The Rollback Fantasy

Daniel Byman, Kenneth Pollack, Gideon Rose. Foreign Affairs. Volume 78, Issue 1. January/February 1999.

Four decades ago a Third World dictator threatened American interests in a crucial region. Unwilling to pay the costs of an invasion or settle for containment, U.S. policymakers convinced themselves that a cheap and easy third option existed: support for some of the dictator's domestic opponents, whose efforts would supposedly spark a popular uprising and topple the regime. The resulting invasion attempt by Cuban exiles at the Bay of Pigs turned into one of the worst fiascoes in the history of American foreign policy.

Incredibly, a similar concept-using the Iraqi opposition to overthrow Saddam Hussein-is one of the hottest foreign policy ideas in Washington today. From congressional leaders to a galaxy of former government officials, from The Weekly Standard, National Review, and Commentary to The New Republic and columnists at The New York Times, The Washington Post, and The Nation, support for the Iraqi opposition has become the leading alternative to the Clinton administration's current policy. One recent letter urging the president to back Saddam's foes more strongly was signed by three former Republican national security advisers, three former Republican secretaries of defense, and seven former Republican subcabinet officials; another prominent supporter of the idea has been Clinton's first CIA director, R. James Woolsey. As a result, in October Congress hurriedly embraced an Iraq Liberation Act authorizing $97 million in military aid to democratic Iraqi resistance groups. The president's staff was unenthusiastic, but Clinton signed rather than pick a fight. And after yet another showdown with Iraq over U.N. weapons inspections last November, the president himself jumped on the bandwagon by touting the administration's "engagement with the forces of change in Iraq" and its intention to work for "a government in Baghdad-a new government-that is committed to represent and respect its people, not repress them."

The basic idea behind the boomlet is simple: The United States should refuse to accept the continued rule of Saddam Hussein and help install the Iraqi opposition in his place. As Senate Majority Leader Trent Lott (R-Miss.) put it, "The doctrine here has to be rollback, not containment." In The New Republic, former Under Secretary of Defense Paul Wolfowitz wrote, "Toppling Saddam is the only outcome that can satisfy the vital U.S. interest in a stable and secure Gulf region, because, to a degree unique among contemporary tyrannies, the Iraqi regime is Saddam Hussein.... The vast majority of Iraqis want him out, though only those with a death wish say so openly. We must help them." Moreover, Saddam could supposedly be replaced at relatively low cost-even, many advocates argue, without committing American ground troops. The world could be free of Saddam, Wolfowitz argued in testimony before Congress, if the Clinton administration would only "muster the necessary strength of purpose."

The fly in the ointment is that this is simply not true. Even if rollback were desirable, any policy to achieve it would have to pass three tests to be considered seriously. It would have to be militarily feasible, amenable to American allies whose cooperation would be required for implementation, and acceptable to the American public. All current rollback plans involving the Iraqi opposition come up short. Those who tout these nostrums as superior to existing U.S. policy are therefore either engaging in wishful thinking or cynically playing politics. Either way, for the United States to try moving from containment to rollback in Iraq would be a terrible mistake that could easily lead to thousands of unnecessary deaths.

The Airpower Approach


Contrary to the impression sometimes given by backers of the Iraqi opposition, there is no single, simple rollback plan. Instead, three competing variants of the idea have been put forward: using airpower to oust Saddam, helping the opposition seize large chunks of Iraq, or sponsoring a guerrilla war by his foes.

The first-the airpower approach-would have the United States and the opposition team up for a decisive military campaign against the Iraqi regime. The problem here is that almost all the heavy lifting would have to be done by the United States, which would incur huge costs without being assured of success in return.

The airpower approach calls for the United States to help the Iraqi opposition recruit, train, and equip tens of thousands of soldiers for limited conventional warfare. When this opposition army was ready to take the field, the United States would launch an air offensive against Saddam's forces to devastate Iraqi combat formations, logistics, command and control, and rear-echelon positions. The opposition troops would creep forward in the wake of the aerial onslaught, forcing Saddam's troops to concentrate for battle and thus become an even better target for destruction from above. As the defending forces retreated or broke up, the opposition would occupy more and more territory, eventually marching virtually unimpeded to Baghdad, where it would set up a new regime.

This plan has two advantages. First, it correctly acknowledges that defeating Saddam's armed forces will be difficult and that the Iraqi opposition cannot do the job without extensive U.S. military support. Second, the plan might be acceptable to American allies, who have sometimes suggested that they would be willing to support a policy to remove Saddam if they were convinced it would be carried through to completion. Although friends of the United States in both Europe and the Middle East have generally opposed the use of force against Iraq during recent crises, few would be sorry to see Saddam go as long as regional stability was not greatly disrupted and national borders remained unchanged. They might just accept his removal by force if the United States insisted on it.

The catch, however, is that the plan requires a truly massive U.S. air campaign, and even then an opposition victory would be unlikely. True, Iraq's conventional military capabilities are now extremely limited, and Iraqi forces could offer only token resistance to American or other Western forces. Those same units, however, would be a formidable adversary for the Iraqi opposition. The Iraqi military currently keeps 400,000 men under arms, including almost 100,000 in the Republican Guard and the Special Republican Guard. Saddam also has 2,000 tanks and 2,100 artillery pieces. Lightly armed opposition troops would stand little chance in open combat against these forces, especially the Republican Guard. As such, for this plan to work the U.S. air campaign would essentially have to destroy all organized Iraqi military resistance. This would require an enormous and costly effort over several months, probably on a scale approaching that of the air campaign in the 1991 Gulf War.1 Although the American people strongly oppose Saddam's regime, they would probably not support a long, bloody campaign against Iraq without an obvious goad other than U.S. frustration.

Airpower alone has never forced a rival regime to capitulate. Few Iraqi units actually fell apart or surrendered en masse even under the pounding they took in Operation Desert Storm. Those that did crack did so partly out of fear of the impending coalition ground offensive, a fear the much less menacing opposition forces would not inspire. And no Republican Guard divisions cracked, although several were subjected to more than a thousand bombing sorties apiece. In short, it is at best unproven and at worst foolish to argue that U.S. airpower alone could eliminate Iraqi ground forces and enable the opposition army to march on Baghdad virtually unopposed.

The plan, moreover, would still require the opposition to capture Iraq's cities from the regime's armed forces. Opposition troops would have to fight these battles largely on their own, because it would be difficult for U.S. planes and missiles to destroy Iraqi ground units hiding in the concrete jungles of modern cities and because U.S. rules of engagement would prohibit bombing urban areas to limit civilian casualties. Even battered Republican Guard divisions would be too much for the weak opposition forces to handle in urban combat. Saddam's foes would probably lose on the ground even if the United States were to devote a massive air campaign to aiding them.

Finally, the airpower approach might lead Saddam to unleash his chemical and biological weapons-his jealously guarded ace in the hole-against the United States and its Middle Eastern allies. Saddam was never close enough to being toppled during the Gulf War to use these weapons of last resort. Slowly pushing him to the wall would give him every reason to lash out with his entire arsenal, while removing all incentives for restraint.

This plan, in short, involves fighting a preventive war against Iraq from a standing start but using opposition troops instead of U.S. ground forces. Its costs and risks would make this an extremely grave and unpopular endeavor, even if it were sold as a way to settle the problem of Saddam once and for all. What backing there was would dwindle further when it became clear that the plan's supposed decisiveness-its greatest selling point-was questionable. The opposition might well lose, leaving the administration with the agonizing choice of either leaving Saddam to crow over his escape, continuing the heretofore futile air campaign until the rubble bounced, or sending U.S. troops to occupy Baghdad. Attempting rollback by relying on airpower, then, would be incredibly risky and more likely to box in Washington than liberate Baghdad.

The Enclave Approach


The second rollback option would have the United States help the opposition umbrella group known as the Iraqi National Congress (INC) seize and hold large enclaves inside Iraq, from which it would proceed to undermine Saddam's rule. This plan, however, is militarily ludicrous and would almost certainly end in either direct American intervention or a massive bloodbath.

The enclave approach calls for the United States to declare the INC the legitimate government of Iraq, provide it with the Iraqi regime's frozen assets, and help it recruit, train, and equip a highly mobile light infantry force of 10,000 to 20,000 troops, most of whom would be former Iraqi military personnel. These forces would then invade and occupy Iraqi Kurdistan, in the country's north, and all of Iraq west of the Euphrates River and south of the 32nd parallel, which would collectively be declared "Free Iraq." The United States would lift economic sanctions on these zones and use airpower to help defend them from Iraqi counterattacks. By providing a haven for defectors and reducing Saddam's area of control to the Mesopotamian river valleys, Baghdad, and some of the northwest, the INC argues, this strategy would either make the Iraqi regime disintegrate or leave it vulnerable to an independent INC military campaign.

The INC has suggested that the entire operation would cost little, take only a year or two, and require such scant direct U.S. military involvement that there would be almost no risk of substantial American casualties. As the INC's president, Ahmed Chalabi, told a Senate hearing last spring, "Give the Iraqi National Congress a base protected from Saddam's tanks, give us the temporary support we need to feed and house and care for the liberated population, and we will give you a free Iraq, an Iraq free of weapons of mass destruction, and a free-market Iraq. Best of all, the INC will do all this for free." Since the plan promises a huge payoff at a tiny cost, it has unsurprisingly become the darling of Congress and Clinton's critics. It could well become popular with the American public too.

The catch is that the INC plan is so flawed and unrealistic that it would lead inexorably to a replay of the Bay of Pigs. U.S. officials would ultimately face the choice of intervening directly or watching the rebels get butchered-as happened with the INC's last attempt to spark an insurrection, in northern Iraq from 1992 to 1996.

As the Middle East Institute's Andrew Parasiliti has observed, "the INC enjoys more support along the Potomac than the Euphrates." Ninety-nine percent of the opposition's combat power comes from two Kurdish militias operating in the north, Massud Barzani's Democratic Party of Kurdistan (DPK) and Jalal Talabani's Patriotic Union of Kurdistan (PUK), and one Shiite militia operating in the south, the Supreme Assembly of the Islamic Revolution in Iraq (SAIRI). Each group has its own agenda, and all have distanced themselves from the INC. The leaders of all three groups now say they are willing to coordinate with the INC but will not follow its lead. Worse, they aver that the INC no longer does much for the opposition cause. Last September Barzani and Talabani accepted an invitation to meet with congressional leaders but canceled when they learned that Chalabi had also been invited.

Internal opposition squabbling aside, America's regional allies hate the plan. The Kuwaitis, the Saudis, the Jordanians, and the Turks adamantly oppose all rollback efforts that do not include the direct use of U.S. military power to ensure victory, fearing that halfhearted efforts will fail and leave them facing Saddam's wrath.

Another problem is that this plan is not just militarily infeasible but downright dangerous. Even if one believes that Kurdish forces could hold the north, the INC approach still would require 10,000 to 20,000 lightly armed opposition troops to control roughly 1,000 kilometers of front inside southern and western Iraq. This is an absurd force-to-space ratio. (U.S. army doctrine, by comparison, calls for a 20,000-man armored division to defend no more than 15 to 35 kilometers of front.) Assuming that the entire regular Iraqi army remains passive-which has not been the case over the last eight years-opposition troops would still face attacks from Saddam's 70,000-man Republican Guard and 25,000-man Special Republican Guard. They would have to hold vast stretches of territory without armor and at a five-to-one numerical disadvantage.

Moreover, a strategy of defending enclaves inside Iraq cedes the initiative to Saddam once the opposition forces reach the Euphrates and the 32nd parallel, letting him choose the time and place of subsequent battles. Even U.S. air support could not protect the opposition troops because Iraqi attacks could come during bad weather or simply at night, when the U.S. ability to interdict ground troops is greatly reduced. The 1996 Iraqi attack on the opposition-held northern city of Irbil is instructive here, precisely because the Iraqis were terrified that the United States would try to defend the city from the air. They therefore built up their forces carefully, attacked at night, smashed the INC and Kurdish forces before daylight, and withdrew the next day, presenting Washington with a fait accompli. The United States simply does not have enough all-weather strike aircraft to keep enough warplanes in the Persian Gulf to respond instantaneously to regime attacks on opposition units inside Iraq for the months and perhaps years that it would take to topple Saddam.

Southern Iraq is also particularly ill suited to the kind of cascading insurrection the INC plan envisages. The south's open desert terrain would be ideal for Iraqi armored operations, and the area southwest of the Euphrates is extremely sparsely populated. New opposition recruits would have to come from Basra, Iraq's second-largest city, which Saddam would be certain to defend furiously. The regime currently has three divisions with 30,000 men around the city and another three divisions within a few days' march. Assuming Saddam could muster 10,000 to 15,000 loyal troops in Basra's defense and remembering that U.S. air support would be of limited help in urban operations, Basra poses an insurmountable obstacle: the tiny INC army could not survive without taking it but would almost certainly be destroyed if it tried.

For the enclave approach to have even a remote chance of success, one of two assumptions would have to prove correct: that the Iraqi regime and its armed forces would simply fall apart when presented with a serious challenge, or that a few inexperienced opposition light infantry brigades could defeat several Iraqi heavy divisions in open combat. Both are wrong. During the Gulf War, Iraq's armed forces were subjected to 39 days of incessant bombing, leaving them thoroughly demoralized. But during the ground war, key units like the Republican Guard still fought extraordinarily hard. These troops would fight once more against the Iraqi opposition, as they have on every occasion since Desert Storm-and in the INC plan they would not even be subjected to an opening bombardment.

The INC thinks its light infantry brigades can easily defeat heavy armored divisions like Iraq's. If this were true, the world's armies would look quite different. To be sure, highly competent light infantry forces with long years of training, such as the U.S. Marines and Israeli paratroopers, have defeated undistinguished armored formations in open terrain. But these are exceptions. Infantry units generally require large numbers of antitank weapons, artillery, armor, and often airpower to stand up to a tank attack. In the vast majority of battles on open terrain during and since World War II, light infantry formations have been overcome easily by even mediocre armored forces. The Iraqi opposition's military track record to date illustrates the point. At Basra and Kirkuk in 1991 and again at Irbil in 1996, Iraqi army and Republican Guard units easily crushed dug-in Kurdish, Shiite, and INC light infantry. The only opposition success, meanwhile-large Kurdish forces pushing two brigades of the Iraqi 38th Infantry Division out of Irbil in 1995-came against the worst formation in the Iraqi army when that unit was demoralized, depleted, surprised, and without much armor or artillery support.

Still, the INC bravely claims that it could emulate the success of Chad in 1986-87, which smashed Libyan armored forces by waging maneuver warfare with highly mobile light infantry. Even a cursory glance at the Chadian victory, however, shows why the Iraqi opposition could never duplicate it. First, Chad enjoyed a favorable force ratio: Libya never committed more than 13,000 troops to the fighting, whereas Chad could call on an army of 10,000 regulars and 30,000 irregulars. Second, the Chadian army was made up of battle-hardened veterans who had fought together for years and developed a high degree of unit cohesion. Third, Chad's military leadership was superb. Hissein Habre proved a first-rate general, and among his excellent field commanders was Hassan Djamous, a commander so good that Western military experts regularly compared him to Rommel. Fourth, the Chadians were naturals at maneuver warfare, and the new American and French armored cars and trucks they got in the mid-1980s let them revert to traditional tribal combat tactics based on rapid movement and swirling envelopments-perfect analogues to modern maneuver warfare. Fifth, the Libyan army in 1986 was one of the most inept on earth. While hardly first-rate, the Iraqi army is more professional, more experienced, and better trained. Chad's success was based on a combination of all five ingredients; the INC has none of them.

In fact, the INC has no military organization at all. If the past is any guide, it probably could not field a competent force even if given years to train and recruit. The finest soldiers the INC could boast would be Iraqi army defectors, collected in dribs and drabs and then thrown together with political dissidents, tribesmen, intellectuals, and other opponents of Saddam. While operating from Kurdistan in 1992-96, the INC itself (as opposed to its Kurdish allies) never had more than a few hundred fighters, and few of these served for more than two or three years. Moreover, the INC leadership has shown little flair for field command, identifying competent leaders, or honing their command skills. INC forces would also have to excel at maneuver warfare to compensate for Baghdad's superior firepower. The INC, however, intends to man its ranks mostly with defectors from the Iraqi military, which is incompetent at such fighting. Such defectors are unlikely to suddenly discover a capability for maneuver warfare that they heretofore utterly lacked. Worst of all, many of these turncoats would be Saddam's spies, who would help scuttle this bid to oust him just as informants have destroyed every past opposition effort to do so. Clearly, relying on the INC's battlefield prowess would be the mother of all mistakes.

The Afghan Approach


The third rollback option-the Afghan approach-would have the United States help the opposition mount an insurgency on the model of the Afghan mujahideen, the Nicaraguan Contras, or the Vietcong. The problem is that one key to successful guerrilla war is a friendly neighboring country willing to provide the insurgents with a secure base, and none of Iraq's neighbors are viable candidates.

The Afghan approach calls for the United States to help the opposition establish a safe haven next to the theater of operations and then recruit, train, and equip a force that can conduct a guerrilla insurgency against Saddam's regime. Guerrilla warfare is the strategy of the weak, who can win only by slowlywearing down their opponents' strength while slowly building up their own. Following Mao's classic three-step approach-start on the defensive, force a stalemate, then go on the offensive-the opposition would start small and avoid conventional battles with the Iraqi army until the rebels had a good chance of winning. After many years the insurgents would topple Saddam's regime and gain control of the country.

This plan, which requires only a modest commitment of outside resources and no risk of substantial American casualties, would certainly be acceptable to the United States, where it could be sold to Congress as a 1990s version of the Reagan Doctrine. Moreover, Mao and others have demonstrated that such a strategy can indeed succeed, although it usually takes more favorable conditions. (Most of Iraq lacks the combination of rough terrain and a stable population base that would enable guerrillas to hide from government forces while living off the land.)

The catch is that the Afghan plan is unacceptable to U.S. regional allies whose cooperation would be needed for it to work. Successful guerrilla insurgencies require a secure rear area where the rebels can train new recruits, shelter dependents, care for their wounded, stockpile weapons and supplies, prepare for new operations, and retreat and regroup in the face of government offensives. Pakistan played this role for the Afghan mujahideen, Honduras did so for the Nicaraguan Contras, South Africa for UNITA, China and later North Vietnam for the Vietcong, and so on. But no country on Iraq's borders will host the opposition today. None of them is prepared to live in a state of open warfare with Iraq for as long as the rebellion takes to succeed.

Most of the candidates for the job can be dealt with quickly. Saudi officials have repeatedly stated that they are not willing to go to war with Iraq indefinitely. Besides, the Saudi territory bordering Iraq is sparsely populated desert-awful terrain for mounting an insurgency. Kuwait, while eager to see Saddam go, is also chary about a long-term state of belligerency-and also has uninhabited desert on its Iraqi border. Jordan, which supported past measures to overthrow Saddam, has consistently balked at allowing direct military moves against Iraq from its territory, and it too has a vast unpopulated desert on its Iraqi frontier. Syria was always only a half-hearted member of the anti-Saddam coalition and has never been willing to participate in direct military operations against Iraq, even during the Gulf War-and it too has mostly uninhabited desert on its Iraqi border.

That leaves only two plausible candidates, Turkey and Iran. Turkey is certainly geographically suitable; the Turkish-Iraqi border area consists of forested mountains and is well populated. Unfortunately, the Turks have removed themselves from the running because they know that any anti-Saddam insurgency based in Turkey and northern Iraq would center around the Kurds. Ankara worries that supporting such an insurrection would fuel Kurdish political aspirations in general and embolden its own Kurdish rebels. Despite repeated pleas from U.S. officials, the Turks prefer to keep Saddam in power and have even favored his efforts to reassert control over Iraqi Kurdistan.

Iran, the final possibility, is the ideal staging ground for an Iraqi insurgency. Indeed, it has sponsored both Kurdish and Shiite guerrillas over the last four decades. Iran's border with Iraq runs through abundant mountains, hills, forests, and swamps; many communities actually straddle the frontier. Yet since the United States and Iran are barely on speaking terms these days, getting them to cooperate in sponsoring an anti-Saddam insurgency would pose a formidable obstacle.

Other problems with the Afghan approach stem from the lengthy time frame it would take to work. First, a long, drawn-out guerrilla war in Iraq would eventually stick in the craw of world opinion. The United States would find itself increasingly isolated at the United Nations and elsewhere, which would affect other foreign policy issues-including maintaining the sanctions and inspection regimes that keep Saddam relatively weak and contained. Second, this plan might eventually get rid of Saddam, but it says nothing about how to manage the challenge he poses in the present and foreseeable future. By embracing the Afghan approach, the United States could easily find itself stuck in a couple of years with a rebellion still in its larval stage and a resuscitated Iraq increasingly free from international restrictions.

Finally, even if the Afghan approach succeeded, it might open up a Pandora's box of new troubles. Destabilizing Saddam is not the same as installing a pro-U.S. government in Baghdad. Instability in Iraq could eventually lead to a military or security-service coup, but this would not necessarily help the INC or any other pro-Western group gain power. Indeed, Iraq might even collapse into civil war. The result of an Afghan-like insurgency, in other words, could be an Afghan-like descent into chaos.

Contain Your Enthusiasm


Since the United States can neither engineer Saddam's fall nor accept him back into the international community, it really has only one option left-the much-maligned existing policy of containment. Still, the Clinton administration's critics are right about one thing: the current containment regime is falling apart. The erosion of U.N. sanctions and the gradual end to Iraq's isolation have allowed Saddam to rebuild his power at home. His refusal to cooperate with the arms inspectors of the U.N. Special Commission (UNSCOM) has let him maintain some of his weapons of mass destruction (WMD) programs. If Saddam's isolation diminishes, he will build up his other military forces as well. The real and pressing challenge the United States faces in Iraq is therefore not whether to abandon containment for a better alternative-there is none-but rather how to shore it up.

This could be done in two ways, each carrying a stiff price. The first approach would be to preserve the containment of Iraq in its current broad form, with the full panoply of sanctions, inspections, reprisals, and diplomatic isolation. This would require the United States to reinvigorate the anti-Saddam coalition, purchasing as necessary renewed cooperation from the international community-including such headaches as China, France, and Russia-through concessions on other foreign policy issues such as NATO expansion, Kosovo, Iran, economic aid, U.N. dues, and so forth. The United States would also have to regularly threaten or use force to respond to Saddam's provocations, taking whatever military steps necessary to coerce him back into compliance. The Clinton administration, in short, would have to push Iraq to the top of its foreign policy agenda and keep it there for the foreseeable future, spending substantial diplomatic capital to keep Saddam in his box.

The second approach would be to shift from broad to narrow containment. In this scenario, the United States would agree to lift the economic sanctions on Iraq, along with the "no-fly" zones and the international travel ban, in return for a new U.N. Security Council resolution prohibiting Iraq from reconstituting its WMD programs or acquiring significant new conventional weapons such as tanks, artillery, combat aircraft, and attack helicopters. To make sure these restrictions were enforced, the resolution would provide for renewed UNSCOM inspections and comprehensive monitoring of Iraqi imports, together with a clear declaration that any U.N. member state would be permitted to use all necessary means to compel compliance. The United States, in short, would end several of the more unpopular elements of the current containment regime in return for blanket authorization for future strikes against Saddam designed to halt his WMD programs or prevent a dangerous conventional military buildup.

Whichever approach one favors, support for the Iraqi opposition can and should be one component of American policy-but only as long as such support is clearly understood to supplement containment, not supplant it. Giving the opposition limited military, political, and financial aid would help make it a thorn in Saddam's side while avoiding the problems of the rollback policies described above. Such an approach would be politically, diplomatically, and militarily feasible and would accomplish a range of U.S. objectives.

The key is for Washington to avoid delusions of grandeur while strengthening Saddam's opponents. The Iraq Liberation Act is misguided less because it authorizes some military aid to the opposition than because it encourages unrealistic expectations and excessively limits which groups can be supported. The first move should be to bolster existing opposition militias such as those of the Kurds, who have fought the Baath regime since it took power in 1968. In recent months, the PUK and the DPK have successfully concluded an uneasy peace with one another, maintaining separate spheres of influence in northern Iraq and concentrating on resisting Saddam rather than killing each other. Despite past betrayals, the Kurds remain willing to work with the United States. The White House should reciprocate by giving the Kurds a steady supply of small arms, funds, and other resources needed to regain their military strength.

The United States should also aid Shiite groups in southern Iraq. Washington has generally avoided direct ties to even the most influential Shiite groups, such as SAIRI, because of their fundamentalism and strong ties to Iran. The United States should certainly try to strengthen, arm, and fund more moderate Shiite groups whenever possible. But it should even consider supporting SAIRI-because such help might induce moderation but primarily because effective enemies of Saddam are too few to spurn.

The United States should also try to foster a more broadly based domestic opposition to Saddam. Here the INC's propaganda is correct: only an umbrella organization can unite the disparate elements of the anti-Saddam movement, especially if the United States provides aid to the Kurds and Shiites independently. Washington must convince both skeptical regional allies and Iraq's Sunni majority that it still seeks a unified Iraq rather than one split into communal-sized pieces and opposes Kurdish or Shiite bids for hegemony over the Sunnis.

If Washington's chosen umbrella organization is to remain the INC, the group must be reinvigorated so that it can increase its influence in Iraq and attract more soldiers to its banner. The United States should help by sponsoring propaganda efforts on its behalf (such as the recently created Radio Free Iraq) and by providing funds and perhaps eventually arms and military training. As important, however, should be U.S. pressure on the INC to remove ineffectual leaders, expand and diversify its membership, and court Iraq's other chief opposition groups.

Such a plan, to be sure, would be unlikely to overthrow Saddam but would help keep him off balance and boxed in even if sanctions and other current components of containment continue to erode. It would foster disgruntlement among the Iraqi elite and possibly even increase the chances of Saddam's being overthrown from within. (Saddam's failure to suppress the opposition in the early 1990s led to several coup attempts and undermined his power base.) Most important, it would help curb Saddam's expansionism. Having an opposition that could legitimately claim to speak for all Iraq would both offer hope to the Iraqi people and symbolically challenge the current regime, forcing Saddam to concentrate on maintaining his control over Iraq rather than risking aggression abroad.

Providing some support to the opposition would also help alleviate the horrendous suffering of Iraq's Kurds and Shiites. Washington has repeatedly encouraged the Kurds to revolt but then abandoned them to Baghdad's mercies: selling them out in 1975, standing by while Saddam slaughtered them in the nightmarish 1987-88 Anfal campaign, sitting on the sidelines after Desert Storm in 1991, and conducting only token strikes after Saddam's encroachment into northern Iraq in 6. The Shiites also suffer. Because of their continued defiance, Saddam has drained the south's marshes and savagely repressed the Shiites. Giving some aid to opposition forces without egging them on to quixotic adventures would demonstrate a U.S. commitment to their welfare and afford them the means to increase their resistance.

Keeping the opposition in business would also give the United States greater flexibility down the road. If broad containment ends, what will follow it? Turkey could either work more closely with Iraq or stay true to Washington. Iran might become a de facto U.S. partner or collaborate with Saddam against Western hegemony. Keeping the opposition viable would leave options open for the United States should conditions somehow change-after, say, an internal revolt in Basra or another major city-and make Saddam's long-term ouster more plausible.

Some U.S. allies will resist any significant effort to work with the opposition, but this can be overcome. Turkey rightly fears that moral and material support to Iraq's Kurds would spread to Kurdish insurgents in Turkey such as the Kurdish Workers Party (PKK). To reassure Ankara, Washington should let it play a major role in the policy's implementation and press the DPK and PUK to emphasize repeatedly that their goal is greater autonomy within Iraq and not Kurdish independence. It should also remind Turkey that the DPK has in fact helped Turkey suppress the PKK in the past and would do so again if supported. If the Turks still balk, the administration should make concessions on other foreign policy issues to get them on board.

The Gulf emirates will also be suspicious, but they like other potential U.S. policies even less. Assurances that Washington does not favor replacing Saddam with a Kurdish or Shiite regime would be both necessary and desirable, since neither option is a feasible way to govern Iraq. The Gulf states would also be comforted by reassurances that U.S. support for the opposition was intended as part of containment, a policy they have generally favored, rather than some new high-risk venture.

This policy should also be acceptable in the United States because it would advance American interests while limiting American involvement. In contrast to the rollback fantasies, using the opposition as part of a reinvigorated containment policy would not require large sums of money or additional military forces. The greatest danger, in fact, would be muting domestic support to ensure that it did not swell into the full-throated but impractical rhetoric that has accompanied the idea to date. Supplementing containment by giving the Iraqi opposition enough help to stay in business and harass Saddam's regime will not grab many headlines. Limited containment will not necessarily get the United States what it wants. But it might help it get what it needs.