|
|
|
|
The New American Interventionism: Essays from Political Science Quarterly
Demetrios James Caraley (ed.)
1999
Public Support for Peacekeeping in Lebanon and Somalia:
Assessing the Casualties Hypothesis
*
James Burk
Among the many lessons that the U.S. military and political leaders drew from the Vietnam War, one was that the military could not deploy effectively or for very long without public support. Another was that public support for military deployments had declined quite predictably (one is tempted to say invariantly) in response to the accumulation of casualties. 1 In principle, it is nothing but wise counsel that rulers ignore the deadly consequences of military deployments only at their peril.
Recently, however, a new lesson has been taught that resembles but should be distinguished from these. Based on accumulating experience with peacekeeping operations, it holds that the public will not support peacekeeping deployments if they lead to the loss of American lives. This lesson is a source of worry and concern for those who believe armed forces are an essential instrument of national security policy, and it is taken very seriously. Since the barracks bombing that killed 241 Marines in Beirut in 1983, national security doctrine has required that there be “some reasonable assurance” of public support before combat forces are committed abroad. 2 The worry is that public support for missions, which seems sufficient, will quickly evaporate when faced with American casualties. It is often noted, for instance, that public support for American participation in the peacekeeping mission to Somalia was strong. But immediately following the deaths of eighteen soldiers in the streets of Mogadishu, negative public reaction forced President Bill Clinton to withdraw U.S. forces. 3 As John Mueller observed, “when Americans asked themselves how many American lives it was worth to save hundreds of thousands of Somali lives, the answer came out rather close to zero.” 4 The concern is that public intolerance of casualties radically constrains the government’s ability to use armed force effectively to defend national interests and to maintain a more peaceful world order. 5 Quick reversals of public support ignore the long-range goals of foreign policy, jeopardize mission accomplishments, and underestimate the logistical difficulties or political costs of rapid withdrawal. Under these circumstances, military and political leaders are understandably wary of undertaking any military action.
The question is whether these worries and concerns are warranted. How reliable is the evidence that the public is so reluctant to accept even minimal casualties that it is fickle in its support of military missions and hamstrings political decisions about the use of force? 6 Is the public so intolerant of casualties that it will only approve peacekeeping operations that are virtually casualty free? And, in any case, does public opinion about casualties wield so much influence over political and military elites that it controls or constrains national security policy making?
In this article, I consider these questions in light of trends in public opinion about American peacekeeping in Lebanon from 1982 to 1984 and, ten years later, in Somalia from 1992 to 1994. The cases are not directly comparable. The intervention in Lebanon occurred during the cold war in an area of the world in which the United States thought it had (and still thinks it has) vital security interests. The intervention in Somalia had no such justification. Somalia was important to the United States as part of its cold war rivalry with the Soviet Union, but afterward American interest in the country flagged. The decision to intervene in Somalia in late 1992 was fairly disinterested, in sharp contrast to the intervention in Lebanon. Nevertheless, these cases are alike in key respects, important for present purposes. They were both highly visible peacekeeping missions attracting public attention, unlike for instance the American peacekeeping mission in the Sinai. They were both peacekeeping missions in which combat casualties unexpectedly occurred. And they were both popularly regarded as missions that failed. For these reasons, these cases may help us to see whether casualties cause the public to withdraw its support of the missions and whether, in fact, the withdrawal of public support drives political leaders to abandon the missions earlier than they might otherwise have done. My argument is that public support for military deployments was neither as unsteady nor as uncritically contingent on the absence of casualties as many have claimed. Nor was it obvious (although it is difficult to prove conclusively) that public worry over casualties was a key factor affecting government decisions about normal security policy.
Before turning to these empirical claims, however, we need to examine more closely the logic of what may be called, the “casualties” hypothesis. It seems perfectly plausible to say that the public is concerned when foreign policy pursuits cause bloodshed and that in a democracy public concerns carry weight. But what reasons do we have to believe it? It matters what we say. To my knowledge, Edward Luttwak has offered the only comprehensive defense of the hypothesis. 7 His argument is provocative, especially its recommendations about how best to frame current military strategy. Yet it is not entirely persuasive. His contention that public intolerance of casualties is a new phenomenon, resulting from recent changes in family structure, is at best historically misleading. Moreover, he offers no reasons for believing that public opinion will exert a determining influence over national security policy. That is a surprising omission given the accumulation of evidence in recent years that relationships between public and elite opinion are extraordinarily complex and that the influence they have on one another defies summary by any single simple rule. 8
Intolerance of Casualties
Put simply, the casualties hypothesis states that American public opinion at present will not support the deployment of military forces abroad if that deployment results in the lives of American soldiers being lost. It is a strong claim and should be distinguished immediately from a related but substantively weaker claim that public support for military operations takes the risks of casualties into account.
We have strong evidence that Americans think the risk of casualties is a crucial, perhaps the most important, factor affecting their support of a decision to use armed force. In an Americans Talk Security survey done in 1988, well over 80 percent of the respondents said that the cost in American lives was an important factor to consider before deploying military force. No other factor ranked higher. The second most important consideration, at just under 80 percent, was the likely cost of civilian deaths in the area of combat. Other factors were mentioned, but were thought less important and some, like the possibility of failure or the cost in dollars, were thought much less important. 9 This is not the first or the only evidence we have on this score. John Mueller deserves credit for systematically documenting the strong negative relationship between the accumulation of casualties and public support for the Korean and Vietnam Wars. His analysis is well known and requires no detailed retelling here. It is enough to recall that in both cases a ten-fold increase in casualties led to a significant drop of 15 percentage points in support of these wars. More recently, Mueller studied public support for the Persian Gulf War. In this case, the ground war was mercifully short, American casualties were remarkably few, and support for the war while the war was going on “soared.” It was lower before the fighting began. Nevertheless, Mueller concludes on evidence from polls querying the matter that this support was not unconditional. It was contingent on the level of casualties. “[A] drop off of support,” he thought, “would have followed a logarithmic pattern as in Korea and Vietnam.” 10 This judgment cannot be tested but is probably accurate given the divided state of public opinion on the eve of the war with Iraq, a division that was not evident when the country initially faced war in Korea and Vietnam.
This evidence, however, shows that public support for military deployments erodes as casualties accumulate and that erosion takes time. It does not speak directly to the hypothesis we are interested in. It does not say that the public will only support what are virtually casualty-free military deployments, a claim that lies at the center of the casualties hypothesis. And it does not hint that public opinion will quickly abandon support for an operation in midstream if casualties are taken. This stronger claim has been made by military and political leaders in private and public, and has affected their willingness to undertake peacekeeping operations short of war. Drawing on U.S. experience in Somalia, for instance, General John Shalikashvili, then chairman of the Joint Chiefs of Staff, noted the public’s unwillingness to support military operations as soon as casualties were incurred. He worried that insistence on casualty-free operations may have a deleterious effect on the judgment of young military leaders. Some might say that his was a defensive posture adopted to prevent the military from being sent on deployments where for a variety of reasons the risk of failure may be high. Obviously, no institutional leader wants to preside over failure. 11 But we should not rush to dismiss the hypothesis as a rhetorical exaggeration rather than an empirical claim to be taken literally and seriously.
Luttwak’s Defense of the Casualties Hypothesis
Recently, Edward Luttwak, a noted scholar of national security affairs, incorporated the casualties hypothesis into a larger argument about what strategic posture the United States should adopt in the post-cold war environment. 12 His argument runs as follows. Historically, great powers, like the United States, have been willing to project force beyond their own borders to secure their own interests. And they have been willing to accept causalities as the price of success. “To lose a few hundred soldiers in some minor probing operation or a few thousand in a small war or expeditionary venture,” he says, “were routine events for the great powers of history.” 13
Not now. Whether we look at the United States or Britain or France or even the Soviet Union before its downfall, we find societies “so allergic to casualties that they are effectively debellicized, or nearly so.” 14 In this situation, Luttwak fears, lesser powers will be quick to flex their muscle, unworried that the resulting violence will be checked by great power interventions. His primary concern, therefore, is to find some means of reorienting great power strategy that respects the public’s disinclination to accept casualties and yet permits great powers to use force to limit violent conflicts while they are still small and more easily contained. This can be done, he thinks, if force structures, presently organized (if necessary) to fight large-scale war are redesigned to wage something like siege warfare, following the model of ancient Rome and relying on advanced technology “to use force remotely, yet accurately and with discrimination.” 15
Luttwak may or may not be correct in his prescriptions for strategy and force structure. That is not our present concern. More important here are the reasons for believing that casualty-free warfare is required. His rationale is sociological. In preindustrial and early industrial societies, there was greater tolerance for casualties than in the advanced postindustrial societies of the present day. 16 That greater tolerance was founded on a family demography that was vastly different from the family demographics of today. In the past, families were larger, and it was “normal to lose one or more children to disease.” So “the loss of one more youngster in war had a different meaning than it has for today’s families.” While always tragic, “a death in combat was not the extraordinary and fundamentally unacceptable event that it has now become.” This changed attitude toward the death of young family members in combat reflects a widely shared expectation that children born into post-industrial families, fewer in number, will survive to old age and that each will consume “a larger share of the family’s emotional economy.” That is why parents and adults generally (to include decision-making elites) “react with astonishment and anger when their children are actually sent into combat situations. And they are apt to view their wound or death as an outrageous scandal, rather than as an occupational hazard.” 17
Luttwak does not say when this change in outlook occurred. Presumably it is relatively recent. And its influence was shown, he believes, in the decision to withdraw U.S. forces from Somalia after the lives of eighteen soldiers were lost, in the reluctance of the United States to send troops into Haiti to overturn its military dictatorship, and in the reticence of European powers to risk ground troops to counter aggression in the former Yugoslavia. 18
Criticisms of Luttwak’s Argument
There are at least three difficulties with the logic of this argument. First, we must suppose, if Luttwak is right, that earlier generations were more tolerant of casualties than ours. Is that really the case? It may be true that great powers incurred more casualties in earlier times. But that is not evidence for his argument. Perhaps the chief difference between then and now is that the public today is better able than it was to hold elites to account for the consequences of their security policy. If so, their ability might stem from an earlier intolerance of casualties. It is not improbable that public revulsion and unrest over the casualties incurred during World War I forced political and military elites since then to take greater care of the lives of soldiers in their command. We know with certainty that soldiers objected to being thrown into battle without regard for their survival; it was not only Russian soldiers who engaged in mutiny. 19 Jay Winter’s recent study of that war documents the extensive grief that civilians felt for the death of loved ones and strangers. Attempts by government officials to assuage or channel those feelings in support of their policies were not always successful. Tellingly, while war memorial art recognized the sacrifice by citizens, the message was “expressed in terms of sacrifice that must never be allowed to happen again.” 20 Birth rates in Britain, France, and Germany were on average much higher from 1910 to 1920 than they are today.
Second, Luttwak attributes a complex change in the pattern of public opinion to just one factor, the decline in birth rates with all that implies for family structure and the emotional economy of family life. But single factor theories, for all their luster, dim as we imagine how many other factors might explain the same result. Luttwak does briefly consider and dismiss the possibility that the changed attitude toward casualties is explained by the growth of democracy or the publicity of mass media. 21 Even if we accept these dismissals, other possibilities remain. A major conclusion of Winter’s study, for example, is that our traditional language for justifying death and consoling the bereaved, revived during World War I, lost its persuasive power after that war and has not been replaced by any other. The modernist culture dominant since then, as shown in the work of Paul Fussell, can express anger and despair at war’s casualties, but it does not console or heal. Military actions may still be justified, but there is no shared theodicy to explain away the losses. 22 Unwillingness to endure casualties may also rest on the perception that there are alternatives to putting armed forces at risk. 23 In his study of public opinion and the Persian Gulf War, John Mueller suggests that before the war began, support for it was divided, in part because many believed sanctions could force Iraq’s withdrawal from Kuwait. Public willingness to support resort to force weakens when a range of nonmilitary options are available. 24
Third, turning from causes to consequences, Luttwak argues that in consequence of the public’s intolerance of casualties, great powers now avoid resort to force; they are effectively “debellicized.” It may be true that public tolerance of casualties is low and that great powers are reluctant to use force in pursuit of foreign policy. It remains to be said how these states of the world are linked. Is there evidence that public intolerance has the causal power to affect (to debellicize) great power policy? Surely, it is a plausible hypothesis. But that is not enough. The causal process must be specified and evidence must be provided to show that it produces the predicted effect. Failing to do that, Luttwak’s predicament is formally similar to the one Arnold Toynbee faced when he argued that war weariness causes peace. His “evidence” was the “long peace” that followed the Napoleonic Wars. But, as Geoffrey Blainey asks, what caused the “long peace” from 1870 to 1914, when there was no war weary generation? Or what caused the outbreak of World War II, which war weariness should have prevented? 25 The problem, however, is not only a formal omission to state the causal mechanism at work. It is substantive as well.
Recent studies of public opinion reject simple models of influence between the public and policy makers, no matter whether the direction of influence attributes dominant power to the public or to the policy-making elite. 26 The relationship between them is highly mediated by experts through systems of mass communication and so is more complex than either simple model would suggest. Public opinion, to include knowledge about foreign affairs, is shaped (literally mediated) by the views of elites (public officials, intellectuals, and journalists) whose judgments and opinions are broadcast electronically and reported in print. 27 That does not mean, however, that elites can manipulate and dominate public opinion. Much depends, as the work of John R. Zaller and Benjamin I. Page has shown, on whether elites agree or are divided among themselves about the appropriate policy and on the relative complexity of the policy issues under debate. In general, the more complex the issue and the more unified elites are about the appropriate policy response, the less likely public opinion will oppose and constrain the policy-making elite. But when the facts of the issue are relatively easy to grasp and elite opinion is divided, then public opinion may wield significant influence and operate as an important constraint on policy making. Most important is elite disagreement. As Zaller has argued, “resistance to [elite] persuasion depends very heavily on the availability of countervalent communications, either in the forms of opposing information or of cueing messages from oppositional elites.” 28
Applying these findings to our problem suggests an alternative to the casualties hypothesis: namely, that public support of particular peacekeeping deployments depends on the degree of elite consensus that the deployment is justified or required. Expectations of consensus and dissensus must be framed in relative terms, of course. Since the Vietnam War, elites have been divided over whether U.S. involvement in world affairs should be, in Wittkopf’s words, “cooperative” or “militant.” 29 Indeed, on some accounts, elite opinion has divided over the wisdom of foreign “entanglements” ever since the country’s founding. We may, nevertheless, attempt to weigh the preponderance of elite opinion. Recently, for example, Eric V. Larson examined the relative unity of public and elite opinion for American wars from World War II to the Gulf War and for the less intense deployments to Panama and Somalia. Based on that half-century of analysis, he concluded that public support for military deployments&-;to include toleration of casualties–“is influenced heavily by consensus (or its absence) among political leaders.” 30 Given his findings, we have reason to expect that mass support for peacekeeping operations and tolerance of casualties are bolstered by evidence of a favorable elite consensus and eroded by evidence of elite disagreement. This elite consensus hypothesis offers a complex alternative to the casualties hypothesis for understanding public support for peacekeeping. It does not preclude considering the effects of casualties on any determination of support for a mission. But it rejects the view, supposed by the casualties hypothesis, that policy makers are virtual hostages to the public’s recoiling from the loss of American life. It suggests that there is room for political leaders to shape public opinion and create a forum for public deliberation and debate.
In sum, the first two difficulties cast doubt on Luttwak’s argument that public opinion is somehow less tolerant of casualties today than it has been. Those difficulties do not allow us to reject the claim. But they render it suspect. The third difficulty is more serious. It shows the inadequacy of Luttwak’s argument that public opinion about casualties constrains decisions made by the policy-making elite. The difficulty is only formal and might be overcome by specifying the connection between mass and elite opinion. When this is done, however, we are led not only to doubt the casualties hypothesis but to consider an alternative, elite consensus hypothesis that is apparently better grounded in current public opinion research. In the following section, we examine public reaction to the Lebanon and Somali deployments to consider which hypothesis has greater support.
Support for Peacekeeping in Lebanon and Somalia
American peacekeeping missions in Lebanon and Somalia were preceded by a radical deterioration of social order caused by civil war. Neither country had an effective government able to exert its authority across the territory. 31
In 1982, Lebanon was a country occupied by no fewer than six indigenous armed forces. In addition, Lebanon had been occupied since 1976 by a large Syrian force, dispatched formally by the Arab League to end the civil war. In the early 1980s, it was occupied also by Libyans, Palestinians, Israelis, and a small number of Iranian Revolutionary Guards. In addition, the United Nations (UN) had peacekeeping forces (UNIFIL) in southern Lebanon to monitor a cease-fire in that area. The complexities posed by this military situation reflected the difficult politics of the Middle East, heightened by the tensions of cold war. Following Israel’s invasion of Lebanon in June 1982 to destroy Palestinian Liberation Organization (PLO) strongholds, the United States quickly intervened to arrange a cease-fire and help guard the withdrawal of PLO forces from Lebanon. In late September–following a massacre of Palestinians by the Phalangist Lebanese Forces, under the eye of the Israeli Defense Force–the United States intervened again as part of a multinational peacekeeping force that aimed to restore the authority of the legitimate government in Lebanon. U.S. Marines were deployed around the Beirut airport. French, Italian, and a small number of British troops were deployed elsewhere in the city. In the spring of 1983, U.S. and Italian troops began to patrol Beirut in support of the official, but contested, government and its army. At that point, they met armed attacks from opposing Muslim forces. In April, the U.S. embassy in Beirut was blown up by a car bomb. In August, U.S. Marines were under mortar attack from Druse militia forces (killing two). Then, on 23 October, 241 Marines were killed in a suicide terrorist attack in which a truck bomb was used to blow up the Marine barracks at the Beirut airport. President Ronald Reagan threatened to retaliate; but sporadic fighting continued without result. In January 1984, despite rising concern over escalating the conflict, President Reagan argued against sudden withdrawal of the multinational force. Nevertheless, the French announced that they would begin to withdraw some of their forces. In February, U.S. forces were also withdrawn to ships off shore and, on 30 March 1984, Reagan formally ended the U.S. role in the multinational peacekeeping force.
In 1991, Somalia too was wracked by a civil war. The military situation in Somalia was less complex than in Lebanon. There were no foreign forces to contend with, and the cold war was no longer available to intensify and in some ways to restrain the conflict. But in human terms, it was no less serious. The fighting disrupted economic life and led to a severe food shortage that threatened starvation for millions of Somalis. Concern for their straits led the UN Security Council in 1992 to monitor a cease-fire of the conflict and to facilitate delivery of humanitarian aid for the country’s relief (UNOSOM I). Relief efforts, however, were seriously hampered when rival clans attacked the relief convoys and stole the food. In December 1992, President George Bush sent U.S. troops to Somalia for peacekeeping duty. They were to protect the relief efforts from armed attack by rival clans, a goal that required proactive peace enforcement from the beginning. In February 1993, the United States declared that security in Somalia was “substantially restored.” It prepared to transfer control of the peacekeeping and humanitarian mission to the UN (UNOSOM II) in May. This was done despite ongoing demonstrations (sometimes violent) against U.S. and French forces by followers of the Somali warlord, Mohammed Farah Aidid, who complained that the peacekeepers favored his rival. Shortly after the UN took over, its forces too were attacked, but more seriously. On 5 June 1993, over twenty Pakistani soldiers serving as UN peacekeepers were ambushed and killed, allegedly by Aidid’s supporters. The UN Security Council condemned the incident and called for the arrest and punishment of those responsible. This began military operations to capture Aidid. Unfortunately for the UN, these operations were not successful. They only increased the number of casualties taken on both sides. On 8 August, four Americans were killed while patrolling streets in Mogadishu. On 24 August, President Clinton sent 400 U.S. Army Rangers to Somalia to capture Aidid. Their efforts were mired in difficulty. On 27 August, U.S. troops conducted a raid in which they mistakenly “captured” UN aid workers. Worse, on 9 September, when crowds turned on UN troops fighting Somali militia, U.S. helicopter fire killed about 200 Somali civilians. The Rangers did apprehend four of Aidid’s close aides; but Aidid remained at large. On 3 October, eighteen Rangers were killed, over seventy were wounded, and one was captured in a fifteen-hour firefight with Aidid’s forces. Television pictures showed Aidid’s men dragging dead American soldiers through the streets of Mogadishu. Four days later, Clinton announced that he would send reinforcements to Somalia immediately, but he also promised to withdraw most troops by 31 March 1994.
Public response to the Somali debacle, we have seen, is frequently cited as evidence for the casualties hypothesis. The perception is that the public supported the deployment so long as there were no casualties, but once the Rangers were killed, support collapsed; and political elites without public support were forced to withdraw the forces, whether or not it was sound policy to do so. Put generally, the argument breaks down into three claims. First, public opinion in support of peacekeeping missions is volatile; rather than a stable foundation on which to base national security policy, it is subject to quick reversal. Second, the most important cause of the erosion of public support for force deployments is the taking of casualties; the loss of eighteen U.S. soldiers precipitated the debacle in Somalia. Third, rapid erosion of public support for a peacekeeping mission causes political elites to suspend the deployment.
Assessing the Casualties Hypothesis
Are these three claims borne out by what we know of the public’s response to peacekeeping missions in Lebanon and Somalia? To answer this question, I identified the population of poll items dealing with these two missions and carefully examined all questions asking about public approval, decisions to remain or withdraw, and assessments of the missions’ success. 32 Ideally, a number of questions with variant wording for each topic would be asked repeatedly over the entire course of both peacekeeping missions. Moreover, the questions asked during the Lebanon mission would be very much like the questions asked during the Somali mission. That ideal is rarely met in public opinion analysis, and it is not met here. Nevertheless, to answer questions about change, trend data are required. And to check the effect of casualties, the trend data must begin before the truck bombing in Beirut or the unsuccessful Ranger attack in Mogadishu and then continue somewhat beyond. While few questions met these criteria, fortunately some did, and they refer to the critical issue of public support for the two missions. Public support is measured here by questions that asked whether the respondents approved or agreed with the decision to send troops on the mission and by questions that asked whether troops should withdraw or stay in the area.
To assess whether public support for these missions was stable or changing, I followed the conventions outlined by Benjamin I. Page and Robert Y. Shapiro in their study, The Rational Public. 33 Public opinion is stable if it has not changed significantly over time. To determine whether public opinion has changed, they first recalculate poll responses to exclude don’t know, not sure, and no opinion responses from the results. 34 After that, any movement in responses from one time to another of six percentage points or more constitutes significant change. If the change is twenty percentage points or more, it is large; and if the change occurs at a rate of ten or more percentage points a year, it is abrupt. Public opinion fluctuates when it reverses itself within a given time interval, operationally defined to mean two or more significant changes in opposite directions within two years or three or more over four years. Change is gradual if it is neither abrupt nor fluctuating.
| Table 1: Public Approval of U.S. Peackeeping Mission in Lebanon, 1982–1984 (in precentages) |
|
|
Questions: (1) NBC News/Associated Press: Do you approve or disapprove of President Reagan’s decision to send American troops to Lebanon? (2) CBS/NY Times: U.S. Marines went to Lebanon as part of an internatioal peacekeeping force to try to prevent fighting there. Do you approve of the government sending troops to Lebanon for that purpose? (3) Gallup: Do you think the U.S. made a mistake sending the Marines to Lebanon or not? (To ensure comparability across all questions, “not a mistake” responses were coded “approve” and “approve” and “mistake” response were coded “dissaprove”). Note: Percentages were recalculated to eliminate “don’t know”, “not sure”, and “no answer” responses. |
Applying these conventions to the data reported in Tables 1 and 2, we can say that public support for U.S. peacekeeping missions in Lebanon and Somalia was definitely not stable. In Lebanon, before the truck bombing of 23 October 1983, opinion was divided over whether to approve or disapprove of the mission, with most disapproving (Table 1). It is impossible to say with certainty whether the level of approval declined from the fall of 1982 to early fall 1983. The 1982 question (1) is not repeated in 1983 and the closest parallel questions (2 and 3) are worded too differently for us to attribute exact meaning to the different levels of approval and disapproval. We can say that from September through November 1983, approval of the mission rose abruptly from 40 percent to 61 percent in question 2 and that the change is evident immediately after the truck bombing in October. A similar but less abrupt rise in mission approval is shown in the data for question 3. Interestingly, that rise in approval reversed itself by January 1984. In early 1984, the level of support for the mission was not significantly different from what it was in early October 1983, before the truck bombing occurred. In brief, in this case, opinion change is significant, large, abrupt, and fluctuating.
| Table 2: Public Approval of U.S. Peacekeeping Mission in Somalia, 1992–1993 (in percentages) |
|
|
(1) CBS/NY Times: Do you think the United States is doing [did] the right thing to send U.S. troops to Somalia to try and make sure shipments of food get through to the people there, or should U.S. troops have stayed out? (2) Times/CNN/Yankelovich: In general do you approve of disapprove of the presence of U.S. troops in Somalia? (3) CBS/NY Times: Given the [possible] loss of American lives and the other costs involved, do you think sending U.S. troops to make sure food gets through to the poeple of Somalia is worth the cost or not? (To ensure comparable interpretation across all questions, “worth it” responses were coded “approve” and “not worth it” responses were coded “disapprove”). Note: Word in brackets in question 1 and 3 represents variant wording. Percentages were recalculated to eliminate all “don’t know”, “not sure”, and “no answer” responses. |
Almost the same can be said about support for the peacekeeping mission in Somalia (see Table 2). From the winter of 1992–1993 to early fall 1993, there is a significant, large, and abrupt decrease in public approval of the peacekeeping mission. One cannot say how large or abrupt the change was. Taken together the three questions used to measure this movement give widely different estimates. Higher levels of approval generated by question 1 may well be caused by asking whether it was the “right thing” to send troops to provide food relief, which was always a popular rationale for the mission. Question 2 asked simply whether one approved “in general” of the presence of troops in Somalia. This question enabled the respondent to take more factors into account. It is quite conceivable that one could believe both that it was right to send troops for food relief and disapprove of U.S. presence in Somalia on other grounds. Question 3 linked mission approval of the humanitarian purpose with the costs, and that depressed levels of support attained when purpose alone is mentioned. Nevertheless, the pattern is the same in each instance. Opinion certainly changed over the period, and the change was, by the conventions adopted here, large and abrupt. There are, however, two differences worth noting between the patterns here and in Lebanon. First, the Somali data show no tendency toward fluctuation. Once the public changed its mind, no later than the fall of 1993, its mind stayed changed. This conclusion is supported by other poll data for 1994 collected at single points in time. Second, perhaps more important, in Lebanon opinion changed in the first place toward greater support for the mission and then by winter returned more or less to its original position. In Somalia the change withdrew support for the mission. Note too that, unlike the Somali mission, support for the Lebanon mission even in the beginning was not very high; opinion was closely divided.
Can we say that these changes in opinion resulted from the unexpected and unwelcome experience of casualties to American soldiers? Obviously, the public did not ignore and was not indifferent to the loss of life in these deployments. Nevertheless, I am hard-pressed to find support in either case for the supposed public unwillingness to support operations because casualties were incurred. At best the data are conflicting. In Lebanon, public reaction to the truck bombing increased support for what never was a very popular peacekeeping operation. Two months later, support began to decline. One might explain the decline as a cooling off of tempers raised by the initial terrorist attack or as recognition that the conflict was continuing without hope of resolution. In either case, support for the operation in February 1984, when Marines were pulled out of Beirut, was not significantly different from what it was in early October 1983, before the truck bombing occurred. One cannot conclude that the loss of 241 Marines, tragic as it was, eroded support for this peacekeeping mission.
In Somalia, support for the mission did erode, and there is evidence that support eroded further and substantially (dropping ten percentage points) after eighteen Rangers were killed in Mogadishu. But even here, one hesitates before concluding that these casualties are the reason why support for the mission declined, increasing support for an early withdrawal. The trend data for question 2 in Table 2 show that by September support for the mission had already declined precipitously (by thirty-four percentage points) from its January high. In short, 77 percent of the decline in public approval of the mission occurred before the Rangers lost their lives. Casualties may still have played a role in the erosion of support for the mission. Throughout the summer of 1993, the UN force faced Somali resistance and suffered casualties. On 9 September, a U.S. helicopter opened fire on a crowd of Somalis, killing about 200 people, including women and children. The UN claimed that those killed were combatants or were being used by combatants as human shields. This incident may also have played a role. Recall the data reported above, from the Americans Talk Security Survey, showing that the public opinion does consider the likelihood of civilian deaths in the area second only to the likely death of American soldiers when deciding to use force in foreign affairs.
Conceding this, one still cannot confidently conclude that casualties had the predicted strong effect. The reason why is that the U.S. Somali peacekeeping operation, which people approved in the winter (Operation Restore Hope), was not the same peacekeeping operation in which the United States was involved in the summer and fall (UNOSOM II). Although the questions remained the same, the mission they referred to in reality had changed. Quite possibly, public support for U.S. involvement in UNOSOM II was never very high. From its beginning, in late spring 1993, that operation aimed more at using force to end the civil war than at providing humanitarian relief. But peacekeeping to end the civil war in Somalia was not as popular as famine relief. A Gallup poll taken in December 1992 asked whether the U.S. role in Somalia should be “limited to delivering relief supplies” or widened “to bring a permanent end to the fighting.” Only 31 percent thought the United States should attempt to end the fighting; 59 percent thought the role should be limited to delivering relief supplies. 35 Most thought that their preferences were being acted on. A Times Mirror News poll taken in January 1993 showed that 81 percent believed the principal objective of American forces in Somalia was “to restore enough order so that famine relief can take place.” Only 10 percent believed it was “to disarm the gangs of Somali gunmen”; 9 percent were not sure. 36 In late March and early April, when asked to rank various possible missions in Somalia, 62 percent of the respondents thought the use of the military for humanitarian purposes was “very preferable.”
In contrast, 45 percent thought it very preferable “to arrest leaders of warring factions.” Just 19 percent thought it very preferable to use “overwhelming force to defeat the primary aggressors.” 37 In June, about two-thirds of the public did approve of U.S. actions to retaliate against Aidid for attacking UN forces. Yet many (42 percent) worried that these actions would cause the United States to be “bogged down in Somalia.” 38 By mid-September, only 36 percent believed that the mission there was “under control.” Most (52 percent) thought the United States was “too deeply involved in Somalia.” The reason they thought so is tied plausibly to their still strong belief, held by 69 percent, that U.S. troops should “only be responsible for making sure that food is delivered.” Few (22 percent) believed the troops should “be responsible for disarming the rival warlords there.” 39 While we would prefer repeated questions to assess the trend, these various questions asked before (some long before) the Mogadishu firefight show that the public never supported the more aggressive mission of UNOSOM II, as they did the mission of famine relief. Admittedly, this argument is more complex and the evidence less telling than in the Lebanon case. Still, on balance it is difficult to find strong support for the casualties hypothesis in the data for Somalia. That is striking, as this case is often supposed to provide clear-cut evidence in favor of the hypothesis.
Given this analysis, it is difficult to conclude that changing public opinion had the predicted effect on political decisions to suspend the peacekeeping missions. Once the putative link between eroded support and casualties is broken, the entire logic of the analysis breaks down. Nevertheless, in both cases, public opinion favored withdrawing U.S. forces from the regions before the president announced a decision to withdraw. In Lebanon, public sentiment for withdrawal increased markedly from 39 percent to 61 percent between October 1983 and January 1984 (Table 3). Over the same period, support for sending more troops, never strong, fell below 10 percent. President Reagan did not announce his decision to remove Marines from Beirut until February 1984. Only a limited number of trend data are available for Somalia (Table 4). They all follow the 3 October firefight and show an unvarying and high proportion–nearly two-thirds–favoring withdrawal from Somalia. The first of these polls was taken the day before President Clinton announced his decision to withdraw from Somalia in six months. Examination of these and other poll data taken at single points in time suggest that the public accepted the president’s proposal for a gradual withdrawal; support for immediate withdrawal was limited.
|
Table 3: Public Pressure to Remove Troops from Lebanon, 1983–1984 (In percent) |
||||
| Deploying Troops | ||||
| Date | Remove Troops | Same Number | Send More | n |
| 22–26 Sep 83 | 42 | 51 | 7 | 1430 |
| 23 Oct 83 | 53 | 23 | 23 | 454 |
| 25 Oct 83 | 41 | 27 | 32 | 588 |
| 26 Oct 83 | 48 | 35 | 17 | 686 |
| 26–28 Oct 83 | 44 | 38 | 18 | 1184 |
| 28 Oct 83 | 39 | 43 | 18 | 491 |
| 3–7 Nov 83 | 42 | 44 | 14 | 1400 |
| 8–13 Dec 83 | 51 | 40 | 9 | 1431 |
| 3 Jan 84 | 63 | 32 | 5 | 571 |
| 4 Jan 84 | 61 | 30 | 9 | 478 |
| 12–17 Jan 84 | 60 | 32 | 7 | 1463 |
|
Question: ABC News Washington Post: Would you say the U.S. should send more troops to Lebanon, leave the number of troops about the same, or remove the troops that are there now? Notes: Percentages and sample sizes were recalculated to eliminate all don’t know, not sure, no answer responses. Not all rows add to 100 due to rounding. |
||||
It is certainly likely that both Reagan and Clinton were aware of the poll numbers before they made their decisions. But other factors than these weighed in the balance. Reagan faced increasing opposition to the Lebanon deployment, especially to his announced policy to retaliate against terrorist attacks, from European allies including Margaret Thatcher and, at home, from congressional leaders. In late December 1983, even Robert Michel, the Republican House minority leader, urged Reagan to remove troops from Beirut, an opinion joined in January 1984 by Republican Senator Charles Percy, who chaired the Senate Foreign Relations Committee. Clinton faced similar pressures in Somalia. In August 1993, the Italians withdrew from UNOSOM II, redeploying their forces in Somalia under their own command. They, along with other allies, objected to the “Rambo methods” of the UNOSOM commander. At home, congressional opposition to the Somali mission mounted as well. In September 1993, two days before the helicopter incident, the Senate passed a nonbinding resolution urging Clinton to outline the goals of the mission by 15 October and to seek congressional approval for further involvement. The House followed with a similar resolution on 28 September. In this context, it is difficult to conclude that public opinion–and especially that public reaction against the unexpected casualties–played a powerful determining part in shaping policy.
|
Table 4: Public Pressure to Remove Troops from Somalia, 1993 (in percent) |
|||
| Deploying Troops | |||
| Date | Withdraw | Remain | n |
| 6 Oct 93 | 65 | 35 | 483 |
| 6–7 Oct 93 | 65 | 35 | 1027 |
| 18–19 Oct 93 | 65 | 35 | 831 |
| 5–7 Dec 93 | 64 | 36 | 1193 |
Overall, examination of the evidence for Lebanon and Somalia lends some, but not much, support to the hypothesis that public intolerance of casualties renders it a changeable and untrustworthy basis for deciding whether to launch or sustain peacekeeping deployments. Opinion was changeable. But it is not clear that casualties (or casualties alone) affected the change. Sudden, unexpected casualties did not cause the public to withdraw its support for either peacekeeping mission. In Lebanon, support for the mission rose, albeit briefly, after the truck bombing, which killed so many Marines. In Somalia, support for the U.S. deployment did fall in reaction to the firefight in Mogadishu that killed eighteen Rangers. But we must remember that public support for the mission had declined significantly even before the Rangers were killed. This evidence makes it difficult to believe that the public withdrew its support for either peacekeeping mission simply because of these tragic incidents or that its withdrawal was a decisive cause forcing political leaders to abandon the missions.
Assessing the Alternative Elite Consensus Hypothesis
If not by casualties, how can we explain the changing patterns of public support for peacekeeping missions? It is possible, of course, that we cannot. Studies of public opinion that point to the orderliness and interpretability of public beliefs have been based on examinations of long-term trends. 40 Opinion about peacekeeping missions form and change and turn into history in a matter of months. Perhaps that is not time enough to become aware of, work through, and reach a settled judgment about the complex issues that most peacekeeping initiatives raise. Or, perhaps, short-term changes over a period of months are due to the random shock of particular events beyond the reach of generalization. Such fluctuations may mask a more stable and orderly pattern that is only evident when we average observations over a period of years, as daily market fluctuations may mask the long-term trend in stock prices. They are interpreted at one’s peril.
The issue, of course, is not so easily disposed. When criticizing the casualties hypothesis, we noted there was a significant alternative to it–the elite consensus hypothesis. Put simply, this hypothesis holds that when political elites (or, more generally, the “experts”) agree on the course of policy, then public opinion, especially over the short term, is likely to go along; but when elites disagree, so will the public. Following this logic, we expect the public to withhold or withdraw its support from peacekeeping deployments when there is division among elites about the necessity for the deployment or the availability of alternatives to military action; and this will occur even if there are no casualties. At the same time, we expect the public to continue support for a deployment, despite casualties, if elite opinion is solidly behind it. How well does this hypothesis explain the observed patterns of public response to the deployment in Lebanon and Somalia?
Consider Lebanon. The first deployment, in August 1982, to protect the withdrawal of the PLO was not controversial. It was limited in scope, incurred no casualties, and was a quick success, ending early. Notice too that the president notified Congress of the deployment, acting consistently with the War Powers Resolution. That resolution requires the president to consult with Congress and gain its authorization for deployments of armed forces into hostile situations or situations where hostilities are imminent. Presidents have been reluctant to grant that they are subject to that resolution; they see it as an encroachment on their powers as commander-in-chief of the armed forces. 41 It remains a source of friction between the two branches of government. That any notice was sent suggests that the president expected no strong opposition to the deployment. The second deployment in late September 1982, following the Phalangist massacre of Palestinians, was quite different. It was always controversial, even within the administration that proposed it. It was strongly opposed by Secretary of Defense Caspar W. Weinberger and the Joint Chiefs of Staff, who thought the mission was vaguely defined. With no certain objective to achieve, it was difficult to define appropriate force size, equipment, rules of engagement, etc. 42 Yet Secretary of State George P. Shultz was equally strong in his support for the deployment and, most important, so was the president. 43 This time, however, President Reagan did not act consistently with the War Powers Resolution; no notification was sent to Congress, and Congress did not authorize the deployment.
In early 1983, elite divisions over this issue deepened and broadened. Marines were attacked while on patrol in Beirut and, on 18 April, the U.S. embassy was bombed, leaving dozens dead and many more wounded. The next day, congressional objections to the deployment were strongly stated. Senator Barry Goldwater (R-AZ) was blunt: it was time, he said, to withdraw the Marines from Lebanon. The Foreign Affairs Committees of both houses approved sending additional aid to Lebanon, but only on condition that any expansion of U.S. forces in Lebanon had congressional approval. Agreement on this forestalled an effort by Senator Claiborne Pell (D-RI) directly to apply the War Powers Resolution to Lebanon. It did not silence congressional criticism of the deployment, which continued through the summer and never subsided entirely.
By late August 1983, when fighting in Beirut killed two Marines–the first American combat casualties–members of Congress renewed calls to review troop commitments as required under the War Powers Resolution. Reagan resisted, creating a dispute between Congress and the administration. Finally, a compromise was reached. Congress passed a resolution that placed the Marines’ involvement in Lebanon under the War Powers Resolution, but also authorized the administration to keep Marines there for another eighteen months. Reagan signed the resolution, but asserted “that he felt no constitutional obligation to seek congressional authorization” to extend the deployment. 44
In short, long before the barracks bombing, political leaders were divided over the wisdom and conduct of this mission, and their divisions grew deeper and more vocal over time. Nor were they silenced by the barracks bombing. Although the House defeated a proposal to end funding for the peacekeeping effort by March 1984, congressional criticism of the deployment intensified through the fall. Division over this issue was also found in newspaper editorials. Immediately after the barracks bombing, the Wall Street Journal argued that it was “time for retaliation” against military targets in Syrian-occupied Lebanon, even if those responsible for the bombing were not positively identified. Action was necessary to demonstrate U.S. resolve against terrorists. 45 The New York Times, in contrast, argued that America was once again being held hostage in the Middle East in pursuit of a “murky diplomatic cause”–removing Syrian backed forces from Lebanon–that is “not a vital American interest.” It asked bluntly what purpose U.S. forces served in Lebanon and under what conditions would they finally depart. 46
Given elite dissensus, it is hardly surprising that public support for the Lebanon operation was divided from the beginning and remained divided until early 1984, when opinion finally crystallized to favor removing the troops (Tables 1 and 3). What is surprising is that public support for the operation rallied briefly in the fall following the barracks bombing, despite casualties and in the face of elite dissensus. The explanation for that rise in support remains matter for speculation. Perhaps it reflected collective anger and frustration with the terrorist attack, so different from acts of war. Perhaps it represented a rally effect associated with the successful invasion of Grenada, which was launched two days after the Beirut bombing. Whatever the cause, the rally was short-lived. By early December, public pressure mounted in favor of withdrawal from Lebanon. Again surprisingly, at this point, public opinion was to some degree in front of (it certainly was not following) the opinion of the still-divided political elite. In this case, the domestic public pressure joined with the pressure of foreign elites to urge a change in U.S. policy. As the year ended, House Speaker Tip O’Neill (D-MA) called for a meeting of Democratic leaders to review their support for the mission, and House Minority Leader Bob Michel (R-IL) called for the Marines to be withdrawn. Administration leaders still defended policies they hoped would unify the country in support of the Lebanon mission. But they failed to do this before the barracks bombing. They were unlikely to succeed, as conditions in Lebanon grew worse, following the barracks bombing. 47 With violence continuing in the region and opposition mounting in Congress, the president’s policy to stay and perhaps retaliate for losses was not sustainable. On 1 February 1984, the caucus of House Democrats called on the president to withdraw the Marines from Lebanon. One week later, he did.
In Somalia, initial support for the humanitarian relief effort was high not only among the public (Table 2) but among political leaders as well. And support was bipartisan. In fact, although it was Bush’s decision to intervene, Congress took the initiative in the summer of 1992 to promote intervention. 48 Nevertheless, as shown by hearings held by the House Foreign Affairs Committee in December 1992, members of Congress were in no mood to write blank checks. Their support was for a humanitarian relief effort to relieve the famine and nothing more. Once that was done, U.S. forces were to withdraw. A similarly restrictive understanding of the mission to Somalia was embodied in the Senate Joint Resolution 45 (SR 45). Passed in February 1993, SR 45 authorized the use of American armed forces in Somalia, as required under the War Powers Resolution, but it prescribed a narrow humanitarian mission for them.
It was not long before continuing clan warfare showed that sticking to limited aims would be difficult. In March, the new administration under President Clinton declared a victory for Operation Restore Hope and turned the problem back to the UN for a new operation to begin in May. The new operation, however, had an expanded mission. Beyond humanitarian relief, it was to end civil war and begin democratic nation building. It was, moreover, to require continuing commitment of U.S. forces, although many fewer than before. As the terms of this transition were negotiated at the UN, the House of Representatives debated its version of SR 45 to authorize the troop deployment to Somalia. Here is the earliest evidence of elite dissensus over the peacekeeping mission. The House resolution adopted a broad definition of the mission, akin to that assigned to UNOSOM II. But expanding the mission beyond what Bush initially outlined and the Senate had approved was strongly opposed, largely along partisan lines, both in committee and on the House floor. Although it passed on 25 May, by a vote of 243–179, it was never “taken up by the Senate because Senate majority leader [George] Mitchell [D-ME] would not allow debate.” 49
What has most to be explained in this case and what the casualties hypothesis left in doubt was the sharp drop in support for the mission between January and September 1993. It is possible, but by no means certain, that public support eroded in the spring with erosion of elite consensus in Congress about what the mission in Somalia was supposed to be and disagreement over who was competent to authorize and define it–the president or Congress. Yet, as we learned from our analysis of the casualties hypothesis, the public never expressed much support for a peacekeeping mission that had other than humanitarian purposes. The public’s mood was captured in SR 45 rather than in the broader resolution passed by the House. The strength of elite support for the broader resolution would be tested when the mission in Somalia changed radically in June 1993 to include a quest to bring Aidid to justice for the massacre of Pakistani peacekeepers. Surprisingly, one looks in vain for evidence in June that elites were divided over this shift to a more aggressive policy. New York Times editorials strongly supported action against Aidid, suggesting that “Mr. Clinton dare not flinch” and that “threatening General Aidid with arrest seems a minimal way of expressing international condemnation.” 50 The Wall Street Journal was less effusive but also supportive. After the peacekeepers attacked Aidid, with more attacks sure to follow, it editorialized that the UN’s mission must so far be regarded a success and that attacks against Aidid showed that UN troops cannot be murdered with impunity. 51 Indirect evidence is also revealing. Strikingly absent from reports about Somalia filed in June and broadcast on the network evening news are any interviews with public figures or leaders from the United States who opposed the actions taken. Protests by Somalis against the more aggressive UN actions were aired. 52 On the surface, public opinion accepted the new policy. We noted earlier that about two-thirds of the public approved of U.S. participation in UN military operations against Aidid, though they were worried that the action might cause the United States to be bogged down in Somalia’s civil war. Here perhaps is evidence that the public may suspend its doubts when elites agree that a given policy should be pursued. By September many minds had changed.
As the United States became more actively involved in efforts to capture the warlord Aidid, the mission was subject to heightened scrutiny and criticism in Congress. In July 1993, Senator Robert Byrd (D-WV) spoke against the mission on the Senate floor, objecting to the expense of the peacekeeping mission–he chaired the Appropriations Committee–and noting that the Senate never authorized use of American troops in Somalia for peace enforcement purposes. As peace enforcement was not contemplated by SR 45, the senator had raised an important constitutional question about which branch of government should authorize foreign deployments of American armed forces. Hostilities intensified through the summer and, though U.S. casualties were low (fewer than eight), Congress returned from its summer recess ready to pick up the cudgel that Senator Byrd had thrown down. Disillusioned with the peace enforcement effort, on 9 September the Senate voted 90–7 to require President Clinton to seek congressional authorization before 15 November to continue the mission in Somalia. The House passed an identical measure on 28 September by a vote of 406–26. Congress was controlled by the president’s party. Still, it would not follow the president’s lead in this matter. Moreover, we know from Eric Larson’s work that the mass media widely reported congressional concerns with administration policy. 53 Elite disagreement was not kept from the public view. Yet, as public support for peace enforcement was never high, we cannot say that decreased public support for the mission resulted from the resurgence of elite opposition in September.
We can say that the October firefight in Mogadishu brought elite disagreement to a climax, making it inevitable that Clinton would not try to expand or to extend the expanded mission in Somalia. It sealed the doom of a policy already so unpopular among congressional leaders that it would probably have been sharply modified or discontinued in any event. In this context, it is perhaps surprising that public opinion favoring withdrawal from Somalia remained the same (at 65 percent throughout the fall). It did not rise as elite disagreement gave way to consensus over what the policy should be. This may be an artifact of the polling, as the first poll on this issue was not conducted until 6 October. By this time, there was already a clear elite consensus favoring gradual withdrawal from Somalia. Alternatively, public skepticism about the mission in Somalia may have exerted pressure on elite opinion to favor withdrawal. But, if so, public opinion was not behaving as predicted by the elite consensus hypothesis. The most we can say is that public support for the mission eroded as elite consensus eroded, that public disillusionment, already strong, grew stronger when casualties were taken, and that public support for gradual withdrawal was unwavering once political elites agreed to limit the mission and leave Somalia by March 1994.
In sum, we cannot conclude with certainty, on the basis of evidence here, exactly how elite and public opinion are interrelated. Our expectation, based on current public opinion theory and research, was that patterns of elite opinion were the guiding force and that public support for peacekeeping missions would be high when elite opinion agreed on the missions and then drop as elite opinion about them became more divided. But neither case offers much support for that hypothesis. Some evidence in the Somali case suggests that when elites agreed on the policy to pursue, the public did suspend disbelief and go along. In both cases, however, it was most often the case that elite and public opinion were divided. Yet when opinion was divided, there was no certain relation between elite and public opinion. In the Lebanon case, at least, elite division did not invariably predict an erosion of public support. We would stand on shaky ground to conclude that public opinion is patterned after elite opinion. What we encountered rather was a highly mediated situation in which presidents and other leaders decided on policies that elites and the public evaluated and then accepted or rejected as their deliberation on the matter led them to believe. The causal order, if any, that explains the influence of one judgment on the other remains unclear.
Conclusion
We began by asking whether there was reliable evidence to support the central claims of the casualties hypothesis. These were, first, that the public will not support military deployments which result in casualties; second, that public support for a deployment would be abruptly withdrawn if casualties unexpectedly occurred; and, finally, that public opinion on this issue is so powerful that it constrains the use of force by–indeed, effectively debellicizes–great powers. Based on this article’s analysis of public and elite opinion during U.S. peacekeeping deployments to Lebanon and Somalia, we may conclude with confidence that evidence for this hypothesis is lacking. While public opinion was not insensitive to the deaths of American soldiers, public approval or disapproval of both missions was, in fact, largely determined before casualties occurred. Opinions about the missions did change. The changes were sometimes large and abrupt. In the case of Lebanon, public opinion fluctuated back and forth in response to the terrorist bombing that killed so many Marines. But when opinion fluctuated about the bombing, it moved in a direction opposite to what the casualties hypothesis predicts. It moved to continue rather than end the deployment. In Somalia, support for the mission had withered already in response to the changing mission before the firefight in Mogadishu.
Stating the matter strongly, patterns of public support for peacekeeping missions reveal no irrational or knee-jerk reactions based on a putative unwillingness to tolerate casualties. They reveal no immature demand for a casualty-free security policy, both wrong to promise and impossible to deliver. They reveal a public prudently cautious about what to expect from the use of force. They are willing to risk it for humanitarian purposes, but they are risk averse with respect to the more complex and often partisan peace enforcement missions.
In light of these findings, why have so many found the casualties hypothesis persuasive, especially within the political elite? One reason may be, as Donald C. F. Daniels suggests, that members of Congress place more credence in what their constituents tell them than what national polls show, and that patterns of constituent contacts might actually support something like the casualties hypothesis. 54 This is a valuable suggestion. Taken seriously, it means that clear specification of the relation between elite and public opinion requires a different kind of study than undertaken here. Local studies are needed to compare public opinion within congressional districts with opinions expressed through constituency contact and to examine whether these influence or are influenced by the policy positions taken by members of Congress. Alternatively, rhetoric about public intolerance of casualties and its influence over policy may simply mask an intra-elite struggle, still unresolved, about the distribution of warmaking powers between the president and Congress. In both Lebanon and Somalia, elite opinion about the deployments divided most often over the application of the War Powers Resolution. The resolution asserts that the president’s power over force deployments is not unilateral, but must be shared with Congress whenever forces are deployed in situations where hostilities are imminent. Presidents may say a situation is not hostile, of course. But hostilities are hard to deny once casualties are taken. On this account, casualties are a hot issue, because they trigger intra-elite conflict. Both of these ideas require further research.
Note, finally, a key assumption of democratic theory–namely, that the public is competent to decide what public policy should be, including national security policy. The casualties hypothesis may seem to share this assumption. It does not. It certainly holds that public opinion is powerful. But by supposing public support for peacekeeping deployments rests on the utopian condition that military operations be casualty free, the casualties hypothesis lays a groundwork for discrediting the quality of public judgment, for denying its competence. By supposing that elites are unwittingly constrained to desire whatever public opinion at the moment desires, the casualties hypothesis adopts the view that democracy is prone to corruption, as Aristotle and others long after have feared. In short, the casualties hypothesis undermines confidence in the public to govern itself wisely and well, at least in the area of foreign affairs. This study suggests that there is no reason to undermine confidence. The claims of the casualties hypothesis lack support and should be laid to rest.
Endnotes
*: Earlier versions of this paper were presented at the Pontignano Seminar on Public Opinion, Democracy and Defence Policy, Siena, Italy, 7–10 October 1996, funded in part by the United States Information Agency, and at the biennial meetings of the Inter-University Seminar on Armed Forces and Society, in Baltimore, October 1995. I am grateful for comments received from the conference participants and from Christopher Dandeker, Robert Goldich, Maria Bina Palmisano, Richard Wells. Back.
Note 1: On the issue of support for deployment, see Harry G. Summers, On Strategy: A Critical Analysis of the Vietnam War (New York: Dell, 1984); and Lewis Sorley, Thunderbolt (New York: Simon and Schuster, 1992). The key source on the relation between accumulation of casualties and trends in support for war is John E. Mueller, War, Presidents and Public Opinion (New York: Wiley & Sons, 1973). Back.
Note 2: This is the fifth element of the so-called Weinberger doctrine. See Caspar W. Weinberger, Fighting for Peace (New York: Warner Books, 1990), 442. Back.
Note 3: See, for example, Donald C. F. Daniel, “The United States” in Trevor Findlay, ed., Challenges for the New Peacekeepers (Oxford, England: Oxford University Press, 1996), 94–95; and Ioan Lewis and James Mayall, “Somalia” in James Mayall, ed., The New Interventionism 1991–1994 (Cambridge, England: Cambridge University Press, 1996), 94–124, esp. 118. Back.
Note 4: John E. Mueller, “Policy Principles for Unthreatened Wealth-Seekers,” Foreign Policy 102 (Spring 1996): 31. Back.
Note 5: General John Shalikashvili, when chairman of the Joint Chiefs of Staff, voiced this concern in testimony before the Senate Armed Forces Committee, as reported by Williams Matthews, “Shali Warns Against Myth of Nonviolent War,” Navy Times (2 October 1995), on-line edition. He is not the only one to have done so. See Benjamin C. Schwarz, Casualties, Public Opinion, and US Military Intervention (Santa Monica, CA: Rand, 1994); Eric V. Larson, Ends and Means in the Democratic Conversation (Santa Monica, CA: Rand, 1996); Charles J. Dunlap, “The Origins of the American Military Coup of 2012,” Parameters 22 (Winter 1992–93): 2–20; and A. J. Bacevich, “The Use of Force in Our Time,” Wilson Quarterly 19 (Winter 1995): 50–63. Back.
Note 6: In the current debate over civil-military relations, this question has also been raised the other way around by Deborah D. Avant, “Are Reluctant Warriors Out of Control?” Security Studies 6 (Winter 1996/97): 51–90. Wondering about the influence of military leaders over foreign policy, she asks: “Are civilian decisions [to employ force] hamstrung by a reluctant military?” Her answer is “not quite” (52). Her analysis and mine, while oriented differently, are not unrelated. We both argue that the decision to use force rests with civilian elites and is substantially affected by whether elites are unified or divided in their outlooks. Back.
Note 7: See Edward N. Luttwak, “Where Are the Great Powers?” Foreign Affairs 73 (July/August 1994): 23–28; “Toward Post-Heroic Warfare,” Foreign Affairs 74 (May/June 1995): 109–122; “A Post-Heroic Military Policy,” Foreign Affairs 75 (July/August 1996): 33–44. Back.
Note 8: On this, see Michael X. Delli Carpini and Scott Ketter, What Americans Know about Politics and Why It Matters (New Haven: Yale University Press, 1996); Benjamin I. Page, Who Deliberates? (Chicago: University of Chicago Press, 1996); Benjamin I. Page and Robert Y. Shapiro, The Rational Public (Chicago: University of Chicago Press, 1992); John R. Zaller, The Nature and Origins of Mass Opinion (Cambridge, England: Cambridge University Press, 1992); and Bruce Russett, Controlling the Sword (Cambridge, MA: Harvard University Press, 1990). Back.
Note 9: Eugene R. Wittkopf, Faces of Internationalism: Public Opinion and American Foreign Policy (Durham, NC: Duke University Press, 1990), 228–229. Back.
Note 10: For the original analysis, see Mueller, War, Presidents and Public Opinion, 60. Analysis of the Persian Gulf War is found in John E. Mueller, Policy and Opinion in the Gulf War (Chicago: University of Chicago Press, 1994), 77. See also Benjamin C. Schwarz, Casualties, Public Opinion, and US Military Intervention, 17. Back.
Note 11: See Avant, “Are Reluctant Warriors Out of Control?”; Stevenson, “The Evolving Clinton Doctrine”; Matthews, “Shali Warns Against Myth of Nonviolent War”; and Colin L. Powell, “Why Generals Get Nervous,” New York Times, 8 October 1992. Back.
Note 12: I focus on the argument of Luttwak, because it is the most fully developed one I know. For references to his argument see above note 7. But the central hypothesis is repeated uncritically in other places. For instance, Bacevich, “The Use of Force,” 60, writes: “An enterprise that yesterday seemed expedient is today not worth the blood of a single American soldier. In the case of Somalia, such thinking led the United States to abandon its commitment altogether.” And Daniel Yankelovich, Coming to Public Judgment (Syracuse: Syracuse University Press, 1991), 62 writes: “Sooner or later public opinion makes itself felt, sometimes directly as in the public pressure that . . . persuaded President Reagan to withdraw the Marines from Lebanon after a number of Marines had been killed by a terrorist bomb.” Back.
Note 13: Luttwak, “Where Are the Great Powers?” 27. Back.
Note 15: Luttwak, “A Post-Heroic Military Policy,” 43. Back.
Note 16: Luttwak, “Toward Post-Heroic Warfare,” 115. Back.
Note 17: Luttwak, “Where Are the Great Powers?” 25. Back.
Note 18: Ibid., 23–24; Luttwak, “Toward Post-Heroic Warfare,” 115. Back.
Note 19: See S. P. MacKenzie, Politics and Military Morale (Oxford, England: Oxford University Press, 1992), 12–13. Back.
Note 21: Luttwak, “Where Are the Great Powers?” To support this claim he draws attention to Soviet conduct of its war with Afghanistan. The Soviet Union was no democracy and yet, he asserts, authorities adopted “inordinately prudent tactics” in order “to avoid casualties at all costs” (24). Moreover, they “never allowed its population to see any television images of the war;” nevertheless, “the reaction of Soviet society to the casualties of the Afghan war was essentially identical to the American reaction to the Vietnam war” (25). Back.
Note 22: Winter, Sites of Memory; see Paul Fussell, The Great War and Modern Memory (New York: Oxford University Press, 1975); and Fussell, Wartime (New York: Oxford University Press, 1989). Back.
Note 23: I am grateful to Ben M. Crouch for suggesting this possibility. Back.
Note 24: Mueller, Policy and Opinion, 62; and Wittkopf, Faces of Internationalism, 181–186. Back.
Note 25: Geoffrey Blainey, The Causes of War, 3rd ed. (New York: Free Press, 1988), 7–9. Back.
Note 26: See Russett, Controlling the Sword, 87–118; also Page, Who Deliberates?; and Zaller, The Nature and Origin of Mass Opinion. Back.
Note 27: As Benjamin I. Page observed in Who Deliberates? 107: “Research has indicated that experts’ views, as presented in the media, have a significant impact on public opinion: when experts speak in favor of a policy proposal, public opinion tends to move toward supporting that proposal.” Back.
Note 29: Wittkopf, Faces of Internationalism; see also Ralph B. Levering, “Public Opinion, Foreign Policy, and American Politics since the 1960s,” Diplomatic History 13 (Summer 1989): 383–393. Back.
Note 30: Eric V. Larson, Casualties and Consensus (Santa Monica, CA: Rand, 1996), xv. Back.
Note 31: Unless otherwise noted, facts about these two cases are drawn from Facts on File Yearbook (New York: Facts on File, 1983–1985, 1993–1995); Keesing’s Record of World Events (London: Longman, 1982–1984, 1992–1994). Back.
Note 32: I collected all public opinion poll data on deposit with the Roper Center that mentioned either “Lebanon” in the years 1982 through 1984 or “Somalia” in the years 1992 through 1994. I conducted this search on line, accessing the Roper Center through CompuServe, and downloading the results. The Roper Center has public opinion poll data back through 1960 available for such on-line searches. I also searched for items before and after the years of focal interest. A few items were found, but there were virtually no items of any relevance to the present research. There were substantially more poll items reported for Lebanon (n=671) than for Somalia (n=267). Many of the items relating to Lebanon dealt with public reaction to Israel’s invasion of Lebanon begun in June 1982, an important topic beyond the scope of this article. Otherwise, almost all of the items for both Lebanon and Somalia related directly or indirectly to the U.S. involvement in peacekeeping. Unfortunately, very few questions were asked repeatedly over the course of either operation, so trend data are not as numerous as we would like. Back.
Note 33: Page and Shapiro, The Rational Public, 44–53. Back.
Note 34: The rationale for excluding these data is that it clarifies tracking changes “in the balance of opinion among those people who express opinions,” which is of course what we are interested in doing here. Page and Shapiro illustrate why with the following example. Take two data points. In the first, responses are 20 percent favor, 40 percent oppose, and 40 percent have no opinion. In the second, responses are 30 percent favor, 60 percent oppose, and 10 percent no opinion. The apparent increase in public favoring or opposing the policy issue, caused by including no opinion data masks that on balance at both points in time, 33 percent of those with opinions favored the policy. It is that kind of masking that their methodology aims to avoid. See The Rational Public, 423–424. The don’t know, not sure, and no opinion responses for the trend data reported here never exceeded 10 percent and ranged most often between 4 percent and 7 percent. Back.
Note 35: Gallup Organization, Question ID: US GALLUP.121192 R2 (Storrs, CT: Roper Center, 1995). Back.
Note 36: Times Mirror News Interest Index, Question ID: USPSRA.011393 R11 (Storrs, CT: Roper Center, 1995). Back.
Note 37: Americans Talk Issues, Question ID: USMS.ATI121 R40-42 (Storrs, CT: Roper Center, 1995). Back.
Note 38: These data come from two sources. Over 18–21 June, Gallup pollsters asked, “Last week, the United States participated in a military operation with the United Nations against one of the warlords in Somalia. Do you generally approve or disapprove of that decision?” 65 percent approved, 23 percent disapproved, and 12 percent had no opinion (Question ID: USGALLUP.062393 R6 [Storrs, CT: Roper Center, 1995]). Over 21–24 June, CBS News/New York Times pollsters posed a broader question. It read: “Mohammed Farah Aidid is the leader of a Somali clan whose troops killed United Nations peacekeepers. UN troops are trying to capture him. Do you think this is a good idea because it shows that violence against the UN will not be tolerated, or do you think this is a bad idea because this gets the UN involved in Somalia’s civil war?” 66 percent thought it a good idea, 23 percent thought it a bad idea, and 11 percent had no opinion (Question ID: USCBSNYT.62893A R69 [Storrs, CT: Roper Center, 1995]). But, in the same poll, when asked, “When United Nations peacekeepers are killed in Somalia, should United States troops retaliate, or is that likely to get the United States bogged down in Somalia?” 41 percent thought US should retaliate, 42 percent though retaliation would bog down the United States, and 17 percent did not know (Question ID USCBSNYT.62893A R68 [Storrs, CT: Roper Center, 1995]). Back.
Note 39: The NBC News/Wall Street Journal pollsters asked “are you concerned that the United States is too deeply involved in Somalia, or do you think that our efforts in Somalia are under control?” (Question ID: USNBCWSJ.93SEPT R19B [Storrs, CT: Roper Center, 1995]). One week later, Time/CNN/Yankelovich pollsters asked “do you think the US troops in Somalia should be responsible for disarming rival warlords there, or should the U.S. troops only be responsible for making sure that food is delivered to the areas affected by the famine?” (Question ID: USYANK.092493 R41 [Storrs, CT: Roper Center, 1995]). Back.
Note 40: See Ronald Ingelhart, Cultural Shift in Advanced Industrial Society (Princeton: Princeton University Press, 1990); Wittkopf, Faces of Internationalism; Yankelovich, Coming to Public Judgment; and Page and Shapiro, The Rational Public. Back.
Note 41: See Louis Fischer, Presidential War Power (Lawrence: University Press of Kansas, 1995); and John Hart Ely, War and Responsibility (Princeton: Princeton University Press, 1993). Back.
Note 42: Weinberger, Fighting for Peace, 151–152. Back.
Note 43: George P. Shultz, Turmoil and Triumph (New York: Charles Scribner’s, 1993), 107–110. Back.
Note 44: Fischer, Presidential War Power, 141. Back.
Note 45: “Time for Retaliation,” Wall Street Journal, 25 October 1983, 28. Back.
Note 46: “America Held Hostage,” New York Times, 24 October 1983. Back.
Note 47: This argument is elaborated in Larson, Casualties and Consensus, 96–97. Back.
Note 48: The facts in this and the next several paragraphs come from the account of Harry Johnson and Ted Dague, “Congress and Somalia” in Walter Clarke and Jeffrey Herbst, eds., Learning from Somalia, (Boulder, CO: Westview Press, 1997), 191–204. Both were participants in the events. Johnston is a member of Congress (D-FL). He is the senior member of the House International Relations Committee and formerly chaired the Subcommittee on Africa. Dague is a foreign affairs analyst for the Congressional Research Service and served as a professional staff member for the subcommittee on Africa when Johnston was its chair. Back.
Note 50: “Bloody Sunday in Somalia,” New York Times, 15 June 1993; and “Drawing the Line in Somalia,” New York Times, 20 June 1993. Back.
Note 51: “Somali Awakening,” Wall Street Journal, 14 June 1993. Back.
Note 52: Based on my own analysis of abstracts held in the Vanderbilt Television News Archive. Back.
Note 53: Eric Victor Larson, Ends and Means in the Democratic Conversation (Santa Monica, CA: RAND Graduate School Dissertation Series, 1995), 272–296. Back.
Note 54: Daniels, “The United States,” 94–95. Back.
Email this citation