David A Baldwin. International Security. Volume 24, Issue 3, Winter 1999.
The debate over whether economic sanctions “work” is mired in a scholarly limbo. One writer contends that recent international relations scholarship has promoted optimism about the utility of such measures and sets out to challenge this trend, while another notes the pessimism that “pervades the sanctions literature” and proceeds to argue that it is unjustified. A third scholar cites the sanctions literature as an example of fruitless academic debate with little policy relevance. Such divergent readings of the scholarly literature are often explained by differences in ideology or fundamentally different theoretical orientations. This does not seem to be the case with respect to the sanctions debate. Under appropriate circumstances, it is quite possible for liberals, neoliberals, realists, neorealists, or globalists to argue in favor of using economic sanctions. If the sanctions debate is bogged down, the explanation does not seem to lie in the essentially contested nature of the subject matter. A second potential explanation is that scholars are talking past one another because they ask different questions, use different concepts, and set the discussion in different analytical contexts. In short, they are talking about different things. This article explores the second explanation.
The basic paradox at the heart of the sanctions debate is that policymakers continue to use sanctions with increasing frequency, while scholars continue to deny the utility of such tools of foreign policy. Two explanations for this paradox have been offered. Some scholars dismiss policymakers as “fools” or suggest that they have not yet learned how to identify situations in which sanctions would be appropriate. Others contend that “the answer lies less in the ignorance of government officials than in the naivete of the research on… sanctions.” There is, however, a third possibility—that policymakers and scholars are addressing different questions. This possibility is suggested by Alexander George and William Simons in their discussion of coercive diplomacy They argue that distinguishing between the policymaker’s choice of a strategy and the policy outcome of doing so is “standard in the study of public policy” The ultimate success or failure of a given policy, they note, “may be determined by several factors outside the policy maker’s direct influence-or even that of his or her opponent.” Thus the question of whether sanctions “work” may be separated from the question of whether they should be used.
The purpose of this article is to separate the question of whether sanctions “work” from that of whether they should be used and to identify appropriate concepts and methods for answering each question. Until researchers agree on which questions to ask and on how to seek answers, the sanctions debate is unlikely to produce useful policy-relevant knowledge. The discussion addresses the following topics: (1) the choice of economic sanctions; (2) the evaluation of economic sanctions; (3) sanctions and military force; (4) the methodological problems of selection bias, strategic interaction, counterfactuals, and symbolic behavior; and (5) the use of sanctions in the Gulf War.
Sanctions and the Logic of Choice
Most sanctions scholars implicitly or explicitly portray their work as relevant to the question of when, if ever, sanctions should be used. It is not always clear, however, that they understand or accept the implications of this orientation. This section discusses these implications in terms of the relationship between choice and alternatives, the relationship between choice and costs, and the question of what policymakers want to know when considering the use of sanctions.
Choice and Policy Alternatives
Choice implies alternatives. If policymakers have no alternatives, there is no choice to be made. This has implications for the concept of sanctions, for the question of whether sanctions are a substitute for force, and for the evaluation of sanctions as a policy option. “Rational decision-making,” as Herbert Simon notes, “always requires the comparison of alternative means in terms” of their respective consequences. Setting economic sanctions in the context of choice, therefore, requires that they be defined in terms of means rather than ends. As tools of foreign policy, they are presumably available to policymakers for a variety of purposes and not restricted to particular foreign policy goals. A given research project, of course, may confine itself to a focus on particular goals such as the use of sanctions to promote human rights or discourage terrorism. Thus an operational definition of economic sanctions employed in a particular research project may rule out, say, such goals as changing the foreign economic policies of other countries. It is difficult, however, to reconcile definitions of economic sanctions that link them inherently to particular foreign policy goals with the logic of choice. Economic sanctions, like war, have their own grammar but not their own logic. Just as Clausewitz defined war in terms of the “peculiar nature of its means,” the logic of choice counsels students of economic sanctions to define them in terms of the “peculiar nature” of their means rather than in terms of particular foreign policy goals.
Are economic sanctions an alternative to military force? This question is often posed as if it were problematic or as if the answer lay in determining whether sanctions would “work” as well as force in a given situation. In the context of the logic of choice, however, the answer to this question is not at all problematic. As one of the courses of action available to policymakers, economic sanctions are always a policy option, that is, one of the policy alternatives they may choose to consider. In this sense, diplomacy and propaganda are also substitutes for—or alternatives to—military force. To describe sanctions this way in the context of the logic of choice implies nothing about the probable utility of such measures; it merely describes the menu of choice available to policymakers. The fact that one finds brussels sprouts disagreeable or too expensive does not make them disappear from the menu in a restaurant. And the fact that sanctions are likely to be ineffective or too costly does not eliminate them from the list of alternatives available to foreign policymakers.
Some writers have suggested that sanctions should be considered as alternatives to military force only when used independently and not when used in conjunction with other measures, such as military force. “Since many have advocated sanctions as an alternative to military force,” one study observes, “a ‘successful’ use of sanctions should enable the sanctioner to achieve its aims without a resort to arms.” Describing sanctions as alternatives to force, however, implies nothing about the degree to which a policymaker may decide to substitute sanctions for force. Even if the resources (e.g., money, time, personnel) committed to using economic sanctions are small in comparison with those committed to using military force, sanctions are an alternative to force to the extent that the resources devoted to sanctions could have been devoted to the use of force.
Rational decisionmaking requires the comparative evaluation of policy alternatives; thus any study of economic sanctions that purports to be relevant to the question of whether they should be used must take into account not only sanctions but the alternatives to sanctions. To justify a conclusion that “sanctions are a notoriously poor tool of statecraft,” it is not enough to describe the disadvantages of sanctions; one must show that some other policy alternative is better. If the menu of choice includes only the options of sinking or swimming, the observation that swimming is a “notoriously poor” way to get from one place to another is not very helpful. And if the principal alternative to economic sanctions is appearing to condone communism, racism, terrorism, or genocide, the observation that they are a “notoriously poor tool of statecraft” may miss the point. In the context of the logic of choice, the evaluation of one policy alternative in isolation from others makes little sense. Despite the dictates of the logic of choice, very few studies of economic sanctions evaluate them in comparison with alternative policy options.
Choice and Costs
Choice implies costs. If economic sanctions were costless, the question of whether to use them would have no policy relevance. When one can have one’s cake and eat it too, there are no choices to be made. Despite the importance of costs in the logic of choice, the literature on economic sanctions often discusses costs in misleading ways or ignores them altogether.
Given that costs are an essential defining characteristic of choice, one would expect any scholarly attempt to determine whether sanctions should be used to take account of the costs of doing so. Unfortunately, this is often not the case. In assessments of sanctions as tools of foreign policy, the costs of using sanctions often receive little or no attention. Even studies that refer to costs often evaluate the “success” of sanctions entirely in terms of their effectiveness in achieving various goals. From the standpoint of the logic of choice, however, a “successful” policy choice is one that maximizes the utility of the policymaker in a given situation. From this standpoint, identification of a “successful” policy choice requires consideration not only of the costs and benefits of using economic sanctions, but also the costs and benefits of using alternative policy options.
From the standpoint of the logic of choice, the neglect of costs is unfortunate; but it is even worse to give the impression that one has taken them into account when one has not. Such confusion arises when assessments of economic sanctions use the terms “effectiveness,” “efficiency,” and “utility” interchangeably. In the context of the logic of choice, the concept of efficiency implies something about the relationship between inputs and outputs, that is, between costs and effectiveness. And the concept of the utility of a course of action implies something about its net value. References to the “utility” or “efficiency” of sanctions imply that costs have been taken into consideration.
Choosing Sanctions: What Do Policymakers Want to Know?
Many scholars state or imply that the question of primary interest to policymakers considering the use of economic sanctions is whether they will “work,” by which they mean whether they will achieve their goals. What policymakers “most want to know,” according to Robert Pape, is “when the strategy of economic sanctions can change another state’s behavior without resorting to military force. ” The logic of choice, however, suggests that this consideration is neither sufficient nor primary. The rational policymaker is interested in the expected costs as well as the expected effectiveness of economic sanctions. Even that knowledge, however, is not enough. In addition, policymakers want to know the expected costs and benefits associated with the alternatives to using economic sanctions (e.g., military force or diplomacy). From the standpoint of the logic of choice, any discussion of economic sanctions that fails to compare their likely cost-effectiveness with that of alternative courses of action provide s no useful policy guidance with respect to deciding whether sanctions should be used in a given situation.
In rational decisionmaking, Herbert Simon has pointed out, it is incorrect to pose a choice between alternative A with “low costs and small results” and alternative B with “high costs and large results.” Instead, one should substitute for A a third alternative C, “which would include A plus the alternative activities made possible by the cost difference between A and B.” Bruce Bueno de Mesquita has applied similar logic to diplomacy and military force: “Thus, leaders expecting a larger net gain through diplomacy than through war… should rationally elect to pursue their goals through diplomatic bargaining and negotiating. This is true even if the expected gross gain from war is larger than the gross gain from diplomacy, provided that the cost differential is large enough (as it frequently is) to make the net effect of diplomacy preferable to war.” The implication of such reasoning is that economic sanctions (or diplomacy) may be preferable to military force even when they are less likely to achieve a given set of goals—provided that the cost differential is big enough.
When economic sanctions are being considered, the question of whether they will “work” in the sense of achieving their goals is only one of several important considerations. What rational policymakers really want to know is: How effective will they be, with respect to which goals and targets, at what cost, and in comparison with which policy alternatives? Scholarly discussions that fail to address all of these questions do not provide a basis for advising policymakers and can be quite misleading if they purport to do so.
The Evaluation of Sanctions
The question of the degree to which sanctions “worked” or were “successful” in a given situation is separable from the question of whether they should have been used in the first place. The fact that rational policymakers may choose to employ sanctions even when they have little chance of success, however, does not imply that attempts to estimate the success rate of economic sanctions are useless. Such estimates are necessary steps toward identifying the conditions under which sanctions are likely (or unlikely) to succeed. Knowledge of such conditions is necessary, but not sufficient, for deciding whether sanctions should be employed in a given situation. In making empirical estimates of the success rate of sanctions, however, one must first answer the following questions: What does “success” mean? And upon what dimensions can it be measured? After considering these two questions, the discussion returns to the issue of sanctions choice and sanctions evaluation.
Concepts of “Success”
Success is a slippery concept. Unlike power or wealth, success is not just one of many goals that may be pursued. To the extent that human behavior is purposeful, everyone may be said to pursue success. This is because success is defined in terms of favorable or desired outcomes. Both the definition of success and the implicit rules used in applying it suggest that costs are an important part of the concept. Successful undertakings are those without excessive costs. Winning a nuclear war by destroying life as we know it, imposing economic sanctions that secure the compliance of the target state only by bankrupting the sender, and amputating one’s leg to remove a wart from one’s toe are unlikely to be described as instances of success. “The operation was a success, but the patient died” does not mean what it seems to say. It is a sardonic expression implying that the term “success” is being misused. If success is defined in terms of favorable policy outcomes, it is necessary to consider both costs and benefits in assessing the success of an undertaking. The concept of a Pyrrhic victory implies a difference between real and apparent victory, and it is as relevant to economic statecraft as it is to military statecraft.
Despite the importance of costs to estimating success, the literature on economic sanctions often defines success entirely in terms of the effectiveness of sanctions in attaining goals. The costs to the sender of achieving these goals are sometimes considered, but they play no role in defining success. The following passage provides a typical view of “successful” sanctions: “What does it mean for sanctions to be successful? In general, we are interested in determining the conditions (if any) under which economic sanctions produce an intended (on the part of the sanctioner) change in policy by the target state. Clearly, sanctions can be used for other purposes, such as to support a war effort, to punish the target, or to make a symbolic statement; but we believe that the focus of the debate regarding the effectiveness of sanctions is on whether they can enable the sanctioner to achieve its goals of altering the behavior of the target.” Clearly, costs are not part of this conception of success; effectiveness is everything. Policymakers, however, are likely to view with suspicion any concept of success that does not include costs.
The narrow “behavioristic” definition of behavior in terms of policy change should also be noted. If behavior is defined more broadly to include beliefs, attitudes, opinions, expectations, emotions, and/or predispositions to act, then supporting a war effort, imposing costs on the target, and manipulating symbols in order to change the image of the sender would be included in estimates of the effectiveness of sanctions.
Logically, the specification of a standard of achievement is prior to the specification of the determinants of, or conditions for, the success of an undertaking. Thus it makes little sense to describe someone as pursuing success without specifying success in doing what. If a prospective student asks for advice on how to succeed in college, one must first determine whether the goal is to get good grades or to get a sound education. “Take easy courses” may be good advice with respect to the first goal, but “take challenging courses” may be more appropriate advice with respect to the second. Likewise, when advising policymakers on how to make economic sanctions successful, one must first ascertain the goals in terms of which success is to be defined. Advising them to “pursue modest goals” or not to “bite off more than they can chew” violates this basic precept of prior goal specification. The same is true of “explanations” of success that treat goal selection as an independent variable. If success is defined in terms of the degree to which an undertaking achieves a given set of goals, the recipe for success cannot take the form of “change your goals.”
Politicians, of course, are not scholars. Appearing to be successful may be an important part of a politician’s utility function. Thus a politician may have an incentive to publicly portray his or her goals as less ambitious than they really are in order to facilitate a claim of victory at a later date. Although unseating Saddam Hussein was not a publicly proclaimed goal of the Gulf War, there is little doubt that U.S. policymakers would have been pleased by such an outcome. And when President Bill Clinton threatened Iraq with military attack in February 1998, he was careful to outline a military goal described by the press as “so modest that he will have little trouble declaring it achieved, even if it fails to deal with the larger threat posed by Saddam Hussein.” It was generally understood, however, that the president’s real goals were more ambitious. The public proclamation of achievable goals, even when those goals are less ambitious than a politician’s real goals, is a common tactic. Thus, “don’t [publicly appear to] bite off more than you can chew” may be sound practical advice for a politician. Scholars, however, are obligated to assess the success of economic sanctions—or any other policy undertaking—in terms of the real goals of the policymaker, not just those that have been publicly proclaimed. Partial success may be all that one can realistically expect in many situations, and scholars should point this out rather than play the politician’s game of redefining success.
Dimensions of Success
The success of an undertaking, such as the imposition of economic sanctions, is measured in terms of the values of the policymaker. The more (net) value a policymaker derives from imposing sanctions, the more successful they are. Unlike business firms, however, policymakers have no common denominator of value, such as money, in terms of which to measure the overall value of a foreign policy undertaking. Thus it is useful to break success down into various components or dimensions so that each can be assessed separately. Although the five dimensions of success considered below are not the only ways to subdivide success, together they capture most of what scholars and policymakers want to know about sanctions.
EFFECTIVENESS. Although the effectiveness of sanctions in achieving goals is not the only dimension on which to measure success, it is certainly an important one. This dimension can be further subdivided into scope, weight, and domain. Scope refers to the range of issues affected by sanctions (e.g., tariff levels, nuclear testing, or respect for human rights). Weight refers to the degree to which sanctions affect the various scopes of the target’s behavior (e.g., low, medium, or high impact). And domain refers to the number of people (countries, international organizations, or whatever) affected by the sanctions. Thus the higher the degree of sanctions impact, the wider the scope, and the larger the domain, the more effective the sanctions are.
COSTS TO THE USER. Because net value is the overall measure of success, costs are an important determinant of overall success. Business firms that lose money are not regarded as successful, no matter how big their gross sales. And techniques of statecraft that involve excessive costs should not be viewed as successful, no matter what their effects are.
When sanctions are used to send a signal, of course, costs are necessary to differentiate a credible signal from “cheap talk.” Thus the grain embargo in 1980 gave President Jimmy Carter’s expression of concern about the Soviet invasion of Afghanistan more credibility than it would have had if he had merely filed a diplomatic protest. And economic sanctions against Rhodesia and South Africa enhanced the credibility of international condemnation of their racial policies. Even so, policymakers do not want to incur unnecessary or exorbitant costs. Although military attack would have added even more credibility to the foregoing signals, such measure were viewed as too costly Ceteris paribus, lower costs increase success.
COSTS TO THE TARGET(S). Inflicting costs for noncompliance on the target of an influence attempt is frequently used as a measure of success. Ceteris paribus, the higher the costs for noncompliance that sanctions inflict on the target(s), the more successful they are. Imposing costs for noncompliance is sometimes confused with failure. Robert Pape, for example, objects to treating the costs inflicted for noncompliance as a dimension of success: “The fact that a target that refuses to concede may suffer substantial costs does not turn failure into success.” Thomas Schelling responds to such reasoning in a brilliant essay entitled “The Strategy of Inflicting Costs.” He uses the example of whether it is worthwhile for one’s adversary to spend money on a bullet if one can protect oneself with the purchase of a bulletproof vest. Schelling concludes: “He has wasted his money if the vest is cheap, made a splendid investment if my vest is expensive, and if asked what he accomplished by buying his bullet should have the good sense to say that he imposed a cost on me, not that he hoped to kill me and was frustrated.” Likewise, when a country using sanctions fails to gain compliance but succeeds in imposing costs for noncompliance, scholars should recognize that the sender imposed costs on the target, not merely that it hoped for compliance and failed to get it.
STAKES FOR THE USER. Not all goals are equally important. Even though economic techniques of statecraft may have made only a small contribution to winning World War II, the value of this contribution must be weighted for the importance of the goal. Ceteris paribus, the bigger the stakes, the more valuable is the contribution.
STAKES FOR THE TARGET(S). The more the target has at stake, the more difficult the undertaking is likely to be. In competitive diving the points awarded for execution are weighted for the difficulty of the dive. When economic sanctions are employed in pursuit of difficult goals, the results should be similarly weighted. The sanctions against Rhodesia and South Africa, for example, were related to demands for fundamental changes in the governance of those states. The wonder is not that it took so long to effect such changes; it is that the changes occurred so rapidly The contribution of sanctions (if any) to these outcomes should be weighted for the difficulty of the undertaking. Ceteris paribus, the more difficult the undertaking, the more valuable is the impact of sanctions.
Estimating the success of economic sanctions is complex and difficult. Effectiveness is only one of several important dimensions to be considered. Although no common denominator of value is available to add up the various dimensions, explicit consideration of each of the above dimensions of success should provide a more comprehensive and policy-relevant estimate of sanctions’ success than estimates based solely on effectiveness in changing the policies of targets. Ceteris paribus, the accomplishment of a technique of statecraft is more impressive the greater the effectiveness, the lower the costs for the user, the higher the costs of noncompliance for the target, the higher the stakes, and the more difficult the undertaking.
Choice and Evaluation Reconsidered
The logical possibility of separating the question of whether economic sanctions should be used from the question of how successful sanctions are likely to be is undeniable. What is logically possible, however, may not correspond with what is psychologically probable. Just because it is possible does not mean that it will be easy. When the question of whether sanctions “work” is posed, the question of whether they should be used is usually lurking in the background. The temptation to move directly from an evaluation of the success rate of sanctions to a judgment of the wisdom of using them seems to be irresistible. Pape, for example, moves directly from a contention that sanctions work “less than 5 percent of the time” to an assertion that sanctions cannot be “a reliable alternative to military force.” Yet raw success rates do not provide an adequate basis for making such judgments. Even when the expectation of success is very low, the use of sanctions is justified if there is no policy alternative with a higher expectation of success. Sometimes policymakers must choose from a set of dismal alternatives.
Sanctions and Force
As instruments of statecraft, economic sanctions are alternatives to diplomacy and propaganda as well as military force. The relationship between sanctions and force, however, is of special interest to students of international relations. This section addresses four aspects of this relationship: (1) success rates; (2) concepts of success; (3) costs; and (4) scope.
Success Rates
The success rate for economic sanctions of 35 percent derived by Gary Hufbauer, Jeffrey Schott, and Kimberly Elliott is often described as “low.” But, as Marc Simon points out, “we should first ask, low compared to what?” With respect to the success rate of (initiating) military force, estimates range from 40 percent to 72 percent. Of course, those concerned with more recent wars might want to weight World Wars I and II more heavily—and the outcomes of those wars provide scant evidence that initiating a war is a useful undertaking. There are, however, additional reasons for skepticism with respect to estimates of the success rate of military statecraft.
Military Success: What “Victory”?
The deficiencies in the treatment of the concept of success in the literature on economic sanctions have counterparts in the literature on military force. “The single most important idea” in thinking about military strategy, according to Bernard Brodie, is that “expressed in the question that Marshall Ferdinand Foch used to ask, De quoi s’ agit-il?—What is it all about?'” “The history of the twentieth century,” observes Raymond Aron, “suffices to remind us that there are many ways to win a war, that various ways are not equivalent, and that final victory does not necessarily belong to the side that dictates the conditions of peace.” Such comments suggest the need for careful explication of the concept of military victory; yet the literature on military force contains few discussions of the meaning of success. Even a book entitled On the Meaning of Victory fails to address the question raised by its title. And the Correlates of War data, which carefully operationalize so many concepts related to war, offer neither conceptual clarification nor operational indicators of this, the single most important concept, relying instead on “the consensus among acknowledged specialists.”
Despite Brodie’s complaint about the tendency to treat wars as “athletic contests,” despite Aron’s warning about the complexities of “victory,” despite Clausewitz’s famous dictum, and despite Schelling’s suggestion that war be treated as a mixed-motive game, the tendency to treat war as a zero-sum game persists in the literature on military statecraft. The idea that “every war has a winner” seems to be deeply embedded in the literature on military force. Evidence of this is provided by the paucity of war outcomes coded as draws and the failure to acknowledge even the hypothetical possibility that all participants in a war might be simultaneous “losers” or “winners,” depending on how close they came to achieving the purposes for which they were fighting.
The persistence of the zero-sum conception of military conflict is puzzling because it is incompatible with many of the topics dominating the scholarly research agenda during the last fifty years. As Schelling notes: “Deterrence…is meaningless in a zero-sum context. So is surrender; so are most limited-war strategies; and so are notions like accidental war, escalation, preemptive war, and brinkmanship. And of course so are nearly all affiance relationships, arms-race phenomena, and arms control. The fact that war hurts—that not all the losses of war are recoverable—makes war itself a dramatically nonzero-sum activity.”
Considering the lack of attention to conceptualizing and measuring military success in the literature on military statecraft, healthy skepticism with respect to the success rates mentioned in the previous section appears justified.
Costs
The case for employing economic sanctions instead of force frequently depends less on the relative effectiveness of sanctions than on the relative costliness of force (which is not to say that sanctions are always cheaper than force). Thus estimates of the costs of using force are needed not only to estimate the utility of military statecraft, but also in deciding whether to use economic sanctions. Although the costs of war have always been an important concern, in the nuclear age, they can be astronomical. It is not difficult to identify students of military statecraft who give serious attention to the costs of using such measures (e.g., Jervis, Knorr, Schelling, Stam, Brodie, George, and Bueno de Mesquita). It is equally easy, however, to find works that purport to say something about the utility of military statecraft but devote little or no attention to the costs of using it.
The case for including costs in calculating the success of a military undertaking is at least as strong as that for including them in assessments of the success of economic sanctions, perhaps stronger. Yet this is often not done. Melvin Small and J. David Singer admit that their failure to take account of costs is a possible weakness in their coding of winners and losers. And Clifton Morgan and Valerie Schwebach’s comparison of the utility of economic sanctions with that of military force attributes costs to the use of sanctions while ignoring the costs of using military force. A useful comparison of the utility of force and sanctions, however, requires consideration of the costs of both.
Scope
One indicator of the utility of a technique of statecraft is the range of issues to which it is considered applicable (i.e., the variety of situations in which the technique is viewed as feasible enough to merit serious consideration). In this respect, military force is less useful than diplomacy or economic statecraft. Whereas diplomacy and economic statecraft are used with respect to a wide range of issues, including human rights, tariff levels, drug trafficking, arms sales, nuclear testing, and war, military force is used with respect to a rather small range of admittedly very important issues. One of the most striking features of the twentieth century is the reduction in the range of issues with respect to which the use of force is regarded as legitimate. In addition, although diplomacy and economic sanctions are frequently employed without military force, military force is almost never employed without diplomacy or economic sanctions. This suggests that diplomacy and economic sanctions are relevant to all of the issues for which force is considered feasible, but that the reverse is not true. Although economic sanctions are often depicted as a poor substitute for force, with respect to the range of issues to which they are relevant, the reverse appears to be true.
Some writers argue that the threat of force is a brooding omnipresence that constantly permeates all aspects of international relations. The idea is that the possibility that force might be used is in the back of every policymaker’s mind. Even if one accepts this point of view, however, the implications are far from obvious. Policymakers are presumably always considering their options. In this sense, it may be correct to say that they always have the military option in mind; but it seems unlikely that military options are the only ones on their minds. Thus it would also be correct to say that diplomatic and economic policy options are omnipresent in international relations. To the extent that policymakers are rational, of course, they would consider the probability—not merely the possibility—that various policy options will be used in a given situation. Thus the extent to which a rational policymaker worries about an adversary’s use of military force will vary from one situation to another.
Are economic sanctions a useful substitute for military force? When should sanctions be used instead of force? To answer these questions, one would have to compare the utility of sanctions and force using the same analytical framework and the same concept of success for each. Both the analytical framework and the concept of success should take account of the costs as well as the benefits associated with using each technique of statecraft.
Methodological Problems and Sanctions
Several methodological issues have been raised with respect to the study of economic sanctions. The following problems are discussed below: (1) selection bias; (2) strategic interaction; (3) counterfactuals; and (4) symbolic action.
Selection Bias
Selection bias occurs when inferences are based on a nonrandom sample of cases that is not representative of the universe of cases from which it was drawn. Some have suggested that the literature on economic sanctions suffers from selection bias because it tends to focus on situations in which sanctions were used or threatened. It seems plausible to suggest that economic sanctions are more likely to be used in situations in which the probability of success is high, thus biasing the sample to overstate the success rate of sanctions. Ceteris paribus, this is probably true; but the bias may be less—and matter less—than the critics imply. Such reasoning is an example of the failure to separate the question of whether sanctions are likely to “work” from that of whether they should be used. Indeed, the discussion in earlier sections of this article suggests that there is little or no justification for assuming that sanctions will be employed only when the probability of success is high. The logic of cho ice suggests that what matters is not the absolute level of expected utility, but rather its magnitude relative to the expected utility of alternative techniques of statecraft (e.g., diplomacy and military force). Indeed, if one assumes, for the sake of argument, that the utility of military force has declined, one would expect economic sanctions to be used more frequently even though their absolute expected utility remained constant, ceteris paribus. For reasons explained earlier, it is rational to employ economic sanctions even when the expectation of success is very low—provided the expected success of alternative courses of action is lower still.
It is not easy to identify or cope with selection bias. As David Collier points out, “Discussions of selection bias by definition presume the existence of a larger set of cases, from among which the cases under analysis have in some sense been chosen.” The identification of the universe of cases is crucial. With respect to economic sanctions, it is not obvious how the universe should be defined. Each of the following definitions is plausible: (1) instances in which economic sanctions were used or threatened; (2) instances in which sanctions were seriously considered, regardless of whether they were used or threatened; (3) instances in which an influence attempt was made, regardless of the technique used; (4) instances in which an influence attempt was seriously considered, regardless of whether it was made or not; or (5) instances of international interaction in which an influence attempt could have been considered, regardless of whether it was or not.
In each of these cases, one could make a plausible charge of selection bias by arguing that influence attempts are made or considered only when the probability of success is above a certain point. Influence attempts are not made at random, but rather in response to some dissatisfaction with an expected future status quo and an expectation that an influence attempt might improve the expected future status quo. Presumably, there are some situations—even if only hypothetical ones—in which everyone is satisfied with the expected status quo and in which no influence attempts are considered. But even such cases of “pure interaction” are subject to selection bias because of what might be labeled the “hermit effect.” People who do not want to make influence attempts (perhaps because they are not very good at it?) or who do not want others to make influence attempts on them (perhaps because they are not very good at resisting such attempts?) sometimes limit their interaction with others by becoming hermits. Although there is no exact counterpart to a hermit in international life, some states do try to insulate themselves from interaction with other states (e.g., North Korea, Myanmar, or Japan during the Tokugawa period). The point is that once one starts down the slippery slope of charging that only those who are good at the game will play it, there is no obvious stopping point.
Another form of selection bias in the study of economic sanctions cuts in the direction of underestimating, rather than overestimating, their success. This is the tendency to rule out of consideration the use of sanctions in situations in which military force is likely to be low in cost-effectiveness (e.g., routine trade negotiations). Sometimes this is done by defining sanctions to exclude such matters, and sometimes it is done by defining the universe of cases to be considered so as to exclude such cases. Either way, the effect is to rule out precisely those cases in which economic sanctions are most likely to succeed. Morgan and Schwebach, for example, base their comparison of the utility of military force and economic sanctions on cases of “serious international dispute” in which “force is the ‘ultimate arbiter.”‘ These cases are crisis situations in which military force has been explicitly threatened or in which there is a higher than normal probability of war. Such militarized disputes would seem to be biased in favor of the use of military force. Although Morgan and Schwebach assert that such cases “clearly correspond with the usual situation in which sanctions are considered,” this is not true. Sanctions are also considered in a wide variety of nonmilitarized situations. To the extent that economic sanctions are used in such militarized crisis situations, the rationale is more likely to rest on the high cost of the alternative of military force rather than on a naive expectation that sanctions will be effective. Contrary to the suggestion of Morgan and Schwebach, policymakers are not fools.
Is there selection bias in the literature on economic sanctions? Perhaps so, but there is more than one kind of selection bias. It is even possible that the different forms of bias tend to cancel one another out.
Also, from the standpoint of the logic of choice, selection bias in studying the success rate of sanctions matters only to the extent that it differs from the bias in studying success rates for alternative techniques of statecraft. The literature on military force is based largely on cases in which it was threatened or used and cases in which its use was imminent. There are few, if any, studies of the utility of force in nonmilitarized disputes, such as routine trade negotiations. Pape describes such disputes as “those for which force would never be considered.” This, however, begs the question of why force is not considered in such situations. The obvious answer is because force is low in utility for such purposes. Confining the study of the utility of force to cases in which its utility is likely to be high would seem to bias estimates of its rate of success.
Strategic Interaction
George Tsebelis has charged that both sanctions policy analysts and policymakers commit what he calls “the Robinson Crusoe fallacy.” Noting that the sanctions success rate is “a poor 40%,” he attributes this to the inability of such analysts and policymakers to identify the conditions under which sanctions will be successful because they do not use a game-theoretical approach. To the extent that this approach serves as a reminder of the need to treat sanctions in the context of strategic interaction by taking account of the likely actions and reactions of other countries as well as one’s own, it is a valuable contribution to the sanctions debate. It is useful to be reminded that “one of the reasons that sanctions have such a low success rate is that ineffective sanctions are the goal of other rational actors (target countries).” But to the extent that Tsebelis’s analysis gives the impression that Strategic analysis requires the use of formal game theory, it is a disservice. Schelling’s work suffices to refute such a contention.
Although Tsebelis’s argument is helpful in understanding why sanctions fail, it is not relevant to comparing alternative techniques of statecraft. Because all of his arguments about the “Robinson Crusoe fallacy” are equally applicable to other techniques of statecraft, his analysis provides no guidance whatsoever with respect to deciding when to employ economic sanctions as opposed to some alternative technique of statecraft in a given situation. It may well be that all techniques of statecraft have low expected probabilities of success, but this does not relieve policymakers of the need to choose among them.
Counterfactuals and Sanctions
Counterfactuals are always difficult in causal analysis, and assessing the success of economic sanctions is no exception. There are, however, some aspects of the treatment of counterfactuals in sanctions analysis that should be noted. The options of “doing nothing” or “choosing to maintain the status quo” are sometimes posed as alternatives to employing economic sanctions. These phrases, however, can be misleading. In a sense, it is impossible for a country to “do nothing.” What is usually implied by such phrases is that a country should do what it would have done had the problem at hand not arisen. The shorthand phrase often used to refer to this hypothetical state of affairs is “the conduct of business as usual.” When the Soviet Union invaded Afghanistan in 1979, the United States responded with a partial grain embargo. Some critics of this action argued that the United States should have “done nothing,” at least with respect to the shipment of grain. The United States, however, had already agreed to sell 2 5 million tons of grain to the Soviet Union during 1980. Thus “doing nothing” in this situation really would have meant going ahead with the largest Russian grain deal on record.
One of the most important uses of sanctions, according to Lindsay, is enabling the sender to “avoid creating negative images of itself. … Had Britain not placed sanctions on Rhodesia [in 1966], most countries would have seen it as a sign of British approval of the Smith regime.” Phrasing the issue in terms of “avoiding negative images,” however, can easily give the impression that no influence attempt was taking place. The same point could be rephrased to make it clear that Britain was attempting to change the behavior of other states (i.e., the behavior of holding images of Britain). The relevant counterfactual condition in this case is not the image other countries held at the time the decision to impose sanctions was being made. It is the image they would have been likely to hold in the future if sanctions or some stronger measures had not been implemented. When policymakers consider whether to use economic sanctions, the relevant status quo is not that prevailing at the time the decision is being taken, but rather the status quo that would prevail if no action or a different action is taken.
Counterfactual conditions should be considered not only with respect to sanctions, but also with respect to other techniques, such as military force. There is a tendency to take the cessation of economic intercourse for granted during wartime. Pape, for example, objects to giving sanctions any credit for the outcome of World War II—”except to the extent that blockade is a normal weapon of war.” In the twentieth century, the assumption that economic relations between adversaries will be suspended during wartime is simply regarded as “normal.” But this is neither a logical nor an empirical necessity. What if trade had gone on unfettered during World War II? What if Germany and Japan could have had free access to U.S. oil supplies? What if they could have resupplied themselves by buying civilian goods and war materiel in the U.S. market? These possibilities are so preposterous that they would be regarded as unthinkable; and that which does not get thought about gets taken for granted. Although it is often asserted that economic sanctions are likely to be ineffective if not backed up by force, it is seldom asked how effective military force would be if not backed up by economic sanctions. One situation is taken for granted; the other is not.
Sanctions As Symbols
Strictly speaking, the confusion about the symbolic uses of economic sanctions is not a methodological issue. It does, however, continue to mar the sanctions debate. The primary source of confusion is the tendency to pose a false dichotomy between the instrumental and symbolic uses of sanctions. Morgan and Schwebach’s observation about the need to understand “whether sanctions contribute to foreign policy goals or serve only symbolic purposes” is typical. Yet when sanctions are used to signal—or send a message—to other countries, they constitute influence attempts in precisely the same sense that firing a shot across the bow of a ship or some other show of force is an influence attempt. Attempting to change how other countries think is an influence attempt regardless of the means used. Using sanctions for signaling purposes is not an alternative to using them as instruments of statecraft; it is statecraft.
Consideration of the symbolic uses of sanctions in the context of the logic of choice would help to reduce some of this confusion. Typically, sanctions are depicted as being used in response to domestic pressures to “do something,” even though policymakers do not expect them to “work.” In such instances, sanctions are often described as “merely symbolic.” Such descriptions imply that the only explanation for using sanctions when their expected probability of success is low must be “domestic politics” or “mere symbolism.” Yet the logic of choice makes it clear that the decision to use sanctions depends as much on the expected utility of other options as it does on the expected success of sanctions. Those who decline to buy a Mercedes because it is too expensive seldom describe themselves as having chosen instead the “merely symbolic” step of buying a Chevrolet. People do not buy Chevrolets because they think they are better than a Mercedes; they buy them because they are more affordable. Likewise, policymaker s do not necessarily use economic sanctions because they think they are the most effective technique of statecraft; they use them because they are more cost-effective. Neither those who buy Chevrolets nor those who use economic sanctions are acting irrationally, and neither action should be described as “merely symbolic.”
Sanctions and the Gulf War
The Gulf War illustrates the importance of discussing the role of economic sanctions in the context of the logic of choice. Although portrayals of the Gulf War as a situation in which sanctions failed while military force succeeded are common, this is misleading with respect to the role of sanctions. There were some, of course, who suggested that sanctions would “work”; just as there were those who suggested that air power would “work.” Depicting the Gulf War as one in which air power failed while ground attack succeeded, however, is no more justifiable than describing the role of sanctions in similar terms. Neither bombing nor sanctions were sufficient, but each made valuable contributions to the overall effort. Consider the following propositions.
First, in choosing to employ economic sanctions, policymakers were at least as concerned with the likely consequences of alternative policy options, especially military force, as with the likely consequences of sanctions. Both General Colin Powell, chairman of the Joint Chiefs of Staff, and General Norman Schwarzkopf, the military commander, worried about the consequences of premature use of military force. No one wanted another Vietnam.
Second, policymakers viewed economic sanctions as an alternative to military force. In the first place, sanctions were seen as part of the menu of choice. Powell was especially concerned with laying out the pros and cons of sanctions “so that the President would have an alternative to going to war.” In addition, during the early stages of the conflict, sanctions were viewed as a preferable alternative in the sense that they were more cost-effective. Some have noted that early on “there were no military options available to the United States” because “the armed forces necessary to force Iraq to leave Kuwait were simply not in place.” This was untrue, however. The United States could have delivered nuclear strikes against Iraq and/or Iraqi troops in Kuwait sufficient to force a withdrawal. The problem was not that such military options did not exist; nor was it that they were unlikely to be effective. The problem was that such options were simply too costly to be feasible. Economic sanctions, however, were feasible and could be implemented the same day the invasion occurred. Thus sanctions were not only an option worth considering; they were a substitute for military force in that they had more expected utility at that point in time.
Third, the decision to employ economic sanctions was not based on a naive belief that they would accomplish all of the U.S. objectives. Time was an important consideration from the beginning. As early as August 1990, President George Bush made it clear that he was skeptical that sanctions would work “in an acceptable time frame.”
Fourth, there is no evidence that policymakers failed to consider the implications of likely countermoves by Iraq (the Robinson Crusoe fallacy). On the contrary, they spent a great deal of time trying to anticipate such countermoves and considering their implications for U.S. strategy For example, one of the early fears was that air strikes might trigger an Iraqi counterstrike against Saudi Arabia before the United States was ready to counter such a move.
Fifth, economic sanctions strengthened the military effort undertaken later. The sanctions provided a useful interim measure during the military buildup, thus helping to “make a military option more feasible.” The lack of serious advocacy of carrying on “business as usual” with Iraq during the buildup suggests widespread understanding of the relevant counterfactual condition.
Sixth, economic sanctions played an important signaling role before war began. Demonstrating resolve is not difficult if one is not concerned with costs. A massive military attack, nuclear or not, would certainly have demonstrated U.S. resolve. The tricky challenge facing the United States was how to do this without appearing reckless. The economic sanctions, combined with vigorous diplomacy and the gradual military buildup, enabled the United States to demonstrate both resolve and restraint simultaneously. Both were needed to forge the coalition that eventually emerged. Reliance on diplomacy alone would have risked appearing to lack resolve, while premature use of military force would have projected an image of the United States as trigger-happy.
Seventh, sanctions imposed significant costs on Iraq for noncompliance. Iraq’s heavy dependence on oil exports meant that the costs of refusing to withdraw from Kuwait were immediate and substantial. The sanctions did not force Saddam Hussein to withdraw, but they did make him pay a heavy price for his intransigence. It is not clear that a more cost-effective policy option was available for this purpose during the early stages of the crisis.
Eighth, although economic sanctions did not bring down Saddam Hussein’s regime, neither did the later use of military force. This is a useful reminder that influence attempts should not be classified as failures just because they do not achieve all of their goals. As Robert Art observes, “A given instrument can carry a state part of the way to a given goal, even though it cannot carry the state all the way there. At one and the same time, an instrument of statecraft can usefully contribute to attaining many goals and yet by itself be insufficient to attain any one of them.”
In the end, dislodging Iraq from Kuwait required the combined efforts of ground forces, air power, naval forces, diplomacy, propaganda, and economic sanctions. Although the importance of each waxed and waned during various stages of the crisis, each contributed to the outcome. Descriptions of the crisis as one in which military force had to be used after sanctions failed is simplistic and misleading. It is instructive to note that the chairman of the Joint Chiefs of Staff did not come away from the Gulf War believing that sanctions are useless or counterproductive, but rather that they are “a useful weapon in the armory of nations.”
Conclusion
Is the sanctions debate hopelessly bogged down or can it evolve into a fruitful and policy-relevant dialogue? The discussion above suggests that progress is possible but that “quick fixes” are unlikely to produce this result. Without fundamental rethinking of the terms of the debate, it is likely to continue going nowhere. Below are four basic requisites for reorienting the sanctions debate suggested by the above discussion.
First, the separability of the question of whether sanctions are likely to “work” from that of whether they should be used must be acknowledged. Were it not for an interest in the wisdom of using economic sanctions, the question of whether they “work” would probably never be asked; and the question of whether they should be used cannot be answered without consideration of the likelihood that they will “work.” Thus the two questions are related even though they are separable. It is nevertheless essential to recognize that they are different questions and that answering one does not automatically provide an answer to the other. A corollary to this requisite is rejection of the idea that only a foolish policymaker would choose to employ economic sanctions when the prospects of success are low. The wisdom of a decision to use sanctions is determined not by whether their expected utility is high or low, but by whether it is higher or lower than that of alternative courses of action.
Second, the importance of costs must be acknowledged. Consideration of costs is essential to answering either question. Regardless of whether one is trying to determine the absolute utility (i.e., success) of sanctions or their utility in comparison with alternative policy options, costs must be considered. Pyrrhic victories are not successes, and partial victories are not defeats.
Third, the need for comparative evaluation of policy instruments must be acknowledged. The wisdom of using economic sanctions cannot be determined by studying the costs and benefits of sanctions in isolation from the costs and benefits of alternative techniques of statecraft. The literature on economic sanctions cannot answer the question of whether they should be used without reference to the likely utility of alternative policy instruments, such as military statecraft and diplomacy. Thus it is not enough for students of economic sanctions to fulfill the requisites listed here. Until the scholarly literature on diplomacy and military force becomes equally sophisticated, the necessary utility comparisons cannot be made. Until students of all types of statecraft ask the same questions using the same concept of success, satisfactory answers to the question of when to use economic sanctions will remain elusive. Progress in the sanctions debate depends on progress in the study of statecraft in general.
Fourth, the complexity of “success” as a multidimensional concept must be acknowledged. Both the literature on economic sanctions and that on military statecraft need more complex concepts of “victory” Simplistic concepts of success defined in narrow “behavioristic” terms that ignore costs and imply zero-sum conflicts are not sufficient. Costs can be ignored in zero-sum games because the costs to one party are by definition benefits to the other, but in mixed-motive games, this is not true. Economic sanctions, like war, have their own grammar but not their own logic. The teachings of Schelling and Clausewitz are likely to be more helpful than those of MacArthur and Moltke.
As the foregoing discussion demonstrates, neither methodological problems nor theoretical orientations are insurmountable obstacles to progress in the sanctions debate. When methodological problems are considered in the context of the logic of choice, they appear less troublesome.
Scholars and policymakers alike should be wary of assertions that “the logic of employing economic sanctions as an instrument of statecraft is well understood.” Similar wariness is appropriate with respect to the assertion that sanctions “still have far to go before they can be a reliable alternative to military force.” The discussion above implies that such assertions are unjustified. It suggests instead that the scholarly analysis of instruments of statecraft has far to go before anything can be said one way or the other as to the extent to which sanctions can be a reliable alternative to military force.