The Use and Abuse of Military Force

As the first decade of the post-Cold War period draws to a close, one thing is certain: military intervention remains a central feature of American foreign policy. In addition to the large-scale intervention that successfully liberated Kuwait in the 1990-91 Persian Gulf War, the United States invaded Panama to protect U.S. citizens and the Canal, oust Panama’s leader, and seat the elected government; entered Somalia, initially to feed its people and then to shape its politics; occupied a Haiti that was hemorrhaging people and ignoring the political wishes of its citizens; bombed Bosnia’s Serbs both to weaken them and to induce them to sign a peace accord; kept the peace in Bosnia in the aftermath of the Dayton peace accords; dispatched air and naval forces to the Taiwan Straits in order to signal China of the U.S. commitment to Taiwan; attacked an Afghan terrorist camp and an alleged pharmaceutical facility in Sudan to retaliate against terrorist attacks and to discourage new ones; bombed Iraq to encourage its compliance with international stipulations and to punish it for ignoring the same; went to war with Serbia over Kosovo; and provided support personnel to a multinational force sent to East Timor.


This accounting makes clear that military force was either used or contemplated by the United States in myriad settings and for various aims. Actual or would-be interventions occurred in nearly every region of the world, and included punitive raids, coercive or compellent attacks, deterrence, preventive and preemptive strikes, peacekeeping, peacemaking, and nation-building. The size, cost, and impact of these interventions ranged no less widely. The intention of this policy brief is to summarize the lessons that should have been learned from these experiences.

The Limits of Air Power

Air power can accomplish many things, but not everything. Six weeks of intense bombing of Iraq and Iraqi forces could not liberate Kuwait during the 1990-91 Persian Gulf conflict; it also took 100 hours of ground warfare. Air power can prepare a battlefield, but it cannot control it. Kosovo underscored a related limitation; although aerial bombardment over the course of some eleven weeks did help persuade Slobodan Milosevic to agree to NATO terms, it seems apparent that the threat of introducing ground forces made a greater impact. Regardless, terrible things happened to civilians on the ground when only air power was employed: thousands of innocent people lost their lives and hundreds of thousands lost their homes and became internally displaced or refugees. In the end, Kosovo was not one but two wars: an air war dominated by NATO and a ground war dominated by Serbian military and paramilitary forces.

Air power might have accomplished more in Kosovo had NATO and the Clinton administration observed some of the traditional guidelines for the effective use of military force. Delay can make any intervention more complicated and costly. Hesitation is understandable when only humanitarian concerns are at issue, as it is much harder to marshal domestic and international support in the absence of an overwhelming cause. But delay also exacts a price by squandering the opportunity to act preventively and with less force. In Kosovo, using air power earlier, after the wholesale violation of the October 1998 agreement—in which Serbia agreed to a cease-fire, limits on its forces in Kosovo, and international monitoring—certainly would have proven more efficient.

Decisiveness is almost always preferable to gradualism. Here again, Kosovo was no exception. Starting at a relatively modest pace diluted the psychological and political impact of the NATO action; it also gave Serbia an almost free hand to pursue its objectives using ground forces. The measures to avoid collateral damage insisted on by European members of NATO inhibited operations and limited their usefulness. All things being equal, it is better to err on the side of too much rather than too little force. This is true of air campaigns—the contrast between the Kosovo intervention and the opening days of Operation Desert Storm could hardly be greater—as well as with ground or combined efforts. U.S. capabilities in Somalia were never increased in step with the mission’s widening in early 1993, while the initial attempt to use force in Haiti—the decision in October 1993 to dispatch only 200 U.S. and Canadian soldiers, followed by the decision to withdraw them when mobs rioted on the shore—ended in humiliation for the United States.

Some of the problems encountered in Kosovo were tied directly to the decision to use air power alone rather than in conjunction with ground forces. It is true that American interests in Kosovo were less than vital, and that persuading the American people and their elected representatives of the need to make large sacrifices, including casualties, would not have been easy. But the Clinton administration never tried. The result is that the air-only intervention failed to achieve one of the principal goals the United States and NATO had set for themselves: guarding the people of Kosovo. Force protection to avoid casualties can and should be a consideration—but not the only one.

Still, it is important to draw a distinction between what proved in the end to be the successful application of coercive force in Kosovo and the failure of it in Iraq in 1998. In the former, significant air power was deployed in an open-ended fashion. The condition for stopping attacks was clear: Milosevic had to meet a specified set of demands. NATO, meanwhile, had to be prepared to stay the course until he met them. Such coercion thus remains a risky form of intervention in that it cedes the initiative to the target, which has to decide whether to hold out or to compromise. But coercion does make for clarity of purpose, because it links intervention with a specific goal.

Punitive interventions are in many ways the opposite: they lack any clear purpose or linkage, and their principal advantage is that the attacking side retains the initiative in that only it decides when it is satisfied. Operation Desert Fox against Iraq in 1998 employed a modest amount of air power for a short and arbitrary period of time with no goal other than to weaken the adversary’s strength to some unspecified degree. Such an action raises serious questions about the use of punitive attacks; in that instance, it would have been far better to have conducted a compellent attack that was not only open-ended and massive in scope but tied to Iraq’s agreeing to accept unconditional international inspections of Iraqi facilities suspected of producing or storing weapons of mass destruction. To be sure, this would have been costlier to carry out and would have provoked significant international opposition. But it also would have had the potential to achieve the important goal of reinstating inspections while humiliating Saddam Hussein in the process, two outcomes that would have justified the diplomatic costs.

Beyond Air Power

Deterrence can work on occasion. Dispatching a carrier task force to the Taiwan Straits in 1996 was a classic example of effective gunboat diplomacy: a long-term strategic commitment was backed up by a sizable show of force. Whether it actually deterred any action by China is less clear, as it is difficult to discern China’s intentions. Regardless, the U.S. deployment may have had an impact on the country’s subsequent behavior by letting China know that any military move against Taiwan would likely be contested by the United States. By contrast, U.S. threats against Serbia over Kosovo failed, suggesting that deterrence requires credibility—which was markedly absent in the latter case, given the history of threats that were not backed up by action.

Carrying out more than one kind of intervention in the same place at the same time can invite trouble. This proved true both in Bosnia, where the presence of a lightly armed United Nations protection force made the use of air power risky, and again in Kosovo, where the presence of unarmed monitors worked to undermine the credibility of the threats to attack. The problem goes beyond the danger of hostage-taking, which is all too real. In addition, it is difficult, if not impossible, to carry out an operation that requires consent at the same time one is threatening (or actually carrying out) a compellent or otherwise hostile operation. The former effort must be ended before the latter can be effective.

New technology is no panacea. To begin with, no system is invulnerable. Even a stealth aircraft can be shot down. Smart bombs go astray from time to time, missing their intended targets and occasionally causing unintended damage and destruction. Bombs and missiles can be fooled by decoys and frustrated by mobility and masking. Efforts during the Persian Gulf War to destroy mobile Iraqi SCUD missiles, for example, were an expensive failure, and attempts to intercept relatively primitive SCUDs with more modern antimissile systems likewise proved unsuccessful for the most part. Of course, no system is better than the intelligence fed into it; accuracy is no virtue if the target is misidentified, as was the case with the Chinese embassy in Belgrade, or if the intelligence assessment is in itself flawed, as may have been the case in the U.S. attack on an alleged chemical plant in Sudan. The latter instance, along with the raid on suspected terrorist installations in Afghanistan, underscores the difficulty of carrying out successful preventive and preemptive interventions when critical, time-sensitive information is difficult to obtain.

Several recent experiences highlight just how hard it is to affect a target nation’s internal politics or political culture with military instruments. Both Iraq and Kosovo suggest that short of occupation, military force is not a very good tool for changing regimes, although a successful use of military force that weakens or humiliates an adversary can help bring about a political environment in which domestic opponents of the regime in question may be encouraged to act. Panama, of course, proved a partial exception to these rules, but even there it took some two weeks to find and arrest Manuel Noriega even though more than 20,000 American soldiers were occupying the general’s small country and faced little organized opposition. Moreover, as was demonstrated in Haiti, even prolonged occupation is no guarantee of desirable results, although the prospects for success clearly are affected by how other foreign policy tools (such as economic assistance) are put to use. Again, military force is good at creating contexts, but what happens within those contexts is more a matter for diplomats and policymakers.

Exit dates ought not to be confused with exit strategies. America’s experience in Bosnia revealed that arbitrary deadlines for getting out are more likely to cause political problems than provide solutions. The notion of an intervention providing a fixed amount of breathing room, after which the local people and governments will be on their own, is absurd; the United States will not be able to turn its back on a humanitarian problem if it gets bad enough or if U.S. strategic interests are adversely affected. This lesson seems to have been learned, for the Clinton administration eschewed any talk of specific withdrawal dates when it deployed peacekeeping forces in Kosovo. Moreover, the lesson of several such interventions, including those in Haiti and Bosnia, seems to be that prolonged commitments do not stir controversy at home so long as casualties are kept to a minimum, which is more likely to be the case when the purpose of an intervention is defined modestly.

What also emerges from recent history is a new appreciation of the impact of decisions not to use force. The clearest example is Rwanda, where nonintervention against horrific violence almost surely forfeited an opportunity to accomplish much good with limited costs. Another example is the decision not to dispatch a force to East Timor when order broke down there in the wake of a vote favoring independence from Indonesia. (In the end, the United States contributed several hundred intelligence, logistics, and communications specialists, but only after the UN authorized and the government of Indonesia invited in an Australian-led, multinational peacekeeping force.) In the case of Iraq, decisions not to use force—or to impose sharp limits on how much force would be used—have given Saddam the time and opportunity to defy UN demands and reconstitute Iraq’s weapons of mass destruction. In Kosovo, meanwhile, the refusal to commit ground forces likely increased the vulnerability of the local populace.

Consequences for Defense and Foreign Policy

Overall, what is most noteworthy about America’s principal military interventions over the past decade is their number and variety. Military force continues to be relevant to a wide range of tasks, which indicates a continuing need for a large and flexible U.S. military. Other governments and organizations are in a position to provide only limited assistance except where armed resistance is modest or non-existent; the lion’s share of any demanding military operation will have to be borne by the United States. That said, America does not have the luxury of maintaining a military tailored only to traditional battlefields, or even to one kind of traditional battlefield. The dismal showing of the Apache helicopters in Kosovo—the difficulties in getting them there and up and running—sounds a warning that the U.S. military needs greater flexibility, which means a force that is easily moved and capable of coping with a wide range of missions in a wide range of environments. Right now, U.S. forces—most notably the Army’s—are too heavy and organized into units that are often too large, too short on sea- and airlift that can use most of the world’s ports and airfields, too light on stores of advanced munitions, and too lacking in the specialized command, control, and intelligence assets and platforms needed to conduct modern combat operations.

Still, even an ideal military cannot succeed if it is undermined by either of two constraints. The first is an unwillingness to allow the military to do its job even when U.S. interests warrant it and the military tool is judged the most appropriate. Not acting entails real costs for the interests at stake or, in the case of humanitarian emergencies, for the innocent people who lose their homes or lives—and for America’s image in the world. Moreover, a narrow foreign policy based solely on self-interest is unlikely to capture the imagination or enjoy the support of the American people, who want international commitments with a moral component.

A different form of reluctance to commit is that involving ground troops. As has been shown, air power is clearly limited in what it can be expected to accomplish. In some instances, such as where high-value targets are sparse or the adversary is exploiting its advantage on the ground, only ground forces will be able to protect the interests involved. Domestic opposition to such a commitment can be reduced and overcome by concerted presidential effort and by designing interventions that justify an American casualty level by the interests at stake. This requires time and political capital, but it is time and capital well spent. Interventions shaped more by politics than by strategy are unlikely to succeed.

The other constraint is in some ways the opposite: over-reliance on the military instrument. Humanitarian crises constitute but one demand on U.S. military capabilities; the United States needs to be discriminating in its interventions, lest it exhaust itself and leave the nation unable to cope with those scenarios where its vital interests are at stake. When it comes to humanitarian situations, several factors should influence the decision to intervene. The first is the scale of the problem: not every repression is a genocide. A second consideration is the likely costs and consequences of acting, both for the immediate problem and for broader U.S. strategic and economic interests. Third is the matter of partners and the extent of military and financial help the United States can expect from others. A final consideration is the likely results of other policies, including but not limited to that of doing nothing.

Such questions are no substitute for situational judgment: there can be no template for intervention. But these questions do provide a framework, and with it some potential guidance. Such considerations would have made America less likely to occupy Haiti or to expand the Somalia intervention into nation-building, but might have encouraged the United States to act earlier in Bosnia and Rwanda, where small interventions could conceivably have prevented genocide. In Kosovo, it would have been wiser to continue diplomacy and deal with a limited humanitarian crisis while looking for ways to weaken or topple the Milosevic regime, or to send in ground forces at the outset and prevent the displacement and killing. In East Timor, forces should have been introduced in tandem with the referendum on its political status. Once this opportunity was missed and violence erupted, the barrier to acting should not have been the opposition of the Indonesian government or the absence of a UN Security Council mandate—hiding behind respect for sovereignty should not be allowed when a government violates the rights of its people in a massive and brutal way, and legitimacy should not be dependent on UN actions. Instead, the reason for delay should have been the absence of partners willing to bear the brunt of the operation, an obstacle that might have been overcome by more forceful diplomacy.

Deciding whether to intervene for what are entirely or mostly humanitarian reasons need not be an all-or-nothing choice. Where interests do not warrant peacemaking or nation-building operations that are costly in both financial and human terms, there may be less demanding options for doing good, such as establishing one or more safe havens in an affected country or designing operations to keep opposing factions apart. The United States would have been wiser to resist the temptation to expand its intervention in Somalia from just delivering food in a large, safe area to full-bore peacemaking. In Kosovo today, the United States should content itself with keeping ethnic Serbs and Albanians safe in their own separate areas rather than undertaking the far more costly and possibly futile effort to recreate an integrated, multiethnic society.

A somewhat restrained approach to humanitarian intervention is unlikely to satisfy either those who wish to put it at the center of American foreign policy or those who wish to push such efforts to the sidelines. But there must be limits on U.S. military action when a situation is not dire, when other foreign policy instruments look promising, when partners are scarce, or when other major powers oppose American intervention. The United States simply cannot commit its military might everywhere all the time. The choice to intervene with military force must be made carefully—and carried out decisively—for American foreign policy to prosper.