The moral poverty of foreign policy is, along with war itself, one of the most enduring and intractable problems in international relations. In recent years, the conventional wisdom on international moral action has been that it

requires a cosmopolitan consensus, a shared understanding of the dignity and equal worth of every human being. Short of achieving a global cosmopolitan majority, the more that Western elites, and especially officials of the world's strongest power, subscribe to this world-view, the easier it should be to achieve effective international action to resolve humanitarian crises, the most important of which are mass killing and genocide.

Samantha Power's disturbing book suggests that the question is more complex. Through a series of careful historical case studies, Power, a former Balkan war correspondent, traces the development of the concept of genocide from the Turkish campaign against the Armenians in 1915 to the present -- along with the repeated failure of the world, and particularly the United States, to prevent such horrors. She argues that this pattern is due not to public or elite indifference to the idea of moral responsibility, nor to a lack of timely warning or feasible intervention options, but rather to structural features of the American political system. Under normal circumstances, she shows, American officials face stronger incentives

to avoid action against genocide than to stop it. Power's book will likely become the standard text on genocide prevention because it thoroughly debunks the usual excuses for past failures, while offering a persuasive framework that can help predict future outcomes and suggest policy responses. It is also engaging and well written; together with the awful fascination of the subject, this should be enough to guarantee that it will be widely read by both students and policymakers.


The crucial puzzle at the heart of Power's book is why the Clinton administration, which entered office more committed to humanitarian intervention and a moral foreign policy than any U.S. administration since World War II, wound up standing by and watching two genocides. Bill Clinton campaigned on the first Bush administration's failure to contain the genocide in Bosnia and staffed his foreign policy bureaucracies with long-standing advocates of "moralpolitik." Yet for more than two years his administration did little about Bosnia until it was embarrassed into stronger action after the Srebrenica massacre of Bosnian Muslims in 1995. And in opposing Serbian ethnic cleansing in Kosovo, U.S. action was fairly prompt but woefully inadequate until British Prime Minister Tony Blair dragged Clinton into committing to defeat the Serbian army on the ground.

The most egregious failure, however, was the Rwandan genocide of 1994. Of all genocides since World War II this would perhaps have been the easiest to stop, and yet the United States not only did not halt the killing but actively prevented other willing powers from taking effective action. After Hutu soldiers murdered ten Belgian peacekeepers on April 7, 1994, the Belgian government made it clear to the United States that it would pull its troops out unless the un presence in Rwanda was reinforced. The United States advised pulling out and also

played a decisive role in persuading the un Security Council to cut the strength of the un Assistance Mission for Rwanda (UNAMIR) from 2,500 to 500 just two weeks after the genocide began. As a result, more than 800,000 people died between April and July 1994. Continuing civil wars in the Democratic Republic of Congo and in Rwanda, partly sparked by the Rwandan genocide, have led to at least another two million deaths.

Power's cases demonstrate, of course, that the record of inaction is not limited to the Clinton administration. Despite abundant warning, the United States took no action during the Armenian genocide or the Holocaust. Although the United States led the drafting of the Convention on Genocide from 1946 to 1948, Washington took 40 years to ratify it -- and then did so with reservations so crippling that it is effectively barred from ever invoking the treaty against anyone. The United States did not act against genocides by the Khmer Rouge against Cambodians in the 1970s or by Iraq against the Kurds in the 1980s. And after 47 years of intermittent ethnic cleansing and genocide in Sudan, the United States still has done nothing there.

Historically, the critical factor in obtaining energetic action in pursuit of morally based foreign policy goals has been not the strength of cosmopolitan feeling, but the balance of domestic political power; Power's findings are broadly consistent with this pattern. She shows that American inaction in the face of genocide has not stemmed from an absence of cosmopolitan values among the public or elites. Depending on how the question is phrased, polls have found majorities as high as 80 percent in favor of at least some intervention to prevent genocide. Nor is feasibility the major hurdle. Although the world almost certainly could not prevent genocide in a strong country -- for instance, China or Russia -- most genocides are perpetrated in very poor countries by very weak regimes. In policy contests over particular interventions it is routinely claimed that the United States lacked enough information to act in time or that genocide prevention was militarily impractical. But these positions often reflect more the interests of those making the arguments than they do genuine, real-world constraints. Rather, Power finds, the central obstacles to humanitarian intervention in general and genocide prevention in particular lie in the organization of the American political system.

The causal chain, though frustrating, is reasonably straightforward. The rest of the world does not act because the United States does not. The United States does not act, in turn, because public support for humanitarian intervention is diffuse and rarely mobilized. Absent clear demonstrations that the public supports intervention, the military is unwilling. Partly because of the military's position, the political right is opposed. Because the political right is opposed, presidents are cautious and believe that pushing for action cannot benefit them politically but can only cost them. Because presidents do not favor humanitarian action, finally, career foreign policy and national security officials learn that strong advocacy of it is both unlikely to succeed and bad for their job prospects. The result is that humanitarian disasters have a hard time even getting on an administration's agenda, let alone generating momentum for action.

The domestic pressures against humanitarian intervention lead to what political scientist Stephen Van Evera calls "non-evaluation": analyses of humanitarian crises are either not conducted at all or are conducted under distorted standards that prevent serious assessments that could provide ammunition for pro-intervention positions. Institutions and individual analysts either learn to play the game or are suppressed. Misapplied "lessons of history" can also play a role. As political scientists Robert Jervis and Yuen Foong Khong have explained, these lessons are often quite simplistically applied: past policies judged successes tend to be repeated, while those judged failures are avoided -- often without serious consideration of differences between the original case and the present problem. Power reports, for example, that every U.S. official she spoke to named the disastrous U.S. intervention in Somalia as the main barrier to action on Rwanda -- but few if any considered whether the emergencies were similar or whether the requirements for intervention would be harder or easier in the latter case.


The Rwandan case is the key to understanding the systematic failure to prevent mass killing. With the emergency so clear and so vast and the effort required so minimal, the United States still did not act, and this failure should serve as a powerful warning to anyone hoping to see a substantial altruistic dimension to U.S. foreign policy any time soon. Power shows that the underlying reason for ducking the issue was the Clinton administration's domestic political calculus. Crucial government decisions were driven less by the situation in Rwanda than by the situation in Washington -- and since the same factors can be expected to operate in the future, the same outcome is likely.

Although some have argued that the failure in Rwanda was due to a lack of timely information, Power shows that senior administration officials quickly understood the magnitude of the emergency but deliberately shut off consideration of virtually any action. In the absence of mobilized domestic pressure, officials sought to hoard their political capital for future crises and avoid any steps that could potentially draw the United States into the conflict. Some officials continued to focus on resurrecting the 1993 Arusha accords between the Hutus and the Tutsis, a project comparable to arranging a truce between the Turks and the Armenians or the Nazis and the Jews. Most important, no high-level meeting of foreign policy principals was ever held, with the result that no one demanded any serious analyses of the crisis. This fact alone virtually assured that no rescue plans could be developed, and mid-level officials who attempted to raise this issue were branded naive or alarmist. The U.S. response to Rwanda, in short, constituted a classic case of non-evaluation.

Because nearly all genocides occur during ongoing wars, non-evaluation exacerbates the fog of war in ways that give credence to the standard excuses for inaction. Apologists for passivity regularly claim that outsiders can never tell until it is too late whether genocide is actually occurring: they cannot know soon enough how many are being killed or how far killers intend to go, and in an ongoing war it takes time to figure out whether the massacres are one-sided or mutual. Therefore nothing can be done because by the time enough is known there are no longer any military options that can be carried out fast enough to matter.

Power shows that such blocking tactics are further assisted by confusion over the concept of genocide. Although the international legal standard requires no particular level of deaths, both the public and policymakers usually take the term "genocide" to apply only to the actual extermination of a very large fraction, often more than half, of the target group. Such a standard makes it virtually impossible to determine that "genocide" is occurring until it is virtually over and there are few survivors who can still be helped. In Rwanda, the U.S. administration carefully avoided using the word "genocide" in order to dampen pressure to act, and even human rights organizations avoided the word for the first two weeks of the killing for fear of being dismissed from mainstream debate. As Power points out, however, this should be merely a semantic issue; large-scale mass killing is a crime against humanity regardless of whether or not most or all of the target community is likely to be wiped out.

Non-evaluation during crises can also contribute to continuing low-quality assessment afterward by providing ready-made models for overly pessimistic judgments but not for critical or countervailing assessments. This is important because a consensus, even an erroneous one, that nothing could have been done in a past crisis can curtail the consideration of options in a future one. Understanding Rwanda is especially important because if little or nothing could have been done about one of the easiest cases that the West has faced in recent decades, there is little point in any genocide prevention effort anywhere.

The most prominent skeptic of aggressive action in Rwanda has been political scientist Alan Kuperman. In his recent book The Limits of Humanitarian Intervention: Genocide in Rwanda (and in a previous article in these pages, "Rwanda in Retrospect," January/February 2000), Kuperman argues that even the strongest imaginable intervention could not have saved more than 125,000 of those who were killed. The critical question is which of Kuperman or Power is right. Power is.

Kuperman bases his claim on three main arguments, all of them dubious or wrong. First, by accepting a low estimate of the Tutsi population of Rwanda and ignoring the simultaneous massacres of Hutu political moderates and members of certain southern Hutu clans, he counts only 500,000 victims of the genocide instead of the generally accepted 800,000. This removes almost 40 percent from the ranks of those who could have been "saved" by any possible intervention.

Second, he claims that the Clinton administration could not have known that the massacres were occurring until it was too late -- not earlier than April 21, 1994, the day after Human Rights Watch publicly labeled the crisis "genocide." But as Power shows, U.S. officials had more than enough information to know in advance that mass killings were imminent, as well as confirmation that they were occurring on a very large scale almost immediately after they began on April 6.

By late 1993, internal U.S. intelligence analyses were predicting that large-scale massacres were likely if war resumed in Rwanda. From December onward, General Romeo Dallaire, the commander of UNAMIR, was inundated with reports

of preparations for mass killings. In January he proposed seizing some of the weapons caches of the interahamwe, the Hutu nationalist militias, but was denied permission by Kofi Annan, who was then undersecretary-general for peacekeeping. On February 23 Dallaire reported that he was "drowning" in information about death-squad target lists. By April 7, the day after the killings began, the U.S. Defense Intelligence Agency (DIA) had intercepted communications from officials in Kigali ordering massacres in outlying districts, and within a day or two more the agency had photographs of mass graves in several parts of the country.

From April 8 onward, news of the killings covered the front pages of

the Western press, with initial reports of tens of thousands dead. Joyce Leader, the U.S. deputy chief of mission in Kigali, and other evacuated embassy officials reported that genocide was taking place. A small U.S. Marine Corps reconnaissance mission also visited Kigali around this time and reported on the scale of the slaughter. On April 10, Dallaire formally requested that UNAMIR's strength be doubled to 5,000. That policymakers did not pay attention to the DIA, Dallaire, Leader, and the other witnesses reflects not a lack of credible information, but rather an unwillingness to get involved.


Kuperman claims, finally, that even if the United States had acted promptly, it could not have gotten to Rwanda fast enough. Power shows, however, that substantial forces were indeed immediately available. In addition to Dallaire's 2,500 peacekeepers, 800 more Belgian troops were available in Nairobi and 300 U.S. marines were stationed in Burundi. By April 9-10 more than a thousand additional French, Belgian, and Italian troops had arrived in Kigali -- to evacuate foreign nationals, though, not to stop the massacres. It is difficult to overstate just how powerful these 4,600 troops could have been in the Rwandan context, compared to the lightly armed, badly led, and low-morale Rwandan army (FAR), let alone the interahamwe. In fact, even after the murders of the ten Belgian paratroopers, Hutu forces remained profoundly reluctant to challenge foreign units. With just the 500 troops that the Security Council eventually left to it, UNAMIR was still able to protect roughly 11,000 potential victims throughout the 100 days of the genocide.

If 125,000 is too low, what is a more reasonable estimate of how many of the victims of the Rwandan genocide could have been saved? Suppose the United States had chosen to encourage, rather than effectively prohibit, prompt military action against the genocidaires. It could have pursued this outcome in a number of ways: by encouraging the Belgians to reinforce rather than retreat; by promptly dispatching the ready brigade of the U.S. Army's 82nd Airborne Division to Kigali and ac-130 gunships to Kampala and Nairobi; by jamming Radio Milles Collines, the radio station used to incite and coordinate the killings; by giving Dallaire command over all the troops immediately at hand and -- most important -- authorizing him to shoot


The government of Rwanda and its army would have collapsed instantly, and the killings in Kigali and its immediate neighborhood would have stopped. Elsewhere the genocidaires could conceivably have tried to accelerate the genocide before intervention forces arrived, but in fact they were already killing as fast they could. It seems far more likely that the pace and scale of killing outside Kigali would have been sharply reduced, for several reasons. First, few if any further genocidal instructions or roving death squads would have gone out from Kigali. Second, the death squads could not have so easily intimidated or killed local leaders who were reluctant to cooperate with them, especially in the southern areas.

Third, the advance of the rebel Tutsi army, the Rwandan Patriotic Front (RPF), would have been greatly accelerated. This is a crucial point, since the genocide was ultimately stopped not by any intervention but because the far was decisively defeated by the smaller and more lightly armed RPF. Indeed, the threat of an RPF advance was enough to scare the Rwandan government into fleeing Kigali within six days of the start of the genocide. Had the United States encouraged action, the organizers of the genocide would have felt far stronger pressure far sooner to abandon the project and flee, lest they be caught by Western forces or by the RPF. If a response had begun on April 9 or 10, anywhere from 300,000 to 700,000 of those who actually died could have survived.

Intervention on behalf of the RPF would not have led to an increase in massacres of Hutus by Tutsis. If ever there was a case of pure black hats versus light gray hats, this was it. (Kuperman's one inexcusable contribution is to repeat the claim that "it is still unclear today" whether reports of mutual massacres were inaccurate.) Stopping the genocide sooner would also have eased Rwanda's transition to peace. A quicker end to the conflict would have reduced the number of accused Hutu war criminals who wound up in Tutsi-run jails and also dampened Tutsi motives for revenge killings. It also would have reduced the number of Hutu refugees who fled the country. This in turn might have cut the ability of the far troops who escaped to Zaire to mount large-scale guerrilla operations during 1995 and 1996, operations that eventually led to the RPF intervention in that country's civil war.


Some Americans do not accept that the United States, or any state, bears any responsibility to stop the murder of foreigners, no matter how large the scale of the killings and no matter how easily preventable they are. Others -- a majority, in Power's view -- think that either the ideals of human dignity on which the United States was founded or simple natural law does require a willingness to act against mass killing. A concern for the United States' international moral authority also requires some correspondence between the ideals that it espouses and those it acts on -- especially at a time when it is proposing to instruct other countries on which techniques of political violence are legitimate and which are illegitimate. U.S. performance in preventing genocide must improve, or its moral authority will prove hollow.

Power's book and the broader history of international moral action yield three main lessons for improving the prospects for genocide prevention in the future. First, public pressure is unreliable. A cosmopolitan majority willing to pay significant costs for international moral action is unlikely to emerge in the foreseeable future. Mobilization by specific ethnic groups is also not assured; as Power points out, refugees from ethnic cleansing and genocide often generate new ethnic lobbies in Western countries, but these lobbies don't usually exist before the genocide.

Second, the prospects for action are best when a nationally prominent politician takes up the cause, threatening the government with a price for inaction. Power points out that in late July 1995, two weeks after the Srebrenica massacre, Senate Majority Leader Bob Dole pushed a bill through Congress that lifted the arms embargo on the Bosnian Muslims. Dole's leadership on the issue embarrassed Clinton and effectively compelled him to take a series of more forceful actions that eventually ended the war. Dole had been to Bosnia, had seen the scale of the devastation himself, and had become personally committed to the issue. But political star power is also unreliable: Dole felt no similar connection to Rwanda and was in fact one of the major congressional opponents of action there.

Third, those interested in humanitarian intervention must develop more expertise in the use of force. Historically, international human rights organizations were usually reluctant to advocate using military force, opting instead for softer sanctions. The 1990s taught the lesson that those who commit genocide have already convinced themselves that they face threats so terrible as to justify murdering women and children. They will not stop unless faced with a credible threat that they themselves will be killed. This lesson has largely been learned by human rights advocates, and much of their discomfort with military intervention has faded. Many are still reluctant, however, to go further and educate themselves properly about the details of military assessment and planning. This reluctance must be overcome, for until it is, governments seeking to avoid the political risks associated with stopping genocide will be able to resort, uncontested, to the same excuses that have worked so well for so long.

After the Cold War, conditions for international moral action appeared more favorable than at any time in memory. Power's is one of the first books to try to explain why this potential has not been realized. Realist scholars have a ready answer: the problem is inherent in international politics and cannot be solved. Constructivist writers argue that it can, and that we are making more progress than may be apparent. Although not an explicitly theoretical work, Power's book fits into an emerging liberal approach to international relations scholarship that focuses on domestic politics to explain why and when states act on their moral beliefs, and she makes the domestic explanation persuasive. Whether this should make us any more optimistic about the prevention of future genocides is unclear.

You are reading a free article.

Subscribe to Foreign Affairs to get unlimited access.

  • Paywall-free reading of new articles and a century of archives
  • Unlock access to iOS/Android apps to save editions for offline reading
  • Six issues a year in print, online, and audio editions
Subscribe Now