Hours before the second deadline for Iraq’s new constitution on August 22, Shiite and Sunni Arab leaders met in a conference room at the Baghdad headquarters of Kurdistan’s President Massoud Barzani. The Shiites wanted the constitution’s preamble to mention Saddam Hussein’s atrocities and the Sunni negotiators were objecting. Guests sipping tea in the adjacent reception room heard voices rise in anger, and then Nabeel Musawi, a Shiite parliamentarian with a long record as a human rights campaigner, came out of the meeting. “The Sunnis,” he said, claim that “Saddam only killed five farmers in the south and some Kurds.” Nabeel’s father disappeared after being arrested by Saddam’s security services in 1981, one of 300,000 Shiites murdered by the Baath regime during its thirty-five years in power. Another deadline was missed.

Three days later, President Bush telephoned Abdel Aziz al-Hakim, a Shiite cleric who leads the Supreme Council for Islamic Revolution in Iraq (SCIRI), Iraq’s largest and most pro-Iranian political party, to ask for concessions on behalf of the Sunni Arab negotiators on the controversial issues of federalism and de-Baathification. Hakim politely thanked the President who, not being well versed in the intricacies of Iraqi politics (or even its broad outlines), was reduced to pleading that his requests be taken seriously. The President then said something about protecting women’s rights and Hakim assured him they were sacred.

The call was pointless. Bush was asking Hakim to make concessions that the Sunni Arab negotiators themselves did not consider sufficient. Hakim’s idea of women’s rights is very different from what Bush wanted, but the President did not know enough to respond to the cleric. The Hakim episode reveals just how clueless the President and his advisers are about the divisions in Iraqi society. Small concessions cannot paper over the differences between the victims of horrific atrocities and those who deny that any crimes took place. There was also no small amount of hypocrisy in the President’s expressions of concern about women. His diplomats had already agreed to soften key protections for women, and two days before expressing his concern to Hakim, Bush had publicly congratulated Iraq on “a democratic constitution that honors women’s rights.” While the President’s personal intervention into the Middle East bargaining was predictably feckless, his ambassador to Iraq, Zalmay Khalilzad, did much to produce the constitution that emerged. Days after taking up his post in early August, Khalilzad summoned Iraq’s top leaders to the capital’s Green Zone, initiating three weeks of nonstop talks that produced the Kurdish–Shiite deal that is expressed in Iraq’s new constitution.

The Shiite and Kurdish leaders who negotiated Iraq’s new constitution did so on the basis of a strong electoral mandate from their voters for the positions they advocated—the only undemocratic part of the process was the Bush administration’s inclusion of unelected Sunni Arabs. But the administration’s insistence on the deadlines in Iraq’s interim constitution—an insistence related to the President’s plummeting poll numbers and not to the concerns of Iraqi leaders, all of whom wanted more time—effectively prevented meaningful participation by Iraq’s peoples or its elected National Assembly. The resulting constitution is not the one that neoconservatives dreamed might precipitate a democratic revolution in the Middle East. In particular, the constitution fails to adequately protect religious freedom and the rights of women.

But the constitution might bring stability to Iraq, a country now on the edge of full-scale civil war. Underneath an Islamic veneer, Iraq’s new constitution ratifies the division of Iraq into three disparate entities: Kurdistan in the north, an Iranian-influenced Islamic state in the south, and, in the center, a Sunni region that has no clear political identity, but that with luck and concerted diplomacy could be governed by a new generation of Sunni Arab leaders. The constitution provides a basis for resolving Iraq’s most contentious issues: oil, territory, and the competition to be the dominant power in Baghdad. If these issues are not addressed, they could set off a widespread civil war. Whether Iraq nominally stays together or formally breaks apart, it was important to find a formula that could reduce the likelihood of a full-scale conflict. The constitution has many flaws, but it provides a peace plan that might work, and it is therefore the most positive political development in Iraq since the fall of Saddam Hussein from power.


At first glance, Zalmay Khalilzad seemed an improbable mediator for this Iraqi constitution. An Afghan-American whose Republican links go back to the Reagan administration, Khalilzad openly shares the fierce political partisanship that characterizes the Bush administration’s efforts in Iraq. In the build-up to the Iraq war, Khalilzad was associated with the neoconservative cabal that plotted the war and then failed to plan for its aftermath; for seven months he acted as President Bush’s envoy to the Iraqi groups that opposed Saddam Hussein. He came to Iraq after serving as US ambassador to Afghanistan, where he was widely known as “the viceroy.” Khalilzad had little time to master the complexities of Iraq’s politics, and on some issues it showed.


If Khalilzad arrived in Baghdad still believing in the Bush administration’s formula of a “democratic, federal, pluralistic and united Iraq,” he swiftly caught on to the reality. Shuttling from faction to faction, he approached the process of drafting a constitution as not so much an exercise in “nation building” as a negotiation of a tripartite peace treaty, which is largely what it was. (I was struck by the similarities to the three-week Dayton talks ending the Bosnian war, in which I took part in 1995. In Baghdad, as at Dayton, the three factions never engaged in substantive negotiations as a group. Instead, agreements were worked out in bilateral talks between two of the factions or by American diplomats shuttling among the residences of the leaders.)

Khalilzad faced two major obstacles: the distrust of all Iraqi parties for one another and the incoherent US policy that preceded his arrival. The Kurds saw the constitution largely as a threat to their continued independence, and examined every proposal from that perspective. As the majority faction, the Shiites controlled the drafting of the text—and whether through inexperience or self-serving intentions, they often simply disregarded agreements others thought had been reached. Naturally, this fed Kurdish suspicions. The Sunni Arabs objected to practically everything that was proposed, frustrating the Shiites and Kurds to the point that they stopped negotiating with them. In the end, Khalilzad had the US embassy prepare drafts, record agreements, and incorporate them into the text.

Khalilzad inherited policy decisions—made both in Washington and in Baghdad—that complicated his task. In May, Condoleezza Rice flew to Baghdad to insist that Sunni Arabs, who had boycotted the January elections and were therefore only minimally represented in the National Assembly, be included as members of the constitution drafting committee. Such inclusiveness in constitution-making is, of course, desirable; but senior Iraqi officials—some of them Sunnis—warned that the particular Sunnis selected by the US to represent their community were not prepared to take part in serious dialogue; nor for the most part were they representative. These warnings were ignored.

The Sunni delegation represented a variety of views, but it was dominated by former members of the Baath Party. The group’s spokesman and de facto leader was a former Baath Party functionary, Saleh al-Mutlaq, who argued against nearly everything that was proposed, and did so in an aggressive way that offended the Kurds and Shiites and some of his fellow Sunnis. Also on his team were Saddam Hussein’s former translator and several other former Baathist functionaries, as well as representatives of the Iraqi Islamic Party, a Sunni religious party. The leaders of the Iraqi Islamic Party ended up supporting the new constitution, but their voices have been drowned out in the anti-Shiite, anti-Kurd rhetoric of the others.

By pandering to unelected former Baathists, the Bush administration made them appear as more authentic representatives of the Sunni Arabs than those Sunnis who had actually been elected, including Iraq’s Vice President Ghazi Yawher (from one the country’s largest Sunni Arab tribes) and the speaker of the National Assembly, Hajem Hassani. Although they were not part of the Sunni negotiating group, both were inclined to agree to a compromise, and Hassani, a liberal who had spent years in California, objected not to the provisions on federalism or de-Baathification but to the inadequate protection of the rights of women. By refusing to compromise, the Sunni Arabs selected by the US forced protracted delays that both emphasized the dominant American role in preparing the constitution and undermined its legitimacy in Iraq and internationally. Their fierce denunciation of the outcome has intensified Sunni Arab hostility to the draft.

The Bush administration also made other mistakes, some of which bordered on the bizarre. Although the administration would have to rely on the pro-Western Kurds to support US positions in the negotiations, US diplomats went out of their way to offend them. The US embassy office in Kirkuk was instructed to snub a Kurdistan government–hosted July 4 reception, unless the Kurds flew the Iraqi flag. The Kurds, who associate the flag with Iraqi genocide, canceled the reception. A few days later, the US embassy’s political counselor, in talking to the foreign press, denigrated Kurdistan’s constitutional proposals, comparing the Kurdish leaders to carpet sellers who set a high price with the intention of settling for much less. Kurdistan’s President Massoud Barzani had the last laugh, since almost all of his proposals were accepted.


The historically contentious Kurds entered the negotiations with advantages that neither the Americans nor the Shiites fully appreciated. Barzani assembled a unified delegation that included Christians, Turkomans, Yezidis, Islamists, and Communists. Shrewdly, he obtained a mandate from Kurdistan’s parliament that gave him no room for compromise on the region’s basic demands: the supremacy of Kurdistan law over federal law, acceptance of the peshmerga, the Kurdish guerrilla army, as the official military force of the region, control of natural resources, and a formula to resolve territorial disputes, particularly the control of Kirkuk. Most important, the Kurds did not need a constitution at all, since their autonomous state already existed.

The Shiites were less well organized and considerably more divided. They broadly agreed that the constitution should define Iraq as an Islamic state and, except for the radical cleric Moqtada al-Sadr, who did not participate, they agreed on the principle of federalism. But the Shiites disagreed on many details. Some wanted two or three Shiite states in the south, including one centered in Basra and another for the holy cities of Najaf and Karbala, while the SCIRI leader, Abdel Aziz al-Hakim, advocated a single Shiite superregion consisting of all nine Shiite governorates. Some Shiites supported the Kurdish position that oil should be controlled by the regions—80 percent of Iraq’s known oil is in the Shiite south—while the prime minister, Ibrahim Jaafari, wanted oil to be controlled by the central government. These differences made it hard to work out agreements with them. Khalilzad regularly intervened to work out a deal, including most notably the one that divided Iraq’s oil between the federal government and the regions.

The Kurds viewed the Iraqi constitution largely as if it were intended for a foreign state. As a result, they were not prepared to block a deal because of concerns to protect secularism and gender equality for others as long as any objectionable provisions about either one did not apply to Kurdistan. (The Kurds were also reluctant to defend these principles after US diplomats had already agreed to the more Islamic formulations.) The Shiites were mostly willing to concede that the Kurds (and any other region) could legally opt out of many provisions of the constitution because they knew this to be the price of having the constitution endorse Islamic law.


Iraq’s new constitution creates a highly decentralized state. The federal government has exclusive powers with regard to foreign affairs, defense policy, monetary policy, and fiscal policy (but not taxation), and it can assign broadcast frequencies. All other powers belong to Iraq’s regions, or are shared between the regions and the federal government. Except for matters within the exclusive jurisdiction of the federal government, regional law prevails when in conflict with federal law. The regions can amend or cancel the application of any federal law on their territory.

Clearly the balance of power favors the regions, especially since they can also maintain their own military forces (called Regional Guards). The federal government is to manage existing oil fields in cooperation with the regions, but oil fields not yet in production are the sole responsibility of the regions. Water originating outside Iraq—including the Tigris and Euphrates Rivers—is controlled by the federal government, while the regions control water that originates on their territory, a provision that gives the Kurds legal control over their plentiful water resources for the first time in their history.

The constitution recognizes Kurdistan as an existing federal region, but leaves the procedures for forming future federal units in the center and south to the next elected National Assembly. This deferral of future federal units is the compromise Bush futilely hoped would placate the Sunni Arabs. But, of course, the Sunni Arabs knew the deferral would not block new regions, since a new assembly with full Sunni Arab participation will still be dominated by an overwhelming pro-federal Shiite and Kurdish majority.

The constitution states that “Islam is the official religion of the state and is a basic source of legislation”—a generally stated establishment of religion that is commonplace in the Muslim world. More troubling is the constitution’s clause prohibiting any law that “contradicts the established provisions of Islam.” Since there is no agreement on what constitutes these “established provisions,” Iraqi secularists fear that this language will allow clerics to impose their own interpretation of Islamic law. The US strongly and successfully objected to a Shiite proposal for a constitutional court (which had similarities with Iran’s clerical Guardian Council), but supported the inclusion of clerics and Islamic experts on the federal Supreme Court.

The Kurds, and other secularists, were particularly appalled by the idea of clerics on the court, but since they had no support from the US, they chose not to make an issue of it. Instead, the Kurds stripped the Iraqi Supreme Court of jurisdiction over Kurdistan’s laws. Here the Kurds’ negotiators were influenced by US constitutional experience. Having seen US justices decide the election of 2000 on the basis of their personal political preference, they had no confidence in US arguments on the value of an independent judiciary.

The provisions on the rights of women are confusing and unworkable. Each Iraqi, for example, can decide whether he or she wants disputes over personal issues such as divorce or inheritance settled according to his or her sect’s religious law or according to the secular civil code. Since Islamic law generally favors men (a sister gets half the inheritance of a brother), there is an obvious conflict if one party chooses religious law and the other the civil code. The constitution does not address this.

Still, for all the attention that has been given to the constitutional provisions concerning women and Islam, the federal constitution is largely irrelevant to the actual treatment of women and the application of Islamic law. Regional constitutions and law will, according to the federal constitution, have primacy concerning these matters. This arrangement enables Kurdistan to preserve its secular status and to keep human rights protections in its constitution that are superior to those in the federal constitution. But this also means that the Shiite region (or regions) will apply a much stricter version of Islamic law, particularly in the treatment of women, than exists in the federal constitution.


The new constitution has received almost uniformly negative reviews in the United States (except, of course, from President Bush, whose opinions on Iraqi developments are no longer taken seriously). Editorials in the pro-war Washington Post and the anti-war New York Times both criticized the constitution for its loose federalism, with the Post arguing that the provisions on oil and regional militaries could lead to civil war or partition. In fact, it was a centralized Iraq that provoked the lengthy Kurdish and Shiite rebellions of the past, and any effort to recreate a more centralized state now would be forcibly resisted or ignored.

But such criticisms ignore the fact that Iraq’s Kurds don’t want to live in pluralistic, multiethnic, centrally governed Iraq, and they don’t have to. For them, Iraq has been an eighty-year nightmare of repression and genocide. Almost unanimously, Iraq’s Kurds want their own country. They would certainly vote down any constitution that denied them the protection of their own army or the benefit of Kurdistan’s natural resources.

The Shiites do not want to live in a secular society. Religion is central to their identity, and in the January elections Shiites voted overwhelmingly for parties committed to creating an Islamic state. They seem prepared to limit such a state to the southern half of the country, but that is all they are willing to concede.

The Sunni Arab negotiators now seem to accept the reality of Kurdistan, but they argue that a federal arrangement south of the Kurdish border will mean the breakup of the country. In my talks with them, they seemed to hope that they could somehow return to the days when Sunni Arabs ran Iraq and were the major beneficiaries of its resources. Today a centralized Iraq would be one in which Shiites dominated Sunni Arabs, especially if the Kurds (who are mostly Sunni) refused to be part of such a state. Since sectarian war is already underway between Sunni Arabs and Shiites in Baghdad and some other parts of the country, it is hard to see how a centralized Iraq run by Shiites could serve the interests of its Sunni population. If federalism goes forward, the Sunni Arabs will almost certainly form their own region and take advantage of the provisions of the constitution that will permit them to maintain their own security forces.

The outcome of the Iraqi constitutional process will therefore very likely be the three-state solution that I described in these pages in May 2004.1 Iraq is well on the way to becoming a loose union of three separate and radically different states (or more, if the Shiites choose to divide themselves into two regions).

This is not, as the constitution’s critics suggest, a complete remaking of Iraq. It is merely the ratification of a breakup that has already happened. And far from igniting a widespread civil war, the constitution provides ways of resolving the very issues that could provoke such a war: oil and territory. The “old oil, new oil” compromise in the constitution stipulates that oil revenues from current production will be distributed equitably among Iraq’s regions, which means that the Sunni region will receive large infusions of money. And while current production is concentrated in the Kurdish north and Shiite south, all of Iraq’s regions have unexploited resources that are likely to produce considerable revenues for them in the future. In the recent negotiations, the fight about oil was not over revenue—the principle of sharing with the oil-poor regions was readily accepted by Kurds and Shiites—but over control. When controlled by Baghdad, Iraq’s oil revenues were used not only to finance development projects concentrated in the Sunni Arab heartland but also to pay for military campaigns that ravaged Kurdistan and the Shiite south. It is hardly surprising that both groups considered regional control of future oil development one of their nonnegotiable conditions.

The constitution also has a formula to resolve Iraq’s most enduring territorial dispute: between Kurdistan and the rest of Iraq over the oil-rich province of Kirkuk. The constitution includes mechanisms to return Kurdish victims of Saddam Hussein’s ethnic cleansing to Kirkuk and for a referendum to decide its status not later than the end of 2007. The United States could promote the peaceful resolution of the Kirkuk question by encouraging power-sharing arrangements among all of Kirkuk’s communities—Kurds, Arabs, Turkomen, and Chaldo-Assyrians. Whether the Americans are capable of the more informed involvement in Kirkuk’s ethnic politics that is now needed is not at all clear.

The strongest argument for the new constitution is that it could avoid civil war. But it has three other virtues: (1) it may hold the country together, (2) it limits Iranian domination to the southern half of the country, and (3) it provides for a more workable military strategy than the one to which the US is now committed.

Kurdish nationalism will not go away. It may be, however, that Kurds will settle for the indefinite continuation of their de facto independence (as allowed by this constitution) and not pursue the riskier option of formal independence. So long as Kurds feel they are under pressure to join in a closer union with Baghdad, the more intense will become their demand for legal independence.

Both secular-minded Shiites and Kurds told me that Iran was behind the positions on the constitution taken by the Shiite religious parties. The south, as I previously wrote in these pages,2 is now largely controlled by Iranian-sponsored militias and pro-Iranian Shiite political authorities. As long as there is a shared sense of identity between Iraqi and Iranian Shiites (and for Iraqi Shiites their religious identity seems much more important than their Arab one), Iran will have a heavy influence on Iraq. But since the central government will have only limited powers, the three-state solution will confine this influence to just part of the country.


President Bush’s military strategy for Iraq can be summed up by a phrase in his June 28 speech to the nation: “As the Iraqis stand up, we will stand down.” According to the Iraqis who run the Ministry of Defense, there is little hope that this will happen soon—or ever.

The Iraqi army nominally has 115 battalions, or 80,000 troops. This figure, often cited by those who see the Iraq occupation as a success, corresponds only to the number of troops listed on the military payroll. However, when the Ministry of Defense decided to supervise the payment of salaries, a third of the payroll was returned. (In Iraq’s all-cash economy, commanders receive a lump sum for the troops under their command; this acts as an incentive for them to maintain ghost soldiers on the payroll.) One senior official estimated that barely half the nominal army actually exists.

Claims about weapons provided by the US to the Iraqi army are even more doubtful. Iraqi Ministry of Defense officials say the Americans have not provided them with records of who has been receiving weapons. Without such controls, soldiers sell their weapons on the open market where some are bought by insurgents. Most weapons captured in recent months come, I am told, from stocks supplied to the Iraqi army and police. Craig Smith reported on August 28 in The New York Times that the US military is now unwilling to provide more sophisticated weapons to the Iraqi military for fear they will be used in a civil war—or against the US.

The problems with the Iraqi army go beyond the many opportunities for corruption. In this deeply divided country, people are loyal to their community but not to Iraq, and the army reflects these divisions. Of the 115 army battalions, sixty are made up of Shiites and located in southern Iraq, forty-five are Sunni Arab and stationed in the Sunni governorates, and nine are Kurdish peshmerga, although they are officially described as the part of the Iraqi army stationed in Kurdistan. There is exactly one mixed battalion (with troops contributed from the armed forces of the main political parties) and it is in Baghdad. While the officer corps is a little more heterogeneous, very few Kurds or Shiites are willing to serve as officers of Sunni Arab units fighting Sunni Arab insurgents. There are no Arab officers in the Kurdish battalions, and Kurdistan law prohibits the deployment of the Iraqi army within Kurdistan without permission of the Kurdistan National Assembly.

Even by paying soldiers salaries that are ten times the military salaries under Saddam Hussein, the United States cannot build an Iraqi army when there is no Iraqi nation. The effort should be abandoned in favor of supporting regional security forces. Thanks to their regional armies, Kurdistan and the Shiite south are stable and reasonably secure. A Sunni Arab military force—responsible not to a Shiite-dominated federal government or an American occupation army but to Sunni officers and a Sunni Arab political authority—is the best hope of combating the Sunni Arab insurgency and its jihadist allies.

A loose federalism will not solve many of Iraq’s problems. Baghdad, with 20 percent of the country’s population, is a mixture of Sunnis and Shiites, along with Kurds, Christians, and other minorities. Precisely because it is mixed, Baghdad is today the center of a dirty war between Sunnis and Shiites. Sunni insurgents target Shiite civilians in increasingly sophisticated and deadly bombing attacks while Shiite police—or, in some cases, militias in police uniforms—arrest Sunnis who end up as corpses, sometimes badly mutilated. Baghdad’s murder rate now exceeds one thousand per month, not including the dead from car bombs, and many of these are victims of sectarian conflict. Making Baghdad a special capital district—as is envisioned under the constitution—will not resolve the city’s conflicts, but it probably won’t make them worse. There is no guarantee that the Sunni Arab region would be able to stabilize itself, but even if it remains chaotic it is hard to see how it could be worse than it is now.

Before Iraq formally becomes a loose federation, the constitution must be ratified. With strong Kurdish and Shiite support, it will presumably win the approval of a majority of Iraqis. But it can also be vetoed if it is rejected by two thirds of the voters in three governorates. Sunni Arabs have a majority in four governorates, but according to the estimates I heard they can only muster the two thirds in the Anbar governorate—especially since the Iraqi Election Commission has ruled that the veto requires the negative vote of two thirds of eligible voters, and not of actual votes cast. Most Shiite politicians say that Moqtada al-Sadr, the radical Baghdad cleric who has denounced federalism, will not provoke the Shiite establishment by mobilizing his followers against the constitution. If he were to do so, the combination of his supporters and the Sunni Arabs might be able to veto the constitution.

If the constitution is rejected, Iraq’s interim constitution requires the political process to restart with December elections for a new interim government, which will have one year to write a new constitution. But there is no chance that Kurds or Shiites will make the concessions necessary to make a new constitutional bargain with the Sunni Arabs. Massoud Barzani told me categorically that “we will never accept less than what is in this constitution.” In an environment of escalating Sunni attacks on Shiite civilians, the Shiites will almost certainly feel the same way.

Most Kurdish leaders say that if the constitution fails, the next talks will be about partition. An independent Kurdistan is no longer unlikely. Arab Iraqi leaders understand that the Kurds want out, and are increasingly weary of having to pay the price for keeping them in. Even Saleh al-Mutlaq, the Sunni negotiator, has said in a recent interview, “If the Kurds want independence, they should ask for it.” Every Shiite leader whom I asked about the issue—including Deputy Prime Minister Ahmad Chalabi—said that they would support Kurdistan’s independence if that’s what the Kurds want. Some Arabs bluntly told me that, at this stage, they would prefer that Kurdistan left.

But the Shiites also may want out. Abdel Aziz al-Hakim’s proposal for a Shiite superstate could be taken as the basis for eventual separation; and the idea of such a state does not attract only religious Shiites. Some liberal and secular-minded Shiites now speak of a Shiite nationalism—as if Shiites were an ethnic rather than a religious community. They see Iraq as a failed state and do not want to spend their lives opposing an endless insurgency in the middle of the country.

If the current constitution is rejected, there will not be another one. And another government with a one-year mandate will not be able to begin to address Iraq’s deep political, economic, and security problems. For all its flaws, this constitution represents the last chance to hold Iraq together. The alternative is not a more centralized state. It is disintegration and chaos.

—September 7, 2005

This Issue

October 6, 2005