Three different civil wars are now raging in Iraq: the first between U.S.-led coalition forces and antigovernment insurgents, the second between the Kurds and other communities in northern Iraq, and the third between Sunni Arabs and Shiite Arabs in the center of the country. The last is the most important because it represents the greatest potential for humanitarian disaster as well as for long-term instability in Iraq and in the region.
Stephen Biddle offers the right diagnosis of the situation but the wrong prescription for treating it. The conflict between Sunnis and Shiites is, as Biddle argues, a communal civil war, not a war based on class or ideology, and the U.S. military's efforts to learn, or relearn, best practices for fighting a counterinsurgency from the Vietnam War are thus beside the point. But his proposal for communal power sharing -- which has been the Bush administration's policy since January 2006 and has become conventional wisdom -- is impractical. Power sharing rarely works well, and in Iraq its prospects are especially bleak: the Shiites are too strong to want or need to share power, there is too little trust between communal elites, and no institution in Iraq is capable of guaranteeing anything to anyone. Worse, the level of violence has passed the threshold where the communities can safely live together. At an earlier stage, this conflict might have been resolvable by compromise. But at this point, that no longer is possible.
Today, all members of both the Sunni and the Shiite communities face real security threats. The violence has escalated dramatically since the bombing of the Askariya shrine, in Samarra, on February 22, 2006, but it had been intensifying for several years. Sunni insurgents have been killing Shiite civilians since 2003, and since Shiite parties won control of the Iraqi government in early 2005, the Shiite-dominated police forces have often operated as death squads. As of late April 2006, the U.S. press alone had recorded 3,500 deaths over the previous two months, and the total number of actual deaths was probably higher. During that period, according to the Iraqi Red Crescent, more than 89,000 Iraqis became refugees. This estimate is likely low too, as it implies a ratio of deaths to refugees of about 1 to 20, and in ethnic-cleansing campaigns such ratios typically run closer to 1 to 100.
Today, no Iraqi Sunni is safe anywhere within the reach of Shiite militias or Shiite-controlled police forces, and no Shiite whom Sunni suicide bombers or assassination squads can get to is safe either. The danger is greatest and the violence worst where the two communities cohabit, as in Baghdad and in parts of the four surrounding provinces -- Anbar, Babil, Diyala, and Salahuddin.
And the situation will get worse, because communal atrocities have hardened sectarian affiliations. Before 2003, virtually all Iraqi Arabs identified themselves as Arabs, in opposition to Kurds and others. Since then, national and ethnic identities have not vanished, but they have been overshadowed by more specific, sectarian identities. Some 92 percent of the votes in the December 2005 elections were.cas2. for sectarian parties, and both communities now use increasingly extreme language, each describing the other in sweeping generalities.
Biddle recommends that Washington suspend its efforts to strengthen the Iraqi state until it can broker a grand bargain among all the communities, coercing them to compromise by threatening to manipulate their relative military power. In practice, such a policy would mean trying to force the United Iraqi Alliance (uia), the main bloc of Shiite religious parties, to surrender the victory it won in last December's elections. The idea would be to threaten to remove U.S. support for the Iraqi police and army if the forces remained split along sectarian lines and refused to reorganize based on loyalty to Iraq. The United States' trump card would be the threat to leave Iraq altogether. (Except for the last point, this essentially is current U.S. policy.)
This strategy is likely to fail. Attempts to compel power sharing among sectarian groups in Iraq will not stop the fighting and could even accelerate it. Prime Minister Ibrahim al-Jaafari has stepped down, but the uia has not split. Despite serious internal rivalries, Shiite leaders have redoubled their commitment to make key decisions among themselves before negotiating with the U.S. government or anyone else. In April, the uia chose Nouri al-Maliki to replace Jaafari. The main Kurdish and Sunni parties promptly accepted the nomination even though Maliki, whom they see as inflexible and excessively sectarian, was their least favorite candidate (their endorsement may reflect the fact that they have little leverage). Shiite leaders will retain control of the all-important Interior Ministry. In early May, it was still unclear whether the Defense Ministry would remain under uia authority or if its control would go to a technocrat not affiliated with the alliance.
A governing coalition has yet to be formed. But it might come to resemble the Kurdish-Shiite accord that underpinned the last government: Baghdad turned a blind eye toward Kurdish activities in the north in return for Kurdish acquiescence on anything the central government did elsewhere. A few Sunni ministers might be appointed, but the Shiites do not want -- nor do they need -- to offer significant concessions. Even if an all-party unity government could be formed, it would not be able to function; the parties' demands cannot be reconciled, and their mutual distrust is far too great.
Trying to create a genuinely Iraqi security force will not work either, because there is no powerful, legitimate political movement loyal to "Iraq," in or out of government. Nor could most members of the security forces be persuaded to identify with such a force if it did exist. Some Iraqi army units, under tight U.S. control, have been deterred from using violence for purely sectarian goals, but others are openly loyal to Kurdish or Shiite leaders. Reforming the police is a lost cause; any U.S. remark about the force's performance is met with heated retorts from Shiite leaders. In March, uia spokespeople demanded that U.S. forces stand aside from further involvement in internal security. Most Shiite leaders do not desire an immediate U.S. departure, but only because they hope to collect more U.S. aid before the civil war escalates further. An additional barrier to coercing Shiite leaders is the fact that Shiite militias are already receiving aid from Tehran on a moderate scale.
Any serious attempt to compel the Iraqis to share power would result in either a quick, ignominious reduction in U.S. troops or an actual U.S. withdrawal followed by a massive escalation of hostilities. Control over every mixed settlement and neighborhood in the country would be up for grabs, which would increase incentives for ethnic cleansing throughout the country. The Shiite-dominated Iraqi government might also find itself forced deeper into a clientelistic relationship with Iran.
In any.cas2., it is beyond the power of any Iraqi government to stop the violence between the communities if they are not separated first. Although the main Shiite militias are controlled by factions within the uia, they do not answer to it or to one another. The most active death squads seem to be those of the Badr Brigades, the armed wing of the Supreme Council for the Islamic Revolution in Iraq, which controls the Interior Ministry. Muqtada al-Sadr's Mahdi Army has also killed many people.
As a result, Iraq is breaking up into communal cantons. As they become unsafe, mixed towns and urban neighborhoods are becoming segregated. No one knows how far this process has gone already; some reports suggest that many towns have already become monoethnic. Shiite and Sunni militias have been inundated by new volunteers, and new independent neighborhood militias are forming, too. Free movement between Sunni and Shiite areas will be increasingly curtailed by checkpoints manned by militias, if not by government forces, as is already happening within and around Baghdad.
Iraq will eventually develop internal communal borders with a few heavily guarded crossing points. Since the ethnic makeup of Baghdad is far too complex for the city to be divided into just two parts, some of its neighborhoods will become isolated enclaves surrounded by barbed wire. This ugly solution has worked before: in Jerusalem, Mount Scopus was a Jewish island from 1948 to 1967. Any such partition of Iraq would likely be de facto, because many Shiite leaders still hope that a unified country can emerge, and no regime in the Middle East would tolerate formal independence for the Kurds.
In the meantime, the United States will remain the strongest military force in Iraq. As such, it will have one remaining duty: the moral obligation to minimize the damage, human and otherwise, caused by ethnic cleansing. This is also a U.S. national security interest: the U.S. government is -- and will continue to be -- blamed by most of the world for all of the harm that befalls the people of Iraq. The shorter that bill of indictment, the better.
Satisfying this obligation would mean using U.S. military strength to protect Iraqi refugees who wish to relocate. U.S. forces must defend the most vulnerable mixed towns and urban neighborhoods from both Sunni and Shiite attackers for long enough to organize transport for those who want to move to safer locations. Otherwise, who controls Baghdad and dozens, perhaps even hundreds, of towns in central Iraq will be determined by full-scale sectarian battles that could go on for months or even years.
Which settlements need to be defended and which communities need to be evacuated are questions that would largely determine the location of the de facto line that would separate Sunni and Shiite communities. Protection and relocation would have to be coordinated with the strongest forces in Iraq, the main Shiite factions. These groups would not be enthusiastic: two of the main uia factions -- the Dawa Party and the Sadrists -- still want a unitary Iraq. But sober Shiite leaders would also realize that such a policy would save many Shiite lives and bring the Shiite-dominated government greater control over more settlements than it could manage otherwise.
Little active cooperation would be required; all that would be needed is enough forbearance on the part of the Shiite militias to let temporary defensive garrisons and evacuation convoys complete their tasks without having to fight. Washington would have to explain its intentions clearly and establish firm limits to its mission both in aim and in time. The tolerance of the Sunni militias would also be needed in areas under their control. But if U.S. forces were scheduled to depart shortly -- leaving the affected settlements in Sunni hands -- the Sunni militias would have little reason to oppose the evacuation of those Shiites who wished to go. So far, few groups have displayed such bloody-mindedness as to suggest that they would take the risk of attacking U.S. forces solely to murder refugees in flight. (Afterward, the number of minorities living on the wrong side of the separation line would be small, which would limit incentives for "rescue" offensives.)
In the longer run, it will be important to ensure that the Shiites remain the stronger side militarily, as any change in the balance of power could encourage Sunni factions to challenge them again. The outcome of a civil war tends to be more stable when the party that is most satisfied is also the stronger one.
Some might say that this policy will legitimate ethnic cleansing. But they would have to face squarely the costs of not protecting refugees; to the extent that the policy did succeed, Iraqis would experience less suffering than if it failed or was never attempted. Others will object that the current U.S. administration is unlikely to adopt these measures. Perhaps, but saving at least some lives would require getting only a few brigade commanders in a few places to think seriously about refugee protection.
Such protection would not mollify the Iraqi Sunnis, who would still be out of power, or angry Sunni Arab governments. But no policy can prevent such discontent. It is also inevitable that whatever rump Sunni statelet remains will continue to be poor, disorderly, and unable to prevent terrorists from operating on its soil. Three years of counterinsurgency in Iraq has stimulated more terrorism than it has suppressed. But if Iraq's sectarian wars were ended, ordinary Iraqi Sunnis might come to realize that the greatest threat to their well-being is not Iraqi Shiites or U.S. troops, but foreign jihadists in their midst. Then, perhaps, they would begin to work at restoring order in their country.
Chaim Kaufmann is Associate Professor of International Relations at Lehigh University.
Copyright 2006 Foreign Affairs