Testimony to House Subcommittee on National Security, Emerging Threats, and International Relations

Testimony to U.S. House of Representatives, Committee on Government Reform, Subcommittee on National Security, Emerging Threats, and International Relations on “Iraq: Democracy or Civil War?”

James D. Fearon

Geballe Professor in the School of Humanities and Social Science and

Professor of Political Science

Stanford University

September 15, 2006

1. Introduction

Let me begin by thanking Chairman Shays and the subcommittee for inviting me to testify. I am honored to have the opportunity to speak with you.

I am not a specialist on the politics of the Middle East, but I have spent a lot of time studying the politics of civil wars. In what follows, I begin with an executive summary. I then discuss some common patterns in how civil wars evolve and how they end, trying to assess where Iraq fits relative to the general pattern. The last four sections consider what this implies for US policy in Iraq looking forward.

2. Executive summary

By any reasonable definition, Iraq is in the midst of a civil war, the scale and extent of which is limited somewhat by the US military presence.

Civil wars typically last a long time, with the average duration of post-1945 civil wars being over a decade.

When they end, they usually end with decisive military victories (at least 75%).

Successful power-sharing agreements to end civil wars are rare, occurring in one in six cases, at best.

When they have occurred, stable power-sharing agreements have usually required years of fighting to reach, and combatants who were not internally factionalized.

The current US strategy in Iraq aims to help put in a place a national government that shares power and oil revenues among parties closely linked to the combatants in the civil war. The hope is that our presence will allow the power-sharing agreement to solidify and us to exit, leaving a stable, democratic government and a peaceful country.

The historical record on civil war suggests that this strategy is highly unlikely to succeed, whether the US stays in Iraq for six more months or six more years (or more). Foreign troops and advisors can enforce power-sharing and limit violence while they are present, but it appears to be extremely difficult to change local beliefs that the national government can survive on its own while the foreigners are there in force. In a context of many factions and locally strong militias, mutual fears and temptations are likely to spiral into political disintegration and escalation of militia and insurgent-based conflict if and when we leave.

Thus, ramping up or “staying the course” amount to delay tactics, not plausible recipes for success as the administration has defined it.

Given that staying the course or ramping up are not likely to yield peace and a government that can stand on its own, I argue for gradual redeployment and repositioning of our forces in preference to an extremely costly permanent occupation that ties our hands and damages our strategic position in both the region and the world.

Redeployment and repositioning need to be gradual primarily so that Sunni and Shiite civilians have more time to sort themselves out by neighborhood in the major cities, making for less killing in the medium run. Depending on how the conflict evolves, redeployment might take anywhere from 18 months to 3 years.

The difficult questions for US policy concern the pace and manner of redeployment: how to manage it so as to maximize the leverage it will give us with various groups in Iraq; and how to manage it so as to minimize the odds of terrorists with regional and global objectives gaining a secure base in the Sunni areas.

2. What is a civil war?

A civil war is a violent conflict within a country, fought by organized groups that aim to take power at the center or in a region, or to change government policies.

How much violence is enough to qualify a conflict as a civil war as opposed to terrorism or low-level political violence is partly a matter of convention. By any reasonable standard, however, the conflict in Iraq has killed enough people to put it in the civil war category. For example, political scientists often use a threshold of at least 1,000 killed over the whole course of the conflict to mark off civil wars. One might consider this too low to capture our everyday understanding. But the rate of killing in Iraq – easily more than 30,000 in three years – puts it in the company of many recent conflicts that few hesitate to call “civil wars” (e.g., Sri Lanka, Algeria, Guatemala, Peru, Colombia).

An insurgency is best understood as a type of civil war (assuming it kills enough people). In Iraq, the civil war began with a disorganized insurgency of Baathists, Sunni nationalists, and foreign jihadis using violence in hopes of expelling the US and destroying or replacing the Iraqi government set up after US invasion. In the last six months, insurgent attacks against Shiites have led to a widening of the civil war, as Shiite militias have responded by attacking Sunnis in the major cities. Sunni-Shiite militia and communal conflict has worsened in part as a direct result of the US strategy of “Iraq-ization” of the war effort.

Following our own example of the US Civil War, many Americans think of a civil war as a conflict that involves virtually everyone in a country and that sees fighting by regular forces along clear frontlines. If one uses this definition, then it might still be possible to maintain that Iraq is “not yet in civil war.” But the argument between the administration and its critics over the definition of civil war is really a domestic political dispute that does not help us understand what is going on there. The violence in Iraq bears a strong resemblance to many internal conflicts around the world that are commonly described as civil wars, and it is instructive to compare them.

Moreover, the US Civil War was atypical – civil wars rarely involve regular armies fighting along clear-cut frontlines. Instead, insurgency and militia-based conflicts like what we see today in Iraq today have been far more common.

3. How civil wars end

Civil wars typically last much longer than international wars. For civil wars beginning since 1945, the average duration has been greater than 10 years, with fully half ending in more than seven years (the median). The numbers are fairly similar whether we are talking about wars for control of a central government, or wars of ethnic separatism.

When they finally do end, civil wars since 1945 have typically concluded with a decisive military victory for one side or the other. In contests for control of the central state, either the government crushes the rebels (at least 40% of 54 cases), or the rebels win control of the center (at least 35% of 54 cases). Thus, fully three quarters of civil wars fought for control of the state end with a decisive military victory.

Quite often, in perhaps 50% of these cases, what makes decisive victory possible is the provision or withdrawal of support from a foreign power to the government or rebel side. For example, the long civil war in Lebanon ended in 1991 after the US and Israel essentially changed their positions and became willing to see the Syrian-backed factions win control if this would lead to peace. International intervention in civil wars is extremely common and often determines the outcome.

Power-sharing agreements that divide up control of the central government among the combatants are far less common than decisive victories. I code at most 9 of 54 cases, or 17%, this way. Examples include El Salvador in 1992, South Africa in 1994 and Tajikistan in 1998.

In civil wars between a government and rebels who are fighting for secession or greater autonomy, negotiated settlements that confer some local autonomy have occurred in about one third of the 41 such wars that began after 1945 and have since ended. This leaves two thirds as cases where the government crushed the regional rebels, or the rebels won military victories that established a de facto autonomous state.

In the rare cases where they have occurred, successful power-sharing agreements have usually been reached after an intense or long-running civil war reaches a stalemate. One of the main obstacles to power-sharing agreements seems to be political and military divisions within the main parties to the larger conflict.

4. Why is successful power-sharing to end civil wars so rare?

If successful power-sharing agreements rarely end civil wars, this is not for lack of trying.

Negotiations on power-sharing are common in the midst of civil war, as are failed attempts to implement such agreements, often with the help of outside intervention by states or international institutions. For example, the point of departure for the Rwandan genocide and the rebel attack that ended it was the failure of an extensive power-sharing agreement between the Rwandan government, Hutu opposition parties, and the RPF insurgents.

The main reason power-sharing agreements rarely work is that civil war causes the combatants to be organized in a way that makes them fear that the other side will try to use force to grab power, and at the same time be tempted to use force to grab power themselves. These fears and temptations are mutually reinforcing. If one militia fears that another will try to use force to grab control of the army, or a city, then it has a strong incentive to use force to prevent this. The other militia understands this incentive, which gives it a good reason to act exactly as the first militia feared.

In the face of these mutual fears and temptations, agreements on paper about dividing up or sharing control of political offices or tax revenues are often just that – paper.

For example: Current US policy seeks to induce Shiite leaders to bring Sunni leaders into the national government and provide them with some spoils of office. The hope is that this will get Sunni leaders to work against the insurgency. There is some evidence that the strategy has been partly effective, at least in terms of bringing significant Sunni leaders in. But why, in the longer run, should Sunni leaders believe that once the US leaves, the Shiites who control the army/militias would continue to pay them off? The same question applies to proposals to change the constitution to “ensure” (on paper) that the Sunni regions gain an equitable part of Iraq’s oil revenues.

Given the vicious fighting that has occurred and the deep factionalization among the Shiites, Sunni leaders would have to be crazy to count on the good will or good faith of Shiite and Kurdish leaders to ensure that a political deal would be respected after the US leaves. The only long-run stable and self-enforcing solution would be for an implicit Sunni threat of renewed insurgency to keep a Shiite-dominated government sticking to power- and oil-revenue sharing arrangements. High levels of factionalization on both sides imply that such an arrangement will probably be impossible to reach without years of fighting to consolidate the combatants and clarify their relative strength.

A second example: Right now representatives of Shiite political factions with ties to different clusters of militias share power in the national parliament and across government ministries. The expectation that US forces would act to prevent illegal grabs of power at the national level, and wholesale attacks by, say, Mahdi Army militias against Badr Brigade militias over territorial control in Baghdad and other cities, is making for an armed and fractious peace between Shiite factions. Regardless of written constitutional rules and procedures, after the US leaves these Shiite factions and their affiliated militias will fear power grabs by the other and be tempted by the opportunity themselves. An intra-Shiite war is thus a plausible scenario following US withdrawal, whether that should come in six months or five years.

In sum, civil wars for control of a central government typically end with one-sided military victories rather than power-sharing agreements, because the parties are organized for combat and this makes trust in written agreements on the allocation of revenues or military force both dangerous and naïve. The US government and Iraqi politicians have attempted to put a power-sharing agreement in place in the context of a new, very weak central government and a violent insurgency and attendant militia conflicts. While the US military could easily destroy Saddam Hussein’s formal army, militias and insurgents are “closer to the ground” and cannot by completely destroyed or reconfigured without many years of heavy occupation and counterinsurgency, if even then.

This means that however long we stay, power-sharing is likely to fall apart into violence once we leave.

5. Likely consequences of US withdrawal: Iraq versus Bosnia

What will that violence look like, on what scale and with what consequences? A central argument against rapid withdrawal of US troops is that this would lead to a quick descent into all-out civil war. The example of Bosnia in 1992 is sometimes invoked, when systematic campaigns of ethnic cleansing caused the deaths of tens of thousands in the space of months.

Though there are some important differences, the analogy is a pretty good one. As argued above, US withdrawal, whether fast or slow, is indeed likely to cause higher levels of violence and political disintegration in Iraq. But rapid withdrawal would be particularly likely to lead to mass killing of civilians.

In Bosnia, massive and bloody ethnic cleansing was the result of systematic military campaigns directed by irredentist neighboring states and their local clients. For Milosevic and the Bosnian Serb leadership, the whole point was to rid eastern Bosnia and Banja Luka in the west of Muslims.

To my knowledge, no significant players on either the Sunni or Shiite side talk about wanting to break up Iraq by creating a homogenous Sunni or Shiite polity. Instead there remains a strong sense that “we are all Iraqis,” even if they may strongly disagree about what this implies for politics.

To date, “ethnic” cleansing in Iraqi cities has been much less systematic, less centrally directed, and more individual than it was in Bosnia in 1992. The breakdown of policing plus insurgent attacks have led to the supply of local “protection” in the form of sectarian militias and gangs. Whether seeking generic revenge, suspected killers from the other side, or profit from extortion and theft, gangs make life extremely dangerous for members of the minority faith in their neighborhood. So Shiites exit Sunni-majority neighborhoods while Sunnis exit Shiite-majority neighborhoods. This is a “dirty war” in which gangs torture and kill suspected attackers or informants for the other side, along with people who just get in the way or have something they want.

Rapid reduction in US troop levels is not likely to cause a massive spasm of communal violence in which all Shiites start trying to kill all Sunnis and vice versa. But it may spur Moqtada al-Sadr to order his Mahdi army to undertake systematic campaigns of murder and, in effect, ethnic cleansing in neighborhoods in Baghdad and other cities where they are strong. Obviously a murky subject, some recent reports suggest that such plans exist.

Gradual redeployment and repositioning of US troops within the region is needed to allow populations to sort themselves out and form defensible lines that would lessen the odds of sudden, systematic campaigns of sectarian terror in mixed neighborhoods. This is one of the strongest arguments against rapid US military withdrawal. Gradual redeployment – or, for that matter, “staying the course” – improves the chances of a less violent transition to a “Lebanon equilibrium” of low-level, intermittent violence across relatively homogeneous neighborhoods controlled by different militias.

If Bosnia in 1992 serves as an instructive historical analogy, Bosnia in 1995 – when the Dayton agreements formally ended the war and initiated a power-sharing arrangement among the combatants – is much more problematic. I consider this comparison farther below.

6. Likely consequences of US withdrawal: Lebanon 1975-91 versus Turkey 1977-80

Turkish cities between 1977 and 1980 experienced major violent conflict between local militias and paramilitaries aligning themselves with “the left” or “the right.” A standard estimate is that more than twenty people were killed per day on average, in thousands of attacks and counterattacks, assassinations, and death squad campaigns working off lists of enemies. Beginning with a massacre by rightists in the city of Kahramanmaras in December 1978, the left-right conflicts started to widen into ethnic violence, pitting Sunnis versus Alawites versus Kurds and Shiias in various cities.

As in Iraq today, the organization of the combatants was highly local and factionalized, especially on the left. The results often looked like urban gang violence. But, as in Iraq, the gangs and militias had shady ties to the political parties controlling the national parliament (like Iraq, Turkey in those years had a democratically elected government). Indeed, one might describe the civil conflicts in Turkey then and Iraq now as “militia-ized party politics.”

Intense political rivalries among the leading Turkish politicians, along with their politically useful ties to the paramilitaries, prevented the democratic regime from moving decisively to end the violence. Much as we see in Iraq today, the elected politicians fiddled while the cities burned. Fearing that the lower ranks of military were starting to become infected by the violent factionalism of society, the military leaders undertook a coup in September 1980, after which they unleashed a major wave of repression against both left and right-wing militias and gang members. At the cost of military rule (for what turned out to be three years), the urban terror was ended.

Could US withdrawal from Iraq lead to military coup in which the Iraqi army leadership declares that the elected government is not working and that a strong hand is necessary to bring basic order to Iraq? Probably not. The Turkish military is a strong institution with enough autonomy from society and loyalty to the Kemalist national ideal that it could act independently from the divisions tearing society apart. Though the army favored the right more than the left, Turkish citizens saw the army as largely standing apart from the political and factional fighting, and thus as a credible intervener.

By contrast, the Iraqi army and, even more, the police force, appear to have little autonomy from society and politics. The police look like militia members in a different uniform, sometimes with some US training. The army has more institutional coherence and autonomy from militias than the police, but it seems Shiite dominated at this point with few functional mixed units. A power grab by some subset of the army leadership would be widely interpreted as a power grab by a particular Shiite faction, and would lead the army to disintegrate completely along sectarian and possibly factional lines.

What happened in Lebanon in 1975-76 is a more likely scenario. As violence between Christian militias and PLO factions started to take off in 1975, the army leadership in Lebanon initially stayed out, realizing that if they tried to intervene the national army would splinter. The violence escalated and eventually the army intervened, at which point it did break apart. Lebanon then entered a long period during which an array of Christian, Sunni, Shiite, and PLO militias fought each other off and on, probably as much within sectarian divides as across them. Syrian and Israeli military intervention sometimes reduced and sometimes escalated the violence. Alliances shifted, often in Byzantine ways. For example, the Syrians initially sided with the Christians against the PLO.

To some extent this scenario is already playing out in Iraq. US withdrawal – in my opinion whether this happens in the next year or in five years – will likely make Iraq (south of the Kurdish areas) look even more like Lebanon during its long civil war.

As in Lebanon, effective political authority will devolve to city, region, and often neighborhood levels, and after a period of fighting to draw lines, an equilibrium with low-level, intermittent violence will set in, punctuated by larger campaigns financed and aided by foreign powers. As in Lebanon, we can expect a good deal of intervention by neighboring states, and especially Iran, but this intervention will not necessarily bring them great strategic gains. To the contrary it may bring them a great deal of grief, just as it has the US.

The Lebanese civil war required international intervention and involvement to bring to conclusion. If an Iraqi civil war post-US withdrawal does not cause the formal break up of the country into three new states, which it could, then ending it will almost surely require considerable involvement by regional states to make whatever power-sharing arrangements they ultimately agree on credible. If Iraq is a bleeding sore in the heart of the Middle East for years (recall that civil wars typically last a long time), then its Sunni-and Shiite-led neighbors may have to come to a region-wide political agreement to be able to enjoy political and economic stability again.

7. “Ramping up” or “staying the course” are delay tactics, not a “strategy for victory”

In broad terms, the US has three options in Iraq: (1) ramp up, increasing our military presence and activity; (2) “stay the course” (aka “adapt to win”); and (3) gradual redeployment and repositioning our forces in the region, so as to limit our costs while remaining able to influence the conflict as it evolves.

The analysis above suggests that none of these options is likely to produce a peaceful, democratic Iraq that can stand on its own after US troops leave. While we are there in force we can act as the guarantor for the current or a renegotiated power-sharing agreement underlying the national government. But, in a context of many factions and locally strong militias, mutual fears and temptations will spiral into political disintegration and escalation of militia and insurgent-based conflict if and when we draw down.

“Ramping up” by adding more brigades could allow us temporarily to suppress the insurgency in the Sunni triangle with more success, and to prevent “Al Qaeda in Iraq” (AQI, which now consists overwhelmingly of Iraqi nationals) from controlling the larger towns in this area. Ramping up could also allow us to temporarily bring greater security to residents of Baghdad, by putting many more troops on the streets there.

But Congress and the Bush administration have to ask what the long-run point is. The militia structures may recede, but they are not going to go away (absent some truly massive, many-decade effort to remake Iraqi society root and branch, which would almost surely fail). Given this, given myriad factions, and given the inability of Iraqi groups to credibly commit to any particular power- and oil-sharing agreement, ramping up or staying the course amount to delay tactics, not plausible recipes for success.

Note that more than ten years after NATO intervention, Bosnia is still at risk of political disintegration and possibly a return to some violence if the international guarantor closes up shop. And in that case the main combatants were not highly factionalized and had already fought to a stalemate by the summer of 1995, before the NATO bombing campaign and the Dayton agreement on power-sharing. Likewise, no one can imagine that Afghanistan would not return to chaos and full-blown civil war if NATO and US troops were to leave.

A long-term US military presence in Iraq is probably less likely to produce a regime that can survive by itself than the international intervention in Bosnia has been, and no more likely than in Afghanistan. Moreover, a permanent US military presence in Iraq will be vastly more costly in terms of lives, money, and America’s larger strategic position and moral standing than the international commitment to Bosnia or Afghanistan has been.

Congress has to ask whether spending more than 60 billion dollars per year in Iraq for a mission that is unlikely to produce a decent government that can stand on its own is the best use of this money for protecting the US from terrorism.

8. Costs of redeployment and repositioning

Even if ramping up or staying the course are not “strategies for victory” as the administration has defined it, this does not imply that immediate withdrawal is the best course of action.

Indeed, in principle it could be that the costs of withdrawal are so high at this point that the best option is to continue the status quo as long as possible. I seriously doubt this is the case. But I would agree that there are potential risks and costs to US national security from reducing our troop presence in Iraq, and that the question of “how to do it” to minimize these risks and costs is extremely complicated.

There are three major areas where reduction of US troop presence in Iraq could have costs and risks that need to be considered in thinking through the best feasible pace and manner of redeployment:

(a) Iraqi civilian deaths;

(b) the threat of “Al Qaeda in Iraq” gaining secure base areas and using them to organize terror attacks against countries in the region and the US; and

On (a), Iraqi civilian casualties: Rapid withdrawal of US forces would most likely cause rapid escalation of the sectarian and intra-sectarian dirty war, making for a sharp rise in civilian deaths well above the current rates.

A more gradual reduction and repositioning of US forces within the region would be far better, as it would allow mixed populations to sort themselves out in the larger cities, and to keep the rate of escalation of militia conflict as low as feasible. Gradual redeployment would also allow the US to prevent (through joint operations and other such mechanisms) the Iraqi army from rapidly becoming a full partisan in the dirty war.

On (b), Al Qaeda in Iraq: AQI is now the principle insurgent enemy of US forces in the Sunni-dominated provinces to the west of Baghdad, although it remains unclear how to interpret the nature and likely trajectory of this organization. AQI apparently consists overwhelmingly of Iraqi nationals, with at best a small fraction of foreign jihadis involved. It has been successful in controlling cities and territory because of its brutality; financing through the hawala system and some foreign sponsors; in some cases mistakes in US counterinsurgency strategy and tactics; the general fear and resentment of Shiite dominance that prevails in the Sunni areas; and general dislike of the US presence in the country.

Reduction of US troop presence in Al Anbar and the other Sunni-majority provinces would almost certainly lead to AQI and other Sunni insurgent forces taking fuller political and military control in these areas. The question for US policy is what sort of threat this would pose US interests, and what could be done about it.

Though we are obviously in the realm of speculation here, I think the common assumption that Al Anbar would become like southern Afghanistan under the Taliban – a home for Al Qaeda training camps producing terrorists to attack the US homeland – merits critical scrutiny.

In the first place, by redeploying our forces within the region, we could retain the ability to prevent large-scale operations of this sort. Second, given that AQI is manned almost entirely by Iraqi nationals who will be fighting the Shiite-dominated Iraqi army, it is not clear that it would see any particular advantage or interest in the global and anti-Western terrorist program of Bin Laden and al-Zawahiri. To the contrary, they might become particularly interested in gaining US covert support (or lack of opposition).

Third, if AQI’s hawala funding is driven by the perception that they are taking on the US, this could begin to dry up following US redeployment and a shift of the conflict to clearer Sunni/Shiite lines. If so, there would be an opportunity for US dollars and Iraqi government dinars to buy away the support of various sheiks and local power-holders who would be forced under AQI’s sway initially.

Likewise, even if “AQI” does gain ground in the Sunni-majority provinces, it is unlikely that it will be or become a coherent political organization any time soon. Rather, high levels of factionalization will persist, limiting its capacity for coherent action on a global or regional scale (if it even has the inclination) and making it easier to acquire information about who is doing what.

On (c), Iranian influence: This seems to me the least persuasive argument about the costs of reducing the US military presence in Iraq.

In the first place, it should be stressed that if the US were to succeed in helping to set up a peaceful, democratic Iraqi government that can stand on it own, there is no question but that Iran would have much more influence with this government and in the Middle East in general than it had under the Saddam Hussein regime. The demographic fact is that democracy in Iraq means, to some large extent, rule by Iraqi Shiites, who have close religious and political ties to Iran. The example of a Shiite-dominated Arab (and possibly Kurdish) state would inevitably have major ramifications to the west of Iraq.

Compared to this scenario (which is the implausible object of current US policy), the scenario of a Lebanon-like civil war in Iraq that follows US redeployment probably implies less Iranian influence in the Middle East as a whole, and more costly Iranian influence in Iraq for Iran itself. Iran would be draw in, much more than at present, to funding and arming Shiite factions against each other and against Sunni insurgents. Even if they manage to establish a Shiite faction in a relatively dominant position in Baghdad, their clients will be highly ungrateful if Iran subsequently tries to steal oil revenues, and they will probably have to face the costs of an unremitting Sunni insurgency. The unlikely event of a military invasion by Iranian forces to grab oil fields in the south could be made even more unlikely by appropriate repositioning of US forces in the region.

Various Iranian leaders have said that they much prefer the US continue to “stay the course” in Iraq, and that they are quite worried about the prospect of an escalated civil war on their doorstep. With respect to Iranian influence and overall strategic position, redeployment of US forces would most likely increase the US’s leverage and would not advantage Iran more than the current policy does.

9. Conclusion

“Staying the course” or “ramping up” in Iraq may put off political disintegration and major escalation of the civil war in progress, but are unlikely to produce a democratic government that can stand on its own and maintain peace after US troops are gone. The most likely scenario following reduction of US troop presence is the escalation of a Lebanon-like civil war. Unfortunately, the odds that this will occur are probably not much better if US troops stay for five (or even more) years as opposed to one.

The evidence supporting this assessment is drawn from the experience of other civil wars. Historically, civil wars tend to last a long time and usually end with decisive military victories. Successful power-sharing agreements to end civil wars are rare. When they have occurred, they have typically required that the combatants not be highly factionalized and that the balance of military power and prospects for victory be well established by years of fighting.

The US has tried to help into being a democratic Iraqi government that depends on power- and oil-revenue sharing among the major religious and ethnic groups. Probably the most common piece of advice these days from US experts and pundits on the question of “how to save the Iraq mission” is that a new political or constitutional bargain must be struck that gives the Sunnis clear assurance of a fair share of the oil revenues. But even if the terms of the constitution are altered – which seems unlikely given Shiite and Kurdish opposition to what is an excellent deal for them – it is not clear why Sunnis would have a good reason to believe that the terms would be respected, especially after the US departs.

In addition, many Sunnis, especially those in the insurgency, seem to believe that they would have a good military chance against the Shiite-dominated government if the US were gone. This belief is hard to change while the US is backing or back-stopping the Iraqi army.

In addition to being logistically problematic, rapid US troop withdrawal from Iraq would yield rapid escalation of militia violence and empowerment of the extremely brutal Sadrist faction on the Shiite side. Redeployment and repositioning of US troops therefore needs to be gradual and tuned to circumstances as they develop, undertaken always with an eye to the deals that can be struck with the various players. The US needs to develop a surge capacity and rapid response forces in Iraq or in the region in order to take on armed groups that get especially ambitious and so threaten to cause quicker escalation of the civil war.

We should not give up on the prospect that Iraqi political leaders will manage to make deals and provide services in such a way as to gain peace and security for the country as a whole. But we should make it clear, at least privately, that their time to do so is limited. In the interim, we need to plan for the possibility that a democratic Iraq that can stand on its own is not going to take root while we are there. This means planning to put ourselves in the best position to influence for the good the evolution of a civil conflict that only Iraqis have the power to end at this point.