Constitution Making Under Occupation: The Politics of Imposed Revolution in Iraq

“We’re short of time—it’s the fault of the Americans,” Kurdish politician Mahmoud Othman said. “They are always insisting on short deadlines. It’s as if they’re [making] hamburgers and fast food.” Othman added: “If we’d had more time, it would have been possible to get Sunni participation. When October 15 comes, many won’t even have seen the constitution.”¹

In light of subsequent history, there are striking and surprising differences of opinion concerning the “permanent” Iraqi constitution (ratified in October 2005) and how it was made. While some analysts (the Crisis Group)² consider the process and its result disastrous, others (Peter Galbraith)³ find both to be much superior to what took place and was achieved in the case of the American-imposed Transitional Administrative Law. I tend to agree with the former view, and I share its advocates’ concern that a constitution in several important respects worse than the TAL was eventually achieved. But I continue to think that the TAL remained a straightjacket in part responsible for the failure and many of its political consequences. At the same time, in two respects the making of the permanent constitution involved positive dimensions as well. First, a partially successful attempt at legitimation through national elections did occur, one that could have provided a completely new basis for democratic constitution making. And, second, an attempt was made, a seemingly serious but certainly an inconsistent one, to include important political actors of the Sunni Arab part of the population. This too was a very important departure from previous practice, even if it came much too late in the game. The very sad story is that the interaction of these two positive dimensions of the process exacerbated its negative features, which were already deeply embedded by the results of the previous stage. This produced a pathological constituent process, a constitution that once again was unable to solve the elementary problem of state rebuilding, and provided for the transition of the country to the civil war in which it finds itself at the time of my writing these lines.

In our normative universe, democratic elections are an irreplaceable part of modern, legitimate constitution making. Very early in the history of democratic constitutions, in the cases of the Federal Convention and the Assemblée Constituante, the delegates were not directly (in the first case) or homogeneously (in the second case, with its original estates) elected, and some critics and even some of the framers themselves considered this a serious problem.⁴ For different reasons, two of the most successful modern constitutions, the Grundgesetz and the Constitution of the Fifth French Republic, also did not involve new elections. Yet as a norm, new, direct elections of the constitutional assembly is nevertheless almost universal today, more so than the idea of ratification in a democratic referendum, for example. Empirically, the immense majority of constitution-making bodies have been democratically elected, at least in the twentieth century. Interestingly, the new two-stage method of constitution making described here and followed in Iraq complicates the picture. It involves the writing of an interim constitution by an unelected body, an interim constitution that will constrain to various extents the makers of the permanent constitution. Nevertheless, in every single example of the model, the assembly that makes the final constitution is freely and democratically elected and retains considerable freedom. Thus, while the model involves two constitutions, one made by a democratically elected body and one made by a different type of agent, the final product that comes out of the process is still supposed to be one made by a body consistent with the demands and recommendations of democratic theory.⁵ Electing a constitutional assembly specifically charged with drawing up a new constitution but under the rules of the interim constitution is the specific formula appropriate to the two-stage model.

The democratic legitimacy of this model depends therefore on relatively early, competitive, and inclusive elections, and this step at legitimation can only come in the second stage.⁶ That is why I write “relatively early,” despite many opinions that in conflictual societies after the collapse of dictatorships elections should be delayed. The first stage of the process always has legitimation problems, even if they are rarely as deep as they were in Iraq. In the second stage, the temptation that the newly elected assembly will vote to disregard the limits placed upon it will always be there, though remarkably such a repudiation, common, for example, in the case of the American type of constitutional conventions in Latin America, has not yet been seen in the two-stage model. Early elections may have something to do with this. Had the time period between the making of interim constitutions by bodies of elites regarded as not fully legitimate been greatly extended, this would have been popularly regarded in many cases as a usurpation, potentially dragging down the authority of the interim constitution in the process. Conversely, in several countries the fact that assemblies elected in relatively early voting chose to accept (explicitly or implicitly) the interim constitutions when they could have (either legally or in a revolutionary manner) amended or repudiated them helped to legitimate the overall process.⁷

In Iraq, free elections did take place as scheduled, and, as we will see, the TAL, the interim constitution, was confirmed with only a partial relegitimation being the consequence. The explanation of this outcome starts with the analysis of the elections themselves. To begin with, they were Sistani’s victory, despite many later claims on the part of the American government.⁸ Good thing too, otherwise the legitimation process I have in mind, partial as it was, would not have occurred at all. Sistani, as already shown, wanted an even earlier date than January 30, 2005, but he managed to get one early enough given his need to organize Shi’ite groupings as a single electoral list, which (working through his agents) he did successfully. The United Iraqi Alliance (UIA) brought together the SCIRI, Da’wa, and several smaller factions under a banner widely associated with the Grand Ayatollah himself. And the result that Paul Bremer tried to desperately avoid was indeed achieved: a Shi’ite victory in the general elections. The UIA received 48.2 percent of the votes cast, somewhat under their expectations but still receiving over 50 percent of the seats. Because of the two-thirds needed to elect a Presidency Council, they could not govern alone, but they were now clearly the major political force in Iraq, at least from the electoral point of view. And for the moment, the Kurdistan Coalition List, or the Democratic Patriotic Alliance of Kurdistan, with 25.73 percent of the votes and seventy-five seats, was by far the second strongest force in the country. They alone were in a position to block amendments to the TAL, even before the election of a Kurdish member of the Presidency Council.

To explain this somewhat inflated result, I first turn to the electoral rule and then to the problem of the Sunni boycott. The making of the rule was farmed out to a UN electoral team led by Carina Perelli.⁹ They chose a single countrywide district list PR for two reasons. First, and probably less important, they chose it because such a system is very proportional, and for a constitutional assembly governability issues should recede to an extent behind those of fair representation of political groups, interests, and streams of opinion that could become important in the foreseeable future. If it is true that the CPA went for the rule because of its desire to limit the vote of large Islamist parties and get smaller parties represented, for once they rather than Sistani’s circle had justice on their side.¹⁰ Excluding small parties via an electoral rule when the fundamental rules of the game are being negotiated is not fair, and it is indeed true that many small territorial lists could have had that effect.¹¹ The problem of course for a constitutional assembly like the one to be elected in Iraq was that it would also be the legislative body in which the executive would be based, and thus governability problems probably should have been considered by the makers of the rule. A single-country-district PR can produce extreme party fragmentation, and a mixed rule that then would perhaps enable the assembly to create both a more inclusive constitution-making panel and a more governmental legislative panel, as in India in 1948 (where they so proceeded on the basis of a single rule inherited from the colonial period), could have served all purposes better. But—and this was the second and more important reason for choosing the rule—a single-country PR did not require a prior census of the population, a problem raised against early elections.¹² Both first-past-the-post voting and multidistrict PR would have required a prior determination of the number of people in each district, with equal districts in the first case and seats proportional to population in the second case. But this problem was perhaps a red herring, since Iraqi experts had long assured both the IGC and Sistani that a reasonable estimate of the population could be made on the basis of ration cards. Since a first-past-the-post rule of the British or American type was not considered, there was thus a choice between multidistrict and a more proportional one-national-district version of PR, and the UN experts chose the latter, with the CPA then decreeing it as Iraq’s electoral rule (CPA Order 96, June 7, 2004).

Note that, unlike in the case of the TAL, no one ever made the absurd point that this purely CPA-enacted rule was abrogated because UN SC Res. 1546 or Prime Minister Allawi did not legislatively confirm it. I note, but without making too much of it, that there was little reason why especially Kurdish constituencies should have objected to this rule, nor was there a possibility, as in the case of the TAL, to have the rule and violate it too. Defenders of the Kurdish cause somehow forget to say that it too was abrogated with the demise of the CPA. It was, all the same, a disastrous rule and a disastrous choice. It was unfortunately foreseeable that a single-country-district PR would be turnout dependent in a way that a multidistrict PR would not be.¹³ Iraq was a country in insurrection, and insurrectionary violence was another reason people used to argue for delaying elections. Of course, according to this logic, such a delay could make the violence even worse, and then one might never be able to have an election. But what one needed was an electoral rule that would be as little affected by the insurrection as possible. The insurrection was geographically concentrated in some areas and not in others. Where it was raging, electoral turnout would be low. In a single-country-district PR system, parties strong in these very areas would be strongly hit by low turnout, and in the single national comparison they would get a lower percentage than their actual support. And this would happen whether or not the insurrection formally discouraged electoral participation.¹⁴ In a multidistrict PR, on the other hand, comparisons would be within each district, and presumably violence would affect the turnout of parties within a territorial district equally. There would be a smaller total turnout, but the seats would correspond to the percentages and the latter thus more or less to the actual support.¹⁵

To understand the Sunni boycott, one must start out with the elementary situation created by the electoral rule: under it, the Sunni part of the population would be dramatically underrepresented and, equally important to the actors, even undermeasured. The Sunni parties and associations only had the choice of being underrepresented or not being represented at all. There were thus also two fundamentally different types of reasons for the boycott: the “expressive” reason of the fundamentalists, who wished to denounce the political process no matter what, and the “instrumental” reasons of less radical forces, who hoped thereby to achieve greater representation. What was obviously spectacular about the call to boycott, its rhetoric and timing, admittedly seems to have had little that was instrumental about it. Linked to a fundamental situation that was not soon going to alter, namely the American occupation, and to events like the second, devastating set of attacks on Fallujah, the boycott seemingly could not be associated with any other set of political goals other than to rally the faithful.¹⁶ Clearly, the calls of the Iraqi Islamic Party to postpone the election pointed in an instrumental direction, especially when coupled with demands for the change of the electoral rule. Since postponement was not going to affect the security situation, only the two demands together made sense. When neither demand was met because of a convergence of American and Shi’ite views on the matter,¹⁷ the Iraqi Islamic Party almost relented. When it finally persisted in its boycott plans,¹⁸ joining the more radical Association of Muslim Scholars and the Sunni Endowments, this clearly did not mean the abandonment of the instrumental goal. Rather, the Sunni moderates calculated that more representation could be achieved through the boycott and through the threat to the legitimacy of the new constitution than through participation in and thereby legitimation of elections that produced very little representation.¹⁹ All the Sunni strategists had to anticipate was a strategic shift of American policymakers under the combined effect of the insurrection and the lobbying by influential Sunni states, especially Saudi Arabia, but also Jordan and the United Arab Emirates. If in the end the strategy did not work, in fact it did work in the intermediate term.²⁰

In the short term, the parties that won the elections gained new legitimacy, especially because the American authorities openly and materially supported the loser, Prime Minister Allawi’s secular Iraqi List. However, the victory of the UIA was not so overwhelming as to revive any plan (if one was still being considered) concerning the extralegal abrogation of the TAL. Very likely a deal concerning this and other matters was already well in the works with the Kurdish parties, even before the elections that were fought out between the UIA and the American-supported Iyad Allawi Iraqi List.²¹ Before the elections, Allawi was already too much of a centralizer for the Kurds and too secular for the Shi’ites, some of whom were already tempted by ideas of regionalism. The Kurds were ready to make concessions on the religious issues, as long as these would not affect Kurdistan. After the elections, the Kurds and Allawi together would have been in any case too small to form a government; indeed, no government could be formed without the Shi’ite UIA. Even the Shi’ites and Allawi did not quite have the two-thirds of the seats necessary to elect a Presidential Council, which was needed to form a government. They had also fought a rhetorically bitter campaign, and the heterogeneous UIA would have been more difficult to keep together if an alliance with the pro-American Allawi was attempted. We can also assume by this time that a section of the Shi’a was thinking in terms of a Kurdish alliance, depending on their interests, either because of regionalism or hoped-for religious concessions, possibly even both.

In any case, the Kurds and Shi’ites together were powerful enough that it was superfluous to add other partners to a coalition that would not have to depend on even some of the unreliable members of the UIA. Thus given the compatibility or compromisability of their interests and the results of the election, the Shi’ites and the Kurds were more or less destined to run the transition government. This also meant that, in line with the strongest Kurdish demands, they had to formally agree to fully abide by the transition rules of the TAL.²² That was probably a greater sacrifice to those closer to Sistani than for others interested in bargaining with the Kurds.²³ But despite their new electoral legitimacy and coalition agreement, they were not quite yet in the position to alone dictate the terms of the permanent constitution.

The Sunni boycott, along with the intimidation of potential voters by the insurgency, was devastatingly successful in the most immediate sense. In the end, only seventeen Sunni Arabs were elected to the National Assembly, and with the Iraqi Islamic Party gone, none of these represented viable political organizations outside parliament.²⁴ This dramatic under-representation, whoever was responsible, only highlighted the exclusion of representatives of what were still very important social strata of Iraq’s population, having more social, cultural, and professional weight than the 20 percent figure usually mentioned in this context. Most importantly, the military insurrection was almost entirely Sunni, and whatever propagandists tell us about their composition, the foundation for the insurrection was and is largely domestic. To the extent its small foreign part was the most nihilistic and destructive, its separation from the Iraqi Sunnis remains an important and viable objective. Thus Sunni exclusion from the political process, though never complete, has been and remains to this day one of the key problems that makes state and regime construction in Iraq next to impossible.

Ali Allawi is nevertheless right: despite early and sporadic recognition of this problem, the U.S. government only started to focus on it after the January elections. Whether the boycotters actually contributed to this shift or only anticipated it (or both), the arrival of Zalmay Khalilzad and even more his determined actions on behalf of Sunni inclusion signaled a dramatic shift in the perspectives of the constitution-making process.²⁵ Allawi speaks of the shift as inaugurating a third stage in American policy in Iraq, a little too neat a conception for something as incoherent and internally conflicted as this policy has always been.²⁶ What I would rather speak of here is a reluctant recognition of the elementary requirements of state rebuilding in a divided country with three very strong and armed political forces struggling for mastery.

State building or rebuilding in Iraq was not in principle impossible.²⁷ On some level, a state is only the knowledge, skill, and competence of a large number of actual or potential agents, and in Iraq this is all readily available. What no longer exists in Iraq is some kind of coherent set of public controls over coercive powers in the country as a whole, especially the means of violence. Bracketing the issue of limited sovereignty because of the occupation,²⁸ since a state with limited sovereignty is possible, whatever control was left to Iraqi governmental powers by the occupier could have been coherently organized only if there was either a fundamental, binding agreement over a state structure, one power forcibly bringing all other powers under its lasting control, or a binding agreement between some of the powers capable of bringing the others under their control. The second of these options presupposed the viability of long-term compulsion against all actors but one, and the third presupposed the viability of long-term compulsion against at least one such actor. In retrospect, the Sistaniled protest shows that pure compulsion was impossible with respect to the Shi’a segment of the population, and the insurrection showed, I believe, though the jury is still out, the same with regard to the Sunnis. Therefore, effective state building as the object of constitution making²⁹ presupposed, both at the time of the writing of the TAL and after the free elections, a comprehensive, inclusive agreement of all the major actors.³⁰ Before any further work on the structure of the regime, the top priority of negotiation would have to be the working out of a state structure acceptable (not preferable) to all sides—obviously, some kind of federal state or federation, but what kind? It is this type of agreement, which some rightly understand as a peace treaty or peace-building agreement, that had to be accomplished before the construction of a new regime.³¹

Another way of putting the matter was that the insurrection had to be brought into a process of constitution making that was partially transformed into a (peace) negotiation. This was a tough problem for Iraq’s occupier. The insurrection was mainly against the occupation (and only later focused more on its supposed beneficiaries), and the occupation continued, supposedly, because of the insurrection. Could this vicious circle be broken? The occupation was now also there to police a state structure negotiated in an imposed, exclusionary bargain, and the insurrection was also protesting that bargain. Sunni elites were against that bargain for the formal but very good reason that they had no part in making it and did not receive any tradeoffs, as did the Shi’a. Nor could they hope for an electoral victory under it, which the United States very reluctantly recognized in the case of the Shi’a. But they were also against it for the generally recognized, solid, and substantive reason that the arrangements were very dangerous for the Sunni provinces in particular. First of all, the ethnic cleansing needed to create the state or region the Kurds really wanted would be in part at the expense of Sunni Arabs, involving a huge refugee problem for the provinces further south. Even more seriously, the asymmetric confederal structure of the bargain already in the TAL established the possibility for a more symmetric one (art. 53c), with any three provinces having the right to form a region, possibly with powers like Kurdistan’s now or then. Assuming regional control over significant parts of the oil resources, which the Kurds have always demanded and at least some of the Shi’a would have an interest to concede if their provinces got the same privilege, the oil-poor Sunni provinces would be impoverished. They had a much better chance to fight such an arrangement early, when they still had the men, arms, and expertise. Aside from ideological and traditional Arab commitments to a more unified Iraq, which were probably very passionate for some, there were also solid material interests supporting such a position.

Again, the problem was foreseen in Washington very early on. However, the answer was not an IGC with memberships arranged according to ethnic quotas.³² Once the United States presided over an exclusionary deal that it refused to reopen, it was this deal that had to be enforced. But enforcing this deal was not possible without the indefinite continuation of the occupation. It gave two reasons for the insurrection to continue, and thus the seemingly vicious circle, the occupation and the arrangement it guarded, could not be broken. Since the occupation could not be ended immediately or even in terms of a realistic timetable without admitting defeat, which would have political and U.S. electoral consequences, the only thing the U.S. government could try to do is change somehow the deal that the occupation continued to preserve. After negotiating a state deal with the Kurds and giving the Shi’ites the election that produced a governmental majority for them (modified, of course, by the results of the state deal though consociational controls), the United States now had to arrange some kind of Sunni inclusion, real or illusory.

There were four great roadblocks in the way. The first was the TAL. With its amendment, ratification, and default rules, along with its concessions to a quasi-confederal structure, it was not clear what, if now included, Sunni representatives totally dedicated as they were to Iraqi unity could actually change in the structure of decentralization conceded to the Kurds. And even if they could live with a Kurdish “federacy,” if the Kurds kept what they had, the Shi’ites would ask for the same, and with their majority they could push it through, if they were united (admittedly a big “if”). In that case, the Sunnis would effectively be helping preside over the final dismemberment of the Iraqi state. Doing so would discredit them in the eyes of their constituency and the insurrectionary forces, and aside from endangering their lives (already in some jeopardy) the result would not lead to any diminution of the insurrectionary challenge.

Second, the process was not under American control, as the making of the TAL had been. Due to his style, personal gifts, and linguistic and regional knowledge, Khalilzad ought to have been a much superior negotiator than Bremer, but all this could not make up for the legal weakness of the American position and their increasing loss of political leverage with respect to their clients. And this was especially so after the Shi’ites and Kurds reached a modus vivendi that replaced the earlier “special relationship” between the United States and the Kurds. In particular, the new majority, the Shi’ite UIA, opposed anything more than a symbolic inclusion of unelected Sunni representatives or members.³³

Third, inclusion would only work if the parties (clearly former enemies) had a sufficiently long time period to bargain and negotiate with one another, try out different alternatives, and slowly develop a minimum amount of trust. The TAL, however, had a very specific timetable for the process as a whole. A draft would have to be “written” by the National Assembly by August 15, a ratificatory referendum would be held by October 15, and new parliamentary elections by December 15. Admittedly, the same TAL permitted a six-month extension of the process, if applied by for by August 1 (art. 61f), and even subsequently all the dates could be amended by article 3 of the TAL. However, U.S. representatives insisted, implausibly, that the insurrection could be dealt a serious blow by the constitutional process only if all the dates were kept. Thus an unusually short time period was provided for the making of the permanent constitution (seven months), some of this was eaten up by the problems of government formation and the formation of the Constitutional Committee itself (three and a half months in all), and it took another two months to include Sunni representatives. Despite this disastrous shortening of the actual timeframe, unforeseen by the TAL and its framers, there was continued strong pressure from the Americans and the largest Iraqi parties not to use the legal possibility of extension according to TAL article 61f.³⁴ Once again, the artificial timetable interfered with the workings of the constitution-making model, this time making the new, more inclusive strategy, which came too late in the game in any case, all the more difficult if not entirely impossible.

And finally, most importantly, whereas in the first stage of the process all the participants were co-opted by American fiat, now the leading parliamentary parties had full electoral legitimacy. Only the Sunni representatives could be portrayed as having been co-opted or imposed. Bremer’s adage that in a democracy you don’t shoot your way into power finally could be applied without self-contradiction, if the Shi’ites and Kurds wished; now, only the Sunni representatives were shooting. But those Sunni representatives weren’t shooting, they would be useless, since their whole purpose was to bring the less radical parts of the insurrection into the political process to produce a kind of peace agreement. Yet it could be said—and it was—that a group shooting its way into power may not be the one that really represents the aspirations of the broad Sunni masses, which may or may not have chosen these individuals to negotiate for them in the constitution-making process.

All these problems were pointed out relatively early, but the American side did not seem to be fully aware of them, or they wished to go ahead with the project of Sunni political inclusion regardless of the difficulties, some of which were caused by their own inflexibility. There were also some favorable factors that the Americans could take into account. The insurrection was now raging, and bringing at least some of it under control also had to be in the interest of the now-dominant Iraqi parties. Most of the Shi’a parties also wanted to end the occupation, and that too required that a role for the U.S. military be no longer needed. The prospect of forming a new Iraqi army that could and would effectively counter the old one, now partially underground, was remote. This too suggested the need to include some important Sunni forces in the political process and perhaps eventually in the new army. Finally, the Grand Ayatollah Sistani himself repeatedly called for new openness to Sunni participation,³⁵ and occasionally so did the Kurdish leaders, probably for quite different reasons.³⁶ Thus initially at least, Khalilzad’s efforts to try to engineer Sunni participation did not have to be seen as merely an American project. It remained, however, a technical problem as to how one would accomplish it, given the fact that Sunni representatives could not now be elected in a regular process.

Many schemes were offered to remedy the situation,³⁷ but the one eventually chosen, after intensive prenegotiations with a new umbrella group, the Sunni National Dialogue Council, was entirely satisfactory in my view and from the point of view of nonpartisan observers.³⁸ Fifteen new Sunni members were added to a fifty-five-member parliamentary Constitutional Committee, plus ten advisors, with the larger group renamed as the Constitutional Commission. Technically, the smaller body would still have to confirm the product of the larger one, but since a consensual structure of decision making had been decided on, this should have been automatic. In other words, members would be asked to confirm the second time around what they voted for in the first place. There were some disputes concerning who the Sunni representatives would be and whether de-Baathification rules would be applied, but finally the decision was left to the National Dialogue Council, which chose fifteen delegates and ten expert advisors from groups including the Association of Islamic Scholars, the Iraqi Islamic Party, the Sunni Endowment, and its own members.

I think it is safe to say that some of the fifteen new members (and their advisors), most likely a majority, sympathized with at least some wings of the insurrection, with some perhaps having political ties and channels of communication to armed groups. But that was an advantage, not a liability, if the whole process was to work. When they were picked, and subsequently, many charges were made concerning their rejectionist mentality and obstructionism, especially in the case of the neo-Baathist Saleh al-Mutlah, leader of the National Dialogue Council, and Adnan al-Dulaimi, leader of the Sunni Endowments.³⁹ The work of the Constitutional Commission, as long as it was allowed to function, does not bear out that there was such obstructionism, a charge made by highly biased parties, which of course does not automatically exclude its possible veracity. The charge of their nonrepresentive nature was much more serious, because it seems that the whole Sunni delegation was indeed picked by the leaders of these two groups plus the Iraqi Islamic Party, and they made no effort at consultations.⁴⁰ But in light of the short timetables, even this mode of selection was less surprising. As it turns out, insistence on these timetables made Shi’ite and Kurdish politicians reject the idea of drafting Sunni members through multiweek, regional caucuses.⁴¹ Perhaps they also did not wish to add legitimacy to the new representatives, but in any case this rejection made their later objections based on a lack of representative character less than fully authentic. Moreover, advocates of the Shi’ite and Kurdish parties forgot that a short period earlier, their leaders too had been co-opted into the Interim Governing Council by the very same Americans whose intervention now included the Sunni delegates, and that incumbency in the IGC meant a positional advantage in all subsequent attempts to be elected and to form a government. Finally, given the fact that the negotiations were to assume the function of peace negotiations in the eyes of some of the same critics, it was a serious question of whom the Sunni members should be representing. If the negotiations were to be effective, it was probably more important that they have channels of communication to armed groups than to the grassroots opinion of their ethnic groups. However, through the network of mosques, the latter was hardly absent. The Crisis Group goes so far as to stress the “tacit community support” for the fifteen and their generally representative nature in terms of their views and composition.⁴² Assuming that subsequent voting is a test for representativity (admittedly a big assumption), the referendum and elections of 2005 bear out the Crisis Group’s assessment.⁴³

With all this said, it remains true that these serious considerations could not change the fact that the type of representation involved by the Sunni presence was now out of sync with the democratic-electoral credentials of their Shi’a and Kurdish negotiating opponents. What was probably missing was a determined effort on the part of the electoral victors, perhaps Sistani himself, to lead the effort at national reconciliation and to lend their own legitimacy to the process of inclusion. In a strange way, what the Americans wanted could have best succeeded in the form of a national, nonpartisan, inclusive antioccupation effort. This is certainly so for the Sunni side. It is unclear how much credit Sunni representatives would have with their community if their main sponsors in an inclusion process were the Americans. For the Shi’ites, it was hard enough to accept Sunni representatives into the political process when violence against the Shi’ite community at the hands of insurgents was already taking place without (as yet) any retaliation. The fact that the Americans were imposing such participation made it even more difficult to go along with it, however rational it may have been even from the Shi’a point of view. Only a determined Shi’ite leadership that turned in a nationalist direction could have changed the picture, but unfortunately the zeitgeist tended toward religiously defined and sectarian identity politics. On that basis, no genuine overture to the Sunni would be forthcoming, unless we count Moqtadah al-Sadr’s occasional flirtation with the AMS, which did not stop his militia from engaging in anti-Sunni acts of retaliation. The project of inclusion remained an American one, and although completely justified, its legitimation problems were probably insoluble. What this project demonstrated, however, is that inclusion could have been possible early in the game, when all the parties were on an equal footing.

Aside from the goal of making a deal around everyone’s notion of a second-or third-best outcome, one crucial experience of negotiated transitions is that deals that can be made and have a chance to last are based on agreements among the opposite sides’ moderate forces, who can control or win the assent of their respective sides’ more radical elements.⁴⁴ Thus there are, abstractly speaking, two types of agreements required: one among the sides and one within each side. In Iraq there were three sides—four if we count the Americans—and thus four or five necessary agreements. And it is not difficult to identify more moderate and more radical forces on each side, even if the picture shifts depending on the issue and if not all of what we have experienced can be easily fitted into any such neat scheme.⁴⁵ In any serious deal, there were two types of agreements to worry about: the deal across the sides and the deal within a given side. No deal would be worth much if the dealmakers on any side were denounced by (all) their radical allies as traitors. At the same time, in a polarized situation the risk of some such denunciation must be accepted, even if the consequences could be deadly. It was important, in other words, for the moderate partners to give each other enough in the negotiations so that they and the deal survived the bargaining process.⁴⁶

This meant that the presence of Sunni players should have dramatically changed the bargaining process. That presence signified that on the procedural level, Iraq now almost had a genuine forum for negotiating a new state bargain. Though not quite. While a great number of details could be handled by the new Constitutional Commission, the really fundamental questions could not. This became a serious problem as the deadline for a possible six-month extension (as permitted by TAL art. 61f) approached. Strong pressure was put on the members to come to an overall agreement within that deadline,⁴⁷ but they could not, since those belonging to hierarchical leadership parties such as the Kurdish delegates did not have the authority to concede anything really important. First, there were some unfortunate attempts to divide the package into two and leave the most important and divisive questions until later, to be decided by majority votes. But this actually would have been equivalent to either (1) not getting the job done and replacing the TAL with an inferior product or (2) to delivering the crucial questions to a future parliamentary procedure with fewer restrictions than the current constitution-making one. It is hard to see, for example, why even the Kurds would accept majoritarian insecurity over the security provided by the TAL. Certainly, the Sunni representatives could not accept any such “compromise.”

Almost all round-table settings presuppose that, aside from the more formalized meetings, there is a possibility for a meeting of the political principals of the really important groups, who would be capable, if anyone is, of making fundamental decisions on the spot, in one or several sessions. With the deadlock of the Constitutional Commission on the really fundamental issues, such as the structure of the state and the place of Islam in it, just such a meeting of principals was called for the weekend of August 6, 2005. This move was important for two reasons. First, the negotiation of the final constitution would take place in three important venues, at least in principle: the inclusive, formal Constitutional Commission; the informal meeting of principals, which was called the “Leadership Council” or the “Kitchen,” presumably equally inclusive and consensual in terms of participation and decision making as the Constitutional Commission;⁴⁸ and finally, the (constitutional) National Assembly itself. This was the right structure, but there was an important proviso and limitation. Because of the artificially imposed time limit and the refusal to extend it, there would not be enough time (less than a month) for the first venue, in effect the “round table,” where expertise and international advice could play its greatest role, to adequately look for and arrive at compromise solutions.⁴⁹ Given the same time limits, it was even more difficult to see what substantive role the third venue, the National Assembly, could play in fashioning the constitutional product. The danger of its becoming a mere rubber stamp was extremely real—not a good precedent for future representative government! Moreover, the National Assembly and its imagined discussions and hearings would be the obvious forum for, on the one side, making the process public and visible to the population and, on the other, permeable to a variety of democratic inputs. Such a public process was foreseen by advocates of some of the political forces, enshrined in the TAL (art. 60),⁵⁰ and promoted by UN representatives, but the time constraints now rendered almost impossible the chances of a public, participatory process focusing on the National Assembly.⁵¹

Second, the kind of constitution-making venues Iraq now had meant that for a fleeting moment the country had the inclusive negotiating format it should have had two and a half or one and a half years before, when the United States and the United Nations respectively could have pushed through a round-table negotiation, involving all the major political forces of Iraqi society, to negotiate an interim constitution, including a state deal, that was instead bargained by the Americans with the Kurds exclusively. Thus it appeared that all the elements characteristic of recent two-stage constitution making were now in place, though certainly not in their proper sequence. The cart was before the horse. This meant, first of all, that Iraqis were now supposed to negotiate, very late in the game and under extreme time pressure, a state structure and, at the same time, the governmental institutions appropriate for that structure—as well as a symbolic national identity—in a final, no longer merely interim package. And second, the change in sequence meant that at best this deal would be a deal among elites, and it would come very late in the overall process, with little opportunity for genuine parliamentary or public discussion. Once a draft emerged, there would be few opportunities and forums left in which to correct it, and it would be hard to treat a “final” constitution as another “interim” one, though not impossible, as the actual history would show. For example, the rejection of the draft in a constitutional referendum could supply yet another opportunity for correction. But such a rejection, with popular choice and input reduced to a simple yes or no, would not be based on experience with malfunctioning, and the corrective would be more a function of a new electoral arithmetic (about which more below) than of constitutional learning. If it happened, the consequences of rejection in a referendum were politically unpredictable and potentially explosive. Finally, sequencing also mattered, because now there was a freely elected parliamentary body in place. The temporal conjunction of a round-table, elite leadership format with a constitutional assembly allowed the parliamentary majority the freedom to manipulate, if it wished, these venues according its perceived interests. The majority was not compelled to make a deal because it reserved the right and had the contemporary opportunity to pass its own option. As opposed to the initially dominant forces of round tables elsewhere, which usually occur much earlier in the constitution-making process, the leaders of the Iraqi governmental parties had electoral legitimacy.

Nevertheless, though out of sequence, an inclusive round-table format was established, and with it a logic toward a fair political compromise first asserted itself. What such a fair deal would have been like is not difficult to reconstruct. As a second, substantive implication of serious Sunni participation in the negotiations, the outcome, this time, in order to really work, had to favor their bargaining position to whatever extent still possible. The point of inclusive negotiations within a proper timeframe was emphatically not to bring the Sunnis to understand better that they simply had to accept a state bargain first negotiated by the Americans and the Kurds, the structure of which was now to be extended to the Shi’ite governorates as well. No time would have been enough for that.⁵² Since the previous state bargain and the subsequent political tradeoffs under Sistani’s pressure favored first the Kurds and then the Shi’ites, one had to find some areas where the Sunnis could make gains, even if many of the results of the earlier arrangements were now no longer reversible. Clearly, given the TAL the Shi’ite majority of the National Assembly would not pass any arrangement that would take away the establishment of Islam as the state religion and the majoritarian, parliamentary structure of the central government. Similarly, the three Kurdish provinces would not ratify a significant diminution of the special rights of the Kurdistan Regional Government. Yet, important compensations had to be found for the Sunni delegates for them to be able to play their proper role, which consisted not only in coming to agreement with Shi’ite and Kurd moderates but also in convincing significant sectors of the radical Sunni insurrection that the deal was a good one or at least the very best one that could be achieved under the circumstances. It may be that Sunni elites still hoped for a fully centralized state, and some insurrection leaders may have even imagined that a new dictatorship could be erected on such foundations. This is the position continually ascribed to them as a group, in a rather self-serving manner, by some supporters of the Kurds. Such a state was now excluded as a possibility, both because of the special status of Kurdistan and because the Kurds would not accept the rest of Iraq being so organized and waiting to bring them again under Arab control. It is therefore much more worthwhile to pay attention to the Sunni bottom line, which was that they could not accept being an impoverished region in the center of Iraq, which was what the various breakup and confederal plans had in store for them. Thus, it was up to the other side to offer them arrangements that would involve guarantees against this worst-case outcome. The guarantees would have to come on three levels: the organization of the state, the organization of the government, and the disposition over natural resources (that is, oil). Substantively, it was important that the Sunnis receive with respect to all three areas a perceivably better deal, and certainly not a worse one, than they did in the TAL, in whose making they did not participate at all. Otherwise, they would have been co-opted into the Constitutional Commission and would be, both in their own eyes and in that of their dangerous constituency, mere window dressing.

There is no need to generate my own idea of a fair constitutional settlement, because, before the consensual process broke down, there were some clues in the press and in the documents of the expanded Constitutional Commission that its members were possibly developing at least important elements of such solution. Surprisingly, many of the crucial steps were taken even before the new Sunni delegates joined the Constitutional Committee, or rather the Constitutional Commission, on July 13.⁵³ This could have been because of the role of experts, international advisors, or because Sunni participation was already seriously discussed before their arrival, or because no Shi’ite group had yet made an irrevocable decision for a generalization of ethnically based federalism. Whatever the reason, in retrospect it seems clear that the solutions as they stood (and were interpreted) in early July could have, if there was sufficient time, gained Sunni acceptance if perceptibly clarified and modified according to their proposals. As to the structure of the state, it was clear that the Sunni delegates now accepted, however reluctantly, that the Kurds were not going to lose their special status, and this meant having a fully autonomous region, with a regional government, constitution, and a regional militia, all in a bilingual Iraq where they would play a strong (but not consociational or power-sharing) role in national government. But it was not likely that they were going to get to expand their region, with the possible exception of Kirkuk, and gain the exclusive right to dispose over the natural resources in their territory. Most importantly, they were not going to get to extend their quasi-confederal regional formula to the rest of Iraq under the misleading name of “federalism.” It is true that the early draft of the permanent constitution available to me⁵⁴ contains regions as well as provinces and allows region formations in addition to the region of Kurdistan, which in this draft is not explicitly mentioned. But there is no sign in that document of a formulation that would restrict the federal government to a few enumerated powers and of nullification rights of the regions regarding most federal laws. On the contrary, it is the regional constitutions that must conform to the federal constitution (chap. 4, art. 18). While the mechanism of region formation seems undecided, the constitutions of the regions would be produced by the National Assembly (chap. 4, art. 7). According to press reports, the constitution was certainly going to establish a second legislative chamber based on the geographic principle of provinces, and there is some trace of this in the draft, which does not, however, provide a scheme for such a body.⁵⁵ The draft has a single-person presidency⁵⁶ rather than a three-person council and thus has no place to involve decentralized units in the management of the federal government other than in a second parliamentary chamber. Thus the political role of the Kurds in the federal government of Iraq would have corresponded to their numerical weight in two chambers and not according to consociational, power-sharing arrangements within the executive. In this sense, the journey from the TAL to the permanent constitution would have been, had things gone right, from consociationalism to constitutionalism, as it was in South Africa.

The direction of change regarding the structure of government parallels this, though for reasons of ideology rather than demography. The changes were in the direction of a more parliamentary government based on the rule of the majority, and that favors the Shi’ites and not the Sunnis. The latter, however, were ideologically committed to all measures that tend to strengthen central government. Paradoxically, in order to deny the confederalist Kurds consociational rights of participation in the executive, the Sunnis must accept the consequences of that same denial with respect to their own participation. Under the TAL (and in practice after the January 2005 elections), it took the consent of the Kurds (and potentially the Sunnis), again assuming a Kurd (and a Sunni) in the three-person presidency, to name a prime minister, making the formation of government if not its later composition dependent on their will. The first constitutional draft no longer contained any trace of this particular cumbersome power-sharing arrangement, which could have made government formation impossible at some point even with a parliamentary majority, an unacceptable state of affairs for a parliamentary government. Not only does the draft have only one president, a more ceremonial one elected by a two-thirds majority of parliament, he must also first offer the leader of the largest party the powerful position of prime minister. The overall relationship between government and state substitutes a federal state with a confederal enclave (Kurdistan), whose center-unit relations are mediated by a geographically based second parliamentary chamber, for the TAL’s uneasy mixture of a centralized government and a confederal state mediated by badly designed consociational elements. Everything would depend on the composition of the upper chamber and its powers and its decision rules. But assuming either a purely provincial upper chamber (with three or four Kurdish provinces) or one based on a combination of regions and provinces, it would have been possible to give sufficient guarantees to the Kurds (and the Sunnis) against any tyranny of the majority at least on the level of lawmaking. There was, to say the least, a potential here for a better federalist formula concentrating more flexibly on the ongoing political decision making than ones focusing, with great rigidity, on the very beginning of the governmental process.

As argued in the previous chapter, amendment rules tend to indicate the nature of the state, and therefore in the case of the TAL indicate an ultimately confederal plus consociational structure from the point of view of the Kurds. In the early drafts, this was going to change. Now minor amendments would take two-thirds of the vote of one national assembly, and major amendments (Netherlands style) would take two-thirds of two assemblies, with an election in between. In both cases, a national referendum would have to approve an amendment by simple majority, but there is no provincial veto of any kind. While there was a need for the president to approve amendments, this again was in the place of the unanimous (that is, consociational) approval of the presidential council.⁵⁷ Nothing was stated to be unamendable this time around, including the rights of regions. In short, large minorities, regional or ethnic, unless they controlled over one-third of the parliamentary seats, would lose their control over constitutional change. This would be the amendment rule of a federal rather than a confederal state. However, from the point of view of majority and minority relations, note that one kind of equality has replaced another. Under the confederal or consociational rule, one group out of three, whether minority or majority, could veto any amendment. Now, it would take either two groups to pass or two to veto an amendment. From the point of view of the Sunni Arabs, they would need an ally either to amend or to block an amendment—but the same would be true for all other groups. But a major amendment would take the alliance of two national groups to accomplish, with the electorate getting to vote on it as well.

Finally, on an issue especially important to Sunni Arabs given the experience of the January 2005 elections, it was almost certain that the electoral rule was going to change in the direction of provincial lists. While not strictly speaking the competence of the Constitutional Committee or Commission (no subcommittee was assigned with this task), clearly it was in this forum that Sunni representatives could strongly advocate their preference for change in this crucial area. While to some extent some Shi’ite representatives held out for the old system, perhaps as a bargaining chip, no deal was possible with the Sunnis without basic reform of the rule.⁵⁸

Such was at least one possible package that was emerging or could have emerged from the Constitutional Commission deliberations. There were alternative drafts, and there no consensus around any of them was achieved in the short time period. All the drafts, moreover, had areas left open for subsequent codification. Many critics of the process overlook that on the most important questions the decisions were political, and since the leaders of the political parties were not part of the commission, the fundamentally disputed issues ultimately had to be brought to them. Because of the pressures of time, undoubtedly, the switch of venues occurred too soon, though it had to occur at some time or another. Every negotiated constitutional process has its “Magdalenka,” “bosperad,” or other informal venue as the basis of the ultimate political compromise. The only problem in Iraq was that the “Kitchen” came both too late (it should have come in the first stage, for the interim constitution) and too early (the Constitutional Commission needed more time) in the process, and it dramatically excluded some of the most relevant members. If Iraq was still, however regrettably, making a peace treaty (which should have come in the first stage), amazingly enough they were now going to make that treaty without the participation of the side they were fighting!

We will never know for certain whether the Leadership Council or Kitchen came to exclude the Sunni side only because of the extreme time pressure they were working under or because the purpose of the meeting of the principals from the outset was to seal a Shi’ite-Kurd exclusionary deal that the Constitutional Commission or Committee were afraid to complete on their own. That the latter option seems more likely is the opinion of the Crisis Group, who say that the goal was to both speed up the process and to confirm that whatever the commission structure and its procedural rules, “the real power to take durable decisions lay with the heads of these two communities.”⁵⁹ Similarly, Jonathan Morrow argues that “scrapping the Committee [ formally the Commission] on August 8 meant that the Sunni Arab Committee members, after no more than one month of trying to develop and assert a coherent constitutional position, were retired en masse.”⁶⁰ I initially gave the Shi’ite and Kurdish leaders more of the benefit of the doubt, perhaps wrongly. What seems shameful is that all this happened in the presence of the American ambassador, Zalmay Khalilzad, who up until this point was presumably working on a three-sided fair deal some of his own principals in Washington considered essential.⁶¹

What seems to have happened in rough outline is more or less this: Sheik H. Hamoudi, the Chair of the Constitutional Commission (and Committee) made a determined attempt (supported by Mahmoud Othman and other senior members) to get support for an extension of the constitution-making process, and thus for the work of the commission, but he failed in the face of American opposition and the resistance of his own party (UIA) leadership.⁶² The TAL’s August 1 deadline for getting a long extension thus came and went; the August 15 deadline for submission remained. Now Hamoudi was forced to call (August 6)⁶³ for a Leadership Council to take over the negotiations, and for this M. Barzani joined the other leaders already in Baghdad. Negotiations moved to the Kitchen on August 8. Three days later, on August 11, one of the key participants, Abdul Aziz al-Hakim, the leader of SCIRI, in front of a large demonstration in Najaf announced the fundamental demand for a southern Shi’ite region of nine provinces. When the August 15 deadline for submitting a draft to the assembly was not met, Sheikh Hamoudi asked the National Assembly to grant a week’s extension by constitutional amendment; this was granted. By this time, the commission was simply a drafting organ for the Kitchen, from which the Sunni representatives were more or less completely excluded. There were two illegal extensions on August 22 and on August 25; on August 25, the Sunnis were called in only to be told of the results, which they had to take or leave. They then formally suspended their participation in the commission and were left to appeal to the United States, the United Nations, and the Arab League, of course in vain. No Sunni group or even major Sunni politician previously in the IGC or the Transitional Government—not Adnan Pachachi, Vice President Yawer, and certainly not the leaders of the Iraqi Islamic Party (whom Galbraith calls proconstitutional)—supported the new draft. More surprisingly, the leaders of the majority of the National Assembly, the Shi’ite and Kurdish parties, who negotiated bilaterally and produced a draft without Sunni agreement, also did not feel confident enough to have that assembly (which they controlled) actually vote on their draft constitution, so they precipitously approved it by executive fiat. Acting entirely extralegally, they compromised both the consensual decision rule previously agreed upon with the Sunnis and the prescription of the TAL (arts. 60, 61a) on which the process up to that point depended.⁶⁴ Thus they violated the TAL through executive fiat rather than through the vote of the freely elected assembly, though of course the executive, having a majority in that assembly, could count on not being challenged by the legislative majority. Technically, they carried out a scarcely disguised coup against the TAL and the National Assembly.⁶⁵ All this was done, as we will see, in the name of a mediocre document full of holes and inferior to the TAL itself, leaving some of the most fundamental constitutional questions for later majorities or qualified majorities to decide.

No one disagrees that Sunni exclusion took place, though remarkably some in the Kurdish camp still refer to the constitution-making process as negotiating a “tripartite peace treaty.” Some of the excluders argue that the fault alone lay with the excluded. “The Sunni Arabs objected to practically everything that was proposed to them, frustrating the Shi’ites and the Kurds to the point that they stopped negotiating with them.”⁶⁶ According to this train of thought, the sole objective of the Sunni representatives was to run out the clock, given the misguided American insistence on rigid deadlines. If no draft could be submitted by August 15, then according to the TAL (art. 61g) new elections would have to be held, and the Sunni parties that were able to force it could redeem their earlier mistake in calling for a boycott and achieve greater representation in the National Assembly.⁶⁷ To counter this strategy, the Sunni representatives had to be excluded, and short extensions of the drafting period had to be sought after all. This argument fails if one were really serious about achieving a peace treaty, not to speak of a working constitution that presupposed a working state, which itself presupposed pacification. One does not make peace with one’s friends, only with one’s adversaries. It was already understood that such a process would take time. A constitutional amendment on August 15 could have been secured for the six-month period originally foreseen by the TAL, giving genuine negotiations a chance, if that was what was desired. Moreover, the only reason the Sunnis had to seek greater representation in a constituent assembly by a roundabout way, if they really did, was because they were not taken seriously in a process that was supposed to be consensual.

In any case, the whole set of self-serving explanations seems spurious in light of one fundamental fact: Hakim’s bombshell announcement. Nothing in the process was more spectacular and decisive than this August 11 demand by the leader of SCIRI for a region of nine southern Shi’ite provinces, when the Kitchen, presumably still containing members from each group, had just begun to meet to iron out the remaining issues left over from the Constitutional Commission.⁶⁸ If implemented, the SCIRI proposal would lead to the creation of a powerful region containing all the ports, 70 to 80 percent of the current oil, and half the population of Iraq, and where Iran would have a decisive influence. If Iraq stayed together in such a “federation” of three very unequal regions, the southern region would dominate it both because of its size and resources and because the Shi’ites would also control, through their majority, the central government. It would be the Prussia of the new Iraq. And if the formation of the region demanded by Hakim led to Iraq’s breakup, a high likelihood since the Kurds were not about to accept Shi’ite dominance on a new basis,⁶⁹ Iran, which always argued for Iraq’s unity, could easily deflect responsibility to the Americans, who indeed had destroyed the Iraqi state, among other things, by encouraging Kurdish separatism in the first place. From the Sunni point of view, a three-region “federalism,” or “confederacy” if the Kurdistan model was adopted, would mean an impoverished, resource-poor Sunni region without any influence in the government of Iraq itself. Self-government and autonomy would appear rather insignificant tradeoffs given the probable losses of revenues from the two oil-rich regions.

Little remembered now are two facts connected to this announcement, facts that demonstrate that Hakim’s demand was by no means consensual within the Shi’ite community itself. One was the immediate opposition to the idea by Prime Minister Jaffari. The other much more important fact was a series of “round-robin” meetings just before Hakim’s speech between the Grand Ayatollah Sistani, Moqtadah al-Sadr, and Hakim. Given Sadr’s attitude before and after, and Sistani’s before and Jaffari’s after, it is not impossible to deduce that the two were trying to convince the leader of SCIRI not to throw his idea of a nine-province megaregion into the constitutional negotiations or at least to stay with the less destructive three-province version of the TAL. Whatever the case, and some say that Sistani by not speaking revealed his tacit acceptance,⁷⁰ Hakim went ahead, with significant popular support. The consequences for the Sunni negotiators had to be devastating. It is generally conceded now that they had come to accept, however reluctantly, the idea of a Kurdish federacy, as it was then constituted, with the possible addition of Kirkuk as the next-to-last straw, perhaps. A Shi’ite region similarly constituted, a superregion in size and power, was beyond the limits of the possible and even imaginable for them. The Hakim demand, advanced in the most radical manner possible, brought home reality in the hardest possible way.

It may be the case, though I doubt it, that Hakim was not making a new demand constitutionally speaking, because the draft as it stood at that particular moment (it is not available to me) already contained, on Kurdish insistence, the possibility of multiprovince regions.⁷¹ That fact, if it were true, would change little. It is generally agreed that the demand “shook up the negotiations.”⁷² Second, even if the abstract option had been in a draft, now the possible meaning and effects of that text were made crystal clear. Most importantly, the text was still the object of negotiation. So far among the Shi’a, political centralism was the prevailing emphasis; after all, they were expected to dominate the central government, which in a democratic Iraq could not form without them. Even in the Kitchen, Prime Minister Jaffari should have counterbalanced Hakim, and the Kurds were expected to be neutral if the Shi’ites themselves were divided on the issue of how the rest of Iraq was to be organized. So in principle even (asymmetrical) territorial federalism and a fairly strong central state were not out of reach for Sunni negotiators, if the new concessions to the Kurds were limited to Kirkuk. With Hakim having gone to the street and having for the moment neutralized Sadr and Sistani, the worst possible option was suddenly very much more likely. If a strategy of seeking new elections emerged, this could have happened because of the new and radical Shi’ite demand. In any case, if in light of this new situation a Shi’ite-Kurdish constitutional bargain was to be consolidated, this required the exclusion of the Sunni from real negotiations. To the extent that their resistance delayed the process, short extensions had to be admitted after all. If these could be used to detach one Sunni actor from the inevitable front of rejection, so much the better. But even such a result was not mandatory from the point of view of the Shi’ite and Kurdish leaders dominating the process. Hence the paucity of concessions they offered to the Sunnis, until they began to worry about the referendum.

The efforts to detach at least one Sunni party would have occurred at the insistence of U.S. Ambassador Khalilzad. Sunni inclusion was his mission, responsibility, and probably personal project as well. Once again, as with the Kurds previously, we have to ask how the representatives of the earth’s one and only superpower could have been so weak. In defense of Khalilzad, one must say that he was presented with contradictory tasks and had to play contradictory roles. He was to engineer Sunni inclusion and had to insist on the rigid deadlines of the TAL, which were now American benchmarks for reasons having nothing to do with Iraq. He had to be present to accomplish anything, but his presence was a huge embarrassment to the Iraqis, who this time wanted to avoid even the appearance of constitutional imposition. What his ubiquitous presence achieved in the end was the worst of both worlds from the point of view of the United States: the appearance of imposition without the reality. The only threat he really had was U.S. withdrawal, but the U.S. government apparently wanted to stay in Iraq. He had no legal authority to order anyone to do anything. Ultimately he could only persuade, but the United States was deemed too self-interested for the words of its ambassador to be taken at face value. Moreover, his success would depend on Iraqis meeting expected “benchmarks” in a timely manner, and this absurd idea made Khalilzad prisoner of the only forces that could produce that result, even if only on an entirely formal level: the Shi’ite and Kurdish parties that dominated all of the venues of constitution making. Only where they disagreed, as on religious issues, did Khalilzad have a chance, and apparently he blew even that chance at constructive intervention. When he chose to go up against the coalition openly, as in the case of amending the TAL on August 15, he suffered an embarrassing defeat,⁷³ all the more unnecessary because the extra week thus granted to the drafters was in line with his attempt to get at least one Sunni party to agree to a revised constitutional draft.

Let me emphasize that Hakim’s demand and the ability of the Kurdish and Shi’ite negotiators to wreck consensual negotiations was a function of the continued occupation, which unlike the constitution-making process was not tied to any schedules or deadlines, arbitrary or otherwise. Evidently, embittering the Sunnis led to the continuation of the insurrection and the possibility of open civil war. As long as the Americans remain to deal with the military consequences, the internal reasons for the Shi’ites seeking a modus vivendi are greatly vitiated. Had the Americans made any threats concerning their stay in Iraq or established any timetables for leaving, the results might have been different. Of course, wishful thinking may play a role here as well, because the Shi’ites cannot really predict what the Americans will do in the face of an insurrection exacerbated by the constitutional disaster they are now causing. At the same time, the Sunnis, who may also be guilty of wishful thinking, may not think that the Americans can stay forever in the face of continued losses, and therefore they may figure they have no reason to accept second-class status within the current quadrangular configuration of forces—a configuration predicated on a U.S. presence. The situation leads to imposition on the one side and bargaining by means of the insurrection on the other side, which weakens moderate and constructive forces on all sides.

Focusing on Khalilzad and the Americans may make it appear that only they constituted an external factor that mattered. Such could not have been the case. Few have noticed how strongly Iran supported the new constitutional draft.⁷⁴ In fact, Iran’s attitude may have counted for more in the outcome than did the vector sum of the Shi’ite attitudes. To the extent that the United States so publicly identified itself with a consensual solution for the constitution-making process, it was much too easy to bring that process down through the acts of a proxy introducing a new demand that made consensus impossible. The temptation to greatly embarrass the Americans was there, and it is hard to believe that Teheran did not take advantage of the opportunity, especially because a policy was available that presented them with a very favorable opportunity of extending their influence in Iraq without having to take any responsibility for the chaotic consequences of pursuing that goal. Undoubtedly, from Iran’s point of view, Hakim’s new demand would not only wreak havoc with American plans for a consensual solution of constitution making but had its own independent rationale of helping to sustain a political crisis where the Sunni insurrection would tie down the Americans for years and force them to accept the consolidation of Shi’ite power.⁷⁵

In any case, once Hakim’s demand for the creation of a Shi’ite superregion was introduced, all attempts to save the process were doomed, including President Bush’s pointless and embarrassing phone call to the SCIRI leader on August 25. First, resisting Kurdish demands to weaken the central state and extend their regional structure to the whole of Iraq had to depend on Shi’ite resistance and their defense of the unity of Iraq, as the main beneficiaries of that unity. When Shi’ite leaders themselves championed regionalism in a more radical version, the Sunni were left alone.⁷⁶ It is not the case that either Kurdish or Shi’ite demands were in themselves irresistible. Galbraith argues that M. Barzani (whose key advisor he was) dominated the process of making the final constitution. And it is true: the Kurds kept everything the TAL gave them, made new gains with respect to a further weakening of the jurisdiction of the federal government and the ultimate disposition of Kirkuk as well, and even managed to gain a kind of mediating position with respect to some issues such as the question of Islam and the state. Most of this was prefigured by the TAL, as the default position guarded by its amendment and ratification rules. The Kurds could not lose anything in the process, and if one wanted their consent on anything new, concessions had to be offered to them. Nevertheless, a more united Arab position against them could have held them to a position more or less in the TAL, one that the Sunnis could now live with. If the Kurds played their cards right, it was because they helped to delegitimate the role of the Sunni delegates and kept their coalition with the Shi’ites intact. This way, a new constitutional bargain between Kurds and Shi’ites could be cemented entirely at the expense of the Sunnis.⁷⁷ This alliance in turn made Hakim’s radical proposal also impossible to resist. The proposal perhaps surprised the Kurds, though they were not in principle opposed to it.⁷⁸ With respect to the issue of “federalism” there now was a Shi’ite-Kurdish bargain, which, similar to the American-Kurdish bargain earlier, was difficult to change, but this time the bargain was one no Sunni party could possibly accept.

That the Shi’ite demands were not entirely irresistible is shown by the fate of another contentious issue. The dominant clerics were also demanding a much stronger constitutional statement on behalf of the role of Islam in civil and family law than the combination of provisions in the TAL provided for. The effort of getting them to compromise was somewhat more successful in this area, where the Americans, Kurds, and secular deputies could concentrate their pressure, however inconsistently. A similar coalition was not available in the area of “federalism.” Here the timing of Hakim’s bombshell, so late in the game and supported by a Kurdish-Shi’ite deal, meant that there was no time to work on a complex compromise formula that could allow all sides to provide input and save face. Here, the American acceleration of the process reaped its bitter fruit. They insisted on the artificial deadlines, but others learned to use them better than they.

In the end of the official or legal process, between August 15 and August 22, only a small unilateral concession was granted to the Sunni in the domain of “federalism,” and it was not enough to gain the support of even the Iraqi Islamic Party. The draft (art. 114) as it stood on August 15 allowed the creation of new regions of any size both from provinces and old regions simply through the request of one-third of the provincial legislature or one-tenth of the voters and the approval of the majority in a provincial referendum. In case of failure, two-thirds of the provincial legislature or one-fourth of the voters would have to request a repeat of the referendum—how soon it was not said. On August 22, the implicit reference to size was gone, but the constitutional right of any province to form regions of undetermined size remained. While the request for forming a region was the same as before, what approval entailed was no longer clearly specified. A new article 115 (in the current version, art. 117)⁷⁹ stated that the “Council of Representatives shall pass a law that fixes the executive procedures relating to establishing regions by simple majority in a period that does not exceed six months from the date of the first session.” Since this law would be passed by simple majority, the mechanism of region formation it clearly had in mind would most likely also operate by simple majority or by majorities in referenda or assemblies. It would be possible, of course, that the majorities involved would indeed operate on any or even a combination of three levels: the provincial electorate, the provincial council, and the National Assembly. At the same time, since forming a region was still defined as a right of provinces (art. 116; current version: art. 118), a future Supreme Court may declare any law interfering with that right unconstitutional. In any case, the crucial point was that everything that might have been objectionable in the August 15 version to the Sunni leaders could be reestablished by simple majority. There might be one rub: as a law, the new legislation dealing with establishing regions might be open to vetoes—in the transitional period the Kurds were able to insist on the veto of any of the three members of the presidential council (art. 135; current text: art. 137, 5th clause). But the same article explicitly exempts laws concerning the establishment of regions from the possibility of a veto.

Thus, the unilateral concessions that went into the August 22 draft were entirely unacceptable to the Sunnis, who assumed that very likely the Kurds and the religious Shi’a together and possibly the latter alone would have at least a majority in the next National Assembly, and even if they, the militant Sunnis, controlled one member of the presidential council, the relevant veto would not apply to the law on region formation (which is how it turned out with respect to Vice President al-Hashimi in 2006). The bottom line of the Sunni delegates was that it would have to be two-thirds, that is, the constitution-amending majority, that would have to work out the rules governing federalism. Since it was possible that a Shi’ite-Kurdish coalition would get two-thirds of the seats in the next parliament, even this solution carried an element of risk, admittedly for both sides, in the debate. But this solution was decisively rejected.

At issue was both the question of region formation and the kind of powers regions would have. Any cursory look at the TAL and the draft of the permanent constitution approved on August 28 (or in the current version) will reveal that it is simply not true that “the list of exclusive federal powers is much shorter” in the latter.⁸⁰ In fact, the list was now almost the same length, if differently organized and numbered, with the addition of control over external water supplies entering the country and the right to take a census of the population and the subtraction of control over natural resources and telecommunications (TAL art. 25 versus Constitution art. 109; current version: art. 113). It is also disputed whether or not the power of taxation is accorded to the federal government; according to some experts, control over fiscal policy in the Arabic formulation contains that power.⁸¹ It is outside this particular regulation that new gains for the Kurds must be sought. First, the right of organizing self-defense and internal security is now explicitly accorded to the region (art. 118, 5th clause; current version: art. 120, 5th clause). This undermines the monopoly in defense and security provided to the federal government by articles 108 and 109. Similarly, the right to have offices for regions in embassies abroad (118, 4th clause) undermines the monopoly in foreign affairs. The right of nullifying federal laws except in the case of exclusive powers remains from the TAL (art. 118, 2nd clause), but now a new provision is added that makes regional law trump federal law in the case of a conflict (art. 112) of whether it is a question of either unenumerated powers of the federal government or concurrent powers. Thus while the Federal Supreme Court can still adjudicate conflicts between the center and the regions (art. 91, 5th clause), it can lawfully side with the center only in the case of its few exclusive, enumerated powers. Nathan Brown is right that with these provisions Iraq can “lurch in a confederal direction, especially if a Federal Supreme Court emerges as a powerful body even mildly friendly to the regions.”⁸² What I do not see is how such a lurch could be stopped by a court otherwise disposed, because it is not given what the Germans call Kompetenz-Kompetenz, the competence to decide questions of competence regarding these questions, in the case of conflicts.

The big issue, of course, was that the management of natural resources—the oil—no longer appears under the exclusive powers of the federal government as in the TAL (art. 25e) but separately, in a very confusing and unclear formulation (arts. 109–110; current version: arts. 110–111). Evidently there was no real compromise here either. The Kurds and Shi’a were to gain important rights when compared to the TAL. The issue was, of course, that even regional federalism could have been made more acceptable if the two oil-rich regions had made clear and precise provisions to fully share their wealth with the third. Given the lack of trust in the country, the desideratum could be guaranteed only by central government control, which the TAL still provided among the exclusive powers of the federal government (art. 25e). It is true that the new constitution (art. 109; current version: art. 110) stated that “oil and gas are the property of all the Iraqi people,” but it adds “in all the regions and the provinces.” The actual dispensation comes in the next article. The federal government was to administer “current fields,” but in cooperation with the regions, on condition that the distribution be fair, and with a quota defined to make up for past wrongs (that is, for the benefit of Shi’ite and Kurdish provinces). Thus even current fields would not be fully under central control, and even that only under legally contestable conditions. There were, however, future fields to be explored, in terms of the brand-new concept of “current,” and about these potentially much more important resources we get no clear regulation. In line with the premise that powers not defined as exclusive or shared powers of the federal government belong to the regions or to provinces (art. 112; current version: art. 113), one possible conclusion to draw is that future fields would be under regional administration. Galbraith certainly draws this conclusion but hopes to avoid it by pretending that all regions are likely to have future oil resources.⁸³ Officials of oil companies, who may know better, expect nothing but contested jurisdictions, chaos, and disorganization deeply undermining the financial capacities of the Iraqi state for a long time to come.⁸⁴

As opposed to all matters having to do with the issue of federalism, where the Shi’ite and Kurdish forces apparently made a deal to support each other no matter what, compromise was possible where the two forces did not fully agree and where opposition to the majority view was not represented by the Sunnis. Thus the solution of the issue of the role of the Shari’a in the constitutional setup had apparently a far different structure, demonstrating what type of compromise was needed, formally speaking, in order to include the Sunnis on the question of federalism. What happened here, in contrast to the debate on federalism, was that the Kurds were in principle opposed to the positions of the Shi’ite clerics on the role of Islam in the state and on the question of the personal status law. Had Ambassador Khalilzad sided with the Kurds, most likely the positions of the TAL could have been reaffirmed. This he did not do, and in particular he supported the key Shi’ite position that no role for Islam in the state could be guaranteed unless there were Shari’a experts as well as secular judges on the Supreme Federal Court (art. 89–91; new version: 91–93), the body that was now to unite the earlier planned two bodies of a Supreme Court and a Constitutional Court (defined as vaguely as it was in the TAL, which left more room, however, for another highest court of appeals, the Federal Court of Cassations).⁸⁵ Khalilzad may have done this to meet deadlines in the face of Shi’ite recalcitrance or because he was hoping for tradeoffs to the benefit of the Sunnis. He got only the first. But when the Kurds went public with secular grievances, Khalilzad too had to moderate his support for the Shi’ite position. In the end, the Kurds were satisfied to go along with Khalilzad because they too wanted to rush, now that they had gotten their way on federalism and Kirkuk,⁸⁶ and as a prize they received the all-important concession of stripping the Federal Supreme Court of the power to review the constitutionality of regional laws.⁸⁷ Whether or not their negative view on constitutional adjudication was reinforced by Bush v. Gore, as Galbraith rather absurdly suggests, I leave for others to determine.⁸⁸ I would rather think that anyone building a confederal enclave would resist extending the jurisdiction of a federal court for structural reasons well explored already by John C. Calhoun, that old enemy of federal judicial review.

The outcome was relatively complex, reflecting a compromise weighted in the direction of the majority’s position, but still a compromise. The constitutional draft once again made Islam as well as democracy and the rights of the constitution standards that all legislation would have to adhere to (arts. 2a, 2b, 2c). As before, these standards were likely to be contradictory, and conflicts would have to be resolved by the constitutional court, the Federal Supreme Court. Here lay the innovation sought by the Shi’a clerics, who understood that the declarations of the TAL remained merely symbolic without proper enforcement. That court, most dangerously from the secular point of view, was to contain both judges and experts in Shari’a jurisprudence (art. 90; current version: art. 91). The number of judges and the form of appointment, however, would be determined by the next National Assembly (similarly to the executive rules for region formation), but by a two-thirds majority! In this area, therefore, the way was open to a future consensual solution of the deferred issue (or to a hopeless stalemate), showing quite clearly that in the area of “federalism” (where a simple majority would decide), deferral was intended only as a smokescreen for the same nonconsensual solution on which agreement today is not possible.

The place of Islam in the state is not a function of the compromise concerning the court alone. Secular Iraqis, especially women, have been especially concerned about the repeated attempts of Shi’ite clerics to change the personal status law of 1959, according to which marriage, divorce, and inheritance law are legislated by the state and are uniformly administered by secular courts. The constitutional draft (art. 39; current version: art. 41) establishes the “freedom” of Iraqis to choose their status “according to their own religion, sect, belief and choice” and leaves the organization of these choices to an ordinary parliamentary law. Thus the majority will be free, as Nathan Brown shows, to give as much or as little weight to a secular alternative and a uniform administration as it chooses, or to establish religious predominance over all cases where even one litigant or interested party may so desire.⁸⁹ The mechanism here, as in the case of federalism, defers the decision in favor of later majoritarian imposition rather than operating in a constitutional, consensual process.

Let us put the two major proposals concerning the relationship of Islam and law together. The Supreme Federal Court can only be formed by a wide compromise. The personal status law, which could be judged unconstitutional by such a court, can be established by simple majority. But a court may not be able to form at all unless the Shi’ites get the number and type of Shari’a experts they want, and that their mode of appointment is acceptable to them. In either case (with a friendly court or with no court at all), a status law could be enacted by a simple majority that does not have to worry about constitutional review. No wonder that advocates of women’s rights and secularism are very upset about the end result of the process and about the betrayal of their cause by the American ambassador, who wound up accepting and even praising the relevant parts of the draft. They should have nevertheless noticed that in the first parliamentary session, each of the three members of the Presidency Council would have veto rights over parliamentary legislation (art. 135, 5th clause, A; current version: art. 137, 5th clause, A). Thus the election of one secular member could interfere with, but only for a single parliamentary period, the establishment of a status law that definitively decides the issue in favor of religious jurisdictions in this area.

There are finally issues having to do with the structure of government, issues we have heard little about in the press but that are of great importance. Here the Shi’ite majority and the Kurdish minority once again did not have the same interests, and probably the divergence again facilitated compromise solutions. The TAL, as I have argued, sought to reconcile an ultimately confederal bargain about the state with a centralistic version of parliamentarianism, by using consociational structures rooted mainly in the three-person presidency rather than by a bicameralism characteristic of a federal state. From the Shi’ite point of view, such a double defense of Kurdish positions had to seem illogical: if one has established a quasi-independent confederacy, why should the same force maintain veto rights over the government of the rest of the country? To the Kurds, the desire for this double defense was based not on logic but experience. It was clear that if they could not control the federal state, they did not mind incapacitating it, a position obviously inimical to the majority. A move in a more federal direction in the earliest constitutional drafts (acceptable to only literal nationalists among the Kurds) would have been based on the combination of a new, second, federal parliamentary chamber and the replacement of the consociational presidency council by a single, more ceremonial president who had to offer the prime ministership to the largest parliamentary bloc. The new draft, probably because of the inability of its framers to agree on a federal formula, solves these questions by somehow combining all these competing alternatives into a single package. There would be a second parliamentary chamber, a Council of the Union, including representatives of regions and provinces not in regions. But the definition of its rules of formation and powers (!) are left to the lower chamber, the Council of Deputies (Representatives) voting by two-thirds majority—and not the current Council of Deputies, but the one elected in the next elections (arts. 63 and 134; current version: arts. 64 and 136). An important consequence of this very unusual delay was the restoration for a single parliamentary period of the three-person, consociational Presidential Council of the TAL, elected on a single slate by a two-thirds majority, this time with a veto for each of the members, obviously as a replacement for the role of a federal chamber in national legislation (art. 135; current version: art. 137). These vetoes, as I have already said, do not exist in the case of region formation but apply to the new law governing personal status, the two-thirds law governing the composition of the Supreme Court, and indeed to constitutional amendments requiring presidential assent.

The amending structure of the new constitution, though rigid enough, is in fact more flexible than that of the TAL. Initiatives for amendments can come from the president and cabinet together or from one-fifth of the lower chamber, the Council of Representatives, which is the only parliamentary chamber given a right of participation in constitutional revision. Where in the case of the TAL no amendment could be made at all that would abridge rights, here (art. 123, current version: art. 125, 2nd clause) the basic principles of chapter 1 and the rights and freedoms of chapter 2 are unamendable only for the first two parliamentary cycles.⁹⁰ Subsequently, they can be amended according to the general rule that applies to most of the constitution: namely, two-thirds of the members of parliament, agreement of the president, plus majority support in a referendum. In the first parliamentary period, this means that the presidential council, one way or the other (by either the amendment rule or the ordinary rule of legislation as now revised), has the option of a consociational veto for each of its members. Finally, and given what is at issue today most importantly, no amendment is allowed that would lessen the powers of regions without the consent of their legislature and population in a referendum, unless one of the very few exclusive powers of the federal government is at issue (4th clause). This means that since the constitution already recognizes the powers of the Kurdistan Region (art. 114; current version: art. 116), no constitutional amendment can touch them.⁹¹ But interestingly enough, powers given by simple majority to new regions in the first session also would become by this clause amendment proof. If Iraq begins to break up according to the dispensation of the current draft, for example, a duly elected parliament would be powerless to legislate any countermeasures, regardless of the majority supporting them. If the state structure of the new constitution did not work, only a revolutionary overthrow could remedy the situation.

After the refusal to legally extend the process of constitution making as permitted by the TAL, the process was nevertheless extended in a manner that was most astonishing and in significant part illegal. In my view, the August 15 amendment that provided for one week’s extension was legal, although even this view has been contested. But the formal date of August 22, by which the draft should have been concluded and approved, came and went. Two illegal extensions were decided upon on August 22 and August 25. The constitutional text, though never voted on, was pronounced final on August 28, only it was not. A new text was brought to the National Assembly on September 13.⁹² These illegal extensions and the absurd claim supposedly based on the TAL (art. 60) that the National Assembly “writes” but does not “vote” on the constitution (“The National Assembly shall write a draft of the permanent constitution of Iraq”) would raise very serious questions in almost any situation concerning the meaning and legitimacy of the constitution being written. But Iraq is not at all an ordinary context for constitution making, and here illegality was an irrelevant detail almost no one cared about (at the time, on the Web, Nathan Brown, Juan Cole, and I were the exceptions).⁹³ Admittedly, it was much more important to register the fact that the thing could not be completed because, whatever the deadlines, Sunni inclusion or even the inclusion of one Sunni party could simply not be given up, both in light of the ongoing pressure of the American Embassy and for rational considerations, which very likely helped to reopen the question. Despite the original calculation of the two-thirds figure for each of the three provinces needed to block ratification, it occurred to the makers of the constitution that perhaps if the Sunni decided to participate and if their rejection were truly unanimous, the Shi’ite-Kurdish constitution might actually fail in three provinces.

Had such fears been there all along and been really serious, I believe that the whole negotiation process would have turned out quite differently. After all, it was the Kurdistan veto that made the Shi’ites make their deal with the Kurds within the TAL, a document they considered humiliating because of that veto, among other things. Would the Kurds have kept all they had in the TAL and have gained some new concessions without the veto? Their own advocates hardly think so. There was no comparable fear of a Sunni veto; if there had been, the Kurds would not have dared challenge the Sunnis as openly as they did. It was only in the end, when their own actions completely antagonized the Sunni community, bringing together radical and more moderate elements who were equally badly treated, that the Shi’ite and Kurdish parties lost a little of their confidence that the ratification rule was only a Kurdish veto after all. I doubt that they really thought a Sunni veto would work, but they wanted to make absolutely sure that it would not. Or, very possibly, what they feared was that they might have to engage in the fixing of the results of the referendum in the eyes of the international press. Thus their attitude changed with respect to the American effort to engineer more Sunni inclusion.

Negotiations concerning changes continued, and several changes were indeed made before the referendum of October 15, one on central control of water resources and another on having two deputy prime ministers. The most dramatic change came on October 12, three days before the referendum. Finally, a concession that appeared to be a genuine compromise was offered to the Sunni parties and was accepted by one of them, the Iraqi Islamic Party. In a highly revealing manner, the compromise came right after the failure of an attempt to guarantee the referendum results through open manipulation of the TAL text. We were told that originally it was the Independent Electoral Commission of Iraq that decided, absurdly and certainly against the most obvious intentions of the drafters and the plain meaning of the text itself, that the two-thirds of the voters of three provinces that would have to vote against the new constitution in order to reject it would have to be two-thirds of the eligible rather than the actual voters. In this case, two-thirds could be mustered perhaps in one province.⁹⁴ But when UN officials had some serious qualms about allowing this ruling to stand,⁹⁵ the National Assembly proceeded to vote by simple majority on October 2 that it must indeed be two-thirds of eligible voters in each of three provinces voting against the constitution if the text is to be rejected.⁹⁶ If the Shi’ites and Kurds wanted to do this, they should have legally attempted to amend the TAL by three-fourths of the votes, preferably when still under the veil of ignorance, and when the Kurds had reasons to preserve the veto they had long fought for. What the governing parties attempted to do instead, under the guise of mere interpretation of the TAL, was to interpret “voters” in the same sentence once as actual (when it had to do with approval by majority) and the second time as eligible (when it had to do with rejection by two-thirds of each of three provinces). Most embarrassingly, under open UN and this time fortunately tacit U.S. pressure, the very same National Assembly was forced, two days later, to withdraw the measure and return to actual rather than eligible voters, in the case of both majority approval and three-province rejection of the draft.⁹⁷

The solution may have raised the possibility of rejection anew. The compromise package offered to the Iraqi Islamic Party could have made a crucial difference in at least one Sunni-majority province, Nineveh, where the vote would be closer; we will never know for sure. In the actual event, 97 percent of the voters of Anbar, 82 percent of Salahddeen, but only 55 percent of Nineveh voted against the draft. Whatever the reasons, the constitutional referendum failed by two-thirds in only two Sunni provinces, and thus a document rejected by the immense majority of the Sunni community was ratified. But was the compromise offer a really serious one?

In its most positive interpretation, the October 12 deal once again reduced the constitution just passed to a provisional one and made the newly elected National Assembly yet another constitutional assembly given the time extension foolishly denied to its predecessor. What was done, technically speaking, seems, to be sure, more modest. The text of the draft that was subsequently voted on in the referendum was amended by a new article, 141 (which the voters never saw),⁹⁸ stating (my italics):

First: The parliament shall form, at the start of its work, a committee from its members, representative of the main components of the Iraqi society. The task of the committee is to present a report to the parliament, in not more than four months, including a recommendation of the necessary amendments that could be made to the constitution. The committee is dissolved after a decision about its suggestions is taken.

Second: The amendments suggested by the committee shall be presented, in one bulk, to the parliament to be voted on, and it is considered to be passed by the approval of the absolute majority of the members of the parliament.

Third: The articles amended by the parliament according to what came in provision (second) of this Article shall be put to the people for a referendum, not more than two months after the passing of the amendments in the parliament.

Fourth: The referendum on the amended articles is successful, by the approval of the majority of voters, and if not refused by two thirds of the voters in three governorates or more.

Fifth: The effect of Article (125) (related to amending the constitution) of this constitution is stopped, and its effect starts again after the amendments in this article are decided on.⁹⁹

Since the three-tiered and in all respects very difficult amendment rule of the new constitution was thereby suspended, everything could be changed for the four-month period, and the road was in principle indeed open to a historic compromise among the main groupings. The rights of regions could be altered, even without their consent. The most basic rights could be altered without waiting for a third parliamentary session. The new rule was more difficult only than the third, easiest amendment route of the constitution that dealt, by default, with mostly matters of the federal government and its branches. However, all this was true only in principle. Whether there was substantive hope for the Sunnis to alter a solid Shi’ite-Kurdish agreement on the major contentious issues was another matter. At the same time, the road was to be made formally very difficult by the restoration of the old three-province veto by two-thirds of the actual voters of each (see the fourth clause, above). As in the case of the TAL, that veto protected most of all those who would benefit from the current constitutional draft, since a rejection of a package of amendments would not return Iraq to a condition without a constitution, or even to the TAL itself, but to the new arrangements approved on October 15. The constitution itself was the new default position! Thus the same text could be viewed as provisional and permanent at the same time, depending on how one judged the likelihood of it being actually transformed. The supposed compromise maintained the dramatic inequality of the Sunni Arabs with respect to the other two main groups, because if in principle the Sunnis too could use the three-province veto (forgetting the experience of Nineveh), it would be useless to them given a situation where the fallback position was a constitution they entirely rejected. They were, in effect, delivered to the good will of their partners (assuming it existed), with the remote possibility that with allies they could hold government formation itself hostage to a prior constitutional deal.

Why the Iraqi Islamic Party accepted such a deal is difficult to comprehend. A few days before the referendum, they could play the card of delivering some Sunni votes, and in return this was the best deal they could get. Perhaps it was not much of a card and therefore not much of a deal. If they were bluffing and could not deliver any voters, which is possible in light of the results, the deal was then probably a concession worth having. But in return they lent the whole process more legitimacy and allowed the branding of their much more realistic partners as hardline rejectionists, which some of them undoubtedly were. If, however, they had the votes to defeat the referendum, the deal was a very poor one. New elections and new constitutional negotiations would have been preferable for the whole Sunni community and would have earned more credit with the insurrectionists, credit that could later have been cashed in on behalf of perhaps a slightly better political formula. Finally, it may very well be the case that they wanted to be in the new government at any cost and were waiting for sufficiently plausible cover, which the October 12 agreement finally gave them. Support for the referendum was then their ticket into a government of “national unity.”

Whether or not there was realistic hope to renegotiate the constitution, even the slightest chance of it depended on Sunni electoral performance. As part of the overtures toward moderate Sunnis, the (Transitional) National Assembly did in fact change the electoral rule, but by adopting a compromise formula. Of the 275 seats, 230 would be elected on non-turn-out-dependent provincial lists and only forty-five on a turnout-dependent national compensational list.¹⁰⁰ Facing in small part the same danger as before, and hoping perhaps after all to be able to renegotiate the constitution possibly on the basis of a role in a governmental coalition, Sunni parties and most armed groups of the insurrection now urged participation in the parliamentary elections of December 15. The apportionment of seats by the Independent Electoral Commission on the basis of ration cards caused some consternation; Sunnis felt that their districts received too few seats, but they nevertheless worked for a strong performance. Their initial hope was that with possible allies such as Allawi’s secular list (now called the Iraqi National List), they could have one-third of the seats in the National Assembly, and with that they could hold the election of the Presidency Council hostage to a constitutional deal, and along with it government formation, because there was no way that the candidate of the largest parliamentary group could be nominated as prime minister other than by that council. This was not to be. While the combined vote of the two Sunni lists, the Iraqi Accord Front dominated by the Iraqi Islamic Party and the Iraqi National Dialogue front led by al-Mutlaq, received over 19 percent and thus more or less an accurate demographic proportion of the vote, the Allawi list lost dramatically with respect to its earlier performance. There was no blackmail potential for the opponents of the constitution unless the Sadrists, now in the UIA, were to join them, and they were certainly unwilling, at that time at least, to break with the Shi’ite alliance supported by Sistani on behalf of the friends of the insurrection. Of course, the support of Kurdish parties for a constitutional deal would have been an adequate substitute, since without them it would be difficult for the Shi’ites to form a government if Allawi and the Sunnis too were in opposition. But the Shi’ite-Kurdish alliance was strong, and most likely there already was a new Barzani-Hakim deal over Kirkuk, in line with the new constitution, and the Kurds were about to get what the Jaffari government so far had denied them in exchange for support on the other major issues, namely a referendum over the fate of the city and possibly the province—a referendum that would allow them to take control of both.¹⁰¹

Thus no pressure could be exerted on the level of government formation for a new constitutional deal. The Iraqi Accord Front was welcome to join a government of national unity and even propose one member of the Presidency Council as a reward for its earlier support, but it had few bargaining chips left outside of the public-relations aspects of having some well-known Sunni leaders aboard (Adnan al-Dulaimi was now with the Iraqi Accord Front). With respect to the continuing insurrection, it probably had fewer ties with it now than did other Sunni parties and groups. In any case, the insurrection was no bargaining chip at all as long as the Americans were willing to fight it on behalf of the new Shi’ite-Kurdish alliance. That is also probably how Iran preferred it, looming in the background, waiting for the inevitable day when it would replace the Americans as the defender of the Shi’ite cause.

The victors together were much weaker than in the previous parliament and did not quite have the two-thirds necessary to name a Presidency Council, pass the laws that required two-thirds of the National Assembly, or amend the constitution (when this would be allowed in the ordinary way during this parliamentary session). They needed one other partner at least, and there was a lot of pressure to build a national unity government including Allawi and the Iraqi Accord Front. That would happen, not under Jaffari, the first nominee of the UIA, but rather Nuri al-Maliki, another Da’wa leader also supported by the Sadrists.

Summary of the December 15, 2005, Iraqi Council of Representatives election results

Here I am not concerned with the disastrous record of this “Government of National Unity,” inaugurated May 20, 2006, four and a half months after the general elections. The Constitutional Committee agreed upon in the compromise of October 12, “representative of the main components of the Iraqi society,” should have been formed by then, but it was not. Nor was it formed by this government, which noisily announced a national reconciliation program. It was only when the Council of Representatives (for the moment the only chamber of the National Assembly) wished to take up the issue of region formation under the constitution (art. 115), which certainly should not have been done before the four-month period for extraordinary amendments was over, that Tariq al-Hashimi of the Iraqi National Accord, now a vice president of Iraq, was able to gain as one of two concessions that the Constitutional Committee would now also begin to meet. The other concession was that whatever law was passed on region formation, it would be suspended for eighteen months.¹⁰² However, another law was passed, with Sunnis and many Sadrists boycotting, providing that once (as provided by the constitution) one-third of a provincial council or one-tenth of the citizens of a province called for the establishment of a region, this could now be done by the majority of a simple referendum of the province’s inhabitants. This law in effect reestablished the constitutional regulation of August 15 that was struck out on August 22 to please the Sunnis, but this time it was reestablished by a simple majority vote that would not have to be ratified as constitutional text. Once a region was thus formed, the result would be constitutional-amendment proof, because the rights of regions would be involved. Of course, the law could still be negated somehow by the results of the deliberations of the Constitutional Committee. But how realistic—and even how legal—would that be?

As to legality, the October 12 amendment to the constitution stated: “at the start of its [the council of representatives’] work” and “no more than four months.” That work could have started on March 19, 2006, when (then interim) President Talabani convened it; on April 22, when a speaker was elected; or on May 20, when the government took office. Or perhaps one should count the four months only from the time when the committee was set up, presumably in October or November. This means that the deadline for submitting a package of amendments has long past (the deadline being, under the most generous interpretation, May 2007). Any amendments submitted now could easily be challenged as irrelevant and illegal.

Regarding political viability, even before the test of the three-province veto, amendments would have to be passed by the very same majority (138 votes exactly) that more or less passed the federalism law last October. Why would they do so? There are only two factors (or their combination) that could help achieve the historic compromise that would lead to a renewal of the constituent process and significant, consensual constitutional amendments. The first would be a possible split in the Shi’ite camp itself, as indicated by the earlier opposition of the Sadrists and the Jaffari government to the Hakim plan of nine-province regionalism, as well as recent proposals by people as different as Sadr himself (who now controls a significant bloc of deputies) and Kanan Makiya to table the “federalism” question for a lengthy period.¹⁰³ In effect, these proposals all mean a return to the TAL’s formula of an asymmetric structure with a confederal status for Kurdistan, leaving the exact nature of the organization of the rest of Iraq undetermined, but with the eventual possibility of elements of provincial federalism as well as the formation of smaller, weaker regions. So far, all such proposals have been swamped by the strength of SCIRI, with the probable backing of Iran. Their slight chance of success has become therefore dependent on what the Americans choose to do.

By “what the Americans choose to do” I don’t mean the strength of their visible pressure, which may actually be counterproductive. If it is true that they want a significant force to stay indefinitely in Iraq, the Americans have little leverage to impose a really fair bargain, however much they talk about benchmarks and inclusion. The Shi’ites need not fear a Sunni insurrection as long as it is more or less neutralized by a superpower, nor need the Shi’ites take the Sunni political wing seriously. The amazing thing is that at least Secretary of State Rice and Ambassador Khalilzad, if not the Pentagon as well, in the end probably came to understand that they were guarding Iran’s prize and that many of the Sunnis they were shooting at were their geopolitical allies. But there is nothing they can do about this as long as they also want to stay in Iraq, because in that case the Sunnis regard them as their main enemy, while the Shi’ites, though they hate the occupiers too, are willing to treat them as allies, however unlikely such an alliance between the United States and close friends of Iran may be. Undoubtedly, leaving suddenly would indeed produce the chaos that would make the already very real failure of the U.S. government visible to all who now choose to pretend otherwise. Though they may have come to this conclusion belatedly, the Democrats in Congress are absolutely right: what the United States could and should do if it wishes to salvage something from the whole lamentable operation is to produce a timetable for withdrawal to pressure the Shi’ites to take their own promise to renegotiate the constitution seriously. In addition, the United States should use a regional agreement supported by an international, Security Council–authorized coalition to make sure that Iran does not take up the slack as they progressively withdraw. But will anyone listen to them after the fiasco that the adventure in Iraq has become?