--a difficult task given the threats they faced from a Sunni-dominated insurgency and a band of Shia fighters loyal to outlaw theocrat Moqtada Al Sadr. On June 7, it looked like Kelly had achieved the impossible.

Almost exactly a year later, however, Iraq's newly elected leaders annulled Kelly's work. At a June 8 Baghdad gathering of the still-intact Badr Corps, President Jalal Talabani, a Kurd, assured "the heroes of Badr" that they will have an enduring role in the new Iraq, provided they recognize the same holds for the peshmerga. "You and the peshmerga are wanted and are important to fulfilling this sacred task, to establishing a democratic, federal, and independent Iraq," Talabani said, a sentiment echoed by Prime Minister Ibrahim Jaafari and sciri leader Abdul Aziz Al Hakim. It was hardly unexpected for the Iraqi transitional government's ruling parties to bless the militias that they command: Among the Kurdish demands that delayed the formation of the government after the January election was a guarantee that the peshmerga remain intact; and Badr veterans control six southern governorships and Iraq's interior ministry, one of the country's most powerful institutions. The peshmerga and Badr are further protected by the fact that their masters are among the few political parties willing to work with an unpopular foreign occupier. "From the Iraqi perspective," Kelly explains, "they're the good guys."

But, this spring, the reason the United States feared the continued presence of the "good guys" has been on vivid display. In the multiethnic flashpoint city of Kirkuk, where the Kurds want to establish their capital, Kurdish forces have kidnapped at least 180 Arabs and Turkmen, prompting American diplomats to cable Washington urgently about actions that have "greatly exacerbated tensions along purely ethnic lines." Further south, Harith Al Dhari, the leader of the influential Association of Muslim Scholars, has publicly accused the Badr Corps of responsibility for a wave of assassinations that have targeted Sunni clerics. (Badr denies the allegation.) In addition to sectarian provocation, Badr, which has ties to Iran, has used its military muscle to ensure political supremacy for sciri, sending militiamen to polling stations and assaulting provincial officials considered disloyal. The militias, in other words, have become a potent force for interfactional violence and social repression.

This, however, may be only the beginning. Kelly--who attempted the last, best effort to give a new Iraqi government a monopoly on the use of violence, a prerequisite for a functional state--laments that there is now practically "zero chance" for removing the militias from Iraq's political landscape. When the United States had perhaps its most promising opportunity to demobilize the militias, right after the invasion of Iraq, it did little. This early mistake allowed militias loyal to U.S.-allied parties to entrench themselves, and it also enabled the formation and expansion of extremist and anti-American militias like Sadr's Mahdi Army. With U.S. influence in Iraq diminishing rapidly and sectarianism taking on an increasingly violent cast, militias have proliferated, with only vague lines distinguishing them from criminal gangs or insurgent groups. As Kelly wrote in his 2004 strategy paper, militias "necessarily create alternate power centers that compete with the established government [and] pose the constant threat of civil unrest or war." Across Iraq, that threat is rapidly materializing. "If there's a sectarian war," says Kelly's ex-boss, David Gompert, "these are the groups who will fight it."

When Larry Crandall answered his telephone in February of 2003, he learned that the Pentagon had not yet given any thought to the prospect of demobilizing either Saddam Hussein's army or the myriad guerrilla and militia groups fighting it. What security professionals call DDR--or disarmament, demobilization, and reintegration of armed factions in the aftermath of a conflict--is among the most arduous challenges of any nation-building mission. Since every conflict arises from, and ends under, unique circumstances, there aren't many lessons from past DDR operations to apply "off the shelf," as Crandall observes. He should know: As a veteran usaid official, Crandall participated in sometimes-stillborn U.S.-run DDR programs in Vietnam, Afghanistan, and Haiti. "But you see what happens when you don't have a DDR program," he contends. "You're going to have some level of armed confrontation with some elements."

On the other end of his phone was a military officer Crandall calls Colonel Jim, who was attached to the recently established Office of Reconstruction and Humanitarian Assistance (orha), the first incarnation of U.S. civilian authority for Iraq. Colonel Jim wanted to know if Crandall would be interested in sketching out a DDR program. "I asked, 'Why haven't you already done this work?'" Crandall recalls. "I didn't understand how screwed up [the reconstruction plan] was." Indeed, just weeks before U.S. forces invaded Iraq, the Bush administration had no idea what it would do with Iraqi soldiers and militiamen after the downfall of Saddam.

Nevertheless, Crandall signed up. In March, he flew to Kuwait City, where orha had set up headquarters at the Hilton Hotel in preparation for deploying to Iraq. In Kuwait, Crandall developed the contours of what he admits was "a pretty modest program." DDR offices would be established in the Kurdish north, the Sunni center, and the Shia south. Orha would inform ex-combatants that they were eligible for benefits if they reported to the offices for screening. There, an interview process would determine "whether they would go to jail, be pensioned out, kept in the service and perhaps retrained, or would go into a reintegration program, where they would be given a set of skills--computer training, for example--and released back into civilian life." The idea was to process the Iraqi military quickly and expand the program to the militia groups. "I didn't think this would necessarily work," he admits. "We wanted to start up the centers to see what would work on a modular, scalable, experimental basis. This was our best guess."

He never got to test it. Orha chief Jay Garner liked the plan, but, in early May, President Bush replaced Garner with Ambassador L. Paul Bremer, and orha became the Coalition Provisional Authority (CPA). Crandall subsequently learned he wouldn't be continuing on to Baghdad and wouldn't receive an anticipated $70 million DDR contract. His contacts in the CPA and the U.S. military told him that senior Pentagon officials thought "no taxpayer money should be spent on a defeated force." Days later, on May 23, Bremer issued CPA Order 2, which disbanded the Iraqi military without attending to the future of its nearly 400, 000 officers and troops. Crandall, stunned, thought that neglecting DDR for both Saddam's army and the militias was a dangerous mistake. But the incoming CPA director for national security, Walter Slocombe, had a more sanguine view of the militias.

When Slocombe, an undersecretary of defense in the Clinton administration, arrived in Baghdad that month, the United States was at the height of its post- invasion power. Many Iraqis still reserved judgment about U.S. intentions. The insurgency, while growing in strength and lethality, had not yet submerged Iraq into chaos. And it would be another two months before the CPA ceded even symbolic political power to the Iraqi Governing Council. Yet the inability to stop the initial wave of post-Saddam looting and the steady increase of violent crime in the months following the fall of Baghdad demonstrated that the United States didn't fully control the country. Iraqi politicians argued that their militias should be used to help keep the peace.

In public, the U.S. response was "unequivocal," as General Ricardo Sanchez, the commanding U.S. general in Iraq, told reporters in September of that year: "We are not going to allow private militias to operate in this country, and ... that has been communicated over and over again." But, behind closed doors in Baghdad, and on the streets in the Kurdish north and the holy areas of the Shia south, there was a different message. Slocombe says that he didn't think the militias were "a very big problem" at that point. Rather, he believed that, sometimes, allowing them to operate "was the right answer." "The police functions at these shrines should be done by religious police who basically work for the clerics who run the shrines," he says. "You can't have foreign troops doing it, and it's even tricky having Iraqi police, much less the [Iraqi] military or National Guard, do it."

Slowly, the larger militias and their political parties realized that the United States would not check their advances. Forces from the Badr Corps and the militia belonging to Iraq's oldest Shia Islamist party, Dawa, began to spread throughout the south. After the August 2003 assassination of sciri leader Muhammed Bakr Al Hakim, Badr marched black-clad militiamen to the shrine of Imam Ali, the holiest place in Shia Islam, while Najaf-based U.S. Marines groused about the "thugs" but did nothing. As the Badrists advanced, Shia clerics and local tribes convened their own bands of fighters to forestall sciri or Dawa domination. Most seriously, Sadr began building his Mahdi Army, which skirmished with Badr in Najaf for control of the Shia shrines and established its hold on the Baghdad slum of two million it renamed Sadr City. Indeed, as Sadr saw sciri grow stronger--with U.S. acquiescence--he felt the need to respond in kind. An Iraqi analyst told journalist Hassan Fattah for The New Republic in the fall of 2003: The United States "allowed the Badr Brigades to have some power, and now he wants some, too."

With the spread of the militias and the growth of the insurgency came an argument that Slocombe began to hear on the few occasions his office broached demobilization with the militia-owning political parties. The politicians referred to their militias as an "insurance policy." As they told Slocombe, "Until you guys can actually maintain security, don't tell us to take our patrols off the street." But the "insurance policy" the militias represented for their parties also applied to politics. The Governing Council parties told Slocombe that they had every intention of disarming--once a stable constitutional order developed. He sympathized. "To be honest," he remembers, "there was an effort to say, 'We're not so stupid as to think we can simply come into a community like Kurdistan and the Shia areas, which had been so badly treated by Saddam, and expect you guys to unilaterally disarm before your potential opponents have disarmed.'"

Slocombe encouraged Governing Council militias to enter the nascent Iraqi security forces, but he was opposed to pressing the issue. "Our policy was: We're prepared to solve this over time and in a gradual way with organizations that will keep their forces under control and not become part of the problem," he says. He considered it the only realistic approach. Iraqi parties "are not going to give up their military power, which they regard, and not entirely unreasonably, as their ace in the hole ... until they're confident that the political process and the security structure and all that is working in a way that's acceptable to them," Slocombe argues. "As long as they don't actively cause trouble, we were prepared to tolerate a gradual approach."

That is, as long as they didn't cause trouble for the United States. What they did to Iraqis, especially moderate Iraqis, was essentially tolerated. Across the south, in late 2003, CPA officials began to notice a worrisome trend. Badr and other militiamen would perform what was euphemistically called "spontaneous de-Baathification"--assassinating those suspected of being members of the former regime. In one particularly controversial episode, Shia militiamen opened fire on a woman believed to have been a high-ranking Baathist, killing her and her baby. Then, in February 2004, Badr sought political "insurance" by flooding a Najaf "caucus" station with chanting, abaya-clad women intended to look like potential suicide bombers, in the hopes of swinging the selection of an interim government toward sciri loyalists. Pro-democracy clerics told their CPA contacts that they feared for their safety.

Additionally, the twin principles of gradualism and ignoring militias that didn't fire on U.S. troops allowed anti-U.S. militias to fester until they were strong enough to attack. That's exactly what happened with Sadr. Late in 2003, Sadr organized a violent protest of pushcart vendors outside the Imam Ali Shrine as a pretext to insert his operatives into the sacred (and strategic) location. "We began to see the beginnings of what would be the Sadrist insurrection," recalls a southern-based CPA official.

But there was little appetite to confront the radical cleric. A bedrock assumption of U.S. military planning was that the true threat to the occupation came from Sunni insurgents. Security in the Shia areas "was taken for granted," as the southern-based CPA official explains, since the Shia were believed to be sympathetic to the United States. When some American officials argued for preemptively striking the increasingly powerful Sadr, they were rebuffed. "My reasoning," explains one such official, "was tantamount to saying we had to open up another front." Only when Sadr launched his insurrection in April 2004 did CPA and military officials see their error. Major General Martin Dempsey, commander of the First Armored Division that battled the Mahdi Army, told reporters in May of that year, "Clearly, in the six months between October and April, when [Sadr] instigated this national attack, he was training troops, gaining resources, stockpiling ammunition.... We missed the opportunity. We probably gave him six more months than we should have."; There were varying degrees of CPA awareness over what was happening with Badr outside of the Green Zone. When reports of Badr violence reached Baghdad, "you have to ask," says a CPA official, "is this person really affiliated [with Badr]?"

Although it would ultimately come too late to stem the growing assertiveness of Badr and the Mahdi Army, by the spring of 2004, the CPA's thinking on militias had already changed dramatically. Months earlier, in November 2003, the Bush administration announced that the Iraqis were to assume a measure of sovereignty in June 2004. There was no more time for "gradualism." When Slocombe ended his six-month tour, Bremer turned to rand Corporation expert David Gompert, a close friend since both men had worked for Henry Kissinger in the 1970s. Gompert reached Baghdad on December 10, 2003. Bremer immediately told him that, despite his other responsibilities, "I need you to get smart about and then tackle this militia problem. We can't wait any longer in trying to deal with it."

Gompert had a head start. Before he left for Iraq, a notice had circulated in the monthly rand newsletter informing interested colleagues that the CPA was seeking assistance. Among those who answered was former Army officer Terry Kelly. Though he didn't know Kelly well, Gompert was pleased: "I knew he was no- nonsense, a quick study, and could have sharp elbows when it was needed." It was needed often. When Kelly arrived in the Green Zone on February 6, 2004, Iraq's largest parties were unsure what political environment the CPA would bequeath them, making them extremely reluctant to draw down their armed forces. Kelly, despite a lack of experience with DDR, was told that he had all of two weeks to put together a policy.

Kelly and Gompert decided to focus "first and almost exclusively" on the militias controlled by Governing Council parties. Badr and the peshmerga represented the largest noninsurgent armed factions in Iraq, and the CPA was convinced it could work with their political masters. What's more, as they saw it, the policy should recognize that these anti-Saddam groups were, "in the truest sense, soldiers of the Iraqi people," which Kelly recalls was "tremendously useful both from a programmatic perspective and from a diplomatic perspective" when dealing with the militias. But their approach meant leaving Sadr out of any prospective arrangement, raising the question of what would happen to the Mahdi Army. It also meant that the CPA wouldn't negotiate with Sunni insurgents to see if they would cooperate. As Kelly explains, "If they're currently shooting at us, we kind of take it for granted that they don't want to engage with us." (This week, Donald Rumsfeld confirmed that American representatives have quietly held talks with insurgent-connected Sunnis in recent weeks in an attempt to draw them into the political process.)

An approach that Kelly and Gompert conspicuously ruled out--as had Slocombe-- was forcible disarmament, which they considered needlessly provocative and politically untenable. "The idea that we were going to huff and puff and rattle our saber and get all these guys to show up and stack arms was just unrealistic, " Gompert says. In order to determine what would work instead, Kelly had the benefit of an unlikely expert: Larry Crandall, who had returned to Iraq in January 2004 to help oversee reconstruction contracts. Not surprisingly, the program that Kelly designed was similar to Crandall's. Militias would be offered three choices for their fighters: positions in the security forces, subsidies and assistance for civilian retraining, or generous pensions befitting Army veterans. On February 22, Kelly presented what he called the Transition and Reintegration (T.R.) Strategy to Bremer and the U.S. military command.

Securing Bremer's approval was easy once Kelly had crafted the policy. Exponentially more difficult was getting the militias and their parties to buy in. Weekly discussions lasted hours. Gompert and Kelly soon heard every anti- demobilization argument that their predecessors had heard and more. "It's very different between the Kurds and the Badr Corps," says Gompert. "In the case of the Kurds, they're very security-minded. They think in terms of, 'How do we provide security for our people?'" From sciri and Badr, by contrast, Gompert and Kelly heard complaints about Coalition missteps, insufficient de- Baathification, and the rise of Sadr, which clarified something for Gompert: "They did not want to leave themselves without armed power at a critical stage in the development of politics, particularly Shia politics."

That led the CPA to seek different results from the Kurds and from sciri. "We never--never--suggested to the Kurds that they not have any people at arms, " says Gompert. "We never said you can keep the peshmerga, and we never asked them to give up the peshmerga." The CPA's objective was to persuade the Kurds to downsize the peshmerga, accept positions in the Baghdad security structure, and adapt the peshmerga to an internal Kurdistan-based protection force--which the Kurds secured for themselves in the U.S.-brokered interim constitution signed in March 2004. But getting the Kurds to agree to the T.R. plan had an additional purpose: enticing the Badr Corps into a process that CPA would use to phase out the Shia militia entirely. "The issue was always the Badr Corps," Gompert emphasizes, "because of the fear of what the Badr Corps would turn into: It would mutate from a former resistance group into a tool of Shia political Islam, a tool that could be used to influence both local politics and national politics."

Kelly was blunt with the Badr Corps' leader, Hadi Al Ameri. "Our philosophy toward all these groups is, as long as you abide by your commitments, we have no problem with you." There was an element of warning as well. Kelly explained that, if the U.S. military discovered Badr militiamen performing "unauthorized security functions"--assassinations, for example--their safety could not be guaranteed. But Kelly thought the CPA should recognize that local Badr commanders possessed significant autonomy from both the Badr leadership and from sciri. "We also made it clear that, just because some yo-yo down in Kufa or Najaf or wherever does something stupid, it doesn't mean we're going to throw the baby out with the bathwater."

This would prove to be a controversial decision. Days after a Najaf-based CPA official, Rick Olson, demanded that Badr's Najaf commander, Haji Hassan, cease his attempts to seize political control in the city, Olson's convoy was riddled with bullets by unknown assailants. He and his entourage were lucky to survive, and Olson had to leave Iraq. Furthermore, the distinction between Badr and the Mahdi Army was not always as ironclad as sciri contended. A CPA investigation found that Badr and Dawa militiamen assisted Sadr in ethnically cleansing a southern gypsy village known as Qawliyya that March, a collaboration first revealed in CPA democracy adviser Larry Diamond's new memoir, Squandered Victory.

There were varying degrees of CPA awareness over what was happening with Badr outside of the Green Zone. When reports of Badr violence reached Baghdad, "you have to ask," says a CPA official, "is this person really affiliated [with Badr]? Is this a group decision? Or is this someone behaving criminally? I was aware of misbehavior of people on the part of Badr--even fairly significant misbehavior--but I was not convinced this was a leadership decision.... I heard about Rick Olson. As far as I know, there is zero evidence that Badr did the ambush." But there's reason to question whether even confirmation of Badr lawlessness would have changed the calculation. For one thing, Badr militiamen did perform the "unauthorized security functions" Kelly warned the leadership against, with no reprisal. For another, the CPA badly needed sciri support for its shaky transition plans. As the southern-based CPA official recalls, "It was a complicated political dynamic in Baghdad. There were a lot of national-level equities, and there was a lot of balancing going on."

Two things ultimately drove sciri to the T.R. plan. First, it grew increasingly afraid of Sunni terrorist attacks against Shia civilians, which the Badr Corps could not prevent. Second, the U.S. response to Sadr's April uprising reassured sciri that the United States would oppose its most feared rival--while underscoring the frightening prospect of the U.S. military eventually turning its guns on Badr. Especially as it saw the Kurds win assurances of their place in the security apparatus, sciri officials turned to Gompert and Kelly and said, "We want opportunities in the security forces. We don't want to be discriminated against." On May 21, Bremer approved a final, detailed T.R. strategy. Shortly afterward, Order 91, issued less than a month before CPA disbanded, outlawed all militias that didn't enter into T.R. agreements with the Iraqi government. After half a year of nonstop work, Gompert and Kelly came home, cautiously optimistic that their plan would succeed.

But problems with implementing T.R. arose even before the Allawi interim government took the reins. "I never assumed sciri would comply," Gompert says. "My strategy was simply to try to get them involved in programmatic implementation" that would create its own momentum. What Gompert did assume was that, as he puts it, "The longer it took to dismantle the Badr Corps, the greater the danger that it wouldn't be dismantled"--and so part of the implementation plan was to disarm Badr more rapidly than other militias. But the plan ran into trouble almost immediately, when, in early June, the interim defense minister, Hazem Shalaan, balked at hiring former Badr Corps fighters, telling Gompert and Kelly, "I don't want Islamists in the army, period."

Then, in August 2004, Sadr launched a second insurrection in Najaf, which ended with the Mahdi Army largely refusing the government's demand to turn in its weapons and Allawi doing little about it. As a result, the dynamic that had drawn Badr closer to T.R. was effectively reversed. Sciri grew bolder about flouting its commitment to disarm. Last December, leader Hakim announced that, if necessary, he would order 100,000 militiamen onto the streets to safeguard the January 2005 election. Making matters worse, Allawi's administration was less interested in drawing the mainly Shia militias into the security forces than stacking them with rehired ex-Baathists, part of his plan to fight the insurgency. Indeed, the prime minister also formed his own militia forces-- Sunni-led counterinsurgent teams like the Special Police Commandos, under the control of his interior minister's uncle. (The U.S. military gingerly termed them "pop-ups," since, as one officer explained to The Wall Street Journal, militias are "illegal.")

Each move by one militia--and insurgent group--to gather strength demanded a countermeasure from its rivals. What Allawi once described as a "watershed for the rule of law" degenerated into a situation where the law played little part in the calculations of powerful political leaders intent on maximizing their advantage--in many cases, the same people entrusted with enforcing the anti- militia law. Under Order 91, any militia-wielding party is subject not only to criminal prosecution but also to potential ineligibility to hold office. Yet the Badr Corps itself has representatives in parliament. Not only is sciri a major component of the ruling United Iraqi Alliance bloc, but ex-Badr militiaman Bayan Jabr is now Iraq's interior minister--the official charged with monitoring T.R. compliance. The party's stated support for T.R. reached true absurdity in April, when now-Vice President Adel Abdul Mehdi of sciri lamented to Knight Ridder, "If you had a political crisis and each militia will go and support their party or political force, then you would have a very critical situation." Badr militiamen guarded his office as he spoke.

The cynicism of Mehdi's comment highlights the assumption behind the CPA's approach to the militias: To work, it required a virtuous circle of broader political development and improving security. If security increased around the country and Iraqis reconciled their deep religious and ethnic divisions, the parties would no longer require paramilitary "insurance policies." Originally, in 2003, the CPA figured the virtuous circle was already on its way; when it didn't materialize, in 2004, CPA security officials crafted the T.R. strategy to get it rolling. "When we say that we reached agreements with all these folks and we're solving the militia problems," says Kelly, "the understanding is the political and security situation must remain stable and continue to mature." Adds Gompert, "I'm not sure there was a strategy that would not have been dependent on a virtuous circle."

But the virtuous circle has given way to a vicious one. Badr intimidation, according to The Washington Post, has driven at least one provincial police force into open collaboration with the rival Mahdi Army. On the streets of Basra, Badr militiamen enforce Islamic dictates, leading one terrified Iraqi to compare them to Saddam's secret police in an interview with a Los Angeles Times reporter. Under Jabr, Badr personnel do join the Interior Ministry, but, as they do, one adviser alleged to Knight Ridder's Tom Lasseter, Sunni officials are being dismissed by the dozens, suggesting that what remains of Kelly's strategy is less integration than infiltration. The result is deepening anti- Shia sentiment among prominent Sunni organizations, such as the Association of Muslim Scholars, which is connected to the insurgency. In turn, as Sunni terrorism against the Shia has intensified, it has given birth to "bands of young men led by young Shia clerics who have had enough," as Nibras Kazimi recently described to The New York Sun. Such Shia militias have marched through Haifa Street, an arterial road in central Baghdad known by American soldiers as Purple Heart Boulevard, announcing, "If anyone in our neighborhood is killed, we will respond by killing people in Haifa."

The failure of the United States to disarm Iraq's militias is emblematic of its larger failures in Iraq--that is, the failure to have understood the enormity of the task of occupying and rebuilding the country. In Gompert's words, "It was only by the end of the CPA that we were beginning to do what we should have been ready to do from day one, to deal with a totally knowable problem. That's not a criticism of my predecessor, it's not a criticism of Jerry Bremer. If it's a criticism, it's a criticism that we as a country did not go in with all the tools that we were going to use." Or knowledge of how to use them. The U.S. military's enduring belief that the Sunni insurgency is the principal threat to Iraq argues strongly that additional troops would not have been used to discourage militia activity--cutting against the conventional wisdom that disaster might have been mitigated if only the administration had invaded with a larger force.

The idea that the allies the United States chose in Iraq could end up posing a threat of a different order--one that entrenches Sunni support for the insurgency and contributes significantly to the unraveling of Iraq--rarely influenced decision-making until it was too late. President Bush's meeting last week with Prime Minister Ibrahim Jaafari, whose Dawa party commands its own militia, suggests it still doesn't, even as the U.S. ability to change direction evaporates. Kelly contends, with some justification, that the plan he crafted can't be considered a "complete failure," since it enabled the militia- wielding parties to at least join the political process. "It was not clear when I got there that they were going to be part of the solution, and they are currently part of the solution," he says. "Or, at least, the solution that we're working with."

By Spencer Ackerman