University of Southern California Dissertations and Theses

Economic sanctions, domestic audiences, and international conflict

(USC Thesis Other)

Economic sanctions, domestic audiences, and international conflict

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. U M I films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send U M I a complete manuscript and there are missing pages, these w ill be noted. Also, if unauthorized copyright material had to be removed, a note w ill indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact U M I directly to order. Bell & Howell Information and Leaming 300 North Zeeb Road, Ann Arbor, M l 48106-1346 USA UMI 800-521-0600 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ECONOMIC SANCTIONS, DOMESTIC AUDIENCES, AND INTERNATIONAL CONFUCT by Howard Evan Dick A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Political Economy and Public Policy) August 1999 Copyright 1999 Howard Evan Dick Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Num ber 9955054 Copyright 1999 by Dick, Howard Evan All rights reserved. UMI UMI Microform9955054 Copyright 2000 by Bell & Howell Information and Leaming Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. Bell & Howell Information and Leaming Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNIVERSITY OF SOUTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES, CALIFORNIA 90007 This dissertation, written by under the direction of hT.Â Dissertation Committee, and approved by all its members, has been presented to and accepted by The Graduate School, in partial fulfillment of re quirements for the degree of DOCTOR OF PHILOSOPHY Dean of Graduate Studies Date ...August ,4,_1999 DISSERTATION COMMITTEE . . . . . A k c ....... ........... Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS Abstract v Introduction vi-xiv Chapter I: Economic Sanctions as Tools of Coercion 1 - 28 Chapter 2: International Conflict and Domestic Audiences 29 - 67 Chapter 3: “Signaling” Sanctions: Sanctions as a Sign of 68 - 119 Future Escalation Chapter 4: Public Opinion and Conflict Escalation 120-254 Sanction Episodes (I) Introduction 120 - 126 (II) Case Study Analysis 126-241 (III) Statistical Analysis 242 - 254 Chapter 5: Conflict Management and Domestic Audiences 255 - 360 in the Balkans Chapter 6: Conclusion and Recommendations 361 -370 Selective Bibliography 37 1 - 383 u Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Case Study Analysis I. Angola 130-133 II. Burma 134-136 lU. Chile 137- 141 IV. China 142 - 148 V. Grenada 149 - 152 VI. Guatemala 153 - 155 Vn. Haiti (1987) 156 - 160 Vm. Haiti (1991) 161 - 168 IX. Iran 169 - 176 X. Iraq 177 - 185 XL Libya 186- 193 XII. Nicaragua (1977) 194- 196 XIII. Nicaragua (1979) 197 - 205 XIV. Panama 2 0 6 -2 1 4 XV. Paraguay 2 1 5 -2 1 8 XVI. Serbia 219 XVII. South Africa 220 - 229 XVIII. Suriname 230 - 233 XIX. Uganda 234 - 237 XX. Uruguay 238 -241 lU Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figures and Tables Figure 3.1: Extensive Form of the Game 79 Figure 3.2: Modeling the Case where d < w 86 Figure 3.3: Modeling the Case where d > z.w 86 Table 4.1: Cases Considered 125 Table 4.2: Case Study Analysis Outcomes 127 Table 4.3: Statistical Results on Generating 247 Domestic Audience Costs Table 4.4: Statistical Results on Causes of Conflict Escalation 253 Figure 5.1: Structure of the Yugoslavian National Government 264 Table 5.1: The Economy of the Former Yugoslavia 267 IV Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Abstract This dissertation investigates the relationship between economic sanctions, domestic audiences, and conflict escalation. The project argues that sanctions can be an effective tool of coercion when three conditions hold. First, sanctions must involve high self-imposed costs. Second, sanctions must signal to a target that the leadership in the sender will suffer a decline in public confidence if sanctions fail. Finally, as the domestic political costs of a failed policy accumulate, the sender escalates the conflict to avoid being perceived as “incompetent.” This research specifies under what conditions sanctions escalate to war and when they effectively coerce a target to backdown. The model also demonstrates that war is an equilibrium outcome even if the a priori expected benefit of conflict is less than the expected cost. Support for the argument is derived from four methodological approaches. First, a dynamic signaling game of two- sided incomplete information shows under what conditions a sanction will work and when it will escalate to war. Second, an analysis of an exhaustive set of 20 sanction episodes drawn from a data set devised by Gary Hufauer, Jeffrey Schott, and Kimberly Ann Elliott provides empirical validity. Third, regression analysis furnishes additional support by showing that self-imposed economic costs via a sanction are correlated with a decline in presidential public approval on the handling of a specific foreign policy, which in turn is correlated with an empirical observation of escalation. Finally, an entire chapter is devoted to international intervention in the Balkans from 1987 to 1995. The cases considered are: Angola ( 1985-1990), Burma ( 1988-current). Chile ( 1970- 1973), China ( 1989), Grenada (1983), Guatemala ( 1977-1983), Haiti ( 1987-1989), Haiti (1991-1994), Iran (1979-1981), Iraq (1990-1991), Libya (1979-1986), Nicaragua ( 1977- 1979), Nicaragua (1981-1990), Panama ( 1987-1989), Paraguay ( 1977-1981). Serbia ( 1987-1995). South Africa (1985-1991), Suriname ( 1982-1990), Uganda ( 1972-1979). and Uruguay ( 1976-1981). The empirical results are based on chronologies of events established from both an analysis of television network news coverage of each episode and an evaluation of over 2000 public opinion polling questions regarding these 20 cases. An organized collection of the polling data, media based chronologies, and televised reports can be obtained from the author. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Introduction fVhv a New Study o f Sanctions? Despite the increasing use of economic sanctions by the United States to advance foreign policy interests, few scholars have throughly examined why sanctions may be able to achieve difficult foreign policy goals, such as destabilizing a government, ending a military campaign, or inducing a significant change in policy. The dearth of thoughtful analysis is disturbing for four reasons. First, it is widely recognized that sanctions inflict disproportionate harm on the poor. ‘ For example, a study by the Public Health School at Harvard University found that the embargo against Haiti “contributed to as many as 100,000 new cases of moderate to severe malnutrition,” as well as contributed to an increase of 1,000 children under the age of 5 dying every month above expected mortality rates. A similar Harvard Study Team in Iraq estimated that sanctions contributed to the deaths of 170,000 Iraqi children in 1991 alone, as well as increased malnutrition and the spread of disease. Numerous studies suggest that more Iraqi civilians have died as a result of sanctions than were killed during the Gulf War.^ Thus from a moral perspective, it may be unjust for America to impose sanctions 'Hufbauer, Gary, Jeffrey Schott, and Kimberly Ann Elliot, Economic Sanctions Reconsidered (Washington D.C.: Institute for International Economics, 1990). Weiss, Thomas, David Cortright, George A. Lopez, and Larry Minear ed. Political Gain and Civilian Pain (New York, NY: Rowman and Littleffeld Publishers, Inc., 1997). • al-Samarrai, T., “Chapter 11: Sanctions against Iraq,” in ed. Cortright, David and George Lopez ed. Economic Sanctions: Panacea or Peacebuilding in a Post-Cold War World?. (Boulder, CO: Westview Press, 1995). Dreze, Jean and Haris Gazdar, “Hunger and Poverty in Iraq, 1991,” World Development, 20:7 (1992) 921-945. Harvard Study Team, “Harvard Study Team Report: Public Health in Iraq after the Gulf War,” (Harvard University Press: Cambridge, MA: 1991). "Sanctions in Haiti: Crisis in Humanitarian Action” (Cambridge, MA: Harvard Center for Population and Development Studies) 1993. “Sanctions Killing 1,000 Haitian Children a Month,” Los Angeles 77/ney November 9, 1993. vi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. if the probability of realizing positive benefits is low. Second, sanctions involve costs not only on the target, but affect economic conditions in the sender state as well. For example, U.S. farmers were harmed more by the 1979-1981 grain embargo than the Soviet Union. Research by Gary Hufbauer suggests that in 1995 alone, U.S. sanctions cost the U.S. economy between $15-$ 19 billion and 200,000 jobs.^ Third, sanctions may not be the optimal means to achieve one's long-term objective. Before imposing sanctions, scholars must rigorously show that sanctions can work. Finally, it is only by explaining how sanctions work can scholars help policymakers more effectively impose this deadly instrument. Thus, one motivation of this research project is to contribute on improving the real-world application of economic sanctions. The major argument in the dissertation is that economic sanctions are most likely to work when accompanied with the credible threat of conflict escalation. But to make the escalation threat credible, sanctions must generate domestic audience costs. Specifically, if an executive imposes a sanction and the sanction fails, the domestic population must “punish” the executive with lower public approval ratings in his handling of that situation. If public opinion falls sufficiently, an executive may try to reestablish his reputation domestically by assuming a more hawkish position. This results in conflict escalation beyond that originally desired by the sender leadership. Most scholars believe that sanctions can act as a substitute for war, particularly when the national interest o f the sender is not at stake. The converse argument is presented ’ Hufbauer, Gary, Kimberly Ann Elliott, Tess Cyrus, Elizabeth Winston, “U.S. Economic Sanctions: Their Impact on Trade, Jobs, and Wages,” Institute fo r International Economics Working Paper, 1997. vii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. here. Sanctions work by credibly threatening escalation. Thus, even if it is not in the national interest of the sender to use military force, if the sender elects to impose sanctions, the sender should be prepared to escalate in the event sanctions fail. The model developed shows that two conditions are necessary, although not sufficient, for sanctions to achieve difficult foreign policy objectives: when sanctions are used to generate domestic audience costs, and when domestic audience costs are large enough to cause the government to prefer escalation over backing down. This research therefore fills an important gap in the literature by developing, and empirically testing, a model of how sanctions can be used as an effective tool of coercion. In the process of developing a more accurate model explaining under what conditions sanctions work, the dissertation makes three additional contributions to the academic literature. First, the dissertation presents the first empirical support for the controversial “domestic audience cost” proposition. This proposition, as first developed by Thomas Schelling in the 1960s and more formally modeled by James Fearon in the 1990s. argues that a self-imposed cost can signal to a target that a sender is “serious” about the target changing her behavior. If the target does not change, the sender’s domestic leadership will suffer a “domestic audience cost,” specifically defined as a loss in public confidence. The larger the self-imposed cost, the greater is the domestic audience cost suffered if a policy fails. As domestic audience costs build, the home government may choose to escalate the conflict in the hopes o f reestablishing V lll Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. public confidence/ Second, by combining models fi’ om the two-level games literature with the conflict escalation literature, the dissertation adds to the growing literature that formalizes a relationship between domestic and international politics. Finally, the dissertation contributes to the understanding of how economic sanctions work within a comprehensive conflict management strategy. Moving slightly away from the formal model of Chapter 3, an in-depth study of conflict management in the Balkans from 1987-1995 shows how sanctions, in conjunction with other tools of conflict management, can coerce a target to accede to the wishes of the international community. The Arsiiment o f the Dissertation: A Brief Overview This project argues that economic sanctions can signal to a foreign government that the sender may be willing to escalate a dispute unless the target changes her behavior. The intuition behind this statement is straight forward. By imposing economic sanctions, the sender government suffers two types of costs. The first is the economic cost associated with the loss of trade. Standard economic models demonstrate that a reduction in international transactions reduces both consumer surplus and social welfare. Export-competiting industries also suffer declines in profits if they ^ Fearon, James D., "Domestic Political Audiences and the Escalation of International Disputes," American Political Science Review, 88:3 (1994) 577-592. Fearon, James D., "Signaling versus the Balance of Power and Interests," Joarna/ o f Cor^ict Resolution, 38:2 (1994) 236-269. Fearon, James D., Threats to use Force: The Role o f Costly Signals in International Crises, Ph.D. Dissertation, University of California at Berkeley, 1992. Shelling, Thomas, The Strategy of Conflict, (Cambridge, MA: Harvard University Press, I960). IX Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. are unable to sell their products abroad/ There is also the opportunity cost of ftiture business ventures. Unless an international dispute is resolved, these type of economic costs accumulate. Eventually, the economic costs may become so large as to encourage interest groups to lobby the sender government to solve the conflict, either by removing the sanction or using force to achieve the goal. The second type of cost is political. If economic sanctions fail, the public will question the "competency" of their government, even if the public, as consumers, do not experience any decline in welfare. Public officials will then be under more pressure to find a resolution to the conflict. If the relative military capability favors the sender over the target, the sender government may authorize the use of force if the cost of war is cheaper than the benefit of winning a war plus the domestic political losses associated with maintaining the status quo. Chapter 3 formalizes the argument. Oreanization o f the Dissertation Support for the thesis is organized around six chapters. Chapter 1 reviews recent works in economic sanctions. The chapter notes that while most scholars find sanctions ineffective, policymakers are using sanctions ever more frequently. The chapter points out the deficiencies in the existing literature and argues that a new model is needed. Chapter 2 reviews various models of conflict escalation, paying particular attention to the work of mainstream rationalists and domestic sources o f war. Standard ’ Krugman, Paul International Economics: Theory and Practice (New York, NY: HarperCollins Publishers). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. rationalist models demonstrate the utility of including “relative military capability” and “level of interest” in the study on the causes of war. However, these models often neglect the role domestic politics play in a leader’s decision to escalate. Recent research on two-level games has started to correct for this deficiency.* These models show that domestic politics will influence the strategies policymakers play when choosing whether or not to exercise military force. One common theme transcending this literature is that presidents with falling or low public approval are more likely to invoke force than those with relatively high approval.^ Using the literature discussed in Chapters 1 and 2 as a theoretical foundation. Chapter 3 presents the formal model. A sequential game of two-sided incomplete information is constructed that explains the relationship between economic sanctions, domestic audiences, and military escalation. Specifically, a game-theoretic structure using “relative military capability,” “level of interest,” and “domestic audience costs” shows under what conditions sanctions are an effective signal of resolve, under what conditions sanctions are not likely to be successful, and under what conditions conflict escalation may arise if sanctions fail. The model developed shows that with complete " Bueno de Mesquita, Bruce and David Lalman, fFdr and Reason (New Haven, CT: Yale University Press, 1992). Bueno de Mesquita, Bruce and David Lalman, "Domestic Opposition and Foreign War," American Political Science Review 84 ( 1 990) 747-765. Pumam, Robert, "Diplomacy and Domestic Politics: The Logic of Two Level Games," International Organization 42:3 (1988) 427-460. ^Aldrich, John, John Sullivan, and Eugene Borgida, “Foreign Affairs and Issue Voting: Do Presidential Candidate ‘Waltz before a Blind Audience?’” American Political Science Review 83:1 ( 1989) 123-141. James, Patrick, and John Oneal, “The Influence of Domestic and International Politics on the fhesident’s Use of Force,” Journal o f Conflict Resolution 35:2 (1991) 307-332. Ostrom, Charles and Brian Job, “The President and the Political Use of Force,” American Political Science Review 80 (1986)541-563. XI Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. information, war can be avoided through the use of a relatively low cost sanction. However, in the presence of incomplete information, undesirable wars may result. While the model is relatively simple, its simplicity yields powerful and robust results. Empirical support for the model is presented in 2 chapters. Chapter 4 empirically tests the model by examining how changes in public opinion correspond to changes in empirical events using both case study and statistical analysis. The sanction episodes considered are: U.S. vs. Angola (1985-1990), U.S. vs. Burma (1988-current), U.S. vs. Chile (1970-1973), U.S. vs. China (1989-1990), U.S. vs. Grenada (1983), U.S. vs. Guatemala (1977-1983), U.S. vs. Haiti (1987-1990), U.S. vs. Haiti (1991-1994), U.S. vs. Iran (1979-1981), U.N. vs. Iraq (1990-1991), U.S. vs. Libya (1979-1986), U.S. vs. Nicaragua (1977-1979), U.S. vs. Nicaragua (1981-1990), U.S. vs. Panama (1987- 1989), U.S. vs. Paraguay (1977-1981), U.N. vs. Serbia (1992-1995), U.S. vs. South Africa (1985-1991), U.S. vs. Suriname (1982-1990), U.S. vs. Uganda (1972-1979), and U.S. vs. Uruguay ( 1977-1981 ). Chapter 5 takes an in-depth look at conflict management in the Balkans, starting from the revocation of Kosovo autonomy in 1987 to the signing of the Dayton Peace Accords in November 1995. The chapter demonstrates how economic sanctions, in conjunction with military force, can be a useful tool of coercion. The case study also supplies clear evidence that the model developed in Chapter 3 has empirical validity by finding that the decision of the U.S. and Europe to laimch NATO airstrikes was driven, in part, by rising domestic audience costs. Chapter 6 summarizes the study’s findings and offers policymakers guidance on how to make sanctions more effective. xii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Finally, the reader is forewarned that for purposes of this study, economic sanctions are defined as the denial of imports, exports, or financial transactions by a sender against a target with the goal of the sender to either destabilize the target government, ending a military conflict between the target and a third party, or increase human rights practices that would require a fundamental change in the structure of target government. The effect sanctions have on other goals, such as deterring terrorism or the development of weapons of mass destruction, is not considered. Acknowledements I am indebted to several people who have contributed to the development of this project. 1 would first like to thank my advisor. Dr. B. Peter Rosendorff, who has offered valuable guidance, suggestions, and comments. My other committee members. Dr. Jonathan Aronson and Dr. Angela James, made excellent recommendations on how to better focus my research. 1 would also like to thank Dr. Anton Lowenberg for his thoughtful ideas on how to improve Chapters 3 and 4. Dr. Gunnar Neilsson generously gave his time to enhance the argument in Chapter 5. The statistical analysis performed in Chapter 4 could not have occurred without the guidance of Dr. Eric Solberg and Dr. Andy Gill. Jim Schank provided invaluable assistance with the collection of data. Conference participants at the Westem Economics Association Annual meeting in Lake Tahoe in 1998, International Studies Association Aimual meeting in San Diego in 1997, and u s e . Department of Economics Seminar in 1998 provided an excellent sounding board for my ideas. O f course, any faults, omissions, or major structural deficiencies xiii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. are my responsibility for not having listened more closely to these advisors. My parents gave me unconditional love and support throughout my life, and especially so during my years in graduate school. Finally, I would like to thank my Gance, Denise Ann Friedman, for her understanding and love. I dedicate this work to her. XIV Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A nation that is boycotted is a nation that is in sight o f surrender. It does not cost a life outside the nation boycotted, but it brings a pressure upon the nation which, in my judgment, no modern nation could resist. Woodrow Wilson I f the government is overthrown, what is the use o f [foreign] loans and credits. Yuan Mu, spokesman for China's State Council on China’ s reaction to Western sanctions in 1989 Chapter 1 : Economic Sanctions as Tools of Coercion Introduction This chapter demonstrates the need for a new model of economic sanctions. Existing research does not consider sanctions in the context of a comprehensive conflict strategy. Nor does the existing literature address the relationship between sanctions and conflict escalation. The lack of scholarly research in this area is unfortunate. As Woodrow Wilson recognized 75 years ago, sanctions can be a powerful instrument in international politics. American lives were likely saved by using sanctions against Serbia between 1991 and 1995 as an alternative to applying military force. Sanctions against Haiti from 1991 to 1994 also enabled the United States to achieve its objectives without a single U.S. casualty. Conversely, an examination of recent deployments of U.S. force shows that sanctions have always preceded military conflict. Since 1989, U.S. troops have seen combat in the Balkans, Iraq, Panama, and Somalia. U.S. forces deployed to Haiti in 1994 avoided combat only because the Haitian government abdicated power just hours before a U.S. invasion. In each of these five cases, the United States had imposed sanctions prior to using force. 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Is this finding a mere coincidence, or is there a correlation between war and economic sanctions? Do sanctions “work,” or do sanctions increase the probability a conflict will escalate? These are the questions that the dissertation is most interested in answering. The existing “sanctions and statecraft” literature focuses on whether or not sanctions “work,” but ignore their role in conflict escalation. This is unfortunate given that sanctions never work in isolation. Once considered in the context of conflict managament, it is not a coincidence that recent applications of U.S. military force was preceded by sanctions. The Grow ins use o f Sanctions as a Technique o f Diplomacv The use of sanctions as a tool of U.S. foreign policy has increased in the past 50 years. Gary Hufbauer, Jeffrey Schott, and Kimberly Ann Elliott foimd that the United States imposed sanctions 75 times between 1945 and 1990. The United Nations Security Council imposed mandatory sanctions against only 2 coimtries in the same period— Rhodesia and South Africa. With the end of the Cold War, however, the use of sanctions has increased. The United Nations Security Council has imposed mandatory sanctions under Chapter VII of the United Nations Charter against 6 states between 1990 and 1992, and another 3 from 1992 to 1996.' From 1993 to 1996, the United States imposed 61 new sanctions targeting 35 coimtries at a tremendous cost to the U.S. ‘ Dashti-Gibson, Jaleh Michelle, Enhancing Enforcement: A Framework fo r Monitoring Multilateral Economic Sanctions. PhJD. Dissertation, Notre Dame University, Indiana, 1998. Hufbauer eL al. Economic Sanctions Reconsidered. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. economy.' According to a study by Gary Hufbauer, U.S. sanctions in 1995 cost the U.S. $ 15-19 billion dollars and approximately 200,000 jobs.^ Even multinational corporations have not been immune to feeling the effects of sanctions. A survey conducted by the European-American Business Council found that U.S. sanctions have impacted the business practices of 94% of U.S. owned companies in Europe while 58% have lost sales."* Yet despite both their high self-imposed costs and deadly effects on target states, most scholars find sanctions to be an ineffective tool o f coercion. This generates an interesting paradox: why has the United States increasing relied on sanctions, absorbing billions of dollars in self-imposed costs, when sanctions are not likely to be effective? Are Sanctions even Effective? The literature has argued that sanctions are not an effective tool of coercion for over 30 years. Johan Galtung concluded in 1967 from his research on sanctions against Rhodesia that “the probably of the effectiveness of economic sanctions is, generally negative.”* Margaret Doxey has argued since the early 1970s that “the record of international sanctions of a non-military kind...suggest that on their own they do not * “A Catalog of New U.S. Unilateral Economic Sanctions for Foreign Policy Purposes, 1993- 1996,” National Association of Manufactures Working Paper, March 1997. ’ Hufbauer, Gary, et. al. “U.S. Economic Sanctions: Their Impact on Trade, Jobs, and Wages.” ■ * Entous, Adam, “U.S. Sanctions Harm Companies, Cost Jobs— Survey,” Reuter News Wire Oct. 2, 1997. ’ Galtung, Johan, "On the Effects of International Economic Sanctions." World Politics 19:4 (1967) 378-416. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. succeed in drastically altering the foreign or domestic policy of the target.”^ Sidney Weintraub’s review o f 7 sanction episodes during the Lyndon B. Johnson administration found that sanctions never achieved their primary foreign policy goal/ Donald Losman’s 1979 analysis of sanctions against Rhodesia, Cuba, and Israel, claimed that sanctions are not effective.® Fredrik Hofinann argued that “when sanctions are used, the manifest goal will probably not be attained.”^ Finally, Klaus Knorr found that sanctions “miserably” failed in 18 out of 22 cases in the 1960s and 1970s.‘“ Following the disastrous experience of sanctions in the late 1970s and early 1980s, the belief among scholars that sanctions fail intensified. Judith Miller foimd from her research on U.S. sanctions against Cuba, Iran, the Soviet Union, and others that “much in recent history supports the popular assumption that while trade boycotts. ‘ Doxey, Margaret, International Sanctions in Contemporary Perspective (New York, NY: St. Martin's Press, 1987). Doxey, Margaret, Economic Sanctions and International Eirforcement (New York, NY: Oxford University Press, 1980). Doxey, Margaret, "Oil and Food as International Sanctions." IntemationalJoumal. 36:1 (Winter 1980-81) 311-334. ^ Weintraub, Sidney ed.. Economic Coercion and U.S. Foreign Policy (Boulder, CO: Westview Press, 1982). * Losman, Donald, International Economic Sanctions (Albuquerque, NM: University of New Mexico Press, 1979). ’ Hofmann, Fredrik, "The Functions of Economic Sanctions," Journal o f Peace Research 2 (1967) 140-159. Knorr, Klaus, The Power o f Nations (New York, NY: Basic Books, 1975). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. embargoes, and other economic sanctions are legitimate alternatives to military action, they rarely work.”'* Robert Paarlberg used the failed U.S. grain embargo against the Soviet Union as evidence that food sanctions "never work”.'" Throughout the 1980s, scholars continued to find that sanctions rarely are effective. An exhaustive study conducted by Gary Hufbauer, Jeffrey Schott, and Kimberly Ann Elliott of 116 episodes of economic sanctions found that sanctions were successful only 34% of the time.'^ The success rate on difficult foreign policy objectives, such as destabilizing a government, stopping a military campaign, or inducing major changes in target country policies was 33%, prompting the authors to note that “by themselves, sanctions are seldom able to roll back military aggression, have limited ability to impair a targeted regime, and never are able to topple a dictator.” A less sanguine outlook is found if one examines only the most recent sanction episodes. Using Hufbauer, Schott, and Elliott’s data set, out of the 35 times the United States imposed sanctions between 1978 and 1990, there were only 4 successful cases. From 1983 to 1990, only 1 successful case was recorded out of a possible 18.'^ Sanctions have not succeed in obtaining a difficult foreign policy objective since 1982. " Miller, Judith, "When Sanctions Worked," Foreign Policy 39 (Summer 1980) 118-129. '* Paarlberg, Robert, "Lessons of the Grain Embargo," F o re i^ Affairs 59 (Fall 1980) 144-162. " Hufbauer eL al. Economic Sanctions Reconsidered. Hvdbwst et.z!L, Economic Sanctions Reconsidered. The 1 successful case was against El Salvador in 1987 to coerce the government to keep in prison a man who murdered 4 U.S. nuns. The criminal was to be released as part of an amnesty program. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The scholastic community has grown even more wary of sanctions in the 1990s. The perceived failure of U.S. sanctions against Burma, China, Cuba, Haiti, Iran, Iraq, Libya, Sudan, and the Federal Republic of Yugoslav (Serbia and Montenegro) has led for calls for ending using sanctions as a coercive tool of American foreign policy.'^ Jeffirey Garten argues that “with the federal government’s ability to conduct foreign policy in a world preoccupied with economic stability and progress dwindling, and Washington having neither the people nor the money to exert the influence it once could,” the U.S. should no longer use sanctions as part of its foreign policy arsenal.'® Richard Haass writes that “the limitations of sanctions are more pronounced than their accomplishments. Sanctions alone are unlikely to achieve results if their aims are large or time is short.”'’ It is such thinking that resulted in the lifting of sanctions against Vietnam in 1994.'* The international community has grown even more cynical of sanctions over time. For example, the U.N. General Assembly annually passes a resolution condemning the U.S. for its sanction policy against Cuba and growing international opposition to sanctions against Iraq has prompted states such as Russia and France to sign, but not implement, multibillion dollar agreements with the rogue state. ” Pape, Robert A., “Why Economic Sanctions do not Work,” International Security 22 (Fall 1997) 90-136. “ Garten, Jeffrey, “Business and Foreign Policy,” Foreign Affairs 76:3 (May 1997) 1. Haass, Richard, “Sanctioning Madness,” F o re i^ Affairs 76:6 (November 1997) 74-85. " Vietnam’s decision to withdraw from Cambodia also contributed to the decision to lift sanctions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. However, while many expect sanctions to fail in achieving diffîcult foreign policy goals, their use is also expected to persist. Joseph Gavin notes that “even experienced policymakers who know about the limited power of sanctions will continue to impose them...despite their economic costs.”* '* At times it even appears that sanctions are being looked upon as an instrument of first resort when a crisis develops— an easy way out for governments anxious to act, and be seen to act by their domestic audiences, whether or not sanctions can achieve the objectives claimed for them.’° Thus, while many scholars expect sanctions to fail, the evidence suggests that the use of sanctions will increase over time. Solvins the Paradox: Current Explanations Why then are policymakers imposing sanctions more now than ever before given that the literature finds that sanctions usually fail and are costly both to the U.S. economy and U.S. diplomatic relations with non-participating countries? A review of the literature suggests three partial answers. The first, and most commonly cited, is that sanctions are successful. It is only that “success” has been incorrectly defined. Finding the goal of economic sanctions often differs then that found in the mainstream literature, David Baldwin argues that relative to other policy alternatives, sanctions have proven highly effective. For example, while most scholars ” Gavin, Joseph, “Economic Sanctions; Foreign Policy Levers or Signals?” Cato Institute Policy Analysis No. 124. November 7,1989.3 ^ Leyton-Brown, David ed.. The Utility o f International Economic Sanctions (London, UK: Croom Helm, 1987). 1. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. believe that the League of Nation’s sanctions against Italy in 1935 were designed to coerce Italy into withdrawing from Abyssinia, Baldwin writes that “the main argument for League action was to test the association’s capacity to stand against Adolf Hitler’s revisionism.”-' Baldwin also writes that sanctions was an optimal strategy relative to going to war or doing nothing. Stefanie Lenway likewise claims that the sanctions were designed to reassure the British and French public that aggression in Europe would not be tolerated, particularly by Germany.” Such differences of opinion regarding the objective of sanctions has lead scholars to different conclusions regarding their effectiveness, with some claiming that the oil embargo against Italy did deter Hitler in the short-run, while others believe that the failure of British and French resolve against Italy encouraged German militarism. The case study in Chapter 5 shows that if the primary objective of U.S. sanctions against Serbia was to end the war in the Balkans, as most scholars claim, then sanctions by themselves were not effective. However, if the objective was to delay Serbian military advances by increasing the cost of war, then sanctions were successful. A second explanation claims that even if sanctions fail to achieve their primary objective, sanctions can still yield other benefits. So-called “secondary” and “tertiary” goals include deterring other states from imitating the behavior of the target, reducing the ability of the target to cause mischief in other countries, increasing the cost of EdUdvna, Economic Statecraft, 157. ” Lenway, Stephanie, “Between Wars and Commerce: Economic Sanctions as a Tool of Statecraft,” International Organizations A\:A (1987). Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. defying international norms, and demonstrating that the sender government is committed to maintaining global stability. Such benefits may cause sanctions to be considered a ‘‘success.” For example, while Weintraub’s research shows that sanctions never obtain their primary goal, he did find that sanctions always accomplish secondary goals.^ An empirical study of sanctions in the 1960s and 1970s by James Barber finds similar results.'"* An inspection of recent sanction episodes adds further support. Sanctions may not have removed Saddam Hussein, but they are successful in hindering his development o f weapons of mass destruction. Sanctions also reduced money laundering through Panama, although Manuel Noriega retained political authority. Finally, senders may impose sanctions to signal moral disapproval of the target’s behavior even if sanctions are not expected to work. FC im Nossal argues, for example, that sanctions can constitute a form of “international punishment” more credibly than diplomatic statements.^ He compares international sanctions to that of a judicial system fining a criminal. Even if the fine may not be sufficient to induce rehabilitation, the international community is expected to impose some cost on the criminal for her behavior. Empirically, supporters of U.S. sanctions against South AJnca used this argument to encourage the U.S. Congress to pass the Comprehensive Anti-Apartheid ^ Weintraub, Sidney ed.. Economic Coercion and U.S. Foreign Policy. - ■ * Barber, James, “Economic Sanctions as a Policy Instrument.” International Affairs 55 (1979) 367-384. “ Nossal, Kim, "International Sanctions as International Punishment," International Organizations 43:2 (Spring 1989). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Act in 1986. Even if sanctions were not likely to work, or even if they were counterproductive, apartheid was considered to be such an egregious violation of human rights that for moral reasons alone, many Americans supported sanctions. Clearly, all three arguments hold some validity. They also help clarify why policymakers may impose economic sanctions despite their perceived failure. However, each argument still leaves unanswered the question of why policymakers promise the public that sanctions will realize goals sanctions alone can not achieve. There may be differences of opinion over the primary objective of sanction, but such differences are rare. Using sanctions as a tool for “moral outrage” only makes sense if they are going to work, otherwise, sanctions should be considered immoral because of the harm they inflict on the poor and weak.*® Finally, even if sanctions do fail at times, other times sanctions do work. Why they work, and by what standard one should evaluate the effectiveness of sanctions is under considerable debate within the literature. Models explaining why economic sanctions can “work” are needed to explain why they are imposed, as well as proscribe how sanctions can be more effectively applied. Models ofHow Sanctions Can fVork The mainstream literature answers the paradox for why sanctions are imposed when they are not expected to work by suggesting that sanctions can work under the proper conditions. A review of the literature shows, however, that the three competing “ Patterson, Jack, “The Political and Moral Appropriateness of Sanctions,” in ed. Cortright and Lopez ed. Economic Sanctions: Panacea or Peacebuilding in a Post-Cold War WorleP.. Weiss, ed. Political Gain and Civilian Pain: Humanitarian Impacts o f Economic Sanctions, 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. approaches lack theoretical and empirical validity. Nor do these three approaches explain how sanctions can be employed in the context of conflict escalation. The three competing models are considered below according to how sanctions punish the target: “collective punishment,” “punishing elites,” or “punishing non-elites.” "Collateral ” Sanctions: Makins Sanctions Effective by Increasins the Total Cost There is a belief among scholars that for economic sanctions to be effective, tremendous economic pressures must be inflicted on the target. The logic behind these models is relatively simple: if the cost of sanctions exceed the benefit of the objectionable policy, target governments will change their policy so that sanctions are lifted. Peter A.G. van Bergeijk has called this the “donkey psychology” approach to the study of sanctions. Just like a donkey, it is assumed that governments can be induced to move in the right direction by means of a stick and a carrot, that is to say, by positive and negative sanctions."’ It does not matter where the donkey is hit, so long as it is hit hard. These models ignore the income distribution effects of sanctions, as well as how sanctions affect the internal political dynamic in the target state— an unfortunate omission given that anyone who has been around a donkey knows that hitting it in the tail may cause the donkey to run in the other direction. But because these are the most common models within the literature, a close inspection is required. ^ van Bergeijk, Peter A.G., Economic Diplomacy, Trade, and Commercial Policy: Positive and Negative Sanctions in a New World Order (New York, NY: E. Elgar, 1994) 16. 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The common thread through this strain of the literature claims that sanctions are most effective when they inflict high marginal costs on the target relative to the marginal gains realized from engaging in the objectionable behavior. These models usually involve two actors (sender, target), two policies (sender sanction, target behavior), and a dichotomoiK outcome (sanctions work, sanctions do not work). Jonathan Eaton and Maxim Engers build a game-theoretic model using these basic assumptions to show that sanctions are most likely to work when they involve high costs, and least likely when they inflict minimal damage."* George Tsebelis finds similar results in his own game-theoretic approach."’ Peter A.G. van Bergeijk finds that the cumulative effect of sanctions will coerce targets to change because the cost of sanctions increase over time relative to the benefit of the target’s behavior.^’ Dean Lacy and Emerson Nlou find that sanctions work under two conditions (1) the target holds nonseparable preferences, such that the cost of sanctions can be weighed against the gain the target derives from her behavior, and (2) marginal costs exceed benefits.’ ’' “ Eaton, Jonathan and Maxim Engers, "Sanctions," Journal o f Political Economy 100:5 (October 1992) 899-928. ^ Tsebelis, George, "Are Sanctions Effective: A Game Theory Analysis," Journal o f Conflict Resolution34:\ (March 1990) 3-28. van Bergeijk, Peter A. G. and Charles van Marrewijk, "Why do Sanctions need Time to Work," Economic Modelling 12:2 (1995) 75-86. " Lacy, Dean and Emerson Niou, “Nonseparable Preferences, Issue Linkage, and Economic Sanctions,” paper presented at the Annual Meeting of the American Political Science Association, Boston September 3-6, 1998. 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Case studies employing the “collective punishment” approach yield identical results. Richard Porter writes that “the goal of sanctions is clear and simple: to impose a reduction in economic welfare on the target country and thereby reduce its willingness to persist in antagonizing the world community.”^- Zachary Seldon’s research of 20 sanction episodes finds that sanctions are most likely to work in cases where sanctions cause the greatest amount of collateral damage. Finally, Hufbauer, Schott, and Elliott find that “the average cost to the target for all successful cases was 2.4% of GNP; by contrast, failed episodes barely dented the economy of the target country, with costs averaging only 1 percent of GNP.”^ ^ It is therefore not surprising that much of the literature proscribes for better monitoring of sanctions and making sanctions multilateral to enhance their effectiveness.^"* The problem with these models is that while collateral damage may be a necessary condition for success, it is not sufficient. The most recent comprehensive sanctions imposed— against Haiti, Iraq, Panama, and Serbia— decreased target GDP by over 50% without achieving their objective. An analysis of the Hufbauer data set shows that since 1982, there has been no correlation between the cost of sanctions and Porter, Richard, "International Trade and Investment Sanctions," Journal o f Conflict Resolution 23:4 (1979) 581. ” Hufbauer et. & [ . Economic Sanctions Reconsidered, 102. ” Dashti-Gibson, Jaleh Michelle, Enhancing Enforcement: A Framework fo r Monitoring Multilateral Economic Sanctions. Martin, Lisa, Coercive Cooperation: Explaining Multilateral Economic Sanctions Princeton University Press, 1992). 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. their effectiveness, as the one successful episode cost the target less than $ 1 million. Over the entire data set, a covariance of only .17 was found between the success score and annual cost of sanctions as a percentage of target GDP.^^ Besides its lack of empirical validity, this part of the literature also lacks a theoretical foundation to explain why collateral damage leads to the target changing her behavior. The “donkey psychology" is only valid if it is the donkey getting hit. However, by definition, collateral sanctions harm different segments of the target population in different proportions. There is no reason to believe that hitting a mule would encourage the donkey to move. For example, if the political leadership is able to benefit from sanctions, it is likely that sanctions will prove coimterproductive. The next two sections consider the income distribution effects of sanctions. It is interesting to note, however, that in non-sanction episodes of transition from authoritarian governments to democratic governments, there is a strong association between economic distress and destabilization. Stephan Haggard’s analysis of 27 democratic transitions in Africa, Asia, and Latin America since the mid-1970s finds that in 21 cases, the country experienced negative economic growth the previous year. His work also shows that both the military and business class were active in seeking a " Hufbauer et. al. use a scoring system to assess the success rate of over 60 sanction episodes. These authors rate both the political outcome and sanction’s contribution to the political outcome on a scale from I (highly unsuccessful) to 4 (highly successful). They then multiply the two together. If the result is greater than 9, then sanctions are found to have been successful. Any number lower than 9 is considered an unsuccessful application. The economic damage done by sanctions in each of their cases. The correlation figure of 0.17 measures the relationship between their “success score” and “cost as a percentage of GDP.” 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. transition as a result of the country’s economic difBculties/° Thus, while declines in national income are correlated with changes in government, there must be something unique about sanctions. I believe that sanction’s income distribution effects along with the political environment in which sanctions are imposed explains the discrepancy. "Smart " Sanctions: Makins Sanctions more Effective by Tarsetins Elites Because political behavior is conditioned, in part, on the economic rewards of that behavior, accounting for who is harmed by sanctions is a necessary exercise. Some people exert more political power than others, and targeting the wrong interest group will result in sanctions failing. It is for this reason that many scholars have advocated imposing so-called “smart” sanctions targeting only political elites in the target. A public-choice model developed by William Kaempfer and Anton Lowenberg suggests that sanctions may change social, political, and moral norms sufficiently to induce a target to remove the objectionable policy if the change results in public demand for abondoning the policy to be greater than demand for the status-quo, holding constant the relative political power of the competing interest groups. Specifically, they model the target’s decision-making process in the context of a simple “demand” and “supply” diagram framework. The “demand” curve shows the price supporters of the objectionable behavior are willing to pay for a marginal increase in the objectional behavior, holding constant the political power of that interest group. The “supply” Haggard, Stephan and Robert Kaufinan, The Political Economy o f Democratic Transitions (Princeton, NJ: Princeton University Press, 1995). 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. curve represents how much opponents are willing to pay to remove the policy, also keeping constant political power.^^ Graphically, the “quantity” of target policy is on the x-axis and “price” is on the y-axis. If sanctions change the marginal benefit and cost interest groups receive from the behavior, the demand and supply curves shift and a new equilibrium forms. According to Kaempfer and Lowenberg, sanctions are most effective if they increase the cost of maintaining the objectionable policy on groups supporting the government. The logic behind the Kaempfer and Lowenberg approach makes intuitive sense. It is when those with political control are directly affected by sanctions that sanctions are most likely to work. Research by Clifton Morgan and Valerie Schwebach finds that sanctions will succeed only when the ruling class is directly harmed.^* A case study analysis by Makio Miyagawa shows that sanctions are not likely to succeed if the ruling class escapes their economic effects.^^ It is from such an analysis that the United States elected to freeze the personal financial accounts of political leaders in both Haiti and Serbia. The United Nations Security Council has also adopted resolutions targeting ^ Kaempfer, William and Anton Lowenberg, International Economic Sanctions: a Public Choice Perspective (Boulder, CO; Westview, 1992). " Morgan, T. Clifton and Valerie Schwebach, "Fools Suffer Gladly: The Use of Economic Sanctions in International Crises," International Studies Quarterly ( 1987) 27-50. ” Miyagawa, Makio. Do Economic Sanctions fVorkH (New York, NY: St. Martin's Press, 1992). Morgan, T. Clifton and Valerie Schwebach, "Fools Suffer Gladly: The Use of Economic Sanctions in International Crises," International Studies Quarterly (1987) 27-50. 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. specific individuals in Libya for sanctions. Clearly, if individual leaders find that the cost o f maintaining the government’s behavior entails personal financial loss, sanctions have a greater probability of inducing those leaders to change the government’s policies. An insightful model by B. Peter Rosendorff shows why targeting the ruling class may achieve difficult goals in a more general setting. He suggests that sanctions work by flattening the income distribution curve; thus making it too costly for the target government to maintain the objectionable policy, even in cases where the government would be required to give up all power for sanctions to be lifted.'*® Specifically, Rosendorff argues that sanctions that target elites work through two mechanisms. First, when sanctions harm elites more relative to the opposition, then the opportunity cost of relinquishing political power falls. This increases the incentive for elites to capitulate to sanctions. Second, when income distribution flattens, tax revenues fall, and the target government becomes less capable of financially supporting repressive arms of the state. Once these institutions are weakened, it is believed that the government can not withstand public demands to end the objectionable policies. Rosendorff supports his argument by analyzing the case of South Afiica. Rosendorff, B. Peter, "A Theory on Transition to Democratic Rule, " unpublished paper, August 1996. 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. While the “smart” sanction approach to success is intuitive, such models are deficient both on empirical and theoretical grounds. Sanctions rarely, if ever, harm elites relatively more than those in the opposition."*' It is the poor who suffer the most from sanctions, either because the government adopts legislation shifting the burden of sanctions to the poor, or because elites engage in “sanction-busting,” importing goods illegally and then selling those goods to common citizens at a premium. For example, a report from the General Accounting Office finds that sanctions are a blunt instrument that can not differentiate between population cohorts, despite all intentions otherwise.*' David Rowe’s research on U.N. sanctions against Rhodesia finds that the government transferred income from blacks to whites as one means to maintain domestic support.*^ Saddam Hussein has relied upon “sanction-busting” to generate money to support coercive arms of the state. The bombing of Srebrenica by Bosnian Serbs was financed, in part, by poor Serbs and Bosnian Muslim who paid warlords exorbitant prices for sanctioned commodities. The Haitian military smuggled goods into the country once the business elite were unable to financially support the government. It is therefore not surprising that although Morgan and Schwebach believe sanctions will only work by targeting elites, they also claim that this rarely happens. ■ * ' Carter, Barry E., International Economic Sanctions: improving the Haphazard U.S. Legal Regime. (Cambridge, MA: Cambridge University Press, 1988). Cortright, David and George Lopez ed. Economic Sanctions: Panacea or Peacebuilding in a Post-Cold War Worlefl. Scharfen, John. The Dismal Battlefield: Mobil tin g fo r Economic Conflict. (Baltimore, MD: Naval Institute Press, 1995). General Accounting Office, Do Economic Sanctions (Washington D.C., General Accounting Office, 1992). ■ ” Rowe, David, Surviving Economic Coercion: Rhodesia’ s Response to International Economic Sanctions. Ph.D. Dissertation, Duke University. 1993. 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Theoretical models of trade lead to the same result: sanctions should harm the poor relatively more than the rich. Using a simple Heckscher-Ohiin trade model, it has been shown that when a developing country trades with a developed country, low-skill workers gain relatively more than high-skill workers in the developing country.'” Because most senders are developed states, targets are developing states, and supporters of the government are high-skilled labor, standard trade theory predicts that a reduction in trade should harm the poor relatively more than the rich.'*^ Thus, comprehensive sanctions should not be expected to work through their effects on elite income. "Dumb ” Sanctions: Making Sanctions more Effective by Tarsetins non-Elites A third model predicts that sanctions can be successful in achieving difficult goals, especially destabilizing governments, if sanctions encourage the masses to mobilize against their leadership. For sanctions to be work in this model, two conditions must be satisfied: relative deprivation must occur and political opportunities must be expanded.” * ® * * Kraegar, Anne. O. ed. Trade and Employment in Developing Countries (Chicago, IL: University of Chicago E gress, 1981-1983). Wood, Adrian iV or/A-SdwrA Trade, Employment, and Inequality (New York, NY: Oxford University Press, 1994). Kaempfer, William and Anton Lowenberg, International Economic Sanctions : a Public Choice Perspective. Dick, Howard, “Does the Success of Sanctions Depend on Relative Deprivation?” unpublished paper. I would like to thank Dr. John S. Odell and participants at a University of Southern California School of international Relations seminar for helpful comments with respect to this part of the dissertation. 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sanctions must first cause relative deprivation, defined as sharp declines in income causing citizens to believe there is “a perceived gap between their value expectations and their capabilities, that is, a discrepancy between what people think they ought to get from society and what they believe they will actually chtain.”^ ’ If this occiurs, interest groups will form to pressure the target govermnent to change her policy. Under certain conditions, if peaceful methods of protest fail, opposition movements will have an incentives to use violence. Given that peaceful forms of protest are rarely successful in states targeted by sanctions, relative deprivation must be severe enough to cause the opposition to employ more violent methods of coercion. The literature provides both theoretical and empirical support to the argument that relative deprivation can lead to mass violence. Ted Gurr first proposed the link between relative deprivation and violent rebellion in his seminal work, ffTiy Men Rebel. Gurr argues that as individuals feel worse-off relative to what they believe they deserve, individuals are likely to organize and demand that the government be replaced. However, because governments in developing countries are not directly responsible to the people, individuals will use force to replace the government."'* Empirical evidence supporting this hypothesis is strong. Erich Weede’s analysis of 52 countries, correlating average income to domestic violence, finds that as income fail, opposition movements grow in strength and are more likely to use Weede, Erich. "Income Inequality, Average Income, and Domestic Violence,” Journal o f Conflict Resolution 25:4 ( 1 981 ). * * Gurr, Ted Robert. Why Men Rebel (Princeton, NJ: Princeton University Press, 1970). 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. violence/^ Lee Sigelman and Miles Simpson show in a time series, cross-national study of 135 states that an increase in income inequality is positively associated with domestic protests/" Kurt Schock’s analysis of 62 states discovers that “economic inequality tends to promote violent political conflict; however, the extent to which it promotes violence is enhanced or constrained by the political opportunity structure.”^' Thus, if sanctions can increase the gini coefficient, either by making the poor poorer or causing economic hardship on the middle class, sanctions may work by causing relative deprivation and encouraging citizens to use violent means to replace their government. Unfortunately for sanction supporters, a growth in the number of individuals willing to use force to destabilize a government is not sufficient. For the opposition to succeed, sanctions must also either weaken existing political institutions or there must have been significant political decentralization prior to sanctions. Theoretical and empirical evidence shows that political opportunities determine the probability groups antagonistic to the status quo prevail, where political opportunities are defined as “the degree to which groups are likely to be able to gain access to power or manipulate the political system.”^ Without such access, opposition movements are not able to force reforms or replace their government. Democratic elections, the right to assemble in Weede, Erich. "Income Inequality, Average Income, and Domestic Violence," Journal o f Conflict Resolution 25:4 ( 1981 ) 639-653. Sigeiman, Lee and Miles Simpson, "A Cross-National Test of the Linkage between Economic Inequality and Political Violence," Journal o f Conflict Resolution 21:1 (1977) 105-125. Schock, Kurt, "A Conjunction Mode! of Political Conflict," Journal ofCortflict Resolution 40:1 (1990) 98-133. Tilly, Charles, From M obiltation to Revolution (Boston, MA: Addison-Wesley, 1978). 21 Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission. protest, freedom of the press, and the presence of semi-official institutions representing the interest of political opponents within the government, are therefore necessary conditions for opposition movements to be effective. Thus, even if sanctions increase financial incentives for individuals to join opposition movements, without the ability to politically participate, opposition groups may not succeed through peaceful means. In states where governments restrict opportunities, the only option available is using violence.^^ This explains the need to support military coups or armed opposition groups for sanctions to have a positive effect. It is interesting to note that pre-sanction characteristics of the target influence the degree to which opposition groups can voice their dissent. Politically weak and structurally decentralized states are less likely to be able to use sanctions as a means to counter the opposition than authoritarian states. Conversely, the more authoritarian the government, the easier it is to use sanctions as an excuse to suppress dissenters. Given that most targets have authoritarian governments, the lack of political opportunities reduces the probability sanctions will encourage an uprising by the masses. Paradoxically, while sanctions may be effective in generating relative deprivation, they may be counterproductive by worsening political opportunities. Governments influence the power of opposition groups through three mechanisms: propaganda, legislation, and coercion. In each case, sanctions may increase the target government’s ability to control society. ” Huntington, Samuel P., The Third Wave (Norman, OK: University of Oklahoma Press, 1991). 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Propaganda has proved effective in convincing domestic constituents that sanctions, and not the government, are responsible for deteriorating economic and political conditions. For even if there is relative deprivation, if people do not believe that it is their government that is the culprit, but alternatively the sender, then opposition movements will not grow in strength. States that do not control all media outlets often censor the press following the imposition of sanctions so as to both shut down the opposition media and portray the foreign country as the “true” cause of the people’s economic hardship. There is also empirical evidence showing that propaganda can cause a “rally-around” effect when a population perceives a foreign threat.It is therefore not surprising that whenever sanctions are imposed, target governments try to foster nationalist sentiments. Circumstantial evidence suggests that the more authoritarian the government, the more likely sanctions spark a successful propaganda campaign. Serbian President Slobodan Milosevic shut down the opposition press, and used the fact that sanctions applied to Serbia, but not Croatia, as evidence that the West was targeting Serbia for economic and political destruction during the Serbia-Bosnian war. Saddam Hussein and Fidel Castro have likewise convinced many of their citizens that their economic distress is due to harsh “economic warfare.” Conversely, in the relatively more open states of Panama and South Africa, opposition movements had better access to media outlets to argue that sanctions signaled that the world community supported their cause. ^ Jowett, Garth and Victoria O’ Donnell, Propaganda and Persuasion (New York, NY : Sage Publications, 1986). 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Legislation is the second means target governments use to control opposition movements. When sanctions are imposed, targets often adopt domestic growth plans that call for an increase in government spending, domestic investment, and domestic production. With government now a greater share of GDP, labor becomes more dependent upon the state, and hence, less willing to challenge government authority. For example, Serbian state employees were found to be the least willing to protest against their government— not surprising since it was the government that provided these people their paychecks. With more people now beholden to the government given its increased role in economic life, there is a decrease in the probability that sanctions will work. By its very nature, authoritarian states have more freedom to adopt such legislative practices relative to democratic systems. Finally, even if sanctions cause relative deprivation, and increase the number of individuals wanting to join opposition movements, authoritarian governments can be expected to use repression to control dissent. Semi-democratic governments are less capable of applying such coercion. Data collected for this project show that all target governments devoted an increasing percentage of GDP to national defense while exposed to costly international sanctions.^* Sanctions may also increase the incentives of the target government to justify the use of repression. ” Sivard, Ruth Leger, World iVfilitary and Social Expenditures, (Leesburg, VA: WMSE Publications, various years). 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Thus, consideration of the type of target government the sender confronts is necessary to determine if sanctions will work. Scholars have found that effective protests are most likely when coercion is moderate, and least likely when coercion is severe. However, extreme coercion can either cause a surge in non-peaceful protests or make the situation so bleak that dissenters do not even try to resist government authority/^ In cases where the objective of the sender is either to stop a war or destabilize a government, unless violent measures are taken by the opposition or the international community, sanctions alone will not succeed. Sanctions against more authoritarian regimes must therefore be accompanied by massive civil disobedience or civil war. But because insurrections rarely, if ever, topple existing regimes, “tightening” sanctions against an authoritarian society often results only in greater poverty and misery on innocent civilians without accomplishing sender objections. Samuel Huntington observes that since 1974. “with the partial and debatable exception of Nicaragua, no authoritarian regime was brought down by a prolonged guerrilla insurgency or civil war.”^ ’ Thus, serious senders should accompany sanctions with aid to opposition groups so that they can more aggressively challenge government repression. Otherwise, sender government’s should not be surprised when a sanction fails to achieve a difficult foreign policy objective. “ Francisco, Ronald, "The Relationship between Coercion and Protest." Journal o f Conflict Resolution 39:2 (1995) 263-281. Gupta, Dipak and Harinder Singh, "Government Coercion of DissiàsnXs," Journal o f Corflict Resolution 37:2(1993) 301-337. Lichbach, Mark Irving, "Deterrence or Escalation: the Puzzle of Aggregate Studies of Repression and Dissent," Journal o f Corflict Resolution 31:2(1987)266-297. ” Huntington, The Third fPave, 164. 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. As an Alternative to War The fact that sanctions may work by creating incentives for internal violence is interesting given that most sanction scholars argue that sanctions are a morally superior alternative to the use of military force. Political leaders have a limited number of tools to select from when trying to achieve a difficult foreign policy goal-do nothing, file diplomatic protests, impose sanctions, or apply military force. If military force is too costly and diplomatic measures not costly enought, the mainstream literature cites economic sanctions as an ideal option. The National Conference of Catholic Bishops summarizes the thoughts of most scholars by arguing that “sanctions can offer a nonmilitary alternative to the terrible options of war or indifference when confronted with aggression or injustice.” Klaus FCnorr writes that “after WWII, a number of strategists developed the idea...that economic war could largely, if not wholly, replace the bloody business of military ground combat as a means to victory.”^ ® Hufbauer. Schott, and Elliott claim that if sanctions are given enough time, military force is not necessary. Robin Renwick finds that political leaders in the sender can use sanctions instead of war to achieve a foreign policy objective.” David Cortright and George Lopez write that “economic sanctions serve as a less-destructive alternative [to war] generally, and are less likely to provoke a dangerous military counterstrike.”® ® " Knorr, Klaus, The Power o f Nations (New York, NY: Basic Books, 1975) 137. ” Renwick, Robin, Economic Sanctions (Cambridge, MA: Harvard University Press, 1981). “ Cortright, David and George Lopez ed. Economic Sanctions: Panacea or Peacebuilding in a Post-Cold War W orlcP. 6. 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. But whether sanctions are, in fact, a substitute for war is an open question. As a tool of foreign policy, sanctions alone have not realized any major foreign policy goals without being accompanied by military force. Haitian military leaders abdicated power as U.S. troops were airborne; Ian Smith’s government in Rhodesia capitulated only after a civil war; Manuel Noriega in Panama was removed through U.S. military intervention; Saddam Hussein was forced out of Kuwait after Desert Storm; and Bosnian Serbs agreed to the Dayton Accords only when NATO airstrikes and a stalemate emerged between Serbian and Croat-Bosnian forces. Finally, the "dumb” sanction approach requires that conflict escalation occurs for sanctions to work. Conclusion The current literature leaves us asking two questions. First, "why are policymakers increasingly using economic sanctions as a tool of coercion in seeking significant behavioral changes in the target given the growing perception that sanctions fail?” and second, “how are sanctions suppose to work?” With respect to the first question, the literature provides one of three answers: sanctions as a tool of coercion are underappreciated, secondary and tertiary goals of the sender can be realized even if the primary objective is not attainable, and the moral outrage sanctions signal is sufficient justification for their imposition. These reasons, however, can not justify sanctions harmful to both the target and sender. Answers to the second question also take one of three forms: sanctions work by imposing collateral damage, targeting elites, or targeting the opposition. Each, however, lack theoreticai and empirical rigor. 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. It is for these reasons that a new model is necessary. It is my hypothesis that economic sanctions work by threatening conflict escalation. After a review of the relevant conflict literature in Chapter 2, a model is developed in Chapter 3 showing that sanctions may work by signaling that the sender will escalate the conflict unless the target accedes to the demands of the sender, even if the cost of escalation is greater than the benefit of realizing one’s objective. It is the threat of liigh future costs, rather than current costs, that explain how sanctions can work. In other words, rather than acting as a substitute for war, economic sanctions complement war. 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. You may not be interested in war, but war is interested in you. Leon Trotsky Chapter 2 : International Conflict and Domestic Audiences Introduction This chapter reviews the conflict escalation literature to gain insight into why states go to war. An analysis suggests that the wars are derived from three sources: psychological differences between the respective leaders, rational calculations based on the expected benefit and cost of war, and domestic political pressures either encouraging or discouraging escalation. Drawing from the most convincing elements of each, and applying them to sanctions, this chapter shows how sanctions can credibly signal a willingness of a sender to escalate if a sanction fails. It is from such a credible signal that sanctions may prove effective or inadvertently lead to an undesirable war. The reader is forewarned that this chapter does not seek to settle the thousand- year-old debate over the cause of war. The existing literature is rich with numerous, sophisticated models offering interested readers valuable insight. This project only uses the existing literature to Justify the underlying assumptions of the model in Chapter 3. However, while no claim is made as to which source “best” explains why wars occur, the chapter does argue that domestic sources are theoretically and empirically valid. An analysis of the literature also finds that mainstream rationalist “critical factors,” such as relative military capability and level of interest, influence the 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. probability of escalation, but alone are not sufficient to explain war. Psychological theories are noted for their contribution and are used to justify why games of incomplete information should be employed in the model. This dissertation does, however, make one important contribution to the conflict literature. Specifically, Chapters 4 and 5, provide empirical support for the "domestic audience cost” proposition. This proposition states that an executive escalates a conflict if both an existing policy fails and the domestic population finds the executive to be “incompetent” in his handling of the international dispute. If escalation does work, then the political leadership reaps domestic political rewards, whereas admitting failure would entail domestic political costs. While scholars have used such an argument to explain U.S. escalation in places like the Balkans, rigorous empirical support of the proposition does not exist. Subsequent chapters provide such support. Psvcholofiical Sources o f Conflict: Miscommunication, Miscalculation, and Heuristics Scholars who study war from a psychological perspective share two similar assumptions. First, they assume that the personalities of political leaders affect the likelihood that hostilities will escalate. Thus, holding all else constant a pacifist U.S. President James Carter is less likely to employ military force than an innate military hawk such as German Chancellor Adolf Hitler. The second common assumption is that misperceptions explain why governments enter wars. Robert McNamara, for example, argues that the United States continued the war in Vietnam because Lyndon B. Johnson’s obsession with defeating communism colored his imderstanding of the true 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. political and military situation.' Scholars who argue that individual psychological factors are the most important explanatory variable on the cause of war include Ned Lebow, Robert Jervis, Ole Holsti, and Alexander George.* These scholars claim that because human beings must rely upon heuristics to make judgments, miscalculations often arise that result in war. The Cuban Missile Crisis has often been cited as one example where misperceptions by political leaders resulted in undesirable conflict escalation. Rational explanations for putting missiles into Cuba, such as using them as a bargaining chip over Berlin, boosting the moral of East European allies, and helping Nikita Khrushchev domestically have been vehemently denied by knowledgeable Soviets.^ Khrushchev’s use of missiles to reestablish a “balance of power” between the U.S. and Soviet Union has also been discounted by both Soviet and American participants. Khrushchev, for example, wrote in his memoirs that the deployment of missiles would have no effect on the “balance of power.” When considering how the missiles would affect the balance of power at the first ExComm meeting. President John F. Kennedy stated that the missiles did not represent a military threat, and that a real threat would come only if the Soviets moved nuclear armed submarines off the American coast— which in fact they did ‘ McNamara, Robert, In Retrospect (New York, NY: Times Books, 1995). * George, Alexander and R. Smoke, “Deterrence and Foreign Policy,” World Politics 41 (1989) 170-182. Holsti, Ole, Crisis Escalation War, (Montreal: McGill-Queen’s University Press, 1972). Jervis, Robert, Perception and Misperception in International Politics (Princeton, NJ: Princeton University Press, 1976). Ned Lebow, Between Peace and War: The Nature o f International Crises (Baltimore, MD: John Hopkins University Press, 1981). ^ GarthofF, Raymond, Reflections on the Cuban Missile Crisis (Washington D.C.: the Brookings Institute, May 1989). 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. following the crisis. Although both Khrushchev and Kennedy discounted nuclear competition as a motivation for deploying missiles to Cuba, many scholars still consider this to be the primary reason behind Khrushchev’s decision to both begin and end the crisis."* This is unfortunate because the empirical evidence shows that not one participant on either the Soviet or American side believed the missiles were designed to influence the “balance of power.” What then was Khrushchev’s incentive to send missiles to Cuba? Khrushchev wrote that the “only goal in placing the missiles in Cuba was to prevent any encroachment on Cuban sovereignty and to assure the capability of the Cuban people to be the masters of their own country.”® Soviet participants echo these claims.® However, at no time during any of the ExComm meetings did anyone speculate that the Soviet’s may have placed missiles in Cuba to deter an invasion. Employing American notions of “rationality,” U.S. thinking throughout the crisis rested on three assumptions: (1) the missiles where somehow related to Berlin, (2) the missiles did not reduce American nuclear superiority, but did damage public confidence in U.S. military superiority, and (3) given President Kennedy’s warnings about placing missiles in Cuba on September 4, 1962 and September 13, 1962, their presence represented an affront to * Fearon, James Threats to use Force: Costly Signals and Bargaining in International Crisis. Fearon’s discussion of the Realist paradigm uses the Cuban Missile Crisis as an example where “balance of power” and “relative interests” explain the outcome to the crisis. ’ Khrushchev, Nikita. “On the Brink: Berlin and Cuba,” Khrushchev Remembers: the Glasnost Tapes (Boston, MA: Little, Brown, and Company, 1990) 171. * Blight, James On the Brink (New York, NY: Hill and Wang, 1989). 32 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. American power and credibility/ The thought that the missiles were placed in Cuba to deter an invasion was never mentioned even as a hypothetical possibility despite previous U.S. threats and preparations to start naval exercises in Operation Mongoose. It was therefore miscommunication and misinterpretation that explained different U.S. and Soviet interpretation of events. * What then explains why the Soviets believed Cuba was threatened while American leaders did not? Given the stakes involved, why was there such a wide gap in the thinking of U.S. and Soviet leaders? Psychological sources to conflict claim that their different perspectives of the world caused Khrushchev and Kennedy to hold different thought processes. It was these different psychological attributes that resulted in miscommunication and near nuclear catastrophe. American policymakers could not understand Soviet concerns over Cuban sovereignty, nor did Soviet leaders appreciate how sensitive Americans would be to having land-based nuclear weapons 90 miles away from Florida. For example, Soviet fears of American “containment” policy, in conjunction to Operation Mongoose exercises, caused the Soviets to fear a U.S. invasion while American officials believed that after the 1961 disaster at the Bay of Pigs, Soviet officials would recognize Trachtenberg, Marc, “White House Tapes and Minutes of the Cuban Missile Crisis: ExCom Meetings October 1962.” International Security (Summer 1985) 164-203. Welch, David, “The Eleventh Hour of the Cuban Missile Crisis: an Introduction to the ExComm Transcripts,” International Security (Winter 1987/88) 5-29. * Bundy, McGeorge, “October 27,1962: Transcripts of the Meetings of the ExComm,” International Security (Winter 1987/88) 30-92. Trachtenberg, Marc, “White House Tapes and Minutes of the Cuban Missile Crisis: ExCom Meetings October 1962,” Welch, David, “The Eleventh Hour of the Cuban Missile Crisis: An Introduction to the ExComm Transcripts.” 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. that the U.S. would not invade Cuba. The Soviets also believed that Americans would become accustomed to having missiles 90 miles from their border, just as the Soviets had come to accept Jupiters in Turkey. Finally, the Soviet government, use to restricting information to their population, thought that no public annoimcement of missile deployment was necessary whereas in the more open, democratic U.S. society, such a public annoimcement would be expected. Even as late as 1990, psychological differences between the American and Soviet thought process have not been resolved, as Robert McNamara claims that he still does not understand the Soviet thinking behind both why they would believe the U.S. contemplated invading Cuba and why the Soviets did not publicly aimounce the deployment of missiles.’ While psychological approaches hold explanatory power, they lack predictive and proscriptive value. Personalities do matter and incomplete information can cause political leaders to apply the incorrect heuristic. But understanding under what circumstances such misperceptions occur or which political leaders are most prone to miscommunication is difficult. Furthermore, political leaders still make decisions based on available information. As Karl von Clauswitz points out, while war is not a “purely rational act of government,” war is “a continuation of policy by other means.” Even an “irrational” leader needs the support of his military, bureaucracy, advisors. ’ See Blight On the Brink, Garthoff E jectio n s on the Cuban Missile Crisis. An excellent work to learn about how differences in the personalities of Khrushchev and Kennedy nearly brought the world to war in the early 1960s can be found in the work of historian Michael Beschloss. See Beschloss, Michael, The Crisis Years: Kennedy and Khrushchev (New York, NY: HarperCollins Publishers, 1991). 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. sources of finance, and often times, his population if he is to undertake a military campaign. Such support can only be derived from rationally understanding the cost and benefit of war. Thus, while incomplete information is adopted into the model of Chapter 3, there is no effort to account for the personalities of political leaders. International Sources o f Conflict: Mainstream Rationalist Arguments Mainstream rationalists argue that war is the result of two states incorrectly calculating their expected utility of war to be positive. Similar to psychological approaches, these scholars claim that misperception and miscommunication explain conflict escalation by causing states to either underestimate the expected cost of war or overestimating the expected benefit. In a world of perfect information, wars rarely arise, but incomplete information increases its probability. Specifically, mainstream rationalists argue that international contests of power are decided by some critical factor, such as relative military capability or level of interest. Rationalist models begin with two or more unitary actors disputing the ownership of some good. Given circumstances, if non-military methods of resolving disagreements fail, one of the actors may choose to obtain the good by threatening war. At this point, either one of the actors “backs down” due to a perceived relative deficiency in at least one of the cri tical factors or the two actors go to war. As Kenneth Waltz notes, “a state will use force to attain its goals if, after assessing the prospects of 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. success, it values those goals more than it values the pleasures of peace.”* '’ Thus wars are thought to occur because of a rational calculation that the expected cost (relative military capability) of war will be less than the expected benefit (level of interest). The first critical factor that determines which actor relinquishes claim to the good in dispute is the cost o f war, often found by comparing the actors relative military capability. Intuitively, the actor with the relatively weaker military is likely to concede the good to the stronger power, as all who have played “King of the Moimtain” are well familiar. The perceived magnitude of one’s own military strength relative to that of one’s opponent influences the decision to continue to lay claim to the good in dispute or backdown. Geofft-ey Blainey, for example, finds that “wars usually begin when two nations disagree on their relative strength and wars usually cease when the fighting nations agree on their relative strength.”' ' In a game of complete information, mainstream rationalist argue that forward looking actors concerned with the prohibitive cost o f conflict, and having an identical value of other critical factors as their opponent, will default on their claim, or at least make partial concessions. But with all else not held constant, or in the presence of incomplete information, “many authors in the realists camp have held that war is more likely when power, or military capabilities, are out of balance, because the stronger side will be tempted to Waltz, Kenneth, Man, the State, and IFar (New York, NY: Columbia University Press, 1959) 160. " Blainey, Geoffrey, The Causes o f War (New York, NY: Free Press, 1973). 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. aggression.”*' The works of Hans Morgenthau and Kenneth Waltz are classic studies into how the imbalance of power increases the probability of war.*^ More recent theoretical research by Randolph Siverson and Ross Miller finds that contextual variables, such as power capabilities, arms races, and polarity provide some insight into the conditions under which escalation takes place...the logic behind the former is simply that equality of power decreases the probability of escalation because the probability of victory is minimized. As power becomes relatively imbalanced, the stronger state is much more likely to use war as an instrument of policy because the probability of victory increases." While Blainey's work suggests that it is a balance of power that increases the probability of war. due to an increase in incomplete information, Blainey still finds that with complete information, relative military capabilities favoring the stronger actor cause the weaker actor to backdown. However, most research in conflict management finds that war is more likely when the variance in military capabilities is large. The second critical factor is the relative value of controlling that good, often measured by comparing the level of interest each actor has for the good. Even if an actor has the military capability to achieve his goal, often times threatening force is not credible because the benefit derived from realizing the objective is less than the cost. For example, while the United States could, theoretically, launch a massive military assault on Canada to force Canada to sell timber at a lower price, the United States lacks the incentive to undertake such a policy. In other situations, the U.S. does find the '* Fearon, Threats to Use Force 11. " Morgenthau, Hans, Politics Among [Rations (New York, NY: Knopf, 1948). Waltz, Kenneth, Theory o f International Politics (Reading, MA: Addison-Wesley, 1979). " Siverson, Randolph M and Ross Miller, ‘ The Escalation of Disputes to War,” International Interactions 19 (1993) 77-97. 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. benefit of deploying force greater than the cost. This “level of interest,” or “resolve,” is an important determinant to whom will be the first to backdown. Intuitively, the actor with the lower intrinsic interest is more likely to backdown holding all else constant. If both the interest and military capability of the actor are relatively low, then that actor will not pursue the good in dispute. Conversely, if both the interest and military capability of the actor are relatively high, the actor may elect to employ military force. Thomas Schelling, and Glenn Synder and Paul Diesing are but three of many scholars that show that level of interest is a critical factor in explaining why wars erupt.'* Robert Powell’s review of the literature also finds that most scholars accept the argument that states with greater resolve are more willing to assume the risk of war.'* Robert Gilpin’s theoretical work on the relationship between relative military capabilities, level of interest, and conflict escalation is representative of mainstream rationalists.'^ Gilpin begins his model by assuming that all states seek to maximize their utility. International stability is maintained when states find that the marginal cost of forcing a change to the status-quo to be equal to the marginal benefit, as any policy would be “unprofitable.” However, given exogenous changes to technology, economic conditions, or the geopolitical structure, there may be shifts in the marginal benefit Schelling, Thomas, The Strategy o f Cor^ict {< Z accianA % t, MA: Harvard University Press, 1960). Snyder, Glenn and Paul Diesing, Conflict among Nations (Princeton, NJ: Princeton University Press, 1977). “ Powell, Robert, “Crisis Bargaining, Escalation, and MAD,” American Political Science Review il:3 (September 1987) 717-735. Gilpin, Robert, War and Change in World Politics (Cambridge, UK: Cambridge University Press, 1981). 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and/or marginal cost functions. The new equilibrium results in a state possessing either too much or too little of international “goods.” For example, if for one actor technological changes relatively increase the marginal benefit of holding a good, or decrease the relative cost of gaining ownership of the good, then marginal benefit will exceed marginal cost and the actor will engage in aggressive international behavior to obtain more “goods.” The international system suffers instability during the transition process while states reallocate resources. The same process occurs for changes in either relative economic conditions or the geopolitical structure.'* Numerous studies also support the proposition that relative military capabilities and national interests are significant explanatory variables. Ole Holsti’s study of 177 international wars between 1648 and 1989 finds that both were important factors in describing the outbreak of war. Christopher Gelpi’s study of 180 disputes between 1948 and 1982 found that greater variance in military capabilities was positively associated with conflict, holding constant the type of government employing force.” Despite both the empirical evidence and intuitive logic of the mainstream rationalist explanation for war, there exists one troubling theoretical question: why do wars occur when the relative military capability or level of interest favors one side over the other? Fearon argues that most wars occur in the rationalist paradigm because of Gilpin ends the story by stating that states overextend themselves in the long-nm, such that even after obtaining a desirable good, the marginal cost of holding that good increase over time, causing a readjustment of world resources in the future. Hence, a global system ruled by anarchy is always subject to conflict. Gelpi, Christopher, "Crime and Punishment: The Role of Norms in Crisis Bargaining," American Political Science Review9l (1997) 339-360. Holsti, Ole, Peace and (Par (Cambridge, MA: Cambridge University Press, 1994). 39 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. “misperceptions and miscalculations arising from psychological pathologies,” or from incomplete information.-" For example, misperceptions have been used by mainstream rationalist such as Steven Brams and D. Marc Kilgour, Bruce Bueno de Mesquita and David Lalman, James Morrow, and Robert Powell to explain how wars arise.-' Lisa Carlson’s research is typical of mainstream rationalists that employ models of incomplete information. Carlson develops a two-sided incomplete information model showing an inverse relationship between cost tolerance and escalation behavior. The model shows that “as the disparity between the players cost tolerance increases, there is an increase in the probability that the lower cost tolerant actor achieves his or her maximum level of escalation at the outset of the conflict,” even in the presence of incomplete information. Carlson then tests her model in a logit regression with 39 cases where the militarily weaker state escalated the conflict and 80 cases where there was no escalation. The regression correctly predicted the outcome in 67% of the cases, suggesting that holding military capabilities constant, as the cost tolerance for conflict increases, states with lower cost tolerances are more likely to backdown.- Conversely, holding cost tolerance constant, relative military capabilities proved significant. Carlson’s empirical evidence lends support the thesis of mainstream rationalists. “ Fearon, Threats to Use Force, 80. Brams, Steven J. and D. Marc Kilgour, “Threat Escalation and Crisis Stability: A Game- Theoretic Analysis,” American Political Science Review il:3 (1987) 833-849. Bueno de Mesquita, Bruce and David Lalman, fPar and Reason (New Haven, CT: Yale University Press, 1992). Morrow, James D., "A Spatial Model of International Conflict," American Political Science Review 80 (1986) 1131-1150. “ Carlson, Lisa, “A Theory of Escalation and International Conflict,” Journal o f Conflict Resolution 39:3 (1995) 511-534. 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The international approach, with its reliance on military capabilities and levels of interest, makes an important contribution to the study of war. However, these models are insuffîcient. Empirically, many militarily weaker states have not acceded to the demands of a stronger power. Relatively weaker states have also been known to start conflicts. For example, Randolph Siverson and Michael Tennefoss find that in 255 conflicts between 1815 and 1965,49 were initiated by less capable military powers.^ Patrick James argues from his analysis of the International Crisis Behavior data that a cost-benefit analysis approach is insufficient, because "of the cases in which the initiator [of conflict] had positive expected utility [of using force], 5 out of 31 resulted in war, while only 1 out of the remaining 44 crises had that outcome.”* '* Thus, while the incomplete information models of Carlson and others improve the quality of the mainstream rationalist framework, they simply do not explain enough. A purely internationalist perspective suggests that "wars would occur only when the disadvantaged side irrationally forgets its inferiority before challenging or choosing to resist a challenge.”^ But to rely on such irrationality deviates from common sense, especially when the military strength of the opponent is visible. This is particularly true in the cases considered in this study. The military power of the United States is known ^ Siverson, Randolph and Michael Tennefoss, “Power, Alliance, and the Escalation of International Conflict” American Political Science Review 78 (1984) 1057-1069. * ■ * James, Patrick, Crisis and War (Montreal, Canada: McGill-Queen’s University Press, 1988). 70-71. “ Fearon, Threats to use Force, 14 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to be greater than that of any other state. Thus, why any target would be willing to challenge the U.S. within the mainstream rationalist literature can only be addressed by examining the critical factor level of interest. However, operationalizing "level of interest” is elusive. How are interests defined? Since an internationalist approach does not consider domestic politics to be important in the decision to use force, there is no explanation for why the United States would have a national interest in Haiti in 1994, but not Burma. Why a national interest in Somalia but not Ethiopia? Did America have a true national interest in liberating Kuwait? Whereas relative military capability can be quantified by comparing military expenditures, defining national interest has become so nebulous that its usage carries little meaning. What is national interest to one scholar is not to another. This does not suggest, of course, that level of interest is not relevant to the study of war. Clearly, any model of conflict must include both the expected benefit and cost. However, these factors alone can not explain the decision to escalate. Domestic Sources o f Conflict Unfortunately, the rigid assumption of a unitary actor employed by mainstream rationalists limits the understanding of what is in the national interest. For them, national interests are determined by elite beliefs. Former Secretary of State George Shultz and former Secretary of Defense Robert McNamara, for example, have argued that political leaders should employ force only when experts find the national interests are at stake. The Realist paradigm laments the fact that domestic politics can interfere 42 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. with what is in the national interest. Walter Lippmann wrote in his early works that if allowed to participate in the foreign policy making of the state, the American public would pressure political leaders to implement policies that were likely to harm the national interest.^ Morgenthau likewise regrets the fact that domestic political conditions can influence executive decision making. Finally, Waltz maintains that the unitary actor is the appropriate tool of analysis because the executive can always supersede the will of the people. According to mainstream rationalists, if leaders state that public opinion is driving their actions, it is only as an excuse for inaction, as the public usually prefers peace to conflict escalation.’’ The problem with such a theoretical argument is that public opinion does shape U.S. foreign policy, and has increasingly done so since the end of the Cold War. For example, U.S. military action in Somalia and Bosnia was driven, in part, by public demands that “something” be done about each crisis. After Americans saw on television U.S. soldiers being dragged through the streets of Mogadishu, public opinion forced President Clinton to withdraw from Somalia. General studies of U.S. foreign policy find similar results. Analyzing the effects of public opinion on foreign policy fi’ om World War I to the early 1990s, Holsti finds that “although the public is not well informed about many aspects of foreign affairs, its opinions are usually stable and “ Lippmann, Walter, Public Opinion (New York, NY: The MacMillan Company, 1961). Though often cited in the literature for opposing public participation in foreign policy-making, Lippmann changed his position in the late 1960s as the American public became increasingly incensed with the Vietnam War. By 1969, Lippmann was writing that political leaders should be listening to the will of the people. ^ McNamara, Robert, fn Retrospect (New York, NY: Times Books, 1995). Shultz, George, Turmoil and Triumph (New York, NY: Charles Scribner’s Sons, 1993). 43 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. reasonable reactions to real world events and can often have an important impact on foreign policies/’ Holsti’s research also demonstrates that since the end of the Cold War, the public’s influence over American foreign policy has increased, citing for example the 'Svidely held belief that the Clinton administration’s frequent threats of strong action, followed by retreats, over Somalia, Bosnia, Haiti, and MFN trade status for China were significantly influenced by public opinion."’* There is also growing theoretical research proving a correlation between domestic politics and international affairs. Robert Putnam’s seminal work on two-level games shows how events at the domestic level constrain the behavior of the executive at the international level.’’ Work by Helen Milner demonstrates how divisions between the Congressional and Executive branch increases the ability of a state to extract concessions from a foreign country. She finds that “in general, external conflict and cooperation reflect the struggle and consensus erected out of domestic politics.”^ ® Peter Katzenstein and John S. Odell have independently shown that domestic politics effects state-to-state interaction on economic matters.*' “ Holsti, Ole, Public Opinion and American Foreign Policy (Ann Arbor, MI: The University of Michigan Press, 1996). ^ Putnam, Robert, "Diplomacy and Domestic Politics: The Logic of Two Level Games," International Organization 4,2:2 (1988) 427-460. Milner, Helen, Interests, Institutions, and Itrformation: Domestic Politics and International Relations (Princeton, NJ: Princeton University Press, 1997) 14. Katzenstein, Peter ed.. Between Power and Plenty (Madison, WI: The University of Wisconsin Press, 1978). Odell, John S. “The Outcome of Trade Conflicts,” International Studies Quarterly 29 (1985)263-268. 44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Not surprisingly, the conflict literature has also employed domestic sources to explain international interactions. Domestic politics has 3 potential effects in conflicts: preventing its use, encouraging its use, or providing incentives for leaders to use force to improve their overall domestic standing when the public does not favor war. Each of these alternatives are considered in respective order. First, domestic politics can play a role in conflict escalation by constraining the ability of a leader to apply force. Democracies are most subject to this type of constraint because the financing for any military conflict must be approved by representatives of the population. Milner points out, for example, that “to start a war without one’s legislatures’ consent, tacit or otherwise, when the legislature must appropriate funds for the war is likely to lead to an inability to prosecute the war successfully.”^' However, obtaining such support is difficult, no matter how “vital” the good is in the minds of democratically-elected leaders. This is because it is only under rare circumstances that the average citizen supports war. As Immanuel Kant keenly observed as early as 1795 in his essay Perpetual Peace [f, as is inevitably the case under [a republican] constitution, the consent of the citizens is required to decide whether or not war is to be declared, it is very natural that they will have great hesitation in embarking on so dangerous an enterprise. For this would mean calling down on themselves all the miseries of war, such as doing the fighting themselves, supplying the costs of the war from their own resources, painfully making good the ensuing devastation, and, as the crowning evil, having to take upon themselves a burden of debt which will embitter peace itself and which can never be paid off on account of the constant threat of new wars." " Milner, Interests, Institutions, and Information 259. ” Kant, Immanuel “Perpetual Peace,” 1795. Cited in Reiss, Hans, Kant’ s Political Writings (Cambridge, MA: Cambridge University Press, 1970) IC O . 45 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Research by supporters of the “Democratic Peace,” proposition, such as Bruce Russett, claims that this explains why no democratic state has gone to war against another democratic state/^ Theoretical models by Michelle Garfinkel also show that democratic states will spend relatively less on military expenditures and more on economic programs given the latter’s higher electoral rewards/^ Thus, while it is often found that the public wants “something to be done,” about a crisis, the public is reluctant to pay for the cost of “doing something,” when military force is involved. The second means by which domestic politics affects foreign policy is by pressuring the government to escalate. While this rarely, if ever, occurs at the onset of a crisis, its theoretical possibility should not be discounted. If the population values a good in dispute, and displays a willingness to use force to achieve that good, then democratic leaders will be under domestic pressure to authorize the use of force. Former Secretary of State James Baker writes, for example, that while the Bush administration did not believe U.S. interests were at stake in Bosnia, U.S. public opinion compelled the administration to tackle the problem. Baker also notes that it was U.S. public opinion that pressured Reagan to impose sanctions against South Africa and forced the Bush administration to maintain the sanctions— despite the belief by policy elites in both administrations that sanctions did not serve the national interest.^* ^ Russett, Bruce, Grasping the Democratic Peace (Princeton, NJ: Princeton University Press, 1993). Recent fighting in the Balkans has challenged the proposition. Garfinkel, Michelle, “Domestic Politics and International Conflict,” American Economic Review 84:5 (December 1994) 1294-1309. Baker, The Politics o f Diplomacy. 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Research by Bueno de Mesquita and Laimer demonstrates how domestic audiences influence foreign policy behavior through these two approaches. While their research shows that domestic political constraints to conflict are often stronger than national interest incentives for war, they also show that the domestic political process sets the value of the good under dispute; and when governments are beholden to follow the will of the people, if public opinion highly values a good, then political leaders will be tempted to pursue that good independent of the probability of conflict escalation. In their view, domestic politics determine how a state establishes what goals are important and how those goals are obtained, either through negotiations or military escalation. Relative military capability only affects the decision a state makes in choosing to confi'ont another state at the initial stage. After the conflict starts, domestic political conditions determine if the conflict escalates.^’ These two approaches to the study of domestic politics and international conflict demonstrate that the models of mainstream rationalists lack robustness. It is important to note, however, that those who cite the domestic sources of war acknowledge that executives may authorize military conflict without pre-war support. This enables us to examine the third variant correlating domestic politics to international conflict— an executive using military force to improve his domestic standing. ” Bueno de Mesquita, Bruce and David Laimer War and Reason, 47 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Diversionarv Theory as an Explanation for War An intellectually challenging question has been raised by scholars interested in studying the relationship between domestic politics and international conflict: “Will a president authorize the use of military force as a means to divert public attention away from domestic issues harming presidential public approval?” Scholars who subscribe to “diversionary” theories of war answer the above question in the affirmative. It has been well established in the literature that an international crisis induces a “rally-around-the-flag” effect improving presidential approval ratings. Research by James Lamare examining the response of domestic audiences after the use of military force shows that domestic approval ratings of an executive initially increased for both parties to the conflict, and remained high for the victor, but fell for the loser. Lamare concludes from his research that available evidence suggest that direct U.S. military intervention in other countries generally results in a significant boost in popular support of U.S. Presidents. The public usually rallies around the president regardless of the uncertainty, the riskiness or the success or failure of his decision to intervene. In times of international crises, the public reaction ordinarily tends more toward unaction then toward rational assessment, at least for awhile." Skeptics of diversionary theory, such as Bradley Lian and James Oneal, also acknowledge that the mainstream literature overwhelming finds that “presidents have routinely used military force to raise their public approval rating.”^ ’ The argument that international conflict mobilizes the population to “rally-around” the government has " Lamare, James, International Crisis and Domestic Politics (New York, NY: Praeger, 1991) 8. ” Lian, Bradley and James Oneal, "Presidents, the Use of Military Force, and Public Opinion," Journal o f Conflict Resolution 37:2 ( 1993) 277-300. Leeds, Brett Ashley and David Davis, “Domestic Political Vulnerability and International D\sç\Aesf Journal o f Cotiflict Resolution A\:6 (1997) 814-833. 48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. even seeped into the sanctions literature with respect to the behavior of domestic audiences in the target. Work by John Scharfen shows that the target population will “rally-around” the government when the state is sanctioned— even if the population is opposed to the government.'**’ Johan Gultung explains the failure of sanctions against Rhodesia, in part, on evidence showing that opponents to the government joined their ranks when the United Kingdom threatened Rhodesia’s sovereignty by imposing an embargo.*' Hence, the association between international conflict and, at least in the short-run, improved domestic public opinion, is well established in the literature. Subscribers to diversionary theory claim that since there is an empirical correlation between the application of military force and an increase in public approval, in times of domestic distress or falling public approval, a president will undertake a military adventure to improve his domestic standing. Empirical support for this hypothesis is strong. Joe Hagan uses data on 38 states drawn from the Comparative Research on the Events of Nations (CREON) and World Handbook H between the years 1959 and 1967 to test for a correlation between domestic problems and war. His results suggest that political leaders do have incentives to divert attention away from domestic problems by agitating an international conflict, given a constraint on the costliness of a military adventure. Considering only American use of military force and comparing the timing of its usage against Gallup poll overall public approval, Charles Ostrom and Scharfen, John, The Dismal Battlefield: Mobilizing fo r Economic Conflict, (Naval Institute Press: Maryland 1995). ■ * ' Galtung, Johan, "On the Effects of International Economic Sanctions," ff^orld Politics 19:4. (1967) 378-416. 49 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Brian Job show that a presidential decision to authorize military force is dependent upon international considerations, the domestic situation, and personal factors. Their regression analysis o f226 episodes of military force between 1946 and 1976 finds that at times of high public approval, American presidents are less likely to employ military force, low public approval is associated with an increased propensity for military force, and the likelihood of using force during an election cycle is statistically positive. Ostrom and Job conclude from their data that “a president’s approval rating is the most important variable in the model from a statistical point of view,” more important than both national interest and relative military capability. The statistical analysis also suggests that there is a positive correlation between the use of force and an increase in the misery index, suggesting that presidents may be trying to divert attention away from domestic problems.^’ Statistical analysis performed by Patrick James and John Oneal confirm the results of Ostrom and Job. James and Oneal improve upon the data analysis employed by Ostrom and Job by accounting for the severity of force in the dependent variable as well as more clearly defining which variables act as proxies for international factors, domestic factors, and the political environment. Using data from the International Crisis Behavior Project on 39 applications of U.S. force between 1949 and 1976, the authors find that a falling presidential approval rating is the most significant explanatory variable in the decision to go to w ar."*^ Other studies showing ^ Ostrom, Charles W. and Brian L. Job, “The President and the Political use of Force,” American Political Science Review 80:2 (June 1986) 541-563. James, Patrick and John R. Oneal, “The Influence of Domestic and International Politics on the President’s Use of Force,” Journal o f Conflict Resolution 35:2 (June 1991) 307-332. 50 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. an association between Mling public opinion and the use of force include Richard Lebow’s study of 13 cases of U.S. use of force and Jack Levy’s work on diversionary theory.** Finally, a peripheral glance at the public opinion polling data collected for this research suggests that in every case where the U.S. employed force, there was a decrease in public approval of the president’s handling of the situation by at least 20%. The Role o f Domestic Audiences. “ Seff-Imposed" Costs, and CanRict Escalation If falling overall public approval rating is empirically correlated with an increase incentive to use force, it is reasonable to suspect that declines in a president’s handling of a situation should also be positively associated with incentives to escalate. Subscribers to this argument claim that “audience costs” are generated when foreign policies fail, prompting presidents to escalate a conflict so as to avoid a decline in public approval. This lead us to a final theoretical explanation for why wars arise. Suppose a situation exists where an executive desires some international good, but the relative military capability of the state is low, the cost of war exceed the benefit o f obtaining the good, and the domestic population does not support war. Can conflict arise under such conditions? Mainstream rationalists and democratic peace proponents would answer the question in the negative. Supporters of “diversionary theory” models Lebow, Richard Ned, Between Peace and War (Baltimore, MD: Johns Hopkms University Press, 1981). Levy, Jack, "The Diversionary Theory of War,” in Midlarslty, Manus ed. Handbook o f War Studies (Boston, MA: Unwin Hyman, 1989). 51 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. would suggest that other variables must be considered, such as overall public approval. However, scholars supporting the so-called “domestic audience cost” proposition use the foundations of diversionary theory to answer the question in the aSirmative. Specifically, these scholars argue that an executive can generate his own domestic political costs by adopting policies, which if fail, cause the public to find him to be “incompetent,” thus providing incentives for him to escalate so as to reestablish his domestic standing. The initial policy signals to the opponent the domestic political cost the executive will suffer if the opponent does not backdown. The more costly the signal, the greater will be the foreign policy failure and corresponding domestic audience cost. Thus by adopting a policy less than war, policymakers can credibly convince an opponent to concede the good in dispute out of concern of future conflict escalation, where the escalation is used to protect the domestic political reputation of the executive, even if the cost of war is greater than the benefit of realizing the good in dispute. These models often assume that both relative military capabilities and level of interest determine which conflicts are initiated, but do not predict which conflicts escalate. Once the process begins, political elites in the competing states find that it is only marginal changes in public support that drives behavior. James Fearon’s work represents the most rigorous and theoretically persuasive research on the domestic audience cost proposition. His research shows that once states enter into a dispute over the ownership of a good, publicly observable measures of 52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. relative military capability and relative interests do not affect future behavior. Instead, relative audience costs for foreign policy failure dictate whether a state chooses to escalate or backdown."*^ Specifically, Fearon demonstrates that regardless of initial conditions, the state more sensitive to audience cost is always less likely to backdown in disputes that become public contests. The intuition is that the greater the domestic cost for oscillating and then backing down, the more informative is the signal of escalation and the less escalation is required to convey intentions.^ Fearon’s intuition that perceptions of policy failure encourage political leaders to pursue a more hawkish foreign policy is acknowledged even by mainstream rationalists like Morgenthau, who regrets this fact. Morgenthau writes, for example, that governments that want to stay in power or simply retain the respect of its people can not afford to give up publicly part of what if had declared at the outset to be just and necessary to retreat from a position initially held, to concede at least the partial Justice of the other side’s claims. Heroes, not horse traders, are the idols of public opinion. Public opinion, while dreading war, demands that its diplomats act as heroes who did not yield in the face of the enemy, even at the risk of war, and condemns the weaklings and traitors those who yield, albeit only half-way for the sake of peace.^^ Based on this assumption, Fearon suggests that in states where the cost of conflict escalation is greater than the value of the good in dispute relative to an opponent, states can credibly persuade the opponent to relinquish claim to that good so long as two conditions hold: (I) the state can generate sufficient domestic audience costs, such that the benefit of going to war to reverse a credibility problem plus the value of the good is ■ '* Fearon, James D., "Signaling Foreign Policy Interests," Journal o f Conflict Resolution 41:1 (1997) 68-91. Fearon, James D., "Rationalist Explanations for War," International Organisations 49 ( 1995) 379-414. Fearon, James D., "Signaling versus the Balance of Power and Interests," Journal o f C or^ict Resolution, 38:2 ( 1994) 236-269. Fearon, James D., "Domestic Political Audiences and the Escalation of International Disputes," American Political Science Review, 88:3 (1994) 577-592. Fearon, James D., Threats to use Force: The Role o f Costly Signals in International Crises. Fearon, James D., “Domestic Political Audiences,” 577. Morgenthau, Politics Among Nations (New York, NY: McGraw-Hill, 1993) 532. 53 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. greater than the expected cost of war and (2) the opponent’s cost of war is greater than the combined value of the good and domestic penalty the government must pay to concede defeat. If a state reaches the so-called “time horizon,” defined as the point in time where condition (1) holds true, then that state is “locked” into a position of supporting war unless the opponent backs down. In a game of incomplete information, states may incorrectly measure their opponent’s domestic audience cost function in such a way that condition (1) holds true for both. At this point, war becomes inevitable, even if the initial cost of war for both states is less than the expected benefit. Fearon continues his analysis by arguing that states can endogenously generate domestic audience costs through two mechanisms. First, states can impose "sunken costs,” which are ex ante costs suffered prior to electing to go to war. For example, if an executive authorizes military exercises or engages in limited combat, the self- imposed costs inflicted on the state will cause the population to lose confidence in the executive if the state’s objectives are not realized. Policies imposing higher levels of “sunken costs” are associated with higher levels of domestic audience costs. The model presented in the next chapter claims that sanctions credibly signal such sunken costs, though Fearon does not mention sanctions in any of his works. Second, Fearon claims that political leaders can “tie their hands,” by committing the state to some action ex post unless the target concedes. Theoretically, the domestic audience cost the executive suffers if he makes a threat and the opponent does not backdown in the face o f the threat prompts the executive to carry-out his threat so as to avoid domestic political humiliation. A recent example may prove useful to understanding Fearon’s 54 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. argument. On September 15, President Clinton gave a national television address where he told the U.S. public that unless the ruling government of Haiti abdicated power, he would authorize the use of military force to restore democracy to Haiti. Independent of whether American national interests were at stake, once Clinton threatened the use of force in such a dramatic fashion, a failure to follow through would have entailed significant domestic audience costs for Clinton. Recognizing this, the ruling military leaders of Haiti relinquished power within four days. Fearon’s work shows that both approaches can credibly generate domestic audience costs, although the method of “tying one’s hands” is found superior to “sunken costs.” Fearon also shows that democracies are better able to adopt policies that generate domestic audience costs than authoritarian governments, given that the former’s political base of support is derived from the population while the latter is not. Research by Bueno de Mesquita and Laimer confirm this finding, noting that democracies can generate higher domestic audience costs than dictators."** Hence Fearon’s work has been used to explain the historical fact that democratic states have rarely, if ever, fought a war against one another. Since democracies can credibly generate domestic audience costs, democracies can better “signal” a willingness to escalate than a non-democracy. In situations where the two parties in dispute are both democracies, signals can be use to establish who would first reach the “time horizon.” ^ Bueno de Mesquita, Bruce and David Laimer, fTar and Reason, 271. 55 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The state that would come in second then backs down. But if one of the states is non- democratic, then difficulty in credibly signaling intent results in both states arriving at the time horizon and war becomes inevitable. However, there are three significant theoretical problems with Fearon’s domestic audience cost model. First, Fearon argues that as the states in dispute move closer to war, only domestic audience costs drive the decision to escalate. Relative military capabilities and level of interest are thought to only influence the decision to initiate a conflict. This is intellectually problematic. Clearly, there is no greater audience cost suffered than that of losing a war, particularly in a democracy. Fearon’s condition ( 1 ) does not account for this fact. If consideration is made of either winning or losing a war, then a model incorporating the domestic audience cost proposition must include mainstream rationalist critical values of relative military capability and level of interest. Second, both the works of Fearon and Bueno de Mesquita suggest that states will not ■ ‘bluff’ in sending a signal of resolve. While the mathematics of this result are intellectually powerful, such a finding is counter-intuitive. Plenty of examples exist where a state tries to “bluff,” by threatening force, and then shy away from the threat. For example, shortly after the invasion of Afghanistan, President Carter imposed a February 14 deadline for the Soviets to withdraw or suffer the consequences of the “severest” of American actions. February 14 passed without any U.S. response. During the Balkan conflict from 1992 to 1995, NATO made numerous military threats of airstrikes without carrying them out. Russia has made military threats against the Baltic states without acting upon them. Political leaders in weaker military states 56 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. routinely threaten conflict escalation only to backdown, as evidenced by threats to use force in the Arab-Israeli and Pakistan-Indian conflict. Empirically, “bluffing” routinely occurs in international relations and therefore will be included in the model in the next chapter. Third, Fearon simplifies his model by making audience cost a function of “time,” although he recognizes that a better measure is the amount of “self-imposed” cost an executive suffers. This project improves on Fearon’s research by considering self-imposed costs rather than time. Such an approach better captures the degree to which an executive puts up for examination his level of “competency.” For example. President Bush put more of his “competency” up for public scrutiny in the 6 months prior to the Gulf War than in the 8 years the United States sought to destabilize the government of Suriname. It was the level of self-imposed cost rather than time that explains why audience costs were generated in the former case, but not the latter. Empirical Support for the Domestic Audience Cost Proposition While the theoretical foundations of the domestic audience cost proposition are well established, evidence demonstrating the theory’s empirical validity is lacking. Empirical support for the proposition requires that three conditions are satisfied. First, there must be evidence that a failed foreign policy adventure is associated with declines in public approval. Second, the public must translate disapproval on foreign policy matters to declines in overall approval. Finally, it must be shown that political leaders escalate a conflict to reverse a fall in public approval. While the literature has successfully demonstrated that the first and second conditions hold empirically, there 57 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. has not been any research to date establishing the third. Chapters 4 and 5 make a significant contribution to the ‘‘domestic audience” literature by showing that declines in a president’s approval rating on specific foreign policy issues is positively associated with an increase in the probability of escalating a conflict. Both empirically and theoretically, the relationship between falling foreign policy approval ratings and declines in voter support has been well established. For example, an analysis of public polling data on the 1980 and 1984 Presidential elections by John Aldrich, John Sullivan, and Eugene Borgida found that international issues “were at least as important as domestic issues in their impact on voting behavior.”^ ’ Studying the same two elections, but within a social cognition framework, Miroslav Nincic and Barbara Hinckley found that foreign policy matters are second only to the domestic economy in voter decisions. The authors also found that foreign policy behavior significantly shapes the “general evaluation,” the public assigns to an incumbent’s competency.^® Theoretical work by Alastair Smith and David Hayes notes that although “domestic issues often play a more prominent role than international ones in determining elections...this does not mean that foreign policy choices are irrelevant, particularly at the margin.”*’ Aldrich, John H., John Sullivan, and Eugene Borgida, “Foreign Affairs and Issue Voting: Do Presidential Candidates ‘Waltz before a Blind Audience,” American Political Science Review 83:1 (1989) 123-141. 124. Nincic, Miroslav and Barbara Hinckley, “Foreign Policy and the Evaluation of Presidential Candidates,” Journal o f Corflict Resolution 35:2 (1991) 333-355. " Smith, Alastair and David Hayes, “The Shadow of the Polls: Electoral Effects on International Agreements,” International Interactions 23 (1997) 79-108. 58 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Support for the third condition, however, has remained elusive. Some scholars refer to the “diversionary theory” literature to suggest that since there is a correlation between failing public approval and conflict escalation, it is only reasonable to suggest that an increase in domestic audience costs are associated with an increase in the probability of war. Later research by Smith theoretically demonstrates that if presidential re-election is assured or there are no prospects for re-election, a president will only engage in foreign policy behavior that serves the “best interests” of the state. However, if a competitive electoral process exists, leaders may engage in aggressive foreign policy behavior to demonstrate to their domestic population that they are "competent,” particularly if the economy is suffering problems or the president is perceived ineffective. If the foreign policy adventure is successful in Smith’s model, the population rewards the president with re-election. If the adventure fails, then the president is perceived to be incompetent and his electoral prospects diminish.^* Note that Smith’s work, along with Fearon’s, assumes that domestic audience costs are only incurred if the public punishes a president for a failed foreign policy adventure.^^ Thus, so long as an executive imdertakes a foreign policy adventure entailing either “sunken costs,” or “tying hands,” and that policy fails. Smith argues that based on the diversionary war literature, an executive will escalate to reverse the decline in public confidence. Using a different approach, Joe Eyerman and Robert Hart Jr. have ” Smith, Alastair, "International Crises and Domestic Politics,” International Studies Quarterly 40(1996) 133-153. ” Smith, Alastair, "International Crises and Domestic Politics,” American Political Science Review 92:3 (1998) 623-638. 59 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. attempted to verify the domestic audience cost proposition by using the “democratic peace proposition,” as a proxy. Specifically, these authors note that since a democratic- democratic dyad is less likely to engage in war relative to a non-democratic-democratic dyad, holding all else constant, the domestic audience cost proposition must be valid,*** However, neither of these approaches are empirically rigorous. Smith’s theoretical argument linking the domestic audience cost proposition to the diversionary theory empirical literature is not convincing. There simply are too many potential intervening variables to go from a decline in public support on one specific foreign policy adventure, to a decline in overall public approval, to the decision to go to war to claim that theory alone verifies the proposition. Rigorous empirical evidence is necessary. Eyerman and Hart’s analysis is misleading because while the domestic audience proposition may explain why democracies do not fight one another, evidence that democracies do not fight does not prove that the domestic audience cost proposition is true. Perhaps other explanations account for their finding, such as cultural or economic ties. A rigorous proof of the domestic audience cost proposition requires that one studies cases where a state has suffered “self-imposed” costs, and correlate those costs against changes in public ophuon in the handling of that specific foreign affair issue, and then trace those changes to the probability of conflict escalation. It is only through such an approach that the domestic audience cost proposition can be empirically ” Eyerman, Joe and Robert Hart Jr, “An Empirical Test of the Domestic Audience Cost Proposition,” Journal of Conflict Resolution 40:4 (1996) 597-604. 60 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. validated. The evidence presented in Chapters 4 and 5 use the self-imposed costs associated with economic sanctions as one means to test this proposition. Specifically, it will be shown that there is a positive correlation between the intensity of economic sanction imposed, declines in public opinion in a president’s handling of a specific crisis, and probability the conflict will escalate. The Role o f the Mass Media in Generating Domestic Audience Costs The domestic audience cost proposition assumes that foreign policy failings are associated with public disapproval. However, the mechanism by which public disapproval is generated is left unexplained. Though not specified in the literature, it seems as though there are two potential explanations to this troubling gap. First, individuals that lose from policy failings will punish political leaders. In the case of "sunken costs,” financial losers from the imposition of sanctions will pressure the government to resolve the dispute either by escalating the conflict (so as to end it) or conceding defeat. Those suffering from the "sunken costs” of participating in military exercises may become disgruntled if they are sent away from their families for months, or even years. Finally, the greater the self-imposed cost, the more likely the public will believe that the credibility and reputation of their leadership is at stake. It is from this basis that domestic audiences costs are ultimately derived. If the self- imposed costs are large, the more likely the population will identify the competency of their leader with the policy outcome. If the self-imposed costs are small, the executive’s abilities will not be scrutinized. 61 Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. Second, media coverage of international conflicts effects the probability domestic audience costs are generated. While there is little disagreement among scholars that the public obtains most of its information about foreign affairs, from the national media, the impact the media has on public opinion is not clear. Doris Graber’s study on how 165 Americans processed information from various news sources suggests that citizens retain an average of only I out of 16 television newscasts seen. However, if either political elites deem a story to be important, or the mass media repeatedly reports on the same story, then media consumers are found to have absorbed a substantial amount of information concerning the reported item, independent of personal characteristics such as initial interest in the story, income level, and education status.^^ Statistical analysis by Donald Jordan and Benjamin Page also finds that while "public opinion concerning foreign policy is rather stable...when collective opinion does move, the news that is broadcast on network television accounts for a large part of the magnitude and direction of change,” holding constant the source, such as an anchor, television commentator, or political leader.^* Such an argument is common among scholars who believe that the national media has power over public opinion and legislative behavior. For example, Shanto Iyengar and Donald Kinder claim that " Graber, Doris, Processing the Mews: How People Tame the Information Tide (Chicago, IL: University of Chicago, 1988). Jordan, Donald and Benjamin I. Page, “Shaping Foreign Policy Opinions” Journal ofC or^ict Resolution Vol 26:2 (June 1992) 227-241. 62 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. television media is the most significant variable in setting the political agenda.^^ Noam Chomsky’s research is indicative of the most extreme elements within the literature by arguing that the national media is primarily responsible for shaping public opinion/^ Policymakers have long recognized the power of the media to shape public attitudes of presidential leadership competency. Research by Lara Brown shows that a “love-hate” relationship exists between presidents and the press. She argues that the media’s portrayal of presidential leadership is instrumental in moving public opinion, and as such, presidents will try to influence media reporting.” In a commencement speech at Harvard University, Vaclav Havel expounded upon the role of the media in forming public attitudes on presidential leadership. Although his comments were directed at the Czech media, it resonated with many American political leaders. Havel claimed that the cynical nature of media reporting has the potential to make or destroy presidential leadership, observing that: I never fail to be astonished at how much I am at the mercy of television directors and editors, at how my public image depends far more on them than it does on myself...! know politicians who have learned to see themselves only as the television camera does. Television had thus expropriated their personalities, and made them into something like television shadows of their former selves. 1 sometimes wonder whether they even sleep in a way that will look good on television."® Iyengar, Shanto and Donald Kinder, News that Matters: Television and American Opinions (Chicago, IL: University of Chicago Press, 1987). Chomsky, Noam, Manufacturing Consent (New York, NY: Pantheon Books, 1988). Brown, Lara “Symbolic Images o f the President in the Press,” unpublished manuscript 1998. “ Cited in Fallow, James, Breaking the News (New York, NY; Vintage Books, 1997) 142. 63 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Individual case studies support the contention that the national media effects the public’s image o f their president. Research by Robert Denton, James Combs, Bethami Dobkin, and David Swanson shows that national media reporting altered public opinion concerning the utility of war against Iraq during the Gulf conflict.®’ Mark Rozell’s study on how news reporting impacted the public’s belief that President Carter was competent finds that the media blamed Carter for the failure of American foreign policy, even though news outlets were not advancing alternative courses o f action. The more the media wrote about Carter’s failures, the more the public came to believe that Carter was incompetent. Rozell writes, for example, that “by late 1979, in light of the Iran and Afghanistan crises, journalists characterized President Carter’s foreign policy agenda as hostage to external events...[and] as these crises dragged on journalists increasingly held Carter responsible for not being able to alter the course of events.”® * Stephen Smith finds that television images of starving children explain U.S. military involvement in Somalia more than any other variable— as well as explain the rapid pull-out o f U.S. troops after American soldiers were cruelly dragged through the streets of Mogadishu.®^ Besides academia, policymakers have also noted the impact media reporting has had on foreign policy. Former Secretary of State James Baker has written that it was both television reporting and changes in public opinion that pressured the Bush Denton, Robert ed. The Media and the Persian G ulf War (Westport, CT: Praeger, 1993). “ Rozell, Mark, The Press and the Carter Presidency (Boulder, CO: Westview Press, 1989) 8. “ Smith, Stephen, “Population, Famines, and Entitlement Theory: Application to the Case of Somalia,” Case Studies in Economic Development (Reading, VIA: Addison-Wesley, 1997). 64 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. administration to take stronger actions against both the Chinese and Serbian government in 1989 and 1991, respectively. He writes that “in Iraq, Bosnia, Somalia, Rwanda, and Chechnya, among others, the real-time coverage of conflict by the electronic media has served to create a powerful new imperative for prompt action that was not present in a less frenetic time,” and that as the global telecommunication revolution continues, "public outcries for the United States to become involved in situations of great tragedy that may or may not be consonant with our national interests will force U.S. Presidents to act when they otherwise may not.”^ Although it is clear that the national media does have an impact on public opinion o f presidential leadership, there is still debate concerning the media’s marginal impact. James Fallow’s study public consumption of news finds that Americans are no longer turning to mainstream media sources as reliable providers of information. He concludes that Americans fault the mainstream media with a lack of in-depth reporting, reporting of non-vital topics, and excessive negativity in their reporting; and as such, the media’s ability to educate the public has waned. Ann Criger’s comprehensive analysis o f media reporting and Congressional agenda setting on ten issues from 1945 to 1980 suggests that “over the long run, the mass media do not tend to set the agendas of Congress or public opinion. The process is too complex and the expectations of media power are too great for this simplistic ‘bullet theory’ approach to communications “ Baker, James A. HI, The Politics o f Diplomacy: Revolution, War, and Peace 1989-1992 (New York, NY: G.P. Putnam’s Sons, 1995). 65 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. effects.”* ^ While these studies are not widely accepted in the mainstream literature, the evidence presented in both studies does give pause to the claim that foreign policy is driven by the national media without consideration of other factors, such as public opinion, national interests, and the expected cost of intervention. But even if the media does affect public opinion, there is still debate over whether public opinion shapes U.S. foreign policy, as required by the domestic audience cost proposition. Mainstream rationalists argue the question is moot. Waiter Lippmann, in his classic study Public Opinion, argues that the complexity of the international political system demands that experts and scholars manage foreign policy, and that the public is incapable of determining U.S. national interests.^ Although he did reverse his position by the end of the 1960s, Lippmann’s work is still cited by advocates of an elite model of foreign policy decision making. Robert McNamara writes that public opinion polls did not play any role in President Kennedy’s or President Johnson’s Vietnam policy.*^ George Shultz claims that American foreign policy should only consider the national interest and not U.S. public opinion.*® Thus, even if the media does influence the generation of domestic audience costs, mainstream rationalists contend that public pressure does not effect foreign policy. Crigler, Setting the Congressional Agenda, 3. “ Lippmann, Walter, Paé/zcOp/mon (New York, NY: The Macmillan Company, 1961). McNamara, Robert, In Retrospect, “ Shultz, George, Turmoil and Triumph (New York, NY: Charles Scribner’s Sons, 1993). 66 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. While these arguments hold interesting intellectual merit, the overwhelming evidence shows that public opinion does affect foreign policy. For even if the media can not uniquely change foreign policy, the empirical analysis must account for changes in media coverage to accurately assess the presence of domestic audience costs. Conclusion This chapter provided a survey of different theories of conflict escalation. These included psychological sources based on heuristics, international sources based on critical factors such as relative military capability and level of interest, and domestic sources such as diversionary war theory and the domestic audience cost proposition. Each source holds intuitive value. Psychological explanations are useful, but are limited in their predictive power. Mainstream rationalists notions of critical values hold strong empirical validity, but are insufficient. Finally, domestic sources provide useful insights by taking one of three forms. First, the public’s reluctance for war constrains an executive’s ability to escalate. Second, domestic audiences may pressure the government to actively intervene. While the public rarely demands that their government goes to war, public pressures has required U.S. administrations to adopt non-military policies, such as sanctions. Finally, even if the public opposes war, if the failed foreign policies of an executive generate “domestic audience costs,” the executive may choose to escalate rather than backdown. The next chapter merges the these strains of the conflict escalation literature with the sanctions literature to build a model demonstrating how sanctions can work. 67 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Diplomacy has rarely been able to gain at the conference table what cannot be gained or held on the battlefield General Walter Bedell Smith Chapter 3 : ''Signaling" Sanctions : Sanctions as a Sign of Future Escalation Introduction The reader will recall that three significant results were established in Chapter 1. First, the empirical evidence shows that economic sanctions almost always “fail”. Hufbauer, Schott, and Elliott found that every U.S. or U.N. application of sanctions from 1983 to 1990 failed except one. Second, it was discovered that the existing sanctions literature does not provide a theoretically rich model to explain why economic sanctions should be expected to work. This was derived from an analysis of the current frameworks and comparing their predictions with the empirical evidence. The most common approach— where the target suffers a larger “cost” from the sanction than the “benefit” obtained from her current behavior, inducing her to change her behavior— was shown to stand on solid theoretical grounds, but lacking empirical rigor. Finally, many pundits and scholars believe that sanctions serve as a substitute for war, even though every application of U.S. force since 1970 was preceded by sanctions. Thus, instead of acting as a substitute for military force, sanctions may actually be a complementary good. Can a failed sanction episode escalate into an undesirable war? I believe the answer is “yes,” and it is from this basis that a model is developed explaining why sanctions can be an effective tool of coercion. 68 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This chapter argues that sanctions are most effective when they credibly signal that the sender political leadership will suffer domestic audience costs if the target does not backdown. Specifically, a failed sanction attempt generates domestic audience costs, and as these costs accumulate, the sender leadership has incentives to apply military force to reestablish his domestic reputation. Thus, even if the value of the good in dispute is less than the cost of escalation, wars may arise if, after the generation of domestic audience costs, the value of the good in dispute plus domestic political costs for policy failure is greater than the cost of war. Sanctions can, and do, work. But when a difficult foreign policy objective is sought, sanctions do not work by imposing a large cost on the target, rather they work by signaling the potential for future costs. The theoretical model establishes the following 5 key results: 1. With complete information, sanctions can work and conflicts will not escalate. 2. With incomplete information, if a large sanction is imposed (separating equilibrium), the target will backdown and there is no conflict escalation as the target believes that future escalation will be prohibitively costly. Sanctions work in this case. 3. With incomplete information, if a small sanction is imposed (semi-pooling equilibrium), the target will sometimes backdown and other times not backdown, and the sender may escalate to war even if the expected benefit of war is less than the expected cost. Thus, sometimes sanctions work while other times sanctions lead to undesirable wars. The model clearly shows that in a game o f incomplete information, there is a positive correlation between economic sanctions, domestic audiences and conflict escalation. 4. The underlying variable driving the results is domestic audience cost. 5. Relative military capabilities and level of interest determine critical parameters. 69 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The empirical analysis in the next two chapters verifies the major assumption in the model— that self-imposed costs, as measured by sanctions, are positively associated with domestic audience costs. The evidence also supports the model’s results by demonstrating a positive correlation between domestic audiences and escalation. This chapter is divided into seven section. I first briefly review the literature examining sanctions as a signal of resolve and present the intuition behind my argument. 1 then formalize both the terms and assumptions used in the game. The third section displays the extensive form of the game and shows the payoff rankings for both the sender and target. In section 4 ,1 more rigorously demonstrate that a relationship exists between domestic audience costs and conflict escalation. The results of the game given complete and incomplete information are presented in sections 5 and 6, respectively. Finally, a concluding section summarizes the results. The findings support the thesis that sanctions can work by credibly signaling a willingness of resolve, but simultaneously show that under some conditions, sanctions may result in escalation when both actors would have preferred the status quo. 1. Sanctions as a Signal of Resolve The existing literature considers sanctions as a signal only in the context of substituting for war. Most scholars write that sanctions inform the target that the sender is willing to use force to achieve his goal, but chooses to impose sanctions first 70 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. given their relatively lower cost. ' Antonia Hayes and Abram Hayes argue, for example, that U.S. policymakers tend to think in terms of an escalation ladder in response to a foreign policy crises: diplomatic initiatives at first, limited coercive measures such as sanctions second, and only if sanctions fail, the use military force.* Andrew Kohut and Robert Toth write that “even when the public feels that the United States has a responsibility to act, the public will always gravitate to diplomatic or economic options over military force, if they are in play as options.”^ Jane Hall notes that “as a crisis drags on, and as other options are demonstrated to be ineffective, expectations grow that the use of force is the only strategy that will work, and military and civilian leaders develop options to use force although the options evolve in relative isolation from, rather then in conjunction with, other non-military options.”^ David Baldwin finds in his research on economic statecraft that a sanction is most effective as a signal when it includes a self-imposed cost because that informs the target that the sender is willing to go to war, if necessary, to achieve his objective.^ Thus, targets suffering from sanctions ‘ Losman, Donald, International Economic Sanctions (Albuqerque, NM: University of New Mexico Press, 1979). * Antonia Handler Chayes and Abram Chayes "Alternatives to Escalation.” in The United States and the Use o f Force in the Post-Cold War Era (Washington D.C.: The Aspin Institute, 1995). ' Kohut, Andrew and Robert Toth “The People, the Press, and the Use of Force,” The United States and the Use o f Force in the Post-Cold War Era, 165. ■ * Hall, Jane “We the People Here Don’t Want No War,” m The United States and the Use o f Force in the Post-Cold War Era, 116. ’ Baldwin, Economic Statecraft. 71 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. are thought to interpret sanctions as a signal that the sender already prefers war to the status quo, and the sender only selects to impose sanctions as a “warning” that war is assured unless the target concedes. [ argue that sanctions are not a substitute for war, but rather a complement. For sanctions to work as a as signal of resolve when they substitute for war, the target must believe that the sender finds the value of the good to be more than the cost of war at the initial stage of the conflict. This is both empirically and theoretically doubtful. Most wars are fought not because of expected positive rates of return, but because of miscommunication and misinformation. Furthermore, even if the benefit o f war exceeds the cost, given that the literature shows that sanctions do not work, it is irrational for state leaders to inflict themselves with self-imposed “sunken costs” only to have to go to war at a later date. Thus, for sanctions to serve as a useful signal, sanctions must complement military force. This is because by imposing a sanction, the sender government credibly commits itself to either winning the conflict or suffering domestic audience costs. If the political costs o f failure are high, the sender prefers to escalate than to backdown. 72 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2. Terms and Assumptions of the Game I now will present a dynamic game of two-sided incomplete information demonstrating how sanctions can signal a willingness to go to war, under what conditions sanctions are successful, and under what conditions undesirable wars result. Let there be some target state (/) who enforces a policy yielding to the target utility x. The sender state (/) can adopt one of three options: do nothing, immediately go to war, or impose any value of economic sanction, s, defined as the loss in trade, between 0 and S, where S represents the maximum sanction a sender can impose given exogenously determined levels of trade linkage between he and the target. If the sender elects to “do nothing,” then he will not take any action and set 5 = 0, s e [0, S]. Once the sanction is imposed, the target suffers a loss defined by the function e(s), a continuously increasing, concave function whose minimum value takes 0 if s=0 and maximum value takes E if s = S, e(s) e [0, E]. Having observed the sender’s action, the target has one o f two options: she can either backdown, or she can fight the sanction at the risk that the sender will launch a military strike. If the target elects to backdown, she loses x and the sender receives d, the value of the good being sought. But if the target selects to fight the sender, the sender is left with two options: he can either backdown, lifting the sanction and no longer attempting to achieve his foreign policy goal, or the sender can escalate the conflict by using force. If the sender backs down, the sender leadership sustains an audience cost as defined by the function a(s), a continuously increasing concave function whose minimum value is normalized at 0 when s=0 and maximized at A when s = S, a(s) e [0,A]. When sanctions are not 73 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. imposed, the sender experiences no audience cost because the sender is not putting his reputation as a problem solver up for examination. Audience costs are endogenously determined by the level of sanction imposed. The more costly the sanction, the larger will be the domestic audience cost if sanctions fail. Note that the model assumes that the audience cost function is identical for all senders.® Data provided in Chapter 4 empirically support the assumption that sanctions are positively correlated with audience costs, holding constant other explanatory variables. All information in this game, F, is common knowledge with two exceptions. First, allow for nature to select between two types of senders, T, = {/, A}, which is known to the sender, but not perfectly known to the target. One type of sender (T, = /), defined as the low-cost sender, values the losses associated with realizing his objectives relative to the benefit of achieving his goal less relative to the other type of sender (T,= h \ defined as the high-cost sender. Specifically, it will always be the case that h> I. Low-cost senders exist for one of two reasons. First, if the domestic population holds internationalist preferences and believes that their government is obligated to be actively engaged in an international dispute, then the population will be relatively more receptive to the sender government taking foreign policy actions, independent of the value of the good. Even if sanctions fail, this population is expected to be more willing to absorb the economic costs associated with sanctions because they are more likely to ‘ One can make the reasonable claim that a(s) for the low-cost sender should he larger than a(s) for a high-cost sender, holding the level of sanction constant However, because of the assumption that a high-cost sender will always lose a war, such that he would chose to backdown rather than go to war, changing the functional form of a(s) based on sender adds nothing to the analysis. 74 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. support their initial application. Alternatively, when there is international support for the sender’s objectives, the relative cost of sanctions fall and the sender is of type low- cost. The high-cost sender, on the other hand, is less willing to tolerate the economic costs associated with sanctions. Because the population is generally isolationist and does not support state involvement in international disputes, or because there is a lack of international support, the cost of sanctions will be relatively high given the same value of c/ as compared to a low-cost sender. The types also differ in their ability to win a military conflict. For the low-cost, internationalist sender, any military conflict will be won because there will always be sufficient support from both the domestic public and the global community. However, a high-cost sender can not win a military confrontation because either the public or the international environment poses an tmsurmountable obstacle for the successful application of military force. As discussed in Chapter 2, it is reasonable to claim that the public or third-parties constrain the ability of an executive to win a war. Thus, while the cost of war is held constant, sender type determines if he will win a military conflict. Nature selects with probability ô that the sender is of type I and probability (1- Ô ) that the sender is o f type h. Second, allow for nature to select between two types of targets, T^ = {g, k}, which is perfectly known to the target, but not to the sender. Specifically, some targets, so-called “strong” targets (T^ = g), will find it cheaper to fight a war relative to continuing her current behavior relative to other target states, called “weak” targets (T^ = k). It will always be the case that g. 75 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Three exogenous factors determine whether a target is strong or weak. First, a strong target has a large military or industrial capability while a weak target is not endowed with such resources. It is intuitive that a war against the U.S. will be relatively cheaper for a state with such capabilities relative to a target fighting the U.S. without resources to protect herself. Second, a strong target receives significant international support, while a weak target is left to fight without international assistance. Again, it is intuitive that large coalitions are more capable of diversifying the cost of war relative to a state left to defend for herself. Finally, in a target where ruling interest groups suppress the opposition, the government is more willing to absorb the cost of war. As noted in previous chapters, empirical and theoretical evidence supports the intuitive argument that governments with different levels of legitimacy hold different valuations for conflict.^ In the extreme case, those groups benefitting from x place no weight on the cost of war. Based on these two types of targets, I set the cost of war for a strong target against a sender at g.y and the cost o f war for a weak target against a sender at k.y. Finally, the cost of war for senders to be as follows: z.w if the target is strong, and w if the target is weak, where z is some parameter greater than 1 capturing the additional expenses incurred to fight against a strong relative to weak target. Nature selects with probability A . that the target is strong and probability (1-A.) that she is weak. All payoffs are known with perfect information. It is only with respect to player types that incomplete information exists. ^ Lichbach, "Regime Change: A Test of Structuralist and Functionalist Explanations." Schock, "A Conjunction Model of Political Conflict" 76 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3. How the Game is Played, the Extensive Form, and Payoffs The game is played in the following manner. Nature selects between the two types of targets by drawing from a uniform distribution between 0 and 1, A . e [0,1], Nature then selects between the two types of senders, also drawing from a uniform distribution between 0 and 1, ô e [0,1]. Thus, nature draws from one of four possible joint probabilities: low-cost sender—strong target (%,), low-cost sender—weak target (tCj), high-cost sender—strong target (iCj), and high-cost sender—weak target The sender moves first by determining if he wishes to demand the target changes her behavior by imposing a sanction, ^ > 0, or if he would rather not start an international dispute and set s = 0, with the option of going to war in a later move.* Upon observing the sender’s behavior, the target chooses from her strategy space, b e (bd, fight}. Specifically, the target can capitulate to the demands of the sender and backdown, or the target can challenge the sender’s request by fighting. The sender makes the final move by either backing down or going to war. Formally, the a strategy for this game is read as {(Sl,cJ, (Sh,Ch) ; (bo), (b^) \p,q). More specifically, a strategy is {(low-cost initial sanction, bd or war for the low-cost sender), (high-cost initial sanction, bd or war for the high-cost sender); (bd or fight for the strong target), (bd or fight for the weak target)! p, q}. Note that p indicates the belief that the target holds that she is playing against a low-cost sender at her move in the game while q represents the belief the sender holds that he is playing against a strong target at his final move in the game. An equilibrium ' 1 will be setting the cost of war to be greater than the benefit horn winning a war, such that the sender will never choose the action {war} in the first move of the game. Different variations of the game, however, may require that such an action is included in the first move. 77 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. is the pair of strategies {(s^, c j, (s„, C h ) ; (bo), (b^)} where each actor is doing as well as it can for itself, given the actions taken by the other actor and given the institutional framework.^ The extensive form of the game is captured in Figure 3.1. Nature moves first by randomly drawing from the 4 possible states of nature. The left-hand side of the tree displays the game when the sender is low-cost while the right-hand side displays the game when the sender is high-cost. The sender first selects what sanction to impose, where the arched dashed lines are information sets showing that the sender does not know with perfect certainty if he is facing a strong or weak target. The target then moves, also in the presence of an information set given uncertainty over sender type. The last move in the game is made by the sender, who may still be uncertain over the type of target. Payoffs are given at the end of each terminal node. Note that nodes are designated by numbers 1-12 for future reference. ’ fCreps, David A Course in Microeconomic Theory (Princeton, NJ: Princeton University Press, 1990)6. 78 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3.1: Extensive Form of the Game NATURE low-cost sender (8), strong target (A .) n. Sender I high-cost sender (1-8), low-costsender (8), high-cost sender (l-8>>-. weak target (1-X) weairtarget ( I -X) strong target (X) Q z 0] Sender 2 - A Sender Sender 4 s=0 s=0 s=0 s=0 U(s 0) sanction s > 0 sanction s>0 U*(s=0) sanction s > 0 U”(s=0) sanction s > 0 Target Target 6 Target 7 Target S bd bd bd bd U(A) fight U C A ) fight U*(A) fight U”(A) fight Sender 9 Sender 10 Sender II Sender 12 bd U(B) war bd UiB} war bd U*(B) war bd U”(B) war 1 U(C) UCQ U*(C) U”(C) 79 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I next establish the payoffs at each terminal node. The utility when no sanction is imposed is normalized at 0. If a sanction is imposed, the sender will realize one of three outcomes: (1) the target backs down yielding the sender a benefit of d minus the cost of the sanction, r,.j; (2) the target fights the sender and the sender backs down at a cost of both the sanction, T,.s and a domestic audience cost, a(s)\ or (3) the target fights and the low-cost sender elects to go to war, where he wins d after suffering the cost of both war, z.w or w, and, T,.s. The sender is assumed to be risk-neutral and uses a linear objective function to find von Neumann-Morgenstem payoffs. Equation ( 1 ) represents the sender’s expected utility. D[destabilize] is defined as an indicator function for when the sender achieves his objective, either by the target backing down or a low-cost sender winning a war. □[destabilize] takes values of either 0 or 1 and is multiplied by the reward the sender receives if he realizing his goal, d. W[war] is an indicator function for when a low-cost sender goes to war. W[war] also takes the value of either 0 or 1 and is multiplied by the cost of war given the type of target, z.w or w. A[audience cost] is an indictor function taking the value of either 0 or 1 depending on if the sender backs down and suffers a domestic audience cost, a(s). Finally, the cost of sanctions is conditioned on sender type, .s. Payoffs at each terminal node for the sender can be derived from equation (1), (l)E [U j(«)]= D[destabilize]r/ - W[war]7^w - A[audience cost]a(^ -T,.s 80 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Similar reasoning is applied for the utility realized at each terminal node for the target. The target also considers a linear objective function to find von Neumann- Morgenstem payoffs. D[destabilize] remains an indictor function taking values of either 0 or 1 and is multiplied by the cost the target suffers if she ends the policies the sender finds objectionable or if she loses a war, x. W[war] is an indicator ftmction taking values of 0 or 1 and is multiplied by the cost of war given target type, Ly or g.y. E(s) is the economic cost suffered from the sanction. Note that I assume e(s) < x at all s. If this condition does not hold, then sanctions do not work from its signaling effect, but rather because the cost o f sanctions would be greater than the objectionable policies of the target, inducing the target to change her behavior. Equation (2) specifies the expected target payoff. (2) E[Uj (•)] = D[destabilize]x - W[war]T^ .y - e(s) Utilities realized by the sender and target at each terminal node with the corresponding pure strategy given equations (1) and (2) with complete information are provided below. 81 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PURE SENDER TARGET NODE STRATEGY UTILITY UTILITY U(s=0) {(0,0), (bd,bd); fight, fight} Ul(s=0) = 0 Ug(s=0) = 0 U(A) {(s, war), (-,-); bd, -} UL(A) = d -/.s Ug(A)= -x -e (s) U(B) {(s, bd), (-,-); fight, -} UL(B) = -a (s)-/.s Ug(B)= -e(s) U(C) {(s, war), (-,-); fight, -} UL(C) = d -2 .w -/.s Ug(C)= -x -e(s)-g .y UfA) {(s, war), (-,-); -, bd} Ul(A1= d -/.s Uk(A )= -x -e (s) U£B1 {(s, bd), (-,-); -, fight} U.(B) = - a(s) - l.s Uk(B}= -e(s) {(s, war), (-,-); -, fight} Ul(Q = d - w - /.s Uk( Q = -x-e(s)-fcy U*(A) {(-, -), (s, war); bd, -} Uh*(A) = d - A .S UG*(A) = - X - e(s) U*(B) {(-, -), (s, bd); fight, -} U„*(B) = - a(s) - h.s Ug*(B)= -e(s) U*(C) {(-, -), (s, war); fight, -} Uh*(C) = - z.w - h.s Ug*(C) = - e(s) - g.y U “(A) {(-, -), (s, war); -, bd} Uh”(A) = d - A.s Uk”(A) = - X - e(s) U” (B) {(-, -), (s, bd); -, fight} U„”(B) = - a(s) - h.s Uk”(B )= -e(s) U “(C) {(-, -), (s, war); -, fight} Uh”(C) = - vf-h.s Uk”(C )= -e(s)-fcy The institutional structure results in the following equalities for the sender: (3)U l(A )= Ul£A1 (4)U h ”(A) = Uh*(A) (5) Ul(B )= Ul(B} (6)U„”(B) = Uh*(B) (7) Ul (C) < UlIQ (8)U h ”(C )< U h*(C) The following equalities hold for the target: (9) Ug (A) = Uk IA1 = Ug*(A) = Uk ”(A) (10) Uo(B) = U d D = Uo*(B) = UK”(B) (11) Uo(C )> U k ( Q (12) U g* (C )> U k”(C) I now turn to ranking the payoffs. I consider the sender’s payoff rankings first. Let it always be the case that the low-cost sender prefers that the target backs down to the status quo, but prefers the status quo to initiating a crisis and then backing down. Specifically, assuming that d > l.s for all jr e [0,5], then for the low-cost sender: (13)U l (A )> U l (s= 0)> U l (B) 82 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. However, the expected utflity for the low-cost sender to escalate is dependent upon how the game is played. No assumptions are made concerning Ul (C) or U,(Q . other than that specified in equation (7). Ultimately, domestic audiences determine where Ul (C) and U,(C) are situated in the payoff rankings. The structure o f the game leads to easily identified payoff rankings for the high- cost sender, assuming d > h.s for all s. Because a high-cost sender always loses a war, suffering greater audience costs than a mere failure of a sanction policy, 1 find that: (14) Uh*(A) > U„*(s=0) > Uh*(B) > Uh*(C) > Uh”(C). Thus, assuming complete information, the proper subgame equilibriiun requires that a high-cost sender choose backdown rather than war under all conditions. The payoff rankings for all targets also are easy to identify. It is always the case that targets prefer the status quo and suffer no sanction. If the sender does impose a sanction, targets optimize their utility where they fight and the sender backs down. However, if war is inevitable, all targets would rather backdown than go to war. 1 include in the rankings the utility the target realizes if a war occurs between the target and a high-cost sender, although this terminal node is never reached. Specifically, the target’s payoff rankings are: (15) Uj(s=0) > Uj(B) > Uj(A) > U/C) > U/(C). 83 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4. Domestic Audiences, Economic Sanctions, and Incentives to Escalate Prior to a discussion of both the complete and incomplete information results, there is one aspect of the structural form still in need of specification— representing the relationship between economic sanctions, domestic audience costs, and incentives to escalate in a graphical form. For the case where d < w, this is represented in Figure 3.2. Note that on the horizontal axis, the level of economic sanction goes from 0 to S. On the vertical axis, the benefits and costs of the known variables are represented. There are many aspects of Figure 3.2. For example, the cost of war a sender against a weak target (w) and strong target (z.w) are drawn as horizontal lines “w” and “z.w”. The graph also displays two different domestic audience cost functions. The first begins at the origin and represents the audience cost a high-cost sender assumes for every level of sanction he imposes if the target fights and he backs down. This audience cost function is always below the cost of war (w) given that for the high-cost sender, it is always the case that the total cost of backing down is less than the cost of war. The second starts at d and shows that if a low-cost sender imposes a sanction and then backs down, he suffers both audience cost a(s) and the opportunity cost of d by choosing not to go to war, which he is guaranteed to win. The higher the self-imposed costs, the greater will be the domestic audience cost suffered. Formally, for all cases where s > 0, [ri[a(s)] / f/(s)] > 0. However, as sanctions increase, the marginal change in audience cost decrease, such that the second derivative is less than zero, [t/’ [a(s)] / rf(s)j < 0. 84 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. There exists two important sanctions in Figure 3.2. When s = s**, the audience cost generated by a failed sanction makes the low-cost sender government indifferent between war and backing down against all targets. Specifically, at terminal node 9, when the low-cost sender confronts the strong target, the utility of war is : (16) UL{(s**,war), (-,-); fight, -}=d-zw- is** while the utility of backing down is: (17) UJ(s**,bd), (-,-); fight, -} = -a(s**) - l.s**. Comparing (16) to (17) to examine which strategy yields a higher utility, I find that (18) UL{(s**,bd), (-,-); fight, -} = UL((s**,war), (-,-); fight,-} = -d-a(s**) =-z.w. I assume that when the equality in (18) holds or the left-hand side is greater, -d-a(s**) > - z.w, the low-cost sender will choose to go to war. ]î-d-a(s**) < - z.w, the sender always selects to backdown against a strong target. By similar reasoning, when s = s*, the domestic audience cost generated by a failed sanction episode is sufficient to make the low-cost sender indifferent between war and backing down against a weak target. A sanction, s, between s* and s** is marked for later reference. At s, the low-cost sender prefers to go to war against a weak target, but not against a strong target. Note that changing the initial value of d only changes the y-intercept of the low- cost sender audience cost function. For example, if d > z.w, then the function starts higher than z.w. This is modeled in Figure 3.3. In the game considered, let it always be the case that d < w. 85 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3.2: Modeling the Case Where < / < w (considered in this chapter) Cost, benefit a(s) zw w a(s) 0 - ^ Sanction level s* s** s imposed 3.3: Modeling the Case Where d > z.w (not considered in this chapter) Cost, benefit a(s) d zw a(s) Q Ï Sanction level s* s** s imposed 86 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5. Complete Information Results {d<w) PROPOSITION I. With complete information, wars will never arise and the actors play one o f the following three strategies that are subgame perfect nash equilibria: (19) {(s**,war), (-, -) ; (bd), (-) |p = 1, ^ = I } (20) {(s*,war), (-, -) ; (-), (bd) | /? = 1, ^ = 0} (21) {(-,-),(0,bd); (fight), (fight)} Proof A low-cost sender must establish domestic audience costs plus the value of the good in dispute to be greater than the cost of war if he is to effectively signal to a strong target that he is “locked in” to war, [d + a(s)] > (z.w). If the target is weak, the low- cost sender must set a sanction such that [d + a(s)] > (w). This occurs only when s > s**. A high-cost sender will always maximize his utility by playing no sanction (s=0) given that he prefers to backdown than go to war for any value of s. Backward induction, along with the intuition of equations (16) and (18), leads to these striking results. I first look at the case where a high-cost sender faces any target. Notice that at nodes 11 and 12, it is always the case that the high-cost sender will choose to backdown given the earlier statement that Uh*(B) > Uh*(C). With complete information, the target knows this, and thus if the target finds herself at node 7 or 8, she will select to challenge the sender, since it must be the case that Uj*(B) > Uj*(A). Moving up the game tree to nodes 3 and 4, the high-cost sender is confi-onted with the following dilemma: either impose some level of sanction yielding utility Uh*(B) = - a(s) - h.s, or do not impose any sanction, and derive a normalized utility Uh*(s=0) = 0. Since it is always the case that - a(s) ~ h.s < 0, for any s > 0, a high-cost sender will never impose a sanction. 87 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Now turn to the case where a low-cost sender confronts a strong type target. Sanctions serve as an effective signal if the benefit of realizing the goal plus the domestic audience cost suffered from a failed sanction attempt is greater than the cost of war. If there is no domestic audience cost, then d < w, and the low-cost sender prefers to backdown. However, including audience costs from a failed sanction, then the executive may prefer war. War involves costs, but after winning the war, he realizes d and avoids suffering domestic audience costs. Suppose [d + a(s)] > (z.w), such that at node 9, the subgame equilibrium requires that the sender selects war over backing down [Ul(C) > Ul (B)]. If the strong target is given the opportunity to move at node 5, she will always maximize her utility by choosing to backdown [Ug(A) > Uo(C)]. The target is aware that if she selects to fight, her utility will be Ug(C) = - ,r -e(s) -y; whereas if she backs down, she obtains a higher utility, Uo(A) = - x -e(s). The sender then prefers to reach node 5 so long as U^fA) =d-l.s>0 = Ul(s=0). However, for the sender to make the sanction threat credible, he must inflict self-imposed costs sufficient to cause the cost of backing down to be greater than the expected benefit of war. Specifically, the sender wants to set a sanction which maximizes Ul(A) subject to the constraint that [d + > Z .W . In the case where d<w,{\ is easy to see from Figure 3.2 that this occurs at sanction level s**. Similar reasoning can be applied to when a low-cost sender confronts a weak target. QED. 88 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis o f Results Proposition 1 clearly demonstrates how the self-imposed cost of sanctions can credibly signal a threat to escalate. A low-cost sender may only need to impose a modest sanction to achieve his foreign policy goals so long as a rational target recognizes that if sanctions fails, the executive in the sender will escalate the conflict to avoid domestic audience costs. In other words, even if war is more costly than the expected benefits ((/ - w < 0), a sender can cheaply realize positive returns from credibly threatening to go to war if the target believes that there is a sufficient audience costs associated with failure. The more sensitive audience costs are to self-imposed costs, the more likely a “cheap” sanction will work. All other variables, such as relative military capabilities and national interest do not affect conflict escalation. But they do have an important influence in determining what level of sanction is necessary for a sanction to be a credible signal. This result also helps explain Hufbauer, Schott, and Elliott’s finding that sanctions are more likely to work if they are imposed quickly with heavy costs rather than in small increments.'® This differs from much of the literature that suggests that Hufbauer, Schott and Elliott Economic Sanctions Reconsidered 99-101. 89 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. small sanctions are initially better than large sanctions because the former allows for more flexibility in negotiations.* ‘ By imposing a large sanction, audience costs are generated quicker, and thus more credibly threaten conflict escalation.’ ’ In the case where [d + a(S)] < z.w, such that even at the technically feasible maximum sanction, S, the domestic audience cost of backing down plus the opportunity cost of is less than the cost of going to war, the proper subgame equilibrium at node 9 specifies that the low-cost sender backs down because no level of sanction can credibly communicate a preference for war against the strong target. A strong target will challenge the sender at all s, such that the sender would prefer not to impose any sanctions. Formally, if [d + a(s)] < (z.w), for all s e [0,S], a low-cost sender will not impose any level of sanction (s=0) if the target is strong and the proper subgame perfect equilibrium is {(0, bd), (-,-) ; (fight), (-) | p = 1, q = 1}. Similar reasoning is used when \î[d + a(S)] < w. Finally, note that the complete information game requires that the low-cost sender credibly generates domestic audience costs to coerce the target to backdown. If the sender does not impose any sanction, s = 0, then the target recognizes that the sender " Doxey, International Economic Sanctions. Careful readers will note that there is one more pure equilibrium strategy to the complete information game: {(0,bd), (0,bd), (fight), (fight) | d < w}. This equilibrium states that when a low-cost sender receives less utility by either going to war or backing down fiom a sanction episode than it would by not imposing any sanction [Ul(s=0) > U^C) > 11^(8) ], a target can deter the sender from ever imposing sanctions by stating that she will always fight. However, this equilibrium neither makes intuitive sense, nor is sequential. In this model, there is no means by which a target can credibly commit herself to fight the sender prior to the sender actually imposing sanctions. Similarly, once sanctions are impose, the target suffers no cost if she “changes” her position. 1 leave for future research the case where the target government suffers domestic audience costs for changing her position. 90 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. does not believe coercing the target is important enough for him to suffer self-imposed costs. This communicates to the target that she can continue with her behavior without fear that the sender will escalate. Furthermore, given that d<w,ii will always be the case at s = 0 that the low-cost sender would never choose war over backing down. The only means by which the sender credibly convinces the target that he prefers war to backing down is by committing himself to suffering domestic audience costs if a sanction policy fails. 6. Incomplete Information Results {d < w) The identical game structure with incomplete information shows that sanctions can work as an effective signal of future escalation imder certain conditions. Under other conditions, however, sanctions can "entrap” a sender into committing military force when at the initial stage of the game war is suboptimal. This result is significant. In a game o f incomplete information, there is a positive correlation between economic sanctions and conflict escalation even when d<w. Specifically, given reasonable beliefs, perfect Bayesian semi-pooling equilibria exist where a low-cost sender may impose a sanction with a positive probability of signaling success, but also a positive probability that he will be forced into a war. This result is contingent upon the high- cost sender attempting to “bluff” the target into backing down, which the high-cost sender does with positive probability. The separating equilibrium and the semi-pooling equilibria are considered in respective order. 91 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (A) The Separating Perfect Bayesian Eauilibrimn PROPOSITION 2: A separating perfect Bayesian equilibrium exists whereby a low-cost sender can credibly use sanctions to signal a willingness to escalate, inducing the target to backdown. A high-cost sender imposes no sanction. Specifically, the following is a separating perfect Bayesian equilibrium strategy: (22) ((s', war), (-, -); (bd), (bd) | /? = I } for the low-cost sender and all targets, (23) {(-, -), (0, bd); (fight), (fight) | p = 0} for the high-cost sender and all target. Proof. The unique separating perfect Bayesian equilibrium is for the low-cost sender to impose a sanction equal to s' = (l/h)*d, whereupon any target recognizes that the sender is of low-cost and all targets backs down. Another separating equilibrium states that upon observing a zero level sanction, s = 0, the target believes with probability one that the sender is high-cost. Assuming that s' < S and s' > s**, the two separating equilibrium strategies are therefore {(s', war), (-, -); (bd), (bd) | p = 1} and {(-, -), (0, bd); (fight), (fight) |p = 0}.'^ The separating equilibrium requires that the target unambiguously knows that she is confronting either a low or high-cost sender. For if she knows she is confironting a low-cost sender, and s'>s**, the target will backdown after observing s'. If she observes s = 0, then the target knows she is facing a high-cost sender. " If necessary, the sender will increase the sanction to s** where s' < s** for the separating equilibrium to hold, otherwise a low-cost sender prefers not to challenge the strong target. Without any loss in the robusmess of the model, it is always assumed that s' > s**. 92 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I first consider the case where the target observes s'= (1/A)*d. A low-cost sender is guaranteed at s'=(l/h)*d that the sanction will credibly signal to the target that the sender is of type low-cost. This is because at s'= (1/A)*d, if the high-cost sender there to imitate the behavior of the low-cost sender, the highest utility the high-cost sender can receive is equal to that of not applying any sanction. Specifically: (24) Uh {([l//i]*d, bd); (bd), (bd)} =d-h.s = d - A.(l/A)*d = 0. Yet such behavior by the high-cost sender can not be sustained off the equilibrium path. Given the same expected utility of doing nothing (s=0), a high-cost sender always prefers not to impose a sanction than imposing s' = (l/h)*d. Intuitively, a high-cost sender prefers a guaranteed utility of 0 at s=0 rather than risk receiving utility Uh*(B) or Uh”(B) in the mistaken chance that the target fights. Finally, it is easy to see that a low-cost sender would never impose a sanction greater than s'=(l/A)*d, because it would only reduce his utility without changing the outcome of the game. Since all targets are worse off fighting a low-cost sender than acquiescing, all targets backdown at any sanction equal to, or greater than, s’. I now consider the case where the target does not observe any sanction (s=0). Only a high-cost sender would choose this action. Since it is always the case that the target backs down at s'= (l/h)*d, the utility of the low-cost fi'om imposing s' is: (25) Ul {([l//i]*d, war); (bd), (bd)} = d - (/ / h)*d > 0 = UJs=0]. 93 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Since /< ft at all s, the low-cost sender will always impose the separating equilibrium rather than maintain the status quo when s' is technically feasible. Only the high-cost sender has incentives to impose no sanction. QED. Analysis o f Results This simple model again demonstrates the powerful effects of self-imposed domestic audience costs. With both complete and incomplete information, at s=0, it is always the case that the target will believe that the sender is of type high-cost and will not engage in conflict escalation. It furthermore shows that even a low-cost sender would almost never impose s=0. Thus, even if the cost of war exceeds the benefit derived from achieving a foreign policy objective, the low-cost sender can use self- imposed domestic audience costs to realize his goals. If the separating equilibrium is technically feasible, the low-cost sender will never impose a sanction s=0. As in the game of complete information, “relative military capabilities,” and “level of interest” significantly effect the sanction necessary to credibly commit the low-cost sender to choose war over backing down in his final move. But the separating perfect Bayesian equilibrium is contingent upon a strong assumption that may not be empirically supportable. Specifically, the separating equilibrium assumes that s' is technically feasible, S > (l/h)*d. One may question, however, whether most low-cost senders are capable o f such a self-imposed cost. In other words, is it sensible to state that S > (l/ft)*d? Empirically, most senders are large economic states whose trade with the target as a percentage o f his GDP is small. 94 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Simultaneously, most democratic senders m ^ place a high value on dy if only because there is a belief in m ^or democracies, particularly in the United States, that “morality” should guide U.S. foreign policy.^* It is reasonable to assume then, that many situations arise where the sender can not fbasiblely impose the separating equilibrium for d will be relatively high and the lack o f trade dependency would limit s. Intuitively, there is only so much “self-imposed” cost a sender can inflict upon himself, and if that level is not high enough relative to foreign policy objectives, no level of sanction can convince a target that she is 6cing a low-cost sender. Based on this result, the separating equilibrium requires that the target economy is dependent upon the sender.'^ Note that if the separating equilibrium is not technically feasible, then the following is a perfect Bayesian equilibrium: (26) {(0,bd),(0,bd); (fight), (fight)}. (B) StmiejMwlinfe F ttftrt Biywwn Equilibriii The analysis must also be concerned if the separating equilibrium can be strictly dominated by another strategy.'^ Even if the separating equilibrium is a perfect Bayesian, it does not imply that the corresponding strate^ will always be played. Kissinger, Henry Diplomacy (New York, NY: Smion & Schuster, 1994). While Henry Kissinger is best known as a Realist, discounting the importance of “morality” in determming the costs and benefits of particular foreign poliqr actions, he does make it clear in his book that American foreign policy has historically been guided by “m orali^ more dian "ratirmal" self-interest Hufbauer, Schott, and Elliott empirically find this result Hufbauer, e t al Economic Sanctions Reconsidered. " Gibbons, Robert Game Theory fo r Applied Economists (Princeton, NJ: Princeton University Press, 1992). 95 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. There are multiple perfect Bayesian equilibria in this game and beliefs will influence which strategy is preferred. It therefore becomes a necessary exercise to find a perfect Bayesian equilibrium that is not strictly dominated given beliefs. Before continuing, however, note that this aspect of the incomplete information game requires two strong restrictions to the model. First, assume that the low-cost sender does not play a strategy designed to affect the play of the high-cost sender. This restriction results in either both senders pooling with respect to the action {s}. or a different separating equilibrium where the low-cost sender imposes a sanction less than s' and the high-cost sender elects not to bluff. Second, I assume that senders do not mix thier behavior with respect to the sanction imposed, but may mix their behavior in the decision to escalate. These restrictions both reduce the number of potential equilibria and facilitate the presentation of the argument demonstrating that sanctions can work by threatening costly future escalation. Given these restrictions, three results emerge: (1) sanctions have a positive probability of signaling resolve, (2) wars may arise between a low-cost sender and a strong target, and (3) wars may arise between a low-cost sender and a weak target. The rest of this section formalizes these findings by showing that if a low-cost sender is to play a sanction other than s' or s=0, he will play either s** or s. Lemma I: Ifd<w, fo r any level o f sanction less than the separating equilibrium sanction s 0 < s < //A*d, i f a low-cost sender elects to impose a sanction, he will choose to impose a sanction o f either s** or s. Otherwise, he will not impose a sanction. 96 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The structure of the game makes this result straightforward. The target suffers a sunken cost for any level of sanction, defined by the function e(s)}'^ Thus, the sender is aware that changing the level of sanction does not affect the target’s strategy of play. Because the target’s cost of war and cost of conceding defeat are exogenous, additional sanctions can only affect the target’s decision by changing the sender’s payoff structure. However, the model reasonably sets as exogenous both the sender’s cost of war and the pre-crisis value of the good in dispute. This leaves the audience cost function a(s) as the only variable that is affected by sanctions. 1 first demonstrate that no sanction less than s* is ever imposed other than s=0, then show that no sanction greater than s** is imposed other than s', and finally, prove that if a low-cost sender chooses to impose a sanction between s* and s**. he selects s. Figure 3.2, along with equations (16) - (18), establishes why no sanction less than s* is preferred by the low-cost sender to either s=0 or s=s'. For all s < s*, there are insufficient audience costs generated to induce a low-cost sender to play the action c = {war}. Independent of beliefs, at every sanction less than s*, all targets play b = {fight}, whereupon the sender elects to backdown. This rather strong assumption serves to maintain the integrity of the model. If e(s) induces the target to change her behavior, then sanctions work not by signaling a threat of escalation, but instead by operatmg the way most sanction scholars expect sanctions to work-by imposing high costs. Chapter 1 recognized that this often fails. However, the model does not discount the possibility that sanctions can work through this mechanism. 97 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Similarly, a low cost-sender will never impose a sanction greater than s** other than the separating equilibrium sanction s'. The intuition can also be easily derived from Figure 3.2. At s**, targets understand that if given a final move, the low-cost sender prefers to go to war than backdown. Any sanction level greater than s** only reduces the sender’s welfare by increasing the economic costs associated from higher sanctions, without changing either his action in the final move or the strategic thinking of the target. Finally, consider the case where the sender imposes a sanction between s* and s**. Why might a sender choose such a sanction, which is labeled as s? First, it may be the case that the only technically feasible sanction is less than s**. Second, and more theoretically interesting, a low-cost sender can effectively coerce both types of targets to backdown at s. Clearly, a weak target may be deterred from fighting a low- cost sender at any sanction greater than s*, although incomplete information results in the weak target sometimes playing b = {fight} hoping that the sender is of type high- cost. A strong target may also choose to backdown after observing s. The intuition here is slightly more complicated. In a game with complete information, s is insufficient to generate audience costs large enough to induce the low-cost sender to choose war over backing down against a strong sender. Thus, a strong sender always chooses to fight after observing s in the presence of frill information. This is not true, however, in a game of incomplete information. Given that the target knows that a low- cost sender is uncertain of the target’s type, and that a low-cost sender prefers to go to war against a weak target than backing down, there is a positive probability that a low- 98 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. cost sender will go to war against a strong sender after imposing s if the low-cost sender finds the expected benefit of war against an unknown type target to be greater than the expected cost. Knowing this, the strong sender backdowns with positive probability. The structure of the game determines s, as well as the probability that the strong target backs down. The low-cost sender never imposes a sanction greater than s because it would not change the outcome of the game while increasing the economic cost of sanctions for the sender. I leave for play of the game to find the exact value of s. Conditions for Imposing s** This section demonstrates the following two are perfect Bayesian equilibria that strictly dominate equilibria when a low-cost sender imposes s=0 or s': (27) {(s**, war), (s**, bd); (bd), (bd) | p > [x / (x+g.y)]} (28) {(s**, war), (s**, bd); (fight), (bd) | [x / (x+^.y)]<p < [x / (x+g.y)], q = I } This section also shows the beliefs necessary for the low-cost sender to find that the strategy defined in (27) and (28) strictly dominate the separating strategy defined in (22). The next section considers if the strategy defined in (27) and (28) strictly dominates semi-pooling equilibria where the low-cost sender imposes s. For these semi-pooling equilibria to hold, 2 conditions must be satisfted-the low-cost sender must prefer to impose s** rather than s' and the high-cost sender must find it in his interest to bluff about his type. These conditions are established later. For now, consider the strategic thinking of the target after she observes s** assuming that all senders are willing to impose s**. 99 Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. After observing s** at nodes 5,6, 7, and 8, is there any information that informs the target whether she is facing a low-cost or high-cost sender? No. Because of the restriction that the equilibrium must pool, both types of senders impose s** and Bayesian updating is not possible. Furthermore, the conditions established in equations (16) - (18) show that at nodes 9 and 10, the low-cost sender will always select war, while at nodes 11 and 12, the high-cost sender will always choose to backdown. Given that these actions hold true anytime a target plays b = {fight}, there is no mixed strategy a target can play to discourage a low-type sender fi’ om going to war. Thus, after observing s**, targets are left with their initial beliefs to determine if they wish to backdown or fight. Lemma 2: Ifd < w , upon observing sanction s = s**, a strong target will fight if 0 < [ X i (x+g.y)] while a weak target will fight i f ô < [ x ! (x+^.y)]. Targets must determine which is better— backing down or fighting relying prior beliefs. I find this by examining expected utilities. Based on the structure of the game, the expected utility for all targets to backdown is equal to: (29) E [U j {(s**, war), (s**, bd); (bd), (bd)}] = - x - e(s). The expected utility of a strong target going to war is equivalent to: (30) E [ U g ( ( s * * , war), (s**, bd); (fight), (-)}]= Ô * (- x - e - g.y) + (l-ô) ♦ [- e(i)]. Setting (29) < (30), a strong target will prefer to fight if and only if (31) 0 < [x /(x + g .y )]. A similar technique for a weak target reveals that she will fight if and only if (32) Ô < [x / (x+t.y) ]. 100 Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. Given that payoffs are known, a sender can costlessly calculate whether a target of any given type will backdown upon observing s**. This results in three possible cases: where both types of targets fight, both types of targets backdown, and the strong target fights while the weak target backs down. These different cases will be used to support the following proposition. PROPOSITION 3: I fd < w, s** is an effective signal o f resolve fo r both types o f senders i/ô < [x / (x+Ly) ], but will not be imposed i/ô > [x / (x+g.y)]. In the case where [x / (x+^.y) ] < Ô < [x / (x+g.y) ], the low-cost sender will impose s** rather than s ' given certain beliefs, even though sometimes sanctions will work, and other times sanctions will result in an undesirable war. Specifically, the following are semi-pooling perfect Bayesian equilibria: (33) {(s**, war), (s**, bd); (bd), (bd) 1 p > [x / (x+g.y)]} (34) {(s**, war), (s**, bd); (fight), (bd) | [x / (x+A:.y)]<p < [x / (x+g.y)], q = 1}. The proof of this proposition is derived from analyzing the three possible cases. CASE I: All targets backdown. If Ô > [x / (x+g.y)], such that all targets backs down upon observing sanction s**, then all senders impose s** rather than the separating sanction s' or s=0. Proof. Intuitively, if the expected utility for the target o f fighting is less than that of backing down, the sender should exploit both his relative military strength level of interest by imposing the semi-pooling equilibrium sanction, s**, given that at both s' and s** all targets backdown. The low-cost sender will always prefer this pooling equilibrium to the separating equilibrium sanction s' = (1/A)*d, which involve greater economic costs. A weak sender can costlessly “bluff,” and also impose sanction s** with the knowledge that all targets will backdown. Specifically, if the target places very 101 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. little weight on the foreign policy objective of the sender (x), or a lot of weight on the cost of war (g._y), then if there is a relatively high a priori belief that the sender is low- cost, then the target would prefer to backdown than risk a costly war against a low-cost sender. Reasonably assuming that d - h.s** > 0, both types of senders will impose s** rather than s '. No other strategy off the equilibrium path considered so far dominates that specified in (34) and (35) given these beliefs. QED. Analysis o f semi-pooling equilibria This equilibrium is truly startling. It suggests that if the target believes that the cost of war would be extremely high or that x is not highly valued, then the target would backdown after observing s** given that she believes a priori with high probability that the sender is low-cost and is likely to escalate if she does not backdown. This may explain why the only case where sanctions have proved “successful” between 1983 and 1990, according to Hufbauer et. al., was U.S. sanctions against El Salvador if the El Salvadorian government granted a pardon to three terrorists convicted of murdering 12 people. From the perspective of El Salvador, the cost of backing down (x) was insignificant relative to the potential cost of escalation ik.g). Note that this finding is consistent with mainstream rationalist arguments that level o f interest influences who wins international contests as well as their claim that the relatively stronger military power is more likely to realize his goal by threatening costly wars. 102 Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. CASE II: All targets fight. If Ô < [x / (x+k.y)], such that a target will always fight upon observing sanction s**, then neither a high-cost sender nor a low-cost sender will impose sanction s**. Off-equilibrium path play shows that the low-cost sender receives a higher expected utility if he imposes s' where technically feasible, or s=0 when s' is not feasible, relative to s** and having to fight the target. Proof. Suppose both types of targets either highly value their current behavior (x) or significantly discount the cost of war (T^.y), such that both target types will prefer to fight than backdown upon observing s**, hoping that they will confront a high-cost sender rather than a low-cost sender. If a high-cost sender knows this ahead of time, which he does because payoffs and a priori beliefs are common knowledge, he will prefer to impose no sanction. However, given the model’s restrictions, if the high-cost sender elects never to impose a sanction under such conditions, now becomes the separating equilibrium. But considering the case where both types of senders pool, such that the high-cost sender has incentives to impose s**, if a low-cost sender knows that if he imposes sanction s** and ô < [x / (x+k.y)], war becomes inevitable. No equilibrium can be supported with such beliefs given play off the equilibrium path. The low-cost sender would either prefer to impose the separating sanction s' to ensure a positive rate of return, or no sanction at all (s = 0), than realize lower utility by suffering s** economically as well as the cost of war. QED. Analysis This result is very interesting. Suppose the goal of the sender is to destabilize a government, end a military campaign, or induce significant reforms that would fundamentally change the structure of the government, such that the target backs down 103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. only if the a priori probability that the sender is low-cost is high. In that case, all targets prefer war to backing down at sanction s**. Sanctions fail to signal resolve in such a case, and their imposition can only lead to a sub-optimal outcome. As most targets are governed by leaders that highly value remaining in power, such that .r is extremely high, and may not be concerned about the relative cost of war, such at k.y is low, incorrect beliefs by the sender may result in war if he chooses to impose s**. CASE III: A strong target fights while a weak target backs down. A semi-pooling equilibrium exists where both types of senders impose sanction s**. and a strong target will fight while a weak target will backdown upon observing s**. The high-cost sender has incentives to “bluff” about his type and imitate the behavior of the low-cost so long as his prior belief o f the target backing down is 1 < [{d - h.s**) / (a(s**) + d)\. A low- cost sender prefers the pooling equilibrium to the separating equilibrium strategy when ^ h)*d-l.s**) / z.w]. The pure equilibrium strategy is specified as follows: (35) {(s**, war), (s**, bd), (fight), (bd)} | [x / (x+k.y)] < p < [x / (x+g.y)], q = I }. Proof. All senders are aware before the first move in the game that in this case, a strong target will fight and a weak target will backdown. If the low-cost sender holds a high a priori belief that the target is weak relative to the cost of conflict escalation or value of the good in dispute, he will have incentives to play s** to coerce the weak target while decreasing the economic cost associated with playing s'. It therefore is necessary to establish the conditions where a low-cost sender would be willing to risk a war against a strong target by imposing s** hoping that the target is weak, as well as consider under what conditions a high-cost sender is willing to risk confironting a strong target rather than not impose a sanction. 104 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Finding the beliefs necessary to support the semi-pooling equilibrium can be accomplished by comparing the sender’s welfare if he plays s' and the welfare of playing s**. Recall that in the separating equilibrium for the low-cost sender, s ' = (l//i)*d, such that the utility he realizes playing the separating equilibrium is equal to (36) Ul (A) = U l ( ^ = d - /.j ' = d - l.(d/h) = d - (I/h) *d. However, if he chooses to play the semi-pooling equilibrium specified in (35), his expected utility is (37) E[Ul {(s**, war), (s**, bd); (fight), (bd) | [x / (x+k.y)] < p < [x / (x+g.y)], q = I }] = (A.)*Ul(C) + (1-A.)*Ul£M = (m -a (s ) - l.s**) + (I - k)*id-Ls**) = [d- /..y**] - [(A)*(afj**^ + d)]. Comparing (36) to (37), a low-cost sender prefers to risk a war against a strong target in the belief that sanctions will credibly signal resolve to a weak target when a priori beliefs are such that (38) {d- is**] - [(l)*(a(s**) + d)\ >d-(//A)*d. Solving for 1, the a strategy employing s** rather than s' is perfect Bayesian if and only if a priori beliefs for the low-cost sender are such that (39) I < {i(\lh)*d\ - is**) / [a(s**) + d\= ([i\IK)*d\ - is**) / [z.w]. So long as (39) holds, then a low-cost sender prefers the semi-pooling s** to s'. Finally, the conditions where a high-cost sender would be willing to “bluff’ and imitating the behavior of the low-cost sender occurs where he is indifferent between imposing sanction s** and not imposing any sanction. This happens when 105 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (40) E[U„*{(s**, war), (s**, bd); (fight), (bd)}] = U„*{(s**, war), (0,bd); (bd), (bd)} (A)*U„*(B) + (l-X)*U„”(A) = 0 (l)*(-a(s**) - h.s**) + (1-X)*(d - h.s**) = 0. Solving for k, a high-cost sender is willing to play s** if and only if (41) X < [(d-A.j**)j / [a(s**) + d\}^ Under the conditions established in (39) and (41) the equilibrium defined in (35) is perfect Bayesian and is not, up to now, dominated by other strategies. QED Analysis o f semi-pooling equilibria It is interesting to find that there exists a semi-pooling perfect Bayesian equilibrium where both sender types impose s** and war is a rational outcome. Unlike the complete information game-where s** generated sufficient domestic audience costs to signal resolve— the same does not hold true in the case of incomplete information. Instead, a low-cost sender is willing to risk a war under the conditions established in (39) when the weak target will always backdown and the strong target will fight. '* A reasonable restriction in the model stipulates that at all levels of s e (s**, s’), there is no sanction which the low-cost sender can impose that will deter a high-cost sender from pooling given the conditions set forth in (41). This restriction is necessary given that a low-cost sender may try to deter a high-cost sender from copying the low-cost strategy in action space s e (s**, s’). At s**, a marginal increase in the economic cost of sanctions will be greater than the marginal increase in domestic audience costs, such that the right-hand side of (41) decreases with an mcrease in s**. I assume that the only separating equilibrium is when the low-cost sender imposes s’. 106 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. It must now be questioned if this perfect Bayesian equilibrium is consistent with the empirical evidence. While such a rigorous empirical test is not performed until Chapters 4 and 5, a preliminary analysis can be conducted by considering the perspective of the target, and then the sender. Is it reasonable to suppose that [x / (x+k.y)] < ô < [x / (x+g.y)]? Probably. If a strong target values her current behavior highly relative to that of war, then [(x) / (g.y + x)j will be close to 1 ; which makes intuitive sense for three reasons. First, by design, many strong target governments are relatively unconcerned about the economic and social welfare of their citizens either because they lack legitimacy or because the population is being repressed. Since war generally harms a society that the government may not value, it is reasonable to expect that g will be extremely low relative to L Second, after destabilization, most strong target governments have little concern about the state of their country. If there is a war, the target government knows that it is going to lose power. Once out of office, what incentive is there for a dictator to worry about the fortunes of his coimtry? Such a defeated leader is likely to be killed (Anastasio Somoza), arrested (Manuel Noriega), or so ostracized by his country (Ferdinand Marcos) that there is little incentive to worry about the state of society. Thus, the cost of war may not be high. Third, target governments often highly value remaining in power, suggesting that x would be high relative to war. Conversely, it is reasonable for a sender to expect that a weak target who faces signiffcant and organized domestic opposition ffnds the cost o f conflict escalation to be high relative to x. Specifically, if there is extensive domestic opposition that is not 107 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. being repressed, lack of popular support suggests that x should be relatively small while conflict escalation can be extremely costly, particularly if a civil war erupts. U.S. support for the Contras in Nicaragua and UNITA in Angola represents two cases where the cost o f escalation was higher for the target government then war would have been if the target was strong. Alternatively, fragmentation within society is usually reflected in the military; thus increasing the probability of a military coup as the crisis between a sender and target escalates. Sender initiated wars not designed to destabilize a government may therefore lead to such results if the target is believed to be weak. It is reasonable to believe, therefore, that [x / (x+k.y)] < Ô < [x / (x+g.y)] often occurs in the “real world” given the different values of exogenous variables. Now consider if a low and high-cost sender prefers to impose the semi-pooling equilibrium s** rather then the separating equilibrium s'. Is it reasonable to suppose for the low cost target that X < {[{\lh)*d\ - is**) / [a(s**) + d\= {[{\lh)*d\ - is**) / [z.w]? Probably. First, the separating sanction s' =(1/A)*d is often not technically feasible, in which case a low-cost sender is always better off imposing sanction s** in the hopes of coercing weak targets to backdown. Second, if a low-cost sender either highly values the good in dispute (d) or finds the relative cost of war low (z.w), a low- cost sender is more willing to risk conflict escalation so as to reduce his self-imposed economic cost (l.s). This seems reasonable for a sender such as the United States that often places significant weight on d and may consider the cost of war low given its relative military capabilities. 108 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Note that this equilibrium is consistent with the argument of mainstream rationalists that interest and relative military capabilities influences the decision to escalate. Interestingly, this equilibrium also suggests that a low-cost sender will only go to war against a strong target, as all weak targets would backdown. This is consistent with the work o f Gilpin, who shows that states are most likely to go to war when there are marginal changes in relative power. Finally, this equilibrium allows tor “blufflng,” as well as demonstrates the importance of domestic audience costs in credibly signaling resolve. Hence, the model has demonstrated the utility of sanctions in signaling resolve through the generation of domestic audience costs while correcting for deficiencies in the literature as mentioned in Chapter 2. Conditions for Imposing s There exists one other semi-pooling level of sanction that a low-sender may wish to impose rather than s** or s' that forms a perfect Bayesian given certain beliefs. The intuition behind a low-cost sender playing s is presented before considering the constraints necessary to support the following as perfect Bayesian equilibria: (43) {(§, war), (s, bd); (fight), ((l-r).bd, r.war) | [x]/ [0.x + ô.k.y] < p< [r] / [dx + àg.y]} (44) {(§, war), (s, bd); (bd), (bd) \p> [x] / [dx + ô.g.y]} where(r) = [(X)*(z.w-d -a (s} )1 / [(1- A)*(d + a(s^ - w )]. 109 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Suppose that a low-cost sender highly suspects that a target is “weak.” Under complete information, a sanction of s* is sufficient to convince a weak target that a low- cost sender would prefer war to backing down at terminal node 10 in the extensive form of the game. However, s* will not generate sufficient audience costs to make choosing war superior to backing down against a strong target. Thus at s*, if the target fights, the low-cost sender prefers war to backing down against a weak target, but backing down to war against a strong target. Incomplete information leads to different results. A sanction of s* will not be sufficient to induce a weak target to change her behavior given that the low-cost sender will not select the action “war” in the final move of the game. Given a positive probability that the target is strong, a low-cost sender would always prefer the separating equilibrium, (s = s'), the status quo, (s = 0), or s**, to a non-credible threat o f escalation at s*. At s*, weak targets can exploit the sender’s lack of information to “bluff,” and act like the strong sender, knowing that the low-cost sender will always backdown. This is because the strategy {(s*, war), (s*, bd); (fight), (fight)} is not perfect Bayesian as strategies off the equilibrium path strictly dominate. Specifically, the expected utility for a low-cost sender playing this strategy is: (45) E[Ul{(s*, war), (s*, bd); (fight), (fight)}] = (A.)*Ul(C) + (1-^)*UjXQ = d - w - U * - (A.)*(z.w - w). But playing the strategy {(s*, bd), (s*, bd); (fight), (fight)} yields a utility of (46) Ul{(s*, bd), (s* bd); (fight), (fight)}] = - a(s*) - l.s*. 110 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Thus, a weak target will always fight after observing s* given that (47) - a(s*) < d - w - (A,)*(z.w - w) for all A , > 0. While low-cost sender never imposes s < s*, he can find a sanction s, s* < s < s**, such that (48) -a(s)> d - w - iKpà^*(z.w - w) where is the sender’s updated belief that he taces a strong target at his last move. To test whether a sender will impose s hoping to both coerce a target to backdown and reduce the economic cost associated with sanctions relative to imposing s' or s**, requires that the game is solved with backward induction using beliefs.'® The results show that a low-cost sender is willing to impose a sanction between s* and s**, but under such conditions, sanctions will sometimes work, but other times playing s can lead the low-cost sender to enter a war against either type of target. The first step to solving this game is to find the updated belief that the sender holds in the last move that the target is strong that makes him indifferent between war and backing down, the so-called updated critical value ( À g ^ J If the target elects to fight, the high-cost target will always backdown. However, the low-cost target will be indifferent between backing down and war when the expected utility of war is equal to that of backing down, given updated beliefs which have been set to A,„t. Specifically, Morrow, James, Game Theory fo r Political Scientists (Princeton: Princeton University Press, 1994) 203-210. ^ As before, assume for now that both the low-cost sender and high-cost sender will pool at s. I l l Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (49) E[U J(s, war), (s, bd); (-), (-)}] = (A .upda.e)*U L (C ) + (I- = d - W - l . S - (% updaJ*(z.W - W ). Note that the strategy {(s, bd), (s, bd); (fight), (fight)} yields (50) Ul{(s, bd), (s, bd); (fight), (fight)} =-a(s)- l.s. Setting (49) equal to (50), a low-cost sender is indifferent between war and backing down so long as his updated beliefs are: ( 5 1 ) {{d + a ( ^ - w )j / ( z .w - w ) = XgnL If lu p d u K > ^cnt then the sender prefers to backdown than go to war. If < Ignt the sender prefers to go to war than to backdown. Such a finding makes intuitive sense. The higher the probability the low-cost sender faces a strong target in his final move, the more likely he backdowns rather than go to war after imposing s. Also notice that it is audience costs that drive the result. As § - s**, then a(s) - a(s**) and the left-hand side of equation (51) equals I, such that the low-cost sender always goes to war if the target fights. As s - s*, then a(^ - a(s*), and the left-hand side o f (51) equals 0, such that the low-cost sender would always backdown. Thus, the sender’s choice of s significantly affects the play of the game. After observing s, how will the target react? Because s is played by both the low and high cost sender, s alone does not provide any information about type. Thus, the probability, p, of the target being subjected to a low-cost sender invasion that makes the target indifferent between backing down and fighting must be found. Note that 112 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. unlike the case where the low-cost sender imposes s** and he always chooses to go to war if any target fights, in the case where s is played, the low-cost sender does not play war in his final move with probability 1. Thus, the strategic thinking of the target demands that one considers when a target is willing to risk a war after observing s. Applying a similar approach as that of Lemma 2, a weak target is indifferent between war and backing down when her expectation that fighting will result in a war is: (52) Pk = M / [< 5.x + ô.k.y] for the weak target. While that making a strong target indifferent between backing down and fighting is (53) P q = [x] / [Ô .X + ô.g.y\ for the strong target. An analysis of (52) and (53) as compared to (31) and (32), shows that both types of targets hold a higher tolerance for a risk of war after observing s rather than s**. This makes intuitive sense given that a low-cost sender prefers not to go to war against a strong target at s. Interestingly, it is the initial belief that determines the magnitude of difference between the target’s tolerance for a risk of war when she observes s relative to that of observing s**. These parameters are known to all parties before the target makes his first move. This information helps establish Proposition 4. PROPOSITION 4: I fd < w, s is an effective signal ofresolve for both types ofsenders ifPg — M / [< 5 .x + ô.g.y], but will not be imposed i f > M / [J.x + ô.k.y]. In the case where [x / (dx+Afc.y) ] < p < [x / (<5.x+dg.y) ], the low-cost sender will impose s rather than s ' or s**given certain beliefs, even though sometimes sanctions will work, and other times sanctions will result in an undesirable war. 113 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The following are semi-pooling perfect Bayesian equilibria: (54) {(§, war), (s, bd); (bd), (bd) | p > [x / (<5.x+Ag.y)]} (55) {(s, war), (s, bd); (fi^t), ((l-r).bd, r.war) \ [x / (dx+dA.y)]<p < [x / (<5jc+dg.y)]} where (r) = [(l)*(z.w - d - a(s) )| / [(1-1) * (d + a(s) - w)|. Proof The proof takes the identical structure as that when s=s**. Without repeating the entire exercise, it is known that with information available on all payoffs and a priori beliefs, a low-cost sender will know which of the three cases he is approaching before choosing what sanction to impose. These are outlined as in the previous section. CASE /: All targets fight. In this case a low-cost sender never imposes s. CASE II: All targets backdown. In this case a low-cost sender always prefers s to either s** or s’, and the high-cost sender “bluffs,” by also imposing s with assurances that he will receive a positive payoff. This supports the semi-pooling equilibrium specified in (54) and shows that it is perfect Bayesian. The reasoning is similar to that employed in Case U for s=s**. CASE III: Strong targets always fight while weak targets play a mixed strategy. With the low-cost sender not selecting war with probability 1 In his final move after imposing s, the weak target has incentives to “bluff,” acting like a strong target hoping to deter the low-cost sender firom choosing war. Knowing that when the low- cost sender holds an updated belief greater than the critical value (X „pd^,g> that he would rather backdown than escalate, a weak target will play a mixed strategy. Specifically, the probability, r, of a weak target fighting the sender that creates the critical belief using Bayesian updating, r = p(fight| target weak) is: 114 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (56) = p(target strong) * p(fight| target strong) p(target strong) * p(fight| target strong) + p(target weak) * p(6ght| target weak) Solving for r and substituting in the critical belief as defined in (51), a weak target plays b = {fight} with probability (57) (r) = [a)*(z.w - d -a (^ )\/ [(1- A .) * (d + a(^ - w)| and chooses not to fight the sender after observing s with probability (58) (1 - r) = [(if + a(^ -w) - (A,)*(z.w - w)\ / [(1- A .) * (d + a(s) - w)|. Thus, while a strong target backs down, a weak target will play a mixed strategy to induce a low-cost sender to be indifferent between war and backing down. Equations (57) and (58) are interesting in several respects. First, as a(s) approaches a(s**), the numerator approaches 0 and weak targets are deterred from bluffing. Thus, low-cost senders would like to impose s as close to s** as possible, given the constraint that as s nears s**, the sender suffers larger economic costs. Second, as the relative difference between a strong target and weak target increases, weak targets are more likely to try to bluff. Finally, the value of the good the sender wishes to obtain, d, has no affect on the behavior of the weak target. Only changes in the cost of war and domestic audience costs affect the target’s strategic decision. Having found the best replies at the last two moves, the next step is to verify that a low-cost sender maximizes his utility by playing s rather than the separating equilibrium s' or s**. 115 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Conditions where the Sender plays s rather than s ' First, recall that the utility of playing the separating equilibrium is: (25) Ul {([l//z]*d, war); (bd), (bd)} = d - (l/h)*d. Given that the target’s behavior makes the sender indifferent between backing down and war, the expected utility of a low-cost sender playing {s, war} can be written as: (59) E[Ul {(s , war); (fight), ((l-r).bd, r.w)}) = (X)*[U(B)] + (l-A.)*(r)*rU(B)l + (1-A.)*(l - r)*lU(C)l = (X)*[- a(s) - s] + (l-l)*(r)*[- a(s) - s] + ( 1-A)*( 1 - r)*{d - s). Setting (59) equal to (25), a low-cost sender will impose s rather than s’ when: (60) (d - l.s) - { [(l)*(a(s) + d)*(z. w - w)| / (d + a(s) -w)} > d - l.s . A similar technique for the high-cost sender finds that he would rather impose s than s = 0 when: (61) (d - A.s) - / [(l)*(a(s) + d)*(z.w - w)| / (d + a(s) - w) ^ > 0. Solving for X, the equilibrium specified in (55) is Bayesian perfect and not dominated by off-the-equilibrium path play by the separating equilibrium when initial beliefs for the low-cost sender are such that: (62) X < [ /.s' -/.§)/ [(Z.W - w)*(a(s) + d)| and for the high-cost sender when: (63) A < [ d - A .S I / [(z.w - w)*(a(s) + d)|. Finally, note that the low-cost sender will select s such that condition specified in (62) holds, otherwise, he would rather impose s’. 116 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Conditions where the Sender plays s rather than s** The mathematical demonstration showing under what conditions the sender prefers s rather than s** follows a similar structure. First recall that the utility of imposing s** for the low-cost sender as defined in (37) is: (37) E[Ul {(s**. war), (s**, bd); (fight), (bd) | [x / (x+k.y)] < p < [x / (x+g.y)], q = 1}] = (À)*UaC) + (l-À)*UL(Al = (X)*(-a(s) - is**) + (1 - k)*id - As**) = [d- A^**] - [(X)*(a(s**) + d)]. Setting (59) equal to (37), a low-cost sender will impose s rather than s** when: (64) (d - As) - /■[(X)*(a(s) + d)*(zw - w)| / (d + a(s) -w)f>[d-is**] - [(X)*(aAr**) + d)l Equation (65) shows that a high-cost sender prefers to impose s than s**, when, based on equations (40) and (61) (65) E[Uw {(s, bd); (fight), ((l-r).bd, r.w)}| > E[U /{(s**,bd); (fight), (bd)}] (d - h.s) - { [(A.)*(a(s) + d)*(z.w - w)j / (d + a(s) -w)} > (A)*(-a(s**) - A.s**) + (1- l)*(d-A .s**). Solving for (64) and (65), the equilibrium specified in (55) is Bayesian perfect and not dominated by off-the-equilibrium path play when: (66) [ (a(s) + d)*(z.w - w) / (d + a(s)-w )|<[ a(s**) + d ] for the low cost sender and (67) [ (a(s) + d)*(z.w - w) / (d + a(s) - w) j < [ a(s**) + d ] for the high cost sender. 117 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis o f semi-pooling equilibria Given proper conditions, there is a positive probability that a low-cost sender will go to war against both a strong and weak target. The analysis also showed, however, that sanctions can be effective as signaling tools against weak targets with positive probability. This equilibrium is consistent with the empirical observation that after a sanction episode, the U.S. has gone to war against both weak and strong targets. I first consider when a low-cost sender prefers s rather than s' and then if the low-cost sender prefers s rather than s**. While beliefs clearly influence the nature of this game, one still must ask if it is reasonable that a low-cost sender would impose s rather than s' by examining equations (62) and (63). Intuition suggests that for low-cost senders, s will often be played in the real world rather than s'. First, if the audience cost function is highly sensitive to sanctions, than the initial belief that the target is strong can be relatively low to support the equilibrium, as only a moderate sanction needs to be applied to “commit” a low-cost sender into applying military force. Alternatively, if a(s) is not sensitive, then the sender will prefer to impose either s=0 or s'. In the next two chapters, it is shown that self-imposed costs do affect domestic audiences costs. Second, if the cost of war for a low-cost sender is relatively similar against a strong or weak target, then the sender is more likely to impose s than s', given that nature must select with lower probability that the target is strong. Finally, as discussed earlier, strong senders may not be able to impose s ', in which case they may be willing to impose s given that a strategy involving s has a positive expected rate of return given the conditions established. 118 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7. Conclusion It is hoped that the rigorous mathematical structure did not distract the reader from the thesis of the chapter: sanctions can work by credibly signaling a willingness to escalate a conflict, inflicting the target with prohibitive costs, and thus inducing the target to backdown after observing a sanction. The separating equilibrium s' always credibly signals resolve, while s** and s can signal such resolve given certain beliefs. The variable driving the results is domestic audience cost. Only when sanctions generate self-imposed costs, increasing domestic audience costs, can sanctions effectively communicate resolve. While this chapter considered the case where d<w, similar logic can be applied to when d>z.woxw<d< z.w. The model also suggests that there may be unintended consequences to imposing sanctions. Because of incomplete information, the imposition of sanctions may result in undesirable conflict escalation. This result is consistent with the work of Fearon and Bueno de Mesquita who argue that domestic audiences explain military conflict. The payoff structures are such that a strong and weak target may play “fight” against a sanction, believing with positive probability that the sender is weak. Sanctions can work by signaling resolve, but it can also lead to conflict escalation. The next two chapters seek to empirically verify these results. 119 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Public opinion in this country is everything. Abraham Lincoln Chapter 4 : Public Opinion and Conflict Escalation in Sanction Episodes Introduction This chapter empirically supports the model developed in Chapter 3. Specifically, verification is sought for three hypotheses: (1) Self-imposed costs are positively associated with domestic audience costs, (2) The accumulation of audience costs are positively associated with escalation, and (3) Targets may backdown if they understand that the sender will escalate to avoid suffering domestic audience costs after he imposes a sanction, even if d<w. Proof of these hypotheses comes from two sources. Section I provides a summary of every major U.S. sanction episode since 1970, and compares the episode to the theoretical model of Chapter 3. Section II employes statistical analysis to provide further empirical evidence. However, prior to discussing the data, both the empirical literature on domestic audience costs is reviewed and the metholodogy I use in the empirical analysis is explained. A Review o f the Empirical Literature on the Domestic Audience Cost Proposition The existing literature that empirically tests the thesis that domestic audiences influence the likelihood o f going to war employ one of three approaches. The most common technique statistically tests overall presidential approval against the use of force, holding constant variables such as the misery index and year. The research performed by Charles Ostrom and Brian Job, and Patrick James and John Oneal 120 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. epitomize this method. However, while such research does contribute to understanding the relationship between domestic audiences and international conflict, deficiencies exist in its methodological structure. For example, factors such as the state of the economy influence overall approval more than foreign policy. Separating the effect of foreign policy on overall approval is extremely difficult, and narrowing the effect of the handling of a specific foreign policy is even more problematic. Can one credibly claim that the airstrikes against Bosnian Serbs were the result of a change in President Clinton's overall approval?' And even if there was a correlation, why did the administration select to employ force in Bosnia instead of Burma? Why did President Bush choose to use force against Panama, but not Columbia, given that the latter supplies more drugs to the U.S. than the former? The second technique that has been employed empirically tests an original model using case studies or statistical analysis. Work by Bruce Bueno de Mesquita and David Lalmer fall into this category. But there are also problems with this approach. The research by Bueno de Mesquita and Lalmer conclude that relative military capability does not affect escalation decisions, although it does effect which target is initially chosen. Furthermore, their empirical evidence is drawn firom self-selected cases, reducing the power of their test. ‘ The data in Chapter 5 even suggest that Clinton’s overall approval may have been increasing at the time of the August-October 1995 airstrikes. 121 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Finally, numerous single case study projects show that "domestic audiences" affect conflict escalation decisions. The work of James Lamare is indicative of this tactic. However, these type of studies do not employ a large enough sample size nor can they eliminate self-selection biases. But there is one problem transcending all the empirical literature that this project seeks to correct. With the exception of some individual case studies, research testing the domestic audience cost proposition, either directly or indirectly, fail to rigorously measure both the dependent and independent variable.* In most instances, either variables are unquantified or proxy variables are used. Additionally, most studies do not quantify “domestic audience costs,” but simply assume their presence. There has yet to be an empirical test of the domestic audience cost proposition in the context originally meant by Fearon and Smith— that of a correlation between shifts in public opinion in the handling of a specific foreign policy problem and conflict escalation. I seek to verify the original concept of the audience cost proposition by examining both public opimon polls and news media coverage surrounding U.S. sanction episodes, without biasing the selection of cases. * One important exception is John Vtuelleris work Public Opinion and the G ulf War. 122 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Method ofAnalysis and Selection o f Cases This project empirically tests the hypotheses using three methodological techniques. The first entails case study analysis of an exhaustive set of sanction episodes between the United States and targeted states since 1970. The second method employs statistical analysis to calculate a monthly predicted audience cost variable by using monthly, case-specific public approval data as the dependent variable and comparing it against an inflation-adjusted accumulative cost of sanction, degree of media exposure, relative economic capability of the target, and time. Another statistical examination is then undertaken to test for a correlation between the dependent variable "escalation" and independent variables relative military capability, monthly predicted audience cost, overall public approval, time, and Cold War alliance. Finally, an in-depth examination of both sanctions and international conflict management in the Balkans from 1987 to 1995 is presented in Chapter 5. To avoid self-selection bias, this project establishes four criteria for the inclusion of episodes. First, only sanction episodes that were initiated after 1970 but before 1993 are included. Cases before 1970 are excluded because quality data is not available. Sanction episodes that started after 1993, such as the 1995 Iran-Libyan Sanctions Act or sanctions imposed against Serbia because of events in Kosovo, are too recent from which to draw conclusions. The second criteria is that cases are only selected where the sender is the United States and non-Soviet Union or Eastern European states are targets. The former reduces the variance in explanatory variables. There is also a practical advantage in using the U.S. as the sole sender given that American public 123 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. opinion polling data and U.S. media abstracts are readily available. Cases where the target is either the Soviet Union or Eastern Europe are excluded because of the impact the Cold War would have on such a conflict.^ Thirdly, in all cases, the objective of U.S. sanctions must be a difficult foreign policy objective. Finally, sanctions against Cuba and Vietnam after 1970 are excluded because such sanctions were imposed only after the failure of military force to achieve U.S. objectives. The agreement ending the Cuban Missile Crisis pledged the United States to never invoke force against Cuba.'' After the Viemam War, it would be unreasonable to expect any president to use force. Hence, these two cases are excluded simply because the possibility of escalation was discounted even prior to new sanctions.^ Based on these criteria, eighteen sanction episodes are drawn from the Hufbauer. Schott, and Elliott data set. These cases are: ^ Freedman, Lawrence, The Evolution of Nuclear Strate©' (New York: The International Institute for Strategic Studies, 1989). ■ * Allyn, Bruce, "Essence of Revision: Moscow, Havana, and the Cuban Missile Crisis," International Security Çm ava 136-172. Beschloss, Michael, 7% e C nm Kcarj: Kennedy and KAnüAcAev (New York, NY : HarperCollins Publishers, 1991). Blight, James, On the Brink (Nevt York, NY: Hill and Wang, 1989). ’ It is interesting to consider, however, if the failure of sanctions against Cuba or Vietnam played a role in the use of military force in either case. Michael Beschloss, cited in Chapter 2, argues that Kennedy’s “fhistration” over the Cuban situation contributed to his decision to authorize the “Bay of Pigs” invasion despite his belief that the mission would fail. The Pentagon Papers show that the inability of sanctions to work did contribute to a desire to escalate the Viemam conflict. 124 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4.1: Cases Considered Target Year Goal Angola® 1985-1991 Oppose Marxist government Burma 1988-1994 Oppose coup, support democracy Chile 1970-1973 Oppose the Marxist government China 1989-1990 Fundamental change in human rights Grenada 1983 Destabilize the government Guatemala^ 1977-1983 Fundamental change in human rights Haiti 1987-1989 Oppose coup, support democracy Iran 1979-1981 Release the hostages Iraq 1990-1991 Withdraw from Kuwait Libya* 1978-1986 Destabilize the government Nicaragua 1977-1979 Destabilize the government Nicaragua 1981-1990 Destabilize the government Panama 1987-1989 Destabilize the government Paraguay 1977-1981 Fundamental change in human rights South Africa 1985-1991 End apartheid, destabilize the government Suriname 1982-1990 Destabilize the government Uganda 1972-1979 Fundamental change in human rights Uruguay 1976-1981 Fundamental change in human rights ‘ Hufbauer et. al. classify the sanction episode against Angola as beginning in 1986. However, the U.S. first imposed sanctions against Angola in July 1985. The data is reconfigured to account for this fact. ^ Hufbauer eL al. classify the sanction episode against Guatemala as ending in 1985. This project respecifies this case to terminate in 1983 after the U.S. committed itself to supplying arms to the military. ' Although sanctions remain in place, this study stops the Libyan sanction episode in April 1986 to coincide with U.S. airstrikes. 125 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Two additional cases are included that not part of the Hufbauer et. al. data set, but fall within the criteria established. Haiti 1991-1994 Destabilize the government Serbia 1991-1995 End fighting in the Balkans These twenty cases are used to test the model devised in Chapter 3. This is achieved through the three analytical approaches. First, brief case studies are performed on each sanction episode. Second, statistical regression analysis provides quantifiable support to the argument. Finally, the next chapter takes an intensive look at the conflict in the Balkans from 1987 to 1995. All three techniques yield identical results: the model developed in Chapter 3 holds empirical validity and self-imposed costs do generate domestic audience costs. /. Case Study Analysis The first technique used to support the model is through case study analysis. In subsequent pages, I provide brief summaries of the 20 episodes and explain how well the model predicted the outcome. Again, those interested in chronologies, an organized collection of public polling data, or television reporting for each episodes are welcomed to contact the author. 1 2 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4^: Case Study Analysis Outcomes Target Target Type A nnual Self- imposed Cost o f Sanctions Audience Costs Present? Effect on overall opinion Presence of Conflict Escalation Angola Middle $4 million No None Yes: rebel aid Burma Strong -$11 million No None No Chile Weak - $5.7 million Slight None Yes: covert aid China Strong $24 million Slight Slight No Grenada Weak $0 Yes None Yes: war Guatemala Weak -$I1 million No None No Haiti Weak - $56 million No None No Haiti (1991) Weak $30 million Yes Slight Yes: send troops Iran Strong $427 million Yes Significant Yes: rescue attempt Iraq Strong $1,100 million Yes Significant Yes: war Libya Weak $113 million Yes None Yes: airstrikes Nicaragua Weak -$13 million No None No Nicaragua (1981) Weak $83 million Yes Significant Yes: rebel aid, mine harbors Panama Weak $43.1 million Yes Significant Yes: war Paraguay Weak - $2 million No None No Serbia Middle $52 million Yes Slight Yes: airstrikes S. Africa Strong $110 million Yes Slight No Suriname Weak - $1.5 million No None No Uganda Weak $28 million No None No Uruguay Weak - $10 million No None No 127 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. An analysis of the 20 cases supports the model developed in Chapter 3. Specifically, self-imposed costs are correlated with the generate domestic audience costs, and conflict escalation is more likely when audience costs are present than when audience costs are not present. A brief summary o f the data are found in Table 4.2. The cases are evenly split between conflict escalation and no conflict escalation. U.S. military force was applied in 5 episodes, 3 times the U.S. limited itself to covert aid, with the Iran hostage rescue attempt and deployment of troops to Haiti representing the final 2 cases. The average length of each sanction episode was 34 months— consistent with the findings of Hufbauer, Schott, and Elliott over the entire sanction data set.’ In 11 cases, the U.S. suffered economically, 8 times the U.S. economy benefited, and I time there was no cost to the U.S. In the 10 cases where there was military escalation, 9 showed evidence of domestic audience costs. Angola was the only exception to this trend. 9 episodes also involved the United States suffering "self-imposed” sunken costs. With the debatable cases of China and South Africa, there were no audience costs suffered in the 10 non escalation episodes; and only 3 of the non-escalatory cases involved positive self- imposed costs: China, South Afnca, and Uganda. This provides evidence that a correlation does exist between the imposition of self-impose costs, generation of domestic audiences, and conflict escalation. ’ Hufbauer, Schott, and Elliott Sanctions Reconsidered. In their study, they found that sanction episodes averaged 3 years. 128 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Finally, sanction episodes did effect a president’s overall approval in 8 episodes. In 6 of the 8, there was conflict escalation, with China and South Africa as the only exceptions. However, in these 2 cases, the effect on overall public opinion was slight and there was no public consensus in favor of economic sanctions. This provides partial empirical support to Alastair Smith’s hypothesis that domestic audience costs generated from a failed foreign policy may affect overall public opinion, and thus provide incentives for a president to escalate. 129 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Case I: U S. vs. Angola (1985-1991) Summary o f Findings U.S. goal: Destabilize the government Type of target: Middle Major sanctions imposed: 1985: Freeze Export-lmport financing 1985: Ban the U.S. military from purchasing Angola oil U.S. annual cost of sanctions: $4 million Other coercive policies: 1986-1991: Provide UNITA rebels military aid Did sanctions work (HSE): No Generation of audience costs: No Historical Background and U.S. Goal American policy toward Angola was relatively “hands-ofF’ until 1985. at which time the U.S. imposed economic sanctions and actively supported the UNITA rebels tying to overthrow the MPLA government supported by Cuban troops. Prior to 1985, UNITA was backed by the South African military. When the South African military withdrew from Angola in April 1985, the Reagan administration sanctioned Angola, as well as pressed Congress to repeal legislation that banned aid to UNITA. Congress lifted the ban in July 1985 and Reagan extended $15 million in aid in January 1986. South Africa rejoined UNITA in trying to destabilize the government in 1986. In 1991, a treaty signed by the MPLA, UNITA, Cuba, South Africa, and the U.S. called for both South Africa and Cuba to withdraw their troops and the holding of elections. Cuban and South Africa withdrew shortly thereafter and sanctions were lifted. 130 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Target Type Angola is considered a middle target in the context o f the model. The data suggest that Angola did spend significant resources on coercive arms of the state, as well as repressed the population. Cuban and Soviet support also contributed to the government’s strength. The civil war, however, posed a significant challenge to the stability o f the government. Relative Military Capability (USAC, 1986): 216.4 Relative Number of Troops (USAC, 1985): 34 Relative Economic Capability (World Bank, 1985): 904 Relative Population Size (World Bank, 1985): 29.7 Generation o f Domestic Audience Costs Given the low level of U.S. sanctions imposed, it is not surprising to find that the public was disinterested in events in Angola. A total of only 14 stories about Angola were reported on network television news in 1985, and 19 in all of 1986. Over half of these reports concerned South African military involvement in Angola or negotiations over Namibia. Other than a few stories reported in January 1986 when UNITA leader Jonas Savimbi visited the White House, little was reported over President Reagan’s failure to destabilize the Angola government. Print media reporting also suggests a dearth of interest. Prior to peace negotiations in 1991, the maximum number o f times “Angola” was cited in a NYT abstract was 31 in January 1986, again as a result o f Savimbi’s visit to the White House. From February 1986 to the start of peace negotiations in 1991, the firequency of NYT abstracts never exceeded 15 per month. No public opinion polling data was deposited at the Roper Center on Angola. 131 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Approval Ratings While there is no public polling data on U.S. attitudes concerning Reagan and Bush’s handling of the situation in Angola, Reagan’s overall approval ratings were high at the start of the sanction episode, falling in November 1986 after the Iran-Contra story was exposed. There is no reason to believe that Reagan’s overall approval rating was influenced by his policies in Angola. Overall Public Opinion July 1986 - December 1991 g 70 (^60 132 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis o f Sanction Episode in the Context o f the Model The small level of sanction imposed implies that s < s*. The model of Chapter 3 suggests that when “sunken costs” are below s*. domestic audience costs will not be generated and the conflict will not escalate to war. An analysis of this case suggests, however, that the conflict did escalate in that military aid was provided to the UNIT A rebels. Thus, while sanctions failed in destabilizing the government and domestic audience costs were not generated, the Reagan administration elected to escalate the conflict by arming UNIT A rebels. At no time, though, did the administration consider using U.S. military force. Such a finding is consisent with the model. 133 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C ase II: U.S. vs. B urm a (1988-current) Summary o f Findings U.S. goal: Oppose military coup, support democracy Type of target: Strong Major sanctions imposed: 1988: U.S., Japan freeze international aid 1989: U.S. suspends Burma’s privileges under the Generalized System of Preferences 1990: U.S. Congress votes to ban Burma imports 1997: U.S. adopts legislation banning new investments U.S. annual cost of sanctions: -$11 million 1988-1991 - $1.4 million 1992-current Other coercive policies: None Did sanctions work (HSE): No Generation of audience costs: No Historical Background and U.S. Goal Following student demonstrations in July 1988, President General Ne Win resigned. However, Ne Win is reportedly to have selected the individuals that would replace him. Students found the new government unacceptable and continued to protest in August and September. When Defense Minister Saw Maung successfully led a coup on September 18, violence increased as the opposition anticipated a crackdown. The new military junta in Burma suppressed human rights throughout the 1990s. The international community imposed sanctions against the new government in an attempt to significantly improve human rights and democratize the country. 134 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Target Type Burma is considered a strong target in the context o f the model. Burma has a population of over 42 million people and uses extreme forms of repression to control domestic dissent. Furthermore, by 1990, Burma was politically supported by major countries in the region, such as Japan, Indonesia, and Thailand. Relative Military Capability (USAC, 1988): 376.25 Relative Number of Troops (USAC, 1988): 12.8 Relative Economic Capability (World Bank, 1988-95): n.a. Relative Population Size (World Bank, 1988): 6.12 Generation o f Domestic Audience Costs The public seemed uninterested in events in Burma. There are no relevant polling data deposited with the Roper Center. Television reporting of demonstrations and the coup between July 1988 and September 1988 was high, but after September, only a January 21 telecast on CBS reported on events in Burma in all of 1989. Television reporting increased significantly the week of the May 1990 election, but following the election, Burma was not reported on again until October 1991, when Aung San Suu Kyi was announced the recipient of the 1991 Nobel Peace Prize. Burma was mentioned on network news 3 times in 1992,0 in 1993, 1 in 1994,6 in 1995, 17 in 1996,2 in 1997,4 in 1998, and 1 in 1999 through May 30. In 1996, network news increased its reporting in response to U.S. sanctions. NYT reporting followed a similar pattern. After publishing numerous stories firom July to September 1988, Burma was rarely cited. With the exception o f the May 1990 election and October 1991 Nobel Peace Prize, Burma as not mentioned more than 10 times in any month through 1995. 135 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Approval Ratings The failure of sanctions against Burma does not seem to have affected overall approval for Reagan, Bush, or Clinton. Overall Public Approval August 1988 - July 1998 90 80 ,70 g 60 5 1 50 40 30 20 ................... aug 88 june 89 apr 90 fsb 9i dec 91 oct92 aug 93 jun94 apr 95 feb 96 dec 96 oct97 Date Analysis o f Sanction Episode in the Context o f the Model The U.S. did not have enough leverage over the Burmese economy to threaten it with self-imposed costs. Nor was there any domestic or international support for escalation. This was especially true given that no other state expanded sanctions beyond those imposed in September 1988. Burma’s four largest trading partners, Japan, Indonesia, Singapore, and Thailand, have worked to increase ties with Burma. Sanctions could not signal a willingness to escalate because the public was indifferent toward events in Burma. In the context of the model, U.S. sanctions were less than s* and insufficient audience costs were generated to credibly threaten escalation. 136 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C ase ni: U.S. vs. C hile (1970-1973) Summary o f Findings U.S. goal: Destabilize the government Type of target: Weak Major sanctions imposed: 1971 : Ex-Im financing is cut by $21 million 1972: All bilateral aid ends, except military aid 1972: All Ex-Im loan guarantees are suspeneded 1973: New investments are banned U.S. annual cost o f sanctions: - $5.7 million Other coercive policies: Financing of costly truckers strike and financial support of political opposition groups Did sanctions work (HSE): No Generation of audience costs: Slight Historical Background and U.S. Goal Marxist-candidate Salvador Allende was elected President o f Chile in August 1970 on a semi-socialist platform calling for the nationalization of foreign industries. Allende also championed closer ties with both the Soviet Union and China. Concerned over U.S. interests in Chile, the CIA and U.S. firms, such as ITT, worked to discredit the Allende presidential campaign. Once Allende was elected, they then turned their attention to opposing the Allende government. There was deep concern among American foreign policy elites that Chile may become a base for Soviet expansion following Allende’s election. The day after Allende was installed, he reiterated his campaign promise to increase ties with the USSR. Within 10 days o f office, diplomatic relations were restored with Cuba. Two 137 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. months later, diplomatic relations were established with China. Growing ties between Chile and communist states, in conjunction with the nationalization of U.S. firms, resulted in the U.S. adopting measures designed to destabilize Allende, including imposing sanctions and financing a truckers strike. Sanctions were lifted after a coup by General Augusto Pinochet in September 1973. Target Type Chile is a weak target in the context of the model. Chile's population and economy were small by comparison to other states. Chile also received only nomial backing from the Soviet Union. Finally, because the military and large segments of the population opposed Allende. his ability to repress the population was limited. Relative Military Capability (USAC, 1970): 405.5 Relative Number of Troops (USAC, 1970): 43.9 Relative Economic Capability (World Bank, 1970): 119.8 Relative Population Size (World Bank, 1970): 21.6 Generation o f Domestic Audience Costs The data suggest that President Nixon suffered a slight decline in public confidence because Allende governed Chile. The success of U.S. sanctions in removing Allende, however, did not improve Nixon’s domestic political standing. This may have been because while Americans wanted “something” to be done about communism in Chile, the public opposed destabilizing a democratically-elected leader. 138 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The public was extremely interested in events in Chile, but did not support U.S. efforts to destabilize the government. Television news reporting of events were extensive from 1970 to 1973. Such coverage is not surprising when taken in the context of the Cold War. Chile was the first government in the Western Hemisphere to become Socialist since the Cuban revolution in 1959 and the U.S. was still fighting the Viemam War. Interestingly, while the media did cover Allende’s policies, network news rarely mentioned U.S. efforts to destabilize Allende and the New York Times only briefly mentioned U.S. sanctions at the time they were imposed. But while the press displayed interest in Chile, the public did not favor destabilizing the government. Public opinion polling reponses show that when told that a military coup had taken place against a popularly elected Marxist president in Chile, 31% disapproved of the coup while only 19% supported it. When asked in September 1974 if it was proper for the U.S. to attempt to destabilize the Allende regime amid allegations that ITT and the CIA were behind the coup, over 60% said it was wrong while only 18% thought it was right. Nearly 70% of Americans thought it was wrong for the U.S. to interfere with the electoral process in Chile. 139 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Approval Ratings President Nixon’s overall public approval was close to 50% when sanctions were first imposed in 1971. By the time sanctions “worked” in destabilizing the government, public approval for Nixon was at historic lows as a result of Watergate. Thus, there is no evidence showing that sanctions affected overall public approval, nor that Nixon’s successful policies in Chile improved Nixon's standing with the public. Overall Public Opinion August 1970 - October 1973 IF'' a. 0:40 eb 73 may aug Date Analysis o f Sanction Episode in the Context ofth e Model Hufbauer, Schott, and Elliott foimd that U.S. sanctions against Chile succeeded. Although the authors do not explain why, it is assumed the explanation lies in the fact that deteriorating economic conditions contributed to the military coup; and that 140 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. sanctions contributed to economic chaos. But sanctions alone do not explain why the economy collapsed nor why a coup was launched. Hufbauer, et. al. acknowledge that AJlende’s economic policies were more important in explaining economic conditions than the sanctions. Furthermore, if neither the CIA nor ITT sponsored the military coup, there can be no direct link between sanctions and the success of U.S. policy. Except under the framework of sanctions as a signal. The model demonstrates that sanctions signal a willingness to escalate. But it is reasonable to suspect that this signaling effect may be interpreted by opposition interest groups that the U.S. would not object to internal violence or a coup. Specifically, if sanctions fail and the U.S. is considering escalation, elements within a military establishment may have incentives to believe that the U.S. may support a coup. While further research is required on the signaling effects sanctions may have on third party incentives to employ violence, it is reasonable to speculate that a correlation does exists. 141 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C ase IV : U.S. vs. C hina (1989-1990) Summary o f Findings U.S. goal: Fundamental change in human rights Type of target: Strong Major sanctions imposed: 1989: Suspend military sales to China 1989: Extend student visas of Chinese students studying in the United States 1989: Freeze Export-Import financing 1989: Suspend Overseas Private Investment Corporation (OPIC) financing 1989: Suspend high-level diplomatic contacts 1989: Vote against World Bank and IMF loans 1989: Suspend talks on China’s entry into GATT U.S. armual cost of sanctions: $24 million' Other coercive policies: None Did sanctions work (HSE): No Generation of audience costs: Slight Historical Background and U.S. Goal Former Communist Party Chief Hu Yaobang’s death on April 15, 1989 sparked large pro-democracy rallies that shook the foundations of China’s political system. Recalled as a peacemaker siding with 1987 student demonstrators, Hu’s death caused students to march in Tiananmen Square demanding political reforms. ' While the actual number is closer to $1-2 million, to maintain as consistent with the Hufbauer, Schott, and Elliott data set, the $24 million figure is derived. 142 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. An inflammatory editorial in the People's Daily on April 26 suggesting that the government may take retaliatory measures against student protesters backfired as 150,000 people flooded Tiananmen Square denouncing the threat. By the first week of May, 300,000 people were staging a sit-in protest. The number of protesters would grow to 1,000,000 per day by mid-May. On June 3, both the 27* Army and the 15* Air Army attacked the Tiananmen Square protesters. Within two days, the uprising was crushed. Official Chinese figures place the number of dead at 300, of which only 23 were students. Conservative estimates by the Chinese Red Cross cite between 2,500 and 3,000. Upper limits reach as high as 7,000. Deploring China’s human rights policies. U.S. political leaders called on China to become more democratic, and imposed sanctions to that effect. Target Type China is a strong target in the context of the model given China’s strong economic, military, and political strength. Relative Military Capability (USAC, 1989): 7 Relative Number of Troops (USAC, 1989): 0.57 Relative Economic Capability (World Bank, 1989): 15.4. Relative Population Size (World Bank, 1989): 0.221 Generation o f Domestic Audience Costs Had President Bush imposed costly sanctions, and the sanctions failed, the data suggest that Bush would have suffered considerable domestic audience costs. However, because the sanctions he imposed were not costly, only slight audience costs 143 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. were generated from the failure of the U.S. to quickly reform human rights practices in China. Media coverage o f the Chinese massacre both on television and in the New York Times was extremely high, indicating the potential for the generation of domestic audience costs. Public opinion polls also show that Americans were interested in events in China and would judge President Bush’s competency based, in part, on his handling of the situation. Public opinion polls show that nearly 80% of the public was following events in China, with 70% having discussed the China situation with friends. Nearly 90% supported the Chinese students and objected to the actions of the Chinese government. The public’s “overall opinion of China” dropped from 72% to 16% between March 1989 and June 1989, and remained low for many months, despite high ratings even during the student protests. The protests even caused an increase from 15% to 24% in those who believed that “communism in China [was] becoming more of a threat to the security of the U.S. than it has been in the past,” between March and November 1989. The percentage believing that China was becoming less of a threat fell in the same period from 34% to 20%. Even as late as December 1989,83% thought it was more important for the U.S. to “stand up for human rights as a condition of our friendship,” than be “conciliatory to the Chinese.” This would suggest that both the public was interested in events in China and that domestic audience costs would build unless the U.S. achieved its goal of stemming human rights abuses in China. However, the public was generally not in support of economic punishment, and, by not imposing costly sanctions. Bush elected not to make 144 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. China an issue where the public would assess his competency. Despite media coverage similar to that of Iran and Iraq, the domestic audience costs suffered from the failed U.S. sanction policy were marginal. Public opinion polls indicate that Americans generally favored Bush’s active engagement with China and only slightly punished him for failing to improve human rights. Polls conducted by Gallup and ABC suggest that over 60% of the public approved of Bush’s handling of events in China immediately following the crackdown. When asked the same question 7 months later by ABC, 55% still “approved of the way Bush [was] handling relations with China.” 64% did not even believe that Bush could have prevented the crackdown. There was no change in the percentage approving mild sanctions from June to December, over 70% in both cases. Even at the time of the massacre, the public by wide margins opposed recalling the U.S. Ambassador or breaking off diplomatic relations. Support for sanctions was also weak. While a majority of Americans favored restrictions on investments, such support was tenuous, with 37% favoring, but 59% opposing “a complete ban on U.S. trade with China”. The only sanctions receiving over 70% of public support were military sanctions and an extension of student visas. Finally, Democrats could not turn Bush’s failure to improve human rights into increased political support because the public was not willing to sanction China, again suggesting that audience costs were marginal. For example, presidential candidate Bill Clinton cited Bush’s alleged support of “dictators from Baghdad to Beijing,” as a reason to vote against Bush, but once elected, Clinton “de-linked” human rights from trade. 145 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Congressional Democrats also tempered their objection to Bush’s weak sanction policy. Bush made it clear early in the crisis that he would support neither vigorous economic sanctions nor revoking of MFN status. At a press conference on June 8, for example, he flatly stated that “I disagree with [supporters of sanctions] 180 degrees. Some have suggested—well you gotta go full sanctions on the economic side. I don’t want to cut off grain. I think that would be counterproductive and would hurt the people." Congress did pass a sanctions bill, but the sanctions imposed suggest support for Bush’s policies of no escalation. Changes in Public Opinion^ President Bush’s overall public approval rating increased as a result of the Tiananmen Square massacre, jumping almost 15% points in June. Overall approval remained high throughout the summer, due to other international events, such as the dissolvement of communism in Eastern Europe. Public approval for Bush’s handling of events in China, however, fell 5% points between June 1989 and February 1990. Thus, while Bush’s overall approval rating was increasing, the public was disillusioned with Bush’s failure to improve human rights in China, * Data on U.S. public opinion on President Bush’s handling of the crisis in China was derived from polling questions from ABC News-Washington Post. Gallup only asked the public their opinion on Bush’s handling of the situation in June 1989. 146 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 0 75 5 o 70 Q . Q . < U S 65 60 55 P ublic Approval: Overall an d in China May 1989 - March 1990 1 I f ------- 1 \ N ' / i i 1 - - - - - 1 I i — may 89 july 89 sept 89 nov 89 Date jan 90 mar 90 Analysis o f Sanction Episode in the Context o f the Model China was a strong target and the U.S. was not prepared to pay for the price of threatening escalation. As the U.S. could not credibly signal a willingness to escalate. Bush imposed only mild sanctions. The American public punished Bush slighly for his policies in China, sensing that the U.S. could not influence Chinese policy, and would suffer significant self-imposed costs if sanctions were imposed. 147 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Coimterfactually, imposing a high level of sanction would have generated higher audience costs. However, aware that sanctions would fail and that it would be necessary to '‘backdown,” Bush imposed only the mildest of sanctions. It is for this same reason that Clinton changed his China policy upon assuming office. Imposing sanctions destined to fail would only result in suffering domestic audience costs. 148 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C ase V: U.S. vs. G ren ad a (1983) Summary o f Findings U.S. goal: Destabilize the government Type of target: Weak Major sanctions imposed: 1983: Prohibition of trade U.S. annual cost of sanctions: $0 million Other coercive policies: Military attack Did sanctions work (HSE): No Generation of audience costs: Yes Historical Background and U.S. Goal The alleged threat Grenada’s Marxists government posed to U.S. national security was established long before Reagan elected to authorize a military invasion on October 25. 1983. Network television news often quoted Reagan as saying that Cuban influence was spreading in Latin America through Grenada and Nicaragua. In an effort to get increased funding for the Department of Defense in 1983, Reagan also authorized the declassification of intelligence photos showing USSR military installations in Nicaragua, Grenada, and Cuba. In 1981, the Reagan administration even considered a CIA plan to use covert force to destabilize the government. But it was not until a military coup on October 19-20, 1983 before the U.S. took economic and military action. The same day as the coup, the United States imposed a comprehensive trade and financial sanction. The next day, a naval task force was deployed fi’ om Virginia. On October 23, a marine task force assigned to leave for Beirut was diverted to Grenada. On October 25, the U.S. invaded. 149 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Target Type Grenada is a weak target in the context of the model. Grenada has less than 100,000 inhabitants, an economy smaller than Green Bay, Wisconsin, and troops strength numbering in the thousands. The Soviet Union was also unwilling to protect Grenada from U.S. threats. Relative Military Capability (USAC, 1983): n.a. Relative Number of Troops (USAC, 1983): n.a. Relative Economic Capability (World Bank, 1983): 36276.6 Relative Population Size (World Bank, 1983): 2614.6 Generation o f Domestic Audience Costs The data suggest that Reagan suffered domestic audience costs because of the Marxist government in Grenada. There was extensive media coverage of events in Grenada prior to U.S. military action. NBC and CBS reported the initial stages of the coup on October 19. The coup was also the leading news story on all three networks on October 20 and on two of the three networks on October 21. Regular reports continued over the weekend of October 22 and October 23, resuming Monday where CBS reported that the United States was preparing an invasion of Grenada. The New York Times was also reporting the coup on its front page throughout the week. Such media coverage for a state smaller than Connecticut with less than 100,000 inhabitants can only be explained by previous actions by Reagan that made U.S. policy toward Grenada a measure of his “competency” to stem the spread of communism. Unfortunately, there is no polling data on U.S. attitudes toward Grenada prior to the invasion. An overwhelming majority of Americans, however, did support the war. 150 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Opinion Overall public approval did increase slightly because of U.S. military action in Grenada from 49% to 53%, but the increase followed a general positive trend during the summer of 1983 and throughout 1984. The killing of 241 Marines 4 days before the U.S. invasion also affected overall public opinion. Overall Public Opinion August 1983 - December 1983 52 ■^8 <46 44 aug 83 sep oct 83 Date dec 83 nov 83 Analysis o f Sanction Episode in the Context o f the Model Although this sanction episode lasted only 1 month, it provides insight into the role audiences costs play in conflict escalation. Hufbauer, et. al. found that sanctions failed. But in the context of the signaling model, sanctions served their purpose. Once Reagan imposed sanctions and authorized troop deployment, a credible threat was made 151 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. that the U.S. would be forced to escalate, if necessary. Furthermore, after television news had been reporting for months that Reagan wanted to increase defense spending to counter the threat from Soviet-backed regimes in Grenada and Nicaragua, backing down would have generated considerable audience costs. In the context of the model, Grenada was a weak target, a(s) was sensitive to sanctions, and s was less than s*. 152 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C ase V I: U.S. vs. G uatem ala (1977-1983) Summary o f Findings U.S. goal: Fundamental change in human rights Type of target: Weak Major sanctions imposed: 1977: Veto loans from international agencies 1977: Military and economic aid denied U.S. annual cost of sanctions: - $11 million Other coercive policies: None Did sanctions work (HSE): No Generation of audience costs: No Historical Background and U.S. Goal The Guatemalan government engaged in horrific human rights practices in the late 1970s and 1980s. State sponsored mass killings, mass rapes, and routine torture of primarily Mayan Indians were reported in both print and television media. After a July 1977 border dispute between Guatemala and Belize nearly erupted into war, the Carter administration reconsidered its aid policy. Financial and military sanctions were imposed in 1977 with the goal of encouraging Guatemala to become democratic. Ronald Reagan’s election brought changes to U.S. policy. Relatively more concerned about the spread of communism in Central America, Reagan believed that sanctions undermined U.S. interests in the region. As early as May 1981, the administration was sending weapons to Guatemala. Following elections in March 1982 153 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. handing power to General Angel Anibal Guevara, the administration worked to remove all sanctions. In October 1982, $170 million in suspended loans was released, and in January 1983, Reagan announced a lifting of the arms embargo. Target Type Guatemala is a “weak” target in the context of the model. Guatemala’s only major ally was the United States. Guatemala also had a small population and economy. The Guatemalan government’s only “strength” was her repression of the population, including killing almost 200,000 of her citizens in the 1980s. Relative Military Capability (USAC, 1977): 2242.2 Relative Niunber of Troops (USAC, 1977): 147.1 Relative Economic Capability (World Bank, 1977): 359.5 Relative Population Size (World Bank, 1977): 34.6 Generation o f Domestic Audience Costs Judging the public’s response to both Carter’s and Reagan’s policies in Guatemala is difficult to assess. No public polling questions are available from the Roper Center to offer guidance. Media reporting of human rights violations in both network television and the New York Times was sporadic. Guatemala was often mentioned in both media sources not for her human rights policies, but in the context of overall U.S. Central American policy, particularly in 1979 when Somoza resigned and in 1981 when Reagan sought Guatemala’s help in destabilizing Nicaragua. Once these stories are excluded, network news rarely discussed events in Guatemala. 154 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Opinion Carter’s public approval was high at the time sanctions were imposed. But with a poorly performing economy. Carter’s approval would fall over time. The contribution made by Carter’s failure to bring democracy to Guatemala to the decline in overall public approval can not be assessed, although it is expected to be extremely small. Overall Public Opinion January 1977 - January 1983 (0 60 < 5 0 jan 77 sept 77may 78 jan 79 sept 79 may 80 jan 81 sept 81 may 82 jan 83 Date Analysis o f Sanction Episode in the Context o f the Model The sanctions imposed did not credibly threaten future costs, and thus failed to reform the Guatemalan government’s human rights practice. Carter administration involved negative “self-imposed” costs while the Reagan administration actively worked to supply arms and financing to the government. In the context o f our model, Guatemala was a weak target, but s < s*. 155 Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. C ase VU: U.S. vs. H aiti (1987-1990) Summary o f Findings U.S. goal: Oppose coup, support democracy Type of target: Weak Major sanctions imposed: 1987: Suspend all conomic and military aid U.S. annual cost of sanctions: - $56 million Other coercive policies: None Did sanctions work (HSE): No Generation of audience costs: No Historical Background and U.S. Goal The poorest state in the Western Hemisphere, Haiti has long been ruled by ruthless dictators. Not even the popular overthrowing of Jean-Claude "Baby Doc’ * Duvalier in 1986 brought about changes in government oppression. Instead of installing democracy, the new government increased human rights abuses. Events had looked promising at first. General Henri Namphy, who assumed control after Duvalier, declared that elections would be forthcoming in June 1986. U.S. officials were so hopeful that democracy would be established that Vice-President George Bush visited Haiti and offered Namphy U.S. political support in August 1986. Unfortunately, Namphy rescinded his promise and canceled the elections in November 1987. The U.S. immediately terminated all economic aid. The U.S. also announced that it would not use force to restore democracy. Elections were rescheduled for January 1988, where Leslie Manigat won as a result of military sponsored vote fiaud. 156 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. By mid-1987, U.S. State Department officials were also concerned over reports that the Haitian military was engaging in drug trafficking. Public opinion polls showed that the public found “drugs” to be the single greatest problem the country was facing throughout 1988 and 1989. By May 1989,27% of Americans were citing drug abuse as “the most important problem the United States was facing”. Panamanian strongman Manuel Noriega had also been indicted on drug charges the previous month. Purging military officers participating in the drug trade now become an additional objective of U.S. sanction policy. When President Leslie Manigat attempted to fire Namphy, in part over differences in cracking down on corruption and drug trafficking, Namphy led a coup and reinstalled military rule. Three months later, on September 18, 1988. Lt. General Prosper Avril led a bloodless coup against Namphy. Although Avril continued to repress the population, his desire to obtain U.S. aid convinced him to fire many military officers in February and March 1989. These firings prompted two coup attempts, both of which failed. Avril was forced out of office in March 1990 when massive street protests erupted following a fatal police shooting of a young girl participating in a peaceful pro-democracy protest. Presidential elections were held in September 1990 and human rights activist Father Jean-Bertrand Aristide was declared the winner. U.S. sanctions were lifted immediately after Aristide’s installment. 157 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Target Type Haiti is a considered weak target in the context of the model. Haiti was the poorest state in the Western Hemisphere and was not aligned with the Soviet Union. It was also only through oppression that the military kept political control. Relative Military Capability (USAC, 1987): 6861.9 Relative Number of Troops (USAC, 1987): 284.9 Relative Economic Capability (World Bank, 1987): 2097.2 Relative Population Size (World Bank, 1987): 39.7 Generation o f Domestic Audience Costs There is no evidence to suggest that audience costs were generated because of the failure of U.S. sanctions. The public seemed uninteresed in Haiti and was not willing to judge the competency of Reagan or Bush based on the effectiveness of U.S. policy against Haiti. No polling questions are available from the Roper Center and the media only discussed Haiti when a coup occurred. Neither television news nor the New York Times devoted much space to the Haitian protests during these years. Not even the March 9 drug indictment the Haitian military leader, Jean-Claude Paul, received much media coverage. There were only 4 newsworthy reports in the NYT indexed for March 1988 despite Paul’s indictment at a time when the public believed that drug abuse was the greatest problem facing the United States. Furthermore, no network carried stories describing Haiti’s role in the drug trade after March— a surprising finding given events in Panama. Comparing Haiti to the case of Panama, where U.S. objectives were similar, one finds that both the low level of sanction and little publicity associated with events in Haiti indicates that audience costs were low. 158 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Opinion There is little reason to believe that U.S. policy against Haiti effected overall approval ratings for either Reagan or Bush. Other events, such as the Iran-Contra scandal, the falling of the Berlin Wall, Gulf War, and state of the economy provide better explanations for presidential approval ratings than U.S. sanctions against Haiti. Overall Public Opinion September 1987 - December 1990 I u S 3 C L 80 75 70 65 60 'A 55 50 45 sept 87 jan 88 may 88 sept 88 jan 89 may 89 sept 89 jan 90 may 90 sept 90 Date Analysis ofSanction Episode in the Context o f the Model Sanctions failed because the threat of conflict escalation was not present. Domestic audiences did not penalize the Reagan or Bush administration for failing to achieve U.S. foreign policy goals, in part because there was little in the way of self- imposed costs. Unlike the similar episodes of Panama or Haiti (1991 -1994), the meager level of sanctions against Haiti was insufficient to generate domestic audience costs and signal American resolve. Given that Haitian democracy was not in the U.S. 159 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. national interest, along with State Department statements indictating that the U.S. would not employ force, it is easy to understand why sanctions failed. Haiti was a weak target, but sanctions were less than that required to induce the U.S. to escalate, s < s*. 160 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C ase V n i: U.S. vs H aiti (1991-1994) Summary o f Findings U.S. goal: Destabilize the government Type of target: Weak Major sanctions imposed: 1991: Suspend all conomic and military aid 1992: Restrict trade 1993: U.N. imposes an oil embargo 1993: U.S. Navy enforces the embargo 1994: All economic transactions banned U.S. annual cost of sanctions: $30 million Other coercive policies: 1991 : Forced return of Haitian refugees 1994: Threaten war and send of U.S. troops Did sanctions work (HSE): No Generation of audience costs: Yes Historical Background and U.S. Goal A coup led by General Raoul Cedras against the democratically elected Father Jean-Bertrand Aristide on September 30. 1991 began with the typical U.S. response, but ended with the deployment of military force for reasons relating to the generation of domestic audience costs. Consistent with previous Haitian coups. President Bush’s first reaction was to end all economic and military aid, but not impose any trade sanctions. Also like other episodes. Bush explicitly ruled out the use of military force, although the military was deployed to manage the flow of Haitian boat refugees attempting to gain political asylum in the U.S. A Bush administration policy to forcibly return all refugees starting November 18 was challenged by opponents of the policy all the way to the Supreme Court, where the policy was ruled constitutional. 161 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. As reports emerged that the poor were suffering from the sanctions. Bush authorized the partial lifting of sanctions on February 9,1992. After another flood of boat refugees sought political asylum in May 1992, administration officials leaked to television networks that an invasion was under consideration. The next day, presidential candidate Bill Clinton gave a major foreign policy speech criticizing Bush's Haitian policy in general, and his policy of forcible return in particular. Having raised the Haitian crisis to a campaign issue, both Bush and the future President Clinton would find themselves under pressure from the public to restore Haitian democracy or be seen as incompetent against a weak state. Domestic audience costs had been generated. Even before assuming office after his presidential victory in November 1992. events forced Clinton to reconsider his campaign pledge of allowing all Haitian boat refugees a political asylum hearing. When satellite photos showed that tens of thousands of Haitians were preparing to leave for the U.S. on inauguration day, Clinton announced that he would not only adopt Bush’s policy of forced returns, but he would more tightly enforce the policy by imposing a naval blockade. Amid falling GDP and an increase in domestic violence, Cedras agreed to meet with U.S. officials at Governors Island, New York to find a negotiated settlement. Signed on July 2, 1993, the so-called Governors Agreement required Cedras to abdicate power and allow Aristide to return by the end of the year. When U.S. troops were denied entrance into Haiti in accordance to the Agreement on October 11, Clinton re imposed an oil embargo. 162 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. By mid-1994, the situation in Haiti became bleak. Administration officials no longer believed that sanctions alone could restore Aristide, the Haitian government continued to terrorize the population, and a report from a Harvard research team found that sanctions were responsible for the deaths of over 1,000 children per month. Amid growing public criticism, Clinton announced new sanctions on May 21 that banned all trade and ordered the U.S. Navy to enforce a comprehensive embargo. Further sanctions adopted in June froze the financial assets of the Haitian ruling elite. Believing that sanctions had failed, Clinton turned to the use of military force. In May, news networks were reporting that Clinton had decided to authorize an invasion, if necessary, to destabilize the Haitian government. In early July, against the backdrop of ordering U.S. Navy ships to leave for Haiti. Clinton announced that he was considering an invasion. Clinton's military advisors also openly discussed the possibility of an invasion. At the request of the U.S., a U.N. Security Council resolution authorizing the use of "any necessary means” to restore democracy in Haiti was passed on July 31. An effort by the U.S. Congress to require Clinton to receive Congressional approval for a military invasion failed in the Senate on August 5. Although a majority of members in both Houses did not support an invasion. Congress eventually decided that presidential prerogatives warranted defeat of the amendment. By September 1994, both administration officials were telling the national press that a U.S. invasion was imminent. On September 15, Clinton gave a national address stating that the U.S. would go to war unless Cedras abdicated power. After last-minute negotiations, Cedras stepped down and U.S. forces "invaded” Haiti on September 19. 163 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Target Type Haiti remains a weak target in the context of the model for identical reasons explained in the case where sanctions were imposed between 1987 and 1990. Relative Military Capability (USAC, 1991): 7786.1 Relative Number of Troops (USAC, 1991): 264.4 Relative Economic Capability (World Bank, 1991): 1746.2 Relative Population Size (World Bank, 1991): 38.2 Generation o f Domestic Audience Costs Public opinion polling data supply clear evidence of growing domestic audience costs. The public was highly interested in events in Haiti, but disapproved of the use of military force, even to enforce a peace agreement. Furthermore. President Clinton's approval ratings for handling the situation in Haiti were significantly below his overall approval. Thus, even though the public opposed the use of force, and Haitian democracy was not considered a strategic interest in 1991, Clinton elected to threaten a military invasion in 1994. I believe that the motivating factor for such a threat was domestic political audiences generated after the failure of costly sanctions. Public interest of events in Haiti was high. Over 50% of the public reported to be following events in Haiti throughout 1994 and 1-4% thought solving the crisis in Haiti was “the most important foreign policy problem facing the United States” in 1993 and 1994. There was also regular media coverage of events in Haiti. Network broadcasts of Haiti was firequent and the NYT cited Haiti in at least 20 abstracts per month, every month during the crisis except 2. Following Clinton’s decision to make the Haitian boat people a presidential campaign issue, media attention increased. 164 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. But even though Americans followed events in Haiti, and Haiti received extensive media exposure, the public did not believe U.S. interests were at stake. When asked if the U.S. has * ’a responsibility to do something to restore democracy to Haiti,” 40% or less said '"yes” and 50% or more said “no” when polled in June, July, and September 1994— both before and after the deployment of American forces. A similar question asking if U.S. "vital interests were at stake” showed that 30% said yes and 55% said no in October 1993 with similar results in every month between May 1994 and September 1994, again both before and after U.S. military involvement. Public attitudes over the appropriate course of action further shows the public's unwillingness to become involved. In July 1994, only 36% of Americans believed that the U.S. should do whatever it takes to destabilize the government. While the public supported the Governor Island’s settlement ending the crisis, support for using troops to enforce any agreement ending the conflict was low. Over 60% opposed the deployment o f600 U.S. troops as part of the Governor Island’s agreement, and even after troops restored Aristide to power in 1994 without the loss of one American life, support for having sent troops barely exceded 50%. In October 1994, not even 2 weeks after their arrival, over 50% disapproved sending troops. Opposition of using troops outside of a negotiated agreement was always above 60% from July 1992 to September 1994. On the other hand, support for sanctions and an oil embargo remained high, even though 58% believed that sanctions would harm average Haitians without restoring democracy. Only 27% believed that sanctions would help restore democracy. 165 Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. Given these conditions, it is not surprising that Clinton’s approval ratings in Haiti were low just prior to his September 15 warning that war was inevitable. When asked if they approve or disapprove of the way he was handling the situation in Haiti in multiple polls between October 1993 and September 14, 1994. roughly 30% consistently said they approved as compared to an overall pubic approval rating of around 50%. It was only after Clinton committed U.S. forces that his approval ratings in handling the Haitian crisis excede 45%. Changes in Public Opinion President Clinton’s approval ratings for his handling of the situation in Haiti was significantly below his overall approval rating. This is evidence that domestic audience costs were generated as a result of his failed sanction policy, even though Clinton’s Haitain policy does not appear to have affected his overall approval rating until September 1994. 166 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Public Opinion: Overall and in Haiti September 1991 - September 1994 o 50 a 40 j ' sept 91 jan 92 may 92 sept 92 jan 93 may 93 sept 93 jan 94 may 94 sept 94 Date Public Opinion: Overall and in Haiti August 9 - November 29,1994 _ 45 a. Sept 25 Oct 19 Nov 6 Aug 9 Sept 7 Date 167 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis o f Sanction Episode in the Context o f the Model This episode demonstrates that sanctions can signal a willingness to escalate by generating domestic audience costs. Clinton inflicted as high a cost on the United States as economically feasible. Once it was clear that sanctions would fail, Clinton was left with two options: remove them or escalate. As lifting sanctions would have resulted in domestic audiences Gndng Clinton “incompetent,” Clinton’s optimal strategy was to escalate. Given that U.S. interests were not at stake, as evidenced by the fact that no other adm inistration considered employing military force against Haiti between 1987 and 1994, domestic audience costs explain why Clinton elected to use force. 168 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C ase IX: U.S. vs. Ira n (1979-1981) Summary o f Findings U.S. goal; Release U.S. hostages Type of target: Strong Major sanctions imposed: November 1979: Ban export of military spare parts November 1979: Ban oil imports November 1979: Freeze Iranian assets April 1980: Diplomatic relations are severed April 1980: Total import embargo imposed April 1980: Freeze $8 billion in Iranian assets U.S. annual cost of sanctions: $427 million Other coercive policies: November 1980: U.S. conducts military exercises December 1979: Soviet Union invades Afghanistan April 1980: Rescue attempt fails September 1980: Iraq invades Iran Did sanctions work (HSE): Yes Generation of audience costs: Yes Historical Background and U.S. Goal The attack on the U.S. Embassy in Tehran on November 4. 1979 shocked and offended much of the world. Taking 95 people hostage, the hostage-takers demanded the return of the Shah to Iran. Iranian political leaders who sought to meet their country's legal obligations by pressing for the release of the hostages were attacked as puppets of the United States. From the perspective o f Iran's religious clerics, the hostage crisis offered an ideal opportunity to vault themselves to political power. 169 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The benefits of the occupation of the embassy were considerable; [Prime Minister Mehdi] Bazaragan and his cabinet were swept away apparently implicated in a pro-USA movement; the nation had to an extent become united in facing the U.S.; disorder in the provinces had been forgotten; student activities had become concentrated on behalf of rather than against the regime; and (Ayatollah Ruhollah) Khomeini had made himself an anti-imperialist hero.' Carter’s strategy to coerce Iran to release the hostages was to impose economic sanctions, be "patient,” and publicly announce that military force was not an option during the months of November and December 1979. Unfortunately, Carter's policy of "patience” proved disastrous. By January, Carter reversed his position on using military force and threatened Iran with an attack during his State of the Union address. After a poorly planned and executed rescue mission on April 25.1980, Carter returned to his earlier position and explicitly ruled out all forms of force. It was only because Iran needed resources to fight Iraq's September 22, 1980 invasion that Iran began serious negotiations. After Carter's defeat to Ronald Reagan in the 1980 presidential campaign, Iran and the U.S. reached an agreement calling for the lifting of sanctions in return for the hostages. On January 21. the hostages were fireed 30 minutes after Reagan was inaugurated as President. ' Economic Intelligence Unit, Country Profile Report (London, UK: Economic Intelligence Unit Limited, Fourth Quarter 1979) 9. 170 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Target Type Iran is considered a strong target in the context of the model given her large population, military capabilities, and Islamic fundementalist government. Relative Military Capability (USAC, 1979): 8.32 Relative Number of Troops (USAC, 1979): 4.94 Relative Economic Capability (World Bank, 1979): 28.5 Relative Population Size (World Bank, 1979): 5.95 Generation o f Domestic Audience Costs Based on an analysis of public opinion polling data, television network news reporting, and New York Times reporting, there is strong evidence to suggest that Carter suffered domestic audience costs as a result of his failed sanction policy. Furthermore, these domestic political costs induced him to authorize a military strike in April 1980. The data show four trends consistent with the theoretical model: (1) the public followed events and thought the issue was important, (2) the public believed the U.S. should "do something” to coerce Iran to release the hostages, (3) sanctions and diplomacy were supported as tools of coercion, but support for force only materialized after its attempted use, and (4) the public punished Carter for his failed policies. Public interest in Iran was high after the hostages were taken. Over 90% of Americans followed events in every month of the conflict. Iran and the ensuing energy crisis was also regularly named as one of the top three “the most important problems facing the coimtry.” All the major network news telecasts reported at least once on the hostage situation during their 444 days in captivity and the New York Times gave firont 171 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. page attention to the crisis everyday. A television program on ABC called Nightline was created to provide daily updates. Finally, the synergism of the hostage crisis, Soviet invasion of Afghanistan, the energy crisis, stagflation, and presidential campaign season insured that Iran would remain on the forefront of the public’s conscious. Second, the public thought that something could be done about the crisis. 64% believed that the U.S. President could make a real difference on getting the hostages out of Iran, while only 28% thought otherwise. Polls also show that Americans expected a "strong” U.S. President to solve the crisis. How best to force the return of the hostages, however, was uncertain. While 64% believed that a president could "make a real difference on getting the hostages” out of Iran, there was little consensus on which policies would most likely achieve that goal. Over 70% of the public in December 1979 and March 1980 agreed that Carter's policy of patience had kept "alive the chance that one day the hostages would be released unharmed”. A similar question asked by ABC News found 55% acknowledging that Carter’s policies "helped keep the hostages alive.” But beyond Carter’s policy of patience, the polls suggest that the public held unrealistic expectations of what Carter could accomplish. All options available to Carter were either not likely to work or were not supported by the public. While over 70% supported a negotiated settlement, a majority did not want to unfreeze Iranian assets nor provide military spare parts as part of any agreement. These polling data are consistent even as late as November 1980, although the public did support the January 1981 settlement ending the conflict. Over 70% also 172 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. supported economic sanctions against Iran, but support for a naval blockade was mixed, with half the country in support and the other half opposed. Despite high support for sanctions, though, there was not a strong feeling that sanctions would work. 52% agreed that the April 8 trade and financial sanctions would have a minor economic and political impact. In late May 1980,59% believed that sanctions would not secure the release of the hostages. Finally, opposition to the use of military force was high. Nearly 50% did not support a rescue attempt prior to April 25, 1980. In December 1979, close to 70% did not believe that military force could work. Unlike the Gulf War, 60% throughout the crisis disapproved of setting a deadline by which Iran must release the hostages or be subject to military strikes. Only in the case where the hostages were killed or harmed did an overwhelming majority favor the use of force. Interestingly, even though there is little evidence that the public would approve the use of force before April 25, 1980, the failed rescue attempt seems to have increased the public’s desire to settle the conflict quickly. Polls show that 70% of the public approved Carter’s mission attempt, even though it ended in tragic failure. 8 months after the rescue mission, 59% still approved of Carter’s actions. 65% of the public even believed that Carter should have attempted a rescue mission earlier, despite not supporting the use of force previously. As the crisis dragged on, an increasing number wanted Carter to act “tougher’ against Iran, but when asked what should be done, the public could not answer. Hence, even though the public had no viable solution, they demanded that Carter find a way to get the hostages back. 173 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The public punished Carter for his failed policies, even though the public was unable to find an alternative, nor blamed Carter for the crisis. Carter’s approval ratings in handling the Iran crisis dropped by nearly 40 points between November 1979 and August 1980 with disapproval ratings reaching the high 60s and 70s. For example, according to a poll conducted by ABC News, between December 1979 and July 1980. the percentage of Americans approving Carter’s "handling of the situation in Iran” fell from 66% to 19%. Those disapproving of his handling of the situation increased from 32% to 79%. The public was clearly unhappy with Carter’s performance, as an increasing number believed that the failure to solve the hostage crisis was damaging U.S. international prestige. At the time of the rescue attempt. Carter’s approval rating for handling the situation in Iran had dipped below 35% and Senator Edward Kennedy was showing signs of successfully challenging Carter for the Democratic nomination. Finally, there was a clear link between Carter’s failed sanction policy in Iran and his overall approval rating. Carter’s failure also played a significant role in his 1980 presidential defeat. Even though the public was split over whether Republican candidate Ronald Reagan or Carter would be best at handling the hostage crisis, the public perceived Carter’s past failures as evidence of his inability to be an effective leader. With Iran a voting issue. Carter’s failures cost him his political career. 174 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Opinion Carter's overall approval rating was closely linked to public approval in solving the Iran hostage crisis. Both declined over time, but Carter's approval for handing the situation in Iran often was higher than his overall approval. Public Opinion: Overall and in Iran September 1979 - December 1980 O) I 50 40 sept 79 nov 79 jan 80 mar 80 may 80 july 80 sept 80 nov 80 Date Analysis o f Sanction Episode in the Context o f the Model This episode provides significant support for the theoretical model on self- imposed costs by showing how a failed sanction episode generates domestic audience costs that encourage a president to escalate. Although the public neither blamed Carter for the crisis nor had a viable solution. Carters approval ratings declined over time. After imposing the maximum sanction possible on April 8. Carter elected to authorize the use of force as his public approval continued to fall. Polling figures showing 175 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Kennedy posing a strong challenge to Carter in the April primaries also contributed to his decision to authorize the failed military strike. Like the case of Chile, Panama, and Uganda, the role sanctions may have played to encourage third parties to employ military force, in this case Iraq, also requires further research. 176 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C ase X: U.S. vs. Ira q (1990-1991) Summary o f Findings U.S. goal: Withdraw fi-om Kuwait Type of target: Strong Major sanctions imposed: August 1990: Freeze Iraqi assets August 1990: Comprehensive U.N. trade embargo and financial asset freeze August 1990: U.S. Navy enforces blockade September 1990: No-fiy zone established U.S. annual cost of sanctions: $1.1 billion Other coercive policies: August 1990: Operation Desert Shield November 1990: Doubling of troops in Operation Desert Shield converts force to one with offensive capabilities January 1991: Operation Desert Storm begins February 1991: Ground war begins Did sanctions work (HSE): No Generation of audience costs: Yes Historical Background and U.S. Goal Less than 2 years after ending a war with Iran, Saddam Hussein again brought Iraq to war in August 1990. Kuwait had been alleged by Iraq to be producing both too much oil and taking oil reserves under Iraqi territory. According to Hussein, this was causing oil prices to be too low for Iraq to generate sufficient revenue to rebuild the war-torn economy. After failing to win Kuwaiti acceptance of an Iraqi plan calling for less oil production, Hussein sent troops into Kuwait on August 2. 177 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The world community’s response to Iraq’s invasion was strong and swift. The same day Iraq attacked Kuwait, the United States froze ail Iraqi assets and imposed a comprehensive trade embargo. Four days later, the U.N. Security Council imposed a embargo, calling on member states to freeze Iraqi assets and end all trade relations. On August II, the U.S. Navy was enforcing a blockade, an action approved by the U.N. Security Council on August 25. Thus, unlike the case of Iran where sanctions were imposed unilaterally and incrementally, within 30 days of the crisis, Iraq was the first state ever economically isolated from the rest of the world. President Bush also generated domestic audience costs by committing the U.S. to liberating Kuwait through the ‘iying-hands” approach. On August 5. Bush told reporters that Hussein’s actions '"will not stand.” By November 1990, over 200,000 troops were stationed in Saudi Arabia as part of Desert Shield. After the November congressional elections. Bush announced a doubling of U.S. forces as the U.S. assumed an offensive position. On November 29. U.N. Security Council Resolution 678 set a January 15 deadline for Iraq to withdraw from Kuwait or be subject to a collective attack. After repeated efforts by the Soviet Union, U.N., and U.S. diplomats to negotiate a resolution failed, the United States began military operations on January 17. 100 hours after the initiation of a ground war on February 22, Iraq conceded defeat and signed an agreement ending military hostilities. 178 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Target Type Iraq is considered a strong target in the context o f the model. Although not supported by the international community, Iraq had the 4“' largest military in the world at the time of the war. Hussein also used repression to control the opposition. Relative Military Capability (USAC, 1990): 21.7 Relative Number of Troops (USAC, 1990): 1.57 Relative Economic Capability (World Bank, 1990): 1139.6 Relative Population Size (World Bank, 1990): 13.77 Generation o f Domestic Audience Costs An analysis of public opinion polling data, television news coverage, and New York Times reporting supplies clear evidence that Bush suffered domestic audience costs as the economic costs of sanctions increased over time. The data also show that the timing of Bush’s decision to impose the November 15 deadline for Iraq to withdraw from Kuwait was based, in part, on falling public approval. Five themes emerge from an examination of the data: (1) the public was following events in the Gulf and considered the issue to be a threat to U.S. national security, (2) the public supported diplomatic and economic sanctions, but did not think such techniques would work, (3) the public was willing to employ force, but only after sanctions proved to fail, (4) after passage of Security Council Resolution 687, there was an increase in the support for using force, and (5) Bush’s approval ratings declined over time. Public awareness of events in Iraq was high. While only 57% followed events the day Iraq attacked Kuwait, from August 8 to the start of Desert Storm, over 85% reported following events “closely”. By December 1990,80% could correctly identify 179 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Hussein as the leader of Iraq. Media coverage was also intense. The New York Times reported on the story every day of the conflict. Network news often devoted 8 minutes or more of every nightly newscast to events surrounding the situation in the Gulf, whereas in the 4 months before the Gulf Crisis, Iraq was mentioned an average of only 2 to 3 times per week. Not even Hussein’s threat to ’‘incinerate one-half of Israel” with chemical weapons received more than 1 days coverage. The public also believed that solving the situation in Kuwait was of importance to the United States. Over 20% thought Iraq was ”the most important problem facing the U.S.” in August, September, November, December, and January. At the height of contentious budget battles in October, 10 to 15% still believed the situation in Iraq was the most important problem facing the country. Over 70% believed Iraq posed a direct threat to U.S. national security and 64% thought Iraq to be an enemy within 3 weeks of Iraq’s invasion. In August 1990,45% said that Bush’s comparison of Hussein to Hitler was fair and 61% thought Bush comparing Iraqi actions to those of 1930s Germany was an accurate analogy. Public expectations of a successful resolution were also high. In September 1990, 68% anticipated that Iraq would be forced out of Kuwait and 46% believed that Hussein would be toppled. It is therefore not surprising that the public supported U.S. involvement in the conflict. 73% in August 1990 believed that the U.S. was right “to have become involved in the conflict”, 75% believed the U.S. should be involved in September, 60% thought so in November, and 62% just days before war was set to begin. 78% also 1 8 0 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. “approved the United States’ decision to send U.S. troops to Saudi Arabia as a defense against Iraq,” in August 1990, though the percentage dropped steadily to a low of 63% in December before inching upwards to 65% two days before the January 15 deadline. Although the public was willing to be involved in resolving the crisis, diplomacy was preferred to any other option. There was wide support for negotiations throughout the crisis, particularly before the January 1 5 deadline. The data also suggest that the public believed diplomacy could work. For example, the public thought that face-to- face meetings between Iraqi Foreign Minister Tariq Aziz and U.S. Secretary of State James Baker would make a diplomatic solution more likely, though the public was divided evenly over whether an agreement must require that Iraq either fully or partially withdraw from Kuwait. Support for economic sanctions was also high. Independent of the question's wording, 70 to 80% of Americans supported both comprehensive sanctions and a naval blockade, including the sanctioning of food and medicine. But Americans were not optimistic that sanctions would work, nor did they believe that sanctions could work quickly. The percentage believing that 'The economic boycott [would] make Iraq withdraw from Kuwait, or not” started at 53% in August, but fell to 30% in December. Americans, however, preferred to use sanctions rather than military force until such time that it was clear that sanctions would not work. At the start of the crisis. 80% thought the U.S. should wait for sanctions to work while 17% believed military action should begin immediately. 70% also believed that sanctions should be given more time when asked in September, October, and November. No poll showed a majority o f the 181 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. public favoring the use of force prior to January 1991. However, when asked what to do if sanctions fail, starting in September, a majority approved force. This may explain the results of a controversial poll taken by ABC periodically from September 9 to January 13 showing nearly 70% believing that “the United States should take all action necessary, including the use of military force, to make sure that Iraq withdraws its forces from Kuwait." The public is willing to use force if they think the national interests are at stake, but only after all other means of coercion have failed. After U.N. Security Council Resolution 687 was passed, an equal number supported continuing sanctions after January 15 as those who supported war. As the January 15 deadline approached, however, Americans “rallied-around-the-flag" and strongly supported war. This suggests that either after January 15. the public believed that sanctions had failed, or consistent with the model of Chapter 3. the deadline increased audience costs beyond a(s**), making war preferable to backing down. Interestingly, the percentage of the public expecting “the United States is going to get involved in a war with Iraq” was high. Over 60% thought war was inevitable in August 1990. Asked the same question by ABC periodically, the percentage saying “yes” increased to 71% by November 15 and 77% by January 2. Four days before the start of Desert Storm, 86% expected war. Even when phrased differently, almost 50% believed that war was inevitable as early as August, with all polls showing at least a 10% increase by January. 182 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Finally, public approval for Bush’s handling of the Gulf crisis fell over time. Polls show Bush’s approval rating for "handling the situation in Iraq” around 80% in mid-August and falling to a crisis low of 55% just prior to U.N. Security Council Resolution 687, after which time, approval increased slightly. Such a finding is not a coincidence. The model of Chapter 3 suggests that as public approval falls during a failed sanction episode, the president has incentives to escalate. After a drop of 25% in approval. Bush elected to “tie U.S. hands,” by commiting the U.S. to war on January 15. Once the war started, approval increased to over 80%. But unlike the Iran- hostage crisis, the Gulf War did not influence either the 1990 Congressional elections or the 1992 Presidential elections. Although people claimed that U.S. policy in the Gulf was an important issue from which they based their vote, once in the voting booth. Gulf policy was secondary to domestic concerns. When asked what issue mattered most after voting, 6% of respondents in the 1990 Congressional Elections said U.S. policy in the Gulf while less than 0.5% in the 1992 presidential election said the Gulf War and its aftermath. Some scholars have even suggested that the war harmed Bush politically because it reinforced his reputation as a president more concerned with foreign affairs than domestic matters.' Thus, while foreign policy failiures may be correlated with electoral failure, foreign policy successes do not correspond to electoral rewards. ' Muller, John, Policy and Opinion in the G ulf War. 183 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Opinion The data show that Bush's overall public approval was closely tied to public approval of U.S. policy in the Gulf from August 1990 to February 1991. This is not surprising given the degree to which Bush implemented policies with high self-imposed costs— both sunken and “tying hands.” What is surprising is that by November 1992, while the public still gave Bush high marks for his handling of the situation in the Gulf, his overall public approval was relatively low. This suggests that while foreign policy failures may cause a deline in overall public approval, foreign policy successes do not improve overall public approval beyond a few months. Public Approval: Overall and in Iraq July 1990 - January 1991 0 7 0 C L July 8 Aug 12 Aug 26 Sep 11 Sep 30 Oct 14 Oct 28 Nov 11 Dec 2 Dec 16 Jan 13 Jan 22 Date 184 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis o f Sanction Episode in the Context o f the Model The Gulf War provides the best example of how sanctions can be used as a signal of resolve. The United States suffered enormous self-imposed costs both economically and politically before January 17. However, to overcome domestic opposition to the use of force. Bush needed to generate domestic audience costs, such that the cost of backing down would be greater than going to war. A comprehensive embargo was placed on Iraq, 500,000 troop sent to the Gulf, and a deadline imposed by which time Hussein must leave Kuwait or be subject to military attack. It was only after all diplomatic and economic measures had been exhausted that the U.S. public was willing to accept war given the domestic political costs Bush would suffer. Polls conducted the week before the war show that over 70% of the public believed that Bush had done everything possible to avoid war. Hussein incorrectly believed that the U.S. was type high-cost and in his words, without “the stomach for war.” 185 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Case XI: U.S. vs. Libya (1978 - 1986) Summary o f Findings U.S. goal: Destabilize the government Type of target: Weak Major sanctions imposed: 1978: Ban military sales 1979: Prohibit the sale of Boeing 747s 1981: Libyan diplomatic mission in Washington DC. is closed 1982: Ban on new oil and gas deals 1982: Ban the import of oil and the export of high- technology goods 1982: Prevent the sale of $600 million in airplanes U.S. annual cost of sanctions: $1 million 1979-1980 $113 million 1981-1986 Other coercive policies: 1981: U.S. shoots down 2 Libyan Su-22s 1981:6* Fleet is order to prepare to evaculate American as the administration threatens airstrikes 1983: USS Nimitz and USS Eisenhower conduct exercises near Libya and are periodically challenged by the Libyan airforce. No shots are exchanged. 1983: U.S. sends aid to Chad government 1985: Naval exercises intensify 1986: U.S. shoots down 2 Libyan aircraft in March 1986: Reagan threatens war and orders airstrikes Did sanctions work (HSE): No Generation of audience costs: Yes Historical Background and U.S. Goal Muammar el Khadafi’s sponsorship of terrorism has always been of deep concern to the United States. Although populated by fewer than 8 million people, this oil rich state has sponsored some o f the most notorious terrorist acts in modem history. 186 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Libya was also financially supporting organizations, such as the PLO and Sandinistas, that threatened U.S. national interests. Khadafi’s support of terrorism began in earnest in September 1980, and reached its height in 1985-1986. The conflict between the U.S. and Libya peaked in April 1986. Blamed for the April 2, 1986 in-flight bombing of TWA Flight 840 between Italy and Greece, as well as the April 5 bombing of a West Berlin nightclub, the Reagan administration elected to launch retaliatory airstrikes against Libya on April 14. However, a review of television news reporting since 1977 shows that the conflict had been building for some time. Relations between Libya and the United States can be categorized into four stages. From 1977 to 1981, the U.S. imposed an arms embargo and minor diplomatic sanctions against Libya for its sponsorship of terrorism. From 1981 to 1982, both sanctions and tensions increased. 3 months after ABC, CBS, and NBC aired reports on July 26, 1981 that CIA director William Casey had approved a plan to destabilize the Libyan government. Khadafi is alleged to hired people to assissinate American officials, an allegation formally denied by the CIA. Three weeks later, on August 19,2 U.S. F- 14s shot down two Libyan Su-22’s over the Gulf of Sidra. Additional sanctions were then imposed on October 28 because of Libya’s role in the civil war in Chad. The assassination of the U.S. charge d’affaires to Paris on November 12 and the discovery of a Libyan-sponsored assassination plan against U.S. Ambassador Max Rabb increased concerns that Libya may try to assassinate Reagan and other top officials. Starting November 21, security aroimd Reagan was augment amid media reports that Khadafi had sent a team to assassinate him, and unlike the previous month, the CIA neither 187 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. confirmed nor denied the allegation. After the assassination of a U.S. Defense attaché in Paris on December 6, Reagan ordered all Americans to leave Libya in preparation of possible U.S. military retaliation. No additional sanctions were imposed, however, and the few workers that abided by Reagan’s order returned. The U.S. did not impose additional sanctions until a February 1982 ban on oil and gas agreements. Reagan expanded this to a formal ban on oil imports and high technology exports in March. The third stage of conflict occurred between 1983 and 1984 as the U.S. became increasingly concerned over Libya’s support of the Sandinistas in Nicaragua and the war in Chad. In late March 1983, administration officials confirmed allegations that Libya was supplying military and economic aid to Nicaragua. Libyan military personnel were also reportedly aiding Chad rebels in their attempt to overthrow the pro-U.S. government in Chad during a military campaign in August. In response, the United States conducted military exercises near Libya periodically from April 1993 to July 1984 when events caused Reagan to escalate the conflict by ordering the U.S. Navy to conduct exercises in the Gulf of Sidra. Throughout 1984, Libya was repeatedly cited as a sponsor of specific terrorist activities throughout Europe and the Middle East. The final stage of conflict occurred from 1985 and 1986, ending in U.S. airstrikes on April 14, 1986. Throughout 1985, worldwide terrorism was on the forefront of American public opinion; reaching a climax after TWA Flight 847 was hijacked in June. In a major foreign policy speech on July 8, 1985, Reagan cited Libya as a major sponsor o f terrorism and called upon the world to join the U.S. in isolating Khadafi. Although Reagan’s call fell on deaf ears, the stage was set for conflict 188 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. escalation. The CIA reportedly was working on the destabilization of the Khadafi regime and U.S. naval exercises off Libya began on November 29. After announcing that Libya was behind airport bombings in Vienna and Rome in December 1985, a White House Press Secretary told reporters on December 30 that Reagan was considering launching a military airstrike. The next day, Reagan formally asked the Joint Chiefs of Staff to prepare military options. On January 7, 1986 Reagan aimounced the imposition of more sanctions. Within one month, there was an increase in U.S. military presence in the Gulf of Sidra and Reagan ordered all Americans to leave Libya. During exercises in March 1986. U.S. and Libyan forces clashed over the Gulf of Sidra. Following the in-flight bombing of TWA 840 and the bombing of a popular West Berlin nightclub in early April. Reagan authorized U.S. airstrikes on April 14. Target Type Libya is considered a weak target in the context of the model. Even though Libya did have considerable military expenditures and a repressive government. Libya also had a small population, suffered from domestic unrest, and received only mild support from the Soviet Union. Relative Military Capability (USAC, 1979): 33.14 Relative Number of Troops (US AC, 1979): 40.2 Relative Economic Capability (World Bank. 1979): 93.96 Relative Population Size (World Bank, 1979): 77.26 189 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Generation o f Domestic Audience Costs Based on an analysis o f both public opinion polling data and media coverage, the data show that Reagan’s failure to stem the tide of Libyan sponsored terrorism generated domestic audience costs. The public was concerned about international terrorism, and believed that the U.S. should do something about the problem. As public interest in terrorism increased, and President Reagan elevated Libyan terrorism near the top of his political agenda, the American public would come to judge Reagan’s competency as a leader, in part, on his handling of the situation in Libya. Media coverage of worldwide terrorism was intense. Television coverage, in particular, was high, whereas that of the New York Times was relatively low. Part of the explanation for the variance in coverage concerns the greater publicity terrorist acts receive from television media. Minus commercials, network telecasts report for 22 minutes. Terrorist activities are designed to capture this influencial media medium. Thus, relative to other episodes, television media spent considerable time on Libya’s alleged sponsorship of terrorism. With the media focused on terrorism, a failure by Reagan to escalate would have resulted in the public losing “confidence” in him. An analysis of public opinion polling data also reveals that Reagan suffered domestic audience costs. Public attention to Libyan terrorist activities was reasonably high. 45% of Americans reported seeing Reagan’s January 7, 1986 press conference announcing the freezing o f Libyan assets and other sanctions. Over 90% heard reports that Libya had sent an assassin to kill Reagan, with 64% believing the story was true. The public has also identified Libya as a major sponsor of terrorism from 1977 to 1998. 190 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. with the percentage indicting Libya increasing from 1977 to 1986 and steadily falling starting in 1987. Finally, the public increasingly believed that Libya was either "not friendly” or an "enemy” of the U.S. Whereas only 27% felt this way in October 1977. 40% thought so in July 1980,66% in January 1982,62% in November 1984, and 67% in October 1985. Thus, as media reporting of Libyan sponsored terrorism increased, the public became more convinced that Libya was a threat to U.S. national security, even though the number of terrorist activities was relatively constant from 1981 to late 1985, and increased after the bombing of Pan Am Flight 103 over Scotland in 1988. Public opinion was mixed over how to best attack Libya’s sponsorship of terrorism. When polled in 1981 over their opinion regarding U.S. sanctions, over 70% stated that they supported such measures. Support for a covert operation to destabilize Khadafi was also strong, though the public overwhelming opposed killing Khadafi. Thus, while the public wanted to see Khadafi removed, the public was not willing to accept the cost associated with such an action. The same holds true for the use of military force. From 1981 to March 1986, support for force never exceeded 50%. Even when asked in 1981 if force should be used if Libya assassinates Reagan, over 85% of the public said “no,” although one suspects that U.S. opinion would have changed had Reagan been assassinated. Over time, though, the percentage supporting force increased. When asked if force should be used if evidence emerged that Libya was behind terrorist activities in April 1986, approximately 60% said “yes.” Support for force after the April 1986 airstrikes was also high. This provides clear evidence that over time, U.S. policy against Libya generated domestic audience costs. 191 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Opinion Despite the intensity of media reporting on Libyan sponsored terrorism from 1981 to 1986, there are no polling questions deposited with the Roper Center asking Americans if they approved of either President Carter’s or President Reagan’s handling of the situation until after the April 1986 U.S. airstrikes. This is unfortunate given that airstrikes were launched, in part, because of public demands that "something be done” about terrorism. Finally, while the airstrikes did increase Reagan’s overall approval in April and May 1986, no other months appear to have been effected. Overall Public Approval July 1978 - May 1986 70 60 1 “ a. Q . < o i s 40 3 C L 30 ' I I 20 J july 78 mar 79 nov79 julySO mar 81 nov81 jûlÿ 82 mâr 83 hov83 jûlÿ 84 riiaKSS nov 85 Date 192 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis o f Sanction Episode in the Context o f the Model Hufbauer, Schott, and Elliott found that sanctions against Libya failed to deter Libyan-sponsored terrorism. However, when analyzed as a signaling tool, the data suggest that sanctions worked as they are intended to work— by threatening future costs. Cumulative self-imposed inflation-adjusted costs against Libya were only exceeded by the sanction episodes of Iran and Iraq. Not even the total cost of U.S. sanctions against South Africa reached Libyan levels. Domestic audience costs increased as more sanctions were imposed and U.S. forces deployed to threaten Libya. After the TWA and West Berlin terrorist acts, the public was ready to impose the type of force it was not willing to approve at any time between 1981 and March 1985. Reagan’s previous policies had failed and only through the use of military force could he "backup” his nationally reported speeches stating that the U.S. would fight terrorism. 193 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Case XII; U.S. vs. Nicaragua (1977 - 1979) Summary o f Findings U.S. goal: Destabilize the government Type of target: Weak Major sanctions imposed: 1977: Carter cuts military aid by $2.5 million 1977: Carter reduces aid by $12 million 1978: Congress forces Carter to resume aid 1979: Carter ends all military and economic aid U.S. annual cost of sanctions: -$13 million Other coercive policies: None Did sanctions work (HSE): Yes Generation of audience costs: No Historical Background and U.S. Goal Nicaraguan strongman Anastasio Somoza was a strong ally in the fight against communism for many years. But Somoza’s repressive human rights policies, including mass killings and torture, were disconceming even to Somoza’s most ardent supporters. No longer finding his repressive techniques acceptable. President Carter suspended aid in 1977. Congressional efforts to resume aid were successful in May 1978, but as repression increased, Congress passed legislation banning aid in the summer of 1978. On January 15, 1979, Carter formally terminated all economic and military aid. By April 1979, what had been low-intensity fighting between government troops and Sandinista rebels erupted into a full-scale civil war. Without U.S. political, economic, or military support, Somoza realized that he was destined to lose the war and he resigned from office on July 17, 1979. 194 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Target Type Nicaragua is considered a weak target in the context of the model. Nicaragua has a small population and low military expenditures. Civil unrest also made political repression difficult and costly. Relative Military Capability (USAC, 1977): 155.95 Relative Number of Troops (USAC, 1977): 343.33 Relative Economic Capability (World Bank, 1977): 933.65 Relative Population Size (World Bank, 1977): 85.41 Generation o f Domestic Audience Costs Unfortunately, no polling agency deposited questions concerning U.S. public attitudes toward Nicaragua with the Roper Center. However, based on frequency and content of media reporting, there is little evidence that domestic audience costs were generated. The New York Times did publish numerous articles on Nicaragua in the final year of Somoza’s reign, as well as print reports over the differences between the Carter administration and Congress over aid in 1978. Television media reporting was also present, although at low levels. Other than an ABC story on the decline in coffee output, the Vanderbilt Archive does not record Nicaragua or Somoza in any abstract in 1977. All 3 television networks did report on human rights abuses and political unrest about 2 times a month between January and March 1978, but there were no reports from April to July. From August 1978 to Somoza’s resignation in July 1979, network television did regularly report on events, devoting no less than 2 stories per month. 195 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Opinion Carter’s successful policy in Nicaragua did not improve Carter’s overall public approval, although it may have prevented a continued decline in public support. Overall Public Approval June 1977-July 1979 >50 14 0 june 77 sept 77 dec 77 m ar 78 june 78 sept 78 dec Date Analysis o f Sanction Episode in the Context o f the Model Hufbauer, Schott, and Elliott found that U.S. sanctions against Nicaragua succeeded. This finding is interesting given that the U.S. goal was identical to that of U.S. goals in Guatemala, Paraguay, and Uruguay, all 3 of which were found to have failed. While there were no domestic audience costs, the loss of U.S. support was critical to Somoza’s decision to leave power. Recognizing that he was a weak target facing a strong sender, Somoza resigned before the conflict escalated. Sanctions therefore worked by credibly signaling the potential for escalation. 196 Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. Case XIII: U.S. vs. Nicaragua (1981 - 1990) Summary o f Findings U.S. goal: Destabilize the government Type of target: Weak Major sanctions imposed: 1981: Reagan freezes economic aid 1981: U.S. ends all aid 1983: U.S. closes 6 consulates 1985: U.S. bans all economic transactions U.S. armual cost of sanctions: $83 million Other coercive policies: 1981: U.S. helps organize the Contras 1983: Joint U.S.-Honduras exercises. Big Pine 1983: U.S. funded Contras begin civil war 1983: Big Pine II begins 1984: CIA helps mine Nicaraguan harbors 1985: Big Pine III begins 1986: Iran-Contra scandal is exposed Did sanctions work (HSE): No Generation of audience costs: Yes Historical Background and U.S. Goal The Nicaraguan people received only a short respite in their confrontation with the United States after Somoza’s resignation as the Reagan administration aggressively sought to destabilize the Sandinista government. On January 22,1981, all aid was frozen, denying Nicaragua $35 million in aid. Aid to the rebel Contras started in 1982. Following the outbreak of civil war in March 1983, the first trade sanction was imposed on May 9, 1983 when Reagan revoked Nicaragua’s sugar quota. U.S.-Honduran joint military exercises began in early 1983 and continued through 1986. In 1984, the Contras, with assistance from the CIA, mined the harbors of Nicaragua in violation of 197 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. international law. The next major sanction was imposed on May 1, 1985 when Reagan invoked the EEEPA. This ended all trade with Nicaragua, prohibited Nicaraguan registered ships from entering U.S. ports, and terminated air traffic between the two states. Finally, in violation of Congressional legislation banning aid to the Contras, members o f Reagan’s cabinet, the NSC. and the CIA diverted revenue generated from the covert sale of weapons to Iran to the Nicaraguan Contras. When the Iran-Contra story was exposed in November 1986, Reagan’s overall approval fell from 63% in October to 47% in December. In 1987, with Cogressional hearings into the Iran-Contra affair in full swing, the U.S. agreed to abide by an August 1987 agreement negotiated by Costa Rica President Oscar Arias Sanchez that called for a cease-fire and presidential elections. Military support for the Contras ended. Following the election of opposition leader Violeta Chamorro de Barrios in February 1990, President Bush lifted all sanctions. Target Type Nicaragua is considered a weak target in the context of the model for reasons similar to that of the last case study. Unlike Somoza, the Sandinistas may have received international support, particularly from the Soviet Union, but such support was too fragile to warrant a change in catagorization. Relative Military Capability (USAC, 1981): 84.19 Relative Number of Troops (USAC, 1981): 39 Relative Economic Capability (World Bank, 1981): 1230.77 Relative Population Size (World Bank, 1981): 79.59 198 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Generation o f Domestic Audience Costs Reagan’s failed policy in Nicaragua generated significant domestic audience costs. Prior to 1983. there was little public interest of events in Nicaragua. However, as media reports of the failure of U.S. policy against Nicaragua increased, and debate in Congress raged over Contra aid, the public showed signs of punishing Reagan both for trying, and failing, to destabilize the Sandinistas. An analysis of the data shows that the public was interested in stopping the spread of communism, but opposed both arming the Contras and using U.S. force to achieve that goal. However, because Reagan had made the overthrow of the Nicaraguan government such a high-profile issue, the failure of sanctions and the subsequent generation of audience costs resulted in escalation. Public awareness of events in Nicaragua was high before Reagan imposed sanctions in 1983. but low from 1981 to 1983. For example, while neither the New York Times nor network television news reported much on Nicaragua in 1981. after the passage of the Boland amendment in December 1982, Nicaragua received regular media coverage from both outlets, with the number of stories reaching a peak in November- December 1986 after the Iran-Contra story broke. Public opinion polls follow a similar trend. The percentage following event in Nicaragua either “closely” or “casually” was 62% in June, 1979 a month before Somoza resigned, but 74% in July 1984 when Congress was debating extending aid to the Contras. The percentage who had “read or heard anything about the fighting between the Sandinista government in Nicaragua and rebels seeking to overthrow the government,” started at a relatively high 59% in August 1983 and exceeded 72% in all of 1985 and 1986. 199 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. However, Americans’ knowledge of the situation was low until 1987. When asked if the U.S. had “been supporting the government of Nicaragua, or not,” only 31% responded affirmatively while 22% answered negatively. Even as late as April 1986. 43% stated that the U.S. was supporting the government while 32% believed otherwise. Another poll found that the percentage believing that the U.S. was supporting the government rather than the Contras increased from 26% in May 1984 to 50% in March 1986. But when the question was worded differently, Americans were more likely to say that the U.S. was supporting the Contras. When asked if they knew “which side the U.S. is backing (in fighting between the Sandinista government in Nicaragua and rebels seeking to overthrow the government), the rebels, or the government.” in January 1984. 25% answered the rebels and 27% the government, but these percentages increased to 46 and 20, respectively, in June 1985. By March 1986. nearly 60% could identify whom the U.S. supported, while only 13% provided incorrect answers. Thus while the public claimed to be following events, it was not knowledgeable of the U.S. position. Americans still found events in Nicaragua to be important for the United States even though they were unaware of which side the U.S. supported. 70% in June 1983 believed that it was “important for the United States to be involved in Nicaragua” to stem the spread of communism. When asked what issues were of concern to them in an open-ended question allowing for multiple responses, 2% to 9% said “Central America, or Nicaragua,” in 11 polls conducted between December 1983 and January 1988. A majority also believed that events in Nicaragua posed a serious risk to U.S. national security. When asked if they would “describe the importance of El Salvador and other 200 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Central American countries— like Guatemala and Nicaragua-to the defense interests of the United States— [as] very important or not very important,” 57% responded '^ery important” in March 1982 and 61% the following month. A poll conducted by ABC found that 30% believed that the * ‘coimtry of Nicaragua is a major threat to the security of the United States,” while 24% stated it was only a minor threat over numerous polls in multiple years. The percentage believing that Nicaragua did not pose a security threat was 30%, except for in a March 6, 1986 poll where 43% believed Nicaragua was not a threat. Interestingly, when asked if Nicaragua was spreading communism, a majority said "yes,” but when asked if poverty or foreign states like Cuba, Nicaragua, and Soviet Union were "the greater cause of unrest in Central America,” 55% said poverty and 25% said foreign subversion in 6 polls taken between 1983 and 1987. Public opinion of Nicaragua was also low, with less than 20% having a positive image. Despite this opposition to the Sandinistas, Americans consistently opposed U.S. involvement in the situation. The percentage believing that the U.S. should facilitate the overthrow of the Sandinistas never exceeded 30% in any public opinion poll, while the percentage opposing undermining the Sandinista regime never dipped below 50%. Support for a negotiated settlement to the war, however, was high. Even U.S. sanctions against Nicaragua received only luke-warm support, as only 53% of the public supported Reagan’s decision to impose a comprehensive trade embargo in May 1985 and 41% opposed his decision-a surprising finding given that the 'rally-around-the- fiag’ hypothesis would suggest that support for sanctions would be higher. 2 0 1 Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. But it was the debate over aid to the Contras where the public most clearly expressed its desire to not be involved. Support for aiding the Contras was always low. In over 46 differently worded poll questions, asked 72 times between 1983 and 1989. not once did support for military aid to the Contras exceed 40%, with most polls showing only 20% to 30% of the public supporting military aid. In these same polls, opposition to military aid stood between 50% to 60%. Interestingly, the Iran-Contra story did not significantly change public attitudes concerning aiding the Contras. Support for non-military aid was stronger, but still well below 40% in all polling questions except when told the aid would send "ffood and medicine." The public opposed the use of force in greater proportion than that of opposing Contra aid. Military exercises in Honduras were disapproved by 60% in 1983 and 1985, and 50% in 1986. Support for the exercises was in the low-3 O s in 1983 and 1985, increasing to 41% in 1986. Opposition to an invasion was above 60% from 1983 to 1987, while support for an invasion went from the high-1 O s to the mid-30s. Given the variety of question wordings, this suggests that there was no change in Americans' aversion to war. Opposition to force remained even if aid to the Contras failed. Nicaragua invaded another country, or the Soviet Union sent warplanes into Nicaragua. Given that the public believed that the situation in Nicaragua was important, and that they disapproved of efforts to destabilize the government, it is not surprising that Reagan’s approval ratings in handling the crisis were low. When asked how they would rate Reagan “on handling the situation in Nicaragua— excellent, pretty good, only fair, or poor?,” 25% of those polled said excellent or pretty good while 65% said fair or 202 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. poor. Asked the same question 2 years later, 31% said excellent or pretty good and 63% said fair or poor. Gallup polls asking respondents if they "approved or disapproved of the way President Reagan [was] handling...the situation in Nicaragua” showed an average approval of 27% and disapproval of 52% between 1985 and 1988. Following the discovery of the Iran-Contra affair, public approval declined in all polls. Changes in Public Opinion The data reveal that Reagan’s overall approval was negatively impacted by his failed policies in Nicaragua, particularly after the Iran-Contra scandal was exposed. At the time sanctions were first imposed in 1981. Reagan’s approval rating was above 50%. However, as the economy fell into a recession, Reagan’s overall approval fell to a low of 35%. When the economy improved, Reagan’s overall approval reached a high of 68% in May 1986. By December, it stood at 65%. One month later. Reagan’s approval fell to 47% as a result of the Iran-Contra scandal. As the graphs below demonstrate, approval for Reagan’s handling of the situation in Nicaragua was 25-30 percentage points below that for his overall approval. 203 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Overall Public Approval January 1981 - February 1990 lan Sljuly 81 jan 82|uly 82jan 83|uly 83jan 84july 84|an 85 jul 85 jan 86july 86|an 87]uly B7jan 88|uiy 88jan 89july 89jan 90 Date 204 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. O verall Approval an d in N icaragua March 1985 - July 1986 6 50 -9 40 - a S«pt2 0«cS Date Analysis o f Sanction Episode in the Context o f the Model The failure of sanctions, along with declining public approval, prompted the administration to escalate the conflict in the hopes of reestabilishing public confidence. These forms of escalation, including arming the Contras, engaging in military exercises, and diverting revenue earned from selling weapons to Iran to the Contras, were opposed by the public. Yet. having made the destabilizing o f the Sandinistas a benchmark for measuring his ability to prevent commimism from spreading, the domestic cost for Reagan of conceding defeat were too high. In the context o f the model a(s) + d > w, although the U.S. elected to escalate with tools other than U.S. military force. 205 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Case XIV: U.S. vs. Panama (1987 -1989) Summary o f Findings U.S. goal: Destabilize the government Type of target: Weak Major sanctions imposed: June 1987: Suspend aid December 1987: End all economic and military aid December 1987: Suspend sugar quota March 1988: Federal courts order asset freeze March 1988: All payments to Panama put in escrow April 1988: All Panama assets are frozen August 1989: Break-off diplomatic relations U.S. annual cost of sanctions: $41.3 million Other coercive policies: March 1988: U.S. carrier Okinawa sent to Panama April 1988: 1,300 troops sent to U.S. base in Panama May 1989: More troops sent to Panama August 1989: Military exercises conducted October 1989: Military assists coup attempt that fails December 1989: Invasion of Panama Did sanctions work (HSE): No Generation of audience costs: Yes Historical Background and U.S. Goal General Manuel Antonio Noriega Morena’s story of intractable hold on power from 1983 to December 20, 1989 when 24,000 U.S. troops forcibly removed the Commander of the Panamanian Defense Force (PDF) reads like an academy award winning drama that would make even hard-core soap opera viewers question just how different television is from the realpolitik of international politics. Prior to mid-1987, Noriega was supported in the U.S. by the NSC, CIA, DBA, and DoD as an invaluable ally in the fight against communism in Latin America, influx of drugs, and advancer of 206 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. U.S. security interests throughout Latin America. Noriega’s support of U.S. policy relating to the Panamanian Canal and stationing of troops at Southern Command, was also welcomed by inner policy circles. He was considered such an important intelligence asset that every President from Lyndon Johnson to Ronald Reagan (Jimmy Carter excepted) authorized the CIA to hire Noriega as a foreign agent.' These powerful supporters shielded Noriega from officials such as U.S. Senator Jesse Helms and Assistant Secretary of State Elliott Abrams who were aghast that America would financially support a known drug kingpin and agent for both Cuba and Nicaragua. Intelligence organizations and Panamanian political activists had long known about Noriega’s darker side. Even as early as I960 there were people questioning the benefits of supporting Noriega." Noriega’s ties to the Colombian Medellin cartel were so well known that in May 1984, Noriega invite the head of the cartel, Pablo Escobar, to stay with him while Colombians recovered from a wave of drug-related assassinations. What most irked people like Helms and Abrams was that in addition to being a major link in the drug trade, Noriega was still a paid agent of Nicaragua and Cuba all throughout the 1980s. Not to mention his own political repression of the Panamanian people. Opponents of Noriega saw no value in helping Noriega stay in power. ' Calderon, Richard, “Panama: Disaster or Democracy.” Foreign Affairs. 66:2. (Winter 1987/88) 340. Robinson, Linda, “Dwindling Options in Panama.” Foreign Affairs. 68:5 (Winter 1989/90) 190. Kempe, Frederick, Divorcing the Dictator: America’ s Bungled Affair with Noriega. (New York: G.P Putnam’ s Sons, 1990). * In I960, a U.S. intelligence officer filed a report stating that Noriega had recently been arrested for raping and nearly beating to death a Peruvian prostitute while on break from school at one of Peru’s elite military academies. Despite the report, the U.S. elected to continue to make contacts with Noriega because of his espoused hatred for Communism. 207 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ultimately, it was the media exposure over Noriega’s drug activities which lead to his downfall. The New York Times launched the issue into the forefront of American politics by running a front page, two-part series on June 11 and June 12,1986 claiming that Noriega was involved in drug trafficking, money laundering, and shipping arms to leftist Colombian rebels. The articles additionally indicted Noriega tor the murder of prominent human rights activist Dr. Hugo Spadafora. By July 1986, drug abuse was cited by the public as the fourth most important problem facing the U.S.. supersede only by the threat of war, unemployment, and the budget deficit. With the New York Times articles now part of the political landscape, politicians would feel increasing pressure to remove the most well known drug kingpin in the world. By April 1987, Noriega had lost all of his political supporters in the U.S. CIA direct Bill Casey was gravely ill, Iran-Contra cost most of Noriega’s closest supporters their political appointments, and the public was clamoring that ‘something’ be done about the influx of drugs. Democrats exploited for their own political gain the historical relationship Reagan and Bush maintained with Noriega by passing a Senate resolution on April 3, 1987 recommending the suspension of all economic and military assistance to Panama under the Anti-Drug Abuse Act of 1986. Three months later, on June 26, the Senate escalated the political stakes by passing a non-binding resolution calling for Noriega to relinquish his post. Noriega’s response to the second Senate resolution accelerated U.S. economic warfare. On Jime 30, Noriega sponsored an anti-American demonstration in front of the U.S. Embassy attended by 12 government ministers. The resulting damage to the 208 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Embassy, estimated at over $300,000, caused the administration to impose its first official sanctions on July 1 : suspension o f economic and military aid valued at $26 million annually. Later that month, the CIA officially removed Noriega from its payroll in a vain attempt to quell public uproar over Noriega’s drug dealings. The United States would not impose additional sanctions until December 1987. Following demonstrations in Panama pressing for Noriega’s resignation, the public began to question the effectiveness of U.S. policy in Panama. The earlier sanctions were obviously not succeeding and Democratic Senators were still criticizing the lack of a rigorous economic attack on Noriega. Reagan extended the April 1987 suspension of aid to an indefinite prohibition on December 22. He also signed Congressional legislation both suspending Panama’s allocation of sugar imports and mandating that U.S. officials oppose all loan applications Panama submits to international institutions, such as the International Monetary Fund (IMF). In a symbolic act on April 1. 1988. Reagan “decertified” Panama under the Anti-Drug Abuse Act of 1986. Although there were no additional sanctions imposed, the American electorate perceived this step as a positive move toward destabilizing Noriega. Vice-President Bush also cited this latest act as evidence that he would be 'tough” on drugs if elected President that November. The most serious sanctions against Noriega, though, came not from the executive or legislative branches of government, but the judicial branch. No longer protected by his political allies in Washington D C., the Justice Department approved requests in 1987 by both U.S. Attorney Richard Gregorie in Miami and U.S. Attorney Bob Merkie in Tampa to seek independent indictments against Noriega and 14 others 209 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. for drug trafScking, money laundering, and other criminal acts. On February 5,1988. Federal grand juries in both cities indicted Noriega on all counts, including creating a ‘vast criminal enterprise,’ selling drugs in the United States, and laundering drug profits in Panamanian banks. It was these indictments which made it politically infeasible for President Reagan and Vice-President Bush to press for anything else than destabilizing Noriega. A presidential election was just around the comer and the Bush campaign feared Democratic challengers could convince American voters that Bush had been coddled this drug dictator during his time in office. Not wanting to be seen by the American public as lacking political toughness compared to Congressional Democrats and non-elected judges, the Reagan administration finally decided to impose harsher penalties against Noriega. On March 11, Reagan suspended all of Panama's trade preferences under both the Generalized System of Preferences (GSP) and Caribbean Basin Initiative (CBI); required that all U.S. payments to the Government o f Panama for use of the canal be placed in an escrow account (amounting to $80 million a year); and announced that the joint American- Panamanian company Northville Industries-Petroterminates de Panama would place its $70 million a year annual payments to the Panamanian government in escrow. It would now be politically impossible for the Reagan administration to do anything less than work for Noriega’s removal from power. The last major U.S. sanction was imposed on April 8, 1988 when Reagan invoked the 1977 Emergenôy Economic Powers Act to require that all U.S. individuals, businesses, and government agencies in debt to the Panamanian government pay their 210 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. dues to a special U.S. account. The same act required all U.S. financial institutions to freeze Panamanian assets, though the New York judicial decision of March 2nd already effectively accomplished this act. The only two significant sanctions between April 8 and December 20 was a breaking of diplomatic relations on September I, 1989 when Delvalle’s official term of office expired and a November 30, 1989 ban on Panamanian vessels entering U.S. ports effective January 31, 1990.^ After Bush was roundly criticized in both the press and public opinion for his handling of the October 1989 military coup, not even 5 months after he called upon the Panamanian people “revolt” against Noriega, domestic audience costs became extremely high and Bush authorized operation “Just Cause.” Target Type Panama is considered a weak state in the context of the model. While Noreiga did employ repression effectively, Panama had a small military budget and economy. Noriega was also oppposed by the population and elements within the PDF. Relative Military Capability (USAC, 1981): 84.19 Relative Number of Troops (USAC, 1981): 39 Relative Economic Capability (World Bank, 1981): 1230.77 Relative Population Size (World Bank, 1981): 79.59 ' Had this last sanction taken effect, it is believed that it would have cost both the U.S. and Panamanian economy up to $80 million a year. 2 1 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Generation o f Domestic Audience Costs The data show clear evidence of the generation o f domestic audience costs. Surprisingly, despite sanctions as early as 1986 and the New York Times June reporting of Noriega’s drug dealings, neither Panama nor Noriega were often cited by television network news nor NYT often in 1986 or early 1987. There were only 10 television reports concerning Panama or Noriega in all of 1986, excluding 6 stories reporting that deposed Philippine leader Ferdinand Marcos may seek refuge in Panama. Of these 10. 3 were concerned with Noriega’s drug activities and 7 with overall Central American policy. Other than 4 stories concerning Panama’s refusal to admit Nazi war criminal Karl Linnas so that the U.S. would not have to send Linnas to the Soviet Union, network news did not report on Panama again until June 1987 after protests erupted over Noriega’s rule. After the June strikes, both media sources provided extensive coverage of events in Panama, often citing Noriega’s drug dealing activities. As media exposure increased, audience costs suffered for not destabilizing Noriega also rose. Public opinion polling data confirms that domestic audience costs were generated as sanctions failed. As early as June 1988, 65% of Americans agreed that "it was been humiliating for the United States to find itself incapable of getting Noriega, dictator o f a small country, to quit office.”. Public approval o f both Reagan’s and Bush’s handling of Panama never exceeded 27% in any o f 5 polling question in 1988 and 1989. Additionally, Vice-President Bush was harmed in the 1988 election campaign fi-om Reagan’s dealings with Noriega. By 1989, over 80% of the public believed that it was important to the U.S. that Noriega give up power in Panama. 212 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The public also wanted to see the U.S. undertake policies to destabilize Noriega, but not use negotiations. Opposition to proposals to drop drug charges if Noriega stepped down was 60-70% in polls between 1988 and 1989. Support for sanctions was fairly strong, but the public clearly did not support a comprehensive embargo. Interestingly, while the public did support Bush’s decision to send additional troops in May 1989,60-70% of the public said that they did not favor the use of military force in polling questions asked through October 1989, less than 2 months before operation "Just Cause”. Yet, after the U.S. went to war, approval for force increased to 80%. Changes in Public Opinion While operation "Just Cause” explains the jump in overall public approval in January 1990, it is difficult to assess how the Reagan-Bush failed sanction policy affected approval ratings prior to January. Overall Public Approval June 1987 - January 1990 2 65 J u n e 87 sept dec m ar 88 Ju n e sept dec m ar 89 june sept dec Date 213 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis o f Sanction Episode in the Context o f the Model The case of Panama provides strong support of the argument of Chapter 3. Reagan and Bush declared their intent of destabilizing Noriega by imposing economic sanctions. Once these sanctions, as well as other methods of coercion, failed, presidential approval fell to an extremely low level, usually in the low 20s. After the failed coup in October 1989, the domestic audience costs became so large that the incentives for military confrontation exceeded the costs. The invasion of Panama was driven as much by domestic politics as perceived “national interests:” the latter of which is questionable given that the drug money laundering and drug exporting industry in Panama was not curtailed even after Noriega’s capture nor were other Latin American governments suspected of selling drugs to the U.S. imposed with sanctions. 214 Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. Case XV: U.S. vs Paraguay (1976 - 1981) Summary o f Findings U.S. goal: Fundamental change in human rights Type of target: Weak Major sanctions imposed: October 1976: End all military aid 1978-1980: U.S. votes to deny Paraguay numerous loans at international lending agencies U.S. annual cost of sanctions: - $2 million Other coercive policies: None Did sanctions work (HSE): No Generation of audience costs: No Historical Background and U.S. Goal The cases of Paraguay and Uruguay serve as excellent examples of non-costly sanctions having little or no signaling value. Despite reports of wide-spread human rights violations since President Alfredo Stroessner assumed office in 1954. no U.S. President expressed opposition to Stroessner’s rule until 1977. As part of a “human rights-oriented” foreign policy. President Carter imposed military and financial sanctions against Paraguay demanding that human rights practices be improved. Because of Stroessner’s strong anti-Communist policies. President Reagan removed the sanctions in July 1981. 215 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Target Type Paraguay is considered a weak target in the context of the model based on its population size, relative military capabilities, and economic strength. Paraguay also lacked support from the Soviet Union. The strength of the government came from repressing opposition groups. Relative Military Capability (USAC, 1976): 3640.2 Relative Number of Troops (USAC, 1976): 140 Relative Economic Capability (World Bank, 1976): 1041.18 Relative Population Size (World Bank, 1976): 78.93 Generation o f Domestic Audience Costs The public seemed uninterested in events in Paraguay, and thus no audience costs were generated. Although Paraguay’s human rights record was poor, there was little media reporting of events. For example, there were a total of 11 network news reports concerning Paraguay between 1977 and 1981. Of these 11 reports: 4 discussed Somoza’s assassination in September 1980; 2 reported on Reagan’s decision to lift sanctions against Argentina, Chile, Paraguay, and Uruguay; 1 was on U.S. efforts to send the Shah to Paraguay amid the Iranian hostage crisis; 1 was on Somoza’s taking refuge in Paraguay, and 1 was on Paraguay’s decision to issues a warrant for the arrest of Dr. Josef Mengele. There were only two stories, both in 1977, that were related to human rights abuses in Paraguay. The New York Times indexed Paraguay an average of only 2.8 times per month between 1977 and 1981. Over half of all stories concerned efforts to capture Mengele and the Somoza settling, and assassination, in Paraguay. Only 2 mentioned U.S. sanctions. Not surprisingly, no public opinion polling information is available on Paraguay to help assess public attitudes. 216 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Opinion There is no reason to believe that Carter’s failed sanction policy against Paraguay affected his overall public approval ratings. Overall Public Opinion October 1976 - July 1981 80 70 60 1 a . Q . ■^ 0 u â 3 Q. 40 30 'V I 20 oct7$n 77apr ju ly octjân'78'apr july octjan 79apr july octjan B O ap r Date u ly octjanSlapr ju ly 217 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis o f Sanction Episode in the Context o f the Model Hufbauer, Schott, and Elliott found that sanctions against Paraguay failed. This should not be surprising: the mild sanctions imposed could not signal a willingness to escalate. Not only were the sanctions minor, they even saved the United States $2 million a year. However, one suspects that had President Carter imposed a total trade embargo, events would have turned out differently; as Carter would then be under more political pressure to achieve his objective. The cases of Paraguay and Uruguay demonstrate the utility of domestic audience costs by contemplating the counterfactual case. The goal of the Carter administration was to fundamentally alter the human rights practices of Paraguay. This would have required Stroessner to abdicate power. The fact that sanctions were so “cost-free” meant that the policy would be out of the public spotlight. Without exposure to the violations of human rights in Paraguay, the public could not punish Carter for his failed policy. Compared to the case of Panama and Nicaragua (1977-1979), were the U.S. objectives were similar, there is reason to believe that a more costly sanction would have resulted in either conflict escalation (Panama) or Storessner’s resignation (Nicaragua). 218 Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. Case XVI: U.S. vs. Serbia (1987 -1995) Summary o f Findings U.S. goal: End military campaigns in Bosnia and Croatia Type of target: Strong Major sanctions imposed: 1991: Arms embargo is imposed 1991: EC imposes economic sanctions 1992: U.N. Security Council mandates that U.N. members imposed comprehensive trade, financial, and cultural sanctions 1992 - 1995: Better sanction enforcement U.S. annual cost of sanctions: $52 million Other coercive policies: 1992: U.N. peacekeepers enter Bosnia to deliver humanitarian aid under Chapter VU of the U.N. Charter 1993: 6 cities in Bosnia are designated as “Safe Areas” 1994: NATO launches airstrikes 1994: NATO launches airstrikes 1995: NATO launches airstrikes 1995: NATO launches airstrikes and attacks Bosnian Serbs with cruise missiles Did sanctions work (HSE): n.a. Generation of audience costs: Yes Chapter 5 provides a comprehensive analysis o f the historical background and sanctions imposed, as well as presents evidence showing that audience costs were generated as the accumulated cost of sanctions increased. Finally, note that Serbia is considered a strong target in the context o f the model. Note that the data below is for Serbia only. Relative Military Capability (USAC, 1991): n.a. Relative Number of Troops (USAC, 1991): 137 Relative Economic Capability (World Bank, 1991): 332.41 Relative Population Size (World Bank, 1991): 24.40 219 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Case XVU: U.S. vs. South Africa (1985 - 1991) Summary o f Findings U.S. goal: End apartheid Type of target: Strong Major sanctions imposed: September 1985: In pre-emptive strike, Reagan imposes ban on the exports of computers, nuclear goods, high- technology, and selective loan programs to deter Congress from imposing harsher sanctions October 1985: Ban on imports of gold coins October 1986: Congress overrides Reagan’s veto of the Anti-Apartheid Act of 1986 December 1987: U.S. firms operating in South Afnca must pay double taxes U.S. annual cost of sanctions: $110 million Other coercive policies: None Did sanctions work (HSE): No Generation of audience costs: Slight Historical Background and U.S. Goal U.S. sanctions against South Alrica have been one of the most researched and analyzed episodes of the utility of economic means of coercion. Many scholars argue that sanctions quickened the end of apartheid. Others claim that the end of apartheid in South Afiica owes more to changes in the demographics of the South African labor force and Western banks refusing to extend new international liquidity. In this project, no attempt is made to determine the political effects sanctions had on apartheid. What is of interest is analyzing how this episode compares to the model of Chapter 3. 220 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. South Africa is an interesting case study given the multitude of objectives sought by the United States. The most obvious objective, of course, was that of ending apartheid. A number of U.S. policymakers were also concerned over South African involvement in both Namibia and Angola, though the 1986 Anti-Apartheid Act sanctioning South Africa did not explicitly mention South African withdrawal from Angola and Namibian independence as benchmarks for lifting sanctions.’ However, unlike Rhodesia, which had a comparable system of apartheid and was sanctioned as early as 1965, both West Europe and the United States were reluctant to impose sanctions against South Africa. There were four reasons for this inconsistency. First, the white-minority government was a strong ally in the fight against communism. Second, even though the United States officially wanted South Africa to withdraw from both Namibia and Angola, elements within presidential administrations from Nixon to Reagan were concerned that such a withdrawal would result in the spread of communism in sub-Sahara Africa. Third, South Africa exported to the United States strategic goods only available from less desirable states, such as the Soviet Union. Approximately 30% of the world's gold is supplied by South Africa. South Africa is also the world's largest producer of key industry inputs such as manganese and chrome ore. South Africa's dominance of the world diamond market was also a cause of concern for U.S. policymakers. Finally, investments in South Africa were in the tens of billions of dollars and these investors did not want sanctions affecting their profits. ‘ Public Law 90-440. October 2, 1986. Section 3 II of the Comprehensive Anti-Apartheid Act of 1986 states 5 conditions the President must certify South Afnca complied with prior to lifting sanctions. 221 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Despite such economic pressure to maintain a "constructive engagement" policy, student leaders in late 1984 began to protest their University's policies of investing in South Africa. By January 1985, student protests spread across the country. Universities such as Columbia, Northwestern, and University of Michigan were forced to withdraw their endowments invested in South Africa because of the student protests. As the media reported their efforts, a growing number of Americans believed that it was time to "do something" about the apartheid system in South Afiica. As public pressure built, political leaders at the local and state level also began to withdraw their investments, as well as impose sanctions prohibiting local agencies from conducting business with South Africa. Consumer boycotts against firms operating in South Afiica resulted in drops in revenue and bad publicity for firms such as Coca-Cola. The growing disappointment with U.S. inaction against South Afiica reached the Congress in late 1985. President Reagan was maintaining that U.S. national interests were best served with a policy of "constructive engagement," whereby through economic interaction, the lives of blacks would be improved and South Afiica would gradually breakdown apartheid due to economic and demographic considerations. The public's concern over South Afiica, however, was too great for Congressional delegates to ignore. After sanctions legislation failed in 1985 and 8 days after South Africa defaulted on principal payments of its foreign debt, Reagan imposed limited sanctions in an attempt to discourage Congress from adopting harsher measures. His efforts to prevent Congress from passing restrictive sanctions legislation failed on October 2, 1986 when the Senate voted to override Reagan's veto of the Anti-Apartheid Act of 222 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1986. The House voted to override the veto on September 23. The only significant sanction passed by Congress after 1986 was a 1987 change in the tax code sponsored by Democrats which eliminated tax benefits for firms operating in South Afiica. All U.S. sanctions were lifted in July 1991 after the election of a black-led government in South Afiica. Target Type South Afiica is considered a strong target in the context of the model. South Afiica had a large military capability and economic infrastructure from which she could fight against sanctions. Repressive government tactics also limited domestic dissent. Finally, while the world public opposed the South African regime, many political elites in the United States and Western Europe supported the white-ruled government. Relative Military Capability (USAC, 1985): 82.7 Relative Number of Troops (USAC, 1985): 23.6 Relative Economic Capability (World Bank, 1985): 50.8 Relative Population Size (World Bank, 1985): 7.8 Generation o f Domestic Audience Costs The evidence shows that Reagan did suffer some domestic audience costs for his failure to end apartheid in South Africa, but such costs were minor. One reason for tliis was because the public could not decide on what was the optimal strategy. For example, while over 80% wanted apartheid to end, there was no public support for the use of military force and only weak support for economic sanctions. Yet, as in the case of Iran, although the public did not know what policy best served U.S. interests, a 223 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. president was made to suffer a loss of domestic support for not achieving U.S. goals. While the public was aware of apartheid and abhorred its tenets, there was no consensus on economic sanctions. Public awareness of the situation in South Africa was high, with over 60% of Americans either "closely" or "casually" following events from 1976 to 1985. The public was also fairly knowledgeable of events in South Africa, with 57% aware in 1983 that South Africa was not a communist state, 50% recognizing that blacks did not have the right to vote in the summer of 1985, and 76% stating that they either know "a lot." or "a little" about the apartheid system in November 1985. Nearly 80% also thought that apartheid was "immoral". American sympathies were for the repressed blacks more than the white government between 1985 and 1990 by approximately a 65% to 12% ratio. Between 1981 and 1986, the percentage believing that "the whites in South Africa have a right to run their government, and it is about time we stood up for the rights of white people abroad," fell from 36% to 28%, while those disagreeing increased from 54% to 67%. Hence, the public not only was aware of events and fairly knowledgeable, but the public also supported the rights of South African blacks. However, when asked questions pertaining to South Africa as a threat to U.S. national security, or the potential effects of communism, the public became more circumspect in its opposition to the white government. Poll results suggest that between 1977 and 1981, the public was willing to tolerate an apartheid government if the alternative was communism, or at least not actively interfere in the internal political situation in South Africa. Simultaneously, only 20% of the public believed that South 224 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Africa was either "unfriendly toward" or an "enemy o f the United States, although the percentage did increase to 37 in both 1985 and 1987. Thus the public was not completely ready to punish South Africa for its racial policies, even though many expected that blacks would need to use violence to overthrow the apartheid system. Public support for U.S. involvement reflected this mixture of wanting to end apartheid, but not destabilize a staunchly anti-communist government that provided positive business opportunities to American firms and investors. Support for U.S. involvement was weak in the 1970s, but demands that "something be done" were strong from 1980 to 1990. 62-63% of the public agreed that "it was immoral for the U.S. to support a government that oppresses blacks." in 3 polls asking the same question in 1981. 1985, and 1986. When asked if the U.S. should take a more active role pressuring South Africa to end apartheid, nearly 50% said "yes". However, when 2 1985 poll questions allowed the respondent to say that the U.S. should not be involved, a majority thought that the U.S. "should stay out of the situation altogether". The public was also highly uncertain of what course of action the United States should follow in both 1985 and 1987. Asked what the U.S. should "do next concerning South Africa,” nearly 20% believed that the U.S. should leave South Africa alone, less than 10% advocated additional sanctions, and 50% did not know what the U.S. should do. The public's unclear position on what to do translated into confusion over which policy was superior: constructive engagement or economic sanctions. During 1985 and 1986, numerous polls asked about these two alternatives. Questions asking about the merits of constructive engagement elicited significant support for that course of action. 225 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. But if asked if economic sanctions should be imposed to pressure South Africa to reform, a majority also responds affirmatively. There was clear opposition to ending all trade with South Africa in some polls, but not in others, although there was general support for reducing trade. However, when informed that South Africa possessed strategic resources, public opinion moved in the direction of opposing sanctions. Support for financial and investment sanctions followed a pattern similar to that of trade, with the public in favor in some polls and opposed in others. Overall, however, it appears that the public tended to oppose financial and investment sanctions. Finally, the public opposed import restrictions of South African krugerrand gold, but export controls on military and nuclear goods were widely supported. It is therefore not surprising that there was no consensus as to whether or not sanctions would work. The population was split nearly equally between those believing that sanctions would help and harm blacks. Beliefs that sanctions could end apartheid was also mixed, with 1 poll question finding that the public did think sanctions were effective while another found the opposite. Overall, therefore, there was no clear mandate from the mass public in favor of sanctions. Special Interest groups likely influenced the initial imposition of sanctions, but with the majority of the public less certain, only slight audience costs were generated. In fact, based on public opinion asking about Reagan's "dealing with the South African situation," we find that Reagan's approval ratings actually increased from 36% to 42% between August 1985 and September 1986, just after Reagan vetoed the Anti-Apartheid Act. Other polls show a slight disapprove of Reagan's handling of 226 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. South Africa, but it was not enough to suggest that domestic audience costs were present. Finally, despite efforts to make South Africa a voting issue in both the 1984 and 1988 election, polls suggest that very few people chose a candidate based on their position on South Africa. Even at the height of the sanctions debate in September 1986, only 1% of Americans said they would "change [their] congressional vote this November [1986] because you disagree with a candidate's position on" South Afnca. Changes in Public Opinion Reagan’s failure to reform the South African government did not impact his overall public approval. However, public approval for Reagan’s handling of the situation in South Africa was 25-30 percentage points lower than his overall approval. Analysis o f Sanction Episode in the Context o f the Model While a politically influential segment of the population forced the U.S. to adopt sanctions, the majority of Americans were more cautious. This limited the generation of domestic audience costs. By stating that he opposed sanctions, and only imposed them as a pre-emptive move to avoid more costly Congressional sanctions. Reagan limited his exposure to public criticism if sanctions failed. If there was a signaling effect to sanctions, it came from consumer boycotts and state government sanctions. Opposition to apartheid was strengthening and it was here where the generation of audience costs was most felt. As grassroots anti-apartheid movements increased, businesses and local governments imposed more sanctions to acquiesce domestic audiences, 227 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Overall Public Approval September 1985 - July 1991 .a sept 85m ar 86 sept m ar 87 sept m ar 88 sept m ar 89 sept m ar 90 sept m ar 91 Date 228 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Overall Approval and in South Africa August 1985 - September 1986 60 feb 66 aug 85 dec oct aug apr june □ate 229 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Case XVni: U.S. vs. Suriname (1981 - 1990) Summary o f Findings U.S. goal: Destabilize the government Type of target: Weak Major sanctions imposed: December 1982: U.S. withholds $1.5 million in aid December 1982: Netherlands withholds all aid U.S. annual cost of sanctions: - $1.5 million Other coercive policies: December 1982: CIA-sponsored coup attempt not approved by Congress Did sanctions work (HSE): No Generation of audience costs: No Historical Background and U.S. Goal The case of Suriname is similar to that of Nicaragua and Grenada. An analysis of NYT reporting in 1982 and early 1983 shows that Reagan often cited Soviet-backed governments in Grenada, Nicaragua, and Suriname as reasons why national defense spending needed to be increased. However, unlike the cases of Nicaragua and Grenada. Reagan chose not to escalate the conflict with Suriname. Sanctions were limited to denying $1.5 million in financial aid and reports that in December 1982 the Senate rejected the CIA’s request to use covert force to destabilize the Suriname government. Thus, while the objective sought in Suriname was identical to that of Grenada and Nicaragua, conflict escalation did not occur. 230 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Target Type Suriname is considered a weak target in the context of the model. While Suriname did receive marginal support from the Soviet Union, Suriname had a small economy and military budget. Relative Military Capability (USAC, 1983): 8781.48 Relative Number of Troops (USAC, 1982): 1100.50 Relative Economic Capability (World Bank, 1982): 3067.96 Relative Population Size (World Bank, 1985): 634.60 Generation o f Domestic Audience Costs The lack of media publicity suggests that audience costs were not generated. There were only 8 news reports on Suriname between 1981 and 1990— 2 on the alleged CIA plot to destabilize the government, 1 of an alleged coup, 3 on an airplane crash, and 2 on winning an Olympic bronze metal. The NYT abstracted Suriname an average of 1.34 times per month, with 2 or fewer abstracts containing “Suriname” in 53 out of the 62 months of the episode. The highest number of abstracted stories was 8, in December 1982 when the Netherlands and U.S. imposed sanctions. Changes in Public Opinion There is no evidence to suggest that U.S. policy against Suriname influenced overall presidential approval. 231 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Overall Public Approval January 1981 - February 1990 80 70 | o Q. Q. < O 40 30 jan 81aug 8 tn a r 82oct 82nay 82Uec S^uly 84feb 8Ssept SSapr 86nov 8Qune 87|an 88aug 88nar 89oct 89 Date Analysis o f Sanction Episode in the Context ofthe Model It is interesting to compare the case of Suriname to that of Nicaragua and Grenada. In all three cases, the objective o f U.S. policy was to destabilize the government and prevent the spread of communism in Latin America. Additionally, all three countries were cited by the Reagan administration as jusifications for increasing 232 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. defense spending. What then explains why U.S. policy against both Nicaragua and Grenada escalated while in Suriname it did not? The model of Chapter 3 offers some guidance. Reagan did not make destabilization of the Suriname government a measure of his “competency’ by either imposing costly sanctions or through his rhetoric. In the case of Nicaragua, both costly sanctions and political rhetoric were used to signal American resolve. In the case of Grenada, it was political rhetoric. In the context of the model, since s was less than s*. domestic audience costs were not generated and Reagan did not have the incentives to escalate as he did in Nicaragua and Grenada. 233 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Case XIX: U.S. vs. Uganda (1972 - 1979) Summary o f Findings U.S. goal: Destabilize the government Type of target: Weak Major sanctions imposed: 1972: U.K. ends all aid 1977: U.S. imposes trade and financial sanctions 1978: U.S. imposes comprehensive embargo U.S. annual cost of sanctions: $28 million after 1977 Other coercive policies: None Did sanctions work (HSE): No Generation of audience costs: No Historical Background and U.S. Goal Ida Amin’s government in Uganda had an extremely poor human rights record with an estimated 500,000 dying during his reign of terror. Citizens were routinely arrested, harassed, raped, and murdered. Amin’s racist views toward whites. Jews, Asians, and Indians chilled many Western leaders. Only states such as Libya and Iraq served as allies to a man many considered to be one of the cruelest in the world. The paranoid Amin maintained political power through 3 techniques. First, his repressive acts prevented any opposition from challenging his authority. Second, Amin would attack neighboring Tanzania periodically to divert attention away from his domestic problems. Finally, Amin controlled the states major exported commodity, coffee, to generate revenue to pay for his security apparatus. Because of Uganda’s 234 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. human rights record, the U.S. imposed sanctions in 1977. It was only after Tanzania used military force to drive Amin out of power in March 1979 that the killings and oppression stopped and sanctions were lifted. Target Type Uganda is considered a weak state in the context of the model. Uganda’s population did not support Ida Amin, the Soviet Union did not serve as an ally, and neigboring states were at war with Uganda. Uganda also had a weak economy. Relative Military Capability (USAC, 1972): 268.25 Relative Number of Troops (USAC, 1972): 178.46 Relative Economic Capability (World Bank, 1972): n.a. Relative Population Size (World Bank, 1972): 20.19 Generation o f Domestic Audience Costs There is no evidence that audience costs were generated from sanctions imposed between 1972 and 1979. Nor does the data reveal that President Carter was rewarded after the toppling of Amin. However, an analysis of New York Times abstracts and television network news reporting does suggests that the public was interested in events in Uganda. Both media sources frequently reported on human rights violations. There were also stories in both sources on how late-night comedians and Saturday Night Live would make jokes about Amin. Thus, while the public was aware of events in Uganda, Carter did not feel any domestic pressure to escalate beyond sanctions. 235 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Opinion There is no evidence to suggest that President Carter’s overall public approval rating was affect by U.S. sanction policy against Uganda. Other factors, such as the Iran hostage affair, the poor state of the economy, and the Soviet Union’s invasion of Afghanistan explain the fall in Carter’s overall approval rating more than the failure of sanctions. Similarly, the data do not show that Carter was politically rewarded after Amin was removed from power. Overall Public Approval February 1977 - April 1979 5 60 < 5 5 Date 236 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis o f Sanction Episode in the Context o f the Model Uganda represents another sanction episode where the “counterfactual” demonstrates the usefulness of the model. The human rights practices in Uganda were well reported both in the New York Times and on television network news. However, by not imposing severe sanction early in the conflict. Carter did not make destabilization of Amin a measure of his “competency.” Sunken costs were limited to the sanctions imposed in 1977 and 1978 and there were no U.S. military exercises. Nor did Carter impose ex post audience costs by engaging in a war of words with Amin.' Given the high level of media publicity, larger sanctions could have easily generated such costs. Note that Hufbauer, Schott, and Elliott found that sanctions against Uganda failed. This is not surprising given that Amin found the cost of abdicating power high compared to the cost of sanctions. It was only through the use of force that Amin was overthrown. Like other cases, what role sanctions may have played in Tanzania’s decision to forcibly remove Amin is a topic of future research. However, what is clear is that the U.S. chose not to escalate the conflict by generating domestic audience costs. ‘ Senator Patrick Moynihan (D-NY), however, did engage in a well publicized war of words with Amin. 237 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Case XX: U.S. vs. Uruguay (1976 - 1983) Summary o f Findings U.S. goal: Fundemental change in human rights Type of target: Weak Major sanctions imposed: 1976: Carter denies $3 million in aid 1977: Congress denies all economic aid and President Carter requests that Congress bans all military aid 1977: Carter reduces aid again. Uruguay rejects all forms of U.S. assistance. U.S. annual cost o f sanctions: - $10 million Other coercive policies: None Did sanctions work (HSE): No Generation of audience costs: No Historical Background and U.S. Goal The cases o f Paraguay and Uruguay serve as excellent examples of sanctions having little or no signaling value. Wide-spread human rights violations reported after a military-sponsored coup in 1976 resulted in the Carter administration imposing sanctions against Uruguay as part of the administration’s 'human-rights-oriented” foreign policy. Because of Uruguay’s strong anti-communist policies. President Reagan removed the sanctions in 1981. 238 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Target Type Uruguay is considered a weak target in the context of the model. Uruguay's only major ally was the United States, and Uruguay had a small population, economy, and military capability relative to the United States. Relative Military Capability (USAC, 1977): 883.62 Relative Number of Troops (USAC, 1977): 75 Relative Economic Capability (World Bank. 1977): 487.60 Relative Population Size (World Bank. 1977): 76.63 Generation o f Domestic Audience Costs Very little was generated in the way of domestic audience costs. Like the case of Paraguay. Uruguay received little media attention. Television network news reported on Uruguay 20 times between 1976 and 1981,7 of which concerned human rights abuses. Reports of Watergate figure E. Howard Hunt’s knowledge of CIA payments to Uruguayan government officials and Reagan’s decision to lift Uruguay sanctions each received 3 mentions on network news. The remaining 7 reported about Uruguay in only a peripheral maimer. New York Times reporting of Uruguay was also sparse. According to the NYT abstract indexing system, Uruguay was mentioned an average of 2.7 times per month between 1976 and 1981, with a 1976 coup accounting for most of the stories. No public opinion polls on Uruguay are available. 239 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Changes in Public Opinion There is no evidence that U.S. sanction policy against Uruguay affected presidential approval ratings. The decline in public approval during the Carter administration owes more to the state of the economy, international turmoil in the Middle East, and the Soviet Union’s invasion of Afghanistan than U.S. sanction policy against Uruguay. Overall Public Approval October 1976 - July 1981 80 70 60 1 % o JQ 3 Q . 40 30 20 f- ri ! / ft/ oct 76mar 77 aug an 78 june nov apr79 sept feb 80 July dec may 81 Date 240 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis o f Sanction Episode in the Context o f the Model Hufbauer, Schott, and Elliott found that sanctions against Uruguay failed. This is consistent with the model. Carter wanted Uruguay to change their human rights policies, but was constrained by his desire to maintain Uruguay as an ally. Thus, the minor sanctions he imposed could not signal American resolve or generate audience costs to credibly threaten Uruguay with conflict escalation. In the context of the model, s < s* , d < w, and sanctions were destined to fail. 241 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. IL Statistical Analysis Introduction The case study analysis suggests that there is a positive correlation between sanctions, domestic audiences, and conflict escalation. A more rigorous statistical test is prefermed in this section to verify these findings. Using the 20 cases established as an exhaustive set of sanction episodes, regression analysis verifies that the accumulated cost of sanctions is correlated to a decline in public approval on a specific foreign policy issue, holding constant other explanatory variables. Once such evidence is supported, a ■ ‘predicted audience cost” value is generated for every month for each episode. Another regression shows that a correlation exists between “predicted audience cost” and conflict escalation. Hence, two regressions are performed, one testing for a correlation between " ‘sunken costs” and “domestic audience costs,” and a second testing for a correlation between “domestic audience costs” and “conflict escalation.” Finally, note that the results presented below are preliminary. Further research is necessary to correct for simulteneity errors and lagged variables. But even with this caveat, the statistical findings appear to be robust. First Statistical Analysis: Testing for the Generation o f Domestic Audience Costs The statistical analysis here verifies the most important assumption made in the model of Chapter 3-that accumulated sunken costs increase audience costs. Specifically, as the cost of sanctions increase, the data show that public approval for a president’s handling o f an international problem declines. 242 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. To measure “domestic audience cost,” public approval was recorded to the following question: “Do you approve or disapprove of the way [name of president] is handling the situation in [name of country].”' I then calculated the difference between the initial approval and that of subsequent months. A drop in public approval was recorded as an “increase” in domestic audience cost. For example, according to Gallup polls, on August 10, 1990, 77% of the public approved President Bush’s handling of the situation involving Iraq and Kuwait. On September 16, 75% of the public supported Bush, on October 14, 65% did, and so on. Based on these data. Bush suffered an audience cost of “2" in September and “12" in October. For the initial month of a sanction episode, the audience cost suffered is assumed to be 0 in cases where data are available.* Missing values were used in cases where no data were available. In cases where data were available, but not in the initial month, the president’s overall approval for his handling of foreign policy was employed as a proxyFinally, for the sanction episodes of Uruguay and Paraguay, I assume that domestic audience costs equal 0 for every month. Although no public polling data supports such an approach, I do this for 2 reasons. First, the lack of media reporting in both cases suggests that public opinion ' For some targets, the wording of the question differed slightly. * Theoretically, one suspects it may even be negative given that a “rally-around-the-flag” effect in the initial state of a major foreign policy crisis is often associated with high levels of public approval. We standardize all cases to be 0 in the first month, however, given that only the cases of China, Iran, and Iraq have data available at the time the crisis began. ^ Specifically, data were collected to the following Gallup Poll question: “Do you approve or disapprove of the way [name of President] is handling foreign policy, in general.” This data set is also on deposit with the Program in Political Economy and Public Policy and covers public approval on foreign policy from 1970 to 1996. 243 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. did not change. Second, both sanction episodes involved “negative” self-imposed costs, such that the United States saved money from the sanction. Because numerous studies show that public opinion is stable over time, it can be inferred that domestic audience costs suffered in both episodes was “0” for all months.'* No other episodes warranted assigning “0 ” values. Four independent variables are expected to influence domestic audience costs. The first is media coverage. The more a story is reported, the more likely the public will find that the president is “incompetent” given that the press is more likely to report on presidential failures than presidential successes. The number of times the New York Times mentioned the name of the country in an abstract downloadable from Lexis- Nexis Academic Universe represents a good measure of media frequency coverage.^ The data go back to 1970 and are consistent in its methodology from year to year. Although television news reporting would have been the optimal means to measure media exposure, the TV News Archive data are inconsistent over time. Crigler’s work also suggests that the TV News Archive may be unreliable for such an approach.® " * Studies showing that public opinion on foreign policy matters is generally stable over time include Hoisti, Ole Public Opinion and American Foreign Policy (Ann Arbor, MI: University of Michigan Press, 1996). Shapiro, Robert and Benjamin Page, “Foreign Policy and the Rational Public,” Journal o f Conflict Resolution 32 (1988) 211-47. * Efforts were made to eliminate stories reporting on accidents and sporting events. ® I ran the regressions using TV news coverage and found even more robust results than those reported using New York Times abstracts. However, statistical consistency demands the more rigorous approach outlined here. 244 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. However, using monthly New York Times (NYT) abstract frequency is not an accurate assessment of public awareness. People remember past events.^ Hence, if Iraq is a "big” story in August 1990, there is no reason to believe that the public will forget about the exposure to Iraq they experienced in September 1990. I therefore index each month by calculating the number of stories in that month plus some factor of previous months, starting with the first month of the sanction episode. For example, in March 1977, NYT records "Paraguay” in 6 of its abstracts. In October 1976, when sanctions were first imposed, NYT recorded 2 stories, 3 in November, 1 in December, 0 in January 1972, and 4 in February. Assuming a 30% recall, I calculate media exposure in March 1977 to be 6 + .3*4 + .3'*0 + .3^*1 + .3‘ **3 + .3^*2 = 7.26. The regression analysis considers the cases where the public recalls 30%, 50%. and 70% of past exposure.* Media publicity is expected to be positively correlated to audience costs. The second independent variable is relative economic capability (REC). Using data from the World Bank, I record annual GDP at $ million market price for the United States and divide it by the annual GDP at $ million market price of the target country. The annual data are recorded in each month of the corresponding year. Unfortunately, data are not available for any months in the sanction episodes of Burma and Uganda and these cases were dropped from the sample. In some years in the Suriname sanction ^ Graber, Processing the A few s. * I also ran regressions using a 25% and 75% recall with similar findings as presented here. 245 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. episode, data were also not available, and those months were likewise dropped.’ Relative economic capability is expected to be positively correlated to the generation of domestic audience costs because the public is more likely to penalize a president who fails against an economically small state than a large state. The third independent variable is the self-imposed cost of sanctions. These data are calculated in terms of real U.S. dollars on a month-to-month bases using the data from HSE in 16 of the remaining 18 cases, and data from various sources for the cases of Haiti (1991) and Serbia. Note that the data considers the time sanctions were imposed. For example, the self-imposed cost of sanctions against Nicaragua in May 1985 is recorded starting in that month. Because the model assumes that sanctions generate audience costs by inflicting “sunken cost,” 1 accumulate the cost of sanctions over time. Monthly CPI data from the Bureau of Labor Statistics are used to keep all costs in real June-August 1983 dollars. Finally, 1 recorded self-imposed costs in terms of thousands of constant U.S. dollars. If the coefficient is positive, it demonstrates that sunken costs are correlated with a decline in presidential approval. The final independent variable is “time.” Fearon’s theoretical work shows that “time” is a significant variable in the generation of audience costs. Specifically, he argues that the longer a conflict is active, the greater will be the domestic audience costs. Thus, I expect time to be positively correlated with domestic audience costs. ’ Efforts were made to collect data from other sources to calculate REC over time. HSE’s measure of REC in the year sanctions were imposed is not consistent with later changes. Data from the IMF and U.S. ACDA both lacked consistency and were often not available. Thus, to maintain the integrity of the data, months where there were missing data were dropped from the sample size. 246 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The Resression Model Measurins Domestic Audience Costs [ assume a simple, linear regression with the dependent variable domestic audience cost and independent variables NYT, REC, constant accumulated “sunken cost” of sanctions, and time. Specifically: (l)Y't = Po + P, X„ + p, X,t + P] X^[ + P; X^t + e Where Y; is the calculated domestic audience cost in period t, X,t measures NYT coverage in period t, including recall of previous coverage, Xi, measures REC in terms of current U.S. current $ million by target $ million, Xj, measures the constant, accumulated “sunken cost” of sanctions in period t, X4 , measures the number of months that passed since the sanction episode began, and P are parameter coefficients. The following results are derived using a sample size of N=235 and are read from left to right. Table 4.3: Statistical Results on G enerating Domestic Audiences Costs Constant NYT 30% NYT 50% NYT 70% REC Sanction Time Adj- R- 0.04103 0 .0 1 1 ** 0 .0 0 1 ** 0.4112 E-4* 0.055* 0.52 -0.8858 0.014* 0 .0 0 1 2 * 0.3486 E-4* 0.072** 0.53 -2.8300 0.018* 0.0016* 0.2005 E-4* 0.105* 0.58 * - significant at the 1 % level ** - significant at the 3% level + - significant at the 1 0 % level 247 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Interpretation o f Results The results support the assumption that more costly sanctions are positively correlated to domestic audience costs. Media coverage, relative economic capabilities, and “sunken costs” proved significant at the 3% level when the public recalled 30% of previous NYT stories and were significant at the 1% level when the public recalled 50% and 70%. Time was significant at the 10% level in the 30% recall case and 3% in the others. The adjusted R-squared was above .50 in each case, suggesting that the regession explains 50% of the variance between the actual dependent variable and the mean. Finally, note that the F-value for each regression was significant at the 1% level. The coefficient signs are also consistent with expectations. A marginal increase in 1 “unit” in each of the independent variables is positively associated with a fall in presidential approval. The variable “sanctions” is particularly instructive. Every increase in sunken costs of $ 1 million is associated with a drop in public approval ratings between .02 and .04, depending on media recall. Each additional indexed NYT abstract is associated with a drop in presidential public approval of approximately .014. holding all other variables constant. Relative economic capabilities did prove statistically significant, but the marginal effects were small. Every 1 unit increase in U.S. GDP over target GDP corresponds to an approximate .0013 fall in approval ratings. Finally, time is shown to have significant marginal effects. The longer a sanction episode lasts, the larger are the domestic audience costs suffered by an executive. The data suggest that presidential approval in the handling of a specific foreign policy falls between .05 and .10 for every additional month a crisis continues. 248 Reproduced with permission of the copyright owner. Further reproduction prohibited w ithout permission. Second Statistical Analysis: Testim for Conflict Escalation The model in Chapter 3 suggests that as domestic audience costs increase, there is an increased probability that the conflict will escalate. Having estabilished that audience costs are correlated to the generation o f domestic audience costs, I now consider if audience costs are positively associated with observed levels of conflict escalation. To perform this test, I developed an ordered probit model with conflict escalation as the dependent variable and employed 5 independent variables. Every month of the 18 sanction episodes was recorded with one of 5 possible levels of conflict. These codings are defined as follows: 0: If there is no conflict escalation 1 : If the U.S. openly supports opposition movements with military or economic aid 2: If the U.S. engages in military exercises, but does not threaten war 3: If based on presidential statements, the president makes an overt threat of war 4: If the U.S. engages in military action. For example, in July 1987, the first month of U.S. sanctions against Panama, a “0" is recorded. In December 1989, when President Bush authorized Operation Just Cause, a “4” is recorded. The decision on how to code each month was based on an analysis of television news reporting, NTY abstracts, and chronologies developed for each episode. Note that because of the statistical nature of ordered probit models, one must use extreme caution in interpreting parameter coefficients. This is particularly true given the distribution of the dependent variable in this case. O f998 months available for data analysis, 587 involved no military conflict escalation, 246 were limited to only aiding the opposition, 140 entailed the U.S. displaying military force without an overt threat of war, 11 months involved the U.S. making overt threats of attack, and the U.S. applied 249 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. military force in only 14 months. With so few cases in the tail, the final threshold parameter settled at a value of less than 2.5. The concern over the interpretative nature is so strong, that marginal effects are not reported, even though the signs are consistent with the model’s predictions. Instead, 1 will only be interested in testing if the explanatory variables are statistically significant and show the expected sign.'" Given the literature review of Chapter 2 and the model in Chapter 3, 5 variables are expected to explain conflict escalation. The first is a “predicted” value of domestic audience costs suffered in each month. This predicted value was developed using the regression results established earlier. While the proxy variable differs from actual values in cases where the data are available, the “predicted” value still provides an excellent measure of the variance in declines in public approval between sanction episodes over time." I expect to find a positive correlation between “predicted” domestic audience costs and conflict escalation. The second independent variable is relative military capability (RMC). 1 expect that the U.S. is less likely to escalate a conflict against a strong target than a weak target. Unfortunately, the structure of the dependent variable does not allow for the use of a multinomial logit regression to directly test Fearon’s hypothesis that relative military capabilities explain which states senders select to target, but does not influence the Greene, William H. Econometrics Analysis (Upper Saddle River, NJ: Prentice Hall, 1997). Judge, George ed. Introduction to the Theory and Practice o f Econometrics 2nd ed. (New York, NY: Wiley, 1988). Vtaddala, G.S. Limited Dependent and Qualitative Variables in Econometrics (Cambridge, MA: Cambridge University Pres, 1983). ' ' There is also the concern that this technique produces large standard errors. However, because the t-values are high, this concern is discounted. 250 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. decision to escalate once the conflict has been initiated. Three different measures of relative military capabilities are employed. First, relative military capability (RMCUS) is defined as U.S. current $ million in military expenditures divided by target current $ million in military expenditures. These data were obtained from the United States Arms Control and Disarmament Agency.'" Second, relative military capability of troop strength (RMCT) was found by dividing the number of U.S. troop by target troops. These data were also drawn from the United States Arms Control and Disarmament Agency. Finally, data from the World Bank on population size is used to calculate a relative population capability (RPC), which other scholars have employed as a proxy for military capabilities.'^ Again, all forms of relative military capabilities are expected to be positively correlated with conflict escalation. The third independent variable is a dummy variable called “Cold War” (CW). CW takes on of 3 values. A communist target was coded 0. A target which was neither an ally nor an enemy in the Cold War is coded 1. A Cold War ally is coded 2. Note that for some targets, codings changed over time. For example. Panama was coded a 2 from July 1987 to February 1988, at which time it coded a 1. I expect a negative correlation between CW and conflict escalation. '" Military expenditure data from the Stockholm International Peace Research Institute (SIPRI) yielded similar results as the data from the U.S. Arms Control and Disarmament Agency. The techniques employed to solve missing data problems are worthy of note. Specifically, the data sources supply data for every target in all months considered for the variables RMCT and RPC, with the minor exception of the number of troops in the Grenada military for the month October 1983. Here, I substituted the RPC as a proxy for RMCT. For the variable RMC, data are not available, at times, for Angola, Libya, Nicaragua (1981), and Suriname. Without any loss in generality, I substitute the subsequent year RMC for the current year when available, otherwise I use the previous year. For the case of Grenada, I substitute RPC. 251 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The fourth independent variable is overall public approval rating. Monthly data was collected for this variable from Gallup public opinion polls asking Americans: “Do you approve or disapprove of the war [name of president] is handling his job as President?” While this variable is included because diversionary war scholars suggest that falling overall approval explains escalation, the model of Chapter 3 suggests that overall approval should be insignificant.'■ * Finally, “Time” is again included as an explanatory variable. The Resression Model Correlating Domestic Audience Costs and Conflict Escalation 1 assume a simple linear regression model with the dependent variable conflict escalation and independent variables relative military capability, predicted domestic audience cost (DAG), overall public approval (APRV), cold war status (CW), and TIME in an ordered pro bit model. Specifically: (2) Y, = P, + p, X„ + p,X ,, + pjXj, + P,X., + € Where Yt is level of conflict present in month t, X„ measures relative military capabilities, assuming one of the 3 forms, Xi, measures the predicted audience cost in month t, using 3 measures of NYT recall, X3 , measures overall presidential public approval in month t, X41 measures the number of months that passed since the sanction episode began, and P are parameter coefficients that estimate relative log probabilities. " Again, this data set is available from the author. 252 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The following results are derived from a sample size of N=998. Note that parameter estimates for “Time” and “Overall Approval” are not reported because both were statistically insignificant in all regressions. Furthermore, dropping these variables did not change either the sign or significance level of any of the other variables. Table 4.4: Statistical Results on Causes of Conflict Escalation Coefficient Parameters (index function for probabilities) Constant RMCUS RMCT PRC DAC3 DAC5 DAC7 CW .594* .877 E-04* .111 E-02* -1.08* .485** .153 E-02* .116 E-02* -1.08* .474** .274 E-02* .128 E-02* -1.00* .594* .877 E-04* .111 E-02* -1.08* .485** .153 E-02* .116 E-02* -1.08* .473** 275 E-02* .128 E-02* -1.00* .593* .879 E-04* .112 E-02* -1.08* .484** .154 E-02* .116 E-02* -1.08* .472** .275 E-02* .129 E-02* -1.00* * - significant at the 1 % level ** - significant at the 5% level 253 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Interpretation o f Results All regressions demonstrated the expected sign. Independent of which measure of relative military capability is employed, there is a positive relationship between an increase in relative military strength favoring the United States and conflict escalation, a positive relationship between increases in U.S. domestic audience costs and conflict escalation, no statistical relationship between overall public approval and escalation, and a negative relationship between Cold War alliance and escalation. The variable 'lim e” was statistically insignificant in all cases but one where it was significant at the 10% level. Given that the model assumes the role time plays is through its effect in domestic audience cost, this result is not surprising. Conclusion This completes the final part of the empirical puzzle. Both case study and statistical analysis supports the three most important results of the model. First the imposition of sanctions can generate audience costs if the sanctions prove costly and they fail to achieve U.S. objectives. Second, as long as the crisis goes unresolved, audience costs are positively correlated with escalation. Finally, as demonstrated in the case study analysis, sanctions can coerce weak states. It is not a coincidence that of the 14 months the U.S. engaged in military conflict, most were against a strong target. 254 Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission. Our vital interests are not at stake...our national interests did not require the United States o f America to fight its fourth war in Europe. James Baker III on using force to stop the war in Bosnia Dick Chaney and Colin Powell went through the dangers associated with using any military force, even just for delivery o f humanitarian relief. Brent Scowcroft and I stressed that we had already exhausted all diplomatic, political, and economic means. James Baker III on supporting the use o f force to deliver humanitarian aid Chapter 5 : Conflict Management atnd Domestic Audiences in the Balkans Introduction Not since the days o f World War Two has Europe witnessed the death and destruction present in the former Yugoslavia from 1991 to 1996. Concentration camps again popped up across the Balkans incarcerating innocent civilians whose only crime was that they were bom a Muslim. Croat, or Serb. "Ethnic cleansing"— a policy whose sole purpose was to remove "undesirable" minorities from the majority population— became a palatable euphemism in the Western media to describe what the Nuremberg Laws of 1935 were designed to do to the European Jews. And just like pre-World War II Germany, the major world powers refused to quickly end this tragic nightmare through the use of military force. This, of course, is not to suggest that the atrocities committed in the former Yugoslavia were of the same magnitude as those in Nazi Germany. But civilized nations must always be concerned when a high educated, industrialized country quickly evolves from a peace-loving state into one reminiscent of that found in Joseph Conrad's Heart o f Darkness. 255 Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission. The crisis in the Balkans raises a three interesting questions that this chapter attempts to answer in the context of using economic sanctions as a tool of coercion. First, could this crises have been prevented? Or, more specifically, why after Marshall Tito's death in 1980. was the former Yugoslavia destined for a civil war? By trying to imderstand why the war broke-out only 11 years after the dictator's death, I wish to see to what degree sanctions could have contributed to controlling the conflict. Second, could the world have done something different to dampen the flames of civil war? Cries by both the European Union and United Nations for the parties to settle fell on deaf ears for over 5 years. Thus, by examining the role of both sanctions and international organizations in conflict management, the reader is provided an historical accoimt o f international intervention in the former Yugoslavia. Third, as sanctions failed and the accumulated "sunken costs’* increased, did domestic audience costs increase to the point where military intervention transpired? More importantly, can it be shown that at the beginning of the crisis, d < w, but that as domestic audience costs were generated, military force was employed? A close inspection of the case of U.N. intervention in the Balkans suggests that the answer is yes. In answering these questions, the chapter has been divided into 5 sections. Section I provides the historical background of events leading up to the breakup o f Yugoslavia. Note that this section provides a general analysis of why Yugoslavia disintegrated following the death of Tito, as opposed to the simple listing a chronology 256 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. of events.' Section 2 scrutinizes conflict intervention techniques, with an emphasis on the role of United Nation peacekeeping forces. Section 3 examines the contribution U.N. forces made in controlling the conflict. Although this project is primarily designed to analyze the utility of sanctions as a tool of coercion, this section becomes necessary given the historical background to the conflict. Section 4 studies the domestic political situation in the United States to see if rising audience costs were present and how such costs contributed to the decision to escalate the conflicL Finally, 1 conclude the chapter by summarizing its findings. The reader is forewarned that an understanding of conflict management and signaling in the Bosnian case requires a detailed discussion of the role of United Nations peacekeeping efforts. Thus, deviating slightly from the strict “sanctions as signals” model developed in Chapter 3 and empirically supported in Chapter 4, this chapter considers the effects of sanctions in the context of conflict escalation. ' Again, for a comprehensive chronolo^f of events in the Balkans from 1987 to March 1996, please feel free to contact the author. The chronolo^ is 30 pages long, single-spaced. 257 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (D Historical Backsround: Causes o f Yueoslavia’ s Collapse as a Unified State Was Yugoslavia simply not meant to be? Many scholars argued that following the death of Marshal Josip Tito, that the disintegration of Yugoslavia was inevitable. This may be so. but scholarship demands that an explanation be provided beyond the overly simplistic, and perhaps inaccurate, claim that only Tito cold keep Yugoslavia intact. This section's thesis is that Yugoslavia disintegrated because the structure both of the federal government and economy were too weak to counter-act nationalist sentiments within the six Republics and two Autonomous Provinces. ( 1 ) Geography, Demographics, and the "Freezer" Theory Yugoslavia was comprised of 6 Republics and 2 Provinces. The 6 Republics were: Bosnia-Herzegovina. Croatia. Macedonia. Montenegro. Serbia, and Slovenia.* Within the Republic of Serbia there exists two Provinces. Kosovo and Vojvodina. The individual Republics have agreed on a number of occasions to unite in a single Yugoslavia as a result of external pressure and Serbian military might— not because there was a genuine desire to live under a single government. Citizens have thus tended to identify more with a specific etfinic group than the State. At the end of World War I. the Allied powers created an artificial coimtry that they called "Yugoslavia." Yet, despite the naive belief that the Republics were similar because of their geographic proximity, each Republic had its own history and cultural - For simplicity. I will refer to Bosnia-Herzegovina from now on as only "Bosnia." 258 Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission. identity-very little of which was replicated in the history and culture of any of the other Republics. Vojvodina and Slovenia, for example, were governed by the Austrian- Himgarian Empire since the mid 1300s. Croatia was also forced to accede sovereignty to the Austrian-Hungarian Empire in the mid 1500s. Macedonia, Kosovo, and Bosnia, on the other hand, were ruled primarily by the Ottoman Empire keep until World War 1 . Finally, Serbia and Montenegro had achieved relative political independence from the Ottomans in the 1600s. It is therefore not surprising to find many cultural, religious, and ethnic differences among the various people inside Yugoslavia. In fact, with the exception of Bosnia, each Republic and Province has a different ethnic group as the majority. The most important ethnic groups in Bosnia, Croatia, Kosovo, Macedonia, Montenegro. Serbia, Slovenia, and Vojvodina are, respectively, Muslim Bosnian, Roman Catholic Croat, Muslim Albanian, Greek Orthodox Macedonian, Greek Orthodox Montenegrin, Greek Orthodox Serb, Roman Catholic Slovenian, and Roman Catholic Hungarian. Such differences have led to contentious animosities between the different Republics for hundred of years. Past conflicts are too numerous to mention here, but some of the more salient ones are worth noting. During the Middle Ages, Serbs were brutally discriminated against by Croatians given the former’ s ethnicity, political independence, and religious background. In the interwar period (1919-1939), Croatians repeatedly claimed that the Serbian King of Yugoslavia ruled as a dictator, denying Croatians civil and political liberties. Finally, a Nazi-puppet government set up in Croatia during World War II carried out a genocide program against Serbs, Jews, and 259 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Gypsies in both Bosnia and Croatia. In the ensuing war. Serbs and Montenegrins absorbed a disproportional share of the casualties in fighting the Axis powers— roughly 530,000 out of slightly over I million.^ Since World War II. the Croatians and Slovenians have complained that Serbia had too much influence over the political system. Albanians in Kosovo have also been vocal over Serbia's abusive behavior toward the Province. Despite efforts by Tito to increase the ethnic mix of each of the Republics to reduce ethnic tension between the different Republics, a national census taken in 1991 indicated that most of the Republics and Provinces were still relatively homogeneous at the start of the war. * Only in Bosnia, where 43.7% of the population was Muslim. 31.4% Serb, and 17.3% Croat, did one ethnic group comprise no more than 50% of the Republic's population. In Croatia. 78% of the population defined themselves as Croat, and the 12% of the Croatian population that was Serb primarily lived in Krajina. Such demographic conditions, combined with historical animosities, has made Croatia and Bosnia the most fought over territories in the former Yugoslavia. By comparison, Slovenia was allowed to secede relatively peacefully, in part because 87% of the population was Slovenian while only 5% were either Croats or Serbs. ' Denitch. Bogdan. Ethnic Nationalism: The Tragic Death o f Yugoslavia. (Minneapolis, MN: University of Minnesota Press. 1994). * Petrovic. Ruza. "The National Composition of Yugoslavia’ s Population. 1991," Yugoslav 5urv^ 1/1992 (1992). 260 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Given such ethnic, cultural, and historical differences, it is therefore not surprising that Republic leaders have invoked nationalism to justify their actions— whether it be claims of independence from a unified Yugoslavia, or Serbia's contention that 'ethnic cleansing' and territorial expansion is necessary to protect Serbian minorities in other Republics and the Provinces. What is surprising to understand is why Yugoslavia remained intact for so long under such stressful conditions. There were two reasons Yugoslavia remained a viable, cohesive entity prior to Tito's death. First. Tito's absolute control of the political system deterred nationalist leaders from espousing ethnic campaigns. Thus, nationalistic sentiments were kept in a "deep freeze" as Tito would impose whatever military or political force was necessary to repress any signs of dissent. For example, when in 1971 the Croatian Assembly suggested a new "confederation." and possibly even "independence." Tito purged both the entire Croatian League of Communists and Croatian Assembly. Additionally, the League of Communists of Yugoslavia (LCY) prevented nationalist extremists from participating in the political process by dictating to lower levels of government who would run for office. The second reason Yugoslavia remained intact was because Tito allowed for enough decentralization in the political process to placate nationalistic movements. Nationalists feelings, however, could not be kept in the heezer forever. The political structure at the time of Tito's death opened the door for Republic leaders to thaw the spirit of ethnic nationalism locked up for 50 years. 261 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (2) The Failure o f Federalism Tito’ s death had three important affects on the structure of government. First, there was no one individual with either the legal authority or political power to interfere in any State agency or Communist Party organization. Second, Republics felt free to invoke veto power over the national government in a way reminiscent of that in the United States under the Articles of Confederation ( 1776-1789)/ Just as in the colonial America prior to the signing of the U.S. Constitution, each Republic had an autonomous economic policy that rarely took into account national objectives. Finally, Tito's death began a process where the LCY leadership lost its power to dictate who would run for local political office. Instead of government being ruled from the "top down"— whereby LCY leaders picked the candidates— the average voter was given more choices as Republic LC leaders democratized the process. This, in turn, provided an outlet for ethnic nationalism. Figure 5.1 outlines the structure of government. Notice that the Republican League of Communist Part\- (LC) and Republican Assemblies had considerable control over the political process. The Assemblies directly elected members into the lower house of the Federal Assembly. They also elected the President of the Republic, who in turn had power over Federal political institutions through his automatic seat on the Federal Central Committee Presidium. Finally, it was the local LCs who were responsible for nominating members to the Collective Presidency. ■ Critchley, W. Harriet, "The Failure of Federalism in Yugoslavia," IntemationalJoturnal (Summer 1993). 262 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The Republics and Provinces also had considerable power over the political, economic, and legal system within their jurisdictions. Prior to early 1991, Federal Executive Council (PEC) jurisdiction was limited to three areas: foreign affairs, defense, and some international economic policymaking, such as the allocation of foreign exchange earnings. During the economic crises of the 1980s and 1990s, the PEC expanded their authority by declaring an economic emergency, at which time the PEC imposed wage and price controls. Constitutional amendments proposed in May 1990 to increase the powers of the central government were rejected by the Republics. In fact, by July 1991. amendments were adopted that restricted the federal government's powers to only defense and foreign affairs— as federal coordination on economic policy was abandoned. The federal government even became dependent upon the Republics for their revenues. This caused considerable financial problems for the federal government when the Republics did not pay their dues. Slovenia was particularly egregious as they refused to pay 15% of their federal obligations starting in 1989 and withheld from the federal government all custom revenue collected as o f June 1991. Hence, following Tito's death, the Republics invoked provisions of the 1974 Constitution that granted Republics jurisdiction over everything within their borders not specifically delegated to the central government. 263 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.1: Structure of the Yugoslavian National Government, 1974-1991 Preridtnt for Life: Josip Tito (until death in 1980n Goveronent CoimnunistPMty CaHectt7e Presidency nom ute \ |/ Pnme Minister / l \ FedersICenlralCoaimttee Presidium ip p n |< Pnme ves Jim ter Federal Executive Council President of Republics ipprove; fEC Federal Assembly Republic Assemblies Federal Chamber Chamber of Republics And Provinces éiecily elect 51Q Communal Assemblies diiectb'elect 12,000 Local Assemblies People vote for candidates League of Communist ofYugoslaviaCongress Republic CentialCommttee ofCommunist Party Lo cal/Regional League of Communist Party dnectly elect 264 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Given the historical animosities between the Republics, and the decentralized nature of the political system, the only institution Yugoslavians could conceivably really behind was the League of Communists of Yugoslavia (LCY). The LCY was very effective in managing the political system from the "top-down" while Tito was alive. Candidates for office were required to obtain approval from the LCY prior to running. Many times, the LCY allowed only one person on the ballot. Tito's death, combined with a severe economic downturn and growing disenfranchisement among the population over LCY policies, eventually led to the collapse o f the LCY. People blamed the LCY for the country's economic problems-even though 'begger-thy- Republic' policies on the part of the Republics are more likely the culprit. The LCY was also severely criticized by Croatia. Slovenia. Kosovo, and Vojvodina, when in November 1988. the LCY endorsed constitutional amendments that revoked the autonomous status of Kosovo and Vojvodina. The LCY's death was similar to that of many Communist parties across Eastern Europe in the late 1980s. In May 1988. at the 13th Party Congress, the LCY voted in favor of the separation of all state and party powers. At the 14th Party Congress in January 1990. the LCY officially repealed its constitutional right to be the only legal party.® Finally, by the early 1990s. many of the Republics had sponsored multi-party elections. Once public pressure brought the LCY to its demise, there was no longer any unifying force to hold the country together. " It is interesting to note that this occurred around the same time as the 1989 "Revolution" m Eastern Europe. It is likely that the political uprising in bordering states influences the LCY’ s decision to open the political system. 265 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (3) Economie Conditions If the "nationalism freezer" was opened because the political structure was too weak to keep the door closed, then it was the economic system that melted away any good feelings among the ethnic groups. Economic conditions polarized Yugoslavia along ethnic lines for two reasons. First income inequalities between the Republics caused wealthier Republics to complain about subsidizing poorer Republics, while poorer Republics argued that the federal government must do more to redistribute wealth. Second, the economy rapidly deteriorated starting in the late 1970s and eventually collapsed toward the end of the 1980s. These two factors paved the way for nationalist demagogues to win political office based on ethnic identifications and calls for economic change. Industrialization and income distribution between the six Republics and two Provinces were unequal. Slovenia and Croatia were industrialized Republics in a country where 25% of the labor force was employed in agriculture. With only 8% of the population. Slovenia produced 20% of Gross Social Product and 25% of both imports and exports.^ Net personal income in Slovenia in the late 1980s was 40% greater than Croatia, the next closest Republic, and over 50% above the national ' Gross Social Product represents the value of goods and productive services minus capital consumption. "Nonproductive” industries, such as education, are not included in GSP, and hence makes it difficult to compare GSP to Western concepts of Gross Domestic Product. Economic intelligence Unit, Country Profile: Yugoslavia 1989-1990 (London. UK: Economic intelligence Unit Limited, 1990). Economic Intelligence Unit. Country Profile: Slovenia 1992-1993 (London. UK: Economic intelligence Unit Limited, 1993). 266 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. average.* Croatian income, however, was 17% higher than Vojvodina. the third wealthiest region, thus making Croatian standards of living more comparable to Western Europe than the other 4 Republics. On the lower side of the income scale, Kosovo and Macedonia were economic basket-cases. In Kosovo, where unemployment ran as high as 50% in the late 1980s and early 1990s, economic conditions led to comprehensive work stoppages by Kosovo Albanians demanding that Serbia relinquish all economic control of the Province. Table 5.1 provides a more complete picture of the economic disparities between the regions. Table 5.1: The Economv of the Former Yugoslavia'^ Republic Population ( 000) Per Capita Social Product ('000 Dinar) Net Personal Income Unemployment Rate (%) Slovenia 1.871 179 3140 1.7 Croatia 4.437 117 2208 7.7 Vojvodina 1.977 133 1885 15.2 Bosnia 4.155 80 1736 23.9 Serbia* 5.574 94 1846 17.7 Montenegro 604 80 1522 24.5 Kosovo 1.760 36 1418 55.9 Macedonia 1.954 75 1399 27.0 Total 22.334 100 2045 16.2 'Excluding Kosovo and Vojvodina * Magas. Branka. The Destruction o f Yugoslavia: Tracing the Breakup 1980-92 (London. UK: Verso, 1993). ‘ Magas. The Destruction o f Yugoslavia. 267 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Maldistribution of income was only one of the economic problems Yugoslavia faced going into the 1980s. Another was the prospect of economic collapse. The 1979 oil shocks and worldwide recession imposed significant hardships on the Yugoslavian economy. As wages and income fell, the central government borrowed from international capital markets. This accumulation o f debt, however, became onerous as the economy slipped into a recession and the government could not pay the debt back.‘° Austerity measures were put into place begiiming in the mid-1980s— ostensibly to improve the balance of trade so that Yugoslavia would earn hard currencies from which to pay its international debt. Unfortunately, these austerity measures exacerbated rising inflation, declining real wages, and falling GSP. Retail inflation increased from 75.6% in 1985 to 198.7% in 1 988. By the end of 1989. inflation was running at over 2000% annually. .A t the same time that inflation was accelerating, the Federal Executive Committee passed legislation periodically in fiscal years 1987-1990 mandating that nominal wages not change. These programs often failed given both the power of workers to set their own wage and the Republic's control over 70% of the nation’ s money supply. Finally, the austerity measures compounded the problem of falling GSP. According to government reports, real GSP declined 1.2% in 1987, 1.5% in 1988.3.2% in 1989. and 4.2% in 1990." In 1988. Yugoslavia ofïïciaily placed a "moratorium" on repayment. " Economic Intelligence Unit Country Profile: Yugoslavia. 1987-1989 (London, UK: Economic Intelligence Unit Limited. 1990). The 1990 figure was obtained from the World Bank. 268 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (4) The Final Days o f Yugoslavia The synergism of a weak federal structure and poor economic conditions opened the way for nationalistic leaders to exploit ethnic hatred to both win political ofHce and press for Republic independence. Presidents elected to each of the Republics in the late 1980s and early 1990s won on a platform of "reforming" the economy and promoting their respective "ethnic" interest— the latter by either increasing political independence (Slovenia. Croatia, Bosnia, and Macedonia) or protecting Serbs in other Republics and Provinces (Serbia and Montenegro ). Kosovo. Slovenia, and Croatia, were the most aggressive in demanding political and economic independence. Macedonia and Vojvodina also sought an increase in autonomy, but followed the lead of the others. Serbia and Montenegro argued for increasing centralization. Bosnia sought to mediate the inter-Republic disputes by proposing 'compromise' plans. In the end. no compromise was possible and Yugoslavia broke-up into 5 independent States: the Republic of Bosnia-Herzegovina. Republic of Croatia, Republic of Macedonia, Republic of Slovenia, and Federal Yugoslav Republic (Serbia and Montenegro). Events in Kosovo, starting in 1981. had a profound impact on the future of Yugoslavia. With the economy collapsing and the political power of Provinces relatively strong. Albanians marched on the streets demanding government improve the economy and grant Albanians more political freedoms.'- Serbia and the federal government responded by increasing repression throughout the Province. Serbia Chronologically, most workers first demanded economic improvements. It was only later that demonstrators started to call for more political independence, in part to a belief that such independence was necessary to help their economic lot. 269 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. unilaterally reduced the political power of the Kosovo government. The criminal code was altered so as to favor Serbs over Albanians.'^ And other forms of repression increased until martial law was finally imposed in April 1981. These actions by the State had four important effects on the future of Yugoslavia. First, by its implicit acceptance of the use of force, the LCY lost legitimacy among many Yugoslavians— both inside and outside Kosovo. As more repressive tactics were used throughout the 1980s. Yugoslavians began to see the LCY as an "enemy" of the people who were responsible for the destruction of the economy; as opposed to the liberating force which freed Yugoslavia from Nazi control. Second. Serbia sent a clear signal that it would not allow Kosovo to separate given the historical importance Serbia placed on Kosovo. Third, tensions between Serbia and both Slovenia and Croatia increased as the latter two expressed outrage over Serbia's action. The most important effect, however, may have been that Serbian and federal repression gave added incentives for Albanians to take to the streets. 1986-1988 saw a dramatic increase in the number of strikes and demonstrations within Kosovo. While there were also strikes in other parts of the country because of the deteriorating economic and political conditions, those in Kosovo were both the most comprehensive and most feared by Serbian and federal authorities. Tensions increased within Yugoslavia when, in 1987. Slobodan Milosevic, as head of the Serbian LC. forced Ivan Stambolic to resign from the Collective Presidency on the grounds that Stambolic was Maga, The Destruction o f Yugoslavia 54. 270 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. not nationalistic enough to protect Serbian interests in Kosovo. Running on a pro- Serbian platform that promised to increase Serbian control over Kosovo. Milosevic won the Serbian Presidency in May 1989. Just prior to his assumption of office, the Serbian Assembly adopted constitutional changes that gave Serbia full control over the internal affairs of the Provinces. These changes were accepted by the LCY. but harshly condemned by Slovenia and Croatia. Slovenia’ s efforts to separate itself from Serbia proved more successful. Serious calls for Slovenian independence started in May 1988 when a newspaper published a stor>’ indicating that the Yugoslavian National Army (JNA) planned to arrest prominent Slovene political opposition figures. This, combined with Slovenia's ethnic homogeneity, relative economic wealth, concerns over Serbian behavior in Kosovo, deteriorating economic conditions, and the general movement against Communism across Eastern Europe emanating from the 1989 "Revolution." led to the collapse of the Slovenian Communist Party and the establishment o f new political parties in Januaiy 1989. By September 1989. the Slovenian Assembly declared the "right to secede." Serbia imposed economic sanctions against Slovenia in November 1989 to pressure Slovenia to rescind its declarations. Instead of caving into Serbia's demands. Slovenia pressed ahead for independence by invoking its own sanctions against Serbia and withholding 15% of Slovenian contributions to the federal budget. Slovenia’ s independence drive received more momentum in April 1990 multiparty elections as economics and autonomy topped the political campaign. The leading vote-getting party in the 1990 elections was a nationalistic party advocating a 271 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. "new" economic and political relationship with Y u g o slav ia .T h e new Slovenian Assembly immediately went to work on keeping its campaign promises. In July 1990. the Assembly approved a resolution proclaiming fiill sovereignty. This was followed by a short military confrontation between the JNA and Slovenian militiamen— that, paradoxical to Serbian objectives, fueled the flames of Slovenian nationalism. A referendum held in December 1990 asked if Slovenia should secede from Yugoslavia if a new federal constitution granting more rights to the Republics was not adopted within 6 months. Fully 94.6% of the electorate voted in favor of the referendum. In February 1991. the Slovenian Assembly adopted the recommendations of the referendum by an almost unanimous vote. When, by June 1991. it was apparent that no new constitution could be agreed upon by the six Republics. Slovenia announced its intention to secede. On June 25, Slovenia formally declared independence. Because very few Serbs lived in Slovenia, the JNA did not believe it was in the national (Serbian) interests to apply military force. Slovenia's drive toward independence forced Croatia also to move toward declaring its own independence. Nationalist sentiments in Slovenia quickly spread to Croatia as the population became concerned over what would happen to a non- independent Croatian Republic inside a Yugoslavia without its historically close ally, Slovenia. Like Slovenia. Croatia abolished the Croatian League of Commimist Party monopoly in early 1989 and new parties were permitted to register. Croatian multiparty Presidential elections which were held two years earlier resulted in the election of a nationalistic President 272 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. elections in April and May 1990 featured candidates who ran on platforms promising economic renewal and a more pure "Croatian" state. The elections handed power to the Croatian Democratic Union (HDZ). a nationalistic Croatian party headed by Franjo Tudjman. In December 1990, Croatia adopted a new constitution that proclaimed both Croatian sovereignty' and Croatia's right to secede. In May 1991. Croatians voted overwhelmingly in favor of a referendum asking Croatians if they wanted to secede. One month later, confronted with Slovenia's decision to declare independence, Croatia also annoimced its independence on June 25. The 12% o f the Croatian population that were Serbs feared that rising Croatian nationalism, in conjunction with its declaration of independence, posed a threat both to their ethnic identity and person safety. Two month after the May 1990 election, the Croatian parliament passed laws restricting cultural and economic opportunities of Croatian Serbs. Croatia also removed Serbs holding high political office. Serbia, with a highly nationalistic Slobodan Milosevic as President, responded by supplying Croatian Serbs with military and logistical support to help them establish a Serbian National Council (SNC) within Krajina. The first act of the SNC was to call for an August referendum asking Croatian Serbs if they wished to remain tied to Croatia or become its own independent state. The results of the referendum— which comprised Serbs living inside Serbian-majority territories of both Croatia and Bosnia— indicated a strong desire for there to be an independent Serbian state. Six months later, in February 1991, the Serbian autonomous region of Krajina formally declared its independence from Croatia. 273 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The Serbian declaration of independence led to a military confrontation between Croatian militiamen and the JNA, who openly identified with Croatian Serbs, in part because a large percentage of JNA personnel were Serbian Krajinians, and in part because Milosevic and the JNA leadership previously announced it would support an independent Croatian Serb state. Fighting escalated in June 1991 following Croatia's declaration of independence. By September 1991.33% of Croatian land was controlled by the Croatian Serbs and JNA. While the war in Croatia killed an estimated 30.000 people before a cease-fire was agreed upon on Januaiy 2,1992, it was the war in Bosnia-Herzegovina that captured worldwide headlines for its atrocities and crimes against humanity. In December 1990, multipart^' elections brought to power the Party of Democratic Action— a mildly nationalistic party lead by Alija Izetbegovic. The SDA. however, was forced to form a coalition government because it had only won 86 seats in the 240 seat assembly.'- Thus, the SDA could not move Bosnia toward independence in the same manner as Slovenia or Croatia. In fact, of all the Republics, it was Bosnia that attempted to push through a new constitution and avoid disintegration. Once Slovenia and Croatia declared sovereignty in June 1991, however, Bosnia quickly moved toward independence. In October 1991. the Bosnian Assembly declared Bosnian sovereignty. In March 1992,99% of voters approved a referendum that called upon the Bosnian Assembly to declare independence, though only 66% of the ■ The Serbian Democratic Party (SDS) won 72 seats and the Croatian Democratic Party (HDZ) won 44 seats. The next largest party, the LC of Bosnia, won only 19 seats. 274 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. population voted as Bosnian Serbs boycotted the election. It is interesting to note that one reason why Bosnia went forward with the referendum was because the European- sponsored Badinter Commission indicated in late 1991 that the EC would recognize an independent Bosnian state on the condition that a referendum on the issue was held. The Bosnian Assembly legitimized the results of the referendum when on April 6,1992, it declared independence. As in Croatia, many Bosnian Serbs feared living under the laws of an independent Bosnian state given historical animosities between the two groups. Bosnian Serbs, with the help of the JNA, quickly took over large pieces o f Bosnian territory, most of which was owned by Serbs. By the summer of 1992. Bosnian Serbs controlled two-thirds of Bosnia. When the JNA officially withdrew from the Bosnian conflict that summer. 80% of the personnel remained behind as part of a new Bosnian Serb militia. With the help of the Serbian government and the JNA. the Bosnian Serbs began an 'ethnic cleansing' campaign to rid Serbian held territories of all ethnic groups other than Serbs. .Another objective of the campaign was to form a Serb "bridge" between Serbia and the Serbian Republic of Krajina. Not wanting to be left out, and under domestic pressure to ensure the protection of ethnic Croats, Tudjman authorized Croatian forces to enter Bosnia in the spring of 1991. By August. Croatia controlled one-third of Bosnia. Only Sarajevo, the capital of Bosnia, remained in the control of the Muslim-led government. 275 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Death and destruction in Bosnia rapidly escalated after the April 1992 declaration o f independence. Within two months of the war. 12.000 died and 1.2 million were homeless. By late June, the estimated number of dead reached 50,000. Two million were homeless by September.'® The killings ended in November 1995 with the signing of the Dayton Accords, though in the spring of 1999. fighting between Serbs and Kosovos has caused NATO to undertake a military campaign against Serbia. As the next part of this chapter demonstrates, efforts by international organizations to mediate the dispute between Serbs. Croats, and Bosnians produced few results prior to the Dayton .Accords. One explanation for this concerns the irreconcilable differences that the parties held. Land is a very scarce commodity. Only one person can legally "own" a piece of land at any one time. Since this conflict was over the control of land, international organizations had their work cut out for them. A second reason was domestic audiences. The leadership in Bosnia. Croatia, and Serbia came to power by promising to their population that their ethnic group's interests would be advanced. Under pressure domestically to perform, there was incentives to continue the war until a stalemate was reached. New York Times. June 7. 1992. New York Times. June 22. 1992. New York Times, September 30. 1993. 276 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (in Actions Taken and Self-Imposed Costs: Sanctions and U.N. Peacekeeoine Missions While world leaders cried out against the atrocities taking place in Yugoslavia, their actions reaffirmed the old adage that "no one can stop people from fighting if they want to fight." This section provides a brief overview of actions taken by the international community in attempting to find a political settlement in both Croatia and Bosnia; as well as advancing some explanations for why many believe such efforts failed miserably. It is my thesis that diplomatic, economic, and pseudo-military measures by international organizations did not resolve the conflict in the former Yugoslavia for 4 reasons: (1) consent for intervention did not exist, as the parties involved did not want to settle the dispute: (2) United Nations (U.K.) and European Community (EC) objectives were not clearly defined, thus resulting in policies which worked at cross-purposes: (3) the U.K. could not decide if the mission was one of peacekeeping, peace-building, or peace-making, and supply troops with the necessary equipment and rules of engagement to fulfill their mission: and (4) the international cotnmunity lacked the political will to commit the necessary financial and military resources to impose a settlement. The crisis was only resolved when domestic audience costs in the U.S. were so high, that Clinton authorized the continuous airstrikes against Serbian targets until an agreement was reached. Thus, this case study will supply significant empirical support for the domestic audience cost proposition. After suffering considerable "sunken costs.” Clinton elected to use military force, even though his approval ratings were falling and the American public opposed military escalation. 277 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (I) Europe Approaches the Problem Alone~and Fails Slovenia and Croatia's decision to declare independence from the Socialist Federal Republic of Yugoslavia on June 25. 1991 began a flurry of diplomatic activity on the part of European institutions. "Both the EC and Conference on Security and Cooperation in Europe (CSCE) acted quickly after the Yugoslavian army attack on Slovenia, [and] the EC soon established itself as the lead institution."’’ While the CSCE acted as an international forum for discussion— a duty handed over to the United Nations by the Fall— it was the European Community that stepped forward and tried to mediate the conflict. There are three reasons for the Commimity's decision to extend its services to the disputing parties. First, the early 1990s marked a high point in the Community's efforts to forge a cohesive economic and political union as demonstrated by the movement towards "Europe 1992" and Western Europe's role in assisting their Eastern European brethren following the 1989 "Revolution". Second, many Europeans saw the problems inside the former Yugoslavia as a "European" problem that could be best solved by other Europeans. Finally, many EC states, such as Germany, Italy, and France had deep historical interests in Yugoslavia. It is for these reasons that the EC first approached the Yugoslavia crisis alone. It is also for these same three reasons that their efforts failed. Steinberg, James. The Role o f European Institutions in Security After the Cold War. Some Lessons from Yugoslavia (Santa Monica. CA: TTie Rand Corporation, 1992) 46. 278 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. EC mediation began in June 1991 and continued, in one way or another, imtil the Dayton Accords. Initially, it appeared that the EC may have been successful in finding a political settlement. On June 25. the EC launched their first mediation mission by authorizing the Foreign Ministers of Italy. Luxembourg, and the Netherlands to go to Yugoslavia. Their mission resulted in the signing of the Brioni Accord on July 8 by all the warring parties. At first glance, the Brioni Accord did an excellent job of separating the parties since it: { 1 ) required Slovenia and Croatia to suspend their declarations of independence until at least October 8: (2) committed all parties to refrain from unilateral actions such as the use of force; (3) stipulated that the Yugoslavian President would have exclusive authority over the Yugoslavian National Army; (4) acknowledged that Yugoslavian parties alone would decide their future and begin negotiations to that effect no later than the first of August;'* and (5) established an EC monitoring mission to supervise the Slovenian— Serb cease-fire.T h e problem with the Brioni Accord, however, was that it failed to install any mechanism to facilitate future negotiations whereby the EC could mediate the peaceful resolution o f the dispute. The Accord also lacked incentives to encourage the parties to compromise. As the October deadline approached, it became clear that progress was not being made in resolving the dispute. Croatia and Slovenia continued to suggest that they intended to declare independence while Serbia. Croatian Serbs, and the SPRY " Note that the EC sponsored another round of negotiations in August to conform with the Brioni Accord. " Later, this was extended to include monitoring of the Croatian-Serbian cease-fire. Steinberg 1992. 279 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. authorities vigorously opposed such actions. The EC responded by initiating another peace-making effort, this time by creating a peace forum. Lord Carrington, former Secretary-General of the North Atlantic Treaty Organization [NATO], was appointed by the EC as chairman of an EC sponsored Conference on Yugoslavia that opened on September 7. 1991. Negotiating with conflicting groups. Lord Carrington's committee advanced a peace plan in October 1991 that envisaged a free association of sovereign states among the six Republics. The European Community tried to induce the parties to accept Carrington's proposal by announcing that if the parties did not accept the plan by November 11. such parties would be subject to comprehensive sanctions. Despite this threat, Carrington's plan was quickly rejected by all the parties. This fact is not surprising when examined in the context of the situation inside the former Yugoslavia. The domestic political dynamics created an atmosphere where a political settlement of the dispute would be unacceptable among both the leadership and their domestic audiences. By September 1991. Croatian Serbs controlled over 33% of Croatia and hesitated to return the land to the Croatian government. If anything, Croatian Serbs were in a military position to augment their territories at the time Carrington forwarded his proposal. Serbia also expressed reservations about granting more autonomy to Croatia when there was evidence suggesting that the Croatian government would not protect the rights of Serbs. Finally, domestic audiences inside both Croatia and Slovenia pressured officials to declare sovereignty. Thus while the EC plan was a bold initiative to bring peace to this troubled land, the parties wanted to fight-and for may months thereafter, they got their wish. 280 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The EC's negotiating leverage and perceived "neutral" arbitrator role quickly dissipated within a few months of the rejection of the Carrington peace proposal. On the economic front, the EC lifted sanctions against Slovenia. Croatia, Macedonia, and Bosnia in December 1991. while maintaining sanctions on Serbia and Montenegro. This cause consternation among Serbs who were already convinced that the EC supported the claims of Croatia and Slovenia at the expense of Serbia. More specifically, in August 1991. the EC formally stated that Serbia was the party most responsible for the conflict-a finding that goes against all principles of "neutrality." By lifting the sanctions against Croatia and Slovenia, the EC implicitly signaled that they supported the disintegration of the SFRY.‘“ More tensions developed between the EC and Serbia on January 15. 1992 when the EC formally recognized Slovenia and Croatia as independent nation-states. The synergism of these two seminal events, in addition to the perceived pro-independence bias the EC exhibited toward Croatia with respect to its (illegal) arms shipments to Croatia, resulted in Serbia no longer trusting the EC to act as an impartial mediator.-' As the Vice-President of the Presidency of the Federal Republic of Yugoslavia wrote to Secretary-General Boutros-Bourtros Ghali on January 28. 1991: ■ ° It is interesting to note that the United States, which imposed sanctions a few days after the EC, maintained sanctions on ail Republics precisely so as not to be perceived as biased in the conflict Memorandum of the Government of Yugoslavia on the Yugoslav Crisis. Received by Secretary-General Boutris Ghali on January 28. 1992. this memorandum lists violations of Security Council Resolution 713. .A . report by the Comminee of the Security Council for Sanctions confirms the Serbian allegations. 281 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Despite the extensive experience of Lord Carrington as the Conference Chairman and the serious efforts of representatives of the EEC and certain member countries, the Conference on Yugoslavia was soon to show that the good offices which the Yugoslav parties had accepted with hope and trust were turning into a biased attitude towards the secessionist republics while pressure was brought to bear on the others to acquiesce...[alsoJ. early on in the Conference proceedings, the representatives of the Yugoslav federation were denied the right to speak and present their views and, at a later stage, they ceased to participate in the Conference on Yugoslavia (1) and for some time now in spite of repeated requires that they be allowed to participate, no invitations have been addressed to them.” It is interesting to note that EC mediation efforts did not stop because of Serbian concerns over EC neutrality. However, with the involvement o f the United Nations beginning in September 1991, the EC's contribution to resolving the crisis declined considerably. Plenary sessions o f the Conference on Yugoslavia continued throughout the year with little success. O f the 8 cease-fire agreements the EC forged between June 25 and January 1, 1992, not one remained in effect for more than a few weeks. The EC simply could not find a formula that all the parties agreed upon. Nor could it remain effective as the negotiation "ballgame" changed when a new manager entered the clubhouse— the United Nations. As it became evident that EC mediation would fail, and that the United Nations held the mantle of leadership, the EC abandoned its "go-it- alone" approach and established a joint UN-EC Conference on the Former Yugoslavia on August 27,1992. = ibid. 282 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (2) The United Sations Becomes Involved The United Nations initially sat on the sidelines as the warring parties committed egregious violations o f human rights. It was only in late September 1991 that the Securit} Council first addressed the situation in the former Yugoslavia. Since then, the Security' Council and General Assembly have remained actively seized of the matter and rapidly expended resources and manpower devoted to stabilizing the situation. An analysis of the UN’s policies suggests that its failure, like that of the EC. was primarily attributed to the lack of an "impasse" in the fighting.^ The unwillingness of the parties to settle their disputes in a peaceful forum did not. however, prevent the U.N. from implementing plans in areas where there was agreement. This section of the chapter demonstrates that the U.N. was highly successful in settling the Serb-Croat conflict from 1991 to 1995 because both sides sought U.N. assistance: but that in the case of Bosnia. U.N. efforts failed miserably for two reasons— unclear objectives and lack of support within the Serbian commimity of Bosnia. When the crisis broke-out in June 1991. U.N. Secretary-General Perez de Cuellar announced that the United Nations had no plans of becoming involved in mediation efforts for a number of reasons. First, no member state had formally put the question of Yugoslavia on the Security Council agenda. Until this occurred, de Cuellar lacked the legal authority to offer his "good offices" or extend any other form of ^ Sheny. George. "The United Nations, international Conflict, and American Security," Political Science Quarterly 5 ( 1986). Sherry argues that for the U.N. to be effective, there first must be an impasse in the fighting. 283 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. mediation to the disputing parties. Furthermore, de Cuellar indicated that he would not bring the matter before the Security Council under Article 99 of the U.N. Charter on account that he considered the crisis in the former Yugoslavia to be an intrastate affair beyond the scope of Chapter VI or Chapter VII. A second reason the U.N. sidestepped the issue was because the EC already was engaged in mediation efforts and it was thought that U.N. participation would be redundant at best, and potentially damaging to EC efforts at worse. Legally, under Article 33. which stipulates that "parties to the dispute [shall] first of all seek a solution by negotiation, inquire, mediation, conciliation, arbitration, judicial settlement [in] regional agencies or arrangements, or other peaceful means of their own choice." de Cuellar was forced to acknowledge the supremacy of EC mediation efforts and could not become involved without permission of the Security Council.’" * Given that both France and the U.K. are members of both the EC and U.N. Security Council, this implied that until the EC desired U.N. involvement, the Security Council would not mandate that de Cuellar offer his "good offices." A third reason for the lack of U.N. involvement was that Croatia, Slovenia. Serbia, and the SFRY did not actively seek the "good offices" of the Secretary-General, nor did the SFRY want to bring the matter to the attention of the Security Council. But with the progressive deterioration of EC negotiations, and the growing violence taking place inside Croatia, the situation inside the former Yugoslavia was formally placed on the Security Council agenda in mid September 1991. U.N. Charter Article 33. 284 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. By mid September, it became clear that for the EC to achieve its objectives, it was necessary to ask for U.N. assistance for two reasons. First, the disputing parties circumvented economic and military sanctions imposed by the EC by obtaining such goods from non-EC member states. Furthermore, as the EC negotiations progressed, the issue of which EC states would supply peacekeepers became a significant roadblock to an agreement satisfactory to both Croatia and Serbia. The legacy of World War H had left a bad scare in the minds of Yugoslavians who suffered from German occupation and the people within Krajina. Croatia, and Serbia would not accept peacekeepers from countries that had strong historical biases toward one side or another. There was also the question of how involved the EC itself wanted to be. While states such as the United Kingdom and France were willing to send a limited number of peacekeepers, there was very little domestic support in those countries for sending the 20.000 to 40.000 peacekeepers that the EC initially thought would be necessary to separate the Serbs and Croats. The only international organization with the depth of countries to overcome these two problems was the United Nations. As an organization of sovereign member states, the U.N. became the tool from which the EC could implement a settlement. This in mind. Belgium. France, and the United Kingdom formally asked the U.N. Security Council to investigate the situation in Yugoslavia. The Security Council responded by passing three Security Council Resolutions in the next three months. The first one. Security Coimcil Resolution 713, passed on September 25. determined that the "situation [in the former Yugoslavia] constitutes a threat to international peace and stability." and thus warranted a comprehensive aims 285 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. embargo under the terms of Chapter VII.^ The Resolution also mandated that Secretary-General de Cuellar "consult" with the appropriate parties in the conflict, though it did not authorize him to extend his "good offices." As a response to 713, de Cuellar met with Lord Carrington, the President of Ministerial Council of the EC, and the Foreign Minister of the Netherlands. Van der Brock. It is interesting to note that de Cuellar did not meet with anyone from the former Yugoslavia, thus suggesting that while the Secretary-General did have some authority to become involved in the crisis, negotiations would remain the province of the EC. The failure to meet with the political leaders of the Yugoslavian Republics also suggests that de Cuellar still treated the crisis as an inirastatu conflict, outside the domain of the Secretary-General's legal authority without supporting Security Council Resolutions. The first official U.N. contact with leaders in the former Yugoslavia occurred October 11 by former U.S. Secretary of State Cyrus Vance, who had been appointed as de Cuellar’s Special Envoy to Yugoslavia on October 8. After this mission. Vance participated as an observer at the second Conference on Yugoslavia sponsored by the EC on November 18. Resolution 713 was the only Security Council Resolution that imposed sanctions under Chapter VII against all the Republics in the former Yugoslavia. It was also one of the most controversial resolutions because many Bosnians believed that it unfairly discriminated against them in their self-defense campaign against Serbian aggression as allowed for under Article 51 of the U.N. Charter. All other resolutions - Security Council. Resolution 713. U.N. Document S/RES/713 (1991). 286 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. were targeted to either the situation in Croatia. Bosnia, or Macedonia with very little overlap.’® While U.N. Resolutions proved remarkably effective in the Croatian-Serbian dispute initially, they are generally perceived to have failed in Bosnia. (3) The United Nations in the Croatian-Serbian Conflict The other two resolutions adopted in 1991 by the Security Council were designed to facilitate a political settlement in Croatia. A November 23 cease-fire negotiated by Cyrus Vance signed by Croatia and Serbia (the first one signed imder the auspices of the U.N.) opened the door for a possible peacekeeping mission in Croatia. Both parties had earlier agreed that U.N. peacekeepers separating the Croatian militia and Serbian-dominated JNA would enhance the security situation in Croatia. But until a cease-fire held, the United Nations would not send peacekeepers into the region. It would be four months between the time the November 23 cease-fire was signed and peacekeepers arrived in Croatia. Security Council Resolution 721, passed on November 27. noted that although the U.N. was committed to fulfill the request of both Croatia and Serbia to send peacekeepers into Croatia to supervise the November 23 cease-fire and collect heavy weapons, the U.N. would not deploy such troops until the Secretary-General confirmed that the November 23 cease-fire was being abided to by both sides.’’ The resolution thus called upon Croatian and Serbian forces to comply ^ In December 1992, the Security Council authorized UNPROFOR's mandate to be extended to act as a preventive deployment force in Macedonia. ^ In the four days between the November 23 cease-fire agreement and passage of SCR 721, Croatian and Serbian forces continued to escalate the war and showed no signs of pulling back. 287 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. with the November cease-fire, though it did acknowledge that the Croatia had demobilized enough for the JNA to retreat from portions of Croatia. Security Council Resolution 724, adopted on December 5, reaffirmed the UN's intentions of approving a peacekeeping operation developed by Cyrus Vance, but again ordered that until the Secretary-General confirmed that conditions were conducive to a successful operation, the Security Council would authorized deployment of forces under the U.N. Plan.'* A report written by Secretary-General de Cuellar on December 11 included a blistering anack on all parties involved in the Croatian-Serbian conflict.-’ The report indicated that conditions were still not conducive to sending peacekeepers into the region. Following another U.N. sponsored cease-fire agreement on January 2, 1992, newly appointed Secretary-General Boutros-Boutros Ghali wrote another report indicating that despite the 15th. and most recent cease-fire agreement, again conditions were not sufficiently stable for a U.N. peacekeeping force to successfully implement As pan of his fifth mission in the former Yugoslavia in early December. Cyrus Vance successftilly negotiated a "U.N. Plan for a Peacemaking Operation" between Serbia and Croatia. There were three key components in this plan. First, paragraphs 15 to 17 authorized the U.N. to set up so-called "U.N. Protection Areas" where there was a large proportion of Serbs in the population. Within these UNPAs. the U.N.. Croatia, and Serbia agreed that military forces would hand over heavy weapons to U.N. personnel. Second, paragraph 20 specifically assigned the U.N. with the task of "facilitating the return to their homes all person displaced by the recent hostilities who do desire." Finally, paragraph 6. in addition to others, guaranteed free movement of U.N. forces within Croatia. The U.N. plan was formally signed by representatives of Serbia and Croatia in the presence of Cyrus Vance in December. The SFRY formally signed accepted the presence of peace-keepers in what was then still considered "Yugoslavia" on December 3 1. Unfortunately for Vance, he did not get approval of the plan by Milan Babic, head of Croatian Serbs. Babic rejected the plan on January 5. 1992. Eventually, under heavy pressure from the JNA and Serbia, he did sign-on. Security Council Resolution 721. U.N. Document S/RES/72I (1991) Security Council Resolution 724. U.N. Document S/RES/724 (1991) "United Nations Plan tor a Peace-Keeping Operation in Yugoslavia." ^ Secretary-General Report 23280. United Nations Document S/23280 (1991). 288 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. their responsibilities as defined under the U.N. Plan/" To enhance the potential that the January 2 cease-fire agreement would hold, the Security Council passed, on January 8, Resolution 727 authorizing the Secretary-General in paragraph 3 to "send immediately to Yugoslavia a group of up to 50 military liaison officers to promote maintenance of the cease-fire agreement.'"' Based on the Secretary-General's January 5 report, these troops were also assigned with preparing for the arrival of tens of thousands of U.N. peacekeepers into Croatia once a stable cease-fire was achieved. Marrack Goulding, U.N. Under-Secretary General for Peacekeeping operations, went to Yugoslavia to negotiate another cease-fire in late January'. By the end of January, the United Nations faced a number of problems in its efforts in Croatia. First, despite the desire on the part of both the Croatian Government and Croatian Serbs to have peacekeepers placed in Croatia, irregular troops continued to tight. Personnel from the JNA also remained in Croatia and supplied Croatian Serbs with military equipment. It was thus very difficult for the U.N. to send peacekeepers into Croatia despite the fact that all parties ostensibly supported U.N. peacekeepers and the U.N. Plan. Second. EC mediation efforts were working at cross-purposes to U.N. objectives. On January 15. the "Presidency of the European Communities announced that the EC and its member states had decided to recognize Croatia and Slovenia as independent states."" This recognition, as mentioned before, hampered the ability of " Secretary-General Report 23363. United Nations Document S/23363 (1992). " Security Council Resolution 727, U.N. Document S/RES/727 (1992). ^-Kessing 38703. 289 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the EC to act as an impartial mediator in the eyes of the Serbians. Given that the U.N. Plan was premised on the fact that U.N. operations in Croatia would not "prejudice the outcome" of a solution to the crisis, the EC's strong show of support for Croatia made peacekeeping a more difficult assignment. With Croatia now a recognized state at a time when a cease-fire was trying to be put into place. Croatian Serbs had greater incentives to leave the negotiating table and escalate their military campaign. Thus, with a decline in the chances of a political solution (the EC's job), finding a political settlement (the UN's job) was made more difficult."’ ^ Despite the rapid decline in the probability that a long-term solution would be found, the month of February marked a high point in U.N. political settlement efforts in Croatia. The war began to bog-down, with neither the Croatian Serbs nor Croatian militia seizing additional territory. Leadership on all sides wanted to see an end to the conflict. Croatian Serbs achieved many of their political objectives, including the establishment of a functioning "Republic of Krajina." Winter conditions both caused many soldiers to want to take a break from the fighting and made some heavy weapons inoperable. Finally. Croatia had been recognized by much of the international communit} . thus obtaining its primary objective in the conflict. These facts facilitated the process of establishing conditions that would permit peacekeepers to enter the state. " This is not to say that I either agree or disagree with the EC's recognition of Croatia. 290 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Security Council Resolution 740 on February 7 added to the growing list of resolutions designed to facilitate quick implementation of the U.N. Plan. In paragraph 5, the resolution "expresses its concern that the U.N. peacekeeping plan contained in [the December 11 Secretary-General Report] had not yet been fully and unconditionally accepted by all in Yugoslavia." The resolution, however, also noted that Croatian President Trojan had "fully and unconditionally" accepted the U.N. Plan. Thus, based on the expectation that Croatian Serbs would accept the plan, the resolution called for the Secretary-General to expedite' preparations for peacekeeping opérations.^"* Secretary-General Boutros-Boutros Ghali respond to 740 by sending Vance on another mission to the former Yugoslavia. Results from that mission led the Secretary-General to write a report on February 20 indicating that all parties had accepted the plan. The report goes on to encourage the Security Council to implement the U.N. Plan for a twelve month interval." Security Council Resolution 743 adopted the recommendations of the Secretars'- General.^* The resolution established the United Nations Protection Force [UNPROFOR] for a twelve month period with the objective of creating "the conditions of peace and security required for the negotiation of an overall settlement of the Yugoslav crisis."'^ Specifically, according to Boutros-Boutros Ghali's February 20 " Security Council Resolution 740. United Nations Document S/RES/740 (1992). Secretary-General Report 23592. United Nations Document S/23592 ( 1992). Security Council Resolution 743. United Nations Document S/RES/743 (1992). paragraph 5. IBID. 291 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Report, the 10.400 troops. 100 military observers, and 2,840 logistical support staff of UNPROFOR would be assigned to achieve the objectives discussed in the U.N. Plan- including supervising a cease-fire, collecting heavy weapons, helping refugees return to their home, and assisting local governments in administering the UNPAs in areas such as police protection.'® Resolution 743 also clearly stated that the Secretary-General should not deploy such forces until cease-fire conditions held— further indicating that the actions of Croatian Serbs. Serbia, and Croatia spoke louder than their words. On April 2, 1992. the Secretary-General determined conditions were sufficiently stable to warrant deployment of peacekeeping troops under Resolution 743. By May 30, UNPROFOR was fully deployed. On September 30, the Croatian and Serbian governments agreed on the complete withdrawal of the JNA from Croatian Serb territories. It is interesting to note that UNPROFOR was designed primarily along the lines of "second" generation peacekeeping. UNPROFOR was not mandated to rebuild the country, nor was it authorized to deliver humanitarian aid, as eventually what transpired in Bosnia. The primary job of UNPROFOR in Croatia was to monitor and observe the cease-fire agreements. There were very minor changes in UNPROFOR's assignment in Croatia between 1992 and early 1995. One of these changes included whether or not to include so- called "pink-zones" in UNPROFOR's mandate. In a report written by the Secretary- " Contributing countries included Argentina. Bangladesh. Belgium. Canada, Colombia, Czechoslovakia. Denmark. Egypt, Finland, France, Ghana, Ireland. Kenya, Luxembourg, Nepal, Netherlands. New Zealand. Nigeria. Norway, Poland. Portugal. Russia Swedea Switzerland, and the United Kingdom. 292 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. General in June 1992, Boutros-Boutros Ghali asked the Security Council to investigate the situation in heavily Serb populated regions of Croatia outside of the UNPA. The Secretary-General also indicated that it may be appropriate to consider trying to extend UNPROFOR's mandate into these zones. Security Council Resolution 762 considered the Secretary-General's report and decided to set up a joint commission to monitor the situation in the "pink zones." UNPROFOR's mandate was not increased to include these regions, nor did 762 authorize the Secretary-General to offer his "good offices" in achieving this objective.' " Thus, between 1992 and 1995. the situation in Croatia remained relatively stable, in large part because both Croatian Serbs and the Croatian Government consented to the presence of UNPROFOR.^" The Croatian Serbs thought that LT^PROFOR would protect them from the Croatian government: and that in the long- run. a situation similar to Cyrus would arise where the people in the area would come to accept that Croatia would be divided along ethnic lines. The Croatian government, on the other hand, had incentives to support L^NPROFOR's presence for two reasons. First, the U.N. had already committed itself to protecting the "territorial integrity" of the Republic of Croatia, thus giving cause for Croatia to believe that eventually the UNPA territories would somedav be turned over back to Croatia. Second, as the Croatian ” Security Council Resolution 762. United Nations Document S/RES/7621992 (1991). The Croatian Government launched a three day offensive against the 'pink areas' from January 22. 1993 to January 25. 1993. without gaining any territory. This was the only major Croatian offensive prior to the Spring of 1995. Other Croatian offensives began in July 1993 and September 1993. Note that the Croatian Serbs strongly expressed their disappointment that UNPROFOR did not deter these attacks. 293 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Serbs were bener equipped and armed, UNPROFOR’s presence allowed for Croatia to build an army capable of defeating the Croatian Serbs. Note that under the cease-fire agreements. L'NPROFOR would also collect heavy" weapons from the Croatian Serbs. But as early as the Fall of 1993. there was growing evidence that the Croatian Government would soon withdraw its consent to the presence of U.N. peacekeepers. On September 24. 1993. the Croatian government informed the U.N. that unless UNPROFOR "promoted" implementation of Security Council Resolutions promising to protect the territorial integritv' of the Republic of Croatia, that the Croatian government would be forced to ask UNPROFOR to withdraw from Croatia by November 30. In response to Croatia's threat, the Security Council passed Resolution 871. which mandated that UNPROFOR. acting under Chapter VII. continue to perform its peacekeeping functions. Thus, while all previous Security Council Resolutions Justified UNPROFOR's involvement in Croatia under the terms of Chapter VÎ, Securitv' Council Resolution 871 now allowed for UNPROFOR to operate in Croatia without the consent of the home government.^' It would not be until the Spring of 1995 that Croatian opposition to UNPROFOR's presence in Croatia would be strong enough to result in a partial withdrawal o f U.N. peacekeepers from Croatia. Januarv 1995 marked the beginning of the end of UNPROFOR in Croatia. In a letter to the Secretary-General dated January 12. 1995, Croatian President Tudjman wrote that his government would no longer consent to the presence of UNPROFOR ■ “ Security Council Resolution 871. United Nations Document S/RES/871 (1993). The resolution also separated UNPROFOR into three parts for logistical reasons, UNPROFOR (Croatia), UNPROFOR (Bosnia-Herzegovina) and UNPROFOR (Macedonia). 294 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Croatia) after its mandate terminated on March 31 because of UNPROFOR's failure to restore the territorial integrity of the Republic o f Croatia. No longer accepting the UN's promises of protecting the territorial integrity o f Croatia, the Croatian Government made it clear to the U.N. that it would take back by force, if necessary, the UNPAs. Within one week of Croatia's threat, the Security Council issued a statement indicating that UNPROFOR's presence was necessary to prevent an outbreak of hostilities. In anticipation of hostilities, both the Croatian Serbs and Croatian Government undertook an intensive campaign to rearm and retrain troops starting in mid-January. The Contact Group (see below) also offered the newly formed Federal Republic of Yugoslavia ( Serbia and Macedonia) a suspension of economic sanctions if it agreed to remain outside of the conflict in mid-February. Fearing that the Croatian Serbs would lose a war to the relatively better armed Croatian Government, the Federal Republic of Yugoslavia (FRY) turned down the Contact Group's ofler on February 20. The stage had been set for another war to break-out in Croatia. A last attempt to save UNPROFOR from being forced to withdraw from Croatia occurred on March 12 when U.S. Vice-President Al Gore meet with Tudjman in Germany in hopes of persuading the Croatian President to continue to permit UNPROFOR to monitor and observe the cease-fire agreements. Under heavy lobbying, Tudjman agreed to allow U.N. peacekeepers to remain in Croatia, but only if the number of troops were reduced by at least one-third and the peacekeepers moved to areas designated by the Croatian government. Gore agreed to Tudjman's conditions. 295 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The United Nations acted quickly to reaffirm its peacekeeping presence in Croatia. On March 22. the Security Council passed Security Council Resolutions 981. 982, and 983, which, acting under the authority of Chapter VII. reconfigured UNPROFOR into three peacekeeping entities: the United Nations Confidence Restoration Operation in Croatia [UNCRO]; the United Nations Protection Force in Bosnia-Herzegovina [UNPROFOR], and the United Nations Preventive Deployment Force in Macedonia [UNPREDEP]."*' Formal deployment of UNCRO did not occur until April 28 as logistical problems, particularly specifying where U.N. peacekeepers would be stationed, impeded quick conversion of UNPROFOR into UNCRO. However, in keeping with Vice-President Gore's promise to President Tudjman, the number of peacekeepers in Croatia was reduced from slightly less than 16,000 to 8,750 between March 1, 1995 and July 1.1995. No longer adequately protected by U.N. peacekeepers, the Croatian Serbs were exposed to attack. On August 5. 1995. the Croatian government laimched a major offensive against Krajina called "Operation Storm." In a matter of three days, Croatian troops overran 65 U.N. observation posts, killed 3 U.N. peacekeepers, and regained territorial control over Krajina. By the time the fighting stopped, it was too late for the FRY or Bosnian Serbs to come to the aid of the Croatian Serbs. Only Eastern Croatia was still under the control of Croatian Serbs. Security Council Resolution 981. United Nations Document S./RES/98I (1995). Security Council Resolution 982. United Nations Document S/RES/982 (1995). Security Council Resolution 873. United Nations Document S/RES/873 (1995). 296 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Was the U.N. mission in Croatia a success? The above analysis would suggest that, at least between 1991 and 1995. the U.N. achieved exactly the goal which it set out for itself. Peacekeepers were deployed under conditions where their safety was ensured. All parties involved consented to the U.N. Plan. Finally, the military situation in Croatia had. more or less, remained dormant with only a few setbacks. One of the problems of the U.N. mission in Croatia, however, was that while it created conditions for a political settlement (i.e. military stalemate), it may have harmed finding a long-term political solution consistent with the Security Council's commitment to protect the territorial integrity of Croatia. By 1995, the Croatian government believed that the status quo was no longer in its own interest, and began to pressure the U.N. to remove peacekeepers from the region. No longer separating the warring parties. UNCRO failed in preventing conflict from resurrecting its ugly head. LTslPROFOR (Croatia) may have also failed to both collect heavy weapons and return displaced refugees, particularly in Croatia.^^ Overall, then, if one believes that the UN's objective of protecting the territorial integrity of the Republic of Croatia was a just and politically correct goal, then U.N. peacekeepers did an excellent job weakening the military prowess of the Croatian Serbs while giving the Croatian government time to prepare a military campaign against the Croatian Serbs. Alternatively, if one believes that peace, stability, and Croatian Serb autonomy were more important objectives than Croatian territorial integrity, then it is clear that the U.N. failed in their mission— Weiss. Thomas. "U.N. Responses in the Former Yugoslavia." Ethics and International Affairs 8(1994). 297 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. especially given that UNCRO and UNPROFOR (Croatia) were mandated under the terms of Chapter VII where consent of the Croatian government was not necessary for the U.N. to act. (4) The United Nations in the Bosnian Conflict The decisions and actions taken by the United Nations following Bosnia's declaration of independence on March 3 clearly demonstrate the difference between what George Sherry has labeled the "first" and "second" U.N."" In terms of the "first" U.N. function— that of acting as a debating forum where the international community can 'discuss' important issues— the United Nations did a remarkably effective job of condemning the brutal atrocities transpiring all across Bosnia. But in terms of the "second" U.N.-that of implementing and enforcing policies to match what is decided in the "first" U.N.— the evidence suggests that the Securit\' Council. General Assembly, and Secretary-General overestimated the probability their policies would prove effective. The disputing parties in Bosnia (Croats. Serbs, and Muslims) could not agree on a political settlement until the Contact Group intervened in the crisis. Unclear objectives and unrealistic goals by both the EC and U.N. raised expectations beyond capabilities. Finally, as an instrument of member-states. the failure of the international community to devote the military forces necessary to maintain a stable peace condemned the people of Bosnia to war for four vears. ‘ Sheny. George, "The United Nations. International Conflict and American Security." 298 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In this section. I examine the response of the U.N. in Bosnia-Herzegovina. While I keep as close as possible to a chronological examination of U.N. intervention, the comprehensive nature of UNPROFOR’S peace-enforcement and peace-making efforts in this facet of the crisis makes such a chronological study difficult. The U.N. Acts Earlv in the Crisis: Sanctions are Imposed Unlike the case of Croatia, the U.N. acted quickly in Bosnia. One week after Bosnians overwhelmingly voted for independence on February 29. 1992. the Secretary- General sent Cyrus Vance to Bosnia to try to prevent war between Serbia, Bosnian Serbs. Bosnian Croatians. and Bosnian Muslims. The day after Bosnia's declaration of independence on April 6. the Security Council passed Resolution 749 which "appealed to all parties and others in Bosnia-Herzegovina to cooperate with the efforts of the European Community to bring about a cease-fire and a negotiated political settlement.'"** Thus although the U.N. was engaged in finding a political settlement, the U.N. still acknowledged that the EC would be primarily responsible for finding a political solution. The resolution also expressed concern that violence had erupted in Bosnia as a result of Bosnia's declaration of independence. Then on April 9. the Security Coimcil issued a statement ( not a resolution) mandating that the Secretary-General do "everything" in his power to head-off a war in Bosnia so long as it was done in complete cooperation with the EC. Security Council Resolution 749. United Nations Document S/RES/749 (1992). 299 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In response to this requesL Cyrus Vance left for his 7th mission to the former Yugoslavia on April 15. This mission, like that of EC mediation efforts, failed. Marrack Goulding left for Yugoslavia on May 4 hoping to extend UNPROFOR's mandate beyond the geographic regions of Sarajevo and Croatia."** Based on Gouldings' mission, and after consultations with Vance. Secretary-General Boutros Ghali wrote a report indicating that so as to remain consistent with Article 33 and not overextend diplomatic resources, the U.N. should let the EC alone try to solve the problem The report also delineated upon the massive human rights violations taking place, including ethnic cleansing. Finally, the report pointed out that the U.N. did not have the military resources available in Bosnia to do anything beyond monitor the situation in Sarajevo.^^ This report led to passage of Security Council Resolution 752 on May 15 which, in paragraph one. "demanded...that [all] parties...cooperate with the efforts of the European Community to bring about a political solution." to the crisis. The resolution also called for a cease-fire and an end to the massive violations of human rights being reported out of Bosnia."** .A .n interesting component of 752 can be found in paragraph nine where the Security Council "requests the Secretary-General to keep under active review the feasibility of protecting international humanitarian relief programmes...and to ensure safe and secure access to the Sarajevo airport." given the airport's importance in As part of the U.N. Plan. Sarajevo was chosen to be the headquarters of UNPROFOR. "" Secretary-General Report 23900. United Nations Document S/23900. Security Council Resolution 752. United Nations Document S/RES/752 (1992). 300 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. allowing relief planes to deliver humanitarian supplies. This clause led to the Secretary-General to write a report on May 26 that suggested to the Security Coimcil that it consider using UNPROFOR to help deliver aid to war-torn regions of Bosnia. Later, the Security Council would accept the Secretary-General's recommendation. Following the lead of the European Community, the Security Council on May 30 passed Resolution 575 that invoked Chapter VII in mandating comprehensive trade sanctions against Serbia and Montenegro.'” Sanctions included the termination of all trade between Serbia and U.N. member states of all goods and services, an air embargo, freezing of assets, and an oil embargo. Security Council Resolution 760, passed three weeks later, exempted humanitarian goods. .A .t least initially, the bark of sanctions may have been worse than its bite. First, few international corporations continued to trade with Serbia and Montenegro at the outbreak of the war. Thus, the decline in trade as a result of sanctions were marginal at best. Second. Resolution 757 did not include m y enforcement mechanism. Nor did it outlaw "transshipments" through Serbia, until Security Council Resolution 787 was passed in November. For example, if an Italian tirm wanted to export a good to Romania through Serbia. Security Council Resolution 757 said that such behavior would be perfectly legal. Thus, it became very easy for some firms to circumvent sanctions with minor changes in shipment paperwork. There is even evidence to suggest that sanctions proved counterproductive for the first couple of years as the Serbian population "rallied around" Serbian leader Milosevic in response " Security Council Resolution 757. United Nations Document S/RES/757 (1992). Other resolutions concerning the sanctions include Resolution 760.787. 820, 821. and 942. 301 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to the "imperialist" Western countries imposing its will on the Serbian people who were fighting "oppression" in Bosnia and Croatia.^® The greatest effect sanctions had was in solidifying support in the U.N. against Serbian aggression. Hence, for the first couple years of the crisis, the sanctions fulfilled the objective of the "first" U.N.. but not the "second.'"' However, as will be explained below, the economic costs o f sanctions became so great on the Federal Republic of Yugoslavia (Serbia and Montenegro) by the Fall of 1994. that negotiators were able to convince the FRY to cooperate in finding a peaceful solution to the crisis in return for the partial lifting of sanctions. The sanctions also led to the generation of domestic audience costs in the United States. As demonstrated below, the American public became restive as both sanctions and threats of military force tailed to end the crisis. By 1995, with no other options remaining and low public presidential public approval for Clinton's handling of the crisis in Bosnia. Clinton pressed for more militant actions, including NATO bombing throughout much of the Summer and early Fall of 1995. “ The New York Times reported on March 27. 1993 that "criticism of the United Nations sanctions, and the notion of tightening them, is almost universal [in Serbia]., even among those who detest the Milosevic Government. The main effect of cutting Serbia’ s economic ties to the world, opposition leaders say. has been to awaken an old Balkan instinct — pride in a defiant struggle for survival. Vak Draskovic, the leader of the Serbian Renewal Movement, the main opposition party in Belgrade, adds that 'I am absolutely against the sanctions because Mr. Milosevic has benefited greatly from them. They are his big excuse for all the problems here.'" '' The sanctions also represented an extension of EC sanctions which were reimposed on May 27 after they were lifted on April 7 in the hopes that lifting of sanctions would deter Serbia from attacking Bosnia. 302 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNPROFOR’ s Humanitarian Assistance Efforts: the "Third Generation" of Peacekeeping Begins Sanctions were not the only policy the U.N. would eventually adopt under the auspices of Chapter VII. Over time, the United Nations redefined the context of the crisis in Bosnia away from one of "stability" to one of "humanitarian." This allowed for the possibility of the U.N. to send peacekeepers into Bosnia— without the consent of all the parties involved— under the legal framework provided by Chapter VII and circumvent Article 2(7). It is this facet of U.N. involvement that most separates it from that of the peacekeeping operation in Croatia.'* The U.N. was now entering the "third" generation of peacekeeping activity and breaking new ground in conflict management; whereas in Croatia. It still functioned much like a "second" generation peacekeeping force. Yet. UNPROFORs efforts to alleviate forced starvation and violations of other human rights were significantly hampered by Bosnian Serb interference. Mass media attention to Western government failures to stem the gross violations of human rights in the Balkans resulted in considerable pressure for leaders to be seen as “doing something." Evidence of crimes against humanity, as defined in the 1949 Geneva Convention, first came to light in April and May 1992. There was substantial evidence that Bosnian Serbs— and Bosnian Muslims and Bosnian Croats to a lesser extent— were engaged in an “ethnic cleansing" campaign, including the mass murders of civilians based on ethnicity, mass rapes, and a brutal policy of forced expulsion. In the months of Jime and July, the 'ethnic cleansing' campaign widened in a Eventually, the U.N. did get some form of consent from tall the parties, but the credibility of such promises must be questioned given what transpired. 303 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. manner not seen since early Nazi Germany. Concentration camps sprang up across Bosnia where Bosnian Serbs sent innocent Bosnian Muslim civilians to live under harsh conditions. Torture became common practice both inside and outside of the camps. And Serbs continuously shelled the Sarajevo airport so as to prevent humanitarian relief from reaching the citizens of Sarajevo. Such crimes had a profound effect in how the U.N. would deal with the crisis in Bosnia. As citizens in Western Europe and the United States were barraged with photographs and video footage showing the gruesome events taking place in Bosnia, public outcry against these horrors tbrced international organizations to shift their objective away from encouraging a peaceful settlement of the dispute in favor o f an active policy to ensure that humanitarian assistance was provided to all groups suffering as a result of the conflict. In other words, the U.N. would focus less on finding a political settlement (i.e. military stalemate), and alternatively focused on delivering humanitarian relief to Bosnians, the generation of domestic audience costs in the West was mostly responsible for this change. U.N. 'humanitarian intervention' escalated rapidly starting in May 1992. In the months of May and June 1992. the Security Council demanded that Serbs relinquish control of the Sarajevo airport to UNPROFOR in almost all Security Council Resolutions. The Serbs eventually ceded control of the airport on June 6. Unfortunately, because Serbs would fire at U.N. conveys going to the airport, the first planes carrying food, medicine, blankets and other humanitarian goods, did not land until June 30. Once UNPROFOR had secured the airport, the United States, United 304 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Kingdom. France. Canada, and Germany launched 'Sarajevo Airlift Operation.' delivering, by airplane. 153.412 metric tons of humanitarian assistance to Sarajevo and the surrounding areas between July 1993 and April 1995. when the operation was suspended because of Bosnian Serb attacks on both relief planes and convoys. This set back, however, did not deter U.N. efforts. The United Nations continued to work toward getting humanitarian aid delivered. First, in part as a response to Security Coimcil Resolution 752, and in part because of a concern over the dreadful humanitarian crisis, the Secretary-General wrote a report on July 10 encouraging the Security Council take "immediate action" to solve the "grave humanitarian situation" in Bosnia by increasing the mandate of UNPROFOR to include the assistance of delivery of humanitarian aid.^"’ In response to the report, the Securit}' Council passed Resolution 764 on July 13 authorizing "the Secretary-General to deploy immediately additional elements of [UNPROFOR] to ensure the security and functioning of the Sarajevo airport and the delivery of humanitarian assistance.”* ■ * The resolution also requested from the Secretaiv'-General that a report be provided detailing how UNPROFOR could be utilized in delivering humanitarian assistance. In response to this request, the Secretary-General wrote a report on July 21 indicating that conditions were too dangerous to send peacekeepers to help deliver humanitarian aid.” Secretary General Report 24263. United Nations Document S/24263. ’■ * Security Council Resolution 763. United Nations Document S/RES/762. This resolution, like many others, also demanded a stop to the war. ” Secretary General Report 24263. United Nations Document S/24333. 305 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The month of August represented a significant turning point in U.N. involvement in Bosnia for two reasons. The first concerns U.N. humanitarian relief efforts. The second concerns U.N. mediation efforts. Massive violations of rules of engagement as defined under the 1949 Geneva Convention began in earnest in the summer of 1992. The mainstream media in the United States and Western Europe ran a number of front page stories describing the concentration camp system and 'ethnic cleansing' campaign being pursued predominately by Serbs. The reports also described in excruciating detail the mass murders and rapes taking place. With such images firmly implanted in the public's mind, governments pressed for the U.N. to "do something" to both alleviate the suffering in Bosnia and placate Western public outrage— particularly in the United States with an upcoming Presidential elections. Recalling Security Coimcil Resolutions 752 and 764. in addition to Secretary-General Reports suggesting how UNPROFOR could be expanded into Bosnia to help deliver humanitarian aid. the Security Council passed Resolutions 770 and 771 in an attempt to alleviate human suffering.^* Securit} Council Resolutions 770 and 771 are arguable the most important resolutions the Security Council passed in the Bosnian crisis. Resolution 770 begins by "recognizing that the situation in Bosnia-Herzegovina constitutes a threat to the international peace and security and that the provisions of humanitarian assistance in Bosnia-Herzegovina is an important element in the Council's effort to restore Security Council Resolution 770. United Nations Document S/RES/770 (1992). Security Council Resolution 771. United Nations Document S/RES/771 ( 1992). 306 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. international peace and order." and continues by authorizing "all necessary means" for UNPROFOR to assist other organizations, especially the U.N. High Commission for Refugees [UNHCR], in delivering humanitarian aid.^^ This marked the first sign that the U.N. would expand its operation in Bosnia beyond Sarajevo. But unlike Croatia, where troops were sent under Chapter VI. these peacekeeping units would work under the authority of Chapter VII. and without the approval of the parties in dispute. It would not be until September 19 that the three leaders in the Bosnian crisis would consent that the Ll^HCR should have unimpeded access across Bosnia so as to deliver humanitarian assistance. And based on historical events, one must question the Serb's and Croat's sincerit>'. Resolution 771 condemned the atrocities taking place and demanded that the concentration camps be open for inspection by international relief organizations. It was from these two resolutions that the U.N. was able to become involved in peacekeeping operations. As Thomas Weiss wote: "For Bosnia- Herzegovina, governments switched their rhetorical arguments from pacific settlements of disputes [Chapter VI] to enforcement [Chapter VII]. albeit in the context of the protection of personnel" in delivering humanitarian assistance.'* Despite these Security Council Resolutions, it would not be until mid September 1992 before the Security Council authorized the deployment of UNPROFOR personnel outside o f Sarajevo. This was because before the Security Council would expand both the mission and number of personnel in UNPROFOR, the Security Council required the ibid. Weiss, "U .N . Responses in the Former Yugoslavia." 307 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Secretary-General to ensure that a stable situation existed. While a cease-fire was not reached, the Serbs. Croats, and Muslims did reach an agreement in Geneva permitting UNHCR to carry out humanitarian relief efforts, without interference, on September 19. This, combined with the Secretary-General's belief that the dangers of U.N. involvement were relatively small compared to the great benefit intervention would provide, prompted Boutros-Boutros Ghaii to write a report recommending to the Security Coimcil that it authorize him to increase the number o f UNPROFOR in Bosnia by a factor of five.'"^ The report was approved by the Security Coimcil on September 14. in Resolution 776. which, acting under Chapter VI. proposed that UNPROFOR be expanded to enable the delivery of over 5.000 tonnes of humanitarian relief a week.“ Once the Geneva agreement was signed, the Secretary-General moved quickly to increase UNPROFOR's role in Bosnia. 1992 ended poorly despite rising expectations generated from the passage of resolutions 770 and 771. Human rights tragedies continued unabated as UNPROFOR stood by helpless. Serbian militia and Croatian military forces systematically carved up Bosnia to the point where 70% of the land was controlled by Serbs. 30% by Croats, and only a few select cities, such as Sarajevo. Zepa. and Srebrenica, remained in the hands of Muslims. Again the UN's words were proving larger than their actions. It was not until November 19 that UNPROFOR was able to deliver humanitarian aid outside of Sarajevo. And even once consent was given. UNPROFOR still found it difficult to ” Secretary-General Report 24540. United Nations Document S/24540. Security Council Resolution 776. United Nations Document S/RES/776 (1992). 308 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. provide humanitarian assistance. It also became clear by the end of the year that the unenforced economic sanctions imposed by Resolution 757 harmed the poor and may have actually fueled the fighting, as smugglers would use their profits to purchase military arsenals.*’' Finally, there was growing opposition within the General Assembly over Resolution 713 since the arms embargo prevented Bosnian Muslims from defending themselves against Serb and Croat aggression.*- With growing domestic disapproval in Western states and sanctions failing, the U.N. moved to impose additional types of punishment on Bosnian Serbs in an effort to stop the fighting. Examples include Security Council Resolution 780 which set up a "Commission of Experts" to investigate human rights tragedies and Resolution 827 which established an International Tribunal. Note that the effectiveness of U.N. humanitarian efforts is considered below. U.N. Sets U P "Commission of Experts." and "International Tribimal" It appeared as though nothing would deter Bosnian Serbs from fighting. Sanctions did not stop the fighting. Humanitarian efforts, while mitigating the suffering, did not resolve the crisis. Diplomatic pressure could not coerce an agreement. Yet. television images of the atrocities taking place in the "civilized. Many New York Times stories report that militias in Bosnia financed their operations by smuggling goods into the countiy. By charge monopoly prices, the militias raised capital to purchase arms while leaving the average citizen poor. General Assembly resolution 47/242. for example, formally requested that Bosnia be made except from 713. 309 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. European" countrv' of Bosnia fueled public demand that perpetrators of crimes against humanity be punished for their actions. The Securit}' Council set up two institutions to facilitate achieving this objective. The first was the establishment on October 6.1992, of a "Commission of Experts" with a mandate to investigate crimes against humanityBy October 12, the Secretary-General would formally appoint enough members to the committee such that they could begin their investigation. However, a lack of access to key individuals in the former Yugoslavia would greatly reduced the effectiveness of any investigation. It was not until April 11. 1994 that the "Commission of Experts" issued their first report indicting numerous individuals for violating the Geneva Conventions. One year later, on April 1. 1995. the "Commission of Experts" issued their final report. From these reports, the International Tribunal would indict many Serbs, Croats, and (a few) Muslims for crimes against humanity. It came to be the second institution, the International War Tribunal, that would assume responsibility for investigating, indicting, and convicting individuals of violating the Geneva Convention. Calls for an International Tribunal came as early as December 1992 when the General Assembly adopted a resolution encouraging the Security Council to establish a war tribunal that would indict, convict, and sentence perpetrators of gross violations of human rights. Based, on part, on pressure firom both the General Assembly and Western democracies— including firom the United States " Security Council Resolution 780. United Nations Document SC/RES/780 (1992). 310 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. where Bill Clinton had just won an election promising to be more "tough" on Bosnian Serb criminals— that "something" be done about the crisis, the Security Council adopted Resolution 808 on February 2,1993 which asked the Secretary-General "to make proposals on setting up an International Tribunal to prosecute individuals" in violation of international human rights treaties, particularly the Geneva Conventions.*” Close to four months later, upon the recommendation of the Secretar>'-General. the Security Council adopted Security' Council Resolution 827 on May 25 that formally authorized the Secretary-General to establish an International Tribunal. It would be another six months before the International Tribunal would be inaugurated in the Hague on November 11. and over a year and a half before the first indictments were issued, on February 13. 19951 The effectiveness of the International Tribunal has been questioned on three grounds. First, the time lag between when the Tribunal was established and when it first issued indictments is significant. A well functioning organization should never take so much time given the pressing need to indict war criminals. Second, as of April 1996. only one person indicted had been brought to trail. The most significant war criminals remained at large at the time of writing, in May 1999. Thus, some question the credibility of an organization that does not fulfill its ultimate objective after years in operation. Finally, there is considerable evidence to suggest that Bosnian Serbs have been imfairly burdened with most of the indictments. Specifically, of the first 21 Security Council Resolution 808. United Nations Document SC/RES/808 ( 1993). 311 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. indictments made, ail alleged criminals were either Serbs or Bosnian Serbs. Even as late as September 1995, only Bosnian Serbs had been indicted by the International Tribunal, despite strong evidence that Bosnian Croats and Bosnian Muslims also violated the Geneva Conventions. Such "bias" in the indictment process compounded Bosnian Serb fears that the international community could not act as a neutral and "unbias" mediator. I believe, however, that the positive benefits of establishing the International Tribunal far exceed its perceived failures. First, once the necessary logistical and legal structures were established, the International Tribunal has been efficient in handing down indictments. Second, the International Tribunal has extended its mandate beyond Bosnia and has sought indictments and convictions for persons alleged to be violating the Geneva Conventions in other conflicts in the world, such as in Rwanda. Thus, even if it "failed" in Bosnia, the International Tribunal appears to be a permanent institution that will help enforce international law for years to come. Third, the mere establishment of a Tribunal was a powerful symbolic signal to the world that international norms must be abided or one faces the prospect of prosecution. Finally, if not for the Tribunal, it is likely that indicted war criminals could have assumed political office following the signing of the Dayton Accords. Overall, then, I find that the establishment of the International Tribimal represents one o f the major success stones of U.N. intervention— though only time will tell if it is able to fulfill its ultimate mandate. 312 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. U.N. Bans Military Flights Bosnian Serbs were in a far superior military position when the war broke-out in April 1992. The JNA had left behind in Bosnia military equipment and trained personnel that could be utilized by the Bosnian Serbs. Additionally. Security Council Resolution 713 greatly inhibited the Bosnians from importing the necessary military equipment to defend themselves, particularly against attack from the Serbian airforce. To help correct for the imbalance between the Bosnian government and Bosnian Serbs in the air. the Security Council passed Resolution 781 on October 9. 1992 that banned all military flights over Bosnia, except for those involved in humanitarian missions.®^ Specifically. UNPROFOR was mandated to monitor air traffic above Bosnia and report on any violations. Despite the fact that the resolution did not include any enforcement provisions, it would not be until March 16. 1993 that a Serbian military aircraft attacked Bosnian targets. Security Council Resolution 781 failed to deter military flights because it lacked any enforcement provisions. Between October 1992 and June 1993, there were 624 observed violations of the flight ban. By June 1994. the total number o f violations had increased to 1160. The most serious military flight violations, however, occurred in 1995 as the Bosnian Serbs began a terror campaign against the U.N. designated 'safe areas' (see below). Significantly increasing it military aerial campaign, the total number of violations rose to 4.847 bv the end of March 1995. It would not be imtil the 1995 air Security Council Resolution 781. United Nations Document SC/RES/781 (1992). 313 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. campaign by Bosnian Serbs, and perceptions by academics and citizens that the Bosnian Serbs were making a mockery of U.N. laws, that the Western powers would enforce provisions of Security Council Resolution 781. Domestic audiences had grown weary of the failure of Western action and wanted something to be done. Appearing incompetent and weak to domestic audiences. Western political leaders had the U.N. approve the use of force to enforce the no-tly zone. U.N. Negotiation EOorts The United Nations came very close, on two occasions, to negotiating a long term political solution to the Bosnian crisis. This, despite the unwillingness of the parties involved in the conflict to reach a political solution; and the unwillingness of many of the Western powers, particularly the United States, to commit the resources and personnel necessary to support an agreement. It was also only the United Nations that was willing to serve in the capacity of mediator. Thus, the fact that the United Nations did not find an agreement acceptable to the Bosnian government, Croatian government. Serbian government, Bosnian Croats, Bosnian Serbs. Croatian Serbs, the United States, Russia, Germany, France, and Great Britain should not lead one to conclude that the U.N. failed in bringing peace to the region. As has long been known to scholars, the United Nations can not force people to live in peace when they want to wage war.“ “ Skjelsbaek. KJell, ’ ’Peaceftil Settlement of Disputes by the United Nations and Other Intergovernmental Bodies," Cooperation and Conflict, Nordic Journal o f International P olitics2\ (1986) 139-154. Ernst Haas. "Regime Decay: Conflict Management and International Organizations, 1945- International Organization 37:2(1983) 189-256. 314 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. There is also little evidence to suggest that diplomacy alone is sufficient to end civil warfare in cases where the disputing parties believe that military means are the most efficient way to achieve their objectives. Instead of criticizing the U.N.. then, one should commend the U.N. for trying to bring peace to a region where the chances for peace were nearly non-existent short of a complete Bosnian Serb victory. Negotiation efforts were initially the domain of the European Community.*’ As in the case of Croatia, the United Nations believed that, consistent with Chapter VIII, the EC should be allowed exclusive jurisdiction over finding a negotiated settlement to the conflict. Security Council Resolution 749, adopted on April 7. "appealed] to Bosnia-Herzegovina and all relevant parties to peacefully settle their dispute by cooperating with the efforts of the European Community.''** Security Council Resolution 752 continued to call upon the parties to negotiate with the EC in good faith. As far as the U.N. was concerned, it did not wish to get itself involved in the negotiation process until all other avenues of diplomacy were exhausted. The EC did not fully reciprocate the division of labor advocated by the United Nations. On July 15. 1992. the EC successfully negotiated the London Accords— a treaty agreed to by all the Bosnian parties which would both impose a comprehensive cease-fire and require that all heavy weapons be turned over to peacekeepers. Yet, during the negotiations, it became clear that neither the disputing parties trusted an Steinberg. James. "The Role of European Institutions in Security After the Cold Wan Some Lessons from Yugoslavia." Manuscript from Rand Corporation. 1992. °* Security Council Resolution 749. United Nations Document SCR/RES/749 (1992). 315 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. exclusively European peacekeeping mission nor were European governments willing to commit the tens of thousands of troops called for in the Accords. To rectify this problem, the EC and Bosnian parties agreed that U.N. peacekeepers should be used to enforce the London Accords. In response, the Secretary-General, in his July 21. 1992 report to the Security Council, encouraged the Security Council to reject the proposed cease-fire agreement, negotiated by the EC. There were four reasons the Secretary-General opposed U.N. intervention under the London Accords. First, the U.N. was not consulted by the EC prior to signing the agreement; second, a stable cease-fire did not exist in Bosnia; third, because the Accord was a regionally negotiated agreement, the Secretary-General believed that the EC should be responsible for sending peacekeepers; and fourth, UNPROFOR was not equipped to handle such a mission at the time.*’ It was the Secretary-General's refusal to agree to the EC brokered London Accord, in conjunction with events in August, that eventually led the EC and U.N. to merge mediation efforts. It was earlier stated that the month of August 1992 represented a turning point in U.N. intervention in Bosnia for two reasons. The first was because of a decision by the U.N. to introduce "third generation" peacekeepers as part of a humanitarian mission. The second reason August 1992 marked a turning point was because it was the month that the EC relinquished its exclusive control over mediating a political solution. In It should be mentioned that there may be a fifth explanation for Boutris-Ghali’ s recommendation. The New York Times reported on July 24. 1992 that "at a closed meeting of the Security Council...Mr. Boutris-Boutris Ghali shook a finger at delegates as he defended his position on Bosnia, saying that the peace-keeping operation cost too much, was organized by the wrong group, and had received undue attention while efforts in Somalia were ignored." 316 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. late August. Lord Carrington resigned his post as chairman o f the Conference of Yugoslavia expressing his frustration over the unwillingness of the parties to negotiate in good faith. Former Foreign Minister of the United Kingdom Darid Owen replaced Carrington and announced the opening of the London Conference on August 26. At the Conference, attended by over 20 states and representatives o f the United Nations, the parties agreed to form a new conference, the Permanent Conference on Yugoslavia, to be co-chaired by Lord Owen and Cyrus Vance, representing the EC and U.N. respectively. The disputing parties also agreed upon 12 principles from which future negotiations would be conducted.™ Thus, the U.N. would now become involved in working toward achieving both a political settlement and solution. 1993 did not fair much better for U.N. negotiators. The war continued with a vengeance, with only periodic cease-fires during the winter months as cold temperatures encouraged soldiers to find warm shelter. Additionally, over 23.000 peacekeepers were finding it increasingly difficult to deliver humanitarian aid to 2.28 million Bosnians as Bosnian Serbs increased their attacks on humanitarian convoys. Y et from out of the despair there did arise a few moments of flickering sunlight as Bosnians temporarily awakened from their nightmare and saw the possibilities o f peace. Two bold and original plans were put forward by the Permanent Conference on Yugoslavia, both of which at one point looked as though they would be accepted by all the disputing parties. ™ For a description o f the 12 principles, please refer to the Chronology in the Appendix. 317 Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission. The first plan was the Vance-0 wen plan, formally introduced at the International Conference on Yugoslavia in January 1993. After consultations with the governments of Bosnia, Croatia. Serbia, the "Serbian Republic of Bosnia-Herzegovina." and Bosnian Croats, Cyrus Vance and Lord Owen put forward the first UN-EC peace proposal. This plan had four elements. First. U.N. peacekeepers would monitor the human rights situation throughout the region. Second, a boundary commission would be established. Third, interim agreements, such as a cease-fire, would be implemented. Finally, Bosnia would be split into 10 autonomous regions divided along ethnic lines— 1 Serb, 3 Muslim, and 6 mixed, including Sarajevo. While all the parties agreed to the first three points, it was the fourth that caused the plan's failure. The Bosnian Serb Assembly voted down the plan on May 15, given that Bosnian Serbs still held a significant military edge over the Muslims and considered the Vance-0 wen plan as providing insufficient territorial rewards for the Serbs. The Muslim government in Sarajevo rejected the plan because it could not accept carving the state into so many different regions. Some also argue that the Muslims held out because of information emanating from Washington DC that President Clinton was considering using American military force to resolve the conflict. In fairness to Vance and Owen, though, it should be pointed out that just prior to the May vote, the three most important parties in the conflict— the Bosnian government. Bosnian Serbs, and Bosnian Croats-did indicate that they would agree to such a proposal. Even Bosnian Serb President Dr. Radovan Karadzic had signed the plan, on May 2, 1995, five month after it was proposed and two months after the Bosnian government and Bosnian Croats agreed to the plan. 318 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. But even had the Vance-0 wen plan been accepted by all three parities, there was still the question of U.N. commitment. On March 3, the Security Council approved the Vance-0 wen plan, but did not necessarily agree to provide the estimated 60,000 to 70,000 troops that the plan envisioned. The United States, in particular, would not guarantee that U.S. troops would participate. In fact. President Clinton had formally denoimced the plan as violating the 'territorial integrity' of Bosnia-Herzegovina. As France and England were not prepared to assume the responsibility of implementing Vance-0 wen alone, prospects for a peaceful settlement were slim. Without Western military support, it became clear that finding any resolution would be difficult. The second peace plan put forward was the so-called Owen-Stoltenberg plan, on July 30. under the auspices of the Permanent Conference on Yugoslavia.^' This plan was very similar to the Vance-0 wen plan with one major exception. Instead of dividing Bosnia into 10 autonomous regions, the Owen-Stoltenberg plan envisioned establishing three autonomous Republics (representing the three ethnic groups) in a newly formed Union of Republics, where each Republic would choose its own name, establish its own government, and control all legal matters within its jurisdiction. There was even some indication that if the Republics so chose, they could secede from the Union and form ties with other states in the region. The cities of Sarajevo and Mostar would be administered by the U.N. and EC respectively. Another important * ’ Upon Cyrus Vance's resignation on June 14. Secretary-General Boutris-Boutris Ghaii appointed Norwegian Foreign Minister Thorvald Stoitenberg as both co-chaurmen of the Permanent Conference on Yugoslavia and Personal Envoy to the Secretary-General. Yogushi Akaski, the Special Envoy to Cambodia, would replace Stoitenberg in the latter role in the early part o f 1994. 319 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. stipulation in the pian was that the Muslim Republic would comprise no less than 30% of Bosnian territory. The three parties initially indicated that they would agree to the plan on July 30. But soon thereafter, the Muslim government argued that the plan rewarded Serbian aggression and would, in the long-run, legally break-up a U.N. member state.^ In addition. Western governments hesitated committing the estimated tens of thousands of troops in Bosnia both for domestic political reasons and because some officials believed that the UJ'l. was stretching itself too thin given all the other peacekeeping activities worldwide, such as in Cambodia and Namibia. Early estimates suggested that Owen-Stoltenberg would cost $4 billion annually and require 40,000 to 50.000 peacekeepers. Unfortunately, the plan was abondoned in late August. U.N. mediation efforts collapsed following the rejection of the Owen- Stoltenberg plan. While the "Permanent" Conference on Yugoslavia did meet periodically, the summer of 1993 marked the last time any major U.N. agreement was reached. .An examination of the evidence, though, suggests that the U.N. can not be blamed for a failure to find a negotiated settlement. U.N. humanitarian efforts, on the other hand, continued to provide necessary humanitarian aid to hundreds o f thousands o f innocent civilians caught up in the tragic war. ^ Bosnia was admitted into the U.N. on May 22, 1992. 320 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. U.N. Sets UP 'Safe Areas' During the Vance-0 wen negotiations, warfare broke-out across Bosnia. One month prior to Bosnian Croats and the Bosnian government signing the Vance-0 wen plan, the Bosnian Serbs launched a massive military assault on the Northeastern Bosnian city of Srebrenica. UNPROFOR. monitoring the situation, did not act to defend Srebrenica— as to do so would have been outside UNPROFOR's mandate and threaten its image as a 'neutral' intervening power. Additionally, Bosnian Serb air campaigns against the city furthered the perceptions of the Western public that Security Council Resolutions were not protecting civilians from brutal attacks. The Security Council had three possible responses to the military attacks on Srebrenica. First, it could let the Bosnian Serbs conquer Srebrenica, thus increasing the probability that the Bosnian government would surrender to Bosnian Serb demands. This strategy, not surprisingly, was perceived as both immoral and politically unfeasible, even if it may reestablish enough stability to the region to allow for effective U.N. peacekeeping operations. Second, the Security Council could order UNPROFOR to defend the city with military support. This strategy, supported by the Bosnian government, was rejected out-of-hand by the U.N. establishment, particularly the Secretary-General, given that doing so would undermine any perception of'neutrality* that is essential for UNPROFOR to successfully conduct its humanitarian missions. The strategy the Security Council elected to follow can be foimd in Security Council Resolution 819, passed on April 17, 1993. This resolution demanded that Bosnian Serbs stop attacking Srebrenica and withdraw their heavy weapons from within 321 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. range of the city. The resolution furthermore "demand[ed] that all parties and others concerned treat Srebrenica and its surroundings as a safe area which should be free from any armed attack or any other hostile act" But what did it mean to be a "safe area?" There is little evidence that the Security Coimcil thought through the implications of using such language. On the one hand, designating Srebrenica as a 'safe area' subtly conveyed to the Bosnian Serbs that attacks on Srebrenica would be unacceptable. Language such as 'safe area' also gave the impression to Western domestic audiences that the U.N. was doing "something" to protect the innocent people of Bosnia. On the other hand, the Security Council Resolution did not have any enforcement mechanism. The resolution did not threaten retaliation. Nor did it authorize UNPROFOR to use force to repel further attacks. The resolution only increased UNPROFOR persoimel around Srebrenica and mandated that they continue to both "monitor" the situation and help deliver humanitarian aid. But it terms of immediate results, the resolution— and Russian pressure— convinced the Serbs to stop their "ethnic cleansing" campaign in Srebrenica. In the long-run. however, the failure to protect the ‘’safe-areas” generated domestic audience costs, providing Western leaders with incentives to escalate again. Srebrenica would not be the only 'safe area' in Bosnia. Security Council Resolution 824. passed on May 6, 1993, designated Sarajevo, Gorazde. Tuzla, Zepa, Bihac, and the surrounding areas o f these cities as U.N. 'safe areas.' Maglaj, recommended by the Secretary-General to be issued 'safe area' status in early 1994, was the only city recommended by the Secretary-General rejected by the Security Council. 322 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. It did not take long before Bosnian Serbs recognized that the Security Council did not authorize UNPROFOR to actually protect the designated 'safe areas.' Mocking the Security Council Resolutions 819 and 824. the Bosnian Serbs continued to shell the major cities still under nominal control of the Bosnian government. "Ethnic cleansing " also continued much as had it did before 819 and 824. Efforts to add "teeth" to the 'safe area' resolutions contributed marginally at best to alleviating the humanitarian crisis. Security Council Resolution 836. passed on June 4, 1993. extended the mandate of UNPROFOR To deter attacks against the safe areas, to monitor the cease-fire, to promote the withdrawal of military or paramilitary units other than those of the Government of the Republic of Bosnia and Herzegovina and to occupy some key points on the ground, in addition to participating in the delivery of humanitarian relief to the population as provided for in resolution 776 (1992)...by any means necessary.^ Yet despite the authorization to use force to protect the 'safe areas.' UNPROFOR never took advantage of this opportunity out of a fear that such an action would harm the UN's credibility as a neutral and unbias mediator to the conflict. The attacks on the 'safe areas' continued to represent a slap in the face of Security Council power and credibilit) . Changing Political Situation By the end of 1993. Western powers expressed deep concern that the U.N. lacked the military and political capability to end the Bosnian crisis. Specifically, there was growing skepticism that the Secretary-General would not use the military force authorized in Resolution 836 to protect civilians from Bosnian Serb attacks. The ^ Security Council Resolution 836. United Nations Document S/RES/836 (1993). 323 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Serbian rejection of Owen-Stoltenberg also suggested to Western powers that the UN- EC Conferences could not mediate the dispute. To correct for these failures. Western powers took two steps to increase the probability of coercing the Bosnian Serbs. The first was a decision by the North Atlantic Treaty Organization [NATO] to threaten airstrikes against Bosnian Serb installations attacking U.N. designated 'safe areas.' At a January 1994 NATO summit. Western leaders 'reaffirmed' their willingness to engage in air campaigns against Bosnian Serb targets, if so requested to by the United Nations Secretary-General. There was also some indications that NATO wanted to attack Serbian positions following a mortar attack on a Sarajevo market on February 5 that was extensively televised by the Western media. Yet. despite these offers, the Secretary-General recommended in a report to the Security Council on February 20 that NATO not be permitted to use force against Bosnian Serbs in retaliation of the Sarajevo mortar attack for two reasons. First, the Russian government had successfully negotiated a stop to the shelling of Sarajevo. Second, the Secretary-General believed that the Bosnian Serbs may respond to the NATO attacks by tfireatening the lives of UNPROFOR personnel— an argument that proved to be correct two months later. NATO threats of retaliation did not deter Bosnian Serbs from shelling U.N. designated 'safe areas.' In late March 1994. the Bosnian Serbs made a strategic decision to shell Gorazde and widen their "ethnic cleansing" campaign. Under pressure to do "something" about the attacks, the Secretary-General reluctantly permitted NATO warplanes to attack Bosnian Serb facilities being used to shell Gorazde. And just as the Secretary-General predicted, Bosnian Serbs responded to the two days of attacks starting 324 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. on April 10 by taking UNPROFOR personnel hostage and threatening to murder them if NATO did not immediately cease their aerial campaign. NATO, under pressure from the U.N. and countries whose troops were being held hostage, agreed to stop its campaign. Bosnian Serbs then released all the hostages. NATO engaged in only a limited number of attacks in 1994 following its disastrous experience in April. Other than one attack in August, one in September, and a few against Croatian Serb units in November, NATO's rhetoric of aerial bombardment did not match its actions. These attacks, like those of April, also resulted in UNPROFOR personnel being taken hostage. The second political decision in 1994 made by Western powers, besides the introduction of NATO reprisal attacks, was a decision to change the main negotiating forum. Concerned that Europeans were not effective in negotiating a political solution, and that the joint EC-UN effort had failed, the United States formed the so-called "Contact Group" in April 1994 with Russia, Germany, England, and France to replace the EC-UN "Permanent" Conference on Yugoslavia as the main negotiating agency. Led by the United States, the Contact Group advanced peace proposals to the three parties (Bosnian Serbs, Bosnian Croats, and the Bosnian government) in Jime 1994 advocating dividing Bosnia into 2 entities. The first would be a Bosniac-Croat Federation and the second would be a Bosnian Serb Republic, called the Republika 325 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Srpska. The former would receive 51% of Bosnian territory and the latter 49%.^‘ * Although the Bosnian Serbs rejected the proposal almost immediately, seventeen months later, they would agree to the American negotiated treaty. A third significant political change that occurred in 1994 which influenced U.N. intervention in the region was a decision by the Federal Republic of Yugoslavia to partially abandon its ally, the Bosnian Serbs, in hopes of both bringing and end to the war and lifting of economic sanctions. As inflation and growing unemployment eroded standards of living in the FRY, Serbian President Milosevic sought an agreement with the Contact Group that would exchange a FRY embargo on Bosnian Serbs for the lifting of the U.N. sanctions against his country. The FRY would also agree to no longer supply arms to Bosnian Serbs, though it did not indicate that the FRY would attempt to reclaim JNA weapons it had previously provided to the Bosnian Serbs. An agreement was reached. Following a decision to cut-off all trade with Bosnian Serbs in the fall of 1994, the Security Council rewarded the FRY for its actions on September 23 by suspending for 100 days sanctions specified in SCR 757 and 820 pertaining to culture. On May 10, the governments of Croatia and Bosnia signed agreements in Washington D.C. that resulted in the formation of the Bosniac-Croat Federation. Although Bosnia and Croatia had been at war for over two years, both parties agreed to fbrge an alliance as a response to the threats posed by Croatian and Bosnian Serbs. The Federation proved to be successful as Croatian and Bosnian forces worked together in attacking Bosnian Serb positions. 326 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. sports, and air travel on the condition that the FRY provide access to U.N. personnel to monitor the FRY-Bosnian Serb sanction.” Satisfied that the FRY was abiding by the sanction, future Security Council Resolutions would extend the suspension.” Note that this was not the first time the Serbian President facilitated U.N. efforts— a fact often overlooked by researchers and newspaper reporters. It was Milosevic who was the leading advocate of the Vance-Owen plan. In fact. Milosevic supported the treaty so much that he went to the Bosnian Serbian assembly meeting considering the treaty in May 1993 and strongly pressed the assembly to accept the agreement. Milosevic also strongly supported the Owen-Stoltenberg plan. But it was only in the Fall of 1994 that Milosevic formally demonstrated his willingness to reduce ties to the Bosnian Serbs by closing his border. A final important political change in 1994 was the signing of a cease-fire agreement negotiated by former U.S. President Jimmy Carter under the auspices of the U.N. on December 23 that committed the factions in Bosnia to stop fighting for at least four months. However, as the next part of this chapter shows, once the four months expired, the conflict would escalate. " Security Council Resolution 943. United Nations Document S/RES/943 (1994). Security Council Resolution 970. United Nations Document S/RES/970 (1995). Security Council Resolution 988. United Nations Document S/RES/988 (1995). Security Council Resolution 1003. United Nations Document S/RES/1003 (1995) provided for complete relief of sanctions as specified in SCR 743. 327 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. What should be made of these political changes with respect to the effectiveness of UNPROFOR and U.N. intervention? First, and most importantly, these events demonstrate that the U.N. does not act in isolation. The United Nations really should be referred to as the United Nations Organization because it operates at the whim of the member states, particularly those on the Security Council. Once the United States elected to shift the negotiation forum to a U.S. led mission, the United Nations had no choice but to accept the U.S. position. Furthermore, despite strong reservations against NATO airstrikes. the Secretary-General was pressured to allow a number of attacks, even though he believed (correctly) that such missions threatened UNPROFOR personnel on the ground. Note that by mid-1995. the Secretary-General's office was forced to give up all control over decisions to launch NATO airstrikes (see below). Finally, as evidenced by Milosevic's efforts, the United Nations can only be as effective as the disputing parties want it to be effective. If the ultra-nationalist Serbian President could not convince the Bosnian Serbs to adopt U.N. sponsored peace agreements, in conjunction with NATO and Contact Group pressure, there were only limited opportunities for the U.N. UNPROFOR Can Not Protect Themselves: the 1995 Bosnian War UNPROFOR had always been exposed to attack by Bosnian Serb, Croat, or Muslim groups given how lightly armed the peacekeeper were. Such lack of a self- defense capability was not an accident. The Secretary-General made a decision that, to keep UNPROFOR as 'neutral' in the conflict as possible, UNPROFOR's rules of 328 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. engagement would limit UNPROFOR to defending themselves only after being directly fired upon— and even then, there were logistical and structural limitations on how far UNPROFOR could go to defend themselves. Finally, with respect to UNPROFOR's mandate of delivering humanitarian aid to needy residence, UNPROFOR was not permitted to apply force to deter Bosnian Serbs from stopping aid convoys. This exposed position of UNPROFOR had already become apparent by April 1994 following UNPROFOR hostage taking by Bosnian Serbs. In 1995, such exposure proved deadly both to UNPROFOR personnel and the U.N. mission in Bosnia. The four month cease-fire negotiated by former President Carter concluded in late April 1995. Almost immediately thereafter, the Bosnian Serbs launched a number of mortar attacks on designated 'safe areas,' particularly Sarajevo. Bosnian Serb forces also fired missiles at U.N. aircraft delivering humanitarian goods as part of the 'Sarajev o Airlift Operation.' forcing the suspension of the airlift for a number of weeks. In response, the Secretary-General's office again succumbed to pressure from NATO and allowed for NATO airstrikes. On May 25. NATO warplanes began a massive campaign to strike numerous Bosnian Serb positions near Sarajevo. Recognizing that it was no match for NATO air power, the Bosnian Serbs attacked UNPROFOR personnel and took over 400 peacekeepers hostage. Bosnian Serbs then took the hundreds of U.N. peacekeepers they had and used them as "human shields" against NATO attacks. Faced with the 329 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. prospect of dropping bombs on UNPROFOR (many of whom were citizens of England. France, and the Ukraine) NATO called off their aerial campaign and began to negotiate for the release of the hostages. Western reporters and political pundits were now claiming that the Western powers were powerless in Bosnia. UNPROFOR lacked the necessary military, logistical, and legal means to coerce the Bosnian Serbs into a treaty— must less protect themselves from small bands of Bosnian Serb militiamen. NATO warplanes were limited in their capabilities as airstrikes posed significant threats to UNPROFOR hostages. Humanitarian relief convoys were also no longer getting through to the innocent Bosnian citizens UNPROFOR was suppose to protect. Concerned that the U.N. mission was failing caused a number of countries to reconsider their participation in UNPROFOR, particularly both as public outrage over UNPROFOR hostage taking increased and the number of UNPROFOR troops killed in action grew to over 150. Britain and France issued statement that they may withdrawal their contingencies unless the U.N. did more to protect peacekeepers. A number of smaller powers, such as Bangladesh, also became concerned that their citizens would be killed in their job as peacekeepers since not enough was being done to protect UNPROFOR. In the words of David Owen, who had resigned his post in the "Permanent" Conference on Yugoslavia in June 1995, unless something dramatic was 330 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. done to protect UNPROFOR personnel, then "if there [was] not a peace settlement by the autumn.-.then the U.N. forces...[would] be forced to leave," out of concern for UNPROFOR lives.” The "dramatic" action taken by the United Nations Security Council to protect UNPROFOR was the establishment of a "Rapid Reaction Force" of 12,500 troops on June 16.” As mandated in Security Council Resolution 998, the so-called Rapid Reaction Force would "rapidly" attack forces threatening UNPROFOR personnel. The "Rapid Reaction Force," however, never had an opportunity to demonstrate their effectiveness as the war would come to an end before the Rapid Reaction Force was trained and in strategic position to implement their mandate. The War Comes to an End Without the protection of the Rapid Reaction Force, UNPROFOR in Bosnia was at the mercy of Bosnian Serbs during the escalation of the war in the summer of 1995. Bosnian Serb troops launched a massive military campaign against the 'safe areas' of Srebrenica and Zepa in July. Specifically, after overrunning numerous U.N. posts and taking 30 Dutch peacekeepers hostage on July 6, Bosnian Serbs attacked Srebrenica with a vengeance; believing that if the 'safe areas' fell to Bosnian Serb control, then the Bosnian government and Western powers would negotiate a peace agreement favorable ^ Los Angeles TY m es, June 12, 1995. ^ Security Council Resolution 998. United Nations Document S/RES/998 (1995). 331 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to Bosnian Serbs. When Srebrenica fell on July 12, the Bosnian Serbs engaged in a massive "ethnic cleansing" campaign, driving out the entire city of Bosnian Muslims and perhaps executing up to four thousand Muslim men. Once Srebrenica fell, the Bosnian Serbs turned their attention to the 'safe area' of Zepa. On July 14, Bosnian Serbs began to shell Zepa and threatened to kill any individual who did not immediately leave the city. The Bosnian Serbs also captured hundreds of U.N. peacekeepers and warned that any attacks by either NATO or the non- operational Rapid Reaction Force would result in the death of UNPROFOR personnel. ' Recognizing that UNPROFOR both would not. and legally could not, assist in the defense of Zepa, civilians of the city surrounded U.N. posts nearby and demanded that UNPROFOR troops hand over their weapons to civilians willing to fight against the attacking Bosnian Serbs. Many of the UNPROFOR personnel did turn over their weapons. When Zepa fell to the Bosnian Serbs on July 19, "ethnic cleansing" again spread its ugly head as Muslims were forced to flee the city, men were executed, and women were raped. A report by the Secretary-General noted that between April 6. 1992 and July 31.1995. Bosnian Serbs had killed, expelled, or imprisoned 90% of the 1.7 million non-Serbs who lived in Serb controlled territory. Bosnian Serb attacks on the 'safe areas' ended in August 1995 as the Croatian offensive resulted in the defeat of the Croatian Serbs. Instead of fighting to gain more territory, Bosnian Serbs were now hard at work absorbing 200,000 Serbs fleeing Croatia ^ It is interesting to note that Bosnians in Zepa took UNPROFOR hostages of their own threatening to kill them unless NATO came to the defense of the city. 332 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and entering Bosnia. The defeat of the Croatian Serbs also deterred the Bosnian Serbs from extending its campaign to other Bosnian 'safe areas.' Recognizing that their dream of a "Greater Serbia" was withering away, and that Serbs in Croatia, Bosnia, and the FRY were sick and tired of war. Bosnian Serb President Radovan Karadzic agreed to let Serbian President Slobodan Milosevic negotiate on behalf of the Bosnian Serbs in the Contact Group. It was only after the defeat of the Croatian Serbs that Karadzic realized that the Serbian population in both Bosnia, and more importantly in Serbia, would no longer support continued fighting. The time had come for peace. Under the auspices of the United States led Contact Group, the three parties fighting in Bosnia (Serbs. Croats, and the government), signed a 60 day cease-fire agreement on October 6, 1995. Less than two months later, on November 21, the same parties initialed the General Framework Agreement for Peace in Bosnia and Herzegovina and the Annexes, otherwise known as the Dayton Accords, ending the war in Bosnia and allowing for a NATO-led peacekeeping force, called the International Force (IFOR), to replace UNPROFOR. The Dayton Accords were formally signed in Paris on December 14. The war was finally over. 533 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Ill) An Assessment of U.N. Peacekeepine Effort Introduction A famous philosopher once said that those who fail to leam from history are doomed to repeat it. Between 1945 and 1989, there was only one case where a member state of the UJN. split into two.“ Since 1989, however, dozens o f regional conflicts have erupted between different nations who were forced to live in the same state, often not by choice, and usually against their will. The U.N. has little history of mediating these kinds of conflicts. Nor does it have much experience in providing humanitarian assistance in war-like conditions. It is therefore imperative that the U.N. leam from its perceived failure in Yugoslavia so that future generations are not condemned to suffer the same fate as Bosnians, Serbs, Croatians, Rwandans, Kurds, and other nations who have historically been oppressed by their governments, community, and even neighbors. Since the crisis in Yugoslavia is likely to represent the kinds of conflicts and wars the world will witness in the future, I would like to make three recommendations on how the U.N. can improve its effectiveness If such another situation arises. Here 1 consider U.N. intervention in Macedonia, Croatia, and Bosnia in respective order. It is my thesis that the most important lesson to be learned is that, following the selection of an objective, the United Nations must behave consist with those objectives. Furthermore. I also find that credible threats of conflict escalation are required. Such credible threats, however, can only be generated by suffering audience costs for policy failure. ' That case was Pakistan-Bangladesh in 1971. 334 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Macedonia The reader will notice that this chapter did not go into detail on U.N. intervention in the Republic of Macedonia. This does not imply, however, that the U.N. did not make a significant contribution to maintaining peace and stability in this part of the former Yugoslavia. A better explanation for why extensive space was not devoted is because U.N. intervention in Macedonia proved successful. Three factors explain why preventive diplomacy deterred Serbs from invading Macedonia. First, the U.N. was able to place peacekeepers on the ground in Macedonia with the full consent of all the major parties in the former Yugoslavia. Unlike the situation in Croatia and Bosnia, all interested parties welcomed the presence of a U.N. preventive deployment force. Recognizing that going to war in Macedonia was not in his country's best interest. Milosevic intentionally agreed (or at least was not opposed) to set up a situation where he would be constrained from undertaking such action. In the context of the model of Chapter 3, Milosevic “tied his hands.” Second, consistent with the frill acceptance of the presence of U.N. peacekeepers, the differing parties in Macedonia did not want to fight. It is much easier to referee a game when the opposing teams do not commit any penalties. The same idea holds true in conflict management. Finally, UNPROFOR performed in a manner consist with its mandate. Specifically, because UNPROFOR was there primarily to act as a tripwire, UNPROFOR limited its actions in Macedonia to observing and reporting troop movements on both sided of the 335 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Macedonian-Serbian border. When UNPROFOR did mediate between Serbian and Macedonian forces— particularly on those rare occasions that Serbian forces crossed the Macedonian border— UNPROFOR was perceived as neutral and unbias by both sides. The most important lesson to be drawn from U.N. experience in Macedonia is that U.N. preventive-diplomacy can be highly effective when disputing parties agree to such intervention. Croatia If the objective was preventive-dip lomacy in Macedonia, what was the objective of the United Nations in the Republic of Croatia? It seems that two important objectives existed. First. UNPROFOR (later UNCRO) was to promote peace and stability in the region by separating the Croatian Serbs and Croatian militia under the authority of Chapter VI. Once the Croatian government and Serbs fully consented to U.N. intervention. UNPROFOR behaved much like a "second" generation peacekeeping unit. Up until 1995. UNPROFOR proved quite successful in this capacity. The problem from the perspective of the United Nations is that its second objective was inconsistent with its first. Not only did the Security Council want to stop the fighting, but it also expressed its strong desire to return the United Nations Protected Areas (mostly fCrajina Serb held territory) to the full control of the Croatian government. As the Croatian government pressed the United Nations to accomplish this second goal o f the Security Council, it quickly became apparent that UNPROFOR may lose the full consent of the Croatian government. To circumvent this problem, the 336 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Security Council elected to mandate that UNPROFOR operate in Croatia under the terms of Chapter VU, arguing that the prospect of "international war" without a U.N. presence warranted such action. So long as the Croatian government would agree to Croatia being divided into two entities— Croatia proper and the United Nations Protected Areas— then UNPROFOR could still behave like a "second" generation unit. But by 1995. fearing that Croatia would become another Cyprus, Croatian President Tudjman believed that the Security Council must commit itself to its second objective (territorial integrity) rather than its first objective (stability). Tudjman strongly indicated to the United Nations that his country would force UNPROFOR to leave his country unless significant changes were made to UNPROFOR's mandate. Once the U.N. accommodated Tudjman's demands. Croatia attacked the Croatian Serb stronghold of Krajina with the knowledge that UNCRO would not fight on the behalf ot' the Croatian Serbs. It therefore seems that once the military situation stabilized in Croatia after the 1991 cease-fire agreement, the Security Council convinced itself that it could promote both the territorial integrity of Croatia and stability between the parties. When Croatia came to the logical conclusion that both objectives were irreconcilable then the U.N. needed to choose between the objectives.** The U.N. chose to agree to Croatia's demands, thus opening up the door to Croatia's summer 1995 offensive.*^ * ' Of course the United Nations did try to reconcile these two divergent objectives in the Permanent Conference on Yugoslavia, but the reality was that such a reconciliation was impossible. Two different groups, the Croatian Serbs and the Croatian government wanted sovereignty over the same parcel of land. There simply was not room for compromise in the immediate future. As pointed out in the section on U.N. intervention in Bosnia, another explanation for non-UN interference is that the Security Council recognized that a Croatian victory in Krajina would likely led to an end in the war in Bosnia. 337 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Were peacekeepers therefore successful with respect to their actions in Croatia? It depends on whether one believes that territorial integrity, stability, or cultural autonomy was the appropriate objective. Specifically, if the objective was "territorial integrity." then the presence of peacekeepers proved highly effective. This is because at the time the war ended in 1991, the Croatian Serbs were in a much stronger military position than the Croatian government. What UNPROFOR did was give the Croatian government time to rearm and train its army in preparation of "Operation Storm" in August 1995. UNPROFOR also seized much of the heavy weaponry of the Croatian Serbs and enforced a trade sanction against Serbian regions of Croatia. In other words. UNPROFOR proved effective because they acted as an interested party in the dispute by assisting Croatia’ s long-term military prospects while harming that those o f Croatian Serbs. This fact, as mentioned before, was not lost on the Croatian Serb community. If the appropriate objective was stability, then the empirical effectiveness of UNPROFOR drops considerably. While the U.N. presence certainly did decrease the magnitude of the conflict between 1992 and 1995— saving thousands of lives in both Croatia proper and the UNPA— there is no question that the U.N. did not stop Croatia from implementing "Operation Storm." But should the U.N. therefore be considered to have failed? Doing so may be too strong of a statement. As mentioned in the beginning of tfiis chapter, the United Nations Organization can only be effective when the disputing parties want to stop fighting. The fact that Croatia attacked Krajina does not necessary imply that it failed, only that the U.N. could not convince Croatia out of 338 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. doing something that it had waited four years to do. Certainly, had UNPROFOR (UNCRO) not been in Croatia between 1992 and 1995, then it is likely that Croatia would have invaded prior to 1995. Finally, suppose the reader believes that the United Nations chose the wrong objective. Specifically. Croatian Serbs claim that the true objective of the United Nations should have been to protect the cultural and territorial integrity of the Croatian Serb community is accepted. Just as how Croatia, Bosnia, Slovenia, and Macedonia claimed that they had a right to succeed from a culturally dominate and oppressive Serbian Yugoslavia, Croatian Serbs likewise argued that they had a right to succeed from a culturally dominate and oppressive Croatia. Obviously, if one believes that this was the true objective of the United Nations, then UNPROFOR clearly failed. However, a reading of Security Council Resolutions suggests that this was never seriously considered to be an objective of the U.N. in Croatia. There is another important lesson, besides selection of objective, that needs to be considered. Unlike the situation in Bosnia, Croatian Serbs very rarely attacked either Croatia proper or U.N. personnel. Why? While it can not know for certain without reading the minds o f the Croatia Serbian leadership, it seems that one explanation is that UNPROFOR was relatively 'neutral' in the conflict prior to August 1995 and that little would be gained from such behavior.*^ UNPROFOR was also acting more like a "second" generation peacekeeping unit, such that both Croatian Serbs ” A more important explanation, though, is that Croatian Serbs already controlled an "ethnically cleansed" territory and that further attacks would not serve Serbian interests. The same did not hold true in Bosnia. 339 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and the Croatian government believed that UNPROFOR was working to the relative benefit of themselves and the determinant of the other. Again, although the Croatian Serb community believed that UNPROFOR was bias in favor of Croatia rather than Serbian interests, Croatian Serbs also believed that without peacekeepers separating the parties, that they would be exposed to attacks by the Croatian government. Croatian Serb hopes that Croatia would become another Cyprus ended with "Operation Storm." Does this suggest that the U.N. should stick to "second" generation peacekeeping activities and not implement "third" generation peacekeeping operations . ’ Not necessarily. However, the experience in Croatia does demonstrate the importance of carefully selecting an objective and then properly employing the correct instruments to achieve that objective. So long as one accepts the "territorial integrity" objective of the Security Council, and that territorial integrity and stability were irreconcilable, then the U.N. proved highly effective in this part of the crisis. Bosnia A review of the events suggest that there are five reasons why most scholars believe that the United Nations failed in the former Yugoslavia. First, and most importantly, the parties in the Bosnia dispute did not want to resolve their differences within the framework of U.N. peacemaking institutions, such as the Conference on Yugoslavia. All three groups thought their actions were just. No one group was willing to trade territory for peace. The crimes against humanity committed also made it emotionally difficult for the parties to come to an agreement. Until there was an 340 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. impasse in the fighting, it was unlikely that any international organization would successfiilly find either a political settlement or political solution. Second, the Western governments were not committed to using military means to force a resolution of the conflict. As Major-General John A. Maclnness. a former commander of U.N. forces in Bosnia, has point out. it would take a minimum of 600,000 to 800,000 troops to impose a political settlement (i.e. military stalemate) if there was no impasse in the fighting. Seeing how the permanent members of the U.N. were reluctant to allocate 40.000 to 50,000 additional troops to support a peace treaty, it was not surprising that they objected to 600,000. The major powers of the world also constrained the prospects of peace by not committing themselves to U.N. sponsored peace agreements, such as the Vance-0 wen plan, prior to successfully negotiating the treaties. Robert Putnman's work, nicely represents the problem the U.N. mediators faced with respect to the Western powers.*” As an organization of member states, U.N. mediators had to target two different constituencies at two levels of negotiation. At the first level of negotiation. U.N. mediators needed to find an agreement that satisfied the demands of Bosnian Serbs. Bosnian Croats, and the Bosnian government. At the second level of negotiations, U.N. mediators needed to obtain approval of any treaty negotiated at the "first" level from Western powers, as it was their troops that would enforcement any agreement. Domestic audiences, however, were not willing to support troops. Finding an agreement that satisfied all the parties involved proved impossible “ Putnam, "Diplomacy and Domestic Politics: the Logic of Two Level Games." 341 Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission. for the United Nations. At the first level of negotiation, it was extremely difficult to find agreements that all three parties accepted unconditionally. The Vance-Owen and Owen-StoItenberg proposals almost accomplished this, but they fell short as Bosnian Serbs believed that they could hold out for a better deal. Additionally, even when agreements could be found at the first level of negotiation. Western powers would sometimes balk at the U.N. negotiated treaties. The United States was particularly egregious in this matter as the Clinton Administration indicated that it would not support the Vance-Owen plan. This type of negotiating arrangement harmed prospects for peace in the region as the United Nations could not credibly guarantee the Yugoslavian parties that agreements signed under U.N. auspices would be fully implemented. It was only when the United States unilaterally decided to form a new negotiating forum, the Contact Group, could mediators increase the probability that the Western powers would accept a treaty agreed upon at the "first" level of negotiations. Three other explanations for why the U.N. "failed" in are also commonly mentioned. One is that UNPROFOR's mission was hampered by the Security Council passing resolutions without enforcement mechanisms. Resolution 757 imposed sanctions without any enforcement provisions. The no-fly zone was not enforced. Resolutions granting "safe" status to six prominent cities in Bosnia were not enforced. UNPROFOR was not equipped to defend themselves. Finally, UNPROFOR's mandate was to deliver humanitarian assistance in a war zone-a task almost impossible to achieve even in more stable conditions, as was demonstrated in Somalia. Under such conditions, the "instrument" that the United Nations chose to accomplish its goal was 342 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. too weak. Stronger mechanisms were required. Another reason for the perceived failure was because there was not one leading voice in the mediation efforts. As negotiations dragged on, it seemed as though everyone wanted to play the role of "neutral" mediator. First it was the CSCE, then the EC, then the Russians, then the U.N., then a joint EC-UN committee, then the Russians again, then President Clinton sent Secretary of State Warren Christopher to 'persuade' European allies to support airstrikes. then the Contact Group, then former President Carter, and finally the Contact Group reached an agreement with all the parties after months of airstrikes. Inside the former Yugoslavia, this was even more problematic. There was no one voice for Serbs living in Bosnia. Sometimes Dr. Radovan Karadzic spoke on their behalf, other times it was Serbian President Milosevic who negotiated at international meetings. Sometimes the Bosnian Serb Assembly agreed with the leadership, other times it did not. It was extremely difficult to get followers to support the leadership on may different occasions, such as supporting peace plans, when the leadership was not given the authority to speak on behalf of the population. The final reason many perceived that the U.N. 'failed' in Bosnia was because Western media would report primarily on areas where the U.N. fell short of its promises rather than report on areas where the U.N. succeeded. The media also greatly increased public expectations of what the U.N. could accomplish beyond the reality of the situation. For example, when the U.N. designated Sarajevo, Srebrenica, Zepa and other cities as 'safe areas,' and these cities were attacked by Bosnian Serbs, the Western media portrayed this as an absolute failure on the part of the United Nations to accomplish 343 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. their mission. Little attention was paid to the fact that without greater military intervention— a strategy opposed by the public, particularly in the United States-there was little UNPROFOR could do to actually make these cities 'safe' from attack. The media reporting was thus consistent with the domestic audience cost proposition. As the failure o f U.S. policy became more apparent, and Western governments increased the importance of the issue by suffering greater self-imposed costs, public demands that "something be done" to end the crisis prompted political leaders to assume more aggressive actions. There were two additional obstacles facing UNPROFOR. First, intense media coverage made it extremely difficult for the Security Council to commit itself to a single, consistent objective. It seems that the Security Council had three objectives in Bosnia: "territorial integrity," "peace and stability," and "humanitarian missions," such as delivering humanitarian aid and investigating war criminals. These three objectives, however, were not compatible. Unlike the situation in Croatia, demarcation lines dividing Serbs, Croats, and Bosnians were not clear cut. This made it extremely difficult for the United Nations to act as a "second" generation peacekeeping force, as (primarily) the Bosnian Serbs would commit 'ethnic cleansing' campaigns in areas with marginal Serbian populations. Additionally, at different points in the crisis, these objectives came into conflict, such as when the Western powers would not allow U.N. negotiators to commit themselves to negotiating an agreement that would allow Bosnian Serbs to succeed from Bosnia either de facto or de jure. Another example where these objectives may have been internally inconsistent was when humanitarian missions 344 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. encouraged the Bosnian civilian population to continue to fight the Serbs as opposed to surrender. In other words, by providing aid and comfort to the Bosnians, the U.N. may have inadvertently exacerbated the war as the Bosnian Serbs needed to apply greater force to achieve their objectives. Note, however, that given the atrocities taking place, most of the Western public and Bosnian population were willing to tradeoff the increase in instability for the humanitarian assistance. The alternative to U.N. humanitarian intervention and peacekeeping forces would have been to allow the Bosnian Serbs to run-over Bosnian territory and force supporters of the Bosnian government, particularly Muslims, to live under the oppression of Bosnian Serbs. This fact may be the most important lesson of the war. There may be times when the United Nations must be willing to accept instability as the price for freedom and long-term stability. Failure on the part of the U.N. to recognize this fact may explain why Security Council Resolution 713, the arms embargo, was not revoked with respect to Bosnia. This leads to the second obstacle facing UNPROFOR in Bosnia. The decision to act as a "third" generation peacekeeping force, where the U.N. would assist in the distribution of humanitarian aid, monitor and "protect" designated 'safe areas,' aggressively investigate war criminals, impose harsh economic sanctions, and enforce military flight bans inherently made it impossible for the United Nations to act as 'neutral' arbitrator to the crisis. Yet the U.N. constructed UNPROFOR in a manner that was designed to maximize its image as a neutral and unbias power rather than maximize the probability that it could achieve its objectives. All of the empirical evidence suggests that no one in Bosnia saw UNPROFOR as neutral and unbias in the conflict. 345 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Specifically, the Bosnian government believed that Security Council Resolution 713 biased the war in favor of the Serbs given the Serbs greater military capability prior to the outbreak of the war. The government also argued that by not protecting the 'safe areas' with military force, and not allowing Bosnians to adequately defend themselves. UNPROFOR was only helping the Bosnian Serbs. Conversely. Bosnian Serbs were convinced that UNPROFOR and the U.N. were bias against their goal of building a "Greater Serbia." Bosnian Serbs point to the fact that until 1996, all individuals indicted by the International Tribunal were Serbs— despite evidence that other groups also committed atrocities. Additionally, while Serbs were hit with economic sanctions for their intervention in Bosnia, the same did not happen to Croatia or Croatian Serbs. It is often forgotten by most pundits that Bosnian Serbs were not the only group to attack Bosnia. Bosnian Croats also set up their own relatively autonomous region inside Bosnia and committed their own ethnic cleansing campaign against Muslims and Serbs— though at considerably less extreme levels. Third, as with Croatian Serbs. Bosnian Serbs believed that they had a legitimate right to secede from Bosnia that was going unrecognized by the United Nations. Finally, Bosnian Serbs were convinced that UNPROFOR humanitarian efforts, such as monitoring the 'safe areas' and delivering humanitarian aid. benefited Bosnian Muslims in greater proportion than Bosnian Serbs in similar circumstances. Western media and public support were also behind the Bosnian government and their objective of maintaining its "territorial integrity." 346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Despite the fact that UNPROFOR could not be completely neutral in the conflict, the Secretary-General and Security Council still attempted to keep UNPROFOR neutral. Secretary-General Boutros-Boutros Ghali's September 20,1993 report acknowledged that "no party in [Bosnia] finds UNPROFOR to be neutral in the conflict;" yet argued that UNPROFOR should still continue to deliver humanitarian aid without applying offensive military use of force to overcome Bosnian Serb obstacles to delivery. Not properly arming UNPROFOR with weapons of offensive capability in the name o f remaining 'neutral' also harmed accomplishing UNPROFOR's humanitarian mission. Thus, another key lesson that can be learned is that selection of "instrument" must be made to be consistent with selection of "objective." Of course this argument is not new, as many scholars have already noted the importance of selecting the correct instrument to accomplish a goal.“ However, if the U.N. is going to move to "third" generation peacekeeping, then it must properly account for the fact that it may not be a welcomed party in the conflict, such that it can reconfigure its peacekeeping force away from a purely neutral force to one more consistent with Security Council objectives. Success of the U.N. in Bosnia So much attention has been paid to the failures of the UJ^. mission in Bosnia that little recognition is given to their successes. I believe that history will someday see U.N. involvement in Bosnia more as a success story than one of failure for four reasons. Baldwin, Economic Stalecrqft. 347 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. First UNPROFOR proved extremely important in reducing human suffering by delivering humanitarian aid. Thus instead of emphasizing the half of the glass that is empty due to obstacles in delivering humanitarian aid, greater attention should be paid to the half o f the glass that is full. 2.28 million Bosnians semi-regularly receive humanitarian assistance through U.N. agencies. Without such assistance, it is quite possible that more people would have died during the war. Second, it was during the Bosnian crisis that the United Nations established an International Tribunal that has regularly worked to indict individuals for crimes against humanity, such as in Rwanda. Additionally, it is because of the Tribunal that war criminals such as Bosnian Serb President Karadzic and Bosnian Serb General Ratko Mladic were not permitted under the terms of the Dayton Accords to hold political office in the new, reconfigured Bosnia. In previous international crisis with massive violations of human rights, no such institution existed that pimished war criminals. Third, relative to what would have likely transpired if the U.N. did not go into Bosnia, Bosnians were far better-off with U.N. presence than without their presence. This is particularly true given that Western powers did not advocate fully arming the Bosnian government such that it could defend itself. More specifically, imagine what would have happened had the U.N. not intervened in the Bosnia crisis under the terms of Chapter VII. Would the international community feel ashamed that it was not doing "something" to punish Bosnian Serbs for violating international norms? Would Serbia no longer feel any constraints not to attack Macedonia? Would more innocent civilians 348 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. been have killed as Bosnian Serbs and Bosnian Croats complete their "ethnic cleansing" campaigns? If not for the presence of UNPROFOR, the answer to all the above questions would be an unequivocal yes. Finally, sanctions did prove in the long-run to be relatively successful in pressuring Serbia to work toward finding a solution to the war rather than add oxygen to the flames of war. Growing costs on the Serbian economy of both the war and the sanctions alienated many Serbians as the war dragged-on. Sanctions also proved to be a useful negotiating tool for the Contact Group. If not for the sanctions and other tools of coercion, domestic audience costs would not have been generated. Without such costs. NATO airstrikes would not have occurred forcing an end to the war. 349 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (IV) The Role o f Domestic Audience Costs in Conflict Escalation Introduction Ultimately, the goal of “conflict management” is not to “manage” a conflict, but to end it. The war in Bosnia stopped only after there was a stalemate in the battlefield and NATO airstrikes relentlessly hit Serbian targets. Numerous threats made by Clinton that he would apply military force went unheard until Clinton was “backed” into a comer, obligated to enforce his threats or be questioned by domestic audiences. In this section of the Chapter, 1 show that there is a strong correlation between domestic audiences and conflict escalation. The Role o f National Interests Many U.S. policy leaders believed that American intervention was not in the U.S. “national interest.” such that d<w. For example, the Bush administration intentionally let the European Community handle the Bosnian conflict due to the belief that it was a “European” rather than an American problem. Secretary of State James Baker makes the U.S. position even more clear when he writes that “unlike the Persian Gulf, our vital interests were not at stake.” Baker continues his description of the Bush administration’s Balkan policy as one of “hands-off,” because the American public would never support U.S. troop involvement in either stopping the conflict or monitoring a peace agreement.** Relative to other international events at the time. Baker, Politics o f Diplomacy 632. 350 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. especially the dissolvement of the Soviet Union after the August 1991 failed coup and events in the Persian Gulf, the Bush administration did not feel U.S. interests were at stake. Nor did the .American public initially believe that U.S. national security was threatened, even though they were aware of the atrocities being committed. In July 1992, 79% of the public reported that they had heard about the fighting in Bosnia and 62% were aware of international relief efforts in early August. About 50% of the public stated that they were following events in Bosnia 'iairly or very closely” in August 1992, with almost no change in the distribution in public awareness through July 1995. The percentage of the American public following events more closely than any other media reported event remained between 3-5% from 1992 to the end of 1994, with a doubling of the percentage during the U.S. airstrikes in May 1994 and peace negotiations in May 1993. The public was also mildly knowledgeable of events, with 20% of the public correctly identifying Serbia as the party conquering Bosnia, with a progressive increase in public awareness over time. 34% also could correctly define “ethnic cleansing” by January 1993 with 40% believing media reports of “ethnic cleansing.” Given the intense television and NYT publicity of the Bosnian conflict, there is clear reason to believe that the public was aware of events in the Balkans. However, this did not translate into the public believing that the U.S. had “national interests” in the Balkans. First, over 50% of the public did not believe U.S. vital interests were at stake in every poll taken from September 1991 to December 1995, with the change in public opinion very slight over time. Second, the public was also 351 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. highly uncertain of whom of the United States should support. A July 1992 poll found that 68% of the public did not know which side they tended to support, with the remaining 32% split equally between Serbs and Bosnians. When asked if they would rather see “Muslims or the Serbians win." in a February 1994 poll. 81% responded “not sure.” 9% “don’t know.” and the rest split 6% to 4% favoring the Muslims. Third, the public did not believe the situation in Bosnia was important relative to other foreign or domestic objectives. Relative to other foreign policy issues. American were not very concerned with events in Bosnia. Americans were more concerned with other international events, such as U.S. troop involvement in Somalia, the transitions taking place in Eastern Europe, peace efforts in the Middle East, and North Korea’s attempt to build an atomic bomb. Only when the U.S. launched airstrikes against Serbian targets or an agreement was about to be reached committing the U.S. to send tens of thousands of troops to Bosnia did more than 20% of the public believe Bosnia was the most important foreign policy issue the country confronted. Relative to all issues the government could address, the public placed very little weight on Bosnia. In no poll between 1992 and 1995 did more than 1% of the public think that Bosnia was one of “the two most important issues for the government to address.” with the exception of in May 1993 and June 1995 following NATO airstrikes. Not even 1% believed it to be “the most important problem facing this country today,” in those same years. Even after the signing o f the Dayton accords, only 4% thought it was the most important issue. It is therefore not surprising that the Bosnian conflict was not an election issue. When asked to choose between 8 voting issues for the November 1992 election, only 352 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1% cited stopping the war in Bosnia. An analysis o f voter issues in polls taken on election day show that less than .5% of the public based their presidential vote on the candidate’s position on the Bosnian conflict. The empirical evidence therefore suggests that neither U.S. government officials nor the public believed that the U.S. had significant national interests at stake in the Balkans when the crisis erupted in 1992. A Desire to ‘Do Something ” This, of course, is not to suggest that benefits were not gained from intervention. Europe had a "vital interest” at stake in containing a regional war: and as discussed throughout the Chapter, international intervention saved the lives of tens of thousands, if not hundreds of thousands. The United States also had "moral” obligations to do something. Thus, while the public did not believe there were national interests at stake, the public still supported U.S. efforts to solve the conflict. An analysis of the data reveals two trends concerning the public’s desire to "do something.” First. 40-50% of the public supported the U.S. doing something about the crisis, and second, support for being actively engaged fell over time. As early as August 1992,43% of Americans believed that the U.S. could not "stand by” while the atrocities were taking place. 47% also disagreed with the statement that "tlie fighting in Bosnia is not America’s problem and we shouldn’t get involved there.” When asked if the U.S. should become more involved in the distribution of humanitarian aid, the response in favor was high. In August 1992,54% o f the public believed that the U.S. 353 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. had a “moral obligation to use military force in Bosnia if there was no other way to get humanitarian aid to civilians and prevent the warring parties there from practicing atrocities.” Between February 1993 and April 1994, 36-41% of the public believed the U.S. "has a responsibility to do something about the fighting between Serbs and Bosnians in what used to be Yugoslavia.” A record high of 37% of the public believed that the United States should also “do more to stop the war in Sarajevo and Bosnia,” in August 1992 while a low of 49% thought the United States has “already done enough.” Over time, however, support for U.S. involvement declined. By August 1994, only 25% thought the U.S. should do more to stop the war, while 65% believed the U.S. had already done enough, with a 12% drop in support falling steadily over time. Another poll shows a similar trend, with a drop of 11% believing the U.S. has a responsibility to act and a 14% point increase in those saying the U.S. does not have such a responsibility. Finally, 55% of the public believed that Bosnia was not the problem of the U.S. in June 1995. an 8% point increase from the same poll question in April 1993. Given the context of the theoretical model, two conclusions can be drawn from this dissussion. First, even though the public did not believe U.S. interests were at stake, they still wanted the U.S. to “do something.” Thus, administration officials would become under increasing pressure to demonstrate their “competency” to domestic audiences. At a meeting with British Prime Minister John Major on May 22, 1992, Baker informed Major that the “Bush [administration] is seeing moimting public concern and criticism over impotent Western inaction in the face of a true humanitarian nightmare...We’ve got to do something about the atrocities there. And we have to get 354 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. humanitarian aid through.” After the meeting. Baker annoimced diplomatic sanctions including a refusal to “recognize Belgrade as the continuation of the Yugoslav state until all its forces were withdrawn from neighboring states and minority rights respected: the permanent withdrawal of [U.S. ambassador Warren] Zimmerman from Belgrade: the closing of two Yugoslav consults: the breaking of contacts with the JNA by pulling our military attaches: and the drawing down of our embassy personnel in Belgrade.”” ’ Thus, the U.S. committment to do “something” about the crisis began and American audience costs would grow over time. Second, while the public became less interested in U.S. involvement, the U.S. government increased its involvement. In conjunction to falling presidential approval ratings for handling of the Bosnian crisis, this supplies strong evidence that as the public finds the existing policy to be failing, presidents attempt to reestablish domestic credibility by escalating a conflict. Public Support fo r if.S. Actions The polling data show that while the public supported the U.S. “doing something,” there was not much agreement on what should be done. Only the imposition of economic sanctions was overwhelming supported. Support for arming the Bosnian or using military force was mixed, with differently worded questions resulting in different outcomes. Baker, Politics o f Diplomacy 645-646. 355 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Support for economic sanctions against Serbia was strong. 75% of those polled in July 1992 agreed that the United States should “enforce strong economic sanctions against the aggressor country until it stops its aggression.” Such support is consistent with the sanction literature’s finding that most sanctions episodes are supported by the public because it is a relatively cheap means to punish a target. But other than economic sanctions, no other policy received strong public approval. Two of the most debated policies during the Bosnian crisis concerned lifting the arms embargo and actively arming the Bosnian Muslims. Support for lifting the embargo was high in 1993. but when told that ending the embargo could escalate the conflict, a majority of the public opposed lifting it. In 1994 and 1995. over 70% of the public also opposed the United States arming the Bosnians. Support for allowing the Bosnians to purchase arms, but disapproval over the United States selling weapons to Bosnia, likely explains why the Clinton administration ignored reports that Iran was illegally selling arms to Bosnians in 1995. Support for the use of military force is difficult to assess. Polling data suggest that there was increasing support for airstrikes overtime, but the presence of U.S. ground troops was strongly opposed. This held true even if ground troops were conditioned on the signing of a peace agreement. Furthermore, the support for ground troops also declined over time. Only the support of using air power to enforce the “no- fly” zone received significant public approval. 356 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. At the start of the conflict, the public was strongly opposed to the use of military force. Over 65% of the public opposed the sending of ground troops to end the civil war in July 1992 while less than 30% favored the application of force. Opposition to using American military force remained high throughout the remainder of the conflict, never dropping below 50% in any poll question, while support never exceeded 35%. ^Tien told that sending troops would prevent the war from spreading, such that it “is better for the U.S. to intervene now and top the problem before it becomes a region- wide conflict,” 35% approved and 54% disapproved using force in May 1993. However, when asked if they would support force under certain conditions, public opinion was more favorable, though still not overwhelming. For example, when replaced with the words “United Nations” instead of “United States,” 20% more Americans favored sending a “[U.N.] force to support the country which is the victim of aggression and to fight and repel the aggressor.” Other polls asking about U.N. military intervention instead of U.S. intervention showed similar increases in support. When told that U.S. troops would be used to help deliver food aid, public support was roughly split equally, but strongly supported when asked if U.S. air power should be used to help in the safe delivery of food. Again, these poll numbers did not change much between 1992 and 1995. The public was also highly supportive of using American troops to help the U.N. withdraw from Bosnia in 1995, even if it entailed U.S. casualties. Support for responding to attacks on U.N. personnel with U.S. airstrikes was also high. However the public did not support airstrikes. Support for using U.S. troops as to assist in the enforcement o f a peace plan was mixed, both by year and by the poll question 357 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. wording. In 1993. during the Vance-Owen and Owen-Stoltenberg negotiations, poll results show both support and opposition to using U.S. troops. Even the same poll question asked over time suggests no firm public opinion on this issue. Interestingly, after the Dayton Accords were signed, there was clear and overwhelming public disapproval of sending U.S. troops, with most polls showing 55-60% public disapproval, and less than 40% approval. Given the “rally-around-the-flag" theories discussed in Chapter 2, finding such low public support suggests that the public had become highly cynical over administration policies and was concerned of another U.S. initiative failing in Bosnia, particularly given the recent attack on U.N. troops. However, when asked if they supported sending American troops as part of an international force where no American soldier would be killed, approval increased to 68%. Asked if they supported force if it cost the U.S. even I life, approval fell to less than 30% and disapproval was 65%. In conclusion, public opinion polls suggest that the public wanted the U.S. to ‘‘do something” about the conflict in Bosnia. As the conflict escalated, including the deployment of U.N. troops, establishment of “safe areas,” and “no-fly zones,” the public became increasingly skeptical of intervention. The public never supported U.S. military involvement during the conflict and increased their opposition over time. However, public opinion still demanded that “something” be done about this issue. After examining President Clinton’s public approval ratings, the empirical validity of the model in Chapter 3 will be clear. 358 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Presidential Approval Ratings Public approval of Presidential actions fell significantly between 1992 and November 1995. when approval rates increased slightly after the signing of the Dayton Accords. Such a finding is not surprising given the events taking place. As the Western world appeared to be impotent, the U.S. public increasingly disapproved the actions o f the Clinton administration— even though the public did not support escalating the conflict. Bush never suffered such costs because he did not make Bosnia a high priority. Clinton did. and suffered the consequences when U.S. actions failed. Public Approval: Overall & in Balkans September 1991 - January 1996 550 ■§40 CL sept 91 feb 92 july dec may 93 cot mar 94 aug jan 95 jun 95 nov Date M 359 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Between early 1993 and the summer of 1994, public approval for Clinton dropped by about 10 points while disapproval increased by 10 points. This occurred in an environment where the public already disapproved of Clinton’s handling of the situation. Poll figures remained relatively constant between 1994 and 1995. Analysis o f Sanctions as a Signal U.S. intervention closely follows the predictions of the theoretical model. Public interest was high enough to require the U.S. to act, but not so high as to support conflict escalation. As efforts by the international community to end the conflict failed, and Clinton’s threats of military force were ignored by Serbia, Clinton suffered domestic audience costs. By the summer of 1995. Clinton had three options: backdown, maintain the status quo. or escalate. Backing down would have entailed significant audience costs and would likely have resulted in a public backlash. Maintaining the status quo would be difficult in light of Croatia’s invasion of Bosnia. Seeing the possibility of ending the conflict, Clinton elected to authorize the use of America airstrikes in the summer o f 1 995. After relentless attacks during August, September, and October, Serbs finally agreed to sign the Dayton Accords. Using such airpower in 1992 would not have been possible. It was only after the generation of domestic audience costs that force became a viable option. 360 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I place immense value in the experiment beingfiilly made, how far an embargo may be an effectual weapon in future. Thomas Jefferson, 1808 Chapter 6: Conclusion and Recommendations Introduction This dissertation explored the relationship between economic sanctions, domestic audiences, and international conflict. I proposed that the imposition of sanctions is correlated with the generation of domestic audience costs. As these costs build over time, political leaders in the sender may choose to escalate the conflict. If properly used, sanctions can signal a willingness to escalate by demonstrating to the target that the executive will suffer domestic audience costs. It is these future costs of escalation that cause a target to backdown. Thus, the credible threat of military force, or at least some measure of conflict escalation, is a necessary condition if sanctions are to achieve difficult foreign policy goals. A sanction by itself is not expected to work. Only in conjunction with other coercive tools can they succeed. The intuition behind the argument is derived from diverse theoretical literatures. Chapter 1 considered the effectiveness of economic sanctions. The mainstream literature finds that economic sanctions are not an effective tool of coercion, and almost never work when the objective is to destabilize a government, end a military campaign, or fundamentally change the human rights 361 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. practices of the target. Such a finding is not surprising given the model developed in Chapter 3. Most target governments will find the value of the commodity in dispute [. r] to be worth more than the cost of sanctions [e(s)\ or even the possibility of conflict escalation. As Yuan Mu noted at the beginning of Chapter 1: "if the government is overthrown, what is the use of [foreign] loans and credits." Furthermore, many targets will rationally expect sanctions to be imposed prior to undertaking an activity deviating from international norms. Thus, sanctions should not be expected to work in isolation. Their usefulness in achieving secondary and tertiary objectives should not, however, be discounted. But for primary goals to be realized, sanctions are only effective if they signal a willingness of resolve. The mechanism by which the signal works is through the generation of audience costs. The audience cost proposition states that if an executive’s foreign policy actions are perceived as failing, the public will punish the executive with lower approval ratings. Regaining domestic support would require conflict escalation, or at least not backing down once the conflict has started. As noted in Chapter 2. while the domestic audience cost literature provides strong theoretical support to the argument, the literature fails in two important respects. First, contrary to Fearon’s findings, we expect mainstream rationalist arguments on the importance of relative military capability to play a role in both the generation of domestic audiences and decision to escalate a conflict. As there can be no greater domestic audience cost suffered than that of losing a war, states should 362 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. not be expected to be “locked in” to war if they expect to lose. Second, attempts by states to bluff are common and must be incorporated into any theoretical model demonstrating how "sunken costs” can signal conflict escalation. The theoretical model developed in Chapter 3 corrects for these deficiencies. 1 constructed a dynamic game with two-sided incomplete information where I found equilibria showing how sanctions can be use as an effective signal of resolve, as well as when that signal fails and a conflict escalates. The most interesting finding is that of Proposition 4 which showed that war is a rational outcome of this game if: (I) the sender values the foreign policy good less than the relative cost of war [w > d], (2) the cost of sanctions is less than the foreign policy good and has a positive probability of working as a signaling tool [s < d], (3) because o f incomplete information, the target suspects the sender is high-cost rather than low-cost, and (4) the audience costs generated from a failed sanction policy plus the value of the foreign policy good is greater than the cost of war [a(s) + d > w]. Thus, war is possible even if no state prefers war! This models makes a contribution, therefore, to both the "sanction and statecraft” literature as well as the conflict literature. Empirical evidence supporting the theoretical model was presented in Chapters 4 and 5. Chapter 4 considered an exhaustive set of sanction episodes between 1970 and 1992 where the United States sought difBcult foreign policy objectives in Angola, Burma, Chile, China, Grenada, Guatemala, Haiti (1987), Haiti (1991), Iran, Iraq, Libya, Nicaragua (1977), Nicaragua (1981), Panama, 363 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Paraguay, Serbia, South Africa, Suriname, Uganda, and Uruguay. Briefcase studies were developed for each episode, taking into consideration media exposure, public opinion polling information, and relative military capability. A comprehensive analysis of these 20 episodes provides strong empirical support to our argument. Once a president elects to make an issue "important," as signaled by the self-imposed cost of sanctions, if the sanction episodes fails, public opinion turns against the president and the president has incentives to escalate the conflict. The cases of Chile, Grenada, Haiti (1991), Iraq, Iran, Libya, Nicaragua (1981), Panama, and Serbia provide the strongest support to the argument. The president is most likely to escalate if the target is "weak,” as predicted by mainstream rationalists. This explains why the United States did not imdertake military escalation against China and South Africa, and only limited military activity against Iran. Conversely, when the self-imposed costs are small, little in the way of audience costs are generated. In these cases, the executive does not have incentives to escalate. U.S. non-escalation against Angola, Burma, Guatemala, Haiti (1987), Nicaragua (1979), Paraguay, Suriname, Uganda, and Uruguay can be imderstood in this context. I then tested our model statistically to provide more rigorous proof. Two statistical analyses were conducted. The first demonstrated a positive correlation between "self-imposed” costs and audience costs. The dependent variable, domestic audience cost, was generated by finding the change in public approval ratings over time in each episode where data were available. I then 364 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. used as independent variables the self-imposed cost of sanctions, media exposure, relative economic capabilities, and time. All variables were found to be statistically significant with the expected sign, independent of which measure of media exposure was employed. A second statistical analysis was then considered— were the dependent variable was conflict level and independent variables were predicted domestic audience cost relative military capability. Cold War alliance, overall presidential approval, and time. Again, independent of which type of media exposure is employed we found all independent variables to be statistically significant with the correct sign with the exception of overall presidential approval. This, however, should not be surprising. As discussed in Chapter 2, a major criticism of diversionary war theory is that overall public approval does not accurately represent public opinion on specific foreign policy adventures. I improve on the existing literature by showing that it is not necessarily overall approval that influences an executive’s decision, but domestic audience costs generated in that specific episode. Finally, time was found statistically insignificant in some cases, but not others. When other statistical analysis was performed dropping time and overall approval ratings, there was no change in the results. Finally Chapter 5 closely examined conflict management in the Balkans firom 1987 to 1995. Deviating slightly from a pure sanctions analysis, I argued that the perceived failure of conflict management techniques increased domestic audience costs in the United States. Once these costs were high, I showed that 365 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. President Clinton rationally chose to escalate the conflict, even though prior to 1994, both the Clinton and Bush administration did not believe U.S. “national interests” were at stake. This chapter also contributes to the literature by demonstrating that U.N. peacekeeping forces can be highly useful. Policy Recommendation: Think Lons-Term The major finding of this project is that sanctions can work when they send a credible signal of conflict escalation. Mixed signals can only result in policy failures. The case of Iran makes this point clear. Despite the demonstration of military force and relatively significant sanctions imposed from November 1979 to March 1980, both President Carter and administration officials publicly stated that the military option was not going to be used. It was only in late March and April 1980 where the threat of military force was made public. However, after the failed rescue mission, the use of force was again dropped as an option. No longer fearful of conflict escalation. Iran would have little incentive to release the hostages until they were in need of resources to fight Iraq and mobilize troops for a potential invasion by the Soviet Union. But even if accompanied with high domestic audience costs, sanctions can still fail. The literature review of Chapter 2 showed that the American public generally opposes the use of military force. Public opinion polls verify the literature’s findings. This can result in the target believing that the United States is high-cost rather than low-cost. Saddam Hussein’s comment that the 366 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. United States does not “have the stomach for war,” as well as Manuel Noriega’s belief that the American public would not tolerate deaths in Panama, caused both targets to believe that the U.S. was isolationist. Only in the case of Nicaragua did the government correctly guess that the U.S. was isolationist, limited in its ability to escalate the conflict by public opposition. This leads us to make the following 6 policy recommendations: /. Make Sure your House is in Order Before you Step Outside Before imposing sanctions, senders must predict how the home audience will respond if sanctions fail. Although doing this is extremely difficult, it is critically important for leaders to make such assessments if predictions are to made about conflict outcomes. If the public is going to “punish” the executive, then the sender must be prepared to escalate. The executive must also generate such costs if a threat of escalation is to be credible. Given the public’s disdain for war, hawkish behavior can not be sustained. Only if domestic audience costs are generated will a target credibly believe there will be conflict escalation. 2. Don’ t Ever Say you are Not Willing to Buy the Farm The explicit rejection of escalation only undermines the effectiveness of sanctions. Given that sanctions alone rarely succeed, targets are most likely to be moved by threats of future costs. The explicit rejection of escalatory action only will make the target suspect that the sender is of type high-cost. 367 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This does not mean, however, that the sender must include escalation as an option. Doing such represents a ’Tying of hands,” that can generate domestic audience costs, backing the executive into a position where he is forced to employ military force to maintain his credibility. A simple comment such as “all options are open,” is sufficient to communicate to the target that escalation is possible, without increasing domestic audience costs. 3. Once you Move, Be Prepared to Stay fo r a While Often sanctions are imposed as a visceral response to a short-term crisis. The public clamors that “something be done” about a problem, without thinking of the long-term consequences. Executives, however, must be prepared to consider the long-term effects of any sanction. Domestic audiences explain why once imposed, sanctions are rarely lifted. Except for the debatable case of Vietnam, the United States has never lifted the type of sanction considered in this project until a primary objective was achieved.' The domestic audience costs o f conceding defeat are high. Thus, if a sanction is going to be imposed, either lift it quickly before domestic audience costs are generated or be prepared to maintain them for the long-term. ‘ But even in the case of Vietnam, President Bush waited until Vietnam had withdrawn from Cambodia. Interestingly, Bush lifted the sanctions after iosmg the 1992 Presidential elections, only after he could not suffer domestic audience costs. 368 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4. Democracies hold Advantages, as well as Disadvantages There are advantages and disadvantages to democracies playing the domestic audience cost signaling game. Democratic states may be better able to generate audience costs, such that the signal is more credible and conflict is avoided. But if the target is unable to correctly read the sender’s domestic political situation, it also exposes the executive in the sender to the possibility of being boxed into a position of ordering military force when, at the initial stages of the conflict, the use of force was not in the national interest. 5. The Neighbors across the Street may be Watching The reader will recall that in Chapter 4, we posed an interesting question: can the imposition of sanctions encourage a third party to use violence against the target? Sanctions may signal to military leaders that a coup would be supported by the U.S., as in the case of Chile and Panama. States considering employing military force against a target may also believe that U.S. sanctions would mean American support. Although we leave this question for future research, an executive must remember that sanctions can signal different things to different interest groups, both at home and abroad. 369 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6. Sometimes Appetizers are Just as Tasty as the Main Course Finally, even if sanctions fail in their primary goal, they can still successfully accomplish secondary and tertiary goals. Often times the value realized is greater than the cost of sanctions. Sanctions may not have toppled Saddam Hussein, but they have slowed his ability to build a nuclear weapon. Sanctions alone could not coerce Iran to release the hostages, but they were highly successful in conjunction with Iraq’s invasion of Iran. 370 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Selective Bibliography Aldrich, John H., John L. Sullivan, and Eugene Borgida, “Foreign Af6irs and Issue Voting: Do Presidential Candidates ‘Waltz before a Blind Audience,” American Political Science Review 83:1 (1989) 123-141. Alejandro, Manelli "The Never-A-Weak-Best-Response Test in Infinte Signaling Games," Journal o f Economic Theory lA'A (1997) 152-173. Alexander, George and William Simons ed.. The Limits o f Coercive Diplomacy 2nd edition (Boulder, CO: Westview Press, 1994). Ascherio, A. e t al., "EflFect of the Gulf War on In6nt and Child Mortality in Iraq," New England Journal o f Medicine 327:13 (September 2, 1992) 931-936. Ausubel, Lawrence and Raymond Deneckere, "A Direct Mechanism Characterization of Sequential Bargaining with One-Sided Incomplete Information," Journal o f Economic Theory 48:1 (1989) 18-46. Baker, James A. Ill, The Politics o f Diplomacy (New York, NY: G. P. Putman's Sons, 1997). Baldwin, David, "Thinking about Threats," Jouma/ o f Cotrflict Resolution 15 (1971) 71-78. Baldwin, David, "The Power of Positive Sanctions," World Politics 24 (1971) 19-38. Baldwin, David, "Inter-nation Influence Revisited," Joama/ o f Conflict Resolution 15 (1971)471-485. Baldwin, James, Economic Statecrqfi (Princeton, NJ: Princeton University Press, 1985). Baliga, Sandeep and Roberto Serrano, "Multilateral Bargaining with Incomplete Information," Journal o f Economic TTteo/y 67:2 (1995) 578-589. Banks, Jef&ey, "Equilibrium Behavior in Crisis Bargaming Games,” American Journal o f Political Science 34 (1990) 599-614. Banks, Jef&ey and Joel SobeL "Equilibrium Selection in Signaling Games," Econometrica 55 (1987) 647-662. Barber, James, “Economic Sanctions as a Policy instrument,” International Affairs 55 (1979) 367-384. Beschloss, Michael, The Crisis Tears: Kennedy and Khrushchev (New York, NY: HarperCollins Publishers, 1991). Bethehem, D.L. ed. The Kuwait Cisis: Sanctions and Their Economic Consequences (Cambridge, MA: Grotius Publications, 1991). Bmmore, Ken, Fun and Games (Lexington, MA: D.C. Heath and Co., 1992). Bhzmger, Richard, "Arms to Go," International Security, \T 2 (Fall 1992) 84-111. 371 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Blume, Andras, "Equilibrium Refinements in Sender-Receiver Games," Journal o f Economic Theory 64:1 (1994) 66-77. Brams, Steven J. and D. Marc Kilgour, "Threat Escalation and Crisis Stability: A Game-Theoretic Analysis," American Political Science Review % \ (1987) 833-849. Bueno de Mesquita, Bruce, "The War Trap Revisited: A Revised Expected Utility Model," American Political Science Review 79 (1985) 156-177. Bueno de Mesquita, Bruce, "The Costs of Wan A Rational Expectations Approach," American Political Science Review 77 (1983) 347-356. Bueno de Mesquita, Bruce, "An Expected Utility Theory of International Conflict," American Political Science Review 74 (1980) 917-931. Bueno de Mesquita, Bruce and David Lalman, War and Reason (Yale University Press: New Haven, CT, 1992). Bueno de Mesquita, Bruce and David Lalman, "Domestic Opposition and Foreign War," American Political Science Review 84 (1990) 747-765. Bueno de Mesquita, Bruce and David Lalman, "Modeling War and Peace," American Political Science Review SI (1987)221-230. Bueno de Mesquita, Bruce, James Morrow, and Ethan Zorick, "Capabilities, Perception, and Escalation," American Political Science Review 91:1(1997) 15-27. Bueno de Mesquita, Bruce, Randolph M. Siverson, and Gary Woller, "War and the Fate of Regimes: A Comparative Analysis," American Political Science Review 86 (1992) 638-646. Calderon, Richard, "Panama: Disaster or Democracy," Foreign Affairs 66:2 (Winter 1987/88) 328-347. Cameron, Charles and B.P. Rosen, "A Signaling Theory of Congressional Oversight," Games and Economic Behavior 5 (1993) 44-70. Campbell, Barry and Danforth Newcomb, The Impact o f the Freeze o f Kuwaiti and Iraqi Assets on Financial Institutions and Financial Transactions (Boston, MA: Graham & Trotman, 1990). Canes, Michael, Country Impacts ofM ultilateral Oil Sanctions and the Conduct o f U.S. Foreign Policy. American Petroleum Institute, Issue Analysis #104. April 1998. Carieton, David and Michael StohL "The Role of Human Rights in U.S. Foreign Assistance Policy," American Journal o f Political Science “ i l (1987) 1002-1018. Carlson, Lisa, "A Theory of Escalation and International Conflict," Journal ofConflict Resolution 39 (1995)511-534. Carter, Barry E., International Economic Sanctions: Improving the Haphazard U.S. Legal Regime. (Cambridge, MA: Cambridge Universi^ Press, 1988). “A Catalog of New U.S. Unilateral Economic Sanctions for Foreign Policy Purposes, 1993-1996," National Association of Manufactures Working Paper, March 1997. 372 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Celentani, Marco and Wolfgang Pesendorfer, "Réputation in Dynamic Games," Journal o f Economic Theory 70:1 (1996) 109-132. Cho, In-Koo and David Kreps, "Signaling Games and Stable Equilibria," Quarterly Journal o f Economics 102 (1987) 179-222. Cho, In-Koo and Joel Sobel, "Strategic Stability and Uniqueness in Signaling Games," Journal o f Economic Theory S O (1990) 381-413. “Collective Sanctions; an Alternative to Military Coercion?" International Relations. 12 (December 1994) 75-91. Cooper, Mary, "Economic Sanctions: Can They Replace Combat in the Post-Cold War Era?," CQ Researcher 4 (October 28 1994) 939-958. Cortright, David and George Lopez ed. Economic Sanctions: Panacea or Peacebuilding in a Post-Cold War W orlcP. (Boulder, CO: Westview Press: Boulder, 1995). Cowen, Tyler and Daniel Sutter, “Politics and the Pursuit of Fame,” Public Choice. 93 ( 1997) 1 9-35. Daoudi, M S., Economic Sanctions, Ideals and Experience. (London, UK: Routledge & Kegan Paul, 1983). Dashti-Gibson, Jaleh Michelle, Enhancing Enforcement: A Framework fo r Monitoring Multilateral Economic Sanctions. PhJD. Dissertation, Notre Dame University, Indiana, 1998. Dehejia, R.H. and B. Wood, "Economic Sanctions and Econometric Policy Evaluation — A Cautionary Note," Journal o f World Trade 26:1 (February 1992) 73-84. Dixit, Avinash, "Strategic Behavior in Contests," American Economic Review 77:5 ( 1987) 891-898. Doxey, Margaret, "Sanctions in an Unstable International Environment: Lessons from the Gulf Conflict" Diplomacy and Statecraft. 2 (November 1991) 208-225. Doxey, Margaret, International Sanctions in Contemporary Perspective. (St. Martin's Press: New York. 1987). Doxey, Margaret "Oil and Food as International Sanctions." International Journal. 36:1 (Winter 1980-81)311-334. Doxey, Margaret Economic Sancttdns and Intematt'onal Erforcement. (Oxfrtrd University Press: New York. 1980). Doxey, Margaret International Sanctions: A Framework for Analysis with Special Reference to the U.N. and South Africa." International Organtation 26 (1972) 527-550. Dreze, J. and H. Gazdar, Hunger and Poverty in Iraq, 1991," World Development 20:7 Q vHy 1992) 921-945. 373 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Drury, Alfred Cooper, Economic Sanctions and Presidential Decisions: Models o f Political Rationality. Ph.D. Dissertation, Arizona State University, 1997. Dyker, David and Vesna Bojicic, "The Impact of Sanctions on the Serbian Economy," RFE/RL Research Report 2 (May 21, 1993) 50-54. Eaton, Jonathan and Maxim Engers "Sanctions," Journal o f Political Economy 100:5 (October 1992) 899-928. Elliot, ICA. and PJ* Uimonen, "The Effectiveness of Economic Sanctions with Application to the Case of Iraq," Japan and the World Economy 5:4 (December 1993) 403-409. Epstein, Larry G. and Michel Breton, "Dynamically Consistent Beliefr must be Bayesian," Journal o f Economic Theory 61:1 (1993) 1-22. Evans, Peter, Harold K. Jacobson, and Robert D. Putnam, Double-Edged Diplomacy: International Bargainng and Domestic Relations (Berkeley, CA: University of California Press, 1993). Eyerman, Joe, and Robert A. Hart, Jr., "An Empirical Test of the Audience Cost Proposition: Democracy Speaks Louder than Words," Journal o f Conflict Resolution 40 ( 1996) 597-616. Fearon, James D., "Signaling Foreign Policy Interests," Journal o f Conflict Resolution A \:\ (1997) 68-91. Fearon, James D., "Rationalist Explanations for War," International Organtations 49 ( 1995) 379-414. Fearon, James D., "Signaling versus the Balance of Power and Interests," Journal o f Conflict Resolution 38:2(1994) 236-269. Fearon, James D., "Domestic Political Audiences and the Escalation of International Disputes," American Poilitical Science Review 88:3 (1994) 577-592. Fearon, James D., Threats to use Force: The Role o f Costly Signals in International Crises, PhT). Dissertation, University of California at Berkeley, 1992. Forland, T.E., "The History of Economic Warfare," Journal o f Peace Research 30:2 (May 1993) 151-162. Francisco, Ronald, "The Relationship between Coercion and Protest," Journal o f Corflict Resolution 39:2 (1995)263-281. Fraunces, M.G., "The International Law of Blockade— New Guiding Principles in Contemporary State Practice," Yale Law Journal 101:4 (January 1992) 893-918. Freeman, HJ., "United States Economic Sanctions against Vietnam: International Business and Development Repercussions," Columbia Journal o f World Business 28:2 (Summer 1993) 12-22. Fudenberg, Drew, David M. Kreps, and David K. Levine, "On Robustness of Equilibrium Refinements," Journal o f Economic Theory AA (1988) 354-380. 374 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Fudenberg, Drew and Jean Tiroie, "Sequential Bargaining with Incomplete Information about Preferences," Review o f Economic Studies 50:2 (1983) 221-247. Galtung, Johan, "On the Effects of International Economic Sanctions," World Politics 19:4. (1967) 378-416. Gartinkel, Michelle, "Domestic Politics and International Conflict," American Economic Review 84:5 (1994) I294-I309. Gaubatz, Kurt Taylor, "Election Cycles and War," Journal o f Conflict Resolution 35:2 (I99I) 212-244. Gavin, Joseph, “Economic Sanctions: Foreign Policy Levers or Signals?” Cato Institute Policy Analysis No. 124. November 7, 1989. Gelpi, Christorpher, "Crime and Punishment: The Role of Norms in Crisis Bargaining," American Political Science Review 91 (1997) 339-360. Gianelli, William. Op-Ed "Review Panama Sanctions for Canal's Sake." The Sunday Herald. May 28, I989-.B3. Gilpin, Robert, War and Change in World Politics (Cambridge, UK: Cambridge University Press, 1981). Grossman, Gene and Elhanan Helpmanm, "The Politics of Free Trade Agreements," American Economic Review, 84:4 (September 1995) 667-691. Grossman, Herschel, "A General Equilibrium Model of Insurrectins," American Economic Review 81:4 (1991)912-921. Grossman, Sanford, "Sequential Bargaining under Asymmetric Information," Journal o f Economic Theory 29:1 (1986) 120-154. Gupta, Dipak and Harinder Singh, "Government Coercion of Dissidents," Journal o f Conflict Resolution 37-2(1993) 301-337. Gurr, Ted Robert Why Men Rebel (Princeton, NJ: Princeton University Press, 1970). Haass, Richard, “Sanctioning Madness,” Foreign Affairs 76:6 (1997) 74-85. Hagan, Joe D., "Domestic Political Conflict Issue Areas, and Some Dimensions of Foreign Policy Behavior other than Conflict," International Interactions 12:4 (1986) 291-313. Haggard, Stephan and Robert Kaufinan, The Political Economy o f Democratic Transitions (Princeton, NJ: Princeton University Press, 1995). Henderson, Conway, "Conditions Affecting the Use of Political Repression," Journal ofConflict Resolution 35 (I99I) 120-142. Hendrickson, D C., "The Democratist Crusade — Intervention, Economic Sanctions, and Engagement," World Policy Journal 11:4 (Winter 1994) 18-30. Hermann, Margaret and Hermann, Charles, “Who Makes Foreign Policy Decisions and How: An Empirical Inquiry,” Intemcttional Studies Quarterly 33 (1989) 361-387. 375 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Herring, George, "America and Vietnam: The Unending War," Foreign Affairs 70:5 (Winter 1991/92) 104-119. Hess, Gregory D. and Athanasios Orphanides, "War Politics: An Economic, Rational-Voter Framework," American Economic Review 85:4 (1995) 828-846. Hibbs, Douglas, Mass Political Violence: A Cross-National Causal Analysis (Wiley: New York, 1973). Hillman, AX., H.W. Ursprung, "Domestic Politics, Foreign Interests and International Trade Policy," American Economic Review 78 (1988) 729-745. Hirshleifer, Jack, "The Analytics of Continuing Conflict," Synthèse 76:2 (1988) 201-233. Hirshleifer, Jack, "Conflict and Rent-Seeking Success Functions: Ratio versus Difference Models of Relative Success," Public Choice 63:2 (1989) 101-112. Hofnann, Fredrik, "The Functions of Economic Sanctions," Journal o f Peace Research 2 (1967) 140-159. Holsti, Ole, Public Opinion and American Foreign Policy (Ann Arbor, MI: University of Michigan Press, 1996). Hufbauer, Gary, Kimberly Ann Elliott, Tess Cyrus, Elizabeth Winston, “U.S. Economic Sanctions: Their Impact on Trade, Jobs, and Wages,” Institute fo r International Economics Working Paper, 1997. Hufbauer, Gary, Jeflfiey Schott, and Kimberly Elliot, Economic Sanctions Reconsidered 2tid ed., (Washington, DC: Institute for International Economics, 1990). Hufbauer, Gary, Jef&ey Schott, and Kmiberly Elliot, Economic Sanctions Reconsidered. (Washington, DC: Institute for International Economics, 1985). Hufbauer, Gary, Jeffrey Schott, and Kimberly Elliot, Economic Sanctions in Support o f Foreign Policy Goals (Washington, DC: Institute for International Economics, 1983). Huth, Paul and Bruce Russett, "General Deterrence Between Enduring Rivals: Testing Three Competing Models," American Political Science Review 87 (1993) 61-72. Huth, Paul, and Bruce Russett, "Deterrence Failure and Crisis Escalation" International Studies Quarterly 32(1988)29-46. International Monetary Fund. International Statistical Yearbook, Various Years. (Washington, DC: International Monetary Fund). James, Patrick and John R. Oneal, “The Influence of Domestic and International Politics on the President’s use of Force,” Journal ofC otflict Resolution 35:2 (June 1991) 307-332. James, Patrick, “Domestic Politics and Foreign Policy: Evaluating a Model of Crisis Activity for the United States,” The Journal o f Politics 56:2 (May 1994) 327-348. Jenkffls, Carolyn Marion, The Effects o f Sanctions on Formal Sector Employment in South Africa (Sussex, UK: Institute of Development Studies, 1993). 376 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Jervis, Robert, "Rational Deterrence: Theory and Evidence," World Politics 42 (1989) 183-207. Jervis, Robert, Perceptions and Misconceptions in International Politics (Princeton, NJ: Princeton University Press, 1976). Johns, Christina and P. Ward Johnson, State Crime, the Media, and the Invasion o f Panama (London, UK: Praeger, 1994). Jowett, Garth and Victoria OTDonnell, Propaganda and Persuasion (New York, NY: Sage Publications, 1986). Kaempfer, William and Anton Lowenberg, "The Problems and Promise of Economic Sanctions," Scandivavian Journal o f Development Alternatives 13 (September 1994) 102— 114. Kaempfer, William and Anton Lowenberg, "Using Threshold Models to Explain International Relations," Public Choice {Jme 1992) 419-443. Kaempfer, William and Anton Lowenberg, International Economic Sanctions : a Public Choice Perspective. (Westview: Boulder, Colorado, 1992). Kaempfer, William, "The Theory oflntematiomd Economic Sanctions - A Public Choice Approach." The American Economic Review, 78:4, (September 1988) pp. 786-793. Katzenstein, Peter ed.. Between Power and Plenty (Madison, WI: University of Wisconsin Press, 1978). Kennen, John and Robert Wilson, "Bargaming with Private Information," Journal o f Economic Literature 31 (1993) 45-104. Kerr, W. A. and J. D. Gaisford, "A Note on the Increasing the Effectiveness of Sanctions," Journal o f IPbr/d rro</e 28:6 (December 1994) 169-176. Kilgour, D. Marc and Frank Zagare, "Uncertainty and the Role of the Pawn in Extended Deterrence," Synthèse 100 (1994) 379-412. Kilgour, D.M., "Domestic Political Structure and War Behavior A Game Theoretic Approach," Journal o f Conflict Resolution 35 (1991) 266-284. Khn, Woo-Sang and Bruce Bueno de Mesquita, "How Perceptions Influence the Risk of War," International Studies Quarterly 39 (1995) 51-65. King, Gary; Robert Keohane, and Sidney Verba, Designing Social Inquiry (Princeton, NJ: Princeton University Press, 1994). Knorr, Klaus, The Power o f Nations (New York, NY: Basic Books, 1975). Kreps, David M., A Course in Microeconomic Theory (Princeton, NJ: Ehinceton University Press, 1990). Kreps, David and R. Wilson, "Sequential Equilibrium," Econometrica 50 (1982) 863-894. Kreps, David M., Game Theory and Economic Modelling (Oxford, UK: Clarendon Press, 1990). 377 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Knigman, Paul and Maurice Obstfeld, International Economics: Theory and Practice 2nd Edition (New York, NY: HarperCollins Publishers, 1991). Lacy, Dean, and Emerson Niou, “Nonseparabie Preferences, Issue Linkage, and Economic Sanctions,” Paper Presented at the Annual Meeting o f the American Political, September 3-6, 1998. Lamare, James, International Crisis and Domestic Politics: Major Political Conflicts in the 1980s (New York, NY: Praeger, 1991). Lax, David and James Sebenius, The Manager as Negotiator (The Free Press, New Yoirk, 1986). Leidy, MichaeL "The Theory of International Economic Sanctions - A Public Choice Approach: Comment," The American Economic Review 19:5 (December 1989) 1300-1303. Lenway, SteAnie, "Between Wars and Commerce: Economic Sanctions as a Tool of Statecraft," International Organizations (1992). Leyton-Brown, David ed.. The Utility o f International Economic Sanctions (London, UK: Croom Helm, 1987). Li, Chien-Pin, "The Effectiveness of Sanction Linkage: Issues and Actors," International Studies Quarterly 370 (1993) 349-370. Lian, Bradley and James Oneal, "Presidents, the Use of Military Force, and Public Opinion," Journal o f Conflict Resolution 370 (1993) 277-300. Lichbach, Mark Irving, "Deterrence or Escalation: the Puzzle of Aggregate Studies of Repression and Dissent," Journal o f Conflict Resolution, 310 (1987) 266-297. Lichbach, Mark Irving, "Regime Change: A Test of Structuralist and Functionalist Explanations," Comparative Political Studies, 14:1(1981) 49-71. Lijpbart, A, "Comparative Politics and the Comparative Method." American Political Science Review. 65 (1971)682-693. Lindbloom, Charles, Politics and Markets (New York, NY: Basic Books, 1977). Lindsay, James, "Trade Sanctions as Policy Instruments." International Studies Quarterly 30 (1986) 153-173. Lipset, Symour Martin, "Some Social Requisites of Democracy: Economic Development and Political Legitimacy," American Political Science Review 53 (1959) 69-105. Lohmann, Susanne, '^vA:age?o\i!6.cs,” Journal o f Conflict Resolution A \:\ (1997)38-67. Lohmann, Susanne, "Electoral Cycles and International Policy Coordination," European Economic Review 'll Q992) 1373-1391. Losman, Donald, International Economic Sanctions. (University of New Mexico Press: Albuquerque 1979). 378 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Mailatb, George J. and Okuno-Fujiwara Vfasahiro, "Belief-Based Refinements in Signalling Games," Journal o f Economic Theory (1993) 241-276. MaOatfa, George J., "An Abstract Two-Period Game with Simultaneous Signaling-Existence of separating Equilibria," Journal o f Economic Theory 46:2 (1988) 373-394. Malloy, Michael P., Economic Sanctions and U.S. Trade (Boston, MA: Little, Brown and Co, 1990). Mansfield, Edward D., Power, Trade, and IKar (Princeton, NJ: Princeton University Press, 1994). Maoz, Z., "Resolve, Capabilities, and the Outcomes of Interstate Disputes, 1816-1976," Journal o f Conflict Résolution!! (1983) 195-229. Martin, Lisa, "Credibility, Costs, and Institutions: Cooperation on Economic Sanctions," World Politics 45-J (April 1993) 406-432. Martin, Lisa, Coercive Cooperation: Explaining Multilateral Economic Sanctions. (Princeton University Press: New Jersey, 1992). MaxwelL Joseph. Qualitative Research Design. (Thousand Oaks, CA: Sage Publications, 1996). Mayer, Wolfgang, "Endogenous Tariff Formation." Bhagwati, Jagdish. ed. International Trade (Cambridge, MA: MIT Press, 1993). McAdam, Doug and David A. Snow, ed.. Social Movements (Los Angeles, CA: Roxbury Publishing Co., 1997). McKelvey, Richard D. and Thomas R. Palfiey, "Endogeneity of Alternating Offers in a Bargaining Game," Journal o f Economic Theory 73:2 (1997) 425-437. McKinlay, R.D. and AS. Cohan, "A Comparative Analysis of the Political and Economic Performance of Military and Civilian Regimes," Comparative Politics 8 (1975) 1-30. McNamara, Robert, In Retrospect (New York, NY: Times Books, 1995). Miller, Judith, "When Sanctions Worked." Foreign Policy 39 (Summer 1980) 118-129. Milner, Helen, "International Theories of Cooperation Among Nations," World Politics 44 ( 1992) pp. 466-497. Milner, Helen V. and B. Peter Rosendorff, "Domestic Politics and International Trade Negotiations," Journal ofC orflict Resolution 4 \:\ (1997) 117-147. Milner, Helen V. and B. Peter Rosendorff, "Trade Negotiations, Information, and Domestic Politics: The Role o f Domestic Groups," Economics and Politics 8:2 (1996) 145-189. Minter, William, "When Sanctions Worked: The Case of Rhodesia Reexamined," African Affairs 87 (April 1988) pp. 207-237. Mitchell, NeQ and James M. McCormick, "Economic and Political Explanations of Human Rights Violations," World Politics 4Q{\9ÎX)4!6A9% . 379 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Miyagawa, Makio, Do Economic Sanctions W o rk P . (New York, NY: SL Martin's Press, 1992). Mogenthau, Hans J., Politics Among Nations: The Struggle fo r power and Peace (New York, NY: McGraw-Hill, Inc., 1993). Moore, Will H., "Action-Reaction or Rational Expectations?: Reciprocity and the Domestic-International Conflict Nexus During the ^oàesÎ3iPro\il<etsi!,” Journal o f Conflict Resolution 29 {1995) 129-167. Morgan, T. Clifton, “Power, Resolve and Bargaining in International Crises: A Spatial Theory,” International Interactions Vol 15:3 (1990) 279-302. Morgan, T. Clifton and Valerie Schwebach, "Economic Sanctions as an Instrument of Foreign Policy: The Role of Domestic Politics," International Interactions 21 (1995). Morgan, T. Clifton and Valerie Schwebach, "Fools Suffer Gladly: The Use of Economic Sanctions in International Crises," International Studies Quarterly (1987) 27-50. Morrow, James D., "A Spatial Model of International Conflict," American Political Science Review 80 (1986) 1131-1150. Morrow, James D., Game Theory fo r Political Scientists (Princeton, NJ: Princeton University Press, 1994). Morrow, JJ)., "Capabilities, Uncertainty, and Resolve: A Limited Information Model of Crisis Bargaining," American Journal o f Political Science 33 (1989) 941-972. Moyer, Homer, Export Controls as Instruments o f Foreign Policy: the History, Legal Issues, and Policy Lessons o f Three Recent Cases (Washington, DC: International Law Institute, 1988). Muller, Edward N., "Income Inequality, Regime Repressiveness, and Political Violence," American Sociological Review 50 (1985) 47-61. Muller, Edward and Erich Weede. "Cross-national Variation in Political Violence," Journal o f Conflict Resolution 34 ( 1990) 624-651. Nincic, Miroslav and Peter Wallensteen, Dilemmas o f Economic Coercion: Sanctions in World Politics (New York, NY: Praeger, 1983). Nossal, Khn, "International Sanctions as International Punishment," International O rgantations 43:2 (Spring 1989). Ostrom, Charles W. and Brian L. Job, “The President and the Political use of Force,” American Political Science Review Vol 80:2 (June 1986) 541-563. Paarlberg, Robert, "Lessons of the Gram Embargo" Foreign Affairs 59 (Fall 1980) 144-162. Pabre, Robert and PauLA. Papayaunou, "Usmg Game Theory to Link Domestic and International Politics," Journal o f Conflict Resolution 41:1 (1997) 4-12. Polachek, Solomon William, "Conflict and Trade," Journal o f Cotflict Resolution 24:1 (1980) 55-78. 380 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Porter, Richard, "International Trade and Investment Sanctions," Journal o f Conflict Resolution 23:4 (1979) 579-611. PoweU, Robert, "Crisis Bargaining, Escalation, and MAD," American Political Science Review 81 (1987) 717-735. Purcell, Susan, "Collapsing Cuba," Foreign Affairs 71:1 (1991/92) 130-145. Putnam, Robert, "Diplomacy and Domestic Politics: The Logic of Two Level Games," International Organization 42 J (1988) 427-460. Rabin, Matthew and Joel Sobel, "Deviations, Dynamics, and Equilibrium Refinements," Journal o f Economic Theory 6i:l (1996) 1-25. Ramet, Sabrina, "War in the Balkans," Foreign Affairs 71:4 (Fall 1992) 89-98 Renwick, Robin, Economic Sanctions. (Harvard University: Cambridge 1981). Richburg, Keith, "Back to Vietnam," Foreign 4(^/rs 70:4 (Fall 1991) 111-131. Roberts, Ken, "Bullying and Bargaining," International Security 15:2 (Fall 1990) 67-102. Robins, Carla Aime, "Dateline Washington: Cuban-American Clout," Foreign Policy 88 (Fall 1992). Robinson, Linda, "Dwindling Options in Panama," Foreign Affairs 68:5 (Winter 1989/90) 187-205. Rodman, KA, "Sanctions at Bay — Hegemoic Decline, Multinational Corportations, and United States Economic Sanctions Since the Pipeline Case," International Organtation 49:1 (Winter 1995) 105-137. Rogers, Elizabeth, "Using Economic Sanctions to Control Regional Conflicts," Security Studies 5:4 (1996)43-72. Rogers, Michael, "Economic Sanctions Discouraging Record," Forbes (October 13, 1986) 8. Rosendorff B. Peter, "A Theory on Transition to Democratic Rule," Unpublished. August 1996. Roth, Alvin E. and Françoise Schoumaker, "Expectations and Reputations in Bargaining," American Economic Review 73 J (1983) 362-372. Rubinstein, Ariel, "Perfect Equilibrium in a Bargaining Model," Econometrica 50:1 (1982) 97-109. Rubinstein, Ariel, "Bargaining Model with Incomplete Information about Time Preferences," Econometrica 53:5 (1985) 1151-1172. Rueschemeyer, Dietrich, Evelyne Huber Stephens, and John D. Stephens, Capitalist Development and Democracy (Chicago, IL: University of Chicago Press, 1992). Rusmow, Dennison, "Yugoslavia: Balkan Breakup?" Foreign Polity. 83 (Summer 1991) 143-159. Russett, Bruce, Grasping the Democratic Peace (Princeton, NJ: Prmceton UniversiQr Press, 1993). 381 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. "Sanctions: Do They Work?" Bulletin o f the Atomic Scientists. 49 (November 1993) 14-49. "Sanctions in Haiti: Crisis in Humanitarian Action,” (Cambridge, MA: Harvard Center for Population and Development Studies, 1993). Scharfen, John, The Dismal B attl^eld: Mobilizing fo r Economic Conflict (Maryland, MD: Naval Institute Press, 1995). Schock, Kurt, "A Conjunction Model of Political Conflict," Journal o f Conflict Resolution 40:1 (1990) 98-133. Schweiler, Randolp, "Domestic Structure and preventive Wan Are Democracies more Pacific," World Politics 44 (1992) 235-269. Seidon, Zachary, Economic Coercion in Theory and Practice. Ph D. Dissertation, University of California, Los Angeles, 1995. Shelling, Thomas, Arms and Irfluence (New Haven, CT: Yale University Press, 1966). Shelling, Thomas, The Strategy o f Conflict, (Cambridge, MA: Harvard University Press, 1960). Sigelman, Lee and Miles Simpson, "A Cross-National Test of the Linkage between Economic Inequality and Political W\o\etKe,” Journal o f Cotflict Reslolution, 21:1 (1977) pp. 105-125. Simon, Marc, “When Sanctions can Work: Economic Sanctions and the Theory of Moves,” International Interactions Vol 21:3 (1996) 203-228. Skaperdas, Stergious, "Cooperation, Conflict, and Power in the Absence of Property Rights," American Economic Review 82:4 (1992) 720-739. Skaperdas, Stergios, "Conflict and Attitudes toward Risk," American Economic Review 81:2 (1991) 116-134. Smith, Alastair, “International Crises and Domestic Politics,” American Political Science Review, 92:3 (Summer 1998) 623-638. Smith, Alastair, “International Crises and Domestic Politics,” International Studies Quarterly 40 ( 1996) 133-153. Smith, Alastair, and David Hayes, “The Shadow of the Polls: Electoral Effects on International Agreements,” International Interactions 23:1 (1997) 79-108. Snyder, GTI. and P. Diesing, Corflict Among Nations: Bargaining Decision-Making and System Structure in International Crises (Princeton, NJ: Princeton University Press, 1977). Sobel, Joel and Ichiro Takahashi, "A Multi-Stage Modle of Bargaining," Review o f Economic Studies 50-3(1983)411-426. Spence, A. Michael, "Job Market Signalling," Quarterly Journal o f Economics 87 (1973) 35-374. Spindler, Z. A., "The Public Choice of Superior" Sanctions," Public Choice 85 (1995) 205-226. 382 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Strack, Harty, Sanctions: The Case o f Rhodesia (Syracuse University Press: New York. 1978). Streeten, P., "Freezing Assets-the USA and the Most Effective Economic Sanctions," International Affairs 69:3 (July 1993) 577. Tilly, Charles, From Mobilization to Revolution (Boston, MA: Addison-Wesley, 1978). Tsebelis, George, "The Abuse of Probability in Political Analysis: The Robinson Crusoe Fallacy," American Political Science Review 83 (1989) 77-91. Tsebelis, George, "Are Sanctions Effective: A Game Theory Analysis." Journal o f Conflict Resolution 34:1 (March 1990) 3-28. United States International Trade Commission, "Lifting the Trade Embargo on Vietnam," International Economic Review (December 1991). "U.S. Economic Warfare Brings Disaster to Panama," New York Times. June 9, 1988:A1. van Bergeijk, Peter A.G., Economic Diplomacy, Trade, and Commercial Policy: Positive and Negative Sanctions in a New World Order. (E. Elgan USA, 1994). van Bergeijk, Peter, "Success and Failure of Economic Sanctions," International Studies Quarterly 30 (1986) 153-173. van Bergeijk, Peter A. G. and Charles van Marrewijk, "Why do Sanctions need Time to Work," Economic Modelling 12:2 (1995) 75-86. van Winkle, Hans A., The Tactics o f Economic Sanctions: a Measure Short o f War (Carlisle Barracks, PA: U.S. Army War College, 1992). Wallerstein, P., "Characteristics of Economic Sanctions," Journal o f Peace Research 5 (1968) 248-267.. Weede, Erich. "Income Inequality, Average Income, and Domestic Violence," Journal o f Conflict Resolution 25:4 (1981) 639-653. Weede, Erich, "Rent Seeking, Military Participation, and Economic Performance in LDC's," Journal o f Cottflict Resolution 30:2 (1986) 291-314. Weintraub, Sidney ed.. Economic Coercion and U.S. Foreign Policy (Boulder, CO: Westview Press, 1982). Wolf, Charles, International Economic Sanctions (Santa Monica, CA: Rand Corp, 1980). Yahfl, Lent, The Holocaust (New York, NY: Oxford University Press, 1990). 383 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Behind the negotiated revolution: economic reform, social change and legitimation in Hungary, 1956-1989

PDF

Individual irrationality and market rationality: Representativeness bias and risk preferences in a double auction market

PDF

The political economy of domestic airline mergers: The neglected international determinants

PDF

Context is everything: An empirical study of institutions, instability and economic growth

PDF

Developing economies and outward foreign direct investments: The case of the South Korean electricity and electronics industry

PDF

Semiperipheral mobility in the world economy: the experience of South Korea's industrial upgrading

PDF

State legitimacy and development: Moving beyond social capital and ethnic theories of African stagnation

PDF

The determinants of electoral participation in the United States, 1952-1992: A theoretical and empirical examination

PDF

The gravity model at work: An empirical study on the effects of governance on bilateral trade flows with special emphasis on intra-Arab trade

PDF

Political economy of economic reform, quality of reform and economic performance: Reform matters, but consistent incentives matter more

PDF

Gateways to the future: Bridging techniques and their effects on technology adoption and standards-setting

PDF

Essays on host country effects of capital flows: Volatility and spillovers

PDF

Diversity vs. unity: Causes and dynamics of ethnic polarization

PDF

Political economy of financial liberalization in emerging markets: A comparative study of South Korea and Taiwan in the 1990s

PDF

The fortunes of fishes and fishers: The political economy of innovation in Atlantic resource management

PDF

An analysis of the developmental state: The case of the Vietnamese textile and garment industries

PDF

Institutional development in the software industry: Intellectual property protection

PDF

Money for the Third Reich: The Nazi's financial legerdemain, 1933-1938

PDF

The political economy of Turkish Islamism and secularism: The role of preference falsification in the establishment of Kemalist secularism and the polarization of Turkish public opinion on the pu...

PDF

Privatization of the telecommunications industry in Saudi Arabia

Asset Metadata

Creator Dick, Howard Evan (author)

Core Title Economic sanctions, domestic audiences, and international conflict

Contributor Digitized by ProQuest (provenance)

Degree Doctor of Philosophy

Degree Program Political Economy and Public Policy

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag economics, general,OAI-PMH Harvest,political science, international law and relations

Language English

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c17-557058

Unique identifier UC11351047

Identifier 9955054.pdf (filename),usctheses-c17-557058 (legacy record id)

Legacy Identifier 9955054.pdf

Dmrecord 557058

Document Type Dissertation

Rights Dick, Howard Evan

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA