Paths to Peace: Domestic Coalition Shifts, War Termination and the Korean War 9780804772372

Citation preview

Paths to Peace

Paths to Peace domestic coalition shifts, war termination and the korean war

Elizabeth A. Stanley

stanford universit y press Stanford, California 2009

Stanford University Press Stanford, California © 9 by the Board of Trustees of the Leland Stanford Junior University All rights reserved No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or in any information storage or retrieval system without the prior written permission of Stanford University Press. Printed in the United States of America Library of Congress Cataloging-in-Publication Data Stanley, Elizabeth A. Paths to peace : domestic coalition shifts, war termination and the Korean war / Elizabeth A. Stanley. p. cm. Includes bibliographical references and index.  ---6269-4 (cloth : alk. paper) . Korean War, 1950–1953—Peace. . Korean War, 1950–1953—Diplomatic history. 3. War—Termination—Decision making—Case studies. 4. Coalitions—Case studies. I. Title. ds921.75.s73 9 951.904'.21—dc 2009007182 This book is printed on acid-free, archival-quality paper. Typeset at Stanford University Press in  ⁄. Minion

For John

Contents

Acknowledgments  xi Abbreviations  xiii

Part I: The Domestic Politics of War Termination 1 Old Baldy  3 2 Domestic Coalition Shifts in War Termination  19

Part II: The Korean War 3 4 5 6 7 8

The Korean States: Powerless Players in the Forgotten War  65 The Soviet Union: Outlasting Stalin’s Preferences  91 The United States, Part 1: Trapped by the NSC-68 Mindset  120 The United States, Part 2: Trapped by Voluntary Prisoner Repatriation  150 China: Trapped by a Hawkish Ally  178 Interacting Domestic Coalitions in Bargains for Peace  210

Part III: Testing and Extending the Argument 9 Domestic Coalition Shifts in War Termination since 1862  239 10 War Termination in Theory and Practice  272 Epilogue  303 Appendixes A. Interstate Wars Included in the Post–World War II Dataset  305 B. Domestic Governing Coalition Shifts since World War II  308

viii

contents

Notes  313    Bibliography  339    Index  383

Maps, Figures and Tables

Maps 1.1 Map of the Old Baldy Area  5 3.1 The Korean Peninsula, Showing Lines of Advance, 1950–53  66



Figures

1.1 Unloading Logs for Bunker Construction at the 2nd Infantry Division Supply Point on Old Baldy, October 1, 1952  4 2.1 Paths by Which Belligerents Change Expectations  50 2.2 Paths to War Termination  52 2.3 How Changed Expectations Create Bargaining Space  56

Tables 2.1 Summary of the Five Information Obstacle Dynamics  43 8.1 9.1 9.2 9.3 9.4 9.5

Domestic Coalition Shifts in the Korean War  222 Cross-Tabulation of War Duration by Coalition Shifts  249 Cross-Tabulation of War Duration by Coalition Shift Direction  249 Cross-Tabulation of War Outcome by Coalition Shifts  250 Cross-Tabulation of War Outcome by Coalition Shift Direction  251 Replicated Hazard Analysis of Bennett and Stam (1996) and Goemans (2000a)  252 9.6 Hazard Analysis with Domestic Coalition Shifts  258 9.7 War-Level Analysis of Domestic Coalition Shifts since World War II  261 9.8 Hazard Analysis of Coalition Shift Directions  264

Acknowledgments

Completing a book about war termination is tough business. Like the wars themselves, recent experience suggests that such books, once begun, are very difficult to end—lasting much longer and costing much more than the author expected at the outset. This endeavor, which began as a Ph.D. thesis and has since gone through repeated changes, would not have been possible without the support of many wonderful people. My mentor at Harvard, Stephen Peter Rosen, and Iain Johnston, Theda Skocpol and Stephen Walt provided very useful early thoughts and recommendations. Financial support to conduct this research and revise the manuscript came from a graduate research fellowship from the National Science Foundation, a post-doctoral fellowship from Harvard’s John M. Olin Institute for Strategic Studies and a junior faculty research leave from Georgetown University. I am especially grateful to my colleagues and students at Georgetown for helping me see this project through to completion and to Georgetown’s Security Studies Program for providing the financial support for research assistance, from Saltanat Berdikeeva, Mindy Chan, Hailey Hoffman, Unyoung Lee, Ethan Marsh, Alexis Pierce and Michael Stanisich. I thank the many participants of various seminars and conferences where I have presented portions of this manuscript over the years, including the American Political Science Association, the International Studies Association, Harvard’s Olin Institute and Georgetown’s Department of Government and Center for German and European Studies. I have benefited greatly from the comments these presentations elicited.

xii

acknowled g ments

I also thank a number of friends and colleagues who read and provided insightful suggestions regarding earlier versions or portions of this manuscript, or supported me in other valuable ways: Gary Bass, Audrey Slesinger Beldio, Nora Bensahel, Beth Blaufuss, Risa Brooks, Michael Brown, Dan Byman, Victor Cha, Audrey Kurth Cronin, Colin Dueck, Robert Dujarric, Bernard Finel, Ben Fordham, Page Fortna, Bob Gallucci, Hein Goemans, Lori Gronich, Kristen Gunness, Lise Morjé Howard, Tim Hoyt, Macartan Humphreys, Dominic Johnson, Gary Kaplan, Charles King, Greg Koblentz, Mark Kramer, Alan Kuperman, Brad Lee, Phyllis Jacobson-Kram, Becky Johnson, Suzanne Neilson, Mariellen O’Hara, Olya Oliker, David Post, Jenny Lind Press, Kristen Quigley, Dani Reiter, Tim Rosenberg, George Shambaugh, Jeremy Shapiro, Sara Sievers, Jennifer Sims, Naunihal Singh, Jack Snyder, Al Stam, Monica Toft, Barbara Walter, Kathryn Weathersby and Jeremy Weinstein. I am especially indebted to Kristin Goss, Brent Mitchell and Tammy Schultz for reading the entire draft of earlier versions of the manuscript and providing critical moral support along the way. I am humbled by the debt I owe to my colleague and friend John P. Sawyer for helping me complete the quantitative analysis in Chapter 9 and for listening to me refining my argument, over and over again. His generous and invaluable support was absolutely essential for completing this project. I thank Geoffrey Burn, my editor at Stanford University Press, for his extra­ ordinary efforts to help me bring this book to publication. I also thank Jessica Walsh and John Feneron for their detailed assistance. I owe a great deal to my undergraduate mentor at Yale, the late Lt. Gen. (Ret.) William E. Odom. In many ways, he is largely responsible for my serving as a military intelligence officer in Korea in 1993—when I first developed an interest in this war—and for my becoming a political scientist after my service in the US Army. I am sad he was unable to see me finally complete this book. Special thanks go to my parents, Deane and Cissie, and my sisters, Alison and Karalyn, for their love and encouragement throughout the process. I also thank Finley and the late Duffy for providing important lap-warmth and love during many hours at the computer. Finally, my deepest gratitude is to my partner, John Schaldach, whose love makes my life more meaningful and balanced, and whose contribution to this book is impossible to describe. This book is dedicated to him. E.A.S

Abbreviations

AFT AIC CCP CIA COW CPV CWIHP DOD DMZ DPRK FRUS GAC GDR GRU JCS KGB KI KMT KWP MAC MGB NATO

Accelerated failure time Akaike Information Criterion Chinese Communist Party Central Intelligence Agency Correlates of War (dataset) Chinese People’s Volunteers (Chinese army in Korean War) Cold War International History Project Department of Defense Demilitarized Zone Democratic People’s Republic of Korea (North Korea) Foreign Relations of United States (See author: U.S. Department of State) Government Administration Council German Democratic Republic (East Germany) Soviet Military Intelligence Joint Chiefs of Staff Soviet Committee for State Security Soviet Committee of Information Kuomintang (Chinese Nationalists) Korean Workers’ Party (Communist Party in North Korea) Soviet Military Administrative Council USSR Ministry of State Security North Atlantic Treaty Organization

xiv

NMC NPKA NSC ODM PH PLA POW PRC ROK SPC TVC UN UNC US USSR

abbrev iat ions

National Material Capabilities (dataset) North Korean People’s Army National Security Council Office of Defense Mobilization Proportional hazards People’s Liberation Army (Chinese army) Prisoner of war People’s Republic of China Republic of Korea (South Korea) State Planning Commission Time-varying covariate United Nations United Nations Command United States Union of Soviet Socialist Republics (Soviet Union)

1

Old Baldy My insides felt like a thousand butterflies were still jumping around in there, and my legs felt like rubber. My mind was still seeing the picture at the top of Old Baldy as the shells burst among us, and hearing the sounds of Chinese bugles blowing as they came charging into our lines. . . . We would bleed and die for an ugly old hill that to the men in the trenches wasn’t worth throwing lives away for, just so we could see Chinese movement to the north. . . .

—Corporal Rudolph W. Stephens, 23rd Regiment, 2nd Infantry Division (1995: 114–15)

Old Baldy is a mountain in central Korea, currently located within the demilitarized zone between North and South Korea. Old Baldy is no longer bald. Like many hills in Korea, it is covered with relatively young deciduous trees and evergreens, all grown in the last fifty years. Old Baldy earned its name during the Korean War, when intensive shelling obliterated all the trees and foliage on the hillside. At that time, Old Baldy’s landscape looked like the surface of the moon, with many old craters from artillery and mortar rounds that had fallen there (see Figure 1.1). As Jerry Ciaravino, who fought there in 1951, remembers, “You knew why it was given that name. It was a big mountain of dirt with no greenery. There was just so much thrown at it that it was just dirt on the top. You were just glad to come back down after you were up there” (Nisbet and Moore 2002: 5). Similarly Corporal Stephens observes, “I believe the only things living on Old Baldy were men in the trenches” (1995: 141). Despite its hard-won name, Old Baldy had no strategic value. There were no natural resources buried inside it, no major communications lines it overlooked, no major industrial center anywhere nearby. In fact, Old Baldy’s only real value—a dubious one, at that—was height. At 266 meters, it was the tallest hill in the immediate area, dominating the terrain to the north, west and south, and it provided a good look-out over another infamous Korean War mountain, Pork Chop Hill. As such, the belligerent that controlled the crest of Old Baldy controlled the tactical situation in the valley below (see Map 1.1). Except for this view down, Old Baldy was not important—not to Korea, and not to the ultimate course of the Korean War. Yet, for this view down, from July



the d omest ic p olit ics of war ter minat ion

f i g . 1 . 1 . Unloading Logs for Bunker Construction at the 2nd Infantry Division Supply Point on Old Baldy, October 1, 1952. Source: Photo by SFC Charles M. Roberts. Available at Wikimedia Commons, http://commons.wikimedia.org/wiki/Image:M39-AUV-Korea19521001.jpg.

1951 until July 1953, China and the United States fought so hard for this blasted, barren hill that, at the peak of the fighting, more than 1,500 men died there every week. During that time, the two sides also faced off in the negotiating tent, ultimately signing an agreement in 1953 that looked remarkably similar to the one proposed two years earlier. Old Baldy changed hands eleven times during the Korean War, with units on both sides suffering massive casualties—often in fierce hand-to-hand combat. As Stephens recounts, the Chinese troops fought with “human-wave attack[s]. . . [they] were dying like flies as they came on. They had the troops to throw away, and they would sacrifice them to win the battle” (1995: 119; also Hermes 1966: 293–95). As a result, as Ming Rose (1999) recalls, “the enemy was killing so many Americans that they had to be piled up like cordwood. Photographs of them were so gruesome that they could not be shown back home.” Indeed, units were sometimes reduced to one-sixth of their normal fighting strength. For example,

old baldy



m a p 1 . 1 . Map of the Old Baldy Area. Source: Wikimedia Commons. http://commons. wikimedia.org/wiki/Image:Old_Baldy_Area.gif

in a particularly costly battle in March 1953, both sides expended 130,000 rounds of ammunition in one night, killing 300 Americans and 800 Chinese (Hermes 1966: 395; Allen 2002: 1).1 The fighting on Old Baldy was not unusual. For two years, from July 1951 until July 1953, this “war of the hills” was waged all around the 38th Parallel while negotiations to end the war stagnated.2 These battles represented some of the most intense fighting in one of the most agonizing and miserable chapters of the Korean War and the casualties were staggering. While no complete tally of battledeaths exists for Old Baldy, numbers for neighboring hills highlight the costs. For example, the United States and South Korea lost 6,400 men in the battles of Heartbreak Ridge and Bloody Ridge, while the Chinese suffered 10,000 killed in the battle of White Horse Hill. In between these costly battles, fighting tapered off to patrol clashes, raids and ambushes for possession of outposts in no-man’s land. Adding to the misery were the bone-numbing cold, the relentless rain, the inadequate food rations, the pounding artillery duels, the invariable sleep deprivation and the inevitable fear. Holding these hills was no walk in the park.



the d omest ic p olit ics of war ter minat ion

To put the costs into perspective, US statistics show that 45 percent of its casualties—22,000 killed, 63,200 wounded—occurred during the “war of the hills” (Foot 1990: 208). The United Nations Command, of which US troops were a part, overall suffered 125,000 wounded and 60,000 killed during this period. Likewise, the Chinese and North Koreans suffered 234,000 killed (Hermes 1966: 500; Hatchimonji 2002). Old Baldy is the quintessence of stalemated war. For two years, China and the United States suffered massive casualties, week after week, even while both governments knew that they would not win the war. Moreover, these casualties happened even though both governments knew that they wanted to end the war and knew that they should. From a strategic view, the losses make no sense—neither side thought that holding this terrain was vital to its strategic interests. This being the case, why did Chinese and American soldiers huddle under devastating exposure to 130,000 artillery rounds in a single night, as the Americans tried to push the Chinese off Old Baldy for the sixth time—long after their own governments had decided that the Korean War was bound for a stalemate? This book is about why those men died. While this story about Old Baldy may seem overly detailed for such an insignificant piece of land, it is instructive nonetheless. If—as most variants of the bargaining model of war assume—“battle-deaths” are a major driver for altering belligerents’ calculus in bargains to end war, then recounting the Old Baldy story reminds us of the human faces behind this anonymous “battle-death” count. One of the biggest puzzles about war is why events like those that occurred on Old Baldy can happen. Why did those governments decide to throw away 294,000 lives fighting for strategically meaningless terrain features, as the conflict seesawed from one hill to the other? What could make this magnitude of loss acceptable? We can look at Old Baldy as a series of battles that were senseless, irrational and quite possibly inexplicable. But this book argues that what happened on Old Baldy during the Korean War was not irrational—at least in the sense that social scientists use that term—nor was it unique. What happened on Old Baldy happens in many wars, and it is explainable. In fact, there is a solid theoretical explanation for this seemingly irrational behavior. This book develops and tests that explanation. In the case of Old Baldy, men continued to die on that hill until other men lost power, thousands of miles away, in Moscow. More generally, this book will demonstrate that the key to ending a war often comes down to changing the composition of the domestic governing coalitions in the states that are fighting it. Old Baldy may make no sense from a strategic viewpoint, but it does from the vantage point of domestic politics.

old baldy



Domestic Coalition Shifts in War Termination Throughout history, shifts in governing coalitions have critically affected war termination. For example, the execution of the Athenian democratic ruler Cleophon and the ascendancy of the pro-Spartan oligarchs in BC 404 led to Athens’ surrender to Sparta, ending the 27-year Second Peloponnesian War. Similarly, the death of Russian Empress Elizabeth in January 1762 led her Prussophile successor, Peter III, to immediately recall Russian armies who were occupying Berlin and conclude the Treaty of Saint Petersburg by May—ending the fighting between Russia and Prussia in the Seven Years’ War. During World War I, the Bolshevik Revolution in 1917 brought Vladimir Lenin and the Bolsheviks to power, who accepted the harsh terms of Brest Litovsk to make a separate peace. The next year, riots in Germany ushered in a new government that then negotiated the final war armistice, as Kaiser Wilhelm II fled to Holland. In the Chaco War, Bolivian President Salamanca was determined to continue at all costs, despite the high command’s decision to end the war. His arrest by the military and forced resignation in November 1934 was the major turning point in the war. The generals placed Vice-President José Luis Tejada Sorzano, who was known to favor peace, in power, and the war ended the next spring. During World War II, France and Italy surrendered shortly after changes in their governing coalitions, in 1940 and 1943 respectively. This phenomenon extends to non-interstate wars, as well. A classic example is the ending of the Wars of the Roses and the installation of the Tudor Dynasty: Richard III died in the Battle of Bosworth Field and Henry VII united the warring factions to end the war. More recently, the leader of the rebel group National Union for the Total Independence of Angola (UNITA), Jonas Zavimbi, was killed in an ambush by government troops in February 2002. Less than six weeks after Zavimbi’s death, a cease-fire was signed, ending the 27-year-long Angolan civil war. Similarly, in Sierra Leone, both sides reached a settlement after Foday Sankoh, the leader of the rebel group Revolutionary United Front, was arrested. Comparable patterns can be observed in extra-systemic wars, such as Charles de Gaulle and Mikhail Gorbachev coming to power in France and the Soviet Union and ending their wars in Algeria and Afghanistan, respectively. Scholars working on issues related to war termination have noted this phenomenon, albeit anecdotally. For example, H. E. Calahan—who argues a change in the relative power among belligerents brings peace—observes that “it seems fair to conclude that a change of regime for the vanquished comes close to being a condition precedent to the making of peace” (1944: 209). James Smith—who argues war ends with an international bargain—notes that “it may indeed be the case that the only way to salvage something from the ruins of war is to have the



the d omest ic p olit ics of war ter minat ion

leader of a particular nation resign” (1995: 106). Among domestic politics scholars, Robert Rothstein concludes that “because it is unlikely that the officials currently in charge can make the necessary changes in policies with which they have become identified, new personnel seem imperative” (1970: 74), while Michael Handel suggests “the termination of a long and stalemated war is frequently preceded by a drastic political change in leadership in the country of one of the belligerents” (1978: 26).3 Bruce Bueno de Mesquita and his colleagues (1992) similarly argue many states in interstate wars experience violent regime changes before the war ends. Finally, in the landmark study of this phenomenon, Fred Iklé (1991) classifies elites in each belligerent state as “hawks” and “doves” and asserts the hawks may need to leave the government before the state can settle. In short, the empirical record includes many examples of domestic governing coalition shifts leading to war termination, and many scholars from different theoretical perspectives have noted this tendency. However, no one has previously attempted to explain the causal mechanisms of this phenomenon in a generalizable manner or to test the argument over a broad group of cases. In this book, I introduce a new theory about shifts in domestic governing coalitions, a state’s elite foreign policy decision-making group. Moreover, I test this theory with detailed historical case studies of the Korean War and quantitative analysis of all interstate wars since 1862. While this book focuses on interstate wars as classified by Singer and Small (1994), Chapter 10 suggests ways this argument could also apply to civil wars. d omest ic coalit ions and inter nat ional bargains for peace

War termination is a political bargain struck between belligerents to dispense with further combat (Bennett and Stam 1998; Filson and Werner 2002, 2004; Goemans 2000a; Pillar 1983; Powell 2002, 2004; Reiter 2003, 2009; Slantchev 2003, 2004; Smith and Stam 2004; Wagner 2000; Werner 1998, 1999; Wittman 1979). As bargaining models about war termination suggest, belligerents will settle a conflict only after they develop an overlapping bargaining space, often modeled at the point at which both sides can agree upon their relative strength. Scholars have often posited a Bayesian rational updating process by which each side updates its expectations about the war and thus changes its bottom line to create an overlapping bargaining space.4 Sometimes wars are short because this bargaining space develops quickly, usually because the battlefield illustrates belligerents’ relative strength in a very compelling manner. Conversely, wars are long because this bargaining space develops more slowly, for example when the battlefield becomes stalemated, belligerents cannot credibly commit, relative power is ambiguous or the issues at stake are indivisible or highly salient.

old baldy



While the Bayesian model assumes the war-ending change in expectations occurs when incumbent leaders change their minds, my research suggests that in many wars, this change in expectations results from a change in the foreign policy leadership itself. Thus, my theory extends the Bayesian model by fleshing out the domestic political aspects of the bargaining model of war. I present a new theory about shifts in domestic governing coalitions and thereby refine the domestic mechanisms for explaining the international bargains that end war. As such, this analysis very consciously builds on a much wider literature within international relations about “two-level games,” one level being domestic and the other international (Putnam 1988; Tsebelis 1990; Snyder 1991; Mayer 1992; Fearon 1994; Schultz 1998; Smith 1998; Milner and Rosendorff 1996, 1997; Partell and Palmer 1999). It also accords with recent scholarship that examines the two-level dynamics in continuing or ending enduring interstate rivalries (Colaresi 2004, 2005; McGinnis and Williams 2001; Mor 1997; Schultz 2005). In this way, the explanation provided in this book is more complete and thus more powerful than previous attempts to generalize about and predict war termination. Like any other social system, war termination is a given end state that can be reached through multiple causal pathways.5 The domestic coalition shift theory provides an alternative causal pathway to war termination beyond the standard Bayesian bargaining models—and accounts for more than a third of interstate war endings since 1862 and 41 percent since World War II. The prevalence of this alternative causal pathway suggests my theory provides greater explanatory power than other forthcoming scholarship that argues for the insignificance of domestic politics in war termination decision-making (Reiter 2009: chapter 1). For sake of comparison, Reiter’s alternative causal pathway—in which states pursue absolute victory to overcome credible commitment problems in the war termination process—accounts for only 22 percent of war endings in the post–World War II period.6 Thus, the domestic coalition shift theory, which includes and expands on the Bayesian model, greatly improves our understanding of war duration and war termination. d omest ic coalit ion shifts in overcoming obstacles to peace

The decision to go to war is a conscious policy choice. However, in order to win, leaders need to implement policies which prepare their militaries, their businessmen and their populations for the hardships and deprivations of war. They frequently need to demonize the enemy to whip up public support, mobilize their militaries and set their economies on a war footing. Enacting these policies is necessary for a state to win on the battlefield, but they make it extremely difficult to simply “turn off ” the decision to go to war. The very policies that might

10

the d omest ic p olit ics of war ter minat ion

make a state effective in preparing for and fighting a war are the ones that preclude an easy end to that war. Indeed, unless the resulting battlefield situation is so compelling that the outcome is obvious to all sides, leaders may find it virtually impossible to extricate their states from these “sticky” policies and the war itself. This book suggests three reasons why a war, once entrenched, is difficult to end. These reasons fall into one of three categories—preference obstacles, information obstacles and entrapment obstacles. • A preference obstacle occurs when leaders do not want to end the war. This happens when leaders receive private benefits from continuing the war, or when fighting the war helps leaders stay in power. • An information obstacle occurs when leaders do not know they should end the war. In this case, the Bayesian updating process does not occur because leaders (1) receive poor quality information, (2) do not necessarily share the same information, (3) use different indicators to assess the war or (4) have personal cognitive or (5) organizational biases that prevent them from using all available information. • An entrapment obstacle occurs when leaders want to end the war but cannot. This happens when leaders face political entrapment from a domestic constituency or external allies. These three obstacles to peace can occur in tandem, although they may not all be present at the same time or have an equally powerful effect. The longer a war goes on, the more the dynamics of “sticky” policies kick in. Because of these obstacles, leaders find it more difficult to shift policies and strategies related to fighting and ending the war as quickly as the Bayesian approach would suggest. In these longer wars, belligerents become “stuck” and a change in expectations is needed to produce a war-terminating bargaining space. This book argues a major source of such change is a shift in belligerents’ domestic governing coalitions. In other words, policies a state enacted to wage war effectively become sticky, and the state will often require a coalition shift to get them unstuck. A coalition shift allows different political actors—with different constituencies, political interests and assessments of the war—to take power. This, in turn, may lead to a shift in the war policy, from continuing it to ending it.7 In other words, domestic coalition shifts can help to overcome the obstacles to peace and thus permit the war to end. While most other empirical studies of interstate war duration and termination have focused only on wars that ended conclusively (Bennett and Stam 1996; Goemans 2000a; Reiter and Stam 2002; Reiter 2009),8 my model accepts the possibility of ambiguous war outcomes and theorizes explicitly about stalemates.

old baldy



Given the increased prevalence of stalemates in the modern era and the dearth of theory about ending these kinds of wars, my model broadens the literature’s current findings by explicitly including them. The longer the war goes on, the less likely the relative power among belligerents will be clearly illuminated on the battlefield. As a result, Bayesian updating is more difficult and belligerents are more likely to need domestic coalition shifts to overcome stickiness in expectations and achieve peace.9

Methods of Analysis To test causal hypotheses about how shifts in domestic governing coalitions influence war termination, this study uses two methods of analysis: detailed qualitative case studies and quantitative analysis. Using mixed methods helps to compensate for some of the weaknesses of the respective methodologies. The detailed case studies in this book demonstrate that my theory accurately identifies the causal dynamics that determine outcomes, while the quantitative analysis demonstrates that the patterns predicted by the theory actually occur over a broad class of cases. Therefore, the two methods provide different and complementary bases for causal inference (Bennett and George 2005: 208, also chapters 2 and 10). While quantitative analysis is useful for identifying correlations among variables to explain the duration of wars, it cannot fully address all of the causal mechanisms outlined in my argument. As Curtis Signorino (1999) and others have argued, the statistical models commonly used in international relations generally do not capture the structure of strategic interaction and interdependence very well. Because actors anticipate the decisions of other actors when they make their choices, the decisions between actors are interdependent and not independent, as many statistical estimators require. Moreover, statistical techniques can only measure actual observed events, not the unrealized potential outcomes that leaders considered but ultimately did not choose. Finally, many of my variables of interest—most notably, the three obstacles to peace—cannot be adequately proxied in a quantitative analysis. Process-tracing in qualitative case studies is particularly well suited to avoid these pitfalls (Van Evera 1997: 64–67; Bennett and George 2005: chapter 10; King, Keohane and Verba 1994: 226–28; George and McKeown 1985: 34–41; George 1979). There are four reasons why process-tracing is particularly useful in tandem with statistical analysis. First, the process-tracing method attempts to identify the intervening causal process between the independent variables and the outcome of a given dependent variable. Process tracing requires that intervening variables

12

the d omest ic p olit ics of war ter minat ion

be connected in particular ways, and this makes it a powerful tool for inference. In testing theories through process-tracing, all of the intervening steps in the empirical case must be as predicted by the hypotheses for the theory to pass. Indeed, process-tracing in one case can be enough to test the theory if its unique predictions are empirically validated. In other words, process tracing allows us to evaluate certain and unique predictions, which provide strong tests of a theory (Van Evera 1997: 31, 76; Bennett and George 2005: chapter 10).10 Second, process tracing is particularly important for generating and assessing evidence on causal mechanisms. Given my theory’s focus on stickiness in Bayesian updating, process-tracing is necessary for evaluating its causal mechanisms. To know whether the three obstacles to peace exist and a coalition shift is a cause of war termination, I need to (1) measure the obstacles directly, (2) assess how battlefield information (and the opposing side’s diplomatic behavior) gets interpreted by the domestic coalition, (3) show how the existing coalition’s decisions about war policy lag behind the war trends that would seem to warrant a war policy change and (4) show that coalition shifts provide the “shock” that brings Bayesian updating back into line with war trends. In short, process-tracing allows us to find out what decision-makers knew when, and whether and how that information was translated into a change in war policy. This minimizes the measurement error on key variables of interest. Third, process tracing allows us to compare competing explanations simultaneously. Although two explanations may predict the same outcome, the logic of the arguments may require different paths to that outcome. In fact, process tracing encourages us to be sensitive to the possibility of “equifinality”—the presence of alternative causal pathways. Only case studies allow a comparison detailed enough to differentiate between two explanations and evaluate their relative importance. As Andrew Bennett and Alexander George argue, “Case studies employing process tracing are particularly useful as a supplement in large-N statistical analyses, which are likely to overlook the possibility of equifinality and settle for a statement of a probabilistic finding regarding only one causal path at work” (2005: 215). Because war termination is characterized by equifinality, processtracing is indispensable for comparing alternative causal pathways in the cases. Finally, process tracing is particularly useful for obtaining an explanation for “deviant” or “outlier” cases that are not well explained by existing theories (Bennett and George 2005: 215–16; Becker 1998: 192–96). Process tracing of deviant cases offers an opportunity to develop contingent generalizations that identify the conditions under which alternative outcomes occur. To some degree, sticky wars and stalemated battlefields are “deviant cases” from the perspective of existing theories of war termination: while Bayesian models accept the updating

old baldy



process can take time, the reasons for the lag are under-theorized. Stalemates are also understudied in the empirical research about interstate war termination, as most empirical studies have focused on wars that ended conclusively.11 Therefore, process tracing within such “deviant cases” can help us enrich the general theory for explaining the process by which all wars end. In sum, process tracing can complement and ameliorate some of the methodological issues with quantitative analysis. Certainly, case studies about the decision-making process require a great deal of data, which creates practical limitations on how many case studies can be completed and limits the findings’ generalizability. Luckily, this weakness is directly addressed by the quantitative analysis, which bolsters the findings’ scope. histor ical case studies

This book examines the theory’s causal mechanisms in the three primary belligerents of the Korean War: the Soviet Union, the United States and China. I have limited the qualitative work to these “primary” belligerents, as they were the most important in making decisions about prosecuting and ending the war. (Chapter 3 will discuss why neither North nor South Korea qualifies as a “primary” belligerent.) Given the importance of the strategic interaction in understanding how an overlapping bargaining space develops in the war termination process, it was critical for me to examine belligerents in the same war. This allows me to control for other realpolitik variables and the strategic interaction among belligerents. Beyond the justification presented in the previous paragraph, there are five reasons for this case selection. First, and most importantly, these cases are strongly associated with the independent variables of interest—the three obstacles to peace and domestic coalition shifts. Such “extreme value” cases can teach us the most, because the variables’ effects should stand out sharply against the background of the case more clearly (Van Evera 1997: 79–80). The variables’ effects are also unlikely to be produced by measurement error or other causes. Such cases offer particularly strong tests because the theory’s predictions are certain and unique. Each of these three cases has a different predominant obstacle to peace, which allows me to explore the dynamics of each obstacle more carefully. In addition, coalition shifts in the three cases vary in terms of shift direction (hawkish, dovish and neutral) and the reason for the shift (exogenous and endogenous shifts).12 Moreover, because some of the belligerents experienced multiple coalition shifts, there is sufficient within-case variation. Thus, these cases provide ample evidence of and variation on the independent variables in my argument. Second, and relatedly, because of the extreme values on the independent variables of interest, these cases provide a “most likely” test for the domestic coalition

14

the d omest ic p olit ics of war ter minat ion

shift theory and a “least likely” test for the alternative standard bargaining models of war (King, Keohane and Verba 1994: 209–10; Eckstein 1975). Because the Korean War is a stalemated war, it meets the scope conditions for the theory, making it likely on a priori grounds that the case will support the theory. Therefore, it would be easy to discard my theory if it did not explain this case. In contrast, the Korean War is a hard case for standard bargaining models. Not only was there power symmetry between the two sides—the UN Command and the Communists—but the final armistice terms accepted in July 1953 were virtually the same as those proposed at the start of negotiations in July 1951. As the discussion of Old Baldy highlighted, why did the belligerents wait so long to make the peace given that the battlefield was stalemated throughout these two years while casualties continued to mount? Third, the three primary belligerents provide variation on regime type, an important variable in recent scholarship about war duration and war termination (Bennett and Stam 1996, 1998; Filson and Werner 2004; Fortna 2004b; Goemans 2000a; Reiter and Stam 2002; Slantchev 2004). As Chapter 2 will explain, the domestic coalition shift theory should apply to all regime types, because all leaders answer to some constituency. However, while regime type is not important for my argument, it is a variable of interest for comparing it to alternative arguments, especially Goemans’ (2000a). Indeed, these cases are useful for testing the relative power of competing theories, because as Part II will show, the theories often make opposite predictions (Van Evera 1997: 83). Fourth, by choosing case studies from the Korean War, I am making an empirical contribution to the literature by examining an important, yet relatively understudied war. The Korean War has been dubbed “the Forgotten War” for good reason. To the extent that the war receives any attention, scholars have generally focused on its dramatic first year—the war’s outbreak, the Inchon landing, the Chinese intervention and the Truman-MacArthur controversy. With the exception of Rosemary Foot’s work (1990), very little scholarship about the last two years of the war—the forgotten part of the forgotten war—exists. Moreover, very few studies have examined all of the belligerents together, taking account of the two-level dynamics that are central to my theory. For this reason, this book examines all of the major combatants in the war, even the secondary belligerents—North and South Korea. In other words, creating a satisfactory explanation of this historically significant, understudied case is in and of itself an important goal. Finally, while the Korean War is a “deviant case” from the perspective of the Bayesian approach, it resembles current situations that are interesting from a policy perspective. As Stephen Van Evera argues, a theory tested in a case that re-

old baldy



sembles current policy-problem cases will more often “travel” to the second case, and thus, “policy prescriptions deduced from the first case can more safely be applied to the second” (1997: 83). Obviously, stalemated wars have policy-relevance today, because of the US’ ongoing wars in Iraq and Afghanistan. More generally, recent scholarship on war termination documents a marked increase in interstate wars ending in stalemate since World War II (Dunnigan and Martel 1987; Fortna 2004b; Pillar 1983; Smith 1995). The Korean War also resembles current and future wars in that belligerents on both sides fought as part of international coalitions. Thus, examining this war provides traction for understanding which factors may be most relevant for explaining war termination in the future. quant itat ive analysis

While the case studies focus on the “deviant” case of a long stalemated war, the quantitative analysis tests the domestic coalition shift theory in all interstate wars since 1862, using two different datasets. I initially use war-level datasets—from Bennett and Stam (1996), Goemans (2000a) and the Archigos project (Goemans et al. 2009)—to establish the theory’s viability over a wider sample of wars, interstate wars from 1862 to 1990. Then, I use a new belligerent-level dataset, which I specifically constructed for this purpose, of the interstate wars from 1945 to 2006. The Archigos dataset allows me to test the effects of changes in heads of state in all interstate wars since 1862. Given the limitations of its coding—explained in Chapter 9—the Archigos data is suggestive but not overly conclusive. While we need to move to a finer-grained measure of coalition dynamics to truly test the model’s effects—one that not only measures multiple coalition shifts within belligerent states but also measures shifts’ effects more accurately over time—the supportive results obtained from analyzing these data increase our confidence in my theory’s robustness. Given the high degree of empirical detail necessary to code domestic governing coalition shifts, the data needed to test my argument empirically are quite extensive. Because information about domestic governing coalition composition and the timing of coalition shifts in wars after World War II is significantly more accessible, I selected this sub-sample of all interstate wars to conduct a more detailed probe of the argument. Theoretically, there should be no difference in the fundamental decision-making processes during wars before 1945, as the findings from the Archigos dataset demonstrate. However, there are some features of the post–World War II environment that make these wars stand out as a cohesive sub-sample for a more nuanced analysis of the domestic coalition shift theory. As already noted, scholars have observed a

16

the d omest ic p olit ics of war ter minat ion

marked increase in interstate wars ending in stalemate in the post–World War II period (Dunnigan and Martel 1987; Fortna 2004b; Pillar 1983; Smith 1995). This increase in stalemates may be driven by structural factors in the international system—the presence of nuclear weapons, the strength of the United Nations and its role in bringing interstate wars to settlement before a decisive victory and the dampening effect of bipolarity during the Cold War—as well as the global increase in democracies. Moreover, these wars occurred during the era of fast communications and round-the-clock news coverage. As a result, they are most similar in terms of strategy, doctrine and technologies to wars fought today. As the 1918 Signal Corps Manual argued, “. . . The art of war has not changed with the passage of years, but it is true that the science of war has changed enormously” (White 1918: 3; also Gray 2006). Thus, to examine which factors are most relevant for explaining war duration and termination in the future, it may be most useful to study recent wars, with the most theoretical and policy relevance to today.

Overview of the Book This book is divided into three parts. The next chapter, which completes Part I, lays out the book’s theory. After reviewing the existing literature about war termination, it explains the theory and presents hypotheses to be tested in subsequent sections of the book. Part II, divided into six chapters, traces the theory’s causal mechanisms in detail through the case studies of the Korean War. Chapter 3 briefly introduces readers to the timeline of the Korean War and explains why the Soviet Union, China and the United States were the most important belligerents in decisions about prosecuting and ending it. Chapter 3 also illustrates, with mini-case studies, North and South Korea’s insignificance in their respective side’s decisions about the war. In the process, it demonstrates how their domestic coalition shifts are unimportant for explaining the war’s ending. The next four chapters examine the argument’s causal mechanisms in more detail in the Soviet Union, United States and China. Chapter 4 traces Soviet decision-making during the war through the lens of the preference and information obstacles. The chapter argues Soviet leader Joseph Stalin did not want to end the war because he was benefiting from it without assuming many of its costs. Indeed, while moderates opposed this war policy, they were relatively powerless to change its course as long as Joseph Stalin remained in power. When Stalin died unexpectedly in March 1953, these leaders took control of the country and quickly enacted their long-held but dormant preference to end the war. Chapters 5 and 6 trace US decision-making during the war through the lens

old baldy



of the information and entrapment obstacles. These chapters explore the Truman Administration’s “NSC-68 mindset,” a set of assumptions about the US-Soviet relationship within which it perceived the Korean War. Fearing Soviet aggression worldwide, the Truman Administration used the Korean War to embark on a massive Cold War rearmament effort and justify its globalist agenda to Congressional and public constituencies. Likewise, believing it needed to stand firm on principles in Korea, it adopted a policy of voluntary repatriation of prisoners of war (POWs). The threat inflation that resulted from selling mobilization stoked hawkish constituencies for an all-out victory in Korea; hence, any efforts to move towards a negotiated settlement increased calls of Truman being “soft on Communism.” Moreover, fearing any escalation would provoke global war, the Truman Administration believed it could neither abandon the POW policy nor escalate the war to bring the Communists to terms. It was only with President Dwight Eisenhower’s inauguration in January 1953 that these obstacles were lifted and the United States could change course towards ending the war. Chapter 7 traces Chinese decision-making in the war through the lens of the entrapment obstacle. Unlike the domestic entrapment in the US case, Chinese leaders were trapped by an external ally—Stalin. This chapter argues that as the war’s domestic costs became apparent, the Chinese government’s policy goals shifted to modernizing and creating a planned economy. A domestic governing coalition shift in the summer of 1952 led the Chinese leaders to want to end the war by compromising on POW repatriation policy. However, these leaders also needed Stalin’s support for Chinese modernization and industrialization, and Stalin was unwilling to end the war. Thus, China was trapped until Stalin’s death. After these belligerent-level chapters, Chapter 8 brings attention to the strategic interaction between the belligerents in the war termination bargaining process. While some observers might view Stalin’s death in March 1953, and the resulting coalition shift in the Soviet Union, as solely responsible for breaking the negotiating logjam, this explanation is too simplistic. Instead, this chapter argues the timing of the war’s end depended upon the strategic interaction of all three domestic coalition shifts in the Soviet Union, the United States and China. This chapter extends the theory with two mechanisms by which domestic coalition shifts can affect the creation of an overlapping bargaining space, and it suggests these mechanisms are likely to operate in tandem when opposing sides both experience coalition shifts at about the same time. In other words, this chapter returns to the bargaining models of war and demonstrates why strategic interaction paired with analysis of domestic governing coalitions leads to the most complete explanation for why, when and how interstate war ends.

18

the d omest ic p olit ics of war ter minat ion

Part III tests the argument beyond the Korean War and teases out its theoretical and policy implications. Chapter 9 tests the argument with quantitative analysis, using war-level datasets of interstate wars from 1862 to 1990 and my newly constructed belligerent-level dataset of interstate wars since World War II. This chapter provides significant support for the hypotheses outlined in Chapter 2; it shows that including domestic coalition shifts in quantitative models of war duration and termination provides more explanatory leverage than existing models. This chapter also examines what effect the direction of the coalition shift (i.e. hawkish, dovish or neutral) has on war termination. The analysis in Chapter 9 enhances confidence that the patterns predicted by the theory actually hold over many cases, and it reaffirms that not all wars end by the Bayesian causal pathway that has dominated the bargaining model of war. Paired with the detailed Korean War cases, this analysis increases the generalizability of this book’s findings. Finally, Chapter 10 summarizes these results and draws out the larger theoretical and policy implications of the argument. This chapter addresses potential ways to extend the research, such as explicitly modeling third-party intervention and extending the analysis to civil wars. Moreover, while this book examines obstacles to peace, the underlying logic is obviously much broader. War termination is only one specific manifestation of a more general phenomenon—major policy change. Indeed, the shift from continuing a war to ending it can be conceived as the most difficult policy change imaginable. Thus, the chapter concludes with general policy recommendations, based on the research findings, for leaders making decisions about prosecuting and ending war—as well as leaders approaching any major policy change.

2

Domestic Coalition Shifts in War Termination

Ending war is a tough business. Historical evidence suggests that wars, once begun, are very difficult to stop. They generally last much longer and cost much more than belligerents expect at the outset. As a result, wars frequently last well beyond the point where a rational cost-benefit calculation would dictate they end. This chapter will try to show that there is a causal link between the difficulty of ending wars and the fact that they are started and ended by politicians. Ending wars requires settling with the enemy as well as at home. At a theoretical level, the question of war termination remains largely unanswered. While a growing literature about bargaining in war has made significant inroads in our understanding, the mechanisms that link domestic politics to the international bargains are still underspecified. Very few theories of war termination adequately encompass both levels of analysis in their explanations. In contrast, this chapter argues the longer the war, the harder it is to end it, because domestic obstacles to peace become institutionalized over time. By focusing on these obstacles, this chapter explains domestic-level impediments to the international bargains that end war. In this chapter, I present two pathways for interstate war termination, building on Bayesian models. As Bayesian models suggest, interstate wars can only end once all belligerents develop similar expectations about the war. Bayesian models suggest that a lag in updating can delay war termination, but they do not explain how states get beyond this lag—why some states get bogged down in updating their expectations while others can update very quickly. In contrast, I look more carefully at this lag and theorize about how states get beyond it. I outline three

20

the d omest ic p olit ics of war ter minat ion

obstacles to peace, as well as coalitional dynamics, which can lead governing coalitions to be unable to end the war—even when such a change is necessary or desirable. As a result, a governing coalition may become so immobile that ending the war is only possible by replacing the incumbent coalition entirely. Thus, while Bayesian models assume the change in expectations necessary to produce an overlapping bargaining space occurs when incumbent leaders change their minds, I argue that in many wars, such change in expectations results from a change in the foreign policy leadership itself. This chapter is divided into three sections and a conclusion. The first section synthesizes the existing war termination literature to provide the theoretical context. It highlights how my theory builds on bargaining models of war by adding domestic coalition shifts—a variable often mentioned but largely undeveloped in theories of war termination. The second section describes my approach, laying out the domestic obstacles to international bargains for peace and how those obstacles can be overcome with domestic coalition shifts. The third section examines what effect the direction of the coalition shift (e.g. hawkish, dovish or neutral) has on the theory.

The Theoretical Context Existing literature about war termination can be grouped into three categories— realpolitik, domestic politics and bargaining models. Of these, bargaining models of war have made the most significant inroads in our understanding. I build on this literature by refining the domestic mechanisms to explain why each side decides to come to the bargaining table. Specifically, I advance a key variable—the effect of domestic coalition shifts—that the literature has mostly ignored and remains largely underdeveloped. realp olit ik arguments: “ the vanquished makes peace ”

The oldest argument for ending war is power. Scholars in this group argue that a change in relative power among belligerents brings peace (Calahan 1944; von Clausewitz 1976; Randle 1973; Hobbs 1979; Beer and Mayer 1986; Timosheff 1965; Coser 1961; Kecskemeti 1958). Realpolitik arguments are by nature rationalist. They assume belligerents can make the necessary cost-benefit calculations about when to terminate a war. Because these scholars view war as having only two outcomes—victory or defeat—war ends when the loser gives up and accepts the more powerful winner’s demands. As H. A. Calahan noted in the first book to describe this argument: “War is pressed by the victor, but peace is made by the vanquished” (1944: 18).

d omest ic coalit ion shifts



However, this approach begs two crucial questions, answered by more recent bargaining models below. First, when is the losing party vanquished enough to give in to the winner? Realpolitik arguments cannot explain why war ends now and not at some other time. Second, why do winners stop then as well? These arguments assume the “winner” will not raise his demands once he realizes his advantage; thus, the argument would only seem to hold with unconditional surrender—when the winner has already been offered everything. Most wars obviously end long before that point. Indeed, most wars end without a clear-cut winner and loser. Lewis Coser recognized this explicitly: “Most conflicts end in compromises in which it is often quite hard to specify which side has gained relative advantage” (1961: 352; emphasis added). One reason is that victory may actually be conceived in many different ways,1 and a “winner” on one dimension may be a “loser” on another. Thus, it may be extremely difficult to determine who the “vanquished” is. d omest ic sent iments for peace

A second group of arguments looks inside belligerent states to explain why wars end. These scholars argue that a change in one side’s domestic sentiment about the war brings peace. They examine various mechanisms within the state to explain war termination, based on their view about which domestic institution or group monopolizes the policymaking process—public opinion, bureaucratic politics or elite behavior. While these scholars focus on domestic mechanisms for peace, they tend to ignore the other side completely. Nonetheless, they highlight how ending a war requires building and sustaining a domestic consensus for peace, and this process takes time. Public Opinion Scholars who have examined public opinion’s role in war termination have focused mostly on democratic regimes, and in particular, the United States. Research on public opinion has suggested that in democracies, public support for war declines monotonically with battle-deaths (Gartner and Myers 1995; Gartner and Segura 1998; Gartner, Segura and Wilkening 1997; Mueller 1985, 1994). While some scholars have challenged what Burk (1999) calls the “casualties hypothesis,” the basic logic underlying it—that the collective public is “rational” and will support the war if the benefits outweigh the costs—remains the dominant view (Randle 1973: 432–33; Page and Shapiro 1992; Wittkopf 1990; Larson 1996; Croco 2008; Jentleson 1992; Jentleson and Britton 1998; Feaver and Gelpi 2004; Gelpi, Feaver and Reifler 2005–6; Kull and Ramsay 2001; Kull and Destler 1999). In contrast, other scholars find little evidence that citizens make such cost/benefit calculations when evaluating a war. Instead, they argue the public cues on elite views about the war in order to evaluate war performance (Zaller 1992; Berinsky 2007).

22

the d omest ic p olit ics of war ter minat ion

Alternatively, they argue public opinion is prone to elite manipulation (Snyder 1991; Mansfield and Snyder 1995; McGinnis and Williams 2001; Mor 1997; Colaresi 2005; Gibbs 1995; Baum 2003: 44–45; Johnson and Tierney 2006: 67–75). Because leaders, elites and the media can manipulate public opinion, it may be this manipulation that has stronger explanatory leverage for explaining war termination than public opinion itself. Bureaucratic Politics Scholars who use bureaucratic politics to explain war termination argue that the struggle to end war is subordinated to a country’s internal politics (Allison 1969: 717–18; Halperin 1970; Sigal 1988; Rose 1994). Each domestic group has different interests that influence its decisions regarding war termination. For example, the military services view war termination from the perspective of maintaining their autonomy and growing their budgets, while civilian political leaders are more concerned with post-war domestic political implications. In contrast, the general staff in the nation’s capital will view its role as compromising between military and civilian views (Halperin 1970). As Leon Sigal argues, officials want “to end the war in a way that best serves their organizational interests. The stands they take on the war abroad depend on the stakes they see in the battle at home” (1988: 23). This perspective highlights that the process of ending a war can often create compromise decisions that are suboptimal and slow to develop. Elite Behavior Scholars who argue for the primacy of elite behavior in explaining war termination assume that governing elites want to remain in power (Rothstein 1970; Iklé 1991; Bracken 1992; Armstrong 1994; Smith 1995; Mendelson 1998). In general, these scholars argue that elite disagreement about war policy leads to war’s end. These arguments link directly to the role of public opinion, because elite disagreement may catalyze public interest in ending war, as well. The most influential scholar in this group is Fred Iklé (1991), who classifies elites as “hawks” and “doves” struggling to influence their state’s war policies. Iklé attributes the difficulty of war termination to the gross asymmetry in wartime political power between hawks and doves. Hawks have a far easier time blocking any settlement than do doves trying to achieve one. However, beyond asserting that hawks may need to leave the government before the state can settle, Iklé provides no mechanism for understanding when and why the doves finally overrule the hawks. Overall, the domestic politics approach leaves three questions unanswered. First, what type of domestic sentiment change is enough to bring peace? Second, when does the change in domestic sentiment have to happen? Third, can a change in domestic sentiment in only one state be sufficient to end the war? In

d omest ic coalit ion shifts



general, these arguments suggest some important variables for explaining war termination, but by only looking within one side, they generally ignore the strategic interaction inherent in settling with the enemy. t wo- or more-sided bargains for peace

A third group of scholars argues that a change in the international bargaining space brings peace (Wittman 1979; Fearon 1995; Schelling 1980; Morrow 1989; Pillar 1983; Wagner 1993, 2000; Smith 1995; Craig and George 1995; Werner 1998; Bennett and Stam 1998; Filson and Werner 2002, 2004; Goemans 2000a, b; Powell 2002, 2004; Reiter 2003, 2009; Slantchev 2003, 2004; Smith and Stam 2004). These rationalist models highlight the interactive nature of war, that it takes two or more sides to end it. Moreover, they address the theoretical problem of enforcement in an anarchic realm. They explain how and why the stronger side makes a credible commitment not to raise further demands and exploit his bargaining advantage in the future.2 Most bargaining models of war focus on information dynamics. In this approach, war is caused by uncertainty about the distribution of power between the belligerents or the costs of fighting, and the war itself is an important source of information for overcoming this uncertainty (Slantchev 2003, 2004; Powell 2002, 2004; Filson and Werner 2002, 2004; Morrow 1989). Because states have incentives to conceal private information about their ability to wage war, both sides may be optimistic at the same time and lead themselves into war. However, at least one belligerent must be wrong about these estimates. Wars end when both sides’ expectations about military victory converge enough to create an overlapping bargaining space—what Slantchev (2003, 2004) calls the “principle of convergence.” Belligerents continuously update their expectations throughout the war, based on observed battle outcomes, rejected diplomatic offers and unreasonable demands from the other side. Many bargaining models conceive of this updating process in mathematical terms, in accordance with Bayes’ rule (Morrow 1994: chapter 6). Given the uncertainty inherent in war, battlefield outcomes and the adversary’s diplomatic behavior help belligerents understand their adversary’s “type.” For example, winning a battle will lead a belligerent to believe he is more powerful than the other side, causing him to incrementally increase his demands. In contrast, a battlefield defeat will lead him to incrementally lower his demands. Similarly, belligerents will assume weaker (or cost-sensitive) opponents will settle sooner and accept worse terms than stronger (or cost-tolerant) ones. Accordingly, belligerents “screen” their adversary’s type through a diplomatic offer designed so that weaker opponents would accept it, while stronger ones would

24

the d omest ic p olit ics of war ter minat ion

reject it. A rejected offer will lead a belligerent to believe he is facing a “stronger type” and thus lead him to incrementally decrease his demands (Slantchev 2004; Filson and Werner 2002; Powell 2004). In practice, this Bayesian updating process may not be as precise as these models predict, and information dynamics cannot provide a complete account of war termination behavior. While Bayesian models accept that this updating process can take time, the reasons for this “lag” are under-theorized. Gartner (1997) argues that states may choose to change their military strategy rather than offer diplomatic concessions, when faced with unpleasant news from the battlefield. Thus, the information updating is not immediately reflected in bargain-converging behavior. Reiter (2009: chapter 11) posits that credible commitment fears may delay war termination, even as uncertainty is being reduced. More specifically, he argues that fears of adversary non-compliance in an agreement, fears of signaling weakness to the other side, and “leader patience” can account for the inefficiency in Bayesian updating. Surprisingly, only one scholar has theorized about domestic politics as a source of this inefficiency, even as Bayesian formal theorists have flagged it as a possible cause. For example, Darren Filson and Suzanne Werner admit their model assumes an attacker can “effortlessly revise upward or downwards her demands,” but a “leader responsive to domestic political concerns, however, is likely much less flexible” (2002: 832). The domestic politics version of the Bayesian argument filters battlefield conditions through cost-sensitive democratic constituencies, who then pressure leaders to change war policy. In turn, this pressure on the leaders causes them to change their expectations about the war (Filson and Werner 2004; Bennett and Stam 1998; Slantchev 2004). However, this argument does not model the lag in the Bayesian updating process that we often see empirically. In contrast, Hein Goemans (2000a) begins to account for lags in Bayesian updating, but only in losing “semi-repressive and moderately exclusionary” regimes, i.e. oligarchies. Goemans argues oligarchs’ expectations about personal punishment after the war serve as a brake on updating their expectations during it. If accepting settlement terms appears to have nasty personal consequences, these leaders as individuals can rationally prefer to continue fighting if there is a chance they can get terms that prevent their punishment. Thus, losing oligarchs can prevent the state from updating war expectations and creating an overlapping bargaining space to end it. While the inclusion of regime type is an improvement on bargaining models that exclude domestic politics entirely, this emphasis focuses attention on the differences between regime types and away from variation within regime types. In the process, such arguments conceal very important domestic political expla-

d omest ic coalit ion shifts



nations for lags in Bayesian updating. For example, by treating all democrats as similarly cost-sensitive and all oligarchs as similarly afraid of political survival, these arguments ignore that different leaders within each regime type face different constraints, based on their governing coalition’s particular obstacles to peace. In other words, it is not appropriate to assume leaders of each regime type will all behave the same way.3 By ignoring variation among leaders within the same regime type, these arguments miss domestic constraints that can explain lags in Bayesian updating. They also ignore the effect of leadership changes in overcoming such lags. In sum, given their ability to account for strategic interaction, I am not rethinking the underlying logic and information dynamics of Bayesian bargaining models. Instead, I build on them by elucidating the underspecified causal mechanism about when and how the overlapping bargaining space opens up. As Robert Powell notes, future efforts to use informational asymmetries to explain war “have to explain why it takes so long to resolve the informational asymmetry” (2004: 345). Accordingly, I look more carefully at the lag in the Bayesian updating process and theorize about how states get beyond it. While Bayesian models assume this change in expectations occurs when incumbent leaders change their minds, my research suggests that in many wars, this change in expectations results from a change in the foreign policy leadership itself. In the process, my theory addresses two other shortcomings in these models. First, most bargaining models do not allow for actors’ goals or interests to evolve during the settlement process. If preference is given, why do actors’ sentiments change during war? Instead, as I will argue, preferences during war appear to move over time in response to several issues, including a state’s “sunk costs” and the information gained from the process of fighting and negotiating.4 Moreover, new leaders often come to power with different preferences and policy goals, which can lead to a new calculation of the costs and benefits of continuing the war. For example, Sarah Croco (2008) argues that “culpable” leaders—those whom the public can clearly link to the decision to go to war—face “settlement costs,” while “non-culpable” leaders can distance themselves from the original war decision and thereby become immune to such costs. Thus, culpable and nonculpable leaders face different utility functions and preferences about continuing or ending the war. Second, and more importantly, my model includes and addresses an entire class of wars—protracted stalemates—that most bargaining models cannot explain and typically exclude from analysis. As Bayesian models indicate, leaders update their preferences as information is revealed through fighting, as long as the battlefield provides clear information. In stalemated battlefield situations,

26

the d omest ic p olit ics of war ter minat ion

this information may be non-existent, limited or ambiguous or leaders may be unable to act on clear information. These wars are empirically too common to be simply ignored as irrelevant: recent scholarship on war termination documents an increasing number of stalemates in the post–World War II period (Pillar 1983; Dunnigan and Martel 1987; Smith 1995; Fortna 2004b). My model acknowledges this fact and accepts the possibility of ambiguous outcomes to war. Indeed, this ambiguity is a key reason why existing leaders have difficulty changing expectations and often require a leadership change to end the war. Thus, my model adds breadth to the current literature by theorizing more explicitly about lags in the updating process that can lead to protracted stalemates.

Domestic Coalition Shifts in War Termination My model focuses on domestic coalition shifts, an intriguing domestic political variable that has been noted by scholars from all three theoretical perspectives (Calahan 1944: 109; Kecskemeti 1958; Rothstein 1970: 74; Handel 1978: 26; Allison 1969: 718; Iklé 1991; Craig and George 1995: 232–33; Bueno de Mesquita et al. 1992; Smith 1995: 106; Randle 1973: 444; May 1973: chapter 5; Staudenmaier 1985). Similarly, Bennett (1997) and Goertz and Diehl (1995) examine regime change in ending enduring interstate rivalries. Despite all of these references—many quoted in Chapter 1—domestic coalition shifts have thus far not been addressed in a systematic fashion. The fact that so many scholars, from so many theoretical perspectives, have mentioned regime change preceding war termination suggests it is an important component in the process. These references about domestic coalition shifts point to a critical, yet underdeveloped, variable for explaining when and why wars end. None of these scholars have examined the variable and its causal mechanisms in a rigorous and generalizable manner as this book does.5 In contrast, I make the domestic coalition shift hypothesis explicit by operationalizing and tracing its dynamics. I retain the idea that war termination is a bargaining process between two interacting sides, but I refine the domestic mechanisms to explain why each side decides to come to the bargaining table. In this way, my explanation is more complete, and thus more powerful, than previous attempts to generalize about and predict war termination. t wo-level games in st icky wars

As with most other major decisions in the international realm, ending a war is a two-level game (Putnam 1988; Tsebelis 1990; Schelling 1980; Mayer 1992; Fearon 1994; Schultz 1998; Smith 1998; Milner and Rosendorff 1996; Partell and Palmer 1999).6 Leaders of the belligerent states must not only settle with the opposite

d omest ic coalit ion shifts



side, but they must also “settle” with their domestic constituencies at home or else be removed from their leadership position. Bargaining at the international level affects bargaining at the domestic level, and vice versa, if only to constrain the “win set” of acceptable settlement terms. Therefore, to be persuasive, an account of the war termination process must look at both levels and the interaction between them to answer the questions of how, when and why wars end. Bayesian models of war termination have mostly focused on the international level game. When incorporating domestic politics, the Bayesian model suggests that battlefield information leads domestic constituencies to reevaluate the war and pressure their leaders to change war policy. This constituent pressure, along with the leaders’ own reevaluation of the war, leads the incumbent leaders to change their war expectations, thus producing an overlapping bargaining space and eventually war termination. While Bayesian models suggest that a lag in updating can delay war termination, they do not elaborate on how states get beyond this lag. By fleshing out the domestic-level game and developing an alternative causal pathway of domestic coalition shifts, this chapter provides an account for Bayesian updating in protracted conflict. Bayesian models appear to best explain wars with a large power imbalance between the opposing sides. In these wars, updating occurs rather quickly, because the battlefield compellingly illustrates the belligerents’ relative strength. As a result, such wars are often short. In contrast, Bayesian models are less effective for explaining longer wars, where realpolitik factors are not enough to manifest clear battlefield outcomes. As Branislav Slantchev (2004: 816) argues, “parity in observable capabilities makes informational problems more severe,” which can lengthen the war. Wars can be long, for example, because the battlefield becomes stalemated, belligerents lack commitment credibility, relative power is ambiguous, or the issues at stake are indivisible or highly salient. In contrast to the straightforward updating behavior usually observed in short wars, long wars tend to be characterized by “sticky” policies. Leaders find it more difficult—for the reasons discussed below—to shift policies and strategies related to prosecuting and ending the war as quickly as the Bayesian approach would suggest. Here, war does not necessarily end when the decision-maker wants it to or should want it to. Rather, belligerents invest political capital in the policies of prosecuting the war and these policies become stuck in a positive feedback loop. Thus, in longer wars, failure of realpolitik factors to produce a bargaining space and political capital sunk costs make it harder and harder to end it. Because wars can become stuck, like Vucinich and Teachman (1993) and Slantchev (2004), I argue that war has negative duration dependence—the longer the war, the harder it is to end it.7 For this reason, the probability that a war

28

the d omest ic p olit ics of war ter minat ion

will end after its second month is far higher than it will after its 24th month. My argument contrasts with those that suggest war has no duration dependence (Bennett and Stam 1996) or positive duration dependence (Filson and Werner 2002). However, Filson and Werner (2002) exclude domestic politics from their formal model entirely and do not test the model empirically. Bennett and Stam (1996) only include aggregated regime type variables that average, for each war, all belligerents’ scores for “institutionalized democracy” and “competitiveness of participation.” In other words, their models do not capture the sticky domestic dynamics that lead war to have negative duration dependence. H realpolitik: The greater the power imbalance between the two sides, the quicker the war. H negative duration dependence: The longer the war has already lasted, the harder it is to end. coalit ion p olit ics

The decision to enter, prosecute or end a war is a political decision—the vector product of many different domestic interests. Thus, rather than characterize each belligerent state as a unitary actor, it is useful to model the actor in these political processes as a domestic governing coalition. A coalition has been defined as “the joint use of resources to determine the outcome of a mixed-motive situation involving more than two units” (Gamson 1964: 5). This definition highlights that coalition policymaking is a collective activity that exercises power by organizing and applying resources to achieve certain ends (Hinckley 1981: 6). Coalitions form the backbone of most decisions and actions of the state; indeed, it is coalitions that make human activity “political.” All leaders must create and maintain coalitions to accomplish anything productive, from raising taxes to ending war. Coalitions form when different actors join together to achieve a common purpose. Minimum winning coalitions—those with the fewest members necessary to control the policymaking process—are preferable, because they require fewer resources and fewer compromises (Riker 1962). However, many factors can lead to larger-than-minimum winning coalitions, including ideology (Axelrod 1970; De Swaan 1973; Hagan 1993); domestic pressures for inclusivity (Hula 1999); uncertainty (Koehler 1972); and crisis situations, like war (Hagan et al. 2001: 203; Gamson 1964). Indeed, the greater the environmental hostility, the more internally cohesive the coalition is (Bennett 1980). The converse should also be true: during a “limited” war, the less threatened a nation feels, the less cohesive its governing coalition will be. A governing coalition’s ability to commit to or change policy depends heav-

d omest ic coalit ion shifts



ily on the effective number of veto players (Cox and McCubbins 2001; Tsebelis 1995, 2002; Hagan et al. 2001). Stephan Haggard and Mathew McCubbins define a veto player as “a person, group or faction who, through their control of an office, post, or branch of government, can reject any proposed changes to existing policy” (2001: 5). As the number of veto players increases, the less likely a single individual or faction can control policy and the more likely divergent interests become involved in political decision-making.8 The more veto players and the greater the distance between their policy positions, the more difficult it can be to change policy. In contrast, sustaining a policy will be much easier (Tsebelis 1995, 2002; Cox and McCubbins 2001). The distributive effects of policy depend on the number of veto players, as well. The more veto players, the more policies enacted will benefit targeted groups or constituencies, instead of the general public. This occurs when veto players are able to demand—and receive—side payments in the form of narrowly targeted policies (Cox and McCubbins 2001). Only groups that are both necessary to a coalition’s success and in a position not to cooperate can demand and receive such side payments (Lamborn 1991: 63). These dynamics can produce two pernicious policy results. The first is “logrolling,” where each coalition member or group gets what it most wants in return for tolerating the unpleasant effects of policies backed by other coalition members (Snyder 1991: 44; Riker and Brams 1973; Hagan et al. 2001: 176; Snyder and Diesing 1977: 351). The second is “squaring the circle,” where coalitions explain away antithetical policy positions. Edward Mansfield and Jack Snyder argue this can lead coalitions to sweep “tough tradeoffs under the rug, pretending that contradictory policies actually make sense or cannot be avoided” (1995: 32). Alternatively, the coalition can blame other states for the policy contradictions it brought upon itself. These coalitional dynamics help explain policy inertia and highlight the obvious relationship between policy stability and domestic coalition shifts. When a policy is functioning because of a coalition agreement, changing the policy requires renegotiating that agreement (Snyder and Diesing 1977: 351). Likewise, the less control a coalition member has over policy choice and implementation, the less willing he is to take political risks to change it. Thus, larger coalitions with more veto players are unlikely to take political risks to change policy (Lamborn 1991: 63–66). For these reasons, policy stability can lead governments to be unable to change the status quo, even when such changes are necessary or desirable. In fact, a government with policy stability may become so immobile that policy change is only possible by replacing the incumbent government (Tsebelis 1995: 321).

30

the d omest ic p olit ics of war ter minat ion

d omest ic gover ning coalit ions

All states are governed by a “domestic governing coalition,” i.e. the elite foreign policy decision-making group in each state. I adapt Margaret Hermann’s concept of an “authoritative decision unit,” as the actors who “if they agree, have both the ability to commit the resources of the government in foreign affairs and the power to prevent other entities within the government from overtly reversing their position” (2001: 56).9 The important consideration here is not just a member’s vote but the resources he can bring to the coalition, such as his official position, personal prestige, recognized expertise and influence within the government. Coalition-building inside the governing coalition recognizes that making foreign policy involves many actors, including contending leadership factions, executive bureaucracies and separate institutions like the legislature and the military. Where there are disagreements among these actors, the resulting policy will reflect the agreements and side-payments that were necessary to gain the policy’s approval (Hagan 1993; Snyder and Diesing 1977; Salmore and Salmore 1978; Steinbruner 1974). In addition, a domestic governing coalition’s members are constrained by broader societal interest groups and other political “patrons,” such as (1) dissenting actors in the regime’s ruling group or party, (2) other political parties in the legislature, (3) military and paramilitary groups, (4) groups reflecting regional or ethnic interests, (5) economic sectors, (6) military-industrial-scientific complexes, (7) public opinion, (8) the media and (9) other interest groups and non-governmental organizations (Snyder 1991; Mansfield and Snyder 1995; Hagan 1993; Fordham 1998). Bueno de Mesquita and his colleagues (1999, 2004) argue that the size of these power-base groups has a strong causal effect on the regime’s decisions about war. Who these constituencies are will vary by regime type: public opinion, the media and opposition groups may have more effect in democracies, while the military, party apparatus or some external ally may be the relevant audience in authoritarian regimes (Peceny et al. 2002).10 From the perspective of the governing coalition members, the primary dynamic in these broader relationships is retaining the support base to remain in power. Thus, coalition-building at this level focuses on maximizing political support, blocking opposition that may be competing for political dominance, and if necessary, adjusting policy to make it compatible with domestic realities. All leaders will depend upon some constituency to stay in power, and different coalition members may answer to different constituencies. Even apparent dictators can be removed if they no longer promote the interests of their selectorate (Morgan and Campbell 1991).11 With an issue as salient as war, all leaders must worry about domestic politics and be careful not to become radically out of sync with their constituency.

d omest ic coalit ion shifts



In this book, I concentrate more on the first kind of coalition-building, within the governing coalition. Leaders are more sensitive to challenges from within the regime than from the wider society, for three reasons. First, the political power and authority necessary to allocate resources and implement policy reside in the governing coalition. Second, dissenting coalition members hold veto power— they are directly able to block policy initiatives, while the external opposition’s veto power is more diffuse. Even if some internal veto players do not actually participate in particular decisions, other coalition members must take care not to alienate them, because they are essential to the coalition retaining power. Finally, coalitions with permanent divisions or factional politics are much more vulnerable to external opposition. For example, actors in a divided governing coalition may try to strengthen their power by aligning with important external actors, thereby increasing the latter’s influence over policy (Salmore and Salmore 1978; Snyder and Diesing 1977; Hagan 1993). I follow Joe Hagan et al.’s (2001) framework for coalition decision rules, the general procedures and norms that coalition members recognize as structuring their interactions. Decision rules clarify how the group makes decisions and stipulate what constitutes an authoritative consensus. Hagan and his colleagues differentiate three kinds of coalition decision rules. The first (rare) category (“unit veto model”) occurs when political necessity or formal voting rules require unanimity among all members. Because any single actor can block the initiatives of all others, the possibility of side payments and the existence of a shared policy orientation are important mediating variables. The second category (“anarchy model”) occurs in coalitions with few, if any, established rules or procedures for making decisions. Here, important mediating variables are predominant actors or actors with exclusive control over resources necessary to implement policy— such as a military’s monopoly over the use of force. Between these two extremes is the third category (“minimum connected winning coalition”). This model applies when the governing coalition is regulated by established voting rules that permit a majority to agree on and implement a particular course of action. Key mediating variables here include pivotal actors and the willingness among coalition members to bargain and logroll. d omest ic coalit ion shifts

As explained above, coalition-building dynamics are an important cause of policy stability, and a shift in the domestic governing coalition may help to overcome policy inertia. I define a “domestic coalition shift” as either (1) a consequential change in the identity of the decision-makers or (2) a substantive change in the type of government. The first is similar to Hagan’s definition of a significant

32

the d omest ic p olit ics of war ter minat ion

regime change, when “one set of leaders is replaced by another, even though the state’s structures remain fixed and the overall structure of the political system is unaltered” (1994: 143). Specifically, such changes in major decision-makers include changes in the actual or nominal head of state, cabinet membership, political parties in a parliamentary governing coalition and junta membership. Notably, this definition excludes changes that involve only military leaders, unless a military leader also holds a cabinet-level post or membership in the ruling junta (and thus serves in a political role in the decision-making process). The latter definition resembles the comparativists’ “regime change,” i.e. a change in the type of regime. Either type of shift can produce a significant change in policies as new members bring new resources, styles, perspectives, interests and preferences to the leadership coalition. Regime type may influence the baseline frequency of these coalition shifts. It is difficult to know exactly how pre-existing election cycles will line up with the timeline of a war, and new elections may not always lead to a new coalition (e.g. the same governing coalition may be reelected). Certainly, democracies are more likely to experience relatively frequent coalition shifts by nature of their institutional structure. Recent work about elections and the use of force in the United States suggests the absence of an electoral cycle effect in US use of force decisions, except when the economy is performing poorly (Gowa 1998; Fordham 2002). However, more broadly, Kurt Gaubatz (1991, 1999) shows that both presidential and parliamentary democracies usually get into their wars early in their electoral cycle, when electoral pressures are lowest. Both findings imply that democracy is not the underlying mechanism driving coalition shifts during war.12 Certainly, regime type appears to have some effect on war duration. For example, Dan Reiter and Allan Stam (2002) argue that democracies are associated with short wars, because they choose to participate in wars they expect to be short (“the democratic selection effect”). Moreover, as many authors have noted since Kant, democratic leaders are more sensitive to the costs of war than other regime types (Bennett and Stam 1996, 1998; Bueno de Mesquita and Lalman 1992; Doyle 1983; Filson and Werner 2004; Gaubatz 1991, 1999; Levi 1997; Maoz and Russett 1993; Reiter and Stam 1998, 2002; Schultz 1999; Stam 1996; Wagner 2000). Research on public opinion has suggested that in democracies, public support for the war declines monotonically with battle-deaths (Mueller 1985, 1994; Gartner and Myers 1995; Gartner and Segura 1998; Gartner, Segura and Wilkening 1997). The dominant view is that the collective public is casualty-sensitive and will support war only as long as the benefits outweigh the costs (Page and Shapiro 1992; Wittkopf 1990; Larson 1996).13 This implies that democracies fight shorter wars because as casualties and other war costs mount, democratic governments be-

d omest ic coalit ion shifts



come more likely to offer concessions in reaction to the war’s waning popularity (Bennett and Stam 1998; Filson and Werner 2004; Slantchev 2004). As Reiter and Stam note, “public consent is like an hourglass: when war starts, the hourglass flips over and democratic leaders know they must win before the sand runs out” (2002: 178). In other words, there are two arguments for why democracies should be associated with shorter wars. The “selection effect” stipulates that democracies choose to fight only those wars they expect will be short. The “cost sensitivity” argument stipulates that democracies settle more quickly than other regimes in response to the war’s waning popularity. If we assume that the democracy variable is capturing both arguments, then including coalition shifts allows us to begin teasing these two mechanisms apart. Why? Because coalition shifts are associated with leaders’ responses to cost-sensitivity, but coalition shifts are not related to the selection effect. The selection effect happens long before the war begins and thus has no relationship to coalition shifts during the war. In contrast, one way democratic leaders respond to cost sensitivity is through coalition shifts. Changing the composition of the government can both signal to the public that it is “doing something” and put in place a government capable of enacting policies that will staunch the costs of war. Because coalition shifts are silent about the selection effect and only capture one aspect of cost-sensitivity, I still expect democracy to be significantly linked to shorter wars. However, because coalition shifts can partially explain cost-sensitivity, a model that includes both regime type and coalition shifts will be better specified than one with regime type alone. Indeed, a model that includes both variables can help us weigh the relative importance of the two mechanisms for explaining why democracies are associated with short wars. Accordingly, the regime type variable should no longer capture the effects of the coalition shift mechanism, so I postulate that it should have less explanatory power when coalition shifts are included in the model. H regime type: Regime type should have less explanatory power when coalition shifts are included in models of war duration and termination. st icky wars and obstacles to peace

There are many reasons why a war, once entrenched, is difficult to end. I argue that these reasons fall into one of three categories—preference obstacles, information obstacles or entrapment obstacles. Essentially, domestic leaders do not want the war to end because the benefits outweigh the costs, they do not know it is in their interest to end the war when the costs start to outweigh the benefits or they want to end the war, but cannot. These three obstacles can occur in tandem,

34

the d omest ic p olit ics of war ter minat ion

although they may not all be present at the same time or have an equally powerful effect. As will be explained below, the dynamics of the domestic governing coalition itself exacerbate the obstacles as well. These obstacles make it more likely belligerents will need a domestic coalition shift before the war can end. By removing these obstacles, domestic coalition shifts may produce the necessary conditions for war to end. The preference obstacle is already well-documented in bargaining models of war; even in rationalist models, if decision-makers do not want to end the war, a bargaining space cannot develop. In contrast, the information and entrapment obstacles interfere with the Bayesian updating critical to making decisions about continuing or ending a war. Preference Obstacles With preference obstacles, domestic coalition members may not want to end the war, for four reasons. First, leaders’ personal stakes may be very high, because their personal reputation, domestic political standing or physical or financial security is attached to continuing the war. If accepting the terms of settlement appears to have nasty personal consequences, leaders can prefer to continue fighting if there is a chance they can get terms that prevent personal punishment—such as losing office, exile or death (Goemans 2000a; Croco 2008). Alternatively, most of the benefits of war—including career advancement, protection from foreign economic competition, operational autonomy, larger budget slices or boosted revenue streams—are disproportionately concentrated in specific groups, such as the military and military-industrial-scientific complexes (Mansfield and Snyder 1995: 25; Snyder 1991: 32–35, 49–52; Davis and Huttenback 1986). Coalition members receiving such concentrated benefits often distribute the war’s costs (taxes and casualties) widely across the population. By pushing war costs off onto others while garnering concentrated perks for their constituents, these coalition members may have no interest in ending the war.14 Second, and relatedly, fighting the war may help leaders stay in power, especially when non-constituents are the ones directly bearing the war’s costs. Such “diversionary wars” can occur when the state is experiencing democratization or leaders face domestic political problems, such as unrest or economic downturn. In these situations, the governing coalition may want to continue fighting to enhance its legitimacy, demonstrate strong leadership, deflect attention away from divisive domestic problems or discredit domestic opponents (Levy 1988, 1989; Hagan 1993; Mansfield and Snyder 1995; Richards et al. 1993; Smith 1996). In other words, the latent “function” of fighting an external adversary is domestic social and political solidarity (Simmel 1898; Aho 1994). The Argentine junta’s decision to fight the Falklands War exemplifies this dynamic. Third, leaders may perceive that the costs of military inaction are higher than

d omest ic coalit ion shifts



the probable costs of continuing the war, such as when the country has been invaded and the only alternative to fighting is losing sovereignty. If the adversary has such high aims, leaders may believe fighting is the only option. Relatedly, although examples seem rare, international considerations could lead coalition members to want to continue fighting, even if they expect to lose, to impress third states with their level of resolve (Jervis 1988: 677). Finally, continuing the war may help leaders in a broader conflict by providing information about the adversary’s military or draining his resources in a tangential fight. Stalin’s motives for continuing the Korean War are an example of this dynamic, as Chapter 4 will explain. Preference obstacles are most problematic when the governing coalition includes members with an unwillingness or psychological inability to criticize their colleagues who hold the hawkish position. These dynamics are actually information obstacles, operating in tandem with the preference obstacle. In coalitions with a power imbalance, where potential critics are less powerful than the hawks, they are often unwilling to express their criticism for fear of perceived penalties that could result—ranging from humiliation to death (Fisher 1969; Schelling 1982; Smith 1995; Iklé 1991: 60–61). The fear of being branded a “traitor” can lead coalition members to falsify their true war-terminating preference in order to preserve their jobs (Kuran 1997; Sigal 1988: 29). Thus, they put their personal preferences above ending the war or changing war policy. The less control a coalition member has over policy choice and implementation, the less willing he is to take personal political risks to change it (Lamborn 1991: 63–66). In contrast, in coalitions with relative power balance, potential critics may be psychologically unable to express their criticism. This resembles Irving Janis’ “groupthink” phenomenon, where “subtle constraints, which the leader may reinforce inadvertently, prevent a member from fully exercising his critical powers and from openly expressing doubts when most others in the group have reached a consensus” (1972: 3). The preference obstacle can also occur in situations where potential critics are willing to speak out, but they lack the necessary power to overrule the hawks whose preferences are creating the obstacle. In these cases, potential critics inside the coalition may be willing to share their criticism, but more powerful hawks are unwilling to listen and may take action to prevent such criticism from arising again. Unlike the unwillingness described above, which is based on coalition members’ fears about expected penalties, here coalition members actually suffer for their dissent. Certainly, the two are related: potential critics are more likely to suppress their doubts and dissent after a leader has reacted negatively to criticism by humiliating, chastising, exiling or killing some of the critics (Smith 1995).

36

the d omest ic p olit ics of war ter minat ion

Information Obstacles With information obstacles, domestic coalition members may not know that they should end the war. In this case, the rational updating process does not occur, because they (1) receive poor quality information, (2) do not necessarily share the same information, (3) use different indicators to assess the war or (4) have individual or (5) organizational biases that prevent them from using all available information. Information obstacle dynamics contrast to the assumptions of Bayesian updating models, reviewed above. Poor Quality Information. First, coalition members may receive information about the war that is ambiguous, incomplete, biased, conflicting or faulty (von Clausewitz 1976: 191; Iklé 1991). Poor-quality information is common during war, even as leaders are bombarded with data—especially in the era of the internet and instantaneous, round-the-clock news coverage. Although reliable intelligence collection and analysis is difficult under any circumstances, it can be near impossible during war. For example, ambiguous information can result when (1) the belligerents’ advantages and disadvantages counteract each other, (2) there are many strategic options still available for decision-makers to try, in the hopes that another strategy might work better and (3) the battlefield itself is ambiguous, such as when a state is winning its air campaign but losing its ground battles. Poor quality information interacts with other aspects of “friction” during war, including unpredictability in the strategic interaction with the enemy and coordination with allies. For example, adversaries often send ambiguous signals to each other during the process of reaching a negotiated settlement (Quester 1970; Jervis 1976, 1989; Smith 1995; Snyder and Diesing 1977). Coordination with allies can also add layers of complexity, including different languages and cultures and the lack of interoperability in battlefield command and control systems (Bensahel 2007). Given such ambiguity, it is perhaps not surprising that even when governments obtain knowledge of the enemy’s war plans and intentions, leaders often fail to take advantage of it or distrust it because of contradictory reports (Iklé 1991). For example, the United States was unable to take advantage of having deciphered Japanese codes before the attack on Pearl Harbor.15 Poor-quality, ambiguous information can lengthen war in three ways. First, as Bayesian models suggest, ambiguous information requires more rounds of “screening” the opponent with diplomatic offers and battles, in an effort to identify his “type” (Powell 2004; Slantchev 2004; Filson and Werner 2002, 2004). Second, the more ambiguous the information, the more likely leaders will rely on prior beliefs, guesses and schemas and the more likely incoming information will be interpreted to conform with desires and expectations (Snyder and Diesing 1977: 310–12; Rosati 2000). For example, coalition members may interpret

d omest ic coalit ion shifts



ambiguous information optimistically, leading them to assume their current war policy is having the desired effect and thus encourage them to continue fighting. Finally, relying on guesses can provide easy opportunities for self-deception. Judgments about the enemy can consciously or unconsciously allow decisionmakers to avoid unpleasant truths, such as the prospect of a prolonged stalemate or defeat. When guesses and predictions get blended in with known facts, it can lead decision-makers to treat the predictions (and their underlying assumptions) as certainties, rather than directing attention to the possibility that the prediction might be wrong. As will be seen below, this dynamic can be complicated by effort justification and other psychological biases in decision-making. Exposure to Different Information. A second reason updating may not occur as Bayesian models predict is because different coalition members may be exposed to different information. At least three dynamics can impede the flow of information. First, given the information asymmetries of the principal-agent relationship, the executive is likely to have some information that is not available to all of his advisors (Downs and Rocke 1994). Similarly, military leaders in the field often have information not available to civilian coalition members at home (Halperin 1970; Feaver 2003). Different coalition members may have differing access to classified intelligence information based on their clearance and “need to know.” Most generally, information monopolies can keep information about the war from the public and sometimes even other parts of the government, especially legislatures (Snyder 1991; Colaresi 2005; Gibbs 1995; Page and Shapiro 1992). Clearly, in the midst of war, those who have access to the flood of raw data have plenty of room to manipulate information and misinform others. As Lord Asquith, British Prime Minister in the early years of World War I, once quipped, the military in the War Office “kept three sets of figures, one to mislead the public, another to mislead the cabinet, and a third to mislead itself ” (quoted in Iklé 1991: 97). Second, coalition members will usually have access to detailed information that originates in their own agency, while only seeing the “sanitized” or synthesized results of such information from sister agencies. To use an example from the US government, the secretary of state will be able to see raw data from foreign opinion polls and cables from US embassies abroad, while the director of National Intelligence is more likely to see raw intelligence information and monitor the dialogue that captures the various nuances in analysis across the intelligence community’s 16 member organizations. In contrast, the secretary of defense is more likely to see only summary reports of such data. Finally, problems with information flow can be aggravated by organizational or structural pathologies in the decision-making process. Organizations and bureaucracies’ formal and informal power structures, standard operating proce-

38

the d omest ic p olit ics of war ter minat ion

dures and internal culture can affect how information is processed and filtered (Allison and Zelikow 1999; Simon and March 1993; Cyert and March 1992; Perrow 1986). For example, it can be difficult for decision-makers within an organization to discuss or even conceive of alternatives that are not culturally sanctioned (Weick 1995), and hierarchical chains of command can thwart information flow. Similarly, unpopular policies can make decision-makers feel like they are under siege, which can lead to closed decision-making procedures that thwart information flow (Janis 1982; Janis and Mann 1977; Anderson 1985). Moreover, there is an inverse relationship between information flow and coercion in a political system. The more centralized and coercive the system, the less likely it is that factual, honest information will reach the top, as officials may hide facts or their true preferences to preserve their jobs (Apter 1965; Kuran 1997).16 Exposure to different information can lengthen war in four ways. First, coalition members can use differing information access to manipulate other coalition members to their policy position. For example, hawkish coalition members can selectively share information that suggests the state is performing well, while hiding information it is not, thereby keeping dovish coalition members on board to continue fighting. Second, this dynamic can also work outside the coalition, as well. Coalition members’ information monopolies can allow them to manipulate public opinion or other actors without information access. If coalition members want to continue the war, they can selectively share information (“spin the story”) to lead others to accept their interpretation of events (Snyder 1991; Mansfield and Snyder 1995; McGinnis and Williams 2001; Mor 1997; Colaresi 2005; Gibbs 1995; Baum 2003: 44–45; Johnson and Tierney 2006: 67–75). Third, organizational and structural pathologies to information flow—such as preference falsification in coercive regimes—can breed confusion and ignorance and generate inefficiencies in decision-making, which can delay a war-ending decision. This dynamic is clearly linked to the self-suppressing tendencies of potential critics in the preference obstacle. Finally, within the coalition, access to different information can lead coalition members to draw different conclusions about war performance, and if their assessments collide, it can create gridlock and policy stability. This final dynamic relates directly to indicators and metrics. Different Indicators and Metrics to Assess the War. A third reason updating may not occur as Bayesian models predict is because different coalition members may use different indicators to assess the war and thus reach different assessments. Indicators are cognitive shortcuts actors use to cope with complexity (Gartner 1997; Johnson and Tierney 2006; Rosati 2000). While most Bayesian models assume all belligerents use a common combat performance metric to update their beliefs, this approach argues that different states—and different coalition mem-

d omest ic coalit ion shifts



bers within states—use different indicators to assess war performance (Gartner 1997). This “dominant indicator approach” focuses on sudden and dramatic changes in the quantitative indicators on which decision-makers rely—such as combat deaths, gains in terrain or leader popularity polls. Because different coalition members often rely on different indicators, they can reach different assessments of the same situation, i.e. when one indicator suggests that a policy is performing well, while another suggests that the policy is performing poorly.17 Not every coalition member will be tracking the same indicators, so new information coming in may not cause all coalition members to change their war policy assessments simultaneously. Indicators not only lessen the cognitive load for decision-makers during the war, but—given their quantitative focus—they can also create a false impression of certainty, for three reasons. First, while quantitative indicators may be based on facts, the process of collecting the data on which they are based is often plagued by measurement error. For example, recent debates about the number of Iraqi non-combatant casualties highlight complications in gathering such statistics. Moreover, decision-makers usually sweep over assumptions that indicators implicitly hide so the indicators can provide aggregated estimates. Unfortunately, some common ways of judging military success—such as territorial gains or improvements in current force ratios—are insufficient or even misleading indicators for guiding the conduct of war (Iklé 1991: 19). Yet coalition members, especially government bureaucracies, continue to rely on them to avoid explicitly articulating “political” judgments about preferences, war aims and risk. As a result, they treat indicators as objective facts that proxy for broader estimates, while ignoring the measurement errors and underlying assumptions inherent in the proxy process. Finally, coalition members often choose self-serving—rather than objective—metrics that corroborate their beliefs, expectations and desires (Johnson and Tierney: 2006: 42; Dunning and Cohen 1992). Different indicators and metrics can lengthen war in two ways. First, if different coalition members rely on different indicators that diverge, their assessments about the war may collide and lead to policy gridlock. As a result, the coalition becomes unable to update its expectations and agree on whether and how to change war policy. Second, because progress in war is often ambiguous, coalition members can choose from a wider range of indicators to assess war performance. This makes it easier to choose self-serving, optimistic metrics that bolster their expectations and assessments about war performance. In theory, if all indicators lead to pessimistic assessments of the current war policy, this could shorten the war. In practice, however, this rarely occurs, because coalition members try to find other metrics that “tell the story” they want to hear.

40

the d omest ic p olit ics of war ter minat ion

Individual Biases. A fourth reason updating may not occur as Bayesian models predict is because coalition members may have individual biases that prevent them from processing war information. Here, information reaches the leaders, but they are unable to absorb and respond to it as Bayesian models would predict. Whether or not the information itself is biased, individual leaders are subject to cognitive, affective and learning biases. Cognitive biases arise from the way the brain copes with information processing constraints, especially when faced with complex problems.18 Affective biases arise from emotional states, such as desire or fear, especially when people feel threatened. These processes may be particularly important when the best option is uncertain or when decision-makers experience physiological or emotional stress.19 Learning biases arise in applying past experiences to shape perceptions of current or future events.20 Individual biases can lengthen war in at least five ways. First, in “belief system defense,” coalition members often take data that supports their expectations at face value, while ignoring, discounting or rejecting information that does not fit with favored or well-established beliefs (Kahneman and Tversky 1979; Jervis 1976: 181–87; Nisbett and Ross 1980: 167–91; Fiske and Taylor 1991; Snyder and Diesing 1977: 310–29). Central or “core” beliefs are protected more strongly than “peripheral” or instrumental ones (Kaufmann 1994: 563). Delusional leaders may be particularly prone to this tendency, as are challengers with the most at stake (Byman and Pollack 2001: 137; Kaufmann 1994: 579). Thus, hawkish leaders may miss important cues that it is time to settle. Second, individuals often see their own behavior as driven by situational constraints, while seeing others’ actions as more voluntary or dispositional. Individuals are also likely to see their own behavior as reasonable, while overestimating the hostility of others. Applying this “attribution bias” to ending war, coalition members may read malicious intent into others’ behavior, furthering the spiral of suspicion and reinforcing hawkish tendencies to escalate or continue the war (Zimbardo 2008: 7, 212; Jervis 1988: 688–90; Kahneman and Renshon 2007). Third, individuals not only recall success more easily than failure, but they also avoid recognizing the likelihood of failure. They tend to give themselves credit for creating success, while blaming outside forces as the cause of failures (Janis and Mann 1977: 81–106; Kaufmann 1994; Johnson 2004; Johnson and Tierney 2006). An important consequence of this “success bias” for war termination is that coalition members with a large stake in the war policy’s success will be slower to acknowledge evidence the state is losing than those without such a stake. Moreover, this bias suggests coalition members will weigh battlefield victories more heavily than defeats, leading to over-optimistic evaluations of war performance and prolonging the war (Smith 1995: 22).

d omest ic coalit ion shifts



Fourth, coalition leaders may incorrectly remember or misunderstand historical events or use history in biased ways (Levy 1994; Johnson and Tierney 2006). For example, leaders can apply historical analogies in simplistic ways, thus hiding important differences between the historical and current situations and drawing misleading conclusions (May 1973; Khong 1992; Schacter 1995). More maliciously, “learning bias” dynamics can interact with information monopolies and allow coalition members to manipulate public opinion. If coalition members want to continue the war, they can selectively draw on history to spin current events or demonize the enemy, thereby leading others to accept their hawkish interpretation. Finally, individual biases become particularly potent during the process of “effort justification,” a common characteristic of sticky wars. Effort justification occurs when individuals who have expended a lot of effort to achieve some goal increase the value of attaining that goal—even if a rational cost-benefit calculation would suggest that persisting now is unwise. In sticky wars, leaders who began the war frequently become so personally invested that they cannot change expectations. As a leader comes to believe his policy is necessary, he is likely to believe it can succeed, even if that belief requires distorting information. The greater the commitment to a goal, the stronger the bias is (Kaufmann 1994; Jervis 1988: 693). By having “too much invested to quit,” leaders may continue investing in policies that are unwarranted, hoping to justify or regain previous investments (Teger 1980; Smith 1995: 74; Croco 2008; Milburn and Christie 1990: 238). Paradoxically, while such “commitment biases” can help leaders avoid the painful truth that the current war policy is unsuccessful, they reinforce leaders’ personal responsibility and commitment to the current policy. This can create a vicious cycle, as leaders continue to justify their effort despite logical or cognitive inconsistency. Paired with “belief system defense” discussed above, effort justification is a strategy for reducing cognitive dissonance (Festinger 1957; George 1980; Jervis 1976; Kiesler, Collins and Miller 1983; Johnson and Tierney 2006). This discussion about effort justification points to a very important feature of the information obstacle: often the reason that leaders do not know to end the war is because they choose not to know. These biases highlight how that choice can be consciously or unconsciously fueled by a reluctance to confront painful facts. Environments with poor-quality information, reduced information flow and uncertainty—such as those during wartime—enhance this human capacity for denial. In other words, the previously discussed information obstacle dynamics can interact with and heighten the effects of individual biases. Organizational Biases. A final reason updating may not occur as Bayesian models predict is because governing coalitions may have organizational or group

42

the d omest ic p olit ics of war ter minat ion

decision-making biases that prevent them from processing war information. For example, small, cohesive policymaking groups under stress can be vulnerable to “groupthink,” where the desire to minimize controversy compromises the quality of discussion. As a result, the group can ignore or misinterpret crucial information; ignore or discount alternatives to the preferred policy; or persist in policies even when confronted with feedback they are risky or not working well (Janis 1982; Hart, Stern and Sundelius 1997). Similarly, individuals may evaluate risk differently when they are in groups than when they make decisions alone—especially if the group is small or predominantly male. With the “risky shift” dynamic, groups are likely to make riskier decisions than individuals. Decision-makers may not want to let their compatriots down or be perceived as the “wet blanket” in the group.21 Both dynamics exemplify aspects of “preference falsification,” when an individual misrepresents his true desires under perceived social pressure (Kuran 1997). Alternatively, organizations’ and bureaucracies’ formal and informal power structures, standard operating procedures and internal culture can affect how information is processed and filtered (Halperin 1974; Allison and Zelikow 1999; Simon and March 1993; Cyert and March 1992; Sigal 1988; George 1980; Vertzberger 1998; Perrow 1986; Weick 1995). Weak states tend to have especially active bureaucracies that can inhibit the policymaking process (Skidmore 1996; Friedberg 1988). War can create organizational neuroses, such as the inability to evaluate information and assess policy in an effective way (Van Evera 1999). Organizational biases can lengthen war in at least three ways. First, they can magnify the war-lengthening effects of individual biases by providing additional motivations for effort justification and belief system defense. Second, organizational biases can magnify structural problems with information flow, thereby generating confusion and creating decision-making lags and inefficiencies. Third, organizational biases can magnify the self-suppressing tendencies of potential critics. For example, as discussed with the preference obstacle, it can be difficult for decision-makers within an organization to discuss or even conceive of alternatives that are not culturally sanctioned (Weick 1995). As these five information obstacle dynamics illustrate, coalition leaders may not update their expectations as Bayesian models would predict. Although information updating pathologies are usually associated with authoritarian leaders, these dynamics also exist in democracies, as the US case in the Korean War will illustrate. All governments, regardless of regime type, experience information obstacle dynamics, which produce subsequent complications for decision-making and policy assessment. Table 2.1 summarizes how the five information obstacle dynamics can lead to longer wars.

d omest ic coalit ion shifts



table 2.1 Summary of the Five Information Obstacle Dynamics Information Obstacle

Poor Quality/ Ambiguous Information

War-Lengthening Effects

More rounds to “screen” opponents’ “type” More reliance on prior beliefs, guesses and schemas—leading to overoptimistic interpretations of information More opportunities for self-deception

Exposure to Different Information

More opportunities to manipulate other coalition members or constituents to a hawkish position More confusion and inefficiencies in decision-making Greater likelihood of conflicting assessments of war performance— leading to policy gridlock

Different Indicators & Metrics to Assess the War Individual Biases

Greater likelihood of conflicting assessments of war performance— leading to policy gridlock Greater likelihood of coalition members choosing self-serving metrics that “tell the story” they want to hear Greater likelihood of absorbing information that supports expectations, while ignoring, discounting or rejecting information that does not (“belief system defense”) Greater likelihood of reading malicious intent into adversary’s behavior (“attribution bias”) Over-optimistic evaluations of war performance (“success bias”) Misapplication of historical analogies to manipulate constituents or draw misleading conclusions (“learning bias”) Effort justification (“commitment bias”)

Organizational Biases

Group-level effort justification and belief system defense Greater likelihood of structural problems with information flow More confusion and inefficiencies in decision-making Greater likelihood of self-suppression among potential critics in the coalition

Certainly, some Bayesian accounts allow for slow updating, as ambiguous information requires more rounds of “screening” the opponent in an effort to identify his “type” (Powell 2004; Slantchev 2004; Filson and Werner 2002, 2004). Moreover, under Bayes’ theorem, if different coalition members had been exposed to different prior information, it could lead to differences in updating in response to new information (Edwards, Lindman and Savage 1963). Nonetheless, Bayesian models predict coalition members’ information sets should eventually converge, especially if they are engaging in frequent policy discussions. Bayesian models also predict coalition members will change their beliefs, or at least question them, when they encounter refuting evidence. Finally, Bayesian models would predict that, once confronted with such disconfirming data, coalition members will actively search for additional information to resolve that inconsis-

44

the d omest ic p olit ics of war ter minat ion

tency (Kaufmann 1994: 565). In short, Bayesian models predict a relatively linear function between actors receiving new information and changing their expectations or beliefs. In contrast, these five information obstacle dynamics would not predict such a linear relationship. While Bayesian models predict information convergence across the coalition, the information obstacle predicts differing rates of learning and updating among coalition members. While Bayesian models predict belief changes after coalition members encounter disconfirming data, the information obstacle predicts self-deception, denial, a lurch towards more self-serving metrics or no response at all. Finally, while Bayesian models predict coalition members will seek out more information after learning disconfirming data, the information obstacle predicts a failure to search for more information, policy gridlock, spin-doctoring to manipulate the public or effort justification. In other words, the information obstacle predicts beliefs remain sticky until some critical point is reached, when they collapse and get replaced by new expectations (Snyder and Diesing 1977: 329). As we shall see below, that critical point may require a coalition shift. As Chaim Kaufmann argues, “Change only occurs when a particular experience is too salient to be ignored, too unambiguous to be discounted, and so squarely in conflict with the prior belief that it becomes consistently cheaper to abandon the belief than to try to resolve or to tolerate the inconsistency” (1994: 563). In short, unlike Bayesian models, the information obstacle would suggest that information updating occurs according to a step function, if it occurs at all. Entrapment Obstacles With entrapment obstacles, domestic coalition members may want to end the war but cannot. Entrapment can occur when the leaders’ hawkish constituencies want to continue the war, and the leaders must acquiesce to remain in power. If leaders know they will likely lose power if they end a war, they can rationally choose to continue it (Downs and Rocke 1994; Goemans 2000a, b; Bueno de Mesquita et al. 2004; Bueno de Mesquita and Siverson 1995; Chiozza and Goemans 2004; Croco 2008). In other words, anticipating removal from power can cause the leaders to continue fighting and prolong the war. Leaders can face this political entrapment from a domestic constituency or from external allies. Internally, entrapment may result from having to answer to hawkish constituencies or, more benignly, from having “spun up” the population to mobilize for war (Russett 1990; Mueller 1985). Unless the war ends quickly, leaders must enact policies—such as demonizing the enemy and mobilizing the armed forces—to ease the process of waging war. These policies require that leaders expend political capital. The more political capital leaders expend, the more they convince

d omest ic coalit ion shifts



their constituents that fighting the war is the right thing to do. Enacting these policies is necessary for winning on the battlefield, but these policies make it extremely difficult to simply “turn off ” the decision for war. The very policies that make a state effective in preparing for and fighting war are the ones that preclude an easy end to that war. Paradoxically, an intolerant, hawkish constituency may actually be the result of the leaders’ propaganda efforts. Many scholars have argued that public opinion is prone to elite manipulation (Snyder 1991; Mansfield and Snyder 1995; McGinnis and Williams 2001; Mor 1997; Colaresi 2004, 2005; Gibbs 1995). In this view, elites have a monopoly over national security information and “expert” opinion, and the population generally accepts elites’ interpretations of events (Zaller 1992; Berinsky 2007). Such manipulation may act as a “push” towards a particular way of thinking, which may either augment or counter the “pull” of pre-existing beliefs or salient events like battlefield defeat (Johnson and Tierney 2006: 67–75). For example, leaders often demonize and exaggerate their conflict of interest with the enemy. After all, it is easier to raise armies, money and morale if a conflict is pitched as a Manichean struggle against evil than as an effort to obtain a marginally better set of peace terms. Framing the war in simple, emotionally compelling terms increases the likely level of domestic support for the coalition’s war policies (Armstrong 1994; Pillar 1983; Rothstein 1970; Baum 2003: 44–45). During major wars, actors often shift to a “mythic perception” of reality, in which simplistic notions like good and evil and “us” versus “them” take precedence (LeShan 2002). Yet while manipulating perceptions helps to sustain the war effort, it hampers making peace. Once established, enemy images are highly resistant to change, further undercutting compromise with the other side (Eldridge 1979; Levy 1996; Frank 1982; Keen 1986; Olweean 2002; Aho 1994, 1998; Iklé 1991; Smith 1995; Mor 1997). As the conflict wears on, this enemy image is reinforced by continued public statements and forceful propaganda. This creates, as Bryant Wedge argues, a “striking tendency to perceive only those items which would sustain the enemy image and be blind to evidence that would contradict such views” (1990: 104). Whether or not leaders come to believe their own rhetoric about the enemy—a process called “blowback”—it can make it difficult for leaders to make more conciliatory or cooperative statements about the other side. The decision to use propaganda to inflate the threat makes de-escalation unlikely. Even if the governing coalition wants to deescalate, other political challengers can capitalize on negative enemy stereotypes and accuse the coalition of “being soft” on the adversary (Colaresi 2004, 2005; Snyder 1991; Vasquez 1993; Van Evera 1984). In short, by demonizing the enemy and using threats to centralize power, leaders can get caught in their own propaganda traps.

46

the d omest ic p olit ics of war ter minat ion

Similarly, leaders can shape perceptions about success on the battlefield, but in the process talk themselves into a corner. Authoritarian leaders can shape perceptions through strict control of information—one way that the entrapment and information obstacles can interact. Yet even democratic leaders can manipulate perceptions of battlefield success, because people usually know less about foreign affairs and are more willing to defer to a leader’s interpretation of events, at least initially (Johnson and Tierney 2006: 70–75; Colaresi 2005: 25–29; Page and Shapiro 1992; McGinnis and Williams 2001; Baum 2003: 44–45). However, raising expectations of victory can create a dilemma. The political costs of failure are higher after a leader builds public support by touting war performance (Baum 2003: 285). As Johnson and Tierney argue, “Sooner or later the public will expect visible results” (2006: 73). Leaders can get stuck in a balancing act between conveying that the mission is currently successful, while simultaneously lowering expectations about the long struggle ahead. In short, mobilizing the state for war can make it difficult for leaders to back down without losing domestic political legitimacy. Entrapment can create unpleasant psychological and political effects for leaders. Psychologically, backing down from a publicly defended position may cause leaders to appear weak or lose face or pride (Smith 1995: 74). Politically, backing down from a publicly defended position can damage a leader’s bargaining reputation internationally (Jervis 1988). While rallying the public for the war can maximize credibility with the adversary, the leader is likely to pay a domestic political price—a domestic audience cost—if he backs down (Putnam 1988; Fearon 1994; Schultz 1998; Smith 1998; Baum 2003). Leaders also risk being driven from power by constituencies that have come to expect the position to be defended at any cost (Colaresi 2004, 2005; Croco 2008). These mechanisms of entrapment are especially problematic in weak states with strong societal pressures (Skidmore 1996; Armstrong 1994). If leaders decide not to cut their losses and move towards ending the war, entrapment can have three effects. First, the governing coalition may decide it will lose power if it makes peace and thus it adopts a high-risk strategy for prosecuting the war. This is the “gambling for resurrection” argument (Downs and Rocke 1994; Goemans 2000a, b; Bueno de Mesquita et al. 2004; Bueno de Mesquita and Siverson 1995; Chiozza and Goemans 2004; Croco 2008). Alternatively, leaders can succumb to the “sunk cost paradox,” effort justification and other psychological aspects of entrapment (Maoz 1990; Jervis 1976; Craig and George 1995; Randle 1973; Smith 1995; Iklé 1991). In theory, rational actors do not let sunk costs influence their decisions, because doing so would not be assessing the situation exclusively on its merits. In practice, however, leaders

d omest ic coalit ion shifts



often treat sunk costs as invested capital on which the public demands a return or show a tendency to feel bound by earlier war policy decisions (Fisher 1969; Albert and Luck 1980; Jentleson et al. 1992: 306; Mitchell 1981: 180–81). Third, because trapped leaders find it so difficult to cut their losses, dislodging them from this position may require a large exogenous shock, such as a major battlefield defeat or third-party intervention on the other side (Stein 1982: 234; Gartner 1997; Mor 1997). People pay more attention to, and are more influenced by, salient events like exogenous shocks. Salient events can register as large changes in leaders’ war metrics. They can also have a disproportionate effect to their actual impact in context, because people see salient events as symbolic of the bigger picture (Johnson and Tierney 2006: 63; Kaufmann 1994: 563; Nisbett and Ross 1980: 40–45). Thus, events like a US Ranger being dragged through the streets of Mogadishu can help “wake up” a coalition from its impulses to justify effort, treat sunk costs as investments or engage in other behaviors of entrapment. Because different coalition members answer to different constituencies, they may feel trapped at different times. During a war, different parts of the power structure have varying commitments to the military effort. As the war continues, groups benefiting from the war are likely to obtain additional influence in the governing coalition and promote the war’s continuation. The more important a constituency’s role in the war effort, the greater the share of the state’s resources it can legitimately claim as its own, and thus the greater the private benefits it may be drawing from the war (Lamborn 1991: 62–63; Iklé 1991: 60; Snyder 1991). A coalition member who answers to such constituencies has the hardest time going against his supporters. Unless he can convince his constituency to change its mind, he will be trapped. This differing rate and type of entrapment can lead a minority faction in the coalition to want peace but not be able to make it, unless it is powerful enough to remove the hawks and create a sustainable governing coalition. While some coalition members will feel trapped by their domestic constituencies, others may be trapped by commitments to allies or external coalition partners. Externally, entrapment may be the result of having hawkish allies or patrons. If an ally intervenes on a belligerent’s behalf, it is difficult to deescalate after that intervention without disappointing the ally. Leaders may decide that they value the ally’s support during or after the war so much that they are willing to stay in the war for the ally’s sake—as the Chinese case in the Korean War illustrates. In addition, war aims often expand as statesmen make promises to win new allies.22 Finally, parallel to the unintended consequences of domestic mobilization, demonizing the enemy and mobilizing allies can make it harder for leaders to back down without losing international stature. Conveying weakness may

48

the d omest ic p olit ics of war ter minat ion

create a reputation as a weak bargainer, which might be exploited by others—the current adversary as well as potential future opponents (Jervis 1988: 693). In both cases, although the governing coalition may have realized it is not going to obtain its initial goals, a radical change in strategy is not politically possible without admitting failure—which politicians loathe to do (Pierson 1984). As a French politician noted, in response to his generals’ suggestion concerning a cease-fire in Algeria: “But that would be capitulation. And it would show that we had been wrong for seven years” (quoted in Rothstein 1970: 69). Because sudden departures from previous policy raise serious questions about the wisdom of having pursued it for so long, entrapment puts leaders in a no-win situation. They realize that they cannot win, but they cannot afford to quit at once (Maoz 1990). The result in both situations is the same: decision-makers may want to settle but cannot because they feel trapped. Two caveats are in order. First, the dynamics of entrapment may seem similar to self-suppressing potential critics in the preference obstacle. In both cases, a coalition member wants to end the war but cannot, but there are distinct differences between them. With entrapment, he suppresses his desire to end the war because he fears penalties from a hawkish domestic constituency or an external ally. With the preference obstacle, however, the potential critic suppresses his desire to end the war because he fears penalties from a more powerful coalition member. In other words, an entrapped coalition member fears penalties from below or outside the coalition, while a self-suppressing coalition member fears penalties from above and within the coalition. Second, while Bayesian updating predicts a relatively linear function between new information and revised expectations or beliefs, the entrapment obstacle predicts coalition members may use new information to update prior beliefs but feel unable to act on those new beliefs. Alternatively, because entrapment dynamics can interact with the information obstacle, coalition members with a large stake in the current policy’s success may be slower to acknowledge evidence that the policy is failing. Thus, while Bayesian models would not predict any correlation between interests and assessments, the interaction between entrapment and information obstacles predicts the greater the interests at risk, the stronger the potential bias. One way to think about the interaction is this: If the psychological commitments and domestic political constraints of entrapment provide the motive, then the complexity and uncertainty of information during war provide the opportunity for psychological biases to persist. The logic of entrapment can motivate a leader’s reluctance to confront painful truths, and this can help coalition members choose not to know they should end the war. One implication is that the longer war continues and the more trapped a leader perceives himself to be, the more individual and organizational biases can impede Bayesian updating.

d omest ic coalit ion shifts



overcoming obstacles to peace

While the three obstacles to peace certainly contribute to the “stickiness” observed in long wars, the dynamics of the domestic governing coalition itself exacerbate them. Coalitional dynamics help explain policy inertia. When a policy is functioning because of a coalition agreement, changing the policy requires renegotiating that agreement (Snyder and Diesing 1977: 351). Likewise, the less control a coalition member has over policy choice and implementation, the less willing he is to take political risks to change it—unless the current policy’s failure is expected to generate more risk (Lamborn 1991: 63–66). Finally, a few actors may have controlled war policy initially through discrete decisions in the early days of the crisis. However, the longer the war continues and its policies become institutionalized, the greater the opportunity for a wide range of actors—including those outside the governing coalition—to participate in the policymaking process (Fordham 1998: 21). Thus, in addition to the three obstacles to peace, policy stability can lead governing coalitions to be unable to change the status quo, even when such changes are necessary or desirable. When new information shows that no policy change can meet all of the current coalition members’ political and psychological constraints, the coalition has two choices: stay the course or loosen the constraints. Various constraints will be ranked differently by different coalition members, and each may become spokesman for the constraint most important to him. To loosen the constraints, there are two possibilities. First, the veto player with the largest ideological or policy distance from the coalition majority may be overruled or removed from the coalition. Alternatively, it may be the weakest coalition member who loses out—either by having his payoff reduced or, more likely, being expelled. Which outcome obtains depends on whether there is a “pivotal actor” in the coalition—such as a strong head of state—and what his constraints are (Snyder and Diesing 1977: 396; Tsebelis 1995; De Swaan 1973). In general, these strategies for loosening constraints are only possible when the coalition core is strong enough to do the overruling or expelling. Otherwise, the policy cannot be revised and the coalition will stay the course. In these cases, a government with policy stability may become so immobile that policy change is only possible by replacing the incumbent government entirely (Tsebelis 1995: 321). These dynamics can help explain the rigid bargaining positions observed in war-ending negotiations at the international level, where a belligerent’s behavior lags behind what a pure Bayesian updating model might predict. Regardless of the length of the war, the overlapping bargaining space necessary to end it can only develop when at least one side changes its expectations about the costs and benefits of fighting. This change could theoretically occur if

50

47% 2.1

INPUT

Battlefield outcomes/ Diplomatic offers

the d omest ic p olit ics of war ter minat ion F I LT E R S Preference Obstacles Information Bayesian updating

RESOLUTION Stickiness re s o lve d

(l by res ess o olv fte ed n by )

Change in decision-makers (coalition shift) Attitude change (within existing coalition)

OUTCOME

Change in expectations (dove or hawk direction)

f i g . 2 . 1 . Paths by Which Belligerents Change Expectations

one (or more) of the belligerents’ existing governing coalitions either (1) changes its attitudes about the war or (2) changes its composition to include more appropriate expectations (see Figure 2.1). The former is the Bayesian updating model. However, a change in attitudes is not very likely because of coalition dynamics and the three obstacles identified: the coalition may have a private benefit from continuing a costly war, may not know its expectations should change, or cannot afford to admit it has changed its expectations. War ends when these obstacles are lifted. The decision to end a war forces policymakers to choose among basic policy goals. The more unfavorable the outcome, the more deeply these choices will cut into fundamental interests and threaten to divide the domestic coalition. Those officials who are identified with the initiation of the war or its early conduct are apt to fear that they would undermine themselves as leaders if they consider any conclusion to the war that did not achieve their principal war aims—unless they are successfully able to redefine the cause and thus what constitutes “victory.” At the same time, those who favor a negotiated settlement must attack—implicitly or explicitly—those who led the nation to war (Iklé 1991). Various proposals for ending the war will thus appear as either threats or opportunities, according to the individual decision-maker’s perspective. Thus, to resolve these conflicting motives, unless a war’s battlefield situation clearly shows the belligerents’ relative strengths in a short period of time, it is unlikely that the governing coalition that started and invested significant political capital in prosecuting a war can end an unsuccessful one. Policies that a state enacted to better wage the war become sticky, and will often require a coalition shift to get them unstuck. The coalition shift allows different political actors—with different interests, assessments of the war and constituencies—to take power. This, in turn, can lead to a change in war policy, e.g. from continuing it to ending it.23 The new leaders may have not only a new calculation of the war’s costs and benefits, but also a new conceptualization of them. Because it introduces new ac-

d omest ic coalit ion shifts



tors, a coalition shift may overcome obstacles to peace and thus permit the war to end. Clearly, this argument contrasts with Bayesian models that assume updating will happen automatically as new information is revealed through battlefield outcomes and diplomatic offers, or that new ideas about war policy will inevitably find their way into the government as leaders change their minds. Instead, I argue new ideas need exposure within the governing coalition, and such exposure often comes with significant personnel changes in the main institutions of power.24 This may be, in fact, the most important way that regime type affects changes in war policy, because democracies are more likely to experience an uninhibited circulation of elites.25 The realities that make a change in attitudes among the existing coalition unlikely are directly addressed by a change in that coalition. Where coalition members place different values on policy priorities, a coalition shift allows one faction to remove members whose preferences created gridlock. Where the rational updating process is not even among all coalition members because of information obstacles, a coalition shift allows one faction to be removed or replaced, facilitating the flow of information and removing spurious inputs. Where coalition members face entrapment, a shift removes the entrapped element. In this manner, domestic coalition shifts provide the procedural opportunity for leaders to toss out those who are against making peace. It also allows the new coalition to distance itself from the decision that brought the state into the conflict, thereby presenting it with more options regarding the future course of the war and enlarging the potential bargaining space. There are four other practical effects as well. First, if the war was not “won,” a coalition shift would allow scapegoats to be identified and removed.26 Second, a shift could provide the previous leaders a “golden parachute” so that they will not be physically or politically harmed after the settlement (O’Donnell and Schmitter 1986: 15–47; Shin 1994: 161–63; Mansfield and Snyder 1995: 36–37). Third, a shift could be necessary for stabilizing the domestic political scene after the settlement; i.e. to make the peace, leaders must include the loyal opposition. Fourth, bringing in a new coalition may also be the best way for the state to rebuild its international reputation (Croco 2008: 28). These four practical effects are not mutually exclusive. Figure 2.2 summarizes the domestic coalition shift theory. To remind the reader, the gray rectangle delineates the standard Bayesian model. In the pure Bayesian model, leaders of a state perceive changes in the battlefield situation (or the adversary’s diplomatic behavior), which causes a change in their expectations about the war. The domestic politics version of this argument filters battle-

52

the d omest ic p olit ics of war ter minat ion

2.2 -- 61%

T h e D o m e s t i c Co a l i t i o n Sh i f t Mo d e l exogenous circumstances

3

coalition shift

battlefield situation

2 pressure on governing coalition

constituency reevaluation

in ers ge ak an -m ch ion cis

de

1

attitude change

change in expectations about war

war termination

attitude change

B aye s i a n Mo d e l f i g . 2 . 2 . Paths to War Termination

field conditions through domestic constituencies, who then pressure leaders to change their war policy—which causes them to change their expectations. Both Bayesian variants assume the change in expectations occurs when incumbent leaders change their minds (“attitude change”). However, as Chapter 9 will show, in many wars—41 percent since World War II, for example—this change in expectations results from a shift in the domestic governing coalition (“change in decision-makers”). It is worth noting that domestic coalition shifts may be produced in a number of ways, both endogenously to the war and exogenously. While endogenous shifts are influenced by the battlefield situation, this does not mean they have no causal weight in explaining war termination. In some wars, the necessary updating of war aims cannot happen without an endogenous shift, which makes the shift a necessary intervening variable to ending the war. Theoretically, there are two kinds of endogenous shifts (labeled as Arrows 1 and 2 in Figure 2.2). Endogenous shifts may be a response to changes in the domestic constituency’s perceptions about the progress of the war that are not shared by the existing domestic coalition (Arrow 1). These may be cases where the coalition has become so immobile that policy change is only possible by replacing the incumbent government entirely. Such shifts often include changes to the head of state through political processes such as coups, revolutions, regime changes or elections that replace

d omest ic coalit ion shifts



the governing coalition from the outside. An example of an Arrow 1 shift is the Bolshevik Revolution in 1917. The new Communist leader, Vladimir Lenin, subsequently accepted the harsh terms of Brest Litovsk to make a separate peace and end Russia’s participation in World War I. Arrow 1 shifts do not necessarily require that all domestic constituencies are ready to end the war. It is important to recognize that domestic constituencies are likely to have their updating impeded by the obstacles to peace, as well. Some constituencies, such as the military or industrial groups, may be gaining private benefits from continuing the war (preference obstacle). Other constituents will lack access to information about the war and come to believe the elite’s interpretation of events (information obstacle). Finally, many constituents will have internalized negative enemy stereotypes and be intensely distrustful of the adversary (entrapment obstacle). Thus, updating among the different domestic constituencies is likely to be as uneven as among the coalition members themselves. Alternatively, endogenous shifts may be a response to changes in part of the existing coalition’s perceptions about the progress of the war and the pressure placed upon the coalition members by their constituencies (Arrow 2). These endogenous shifts are less likely to include head of state changes; rather, they are likely to result from heads of state recalibrating their governing coalitions to remove veto players who are stuck by the three obstacles. A common Arrow 2 shift among losing states is to remove hawks in the existing domestic coalition to allow the remaining doves to make a “bad” peace. For example, Menachim Begin and other Gahal Party ministers opposed to peace resigned from the Israeli national unity government in August 1970, thereby allowing Prime Minister Golda Meir to accept a negotiated settlement to end the 17-month Attrition War the next day. Finally, exogenous shifts (labeled as Arrow 3 in Figure 2.2) may include the death, illness or incapacitation of a coalition member or changes in the coalition composition for purely domestic political reasons, such as economic crises or political scandals. For example, Stalin’s death in January 1953 allowed the Communist states to offer crucial concessions necessary to end the Korean War. While exogenous coalition shifts occur for reasons that have no relationship to the war, their effect on war policy can be the same as endogenous shifts. Either way, the coalition shift can be the breaking of the dam that releases the forces of change. Importantly, my model excludes domestic coalition shifts that are forced upon a state by its external enemies. This is consistent with the “domestically-instigated regime change” approach in Bueno de Mesquita et al. (1992). Excluding these “victor-imposed” shifts helps guard against the possibility that the causal arrows are reversed—that the end of the war is leading to coalition shifts, rather than vice versa.

54

the d omest ic p olit ics of war ter minat ion

The white box around the domestic coalition shift model highlights how this alternative causal pathway to war termination builds on the Bayesian model. Domestic coalition shifts can explain how states get beyond the lag and stickiness of long wars. However, domestic coalition shifts are neither necessary nor sufficient for getting beyond these lags, because we cannot know a priori how much each shift will change the belligerents’ expectations and what effect the strategic interaction with the opposing side will have. Theoretically, a domestic coalition shift in any belligerent could produce the bargaining space. However, realistically, the power differential between allies in multilateral wars is not always equal. Thus, a coalition shift in the state that has the most responsibility for or influence on its side’s decisions about war policies (the “primary belligerent”) is more likely to affect war termination. Most wars have only one primary belligerent on each side, but some wars have more than one primary belligerent on each side—especially when there are great powers fighting together.27 In sum, the longer a war goes on, the less likely the relative power among belligerents will be clearly illuminated on the battlefield. As a result, Bayesian updating is more difficult, and it is more likely belligerents will need domestic coalition shifts to overcome stickiness in expectations and achieve peace. Belligerents that appear to be losing or stalemated may be more likely to experience coalition shifts, because they are most in need of updating their expectations about the war’s costs and benefits. All domestic coalition shifts, but especially those among primary belligerents, are expected to hasten the war’s end. H shift: The longer the war, the more likely a domestic coalition shift is needed to end it. H loser: Domestic coalition shifts are more likely among “losing” and “stalemated” states. H primary shift: Domestic coalition shifts in primary belligerents will have a greater effect towards hastening the end of the war than shifts in other belligerents.

Extending the Model: Coalition Shift Directions Now that I have explained the role of domestic coalition shifts in ending war, in this section I return to Iklé’s concept of “hawks” and “doves” to examine what effect, if any, the direction of the coalition shift (e.g. hawkish, dovish or neutral) has on the theory. Like scholars who examine the role of hawks and doves in interstate rivalry termination (Mor 1997; Colaresi 2004, 2005; Schultz 2005), I seek to expand the framework laid out by Iklé (1991) to explain interstate war termination.

d omest ic coalit ion shifts



As a reminder, Iklé argues war ends when one belligerent’s doves in effect ally themselves with the other side’s doves and get their respective governments to agree to stop fighting, over the opposition of hawks on both sides. He attributes the difficulty of war termination to the gross asymmetry in political power between hawks and doves in wartime: hawks have a far easier time blocking any settlement than doves have trying to create one. Whenever hawks on either side refuse an opportunity to settle, they reinforce the insistence of the other side’s hawks to continue fighting. By contrast, he argues, doves can only move their states towards settlement when they have clear signals that enemy doves will do the same. Although Iklé’s approach provides a useful attempt to characterize different groups within a governing coalition, it provides few generalizable propositions. My model, which I nuance here to consider the direction of governing coalition shifts, attempts to fill this void. In this section, therefore, I postulate about what effect coalition shift direction has on war duration. In relation to war termination, domestic coalition shifts can be in one of three directions. A “hawkish” shift is one in which the new coalition promises to take actions that move away from ending the war, such as expanding the military scope of war, mobilizing additional resources towards war or rejecting compromise offers. This category includes increasing battlefield activities to pursue “peace with honor,” because while peace may be the stated objective, the coalition is still planning to take escalatory actions. A “dovish” shift is one in which the new coalition plans to take actions that move towards ending the war, such as reducing the military scope of war, reducing resources for the war in favor of other domestic or foreign policy aims, or offering peaceful settlement terms or compromises. This category includes scapegoating previous leaders, as a means to save face while accepting peace. A “neutral” shift is one in which the new coalition intends to maintain the status quo and carry on the same policies vis-à-vis the war as the previous coalition. It is important to note that the direction of these shifts is dependent on the ex ante intent of the new coalition rather than any subsequent policies pursued by it.28 Initially, a coalition may shift in a more hawkish direction. For example, during the initial mobilization for the war, the governing coalition is likely to spin up the public with a “rally round the flag” effect (Mueller 1985; Russett 1990)— causing coalition members who were initially against the war to adopt hawkish positions for political expediency or leave the coalition. As John Vasquez argues, hostility from the adversary helps to “create a constituency for hardliners” and reinforce the hawks’ position within the government (1993: 199; see also Colaresi 2004, 2005; Snyder 1989; Hagan 1993). During the high-threat environment of war, hawks have an easier time selling their policies than doves. While the “rally”

2.3 -- 61 %

56

the d omest ic p olit ics of war ter minat ion B Hawkish change

A Hawkish change

A Dovish change

Side A’s bargaining position

B Dovish change

Overlapping

Side B’s bargaining position

bargaining space

f i g . 2 . 3 . How Changed Expectations Create Bargaining Space

effect itself is short-lived, the hawkish shifts it often engenders can be precisely what lead to the obstacles that cause the war to get sticky. Moreover, as the war goes on, if many strategic options are still available, coalition members may feel obligated to try alternative strategies on the battlefield to see if another one will work better. Such reasoning would cause the coalition to try more escalatory policies (e.g. Bueno de Mesquita et al. 2004), which may cause any remaining dovish coalition members to leave the government. Of course, trying more options extends the war. Eventually, if these other options do not bring success, hawks may be delegitimized, which can lead the coalition or domestic constituencies (or both) to reevaluate war policy. This can result in a shift in a dovish direction. Figure 2.3 demonstrates how these domestic level shifts can affect the creation of an overlapping bargaining space with the other side. There are two mechanisms by which domestic coalition shifts can lead to the creation of an overlapping bargaining space between belligerents. The first mechanism occurs when the new governing coalition modifies its own bargaining position vis-à-vis other belligerents. In contrast, the second mechanism occurs when the coalition shift direction provides information for decision-makers on the opposing side to update their expectations and thus change their bargaining position. first mechanism: a coalit ion shift leads to a new bargaining p osit ion

The first mechanism occurs when the new governing coalition modifies its own bargaining position vis-à-vis other belligerents. While the new bargaining position may be affected by beliefs about the adversary, it may also be completely independent of any consideration of the adversary’s behavior. Thus, this mechanism operates unilaterally and independently of strategic interaction with the adversary. In this mechanism, because hawkish coalition shifts may lead a state to take actions that move away from ending the war, we would expect them to be asso-

d omest ic coalit ion shifts



ciated with lengthening the war. Hawkish shifts, therefore, should occur early in the war and be associated with prolonging wars. Similarly, because dovish coalition shifts may lead a state to take actions that move towards ending the war, we would expect them to be associated with shortening the war. Dovish shifts, therefore, should occur temporally close to the war’s end and greatly enhance the probability of the war ending. Unlike hawkish and dovish shifts, neutral shifts are not intended to lead to a change in war policy; thus, their relationship to the overlapping bargaining space at the international level may not be as clear. Neutral shifts are frequently exogenous shifts. For example, leaders can leave for institutional reasons completely orthogonal to the war, such as serving the maximum number of terms in office or fulfilling a term of service as an interim (“acting”) leader. Alternatively, neutral coalition shifts can result from consequences in other policy areas, such as poor economic performance or a political scandal. Neutral shifts can also arise independent from the war for very random reasons, such as death, sickness, assassination or retirement. For these reasons, we would not expect neutral shifts to be clustered at any particular time in the war. Although a coalition shift occurs, the new coalition—which often includes a predetermined successor or members of the same political party—does not intend to change its predecessors’ war policy. However, neutral shifts do bring new leaders, who may have a new way of looking at the war. While the new governing coalition intends to carry on with the same war policy, in practice it may look at the war’s costs and benefits differently and may want to free up resources and political capital to pursue other policy goals. Thus, like interstate rivalry research which conceptualizes regime change as a political “shock” that may cause the rivalry’s end (Goertz and Diehl 1995; Bennett 1997), I argue neutral shifts have the potential to bring to power leaders who will end the war—even if they intend to maintain the status quo. As Scott Bennett argues, “The occurrence of a shock at least raises the possibility that the contract range will expand, while in the absence of a shock the contract range remains unchanged” (1997: 1205, FN 2). In other words, since neutral shifts at least offer the possibility of change, there is a greater chance that a neutral shift may bring the war’s end. Indeed, since the new leaders by definition came to power for reasons not related to the war policy, they could view the war as a distraction. As a result, they may end up taking actions that move towards ending the war in order to focus on other things.29 Thus, we would expect neutral shifts to be associated with ending the war, but not as quickly as dovish shifts. H hawk: Hawkish coalition shifts will occur early in the war and be associated with prolonging the war.

58

the d omest ic p olit ics of war ter minat ion

H dove: Dovish coalition shifts will occur late in the war and be associated with shortening the war. H neutral: Neutral shifts will shorten the war, but not as quickly as dovish shifts. Because war is a strategic interaction, coalition shifts on both sides may be necessary to create an overlapping bargaining space. How many coalition shifts will be necessary to create the overlapping bargaining space depends entirely on how much distance exists between the reservation prices of the two sides.30 If the two sides’ reservation prices are not very far apart, then a dovish coalition shift on one side may lead the new coalition to change its reservation price enough to bring an overlapping bargaining space within reach. However, if the distance between the two sides is greater, coalition shifts in more than one belligerent may be needed. During sticky wars that endure for many months or years, multiple coalition shifts may be necessary to dislodge the obstacles to peace and create the overlapping bargaining space. These multiple shifts could theoretically occur in multiple belligerents or within one belligerent over time, each shift moving that belligerent’s reservation price closer to the others’. second mechanism: a coalit ion shift leads to a new bargaining p osit ion on the opp osing side

A second way that coalition shifts can work to create an overlapping bargaining space between belligerents is through signaling. This mechanism is clearly related to the strategic interaction between the opposing sides, because actions that one side takes are directly linked to the other side’s response. In this mechanism, the direction of the coalition shift communicates to belligerents on the opposing side, conveying information about the first side’s future intentions and likely bargaining position. For example, if the new governing coalition campaigned with a platform to end the war, it might help doves on the other side move their own state towards peace. Alternatively, if the new governing coalition campaigned to escalate its war aims, it might help the other side recalibrate its understanding of the costs necessary to fight an adversary with a renewed will and a greater willingness to commit resources to win. While Bayesian models have yet to include this source of signaling, domestic coalition shifts can provide information for updating beliefs, just as battle outcomes and diplomatic offers do. In fact, the direction of a coalition shift reveals information about the adversary’s “type”: hawkish shifts suggest that the adversary is strong or has a high tolerance for war costs, while dovish shifts suggest that the adversary is weak or cost-sensitive. Certainly, signaling is vulnerable to manipulation, deception or bluffing (Jervis

d omest ic coalit ion shifts



1989; Dixit and Skeath 1999: chapter 12). However, following Jervis (1989), I argue that domestic coalition shifts are “indices” rather than “signals.” Signals are statements or actions which actors use to influence the receiver’s image of the sender, and they can be easily manipulated. In contrast, indices are statements or actions “that carry some inherent evidence that the image projected is correct because they are believed to be inextricably linked to the actor’s capabilities or intentions” (Jervis 1989: 18). As a result, most observers believe indices cannot be manipulated or used for deception, and they often rely heavily on them when signals are suspect. Indeed, Jervis (1989: 34) includes domestic events, such as elections, as an example of an index. Therefore, if a belligerent looks to its adversary’s domestic coalition shift, it may believe the information being communicated is very credible. With this in mind, say, for example, that Belligerent A experiences a dovish coalition shift. Seeing this shift, Side B may surmise that A will probably lower its demands, especially if the new government campaigned on a platform to end the war or to redirect its policy focus and resources to other issue areas. Alternatively, if A’s new governing coalition campaigned to escalate its war aims, seeing this new hawkish coalition take power might help B realize war costs are likely to escalate. If they are capable of perceiving and learning from this information, it can even lead them to change their own bargaining position. Thus, with this second mechanism, a domestic coalition shift can lead the other side to change its bargaining position. For this reason, hawkish shifts may signal to the opposing side that the costs of the war will dramatically increase in the future. This could cause the other side to sue for peace, thereby shortening the war. In contrast, dovish shifts may indicate weakness, so we might expect the opposing belligerents to push for even more concessions beyond what the new coalition is willing to give—causing the war to lengthen. Because neutral shifts by definition are not related to the war, they are less likely to provide information to the other side. Thus, we would not expect them to operate through this second (signaling) mechanism. H hawk*: Hawkish coalition shifts will cause the other side to sue for peace and thus shorten the war. H dove*: Dovish coalition shifts will cause the other side to push for more concessions and thus lengthen the war. H neutral*: Neutral coalition shifts are unlikely to provide information about the war to the opposing side and thus should not affect the war’s duration through the second mechanism. Given the ambiguities inherent in international signaling, we would expect the second mechanism’s effect to be less powerful than the first’s. Moreover, the

60

the d omest ic p olit ics of war ter minat ion

second mechanism can only work when the information obstacle is not blocking the opposing side’s rational updating process. Nonetheless, these alternative predictions reinforce the importance of accounting for strategic interaction in models of war termination. Given the theoretical validity of both sets of predictions, I will return to them in Chapters 8 and 9, to weigh their relative importance empirically.

Conclusion War termination is a bargaining process between two interacting sides. While bargaining models of war suggest that wars end after an overlapping bargaining space develops, my argument suggests how, when and why it develops. This approach offers real insights about the timing of war termination, including the under-theorized category of protracted wars, which bargaining models have had little success in explaining. I have presented two alternative pathways for interstate war termination, building on Bayesian models. While Bayesian models suggest that a lag in updating can delay war termination, they do not explain how states get beyond this lag—why some states get bogged down in updating their expectations while others can update very quickly. In contrast, I look more carefully at this lag and theorize about how states get beyond it. In this way, my explanation is more complete than previous attempts to generalize about war termination. Its major theoretical contribution is its attention to domestic coalition shifts as an important mechanism for why states come to the bargaining table to end interstate war. Ending war requires settling with the enemy as well as at home. This chapter has suggested three reasons why a war, once entrenched, is difficult to end. A preference obstacle occurs when leaders do not want to end the war. This happens, for example, when fighting the war helps leaders to stay in power or gain other benefits. An information obstacle occurs when leaders do not know they should end the war. In this case, the rational updating process does not occur, because leaders (1) receive poor quality information, (2) do not necessarily share the same information, (3) use different indicators to assess the war or (4) have personal cognitive or (5) organizational biases that prevent them from using all available information. An entrapment obstacle occurs when leaders want to end the war but cannot. This happens when leaders face political entrapment from a domestic constituency or external allies. While the logic of these three obstacles is distinct, they can often occur in tandem. For example, leaders that hold strong preferences are also likely to fall victim to information processing biases to reinforce pre-existing beliefs. Similarly, the psychological commitments and domestic political constraints of entrapment can make leaders reluctant to confront

d omest ic coalit ion shifts



painful truths, thereby creating ample opportunities for the various dynamics of the information obstacle to take hold. In addition to the three obstacles, policy stability can lead governing coalitions to be unable to change the status quo, even when such changes are necessary or desirable. Indeed, a governing coalition with policy stability may become so immobile that policy change is only possible by replacing the incumbent coalition entirely. For all of these reasons, wars do indeed become sticky: the longer the war, the harder it is to end it. Thus, while Bayesian models assume the change in expectations to produce an overlapping bargaining space occurs when incumbent leaders change their minds, I argue that in many wars, such change in expectations results from a change in the foreign policy leadership itself. The longer a war goes on, the less likely the relative power among belligerents will be clearly illuminated on the battlefield, and the more likely belligerents will need domestic coalition shifts to achieve peace. A coalition shift allows different political actors—with different political interests, assessments of the war and constituencies—to take power. This, in turn, can lead to a shift in the war policy, from continuing it to ending it. A coalition shift helps to overcome the obstacles to peace and thus permits the war to end. All domestic coalition shifts, but especially those among primary belligerents, are expected to hasten the war’s end. Finally, the last part of this chapter extended the model to consider what effect coalition shift directions have on war duration. I outlined two mechanisms by which domestic coalition shifts can affect the creation of the overlapping bargaining space at the international level. The first mechanism suggests coalition shifts can lead the new governing coalition unilaterally to change its own bargaining position vis-à-vis other belligerents. In contrast, the second mechanism suggests the direction of a coalition shift serves as an index to belligerents on the other side, conveying information about the first side’s future intentions and likely bargaining position. Given the ambiguities inherent in international signaling and the potential for information obstacles to block the opposing side’s ability to learn from coalition shifts, we would expect the second mechanism’s effect to be less powerful than the first’s. The next part of the book will examine these dynamics in the Korean War.

3

The Korean States Powerless Players in the Forgotten War

The Korean War is often dubbed the “forgotten war” (e.g. Blair 1987)—wedged between the sheer global magnitude of its predecessor, World War II, and the American domestic ordeal of its successor, Vietnam.1 In studies of the war, scholars have generally focused on its dramatic first year—the war’s outbreak, the Inchon landing, the Chinese intervention and the Truman-MacArthur controversy. Despite a recent revival of interest, as nations commemorated the war’s 50th anniversary and as Soviet and Chinese archival materials supplement the story,2 very little scholarship about the last two years of the Korean War exists. These years are the forgotten part of the forgotten war.3 However, study of the war’s last two years brings to light a tragic picture. When the Korean War finally ended in 1953, total casualties were estimated at 1.5 million killed and 3.6 million wounded (Kang 2005: 196). The tragic irony is that the final armistice terms the United Nations Command accepted in July 1953—after 575 meetings—were little different from those they proposed at the start of negotiations in July 1951 (Foot 1993: 193). Given that the battlefield situation remained relatively constant throughout these two years, why were these terms not accepted two years earlier? The six chapters in Part II will try to answer this question through case studies of the war’s primary belligerents—the Soviet Union, the United States and the People’s Republic of China—and a detailed examination of the strategic interaction between them. This chapter sets the stage for the discussion in three ways. First, it provides a brief background on the war, with key dates in the armistice process. Second, it explains my decision to include the Soviet Union as a primary belligerent, even

m a p 3 . 1 . The Korean Peninsula, Showing Lines of Advance, 1950–53. Source: Central Intelligence Agency, Map #802191 (1993). Data sources for lines of advance: McMahon (2003: 54); Hickey (1999: xiv–xv); Chen (2001: 86).

the korean states



though it was not a declared participant in the war. Third, it explains the decision to exclude the two Koreas as primary belligerents: although the war was fought on their soil, neither Korean party had much weight in its respective side’s final decision about when and how the war was ended. The chapter concludes with mini-case studies of the Koreas to demonstrate that although both states experienced domestic coalition shifts during the war, these shifts did not affect the international-level strategic interaction. In contrast to the primary belligerents, whose domestic coalition shifts are hugely important for explaining when, how and why the war ended, the Koreas’ powerlessness meant that their domestic politics are less relevant to war termination.

Overview of the War The Korean War began on June 25, 1950, as the North Korean People’s Army (NKPA) surged south across the 38th Parallel. On June 30, US President Harry S. Truman authorized the use of US ground troops to support the South Koreans.4 One week later, at the United Nations’ Security Council’s request, General Douglas MacArthur was appointed commander of all UN forces fighting against the North Koreans. During that summer, the North Koreans pushed their way down the peninsula, eventually pinning UN forces into a small area around the southeastern port of Pusan (see Map 3.1). By the fall, the United States had brought enough forces into the region to change the balance. On September 15, MacArthur conducted a successful amphibious landing at the port of Inchon, west of Seoul and deep behind enemy lines, enabling UN forces to break out of the “Pusan perimeter” and push rapidly northward (Stueck 1995: 85–87; Hickey 1999: 72–81). On September 30, South Korean troops crossed the 38th Parallel into North Korea, and American troops followed one week later (Goulden 1982: 233–45; Matray 1979: 72–81). Fearing for its own security, China gave warnings that it would intervene if the UN continued its advance north. On October 18, Chinese leader Mao Zedong issued orders for the Chinese “volunteer” forces to “resist the attacks of US imperialism” (Stueck 1995: 99; Foot 1985: 85). Chinese forces entered North Korea on October 19 and fought intermittent battles, eventually breaking off contact with UN forces on November 16. MacArthur continued his drive northward towards the Yalu River, launching “the end-of-war offensive” on November 24. That same day, the Chinese counterattacked with hundreds of thousands of soldiers, thoroughly routing the UN forces (Zhang 1995: 95–115). Two days later, MacArthur announced that his forces faced an “entirely new war” (Rees 1964: 157). Indeed,

68

the korean war

in two days, November 30 and December 1, UN casualties exceeded 11,000 (Foot 1985: 101). The new Communist offensive pushed southward again across the 38th Parallel on January 1, 1951, and by late January, had pushed the UN back south of Seoul. With the new war’s full dimensions clear, Britain and India pressed the United States to begin negotiating a cease-fire, based on a deal that would include admitting China to the UN and abandoning Taiwan (Stueck 1995: 138–42). On January 13, the United States agreed to the British- and Indian-sponsored UN cease-fire proposal. However, on January 17, China rejected this proposal, arguing this would only give UN forces breathing space (Foot 1990: 29). In April 1951, Truman dismissed MacArthur for speaking out publicly to preempt a US statement about its willingness to settle. The new UN ground commander, General Matthew Ridgway, revived the UN side’s battlefield fortunes as the Communist offensives in April and May stalled out. By June 1951, the battle line had more or less stabilized and each side began to realize that moving beyond a stalemate would be extraordinarily costly. On June 23, Soviet UN Ambassador Jacob Malik suggested in a radio address that both sides agree to an armistice at the 38th Parallel (Stueck 1995: 208; Rees 1964: 262). Armistice negotiations between the UN Command, China and North Korea began on July 10 in the Korean town of Kaesong. To contemporary observers, agreement seemed only a matter of time: “The Joint Chiefs thought the talks would last about three weeks, but being military men they also had to be realistic. It might take as long as six weeks to get an armistice” (Rees 1964: 285). By the end of July, the belligerents agreed on the armistice talks agenda: 1. Adoption of an agenda; 2. Fixing a military demarcation line to establish a demilitarized zone as a basic condition of a cease-fire; 3. Concrete arrangements for realizing an armistice, including the composition, authority and functions of a supervising organization; 4. Arrangements relating to prisoners of war (POWs); and 5. Recommendations to the belligerent governments on both sides (Vatcher 1958: 33–34, 38–43; Joy 1955: 18–29; Hermes 1966: 26–31; Goodman 1978: 14–22; Bailey 1992: 76–77; Ridgway 1967: 202). That it took the negotiators several weeks to agree upon an agenda foreshadowed future frustrations. On August 24, using as a rationale the latest in a series of “incidents” around the negotiating area, the Communists broke off formal negotiations (Stueck 1995: 227–35; Foot 1990: 42–50; Goldhamer 1994: 76–82; Boose 2000). Formal talks resumed in October 1952 in Panmunjom. The next

the korean states



few months were very productive, so that by March 1952, all items were resolved except for agenda item 4—prisoners of war (Foot 1990: 53, 74–76: Collins 1969: 338–39). In effect, POW repatriation stalled the talks for an additional 15 months. In January 1952, the UN Command had proposed non-forcible (voluntary) repatriation for prisoners of war (Joy 1955: 178). In contrast, the Communists insisted that all prisoners must be returned, even those who did not want repatriation. By March, the Chinese negotiators hinted that they might accept some compromise solution: a UN screening of the prisoners and subsequent removal from the lists of those who would violently resist repatriation (Bailey 1992: 94). This Communist concession came after the UN team had suggested such a screening would result in the return of 116,000 Communist prisoners—including almost all Chinese POWs (Rose 1994: 291–92). In early April, therefore, the UN screened prisoners. The results shocked everyone: only 70,000 of the 132,000 agreed to return, including only 5,000 of the 21,000 Chinese prisoners (Bernstein 1983: 281–84). When the UN presented these figures on April 19, the chief Communist negotiator, “struggling to conceal his shock and outrage, immediately moved for a recess” (Bernstein 1983: 284; Foot 1990: 99). On April 28, the UN delegation offered a package proposal about prisoners, airfield reconstruction and Soviet membership on the neutral nation commission. On May 2, the Communists officially agreed to forgo Soviet membership in exchange for the ability to rebuild airfields. With this agreement, the POW issue became the only remaining issue in contention (Foot 1990: 102), and on May 7, talks broke down because of it. The remaining eight months of the Truman Administration saw little progress in the negotiations. Over the summer, the UN Command increased its air campaign, and on September 28, it delivered its final offer on the prisoner of war issue. On October 8, this offer was rejected, and the UN announced an indefinite recess to the talks (Boose 2000). There was no movement on negotiations until March 30, 1953—three weeks after Soviet leader Joseph Stalin died—when the Communists conceded on POW repatriation. The previous month, UN Commander Mark Clark had proposed an exchange of sick and wounded prisoners, and on March 27, Chinese premier and foreign minister Zhou Enlai agreed, adding that such an exchange “should be made to lead to the smooth settlement of the entire question of POWs, thereby achieving an armistice in Korea for which people throughout the world are longing” (Hermes 1966: 412; Foot 1990: 167). Three days later, Zhou publicly proposed that POWs unwilling to be repatriated should be transferred to a neutral state (Chen 2001: 112).

70

the korean war

Armistice negotiations resumed on April 26, and by June 8, the Communists agreed to the principle of voluntary repatriation. On June 18, South Korean President Syngman Rhee—in a last-ditch effort to undermine the negotiations—unilaterally ordered South Korean guards to release 25,000 North Korean prisoners resisting repatriation. During the following month, Communist troops launched an offensive at ROK positions (Zhang 1992: 140; 1995: 243), while Eisenhower’s Administration offered Rhee various carrots to secure his agreement to honor the armistice (Allen 1994: 106; Keefer 1986: 286–87; Kotch 1983: 239–59; Goulden 1982: 640–42; Kaufman 1986: 328). The armistice ending the Korean War was finally signed on July 27, 1953. By the time the armistice was signed, about three million civilians were killed in the North and the South, and many more became refugees. Total battlefield casualties for both sides were 3.6 million wounded and 1.5 million killed (Kang 2005: 196). US casualties numbered more than 157,000, including 54,200 dead; South Korea suffered more than 1.3 million casualties, including 415,000 dead. Other UN casualties numbered more than 16,000, including 3,100 dead. Communist casualties were estimated at more than 2 million (Kaufman 1999: 15; Foot 1990: 208). Sadly, the last two years were extremely costly for all belligerents. Had the war ended two years earlier, when armistice negotiations began, the United States would have been spared 45 percent of its casualties—22,000 killed and 63,200 wounded. While Communist statistics are less certain, at least 234,000 Chinese and North Koreans would have lived had the war ended in July 1951 (Foot 1990: 208; Hermes 1966: 500; Hatchimonji 2002).

Including the Soviet Union as a Primary Belligerent The Soviet Union was not a declared participant in the Korean War, and for this reason, it is often excluded from Korean War histories or quantitative analyses of war. For example, Bruce Cumings—one of the preeminent “revisionist” historians of the war—argues that the Soviet Union played a small role, because Soviet influence over North Korea was “flimsy.” In his estimation, North Korean leader Kim Il-Sung could have acted independently of Moscow during the war, because it “was by no means reliant solely on Soviet arms” (Cumings 1990: 447).5 Similarly, many quantitative datasets used in international relations scholarship, such as the Correlates of War, exclude the Soviet Union because it neither was a declared participant nor suffered enough battle-deaths to qualify for inclusion as a belligerent.6 Indeed, skeptics might wonder how useful the Soviet case is for theorizing

the korean states



about war termination. How often does a state have almost complete control over war decisions, reaping benefits without incurring the costs of fighting? Four states in other interstate wars since World War II had this dynamic operating to some degree: France and Britain in the Sinai War (1956), Greece in the Cyprus War (1974) and Syria in the Lebanon War (1982). Thus, 15 percent of the interstate wars since World War II involve states controlling war decisions without committing large numbers of ground forces to the war. Moreover, this dynamic is quite common in civil wars, where an outside patron to one of the civil war factions supplies arms and effectively controls war decisions. For example, Yugoslav President Slobodan Milosevic played a central role in Bosnian Serb war decision-making during the civil war in Bosnia-Herzegovina in the 1990s. Likewise, the United States used its role providing arms to the mujahedeen to influence war policy during the Soviets’ war in Afghanistan in the 1980s. Finally, regardless of how “representative” the Soviet case is, including it allows me to examine the dynamics of the preference obstacle (see Chapter 4) and thoroughly explore the empirical record of all belligerents in the Korean War. I contend the Soviet Union was a primary belligerent in the war for three reasons. First, although the Soviet Union did not fight on the ground in Korea, it provided massive quantities of armaments and supplies and significant numbers of military advisors to the Chinese and North Koreans. The North Korean army was formed with Soviet occupation forces’ weapons and equipment (when the Soviets withdrew in 1948), and the Soviets continued to provide all weapon upgrades and ammunition before and during the war. Before the war, the Soviets also supplied sufficient equipment for six infantry divisions, three mechanized units, eight battalions of border troops and combat aircraft (Zubok and Pleshakov 1996: 55; Shrader 1995: 65–66; Whiting 1960: 42–43; Griffith 1967: 344–45; Simmons 1975: 191). Richard Thornton argues this pre-war buildup “doubled the number of North Korean planes, tripled its infantry divisions, more than doubled and possibly even quadrupled the number of tanks and massively increased the NK[P]A’s artillery and mortar capability” (2000: 110). In terms of advisors, since December 1946, North Korean units had regularly assigned Soviet advisors. When the war broke out, the Soviets also sent technical advisors to help maintain and operate Soviet-supplied arms and equipment. There were roughly 10,000 Soviet and Soviet-bloc anti-aircraft artillerymen, advisors and technicians in North Korea at the time of the armistice, and the Soviets also staffed field hospitals (Hermes 1966: 283–84; Shrader 1995: 92; Ognev 1993: 58). As Chapter 4 will detail, evidence also suggests the Soviets planned the North Korean invasion (Stueck 1995: 31; Thornton 2000: 110–11). Yu Song-Chol, who served as Kim Il-Sung’s Russian translator, argues the draft invasion plan

72

the korean war

was “prepared directly” by the Soviet military advisory group, headed by Lt. Gen. Vasiliev.7 Vasiliev passed the plan to Kim Il-Sung, who passed it to Kang Kon, the chief of the NKPA general staff, who in turn passed it to Yu and instructed, “You translate this and set up plans” (Yu 1993: 152; Chu 1993: 116; Thornton 2000: 110). Once the North Koreans invaded, however, Stalin ordered the Soviet advisors to pull back from the front lines (Im 1993: 138). The Soviets provided most of their military aid to the Chinese after the Chinese People’s “Volunteers” (CPV) intervened in Korea in October 1950. Although the Soviets sent 3,000 military advisors and technicians to China in April 1950,8 they did not provide Soviet armaments and materiel until August 1951. Thereafter, both the Chinese People’s Liberation Army (PLA) and the CPV in Korea were reorganized and re-equipped with modern Soviet equipment. The Soviets equipped 64 Chinese infantry and 22 air divisions, and they also provided 80 percent of the ammunition used by Chinese forces in Korea—3 million tons of military materials and 250 thousand tons of ammunition (Hao and Zhai 1993: 162; Shrader 1995: 66; West 1993: 9; Futrell 1983: 245). In February 1951, Moscow agreed to build factories to repair and service aircraft in China. China converted these factories to make new aircraft, so that by the end of the war, its air force had more than 3,000 planes, including advanced MiG jets. The Soviet Union also transferred technology for producing seven kinds of small arms, including machine guns (Jung and Halliday 2006: 361). However, this “military assistance” to China was not free but came with many strings attached, as Chapter 7 will detail.9 Second, the Soviet Union engaged in an intense air war with UN forces for more than two years (Halliday 1993; O’Neill 1996). Although information about the Soviet air war in Korea was tightly guarded as a “state secret” throughout the Soviet era, some 26,000 Soviet airmen fought in the Korean War and claimed to have destroyed 1,309 UN aircraft. Two Soviet air divisions were sent to defend the Yalu River in November 1950, and in March 1951, a division of the Moscow Air Defense District was deployed to Manchuria. The Soviet aircraft bore North Korean markings. The Soviet pilots were required to wear Chinese uniforms and trained to use Chinese and Korean phrases on the radio, but during crisis, they often reverted to Russian. They were also prohibited from flying over enemy territory, to avoid being taken prisoner if shot down. Moreover, Stalin ordered the Soviet air force to train Chinese pilots as quickly as possible to replace Soviet air crews. The Soviets sent their best pilots to train the Chinese: while it typically required at least 300 hours of training to fly a jet aircraft, the Soviet instructors adapted a 60-hour course to teach the Chinese to pilot a MiG-15 (Zhang 1995: 177; Lobov 1991: 5; Zaloga 1991: 76; Edwards 2000: 68–69; Weathersby 1993c: 437–38;

the korean states



1996; 2002: 19–20; Hao and Zhai 1993: 162; Junji 1991: 21–23; Smirtchkov and Vetalyayev 1991: 25, 29; Park 1996a: 207–10; Volkogonov 1993; Petrov 1994b: 62). Third, the Soviet Union played a central role in formulating diplomatic strategy for the Communists during the war, and thus drove decisions about prosecuting and ending the war. Stalin had final say in decisions about military and diplomatic strategy, regularly corresponding by telegraph with Chinese and North Korean leaders. Stalin himself negotiated with Mao and Kim Il-Sung over the amounts of various supplies to be delivered, their delivery schedule and the terms of payment. Although the Chinese were responsible for operational management of the war after November 1950, Stalin closely followed events on the ground and periodically intervened with specific instructions. Most importantly, Mao and Kim requested and received Stalin’s instructions prior to any significant diplomatic or strategic moves (Weathersby 1996, 1998; Chen 2001: 59; Sheng 1995: 309; Shen 2003–4; Thornton 2000: 380; Goncharov et al. 1993: 42). Indeed, as Chapters 4 and 7 will show, Stalin’s sway over Mao and Kim was a major factor in the war continuing. Perhaps because Kim’s overly optimistic projections about quick victory were soon proven wrong, Stalin frequently intervened with and overruled the North Korean command—usually siding with the Chinese. Recent archival evidence shows Stalin often arbitrated between North Korea and China, intervening to “help” them establish a combined command structure under Chinese control and phase offensive campaigns per Chinese military advice. Stalin even interjected to place the Korean railroad system under Chinese control, over Kim’s objections concerning Korean sovereignty (Shen 2003–4; Weathersby 1996; 1998: 91; Chen 2001: 59; Sheng 1995: 309). Stalin also intervened with the Chinese, reserving Sino-Soviet affairs for himself among top Soviet decision-makers. Until his death, Stalin dealt personally with all problems, big and small, between the Soviet Union and the Chinese Communists (Li 1992: 61; Goncharev et al. 1993: 14–15; Zubok and Pleshakov 1996: 56; Petrov 1994a: 5; Shi 1993: 70, 88–89). In sum, the Soviet Union is included as a primary belligerent in the war because it bankrolled the Communist war effort with materiel and advisory support, provided most of the pilots for the Communist air war and—most importantly— played a central role in formulating decisions about prosecuting the war. Therefore, any analysis of the war termination process must include the Soviet Union.

Excluding the Korean States as Primary Belligerents If the last two years of the Korean War have received scant treatment in popular histories and scholarly analysis, then the role of the two Korean states in these

74

the korean war

accounts has been virtually non-existent. Recent scholarship, based on newly opened archives, is starting to piece together the domestic political story for both Koreas, but the process is far from complete.10 Undoubtedly, as additional information gets released and Korean scholars’ work becomes translated, our understanding of the Korean states’ perspectives will be strengthened. Yet, even if the historical account were more complete, North and South Korea still do not qualify as primary belligerents in the Korean War. Although the war was fought on their soil and both sides suffered tremendous civilian and military casualties, neither Korean party had much power in its respective side’s decisions about when and how the war ended. Their respective powerlessness provides an ironic parallel: while North Korea wanted to end the war much sooner, South Korea wanted to fight on much longer, but both were overruled by their respective side’s more powerful primary belligerents. Clearly, without their great power patrons, neither Korean state had the capacity to wage this war. The NKPA existed almost entirely because of Soviet arms and advisory support; it seems unlikely it could have launched the invasion without that assistance. Similarly, after the North Korean attack, the ROK army would have been decimated and pushed off the peninsula if not for US assistance. After the UN’s successful amphibious landing at Inchon in September 1950, it was the North Korean’s turn to be almost overrun—until China intervened to save it. Therefore, both Korean states were materially powerless and depended heavily on their patrons. Part of this relative powerlessness resulted from the way North and South Korea were perceived by other states in the conflict. All three primary belligerents viewed the Korean states as pawns in a Cold War battleground. For example, in her social constructivist account of the war, Jennifer Milliken (2001) argues US officials typified North Korea as a “Soviet satellite state”—equating North Korea with Eastern European states “whose governments were under complete Soviet domination and control, and which therefore acted as ‘puppets’ in Soviet courses of action” (2001: 49). In the same way, US officials typified South Korea as a “bona fide client state”—equating South Korea with Japan and West Germany, states “fully ‘under US protection’ and dependent on US backing if it was to combat Communism successfully” (2001: 51). Conversely, China framed the conflict as a “US war of conquest in Asia” with the United States planning to conquer the entire Korean peninsula and extend US imperialist domination in Asia. Thus, the Chinese typified South Korea as a “US-backed ‘reactionary puppet regime’,” while framing North Korea as a “brother country” carrying out “a revolutionary struggle for national liberation” (Milliken 2001: 117). While Milliken does not examine Soviet framing of the Korean actors, Stalin’s dismissal of North Korea’s interests and losses—detailed below—suggests the Soviets had a similar perspective.

the korean states



Typifying the Korean states this way effectively eliminated them from the strategic interaction between the two sides in the war. Neither North nor South Korea was perceived as an independent actor, with its own interests and decision-making processes. As a result, both Koreas were almost irrelevant to the international-level bargaining process. For example, given its understanding of North Korea as a “Soviet satellite state,” the United States missed the possibility to interact directly with the North Koreans or exploit differing interests and perspectives in the Communist camp. In sum, North and South Korea were powerless in their respective side’s decisions and actions in two ways: not only were they materially powerless, as their forces were incapable of waging war without their larger patrons’ help, but the way the three primary belligerents conceived of them as client states rendered them politically powerless as well. Within this perceptual context, even if the Korean states had been more powerful materially, it is unlikely they could have influenced their respective side’s war policy. The rest of this chapter provides a brief overview of the wartime domestic politics in both Korean states to demonstrate their relative insignificance for explaining when, how and why the war ended. north korea

With the exception of its idea to launch the war, North Korea was never central to Communist decisions about war policy. Indeed, after getting Soviet support to launch the attack in April 1950, the North Koreans consistently forfeited the driver’s seat in military and diplomatic strategy to the Soviets and the Chinese. As noted previously, the Soviets planned the invasion, and Stalin supervised high-level diplomatic and military strategy decisions. Then, after the Chinese intervened to save Kim Il-Sung from failure, they handled operational and tactical decisions, first on the battlefield and then in the armistice tent. By December 7, 1950, the day after Chinese troops recovered Pyongyang on their southward offensive, Kim ceded command to CPV military commander Peng Dehaui (Shen 2003–4: 13–14; Suh 1988: 122). By this point, Kim’s depleted 75,000 man army was outnumbered 6:1 by the CPV. As Jung and Halliday argue, “Mao had taken over Kim’s war” (2006: 359). Furthermore, although North Korean Nam Il was the titular head of the Communist negotiating team for armistice negotiations, the North Koreans had no real power. Behind the scenes, the Chinese established a “negotiation direction group,” led by Vice Foreign Minister Li Kenong and Qiao Guanhua, the head of the Foreign Ministry’s public information bureau. Reluctantly, the North Koreans agreed that this group, which communicated daily with Beijing, would direct the negotiations (Stueck 1995: 224; Chen 2001: 99). As Shen Zhihua put it, “the

76

the korean war

way the Chinese, backed by the Soviets, forced the North Koreans to accept their views left a shadow on the heart of Kim Il-Sung” (2003–4: 19). In other words, once the larger states got involved in the war, the North Koreans held little sway in Communist decisions to prosecute and end it. The DPRK regime was led by Kim Il-Sung, the chairman of the Korean Workers’ Party (KWP). When the KWP was created in 1949, merging the Communist parties in both North and South Korea, it brought together three factions: (1) the Soviet/Manchurian faction, led by Kim Il-Sung, which included Koreans who had lived for extended periods in the USSR before 1945 and the “Kapsan” guerilla fighters during Korea’s Japanese occupation; (2) the South Korean faction, led by Pak Hon-Yong, which had operated primarily in South Korea until 1948;11 and (3) the Yenan faction, led by Kim Tu-bong and Mu-Chong, which had fought with the Chinese Communists (Stueck 1995: 22; Park 1995: 242–43; 1996b: 108; Simmons 1975: 23–24). After the KWP’s creation, the Soviet faction controlled the military’s highest ranks,12 while the Yenan faction was a majority in its secondary tiers. Nonetheless, Pak Hon-Yong’s experience in South Korean politics quickly catapulted him to the second leadership position; he became KWP vice chairman, first deputy premier of the government and foreign minister. As a result, the South Korean faction assumed almost half of the KWP politburo and central committee seats (Park 1995: 243–44; Park 1996b: 261; Thornton 2000: 111).13 There has been some debate about which leader—Kim Il-Sung or Pak HonYong—was responsible for the June 1950 attack. Some observers argue the war resulted from a power struggle between the two leaders, with Pak launching the attack to eclipse Kim and the Soviet faction (Simmons 1975: 105–9; Suh 1988: 120–21). Certainly, Pak was in a pivotal position to influence the decision, given his near-monopoly of information on conditions in the south, and Stalin was persuaded by Pak’s estimate of overwhelming South Korean support for the invasion (Petrov 1994b: 65; Shin and Shin 1993: 106). However, more recent scholarship suggests Kim was the driving force behind the invasion, with Pak’s support. Both leaders attended all of the meetings with Stalin to request Soviet assistance for the invasion. Moreover, while Pak provided data about guerilla activities and spy penetration in the south, Kim had other sources of information that supported Pak’s assessments (Park 1995: 245; 252; 1996: 264–66; Kim 1996: 267–69).14 The only high-ranking North Korean leader who opposed the invasion was Choe Yong-Gon, the defense minister and NKPA commander, who was consequently left out of the invasion planning (Park 1995: 258; Kim 1996: 287–304; Thornton 2000: 111–12). The day after the invasion, North Korea experienced the first (hawkish) shift in its domestic governing coalition. On June 26, 1950, Kim Il-Sung created a seven-

the korean states



man Military Committee to coordinate the war effort and direct all government functions during the war. He also streamlined the chain of command to bypass Defense Minister Choe. Because of Choe’s support in the army, Kim was unable to remove him from his official posts. However, the de facto chain of command now ran from Kim Il-Sing to Vice Premier and “battlefield commander” Kim Cha-ek to NKPA Chief of Staff Kang Kon (Park 1995: 259, 261–66; Yu 1993: 153; Thornton 2000: 111–12). In addition, the Military Committee was vested with authority “to mobilize ‘the entire forces of the country’ and all citizens, organs of power, political parties and social organizations were ‘obliged to submit themselves completely’ to it” (Nam 1974: 85). Besides Kim Il-Sung, the Military Committee included three other members of the Soviet faction—Vice Premier and Industrial Minister Kim Cha-ek, Defense Minister Choe and Chong ChunTaek—and two members of the South Korean faction—Foreign Minister Pak and Vice Premier Hong Myong-Hui. The seventh member was Interior Minister Pak Il-U, the only representative of the Yenan faction (Kim 1975: 177; Nam 1974: 85). After North Korea’s initial battlefield success, only to be pushed back north towards the Yalu River in the fall, China intervened and took over war policy decision-making. Two weeks after China pushed UN forces back south of Pyongyang, the KWP Central Committee on December 21 held its third joint plenum. There, Kim publicly admonished military leaders, provincial party bosses and eight major party leaders for North Korea’s poor war performance. Of those criticized, the highest ranking included Yenan faction leader Mu-Chong, Soviet faction member and Vice Defense Minister Kim Il, and the KWP party organizer, Kim Yol, a Soviet-born Korean.15 During his speech, Kim Il-Sung fired all eight major party leaders from their positions and expelled them from the party. He reserved his strongest criticisms for Mu-Chong, accusing him of refusing to defend Pyongyang during the retreat north and committing “barbarous atrocities” against the Korean people (Kim 1975: 178–79; Nam 1974: 87–88; Suh 1988: 122–23; Lim 1982: 198, 204–5; Simmons 1975: 186).16 Kim Il-Sung’s scapegoating of subordinates at the joint third plenum is the second North Korean domestic coalition shift during the war. While it may appear to be a dovish coalition shift, as he tried to shift blame for the military defeats onto others, Kim did not scapegoat these leaders intending to change war policy towards ending the war. Instead, on closer examination, it appears the major purpose of the change was to eliminate his rival Mu-Chong, the Yenan (Chinese) faction leader. Given China’s recent intervention to save North Korea from complete collapse, Kim Il-Sung worried that this faction, now supported by China’s military, would become a greater threat to his leadership. Eliminating this faction’s leader was the easiest way to avoid that outcome, while publicly

78

the korean war

criticizing leaders from all three factions allowed him to deflect attention from this purpose (Kim 1975: 179). Indeed, shortly after the plenum, every other expelled party leader was reinstated, albeit often to different positions. Only Mu-Chong was permanently expelled, stripped of military rank and imprisoned (Kim 1975: 178–79; Suh 1988: 123; Lim 1982: 198, 205; Nam 1974: 88).17 As Joungwon Kim argues, the purge “shifted the onus of defeat from Kim Il-Sung and the Soviet Union, while maintaining the fiction that it was the South that had initiated the war, and it prevented the Yenan faction from being able to team up with the Chinese to claim that they had ‘saved’ the nation” (1975: 179). Therefore, because this coalition shift was predominantly about Kim Il-Sung consolidating his power and minimizing the Yenan faction’s influence, it is coded as neutral. The third North Korean coalition shift occurred on January 31, 1951, when Vice Premier, Industrial Minister and battlefield commander Kim Cha-ek died. Kim Cha-ek was one of Kim Il-Sung’s longtime friends and “number two man” in the Soviet faction, but the circumstances surrounding his death are unclear. Some authors claim he was killed in a bombing raid, while others argue he died of carbon monoxide poisoning after a late-night visit with Kim Il-Sung (Kim 1975: 179; Lim 1982: 208–10). Given the exogenous nature of this shift, it is also coded as neutral. In early June 1951, as the fifth Communist offensive stalled, Kim traveled to Beijing and then Moscow to argue for continuing the offensive, but Mao and Stalin overruled him again. The Communist leaders reached the “consensus” that they would now work for an armistice through negotiations, with the goal of restoring Korea’s prewar status quo (Chen 2001: 97–99).18 After Soviet UN Ambassador Malik’s June 23, 1951 speech, arguing for armistice negotiations, “widespread resentment against the Soviet Union grew within the KWP” (Simmons 1975: 200, 209). Without the ability to fight alone, however, the North Koreans had little say in the matter. Throughout that fall, the North Koreans increasingly realized their interests were not identical to those in Beijing or Moscow. As a result, Kim Il-Sung publicly criticized the foremost Soviet Korean, Ho Ka-I, at the KWP Central Committee’s Fourth Plenum in November 1951,19 and recalled North Korea’s ambassador in China. A new replacement did not arrive in Beijing until March 1952 (Simmons 1975: 209, 211). By January 1952, the North Koreans began indicating they were ready to end the war. North Korean territory had been devastated by the war, and a severe flood in July 1951—the worst in 40 years—had ravaged the military’s logistical system, rail and road infrastructure and agricultural land. As a result, there was famine among the civilian population (Zhang 1995: 165, 172). On January 22, For-

the korean states



eign Minister Pak told Chinese military commander Peng he thought “the Korean people throughout the country demand peace and do not want to continue the war.” That month, Peng estimated 10 percent of the North Korean population was starving, and “the majority of the peasant population will be able to subsist only until April-May.” Moreover, the DPRK had collected “too large a percentage” of the previous year’s grain yield as an in-kind tax. In his February 8 telegram to Stalin relaying these conversations, Mao added that Kim had bluntly told him that Kim had “no desire to continue the war.”20 Similarly, Razuvaev, the Soviet ambassador to North Korea, reported to Moscow in March 1952 that “when Kim was discussing the reasons for the deadlocked negotiations with [DPRK negotiator] Nam Il, he advocated signing the cease-fire agreement and turning over all unsolved problems to a political meeting for further research.” According to Kim, “delaying negotiations was not beneficial, because the US Air Force was continuing to inflict damage on the DPRK. He does not see the rationale behind continuing the debate over POWs, because this debate is now leading to great losses.” Razuvaev reported that Kim had instructed Nam Il “to get a clear sense of the Chinese attitude on this question” and suggested making concessions on the POW problem “in the name of [PRC negotiator] Li Kenong.”21 Stalin’s response to North Korea’s hardship was not particularly sympathetic. In a terse April 1952 cable to Kim, he said: “It became known to me that the Korean people are in need of bread. We have in Siberia 50,000 tons of prepared wheat flour. We can send this flour as a gift to the Korean people. Telegraph your agreement.”22 Given such “enthusiastic” support from the Soviet patron, it is not surprising that around the same time, Defense Minister Choe Yong-Gon told the North Korean military staff: Unless the USSR makes a definite promise to intervene in the Korean War, the only alternative is for North Koreans and Chinese to come to terms with the UN. An armistice will permit the communization of South Korea from within. . . . The Soviets plan to keep Asians fighting Americans while the Soviets sit back and talk about Soviet peace. They do not care about the unification of Korea; they merely invested material and professional aid in order to pin down UN forces and weaken US strength. (Simmons: 1975: 215, emphasis added)

By July, Kim worried that because of the stalemated battlefield, armistice negotiations could be drawn out indefinitely. Kim moaned that the UN “almost without suffering any kind of losses constantly inflicts on us huge losses in manpower and material values”—noting that in one 24-hour period of “barbaric bombing” on Pyongyang (from July 11 to 12), 6,000 North Koreans had been killed and wounded. Urging Stalin to approve offensive military operations to

80

the korean war

revive the negotiations, Kim argued that the Communists desperately needed “to move decisively towards the soonest conclusion of an armistice.”23 Similarly, in correspondence with Mao, Kim argued forcefully for active military operations so the enemy would cease its “furious bombardment.”24 Telegrams of this kind continued throughout 1952 and 1953. Nevertheless, North Korea was consistently bypassed in tactical and strategic decisions on the battlefield and in the negotiating tent. The over-riding truth was that a protracted war helped Soviet and Chinese interests, so they did not adopt the North Korean position. As Jung and Halliday put it, “Mao had hijacked the war, and was using Korea regardless of Kim’s interests” (2006: 362). Nam Il was the North Korean representative at armistice negotiations, but his presence was almost entirely for the UN negotiating team’s benefit. Telegrams among the Communist leaders clearly demonstrate that Mao and his subordinates were in charge of armistice negotiations (Chen 2001: 99). For example, during POW repatriation discussions in early 1952, the Communists were willing to give ground on Korean prisoners, but they refused to show flexibility on Chinese prisoners who did not want repatriation (Stueck 1995: 270–72). Moreover, although Chinese leaders traveled to Moscow during this time to discuss settlement terms, the North Koreans were not included.25 As Chapter 7 will explain, when Chinese Foreign Minister Zhou met with Stalin in the fall of 1952, both leaders mentioned the North Koreans’ plight, but only as evidence for their own preferences, not actually to take the North Korean viewpoint into account. As Stalin remarked dismissively in those discussions, “The North Koreans have lost nothing, except for casualties.”26 Finally, even after the Communist concession about POW repatriation, North Korea was unable to influence the armistice’s timing. After South Korea unilaterally released prisoners in June 1953 to sabotage the agreement, North Korea pushed to sign the agreement anyway. However, CPV Commander Peng, with Mao’s support, overruled Kim Il-Sung once again to launch a new military campaign to punish the ROK army, extending the war an additional six weeks (Shen 2003–4; Chen 2001: 116). In sum, although the war was being fought on its territory, the DPRK suffered until the armistice in the role of invisible junior partner. If anything, the North Korean experience provides an extreme example of a secondary belligerent entrapped by its more hawkish external allies in a war it no longer wanted to fight. While the North Korean governing coalition experienced shifts during the war, they had little effect on the Communists’ decisions about prosecuting and ending it. With the exception of the initial hawkish coalition shift, North Korea’s coalition shifts had no effect on decisions to prosecute and end the war. Although Kim

the korean states



Il-Sung used the battlefield situation to justify changes in the governing coalition, this merely provided convenient cover for consolidating his rule and sidelining rival factions.27 Thus, given North Korea’s relative insignificance in Soviet and Chinese war policy decisions, there is little relationship between North Korean domestic coalition politics and the strategic interaction at the international level. south korea

If China and the Soviet Union basically ignored North Korea’s pleas to end the war sooner, the United States had to rein in South Korea for attempting to prolong it. Led by President Syngman Rhee, the ROK government remained consistently more hawkish than its US patron. Although the United States held the upper hand in the relationship, because South Korea depended on US support for its survival, the United States felt more constrained by Rhee than the Communist great powers felt by Kim, for two reasons. First, because the UN was fighting for the ROK, the United States felt more constrained by international opinion—and thus deferred to its client more in public—than the Communist great powers did. More generally, as the South Korean case shows, it may be easier for hawkish secondary belligerents to act as veto players to settlement than for dovish secondary belligerents to compel more powerful hawkish allies to settle. Despite this, there is still little relationship between South Korean domestic politics and the international bargaining process. Like Kim Il-Sung, Rhee used the ongoing war effort to affect domestic politics—not to enhance the war’s prosecution but to provide cover for consolidating his rule and sidelining rivals. Rhee was elected president when South Korea gained independence in August 1948. Over the next two years, Rhee faced intense pressure from a variety of sources, including Communist-instigated army and mass rebellions and more powerful opposition parties who controlled the National Assembly. These parties, including the Korean Democratic Party and the Democratic Nationalist Party, held more seats in the National Assembly but lacked commensurate representation in Rhee’s cabinet and thus tried to undermine his rule. Rhee’s external support from the United States also declined during this time, as the US and ROK governments clashed over economic policy and US defense commitment to South Korea. In some ways, Rhee was caught between a rock and a hard place: while the United States helped Rhee remain in power, its demands in exchange for that support undercut his internal powerbase and furthered dependence upon the United States. For example, the United States wanted Rhee to increase taxation to curb inflation, but this would have increased the power of the National Assembly opposition parties. As Rhee refused to accede to some of these

82

the korean war

demands, the United States began phasing out its support, so that by January 1950 the American defense perimeter in Asia excluded South Korea (Kim 1975: 116–30). In many ways, therefore, the Korean War came at precisely the right time for Rhee. Not only did the war re-cement US support, but Rhee benefited directly from the growth of the Korean army, which provided a new organizational base of support and a new funding source. In exchange for looking the other way as senior military officers diverted American military aid, Rhee’s new political party (the Liberal Party) used some of these diverted funds to consolidate control at the local level (Kim 1975: 133–35). The war considerably enhanced Rhee’s power and prestige, and as James Palais argues, Rhee “used the opportunity to obliterate the Left by force” (1973: 324). The South Korean government experienced six domestic coalition shifts during the war, with Rhee consolidating his power throughout. All shifts are coded neutral, as the new coalition in each case intended to maintain the same war policies. The first shift occurred after the National Assembly passed a resolution demanding the dismissal of Prime Minister and Defense Minister Shin Song Mo and Home Minister Cho Pyung Ok. Chang Myon, the first ROK ambassador to Washington, replaced Shin as prime minister on November 23, 1950, but Shin retained the defense portfolio (Stueck 1995: 108). During this time, as UN forces advanced north towards the Yalu River, the UN debated how to create a unified national government in Korea. US officials wanted to hold elections to fill the 100 empty National Assembly seats designated for the northern areas, since the assembly had just been elected for a two-year term that spring and Rhee was sure to balk otherwise. However, other countries, especially Britain, wanted to hold national elections as a way to depose the internationally unpopular Rhee. Rhee tried to preempt this possibility by announcing he was assuming “temporary civil control” over the entire country. Before the UN reached its final decision, however, the Chinese counteroffensive rendered the debate moot (Stueck 1995: 108–10). The second ROK coalition shift occurred in the spring of 1951, in response to two domestic political crises: (1) the Kochang incident in February 1951, in which a ROK army battalion massacred 500–1,000 civilians suspected of being Communist guerillas, and (2) the Citizens’ Army incident in March 1951, in which National Defense Corps officers embezzled more than 2.3 billion won and thousands of tons of rice, “driving nearly 1 million soldiers to the verge of death from starvation and disease” (Kang 2005: 198; see also Stueck 1995: 213). The National Assembly got wind of the massacre and started an investigation, followed shortly thereafter by similar international press inquiries. Over the next few weeks until

the korean states



May 16, as details of the twin scandals came to light, a wave of cabinet members resigned or were dismissed—including Vice President Yi Si Yong, Defense Minister Shin Song Mo, Interior Minister Cho Pyung Ok, and Justice Minister Kim Jun-Yun. While these scandals certainly were related to the war, Rhee did not intend to use their dismissals to change war policy. Such intentions would have signaled a dovish shift. Instead, he used the occasion to install appointees who had served in the Japanese regime and were “mere instruments of the president” (Stueck 1995: 213; Pak 1980: 141). Nonetheless, the new vice presidential selection reflected the toll these scandals had on Rhee’s popularity: the new vice president, Kim Song-Su, was the leader of the opposition Democratic National Party (Pak 1980: 145). While this domestic turmoil was unfolding, the Communist offensives near the 38th Parallel were beginning to stall. In response to rumors of impending cease-fire negotiations, the ROK government launched a massive effort to mobilize the public to resist stopping the war short of the Yalu River. Mass protests occurred in Pusan, the temporary seat of the government, and the usually hostile National Assembly unanimously passed a resolution to continue fighting for an “independent and unified country.” At the end of June, the ROK cabinet announced that a cease-fire must include a Chinese withdrawal from Korea, the disarming of the North Korean Communists, a UN guarantee that no third party would assist North Korea militarily or financially, and recognition of “the national sovereignty and territorial integrity” of the Republic of Korea. Rhee promised the United States he would not make any more provocative statements to the press, but he threatened “drastic action” should the UN try to stop him from unifying the country. Perhaps for this reason, the United States decided to include a ROK representative in the UN team for the armistice negotiations (Stueck 1995: 215; Zhu 1998: 242). Over the next year, as the armistice negotiations eventually resolved every issue except POW repatriation, Rhee continued to build his Liberal Party, financed by the diverted army funds, so that by early 1952, 93 members of the National Assembly had been persuaded to join. Unfortunately for Rhee, however, these assembly members refused to endorse Rhee’s preferred form of government (a presidential system). Instead, they favored a parliamentary government, which would assure their own appointments to cabinet posts. Outside of the assembly, the Liberal Party’s base was formed by combining Yi Pom-Sok’s Youth Corps with remnants of Rhee’s National Society party. Rhee had appointed Yi as home minister during the May 1951 coalition shift in exchange for this support. The Liberal Party performed very well in local-level elections in February 1952, thanks to the Youth Corps’ support and Yi’s control of the local police, which publicized the Liberal

84

the korean war

Party and hampered electioneering by opposing groups (Kim 1975: 135–36). Because this election had no effect on the governing coalition, however, it does not count as a shift. Although Rhee had finally created a political party with local roots, he still had no control over the hostile National Assembly. This was becoming a problem, for Rhee’s first term of office was to expire in 1952. According to the constitution, the new president was to be elected by the assembly, and it was clear he would not be their choice. Thus, in November 1951, Rhee introduced constitutional amendments to permit the president to be elected by popular vote and to create a bicameral legislature. Both measures were designed to increase Rhee’s power and weaken the legislature, the first by removing the legislature’s power to elect him and the second by dividing and expanding the assembly to make it easier to control (Palais 1973: 314). As expected, because constitutional amendments needed two-thirds’ support from the assembly, the proposals did not pass. However, neither did the assembly’s own proposal for creating a parliamentary system. About this time, the governing coalition experienced its third neutral shift, when Prime Minister Chang Myon resigned to run in the upcoming presidential election.28 He was replaced by Chang Taek-Sang, who was appointed to conciliate the anti-Rhee assemblymen (Kim 1975: 137; Pak 1980: 102, 107). The fourth neutral coalition shift occurred when Rhee declared martial law in May 1952. That month—just as the armistice negotiations broke down over POW repatriation and riots in the UN POW camps implied inhumane treatment of Communist prisoners—Rhee re-introduced his constitutional amendments with a number of concessions to parliamentary advocates to secure their support. The Liberal Party, in virtual control of the local governments around the country, also began stirring up popular support for the amendment about direct presidential election. Petitions of support for the “right of the people” to elect the president came pouring in from local assemblies, and mass demonstrations (probably organized by Home Minister Yi’s Youth Corps) surrounded the National Assembly building in Pusan, demanding the assembly pass the amendment. The bureaucracy, still loyal to Rhee, cooperated by closing schools and shops to allow for the demonstrations. As the demonstrations got rowdy, on May 25 Rhee declared martial law in the temporary capital of Pusan and 22 surrounding counties. Under cover of martial law, Home Minister Yi then ordered the army to “detain” 47 assemblymen on suspicion of having received funds from the Communist Party. These assemblymen were held at the assembly until they voted through Rhee’s amendments on July 4. Afterwards, the charges against most assemblymen were quietly dropped (Kang 2005: 198; Stueck 1995: 277; Kim 1975: 137–38; Pak 1980: 102–11; Henderson 1968: 167–68).

the korean states



The South Korean domestic political crisis sparked a wave of international protest. Coming directly after the POW screening debacle, the crisis constrained the UN’s ability to present itself as the innocent party in the process of ending the war. US and UN leaders, including President Truman, tried to persuade Rhee to restore constitutional government, but he insisted that a plot existed to overthrow him and that foreigners should stay out of the ROK’s internal affairs. Rhee’s behavior, as well as the conditions in the Koje-do POW camp, eliminated the possibility of expanded UN military action to bring the Communists back to the bargaining table. Other UN members were simply too critical of Rhee to sign on for escalating the conflict (Stueck 1995: 277–78; Pak 1980: 105). In this way, the fourth ROK coalition shift did reverberate into the international bargaining process. On July 4, 1952, the National Assembly—surrounded by the police—finally held a stand-up vote to allow direct elections of the president. On July 18, Rhee announced the presidential election would be held on August 5, leaving the opposition with only 18 days to campaign. As a result, the Democratic National Party (DNP) did not even try to run a candidate against him (Kim 1975: 139; Pak 1980: 187). As Chi-Young Pak argues, “the consequences of the 1952 constitutional crisis dealt a fatal political and psychological blow to the opposition forces. After the unsuccessful constitutional struggle, the DNP and other anti-Rhee groups lacked the vigor, decisiveness and harmony they needed to be effective in the popular presidential election” (1980: 110–11). The fifth neutral coalition shift occurred with Rhee’s re-election as president on August 5. During the short campaign season, Rhee sacked his loyal home minister, Yi Pom-Sok, to distance himself from Yi’s decision to use the police in securing passage of the constitutional amendments. Rhee also refused to back Yi’s vice presidential candidacy, endorsing instead Ham Tae-Yong, a Christian minister from North Korea. Not surprisingly, Rhee won the election with 72 percent of the popular vote, while Ham won with 41 percent (Pak 1980: 188–89; Kim 1975: 138–40).29 Once again, Rhee used the ongoing war effort for personal gain, to consolidate his own rule rather than to influence war policy itself (Palais 1973: 325). As this summary of South Korean domestic coalition shifts shows, Rhee’s domestic political maneuverings had almost nothing to do with the ongoing war effort. All six neutral shifts helped Rhee consolidate his power, not wage the war more effectively. (The final ROK neutral coalition shift occurred October 6, 1952, when Prime Minister Chang Taek-Sang retired and was replaced by Paik Too-Chin.) None of these shifts occurred with the new coalition intending to change war policy. Throughout the war, Rhee remained hawkish about prolong-

86

the korean war

ing it until he could unify Korea under his control. Instead, Rhee believed South Korea was not well represented in armistice negotiations and seldom consulted. Towards the end of negotiations, therefore, Rhee’s government protested and mounted several domestic demonstrations against the UN Command’s proposals (Hermes 1966: 426–27; Zhu 1998: 242). After the March 1953 Communist concession on POW repatriation, Rhee tried even harder to assert his hawkish views. He wanted the United States to continue the war or if that were not possible, to offer a mutual security pact to the ROK. In contrast, the United States hoped to avoid a formal commitment. In conversations with US policymakers, ROK officials threatened to expel UN forces that refused to fight back north to the Yalu River or to remove ROK troops from the UN Command. At home, Rhee launched a campaign to mobilize support for continuing the war, including pro-unification editorials in South Korean newspapers and mass rallies in Seoul, Inchon and Pusan. On April 2, 1953, the National Assembly unanimously passed a resolution opposing an armistice without unification (Stueck 1995: 316–17; 2002: 189–90; Hermes 1966: 426–27; Zhu 1998: 242). While US officials believed most South Koreans wanted peace, they remained wary of Rhee’s unpredictable nature. Moreover, because ROK army units manned two-thirds of the front line by this point, Rhee’s perspective had some weight with the US decision-makers. In addition, the US Republican right wing (especially in Congress) echoed Rhee’s hawkish views, and President Eisenhower felt constrained by both. Thus, in late April, the UN commander, Gen. Mark Clark, gave in to Rhee’s demand for a representative at the armistice negotiations who would report directly to Rhee, and the US ambassador to Korea, Ellis Briggs, advised a mutual security pact might be necessary to gain Rhee’s acquiescence (Stueck 1995: 317; 2002: 190). Perhaps most significantly, largely because US officials feared Rhee’s reaction, they did not inform him of their final negotiating offer before it was presented on May 25 to the Communists at Panmunjom. This proposal was obviously not satisfactory to the South Koreans: it failed to release anti-Communist POWs immediately after an armistice, and it permitted Communist officials and Indian troops—which Rhee did not consider neutral—into ROK territory to supervise and question those prisoners. More symbolically, by not sharing the proposal with Rhee until virtually the moment it was presented to the Communists, it was “a humiliating reminder of America’s refusal to treat South Korea as an equal partner in the war” (Stueck 1995: 331). When Clark outlined the proposal to Rhee, Rhee told him, “I am deeply disappointed. Your government changes its position often. You pay no attention to the view of the ROK Government” (Clark 1954:

the korean states



268–69). About this time, the US government also authorized Clark, in the event of internal ROK instability, “to take whatever other steps you deem appropriate to safeguard the integrity and security of your forces.” Clark took this to mean that “if the present ROK government cannot be forced to accept the armistice terms,” he could establish a more “amenable” regime (FRUS 1952–54, 15: 1153). After presenting the final offer to the Communists, Eisenhower offered on June 7 to negotiate a mutual security pact promptly after the armistice was signed. He also offered economic assistance to rebuild Korea. But these moves to placate Rhee came too late. Rhee enhanced his ability to take actions independent of the UN, such as placing all ROK military units under the Ministry of National Defense, rather than the army, whose chief was strongly pro-American (Bernstein 1978: 38–47; Kotch 1983: 239–59; Stueck 1995: 316, 330–39; 2002: chapter 7). On June 18, in a last-ditch effort to undermine the negotiations, Rhee unilaterally ordered South Korean guards to release 25,000 North Korean prisoners resisting repatriation. Rhee claimed in his press release that “most of the United Nations authorities with whom I have spoken about our desire to release the prisoners are with us in sympathy and principle” (FRUS 1952–54, 15: 1199). Interestingly, although the Communists sent a pointed letter to the UN Command asking it to rein in the ROK, their response was relatively mild, suggesting how ready they were to conclude the armistice. The Communists directed most of their propaganda solely at Rhee and his government, rather than the United States or UN. Moreover, the Chinese military launched directed attacks against South Korean units, pushing ROK forces back six miles. During this offensive, ROK forces suffered 14,000 additional casualties, according to UNC figures (Stueck 1995: 338–39; Zhu 1998: 247–52; Foot 1990: 186–87; Clark 1954: 292). Obviously, Rhee’s decision to release the prisoners unilaterally postponed the armistice by several weeks, increasing casualties by tens of thousands. Yet playing this veto card was not irrational. Over the next few weeks, Rhee extracted a number of concessions from the United States to secure his agreement to honor the armistice (Allen 1994: 106; Keefer 1986: 286–87; Kotch 1983: 239–59; Goulden 1982: 640–42; Kaufman 1986: 328; Zhu 1998: 252–53; Stueck 2002: 191–93). As UN Commander Clark put it, Rhee “knew that no matter what happened we could not, after three years of war, after all the blood and treasure we lost, let Korea go to the Reds by default because of a quarrel ‘in the family’” (Clark 1954: 272). William Stueck argues that Rhee shrewdly “took care to prevent the crisis from escalating out of control” by leaving provocative comments to his foreign minister and “positioning himself to mend fences” (1995: 338). Rhee derived substantial advantages from his last-ditch spoiler actions. First,

88

the korean war

he got the United States to insist at Panmunjom that no Communist or Indian officials or troops attached to the Neutral Nations Supervisory Commission would operate in South Korea. Moreover, both the ROK and US governments agreed to withdraw from the post-armistice political conference after 90 days if no concrete progress was made (Clark 1954: 287–88). Second, he received a mutual defense pact, long-term economic aid and US support for expanding the ROK army, navy and air force (Zhu 1998: 252). More symbolically, he forced the United States to bargain openly with him about these issues, thereby increasing his prestige at home. Finally, he set the stage for future interactions with the United­ States, ensuring the more powerful ally would no longer take him for granted. In a letter to Eisenhower on the day the armistice was signed, Rhee thanked him “for the statesmanlike vision with which you have brought the relationships of your powerful nation and of our weaker one onto a basis of honest mutuality and two-way cooperation” (quoted in Stueck 1995: 338–39). In sum, while South Korea experienced a number of domestic coalition shifts during the Korean War, these shifts had almost no effect on the war termination process. The one possible exception was the 1952 constitutional crisis, which may have eliminated the possibility of expanded UN military action to pressure the Communists in the negotiating tent, because other UN members were too critical of Rhee to sign on for escalating the conflict. (However, as Chapter 6 will demonstrate, at that point, the US governing coalition was unwilling to escalate much anyway.) Given South Korea’s relative insignificance in UN decisions about prosecuting and ending the war, Rhee’s domestic political maneuverings had little to do with the ongoing war effort. All six neutral shifts helped Rhee consolidate power, and none brought to power a new coalition intending to change war policy. If South Korea’s domestic politics had any effect on when, how and why the war ended, it was expressed in the last two months of the war, when Rhee unilaterally released North Korean POWs. This last-ditch effort to undermine the armistice delayed the agreement by six weeks and affected the details surrounding treatment of non-repatriated POWs.

Conclusions While some scholars focus on battle-deaths as the major determinant of a belligerent’s role in a war, this chapter has argued for a more nuanced coding of belligerents’ importance in decisions about prosecuting and ending war. This chapter has explained my decision to include the Soviet Union—and conversely, to exclude the Korean states—among the primary belligerents of the Korean War. While the Soviet Union was not a declared war participant, it bankrolled the

the korean states



Communist war effort with materiel and advisory support, provided most of the pilots for the Communist air war and—most importantly—played a central role in formulating war policy. In the Korean states’ case, although the war was fought on their soil, North and South Korea were powerless in their respective side’s decisions and actions. Not only were they materially powerless, as their forces were incapable of waging war without their larger patrons’ help, but the way the primary belligerents viewed them as client states rendered them politically powerless as well. Within this perceptual context, even if they had been more powerful materially, it is unlikely they could have influenced their respective side’s decisions about how to fight and end the war. While both Korean states experienced domestic coalition shifts during the war, there was almost no relationship between these shifts and the international bargaining process. Both Kim Il-Sung and Syngman Rhee used the ongoing war effort to affect their domestic political scenes. However, changes in their respective governing coalitions were intended to provide cover for consolidating their rule and sidelining rivals—not to enhance fighting or ending the war. Ironically, if the Korean parties had any discernible effect on the war’s prosecution and termination, it appeared at the very beginning and very end: North Korea was certainly responsible for launching the war, while South Korea delayed the timing of the armistice by six weeks. Thus, given their relative insignificance in their respective side’s decisions, there is little relationship between Korean domestic coalition politics and strategic interaction at the international level. This analysis has two wider implications. First, the ROK case suggests that hawkish secondary belligerents may inherently have more power than dovish ones. While Rhee was not any more powerful than Kim in his side’s war policy decisions, his last-ditch veto allowed him to extract a number of concessions from the Eisenhower Administration. Compared with Kim’s unsuccessful pleas to end the war sooner, this suggests it may be easier for hawkish secondary belligerents to act as spoilers to settlement than for dovish secondary belligerents to compel more powerful allies to settle. Second, this chapter demonstrates that how belligerents are framed often affects how much power they have in the decision-making process. The primary belligerents viewed the Korean states as clients, pawns in a Cold War battleground—to some degree creating a self-fulfilling prophecy. Typifying the Korean states this way effectively eliminated them from the strategic interaction between the two sides. Neither North nor South Korea was perceived as an independent actor, with its own interests and own decision-making processes. As a result, they were almost irrelevant to the international level bargaining process. For example, it never occurred to the United States or other UN actors that they could interact

90

the korean war

directly with the North Koreans or exploit differing interests and perspectives in the Communist camp. Yet such differentiation among the various actors in the other side’s coalition could have held the potential for a faster termination of the war. We will return to this idea in Chapters 8 and 9, when we explore the role that secondary actors can have in signaling to the other side.

4

The Soviet Union Outlasting Stalin’s Preferences

The next four chapters examine the dynamics of the three obstacles to peace— and the role of domestic coalition shifts in lifting them—among the primary belligerents in the Korean War. It may seem counterintuitive to begin with the Soviet Union, because its domestic coalition shift—a result of Stalin’s death in March 1953—occurred last of the various belligerents’ coalition shifts during the war. But removing Stalin and his preference for prolonging the war were essential for ending it, because while the other primary belligerents were ready to settle sooner, the Communist side remained constrained by Stalin. Therefore, first understanding the Soviet dynamics provides an important context for considering the US and Chinese cases. As Chapter 2 explained, not every belligerent will necessarily experience all three kinds of obstacles. Remember that a preference obstacle occurs when leaders do not want to end the war; an information obstacle occurs when leaders do not know they should end the war; and an entrapment obstacle occurs when leaders want to end the war but cannot. While the logic of these three obstacles is distinct, they can often occur in tandem. For example, as the Soviet case demonstrates, leaders who hold strong preferences are also likely to fall victim to information processing biases or organizational pathologies that reinforce preexisting beliefs. The Soviet Union under Stalin provides strong evidence of the preference obstacle, with “supporting” hindrances from the information obstacle. This chapter argues Stalin did not want to end the war, because he was benefiting from the war without assuming many of its costs. While other leaders in his governing

92

the korean war

coalition valued other preferences more highly, Stalin’s record of squashing opposition, through numerous leadership purges, led those who opposed his war policy to self-suppress out of fear of punishment. Even the information that Stalin and other leaders received about the war was tailored to support Stalin’s known preferences; thus, information obstacle dynamics exacerbated the preference obstacle. When Stalin died unexpectedly in March 1953, these other leaders took control and quickly enacted their long-held but dormant preferences, leading them to end the war. In other words, the domestic coalition shift removed the stickiness of Stalin’s preferences and allowed the war to end. The rest of this chapter is divided into six sections. The first section explains the Soviet governing coalition, suggesting that while Stalin was the preeminent decision-maker, there were two factions within his government. The second and third sections explain the Soviet decisions to support the North Korean invasion and draw out the war. The fourth section explains how information processing dynamics in the Soviet regime exacerbated the effects of this preference obstacle. The fifth section describes the Soviet domestic coalition shift that resulted from Stalin’s death and the actions the new coalition took to end the war. Finally, the sixth section examines some alternative explanations for Soviet behavior.

The Soviet Governing Coalition If the ideal type of “dictator” has ever existed empirically, Joseph Stalin would certainly be a strong contender for the role. Responsible for the death of 27 million Soviet citizens during World War II, numerous purges and imperial conquest of vast swaths of territory, Stalin was ruthless, autocratic, suspicious and vindictive. By most accounts, Stalin had the final power in all decisions in the Soviet Union, both on foreign and domestic matters (Stueck 1993: 114; 1995: 32; Zubok 1992: 5; Medvedev 1989). Yet Stalin was not purely the absolute dictator that conventional wisdom would have us believe. In the face of the tremendous task of governing an empire as huge as the Soviet Union, even Stalin’s power was curtailed by his own physical limitations and his dependence upon the complex bureaucracies necessary to run a planned economy. As Ronald Letteney argues, “Stalin’s necessary reliance upon subordinates must have meant that there was room in the shaping of policies for their influence” (1971: 5). Although Stalin worked hard to ensure that other leaders could never amass enough information and power to overthrow him, there were factions among the elite decision-makers in the Soviet Union. Stalin remained on top by strictly controlling information and purging or demoting rivals whose influence became too threatening, but the factions existed, nonetheless.

the sov iet union



There has been some debate about the role of other leaders in foreign policy decisions. On the one hand, Nikita Khrushchev argues that Stalin “jealously guarded foreign policy . . . as his own special province” and that “the rest of us in the leadership were careful not to poke our noses into these matters” (1974: 188–89, also 191, 355–57). Moreover, Kathryn Weathersby contends that archival documents reveal the absence of “debates” about foreign and domestic policy decision-making. She notes that “in the hundreds of [Central Committee and Foreign Ministry] files I have examined, I have never seen a document indicating a policy debate of any kind” (1993a: 6). For example, the Politburo sponsored no public conferences about national security until 1983, and no substantive debates on strategy, analogous to Eisenhower’s 1953 “Solarium” exercise, were held until the July 1988 Scientific and Practical Conference (Zubok: 1992, FN 20). On the other hand, William Stueck contends Stalin made foreign policy choices in light of how they would influence his personal power at home. Perhaps because of this insecurity, he allowed a good deal of internal debate on foreign policy (1993: 115). Vladislav Zubok and Constantine Pleshakov argue Stalin “often drew the ministers of defense, armament, and security, as well as senior intelligence officials” into his decision-making circle; the top three, four or five—always including Vyacheslav Molotov—would serve as members of ad hoc working groups for certain foreign policy issues (1996: 84). Even more broadly, Ronald Letteney argues “factionalism within the Soviet leadership was an element of major significance during Stalin’s reign,” with the Soviet Politburo divided into two cohesive factions with respect to foreign policy (1971: 25; also Ra’anan and Lukes 1990: 14). These two perspectives probably conflict less than it might appear: while Stalin probably did not sponsor “policy debates” as democracies would conceive of them, informal debates within Stalin’s inner circle certainly occurred. Stalin was not operating completely alone. Given Stalin’s paranoia, however, these discussions were probably not written down—especially evidence of dissent. The Soviet leadership was divided into two factions—hardliners and moderates—each with its own general political philosophy, worldview and foreign policy prescriptions. The factions competed not only for promotions, but also for domestic resource allocation and influence over policy decision-making. While the factions were most visible at the Politburo level, they extended down to middle-level bureaucrats and local administrators, easing the policy implementation process. Each faction used different journals to expound their views: the hardliners used Pravda (the official newspaper of the Soviet Communist Party) as their mouthpiece, while the moderates used Izvestiia (the official organ of the Soviet government) as theirs. Usually, the military press and the Communist Informa-

94

the korean war

tion Bureau seconded hardliner positions, while Soviet trade union publications, such as Trud, faithfully reflected the moderates (Stueck 1993: 115; Ra’anan and Lukes: 1990: 15–19; Letteney 1971: 26–27). The hardliner faction believed capitalist states were inherently hostile and aggressive; they advocated for continuing past militant policies. They argued worldwide Soviet pressure on the United States would cause it to retreat to the Western hemisphere. The hardliners included former foreign minister Vyacheslav Molotov, Stalin’s leading ideologue Mikhail Suslov, Chief of Heavy Industry Lazar Kaganovich, and Marshal Klement Voroshilov, the Politburo’s “military man.” It is not surprising these men represented broader constituencies within the military and heavy industry. In their speeches, Molotov, Kaganovich and Voroshilov frequently stressed the need for “vigilance,” which in Soviet-speak was synonymous with priority for heavy industry and defense spending. Stalin tended to side with the hardliners, perhaps because Molotov and Voroshilov were his closest personal relationships within the Politburo. In fact, from 1949 until his death in 1953, Stalin sided mostly with the hardliners against the moderates on foreign policy questions and took actions to check the influence of the moderate camp (Letteney 1971: 30; Snyder 1991: 230–31). In contrast, the moderate faction argued the Soviet Union needed to reduce international tension, because Soviet pressure strengthened the Western alliance.1 The moderates asserted that Soviet conciliatory moves would make it more difficult for the United States to maintain and consolidate its alliance; hence, they promoted limited cooperation with the capitalist West (“peaceful coexistence”). They believed the contest between the Communists and capitalists would be decided by long-term economic competition rather than by any immediate military conflict. Accordingly, they advocated shifting resources from heavy industry and the defense sector towards internal development, light industry and consumer spending, to support economic development over the longer term. The moderates included former foreign trade minister Anastas Mikoyan, chief of the secret police Lavrentii Beria, and Stalin’s heir apparent Georgii Malenkov—the only man in late 1949 besides Stalin to sit on all three of the top Soviet councils: the Party Secretariat, the Council of Ministers and the Politburo. The moderates sought a constituency for these views among the urban middle class, technical intelligentsia and managerial elite (Ra’anan and Lukes 1990: 14, 28; Letteney 1971: 27–30, 36, 188, 208; Stueck 1993: 115; Snyder 1991: 231–32, 245–46; Nicolaevsky 1975: chapter 3).2 During 1949 and the first half of 1950, the two factions appeared to be fairly evenly matched. Their efforts often seemed to neutralize each other, so that the Soviet government took few foreign policy initiatives. By the time of Stalin’s 70th

the sov iet union



birthday celebration in December 1949, the two camps compromised on some of the outstanding issues in foreign policy—setting the stage for the subsequent decision to support Kim Il-Sung in the Korean War. In the compromise, the hardliners accepted the line of “peaceful coexistence” over the long run, while moderates agreed in the short run to intensified “vigilance.” In concrete terms, this meant the Soviet leadership acknowledged communism and capitalism would coexist in Europe for an extended period of time, implying the need for a more moderate line vis-à-vis the United States in Europe. However, there would be no concomitant rearrangement of domestic economic priorities away from military production and heavy industry, since “vigilance” was necessary to build up East European militaries. After this compromise, there was a notable decline in the factional debates in the Soviet press (Dunmore 1984: 125; Letteney 1971: 213–18). Despite the existence of these factions, Stalin was a very suspicious person, bordering on paranoid. Stalin was concerned about his deputies’ power. By the time the Korean War began, Stalin had a “genuine weariness” but realized “that a man in his position cannot really retire, that one cannot be a part-time despot” (Ulam 1973: 705; also Mastny 1996: 12; Nogee and Donaldson 1992: 88; Zubok and Pleshakov 1996: 22–25; Taubman 1982: 197; Stueck 1993: 14; Alliluyeva 1967: 196–200; Khrushchev 1974: 244). Stalin’s superior information network only fed his worst fears and a sense of isolation (Zubok and Pleshakov 1996: 24; also Medvedev 1989: chapter 8). Moreover, there has been debate about Stalin’s psyche and its influence on Soviet politics in his last years.3 Given the extent of his ruthlessness, and the likelihood his actions did create true enemies, his paranoia certainly had a rational component. As will be discussed further below, frequent purges and juggling of high-level personnel, combined with tight control over access to information, helped him bolster his position of power. Nonetheless, Stalin’s suspicious worldview and paranoid nature permeated the coalition’s decisions and created a culture of fear, which made it difficult for others to speak out against him.

The Soviet Decision to Support the War Since the spring of 1949, when Kim Il-Sung and his foreign minister, Pak HonYong, visited Moscow, Kim had begged Stalin for his blessing to initiate a “reunification of Korea.” This started as Kim’s war, and he gained Soviet approval only after persistent appeals—including at least 48 telegrams. Stalin initially feared such a war could result in direct military conflict between the Soviet Union and the United States. However, on January 30, 1950, Stalin said that while he still had “reservations” about the North Korean invasion, he “did not object in principle”

96

the korean war

and would “assist” Kim Il-Sung in the matter of unifying Korea by military means (Weathersby 1995: 9; also 1993a; 1993c: 429; 1995: 2–3; Stueck 1993: 119; Simmons 1975: 32; Zubok and Pleshakov 1996: 55; Taubman 1982: 212–13; Volkogonov 1993; “On the Korean War” 1993: 15; Khrushchev 1993: 62; Hao and Zhai 1990: 100; Kim 1996: 251–55; Park 1996a: 133–55).4 This section argues Stalin’s change of heart was advocated by the hardliners since mid-1949 and implemented over the moderates’ objections. The Soviet decision to support the North Korean invasion departed from its original post-war policy towards the Korean peninsula. From 1945 until early 1950, Soviet policy in Korea was to maintain a balance of power by prolonging the peninsula’s division and thus protect Soviet strategic and economic interests. By retaining control over North Korea, Stalin could protect Soviet borders and ensure unimpeded access to economic resources there. Although Stalin had no interest in seeing the peninsula reunified, Kim used Stalin’s “doublespeak” of offering Soviet “commitment” to unification and independence to push for the literal meaning of the Soviet proclamations (Weathersby 1993a: 17; 1993b; 1993c; 2002). In early 1950, however, the Soviet leadership decided to support Kim Il-Sung’s invasion, for five reasons. The underlying motive in all five reasons was clearly Soviet self-interest. As Sergei Goncharev, John Lewis and Xue Litai argue, “Simply put, Kim Il-Sung was merely a pawn in Stalin’s grand chess game. . . . Only if that war would promote his larger schemes would he agree to it” (1993: 142). First, many scholars argue the Soviet relationship with the Chinese Communists drove the decision. Stalin used the war to maneuver China into conflict with the United States, removing the possibility of friendly relations between them and subordinating China to the Soviet Union (Ulam 1992: 81–82; Thornton 2000: 380). Even more personally, Mao rose to the top of the Chinese Communist Party (CCP) through a series of struggles with rivals, many of whom were considered Stalin’s chosen leaders. Thus, Stalin was insecure about a Chinese leader perceived to be not under his control (Swank 1953: 15–37; Jo 1993: 273; Thornton 2000: 25; Jung and Halliday 2006: 331–72). To provide some background, Soviet behavior towards the Chinese Communists certainly left them reason to be wary, and thus it is not surprising they wanted to avoid dependence on the Soviet Union. Under the post–World War II Yalta Agreement and the 1945 Sino-Soviet Treaty worked out with the Chinese Nationalists (Kuomintang, or KMT), the Soviet Union had secured buffer zones around the Soviet periphery, including Outer Mongolia and sections of Manchuria and Xinjiang. Stalin realized this arrangement could only hold if the KMT remained in power; thus until the eve of the PRC’s creation, he continued to sup-

the sov iet union



port the KMT during the Chinese civil war. For example, in January 1949, Stalin tried to persuade Mao to accept Soviet mediation to end the war with a divided China—Stalin’s optimum outcome. When this failed, he sent Mikoyan to meet the Chinese leaders in person to try to convince them to allow the KMT to escape across the Yangzi River and establish two Chinas (Murray 1995: 2–7; Thornton 2000: 27–28). Instead, the Chinese Communists successfully drove the KMT to Taiwan and established the People’s Republic in October 1949. During this time, the CCP sent numerous signals about renegotiating the Sino-Soviet Treaty on a more equitable basis. In the winter of 1949–50, Mao and Foreign Minister Zhou Enlai traveled to Moscow to hammer out a new treaty with Stalin (Petrov 1994a: 14–18; Chen 1991; 1994: 78–85; Mastny 1996: 88–91; Zubok and Pleshakov 1996: 58–62; Jung and Halliday 2006: 343–49). Stalin, of course, was reluctant to give up the Changchun railroad, ports and naval facilities he had secured in Manchuria after World War II. While the February 1950 “Treaty of Friendship, Alliance and Mutual Assistance” outlined a more comprehensive bilateral relationship (including security guarantees) than similar Soviet agreements with Eastern European states, Stalin was able through secret protocols to protect his interests, nonetheless. As Jung Chang and Jon Halliday argue, “The published treaty was a formality. The essence of the treaty was in secret annexes. . . . Mao went to great lengths to conceal how much the treaty gave away” (2006: 348). For example, one clause gave the Soviet Union exclusive access to all industrial, financial and commercial activities in Manchuria and Xinjiang, while another gave it a monopoly on all of China’s “surplus” strategic minerals for 14 years (Gaddis 1997: 72; Goncharev et al. 1993: 110–29; Jung and Halliday 2006: 348–49).5 With friends like these, it is no wonder Mao sought to preserve some degree of independence, which he hoped to accomplish through diplomatic relations with the United States and other Western countries. During his visit to Moscow in December 1949, Mao sent a cable home to CCP Vice Chairman Liu Shaoqi and Zhou, authorizing them to proceed with establishing relations with capitalist countries. While the cable specifically addressed approaching the Burmese government to sever its relations with the KMT and establish relations with the CCP, Mao’s guidance was more general: “this procedure for the talks is absolutely necessary and we must act in this way toward all capitalist countries . . . this way the initiative will always be in our hands” (quoted in Goncharev et al. 1993: 238). Some scholars argue Mao assumed the Soviets were reading his cables and used this lever to pressure Stalin into signing the new Sino-Soviet Treaty. However, in the process, this cable triggered Stalin’s “worst fear”—that Mao would establish relations with the United States—and Stalin’s subsequent decisions in Korea

98

the korean war

should be viewed in this context (Thornton 2000: 46; Jung and Halliday 2006: 346). Whether or not this claim about the cable’s significance is correct, China’s eventual entry into the Korean War clearly forestalled friendly Sino-American relations—an outcome to Stalin’s benefit. A second, related reason for agreeing to support the North Korean invasion stemmed from Stalin’s insecurity about his position as head of the international Communist movement, given the potential rivalry with Mao for that role in East Asia. This was a relevant concern because in addition to Kim Il-Sung’s desires, Ho Chi Minh was seeking Soviet and Chinese support in his Indochinese war against the French. Stalin was worried about being upstaged by the Chinese in Korea and East Asia more generally, which would weaken Stalin’s position as leader of the Asian Communist camp (Zubok and Pleshakov 1996: 61; Gaddis 1997: 72; Jung and Halliday 2006: 350–51; Weathersby 1993b: 27; Shulman 1963: 141; Letteney 1971: 310; McCagg 1978: 309–10). Third, Soviet hardliners and Kim Il-Sung convinced Stalin that the United States would not intervene, triggering Stalin’s opportunistic tendencies. Kim and the hardliners argued the United States had not used force to prevent a communist victory in China, and US Secretary of State Dean Acheson’s January 12, 1950. public speech had excluded Korea from the US “defense perimeter.” The Soviet ambassador to Pyongyang, Terentii Shtykov, backed Kim’s claims and helped him persuade Stalin of the validity of Kim’s assessments (Goncharev et al. 1993: 133; Letteney 1971: 307–8; Gaddis 1997: 72–73; Weathersby 1993c: 433; Stueck 1995: 69; Petrov 1994b). At his April 1950 meeting with Kim, Stalin acknowledged that “information coming from the United States” indicated a prevailing mood not to interfere. Stalin may have learned—through his British spies in Washington—about the classified December 1949 change in US strategic policy for the Far East, which included Japan and the Philippines in the US defense perimeter but excluded the Asian mainland (Gaddis 1997: 72; Weathersby 2002: 11–12). Fourth, Soviet hardliners and Kim also expressed confidence about the prospects for Korean unification, stressing the magnitude of potential armed partisan opposition to Syngman Rhee’s regime in South Korea. Starting in mid-1949, the hardliners began emphasizing the need and desire to reunify Korea under North Korean leadership and suggested Rhee’s regime was not strong enough to stop reunification. By the following spring, Pravda claimed South Korean partisans, during February 1950, had carried out more than a thousand military operations, which the ROK government failed to stop (Letteney 1971: 312–13, 316). Similarly, in their meeting with Stalin that same month, Kim and his foreign minister Pak predicted that an uprising of 200,000 South Korean Communists would support the Korean People’s Army during the invasion (Kim 1996: 268; Goncharev et al. 1993: 143–44).

the sov iet union



Finally, in yet another similarity between their views, Soviet hardliners and Kim argued a North Korean attack could buttress against South Korean and Japanese aggression. Before the Soviet foreign policy compromise in December 1949, the hardliners argued the United States was preparing South Korea and Japan as “military springboards” for aggression aimed at the Soviet Union and Asian “national liberation movements.” Now, the hardliners imputed a US-Japanese threat to Soviet interests in Asia to justify their plans for Korean reunification (Letteney 1971: 314, 319–21). Obviously, this line of argument contradicts the previous argument about US non-intervention if North Korea invaded the South. The hardliners were arguing it both ways: while they asserted the US position in Asia threatened Soviet interests, demanding aggressive action, they simultaneously implied the United States was indecisive and could easily be forced out of Asia. As these five reasons suggest, the hardliners were a significant force behind Stalin being persuaded to back Kim’s plans to unify Korea by force. In addition, there appears to be at least one domestic political reason for the hardliners’ pushing for war: a desire to maintain the Soviet’s internal focus on heavy industry and military production—despite the December 1949 foreign policy compromise. The hardliners needed “to maintain sufficient international tensions to justify the sacrifices of the population inherent within a program of increased ‘vigilance’ under conditions of peaceful coexistence” (Letteney 1971: 44). Finally, the decision to support Kim’s venture could have resulted from coalitional log-rolling over twin decisions about Korea and Yugoslavia. While the Soviet leadership rejected the hardliners’ scheme for invading Yugoslavia to bring Tito back in line—the moderates’ preferred position—it agreed to support the North Korean venture (Letteney 1971: 287). This way both factions “won” something. Thus, Stalin threw his support behind the hardliners, as their views complemented his own, and he gave Kim Il-Sung the green light. The moderates objected to this plan, dissenting to the war policy through speeches and journal articles. For example, the moderates did not impute any aggressive intentions to South Korea and argued the weak Rhee regime should be given more time to deteriorate. They minimized the aggressive intent of the American occupation in Japan, explaining it primarily in economic rather than military terms. They asserted the KMT’s defeat in China made the United States more determined to remain in Asia and cautioned it might intervene in Korea. They also argued, through a series of Izvestiia front-page editorials, that peace was needed to build communism and “socialism does not need war as proof of its historical superiority” (Letteney 1971: 45, 312, 314, 319, 329, 333). Nonetheless, the moderates were overruled on Korea, while their positions on other issues (like Yugoslavia) prevailed. Cynical observers could argue the moderates’ conciliatory line was perhaps the only real option in Western Europe,

100

the korean war

anyway—in light of the Marshall Plan’s success, the failure of the Berlin blockade and the formation of an independent West Germany and NATO (Gaddis 1997: 72). To some degree, then, perhaps Stalin had no choice but to agree with the moderates’ perspective in Western Europe. However, the moderates’ disagreements about the Korean War did not end with its outbreak, as the next section details.

The Preference Obstacle: The Soviet Decision to Prolong the War To return again to the domestic coalition theory, a preference obstacle occurs when members of the governing coalition do not want to end the war, either because they receive private benefits from continuing the war or fighting helps them to stay in power. Preference obstacles are most problematic when the governing coalition includes members who are unwilling or unable to criticize their hawkish colleagues who want to continue the war. In coalitions with a power imbalance, where potential critics are less powerful than the hawks, they are often unwilling to express their criticism for fear of expected penalties. The fear of being branded a “traitor” can lead coalition members to falsify their true war-terminating preference to preserve their jobs. The less control a coalition member has over policy choice and implementation, the less willing he is to take personal political risks to change it (Lamborn 1991: 63–66; Kuran 1997; Sigal 1988: 29). Alternatively, they may be willing to speak out but lack the power necessary to overrule the hawks whose preferences are creating the obstacle. In these cases, even if potential critics try to speak out, the more powerful hawks are unwilling to listen and may take action to prevent criticism from arising again, by humiliating, chastising, exiling or even killing some of the critics. Such actions are more likely to lead any remaining potential critics to “self-suppress” their doubts and dissent. To preview the rest of this argument in the Soviet case, after siding with the hardliners to support the war, Stalin remained reluctant to risk provoking direct military conflict with the United States. Until the Chinese intervention, Stalin consistently took steps to prevent such escalation, while simultaneously cajoling China into battle—either because the cautious opportunist believed such intervention would not cause global war or he had more Machiavellian designs to stop Mao from having his cake and eating it, too. By the time the war bogged down in mid-1951, Stalin had orchestrated the situation so that he was not bearing the war’s costs directly, yet gaining plenty of benefits from it. While the moderates opposed this war policy, they were relatively powerless to effect its change until after Stalin’s death.

the sov iet union



orchest r at ing the north korean invasion

Despite his decision to support Kim’s attack, Stalin remained reluctant to risk provoking a military conflict with the United States. He therefore made it very clear to Kim that the Soviet Union would not send troops to his assistance. As Stalin told Kim in April 1950, “If you should get kicked in the teeth, I shall not lift a finger. You have to ask Mao for all the help” (quoted in Jung and Halliday 2006: 354; also Weathersby 2002: 12; Thornton 2000: 107–8; Gaddis 1997: 74; Volkogonov 1991: 540). Stalin also insisted Kim consult with Mao in person before the attack. By handing over the “final decision” to Mao, Stalin tried to minimize his involvement in the war. Kim informed Mao of his intentions on May 13, five weeks before the planned invasion—Mao’s first direct knowledge of the attack. With “little room to voice objections to the fait accompli presented by the Koreans,” Mao agreed to provide assistance but warned Kim about possible Japanese or American intervention (Weathersby 2002: 13; Park 1996a: 195; Gaddis 1997: 78; Wolff 2000: 25; Kim 1996: 264).6 Soviet military advisors drafted the invasion plan, participated in reconnaissance and planning at the divisional level, and supervised the staging of troops from June 12 to 23. In early May, the Soviet advisors discarded the North Korean invasion plan and wrote a more aggressive one, which was translated into Korean and kept among the Soviet-born Koreans “to keep Mao ignorant of their plans” (Thornton 2000: 110; also Weathersby 1996: Document 14). According to Gen. Yu Song-chol, who translated the plan, Soviet advisors drew up the plan in about three days without consulting the Koreans: “They did everything themselves. They did not study the terrain and did not know it. . . . Because of that, they made a lot of mistakes” (quoted in Goncharev et al. 1993: 150). Despite the Soviets’ elaborate involvement in planning, Stalin pulled back all of the Soviet advisors from the front lines before the campaign began. As Khrushchev recounts in his memoirs, when he asked Stalin why, Stalin replied: “It’s too dangerous to keep our advisors there. They might be taken prisoner. We don’t want there to be evidence for accusing us of taking part in this business. It’s Kim Il-Sung’s affair” (Khrushchev 1974: 370; Stueck 1995: 36). Two weeks after the invasion, Kim asked Stalin to return Soviet advisors to the front; Stalin agreed but instructed they pose as Pravda correspondents so as to avoid appearing involved (Volkogonov 1993: 29; Weathersby 1993c: 434). Richard Thornton attributes a more nefarious purpose to Stalin’s haste to get the North Korean invasion underway and to remove Soviet support for the offensive. He argues Stalin was trying to preempt Mao, who was simultaneously preparing for an amphibious assault against the KMT on Taiwan. Mao was try-

102

the korean war

ing to attack Taiwan before the Korean War, and Stalin wanted to obstruct Mao in achieving a victory over the KMT. Even more ambitiously, Stalin wanted to guard against a Sino-American rapprochement, which required prodding China into conflict with the United States. This necessitated the North Korean assault failing. As Thornton argues, The ways in which the Soviets devised and executed the war plans, allocated the NKA’s weapons mix, and parceled out logistical support ensured failure. . . . The responsibility for failure was Stalin’s, not Kim Il-Sung’s, whose troops fought tenaciously, but in vain. Yet within Stalin’s failure lay the seeds of a greater success—a Sino-American conflict that would crystallize the structure of global politics for the next two decades. (2000: 227–28; also 113–18; 229)

Whether or not this was Stalin’s ultimate objective, Nikita Khrushchev agrees about its effects: If we hadn’t refused [Kim] aid in qualified personnel to assess the distribution of forces and to direct operations, there’s no doubt that North Korea would have been victorious. I think if Kim had received just [one] more tank corps, or two at the most, he could have accelerated his advance south and occupied Pusan on the march. The war would have ended then and there. (1970: 370)

Similarly, Stalin’s last-minute decision on June 21, to modify the offensive plan from a local operation at Ongjin peninsula to an attack across the whole front line, “greatly increased the likelihood of [American] interference” (Weathersby 2002: 15). After the United States intervened, Stalin tried to distance the Soviet Union from responsibility for the invasion (Zubok and Pleshakov 1996: 65; Jung and Halliday 2006: 355). Even more importantly, when the United States moved its Seventh Fleet into the Taiwan Strait in response to the North Korean attack—an act Mao pronounced as tantamount to a declaration of war—Stalin prevented the Soviet representative from attending the UN Security Council meeting that put the Korean “police action” under UN Command. In this way, Stalin “effectively precluded any formal declaration of war by either the Chinese or the American side and so avoided activating the Sino-Soviet treaty” (Goncharev et al. 1993: 215). As Jung and Halliday argue, “if Stalin decided not to use his veto [in the UN Security Council meeting], it can only have been for one reason: that he did not want to keep Western forces out [of Korea]. He wanted them in, where Mao’s sheer weight of numbers could grind them up” (2006: 354, emphasis in original).

the sov iet union



conv incing the chinese to intervene

By October 1950, Stalin turned to China to send its promised troops, continuing to rule out a direct Soviet intervention. The North Korean People’s Army (NKPA) faced total defeat after the UN Command’s successful amphibious landing at Inchon. On September 29, Kim sent Stalin a cable pleading for direct military assistance, but 70 minutes after receiving it, Stalin urged the Chinese leaders to immediately send their “volunteers” instead. Stalin told Mao the situation in Korea was grave and without Chinese support, the Korean Communist regime would collapse and need to establish a government in exile in China (Wolff 2000: 26; Weathersby 1993c: 452–54; 2002: 17; Hao and Zhai 1990: 109; Zhang 1992: 95; Pollack 1989: 217; Stueck 1995: 104–5). He asked Mao to dispatch troops to Korea but did not mention what support the Soviet Union would offer China in return.7 Mao demurred, contending that Chinese intervention would “provoke an open conflict between the USA and China, as a consequence of which the Soviet Union can also be dragged into war.”8 Stalin tried again to pressure the Chinese to intervene. He first reassured Chinese leaders that “the USA, as the Korean events showed, is not ready at present for a big war” and nor was Japan “capable of rendering military assistance to the Americans.” However, Stalin warned, “If a war is inevitable, then let it be waged now, and not in a few years when Japanese militarism will be restored as an ally of the USA.”9 A few days later, in his October 10 meeting with Chinese Foreign Minister Zhou, Stalin also threatened that if the Chinese would not intervene, he was ready to order the NKPA to evacuate. Stalin proposed China accept the main NKPA forces and their weapons into Manchuria, and the Soviets would take the “disabled and wounded” and Koreans of Soviet origin. “Both of us have to share this heavy burden,” he told Zhou. “I suggest that we tell Kim Il-Sung about our advice, which is to withdraw” (quoted in Thornton 2000: 342; see also Goncharev et al. 1993: 189–92; Gaddis 1997: 80–81; Jung and Halliday 2006: 358). Stalin also reneged on his bid of combat air support, which he had promised the Chinese leaders in July. While offering Soviet airplanes and tanks to China, he now stipulated they could only be used to train Chinese troops and categorically forbade their use at the front. In addition, Stalin argued Soviet air support would not be possible for at least two and a half months after China’s intervention (Jung and Halliday 2006: 358; Sheng 1995: 303–4; Chen 1991: 29–30; Hao and Zhai 1990: 111; Khrushchev 1993: 65–66; Weathersby 2002: 19–20). As Goncharev and his colleagues argue, “only after Stalin was certain that the Chinese were fully engaged in the fighting did he secretly introduce fighter planes and send his troops to China” (1993: 216–17). In other words, Stalin continued to avoid direct confrontation

104

the korean war

with the United States, while convincing China to do its “dirty work” at bargain prices. prolong ing the ar mist ice negot iat ions

Although Stalin had instructed Soviet officials soon after the war began to respond positively to peace overtures, after China intervened, he told officials that pursuing negotiations was wrong (Weathersby 1996). After the Chinese campaign removed the threat of an American advance on Chinese and Soviet borders, Stalin focused instead on the benefits he received from a longer war. In a telegram to Mao on June 5, 1951, Stalin argued that “the war in Korea should not be speeded up, since a drawn out war, in the first place, gives the possibility to Chinese troops to study contemporary warfare on the field of battle and in the second place shakes up the Truman regime in America and harms the military prestige of Anglo-American troops.”10 Although we do not have Mao’s reaction to Stalin’s enthusiastic backing of this “training opportunity,” it was apparent that a longer war suited Stalin’s interests. The three Communist allies had different views about the armistice negotiations. When the talks began in July 1951, the North Koreans did not want an armistice, but without the ability to fight alone, they had little say in the matter.11 As Chapter 7 will explain, until the summer of 1952, the Chinese viewed armistice negotiations as another tactic in the ongoing conflict—a tactical maneuver, not a strategic policy shift. But for the Soviet Union, armistice negotiations were only useful as long as they prolonged the war. As a result, Stalin and the hardliners consistently maintained a hawkish stance towards the negotiations. After the initial opportunity for a negotiated settlement collapsed in August 1951, Stalin considered it in the Soviet interests for the war to continue, as long as UN forces did not advance into North Korea again. Russian documents indicate Stalin’s main concern was to ensure the Communist side did not give the impression of weakness in the negotiations, since the Soviet Union could only benefit from a stalemated war (Gaddis 1997: 108–9; Weathersby 1998: 102; Taubman 1982: 221). Thus, after the negotiations started, Stalin instructed Mao to continue “using flexible tactics in the negotiations,” but at the same time, “pursue a hard line, not showing haste and not displaying interest in a rapid end to the negotiations.”12 As Chapter 3 explained, several times after January 1952, the North Koreans begged Stalin for peace. Yet each time Stalin advised Kim to hold on. A stalemated war provided the Soviets with four benefits. First, it tied down American forces, which hindered US efforts to engage militarily in Europe, drained US economic resources and caused domestic political problems for Truman. Tying down US troops in Asia could also alter the balance of power in Eu-

the sov iet union



rope (Jung and Halliday 2006: 353; Zubok and Pleshakov 1996: 70). Second, it created a rift between the United States and its allies over tactics in the Korean War (Nogee and Donaldson 1992: 107). Third, it provided the Soviets with an excellent opportunity to gather intelligence on US technology and military organization. Not only could it field test its new equipment against American technology, but it could gain information from US prisoners of war (Jung and Halliday 2006: 353; Lashmar 1996; Weathersby 1998: FN 4; Jolidon 1996–97). Finally, it created hostility between China and the United States and tied China more firmly to Moscow through dependence on Soviet military and economic assistance. The Korean War also helped to solidify hostility towards the West and impose a Stalinist political ideology and structure at all levels in China (Hamrin 1994: 97; Thornton 2000; Zubok and Pleshakov 1996: 70). In short, because the Soviets were not deployed in the ground war, Stalin was able to push the human costs of the war onto his allies. As a result, from Stalin’s perspective, there was no reason to end the war. It was clearly Stalin’s preference to delay the war’s end, and he consistently took steps—within his own coalition (explained below) and in his interactions with the Chinese (explained in Chapter 7)—to ensure this preference prevailed. the p owerlessness of cr it ics in the sov iet gover ning coalit ion

Not all Soviet leaders agreed with Stalin’s support of a prolonged war. Although Soviet factionalism had been muted during the first months of the war, it reemerged in late 1950. The moderates stressed how the United States was benefiting economically from the war, and they displayed a restrained and sometimes even favorable attitude towards the UN. Moreover, starting in the fall of 1950, moderates indicated they believed that Soviet interests would be served by a negotiated settlement. Specifically, articles in Izvestiia appeared more concerned than those in Pravda about the dangers inherent in an intensified arms race as a result of the Korean War. They also worried the West would launch a preventative war somewhere else. In contrast, hardliners emphasized discord among the states fighting as part of the UN Command and were harshly critical of the UN itself. Like Stalin, hardliners clearly indicated they opposed a negotiated settlement. Specifically, articles in Pravda stressed the West’s untrustworthiness and blamed world tensions during the Korean War on US policy. Pravda articles also called for increased vigilance rather than a curtailment of the arms race, as Izvestiia did (Letteney 1971: 336–39). While the moderates opposed prolonging the war, the evidence suggests they were relatively powerless to change that policy as long as Stalin remained in

106

the korean war

power. First, given Stalin’s extensive history of purging rivals and subordinates who disagreed with his policies, a culture of fear permeated the leadership circle. As a result, the moderates probably “self-suppressed” to some degree. Many of these leaders had direct personal experience with—and indeed, helped orchestrate—Stalin’s capricious personnel changes, and thus they knew the potential consequences of disagreement. For example, two moderate leaders, Beria and Malenkov, owed their careers to Stalin. Both men benefited from the purges and staged political trials of the 1930s; Malenkov had even helped to implement the Great Terror for Stalin. After the purges ended, Malenkov and Beria became Stalin’s top lieutenants, with Beria controlling the internal security forces and surveillance apparatus and Malenkov the Central Committee apparat and Communist Party cadre selection—critical instruments in Stalin’s absolute hold on power. Indeed, the power to appoint is the power to control, a concept known in the Soviet system as nomenklatura (Ra’anan and Lukes 1990: 27–28, 41, 106; Zubok and Pleshakov 1996: 141–43; Ra’anan 2006: 8–9). Given their roles in this powerful machine, Beria and Malenkov knew the potential consequences of disagreeing over policy, perhaps better than anyone. In addition, Beria and Malenkov also might have had the most to lose, in terms of influence and political power, for disagreeing too much. Was opposing some war in Asia worth the cost of losing these positions in the leadership circle? Conversely, some moderate leaders had direct experience with the other side of Stalin’s personnel policies. For example, in March 1949, Anastas Mikoyan was relieved of his duties as minister of foreign trade, although he remained a Politburo member and deputy chairman of the USSR Council of Ministers. The truth is that Stalin’s purges did not end with the Great Terror. The 1949–50 purge of Communist party leaders in Leningrad and Moscow, the 1952 “Kremlin Doctor’s Plot,” and Stalin’s frequent juggling of high-level personnel over the last years of his rule suggest a never-ending quest for security in his regime (Taubman 1982: 203; Alliluyeva 1967: 61, 193, 208; Khrushchev 1974: 245–61, 278–81; Hahn 1982: 118–29, 136–49; Ulam 1973: 705–14; Mastny 1996: 75–76; Zubok and Pleshakov 1996: 144–45; Kostyrchenko 1995; Vaksberg 1994: 247–78; Brent and Naumov 2004: chapter 8; Nicolaevsky 1975: 110–14). Moderate leaders were also strongly influenced by Stalin’s major “cleansing” in Czechoslovakia. When top Czech Communist officials were hanged in Prague as “spies” in December 1952, “Molotov, Mikoyan, Beria and Malenkov knew very well that the same would have happened to them, had Stalin lived longer” (Zubok and Pleshakov 1996: 145; also Medvedev 1989: 795–96; Ra’anan 2006: 3, 11–12). Indeed, in the last months before his death, Stalin was preparing for another round of high-level purges (Taub-

the sov iet union



man 1982: 197–98; Simmons 1975: 227; Garthoff 1953: 28; Mastny 1996: 161; Hahn 1982: 136, 141–43, 149, 156; Medvedev and Medvedev 2003: 38). For both of these reasons, moderates certainly understood and probably feared the consequences of speaking out, which could have led them to “self-suppress” their dissenting views. Second, Stalin took actions to curtail the moderates’ power vis-à-vis war policy. During the war, Stalin decreased the influence of two moderates, Malenkov and Beria, by bringing Khrushchev to Moscow from his post as Ukrainian first secretary. Khrushchev quickly emerged as Malenkov’s competitor on agricultural policy (Hahn 1982: 137; Dunmore 1984: 125; Ra’anan 2006: 9; Azrael 1966: 105–8; McCagg 1978: 307). As Khrushchev remembers it: “One of the reasons Stalin had called me back to Moscow was to influence the balance of power in the collective and to put a check on Beria and Malenkov. . . . I was constantly running up against Beria and Malenkov” (1970: 250). Moreover, in Stalin’s last months, he diffused Beria’s and Malenkov’s influence by establishing rotating chairmanships of various institutions (Ra’anan 2006: 12–13). As James Richter contends, “Not only did Stalin pit the members of his inner circle against each other to ensure that they did not conspire against him, but in addition he compartmentalized, manipulated, and terrorized the various components of his regime to eliminate the possibility of any organized resistance” (1994: 31). The third reason that the moderates had difficulty changing policy is that Stalin gradually excluded his top deputies. While there were few formal membership changes in the Politburo, the Secretariat or the Council of Ministers, Stalin held meetings less frequently, with only four meetings convened in 1952, fewer than any other year. Moreover, Stalin turned increasingly to his personal secretary, Aleksandr Poskrebyshev, and to junior party members, whom he could use against his colleagues in the Politburo. These second-tier leaders had few ties to Malenkov and Beria: “that these men were being used by Stalin against his leading deputies is suggested by the retaliation against them after Stalin died” (Hahn 1982: 145–46, also 136–37; Mastny 1996: 140; Medvedev and Medvedev 2003: 36; Khrushchev 1970: 281, 308–9; Ra’anan 2006: 8). Fourth, starting in 1951, Stalin removed Beria’s protégés from his main power bases, the internal security bureaucracies and the Georgian party organization, and replaced them with people hostile to Beria. Stalin also directed Poskrebyshev to withhold certain reports from Beria. As Peter Deriabin, a former Soviet counterintelligence officer, argues, “Beria was upset and could see his power slipping away and probably felt he was being set up by Stalin to be blamed for internal problems in the future” (quoted in Ra’anan and Lukes 1990: 32; also Dunmore 1984: 125; Medvedev and Medvedev 2003: 38, 41–42; McCagg 1978: 307). Indeed,

108

the korean war

Stalin implicated Beria in the alleged “Kremlin Doctors’ Plot,” a group of Jewish doctors accused in 1952 of killing Andrei Zhdanov (Zubok and Pleshakov 1996: 144; Hahn 1982: 107; Brent and Naumov 2004: chapter 8; Medvedev and Medvedev 2003: 1–2, 18–23; McCagg 1978: 307; Nicolaevsky 1975: 110–14). By 1952, Stalin had “compelling reasons” for getting rid of Beria, because as security chief he controlled an important instrument of terror—and thus, power (Mastny 1996: 160). Many authors argue Beria especially was being targeted as Stalin began preparations for a new round of purges in the last months before his death (Simmons 1975: 227; Garthoff 1953: 28; Mastny 1996: 161; Hahn 1982: 136, 141–43, 149, 156; Medvedev and Medvedev 2003: 38; Nicolaevsky 1975: 113–14). Finally, Stalin surprised his deputies at the 19th Party Congress in October 1952 by reorganizing the party’s central institutions to minimize the influence of his top deputies. These institutional changes constitute a neutral governing coalition shift, since change in the coalition’s composition was intended to sideline rivals, not to change war policy. At the congress, the first in 13 years, Stalin abolished the old 12-member Politburo and created a Central Committee Presidium, with 25 members and 11 candidate members. Stalin selected Presidium members without consulting other Politburo members, and he included some of Malenkov’s obvious enemies. By admitting more junior leaders and increasing its size, Stalin diffused the Presidium’s influence and rendered it less suitable for making top-level policy decisions. The Central Committee also lost influence in policymaking through a similar increase in size (Hahn 1982: 147; Garthoff 1953: 2–3, 5–12; Medvedev and Medvedev 2003: 2, 36–43; Khrushchev 1970: 279–80; Azrael 1966: 108). In further decisions between November 1952–January 1953, Stalin also removed some top deputies from the Council of Ministers, while assigning a group of leaders different signature authorities and rotating chairmanships. These decisions effectively created the structure for a diffused “post-Stalin pseudo-collective leadership” (Ra’anan 2006: 10–14).13 Each of these actions implies that Malenkov, Beria and other moderates had become quite dissociated from Stalin’s policies at the end. Indeed, some authors have even suggested that by December 1952, Malenkov and Beria were planning Stalin’s overthrow or murder (Mastny 1996: 163, 168–69; Brent and Naumov 2004: chapter 10; Medvedev and Medvedev 2003: chapter 1; Ra’anan 2006: 5–7).14 In light of Stalin’s tendency towards secrecy, it is even possible that the moderates did not participate actively in Soviet decisions related to the Korean War. Regardless, the evidence demonstrates a power imbalance existed in the Soviet governing coalition. Whether the potential critics of Stalin’s war policy were unwilling to express their criticism for fear of expected penalties, or whether they did speak out but simply lacked the power necessary to overrule Stalin, the result was the same. Stalin did not want to end the war, and Stalin’s preference prevailed.

the sov iet union



How the Information Obstacle Exacerbated the Preference Obstacle Information processing dynamics exacerbated the effect of Stalin’s preferences in two ways. First, and most importantly, because Stalin viewed information as power, he strictly controlled other coalition members’ access to it. Controlling information allowed Stalin to control his subordinates. For example, declassified archives show that, among the Soviet coalition members, only Stalin communicated by telegram with the North Koreans and Chinese during the war. Stalin alone received personal communications from his ambassador in Pyongyang and the Soviet military forces and technicians there to assist Kim. This suggests Stalin had access to information from the front that he could keep secret from other leaders. Moreover, Stalin had many personal channels of “special information,” with the intelligence services reporting the best information directly to him. Similarly, Stalin shrouded the rationale behind his decisions in secrecy. No individual was allowed to know the whole picture—just pieces of it (Ra’anan and Lukes 1990: 32; Chuyev 1991: 334–35; Jeon 1996–97; Zubok and Pleshakov 1996: 86). As Vladimir Petrov argues, While Mao sometimes shared his reasoning with his close associates, Stalin consistently hid his, often forbidding them to keep minutes of the more sensitive conversations he had with foreign leaders. . . . Obsessed with security, he used various assumed names and informed others on a need to know basis only. (1994a: 5)

Thus, other coalition members did not necessarily have access to as much information about the war as Stalin, which certainly impeded full discussion of policy options among the leaders. Stalin’s practice of jealously guarding information trickled down into the Soviet intelligence system, as well. Soviet intelligence analysts competed for the ear of Politburo members, which not only prohibited a smooth coordination of information flowing to top leaders, but also hindered analysts in synthesizing information effectively. Paradoxically, intelligence collection and analysis under Stalin was both highly centralized and severely stove-piped. After centralizing all foreign intelligence departments in 1947 under one umbrella agency, the Committee of Information (KI),15 Stalin eventually decided interagency rivalry in intelligence would serve him better. Thus, in November 1949, he transferred military intelligence to the General Staff. By November 1951, he returned civilian intelligence to the Ministry of State Security (MGB, the KGB’s predecessor) and disbanded the “large” KI for a “small” one (Zubok 1992: 3–7; Andrew and Gordievsky 1990: xi). Officially affiliated with the Ministry of Foreign Affairs, the small KI had its

110

the korean war

own budget and direct communication lines with the Kremlin. Like the large KI, its main task was reducing disinformation, not forecasting or commenting on alternative policy options. Lacking many of the senior career officials who had been “borrowed” from the MGB and GRU (Soviet Military Intelligence) for the large KI, the small KI was younger and less experienced. While the small KI’s officers were entitled to receive foreign intelligence information, in practice the MGB and GRU withheld information, especially human intelligence. This left the small KI with embassy reports and open sources, such as newspapers, official announcements, documents and communiqués (Zubok 1992: 8). These organizational pathologies in the intelligence system not only reinforced Stalin’s tendencies towards secrecy, but in practice they meant Stalin had few other advisors with whom he could double-check his understanding of the facts—since these others leaders did not have access to the information. With such an avalanche of information directed exclusively at him, Stalin certainly did not have the time to digest it. Inundated with raw, unanalyzed data, Stalin was able to pick and choose among the data, paying attention to information that conformed to his pre-existing beliefs and preferences and disregarding information that did not. Thus, Stalin’s preferences were reinforced by the information he received, allowing him to manifest many of the common psychological biases discussed in Chapter 2. As a result, “Stalin was both overwhelmed and ignorant at the same time. . . . Often Party leaders’ decisions were not the result of an analytical process—sorting out facts and factors and choosing the optimal mix—but of heuristic reasons, based on selected events or even momentary moods” (Zubok 1992: 3). As Zubok and Pleshakov contend, “it was not the intelligence that led him. On the contrary, it was Stalin’s mind that employed and directed the intelligence . . . his superior system of espionage fed Stalin’s worst fears” (1996: 24). Second, given the regime’s culture of fear and history of coercion (e.g. Medvedev 1989), it is not surprising that information processing in Stalin’s regime exemplified the inverse relationship between coercion and information flow. The information Stalin and other leaders received about the war was tailored to support and reinforce Stalin’s known preferences and beliefs. For example, in May 1949, when the hardliners were trying to make the case that South Korea was planning to launch an attack on the North, Soviet Ambassador Shtykov used intelligence information out of context to conform to the hardliners’ preferences. In his May 2 report, Shtykov noted accurately that ROK forces were being expanded with US assistance and that Rhee’s government was taking steps to increase its combat readiness. However, Shtykov “failed entirely to note the disparity in aims” between Rhee and his US patron, which at that time was trying to disengage from Korea. As Weathersby put it, “instead—perhaps out of an un-

the sov iet union



derstandable impulse towards self-protection—he merely repeated Stalin’s own conclusions, adding supporting detail.” In other words, despite having an extensive intelligence network that would have known of the sharp divergence between American and South Korean intentions, “Shtykov did nothing to correct his boss’ misperceptions” (2002: 5). Similarly, declassified small KI documents suggest that the authors clearly tailored their conclusions to the existing Soviet policy guidelines. The conclusions of the most important classified documents match closely, sometimes verbatim, major articles and editorials in Pravda, the official organ of the Communist Party’s Central Committee and the mouthpiece of the hardliners (Zubok 1992: 2). One of the reasons why analysts presented what Stalin wanted to hear is that he was quite specific about the intelligence information he sought and directed his subordinates to obtain only that information. According to one knowledgeable KGB source, “We almost never went out looking for information at random. Instead, orders to look for certain specific things would come from above . . . our agents abroad were directed to find out at all costs. This highly authoritarian style produced excellent results” (quoted in Zubok and Pleshakov 1996: 24). Ironically, the Soviet Union had especially deep penetration with its early Cold War espionage activities, leading to better information than that available in the West. However, this intelligence was not used to see the situation clearly but to reinforce Stalin’s existing beliefs (Andrew and Gordievsky 1990: 469; Watts 1990). Even had Stalin been more receptive to objective information, the system did not select for open-minded thinkers who could provide honest, balanced perspectives. As Georgi Arbatov writes in his memoirs, by the 1950s, even the brightest young Soviet minds had been “clogged and blunted by the long years of Stalinism” (quoted in Zubok 1992: 2). Stalin’s paranoid, xenophobic and ideological nature spread down the bureaucratic hierarchy. After the Great Terror, the guiding criterion for selecting new bureaucrats, diplomats and intelligence officers was “honesty”—Molotov’s word—which Zubok and Pleshakov translate as “the absence of independent thinking and unquestioning loyalty to Stalin and his regime” (1996: 23, 88). Thus, information processing pathologies in the Soviet regime exacerbated the preference obstacle. By jealously guarding information, Stalin was able to keep those who might disagree with his preferences in the dark. Moreover, because Stalin was inundated with raw data, he could focus on that information which conformed to his pre-existing preferences and disregard that which did not, without other coalition members’ opinions being voiced as a corrective. Finally, much of the information Stalin and other leaders received was tailored to support Stalin’s known preferences, rather than to provide an accurate picture

112

the korean war

about the war. Paired with Stalin’s actions to mitigate the moderates’ influence, these information dynamics leave little reason to wonder why Stalin’s preference to prolong the war prevailed so successfully.

The New Soviet Governing Coalition: Outlasting Stalin’s Preferences There was no movement in the Communist position until after Stalin’s death on March 5, 1953. His death unleashed a radical change in the Soviet approach to the Korean War. As Weathersby argues, “despite the great uncertainty and anxiety within which the new collective leadership operated, it nonetheless moved immediately to bring an end to the war in Korea” (1998: 108). Indeed, it is remarkable they could take such decisive action within ten days of Stalin’s funeral. The leadership troika that emerged was—by the very nature of all trios—an awkward, dysfunctional structure prone to in-fighting and power grabbing. Nonetheless, these new leaders took control of the country and quickly enacted their long-held but dormant preferences, leading them to end the war. When Stalin died, the question of succession had not been on the agenda. All of Stalin’s last minute-maneuverings, from the planned purge to the massive reorganization of state structures at the 19th Party Congress, meant a decision about his successor was wide open, since he had added new blood to mitigate the effects of his previously chosen top deputies (Garthoff 1953: 26, 42; Ra’anan 2006: 12–13; Medvedev and Medvedev 2003: 34–44). As Zubok argues about the months after Stalin’s death, “Stalin’s former lieutenants, the oligarchs of the regime, mauled and bruised each other, seeking to change the power balance by appealing to the party and state elites, heads of the central [Communist Party] apparatus, secretaries of regional party committees, leaders of powerful branches of the economic, military and security structures” (2000, emphasis added). That these new leaders could agree to act decisively to end the Korean War only reinforces what a high priority they assigned to a settlement. The new Soviet leadership was led by the troika of Malenkov, Beria and Vyacheslav Molotov.16 Stalin’s “mantle” was divided among them “so that no one would be able to wear it whole” (Mastny 1996: 171). Malenkov became the chairman of the Council of Ministers, Stalin’s formal successor as head of state. Malenkov also assumed the role of Communist Party first secretary, but as other aspirants joined forces to diminish his power, he was relieved “at his request” within six days and the position was formally abolished. Beria assumed control of the interior ministry and merged it with MGB, essentially returning to his power base in the security services. Molotov, who had lost almost all real power

the sov iet union



in 1949 when he left his post as foreign minister, returned to that position. Nikita Khrushchev controlled the central party apparat and by September 1953 also revived the position of party first secretary (Goldgeier 1994: 21–22; Zubok and Pleshakov 1996: 154–54; Larson 1997: 42, 57; Hahn 1982: 104–9; Garthoff 1953: 6, 13, 17, 19, 32; Richter 1994: 31; Nogee and Donaldson 1992: 115; Taubman 1982: 231; Ra’anan 2006: 15–28). Within 24 hours of the announcement of Stalin’s death, the new leaders discarded his painstaking October 1952 reorganization of the party organs. On March 6, they reduced the newly created Presidium to ten members and four alternates, and they relieved the junior leaders on which Stalin had relied during his last year. They also restructured state institutions, with Presidium members assuming direct control over key ministries, as had been the practice from 1945 to 1949. In this manner, they shifted the policymaking functions from the party to the state and consolidated the chief instruments of power directly into the hands of the old Politburo members. In other words, the new leaders essentially voided Stalin’s bureaucratic maneuvering. Further changes—including a public repudiation of the alleged “Kremlin Doctor’s Plot” and the Georgian party purge, both of which had sullied Beria—were announced over the next few weeks. As Garthoff modestly observed two months later, “it should be noted that a certain degree of flux has prevailed” (1953: 36, also 17–20, 37–38; Larson 1997: 42; Nogee and Donaldson 1992: 115; Volkogonov 1991: 575; Ra’anan 2006: 22–27; Azrael 1966: 122–24). Again, the fact that so much bureaucratic jockeying and subterfuge was occurring makes it all the more remarkable that they could simultaneously take actions to end the Korean War. The new leaders had wanted to end the war for a while. As noted, Beria and Malenkov led the moderate faction, which advocated more conciliation towards the West and had supported a negotiated settlement in Korea. Although militant Molotov had always supported a hard line towards the West, even he “believed that Stalin should never have approved the North Koreans’ plan to reunify their country” (Larson 1997: 45). Moreover, this new troika “were seriously concerned that the new Republican administration . . . might escalate the war,” and they looked differently at information about US potential use of atomic weapons in Korea (Zubok and Pleshakov 1996: 155; Mastny 1996: 164). The major foreign policy goal of the new troika was to avoid a Soviet-American conflict; the new leaders quickly expressed interest in disarmament and détente with the United States. They wanted to curtail military and industrial programs, so as to free resources for neglected domestic programs and agriculture. They were also more focused towards Europe and the growing crisis in the German Democratic Republic (Azrael 1966: 125–26; Mayers 1986: 113; Zubok

114

the korean war

and Pleshakov 1996: 148; Stueck 1995: 327; Richter 1994: 35, 37; Larson 1997: 45).17 The new troika quickly began easing international tensions, a policy that continued even after Beria’s ouster later that year. Beginning in April, hardline policies were relaxed in the GDR.18 By May, the new leaders restored diplomatic relations with Israel, Yugoslavia and Greece and dropped territorial claims to Turkey. In June, they supported a new Hungarian leader, Imre Nagy, in his embarking on a “New Course” prioritizing consumer goods (Richter 1994: 34; Hahn 1982: 157). Part of their different worldview came from their previous government experiences. For example, Beria and Malenkov had been in charge of constructing several military industrial complexes, for atomic weapons, cruise missiles and air defense systems. Beria, especially, had a deep history administering atomic research and development. One result of this background is that for Beria and Malenkov, “the Cold War was first of all an arms race. . . . Beria was the first among the Kremlin leadership to recognize that the Bomb provided the ultimate answer to all problems of security and international policy” (Zubok and Pleshakov 1996: 149; also 141–42, 153; Goldgeier 1994: 2–5). Just as US President Eisenhower’s reliance on massive retaliation mitigated the need to fight Communists on the ground in the Korean War, the Soviet leaders’ focus on the arms race may have lessened their need to back Asian Communists in the war, as well. Because of these other policy priorities and their different conceptualization of the costs and benefits of prolonging the Korean War, the new Soviet leaders moved immediately to end it. First, at Stalin’s funeral on March 9, Malenkov emphasized “the possibility of the prolonged coexistence and peaceful competition of two different systems, capitalist and socialist.” Then, in a March 15 speech to the Supreme Soviet, Malenkov unveiled a “peace initiative.” He declared, “There is not one disputed . . . question that cannot be decided by peaceful means on the basis of mutual understanding of interested countries,” and singled out the United States as among those countries to which this statement applied (Folliot 1956: 12–13; Thornton 1982: 239; Zubok and Pleshakov 1996: 155; Mastny 1996: 172; Stueck 1995: 308). Finally, on March 19 the Council of Ministers adopted a resolution to end the war, with attached letters to Mao and Kim Il-Sung. As Weathersby notes, “in tortuously convoluted language reflecting the great psychological difficulty of altering the policy pursued by Stalin,” the resolution declared that the Soviet Government has reached the conclusion that it would be incorrect to continue the line on this question which has been followed until now, without making those alterations in that line which correspond to the present political situation and which ensue from the deepest interests of our peoples, the peoples of the USSR, China and Korea . . . (1996: Document 111)

the sov iet union



The resolution went on to outline statements that the Chinese, Korean and Soviet governments should make to indicate their willingness to resolve outstanding issues and reach an armistice agreement. We shall return to the Communist actions that resulted from this domestic coalition shift in Chapter 8. In sum, since the majority of the new leaders had argued against the war from the start and disagreed with the decision to prolong it, it is not surprising that once in power, they would want to enact these long-held but dormant preferences. What is remarkable is that they could take such decisive action towards ending the war within ten days of Stalin’s funeral. It speaks volumes to the noncontroversy of this major policy shift among Stalin’s successors and to the relative powerlessness they must have felt under Stalin’s rule.

Alternative Explanations This section briefly examines four alternative explanations for Soviet behavior in the Korean War. The first alternative explanation, most different from the one advanced here, is that Stalin was an absolute dictator. This explanation assumes that Stalin alone was responsible for all war decisions, because other members of the Communist Party were irrelevant to decision-making about the war. Yet, as the empirical evidence in this chapter demonstrates, Stalin did not make decisions about and implement policy alone. The foreign policy compromise in December 1949 is a direct result of Stalin mediating between two ideological factions and marginally siding with the hardliners. Moreover, in choosing to support Kim Il-Sung’s venture, Stalin overcame his personal concerns about potential US intervention in large part because of the arguments for war advanced by the hardliners. Finally, because the Soviet Union was led by a troika after Stalin’s death, this explanation does not speak to Soviet behavior during that time. For these reasons, this explanation is less nuanced for explaining Soviet behavior during Stalin’s rule than the domestic coalition theory. A more generalized version of the first alternative explanation would focus on regime type to explain Soviet behavior. For example, Goemans (2000a) argues dictatorships will rationally update their preferences about the war and settle as soon as the costs of prolonging the war outweigh the benefits. Goemans argues dictators will behave like unitary rational actors—even when they are losing— because they have sufficient capacity to repress potential opposition and thus need not worry about personal punishment resulting from defeat in war. The Soviet Union is usually coded a dictatorship during Stalin’s rule and an oligarchy after his death. However, with Goemans’ non-standard coding rules, he would consider the Soviet Union a dictatorship during both periods.19

116

the korean war

Thus, using his coding, Goemans would predict the Soviet Union would behave rationally during Stalin’s rule, and it did. However, Stalin did not continue a losing war because he could repress potential opposition and thus need not fear personal punishment for losing. Instead, Stalin continued the war because the majority of its costs were being borne by other states, and indeed, Stalin had explicitly orchestrated the conflict so that was the case. Thus, while both Goemans and I would predict Stalin rationally to choose to continue a losing war, we differ about which factors were most important in his cost-benefit calculation. What drove the Soviet decision to continue the war during Stalin’s rule was not the regime type, but the fact that the majority of the war’s costs were not being borne by Soviet decision-makers. Using Goemans’ non-standard regime type coding, the successor regime’s behavior accords with Goemans’ theory. Given its decision to end the war promptly, because the costs outweighed the benefits from their perspective, the successor regime behaved as a rational unitary actor would. In contrast, standard regime type coding would classify the troika as an oligarchy, in which case its behavior would not support Goemans’ theory. (Goemans argues oligarchies in a losing war will gamble for redemption in an attempt to turn the battlefield around, in order to stave off personal punishment after defeat.) Regardless, the domestic coalition theory better explains how the troika’s cost-benefit calculation differed from Stalin’s—and for at least two members of the troika (Malenkov and Beria), had even differed before the war began. The new leaders viewed a prolonged war as a cost they did not want to bear in light of other policy priorities, in line with their different conceptualization of the war’s costs and benefits. A third alternative explanation, advanced by Stalin’s biographer, Dmitrii Volkogonov, is that Stalin had decided to end the war on February 28, but that night, he suffered the fatal stroke that killed him on March 5. Stalin had dinner that evening with Malenkov, Beria, Khrushchev and Bulganin. To recount Volkogonov’s entire argument: As usual, they discussed a wide range of topics. Bulganin gave an account of the war in Korea, confirming Stalin’s view that the situation had reached a stalemate. He decided he would tell Molotov the next day to advise the Chinese and North Koreans to “try to get the best deal they could in the talks,” but in any event to try to bring the armed conflict to a halt. (1991: 570)

Although Volkogonov conducted detailed archival research for his study, he does not offer any references for the assertion in this passage. Moreover, besides this passage, Volkogonov includes only two paragraphs about the Korean War in the entire 600-page biography (1991: 540–41). Thus, not only is it impossible to assess

the sov iet union



whether Stalin actually said this, but the miniscule role the Korean War plays in Volkogonov’s account makes it difficult to evaluate the merits of this assertion. Given the evidence available today, there is no way to confirm or deny Volkogonov’s account. If Volkogonov is right, Stalin had changed his mind and decided to end the war, which would suggest the preference obstacle was about to be lifted. However, it is impossible to tell if this counterfactual may have come to pass. What we do know is that Stalin took no actions to end the war before his death. Therefore, even if his expectations had changed, they did not produce the war’s end. Without a reference, we cannot know for sure how Volkogonov knows this “fact” he reports. Presumably, it must have been from the other men present at the dinner—and these men were precisely the ones who wanted to end the war, anyway. It is certainly possible that these men, who assumed power after Stalin’s death, said that Stalin said this. Then it would not look so drastic for them to be going against his wishes. It was in their interest to have it seem like Stalin wanted to end the war, because it provided justification for them to enact their own longheld but dormant preferences for ending the war. Fourth, Jung and Halliday build on Volkogonov’s account to argue that Stalin had decided to end the war because he felt “unremitting pressure” from Mao to share nuclear technology and did not want to. In their depiction of events, President Eisenhower “suggested in his State of the Union address [on February 2, 1953] that he might use the atomic bomb on China. This threat was actually music to Mao’s ears, because he now had an excuse to ask Stalin for what he wanted most: nuclear weapons” (2006: 368). Jung and Halliday argue that after this speech, Mao dispatched his top nuclear scientist, Qian San-qiang, to Moscow. In their words, Mao’s message boiled down to this: Give me the Bomb, so that you will not be drawn into a nuclear war with America. This confronted Stalin with a serious dilemma, as Russia had a mutual defense pact with China. Stalin did not want to give Mao the Bomb, but he was worried about Eisenhower. It was under this unremitting pressure—from Mao as much as from the West—that Stalin, it seems, decided to end the Korean War. (2006: 368)20

Jung and Halliday’s account is speculative and somewhat misleading. As Chapter 8 will explain, Eisenhower did not make nuclear threats in his State of the Union address. US nuclear threats did not occur until May 1953—and it remains unclear whether those threats were ever even received by the Communists. The only threat Eisenhower made in this speech was his announcement that the US Seventh Fleet would be removed from the Taiwan Strait. Therefore, while Mao may have wanted nuclear technology and sent his nuclear scientist to Mos-

118

the korean war

cow, the evidence does not support the assertion that Mao credibly threatened to draw the Soviets into a nuclear war with the United States unless China received atomic weapons from Stalin.

Conclusions This chapter has explained the Soviet role in prolonging the Korean War. The Soviet Union under Stalin provides strong evidence of the preference obstacle, supported by the information obstacle, as reasons for prolonging the Korean War. Stalin did not want to end the war because he was benefiting from it without assuming many of its costs. A stalemated war provided the Soviets with four benefits. First, it tied down American forces, which stalled US efforts to engage militarily in Europe, drained US economic resources and caused domestic political problems for Truman. Second, it created a rift between the United States and its allies over tactics in the Korean War. Third, it provided the Soviets with an excellent intelligence-gathering opportunity. Finally, it created hostility between China and the United States, tying China more firmly to Moscow through dependence on Soviet military and economic assistance. After siding with the hardliners to support the war, Stalin remained reluctant to risk provoking direct military conflict with the United States. Until the Chinese intervention, Stalin tried to prevent such escalation, while simultaneously cajoling China into battle. Whether this was because he believed such intervention would not cause global war or because he wanted to stop Mao from establishing friendly relations with the United States and becoming the bulwark of Asian communism, Stalin’s priority was taking care of Soviet interests. By the time the war bogged down in mid-1951, Stalin had orchestrated the situation so that he was not bearing the war’s costs directly, yet gaining plenty of benefits. Thus, he consistently took steps—within his own coalition and in his interactions with the Chinese (explained in Chapter 7)—to ensure this preference prevailed. While the moderates opposed this war policy, they were relatively powerless to effect its change as long as Stalin remained in power. Stalin actively curtailed their influence by excluding, rotating and demoting the moderates and by strictly controlling information. Given the culture of fear, the moderates also “self-suppressed” to some degree. Information processing pathologies in the Soviet regime exacerbated the preference obstacle. Not only was Stalin was able to keep those who might disagree with him in the dark, but much of the information leaders received was tailored to support Stalin’s known preferences, rather than to provide an accurate picture about the war. For all of these reasons, it was not until Stalin’s death that the moderates

the sov iet union



could finally enact their preferences and end the war. The fact that they decisively moved to end the Korean War in less than ten days after Stalin’s funeral speaks volumes to the non-controversy of this decision among Stalin’s successors, as well as to the fact that it was Stalin who kept the war going. These new leaders took control of the country and quickly enacted their long-held preferences, leading them to end the war. In other words, the domestic coalition shift dislodged the stickiness of Stalin’s preferences and allowed the war to end.

5

The United States, Part 1 Trapped by the NSC-68 Mindset

If the Soviet Union under Stalin exemplifies the preference obstacle to ending war, then the United States under President Harry S. Truman provides strong evidence of the two other obstacles, information and entrapment. Chapter 2 explained that an information obstacle occurs when leaders do not know they should end the war. The information obstacle hinders Bayesian updating, impeding the ability to use all available information and update expectations. With the information obstacle, such rationalist updating does not occur, because leaders (1) receive poor quality information, (2) do not necessarily share the same information, (3) use different indicators to assess the war or (4) have individual or (5) organizational biases that prevent them from using all available information. In contrast, the entrapment obstacle occurs when leaders want to end the war but cannot. Entrapment can occur when leaders’ hawkish constituencies want to continue the war, and the leaders must acquiesce to remain in power. Leaders can face this political entrapment from a domestic constituency or from external allies. This chapter examines domestic entrapment, while Chapter 7 will consider entrapment by an ally. Information and entrapment obstacles can occur in tandem, as Korean War decision-making in the US governing coalition demonstrates. Entrapment dynamics can lead decision-makers to become “too invested to quit” and thus process information so as to justify a course of action despite evidence against it. For example, they can ignore, discount or reject information that does not fit with favored or well-established beliefs, while taking at face value data that does. The psychological commitments and domestic political constraints of entrap-

the united states, part 1



ment can provide the motive, while the complexity and uncertainty of information during war can provide the opportunity, for psychological biases to persist. In other words, the logic of entrapment can make leaders reluctant to confront painful truths, causing them to choose not to know they should end the war. Unfortunately, the longer the war continues and the more trapped a leader perceives himself to be, the more information obstacles can hinder rational updating. This chapter examines the interacting dynamics of information and entrapment obstacles, at the macro-level, in US decisions about prosecuting and ending the Korean War. The Truman governing coalition embodied what I call the “NSC-68 mindset,” a set of assumptions about the US-Soviet relationship within which it perceived the Korean War. The NSC-68 mindset not only blinded the governing coalition to information that may have helped it end the war sooner, but also drove a domestic mobilization effort from which it was difficult to back down. Fearing Soviet aggression worldwide, the Truman Administration used the Korean War to embark on a massive Cold War rearmament effort and justify its globalist agenda to Congressional and public constituencies. However, winning support for this mobilization required building a coalition with moderate Republicans and Asia-lationists, leading Truman to overturn previous Asia policies and inadvertently draw China into the war. Furthermore, because it believed the only course was to conduct a limited war in Korea while preparing for the global war expected in Europe, the Truman coalition got caught in a dilemma. The threat inflation that resulted from selling mobilization to the American people stoked hawkish constituencies for an all-out victory in Korea; hence, any efforts to move towards a negotiated settlement renewed calls of Truman being “soft” on Communism. This chapter lays out half of the US story in the Korean War, discussing the interacting dynamics of entrapment and information obstacles at the macro (strategic and philosophical) level. Chapter 6 will examine those dynamics at the micro (operational and policy) level, in the US decision to adopt a policy of voluntary repatriation of prisoners of war (POWs). However, it was only with a domestic governing coalition shift—the election of President Eisenhower—that entrapment and information obstacles at both levels were lifted and the United States could change course towards ending the war. While Eisenhower retained the same war aims that Truman had espoused, his governing coalition brought different actors—with different constituencies, worldviews and conceptualizations of the war—that left them freer to consider an expanded palate of strategic options. Notably, as a Republican, Eisenhower could more easily accept terms Truman could not, while placating the hawkish Republican right wing with other policy issues.

122

the korean war

This chapter is divided into two sections and a conclusion. The first section explains the US governing coalition during the Truman Administration, as well as the wider domestic political context in which it operated. The second section outlines how the NSC-68 mindset developed and became entrenched during the North Korean attack, deepened with Chinese warnings as the UN Command advanced north, and blossomed in the “entirely new war” that resulted from the Chinese intervention. It also examines the Truman Administration’s responses to these events, in light of the domestic environment. Chapter 6 will discuss the voluntary POW repatriation policy, the Eisenhower coalition shift and alternative explanations for US behavior during the Korean War.

The United States Governing Coalition The Truman governing coalition relied heavily on a departmental approach to solving problems, with Secretary of State Dean Acheson playing a dominant role in foreign policy decision-making. Before the Korean War, the governing coalition adopted a Europe-first internationalist approach to containing Communism, relying on moderate Republican internationalists to help push these policies through Congress. To sell these programs to Congress and the public, the Truman governing coalition inflated the Soviet threat and relied on anti-Communist propaganda, but this decision had longer-term political implications that eventually constrained Truman during the Korean War. w ithin the t ruman gover ning coalit ion

When the Korean War started in June 1950, Truman had been president for more than five years but was rarely involved directly in the foreign policymaking process. Truman relied more on regular government organizations—the cabinet departments, Joint Chiefs of Staff (JCS), Central Intelligence Agency (CIA) and National Security Council (NSC)—in foreign policy decision-making than his predecessor had. Truman was aware he lacked formal foreign policy training and thus was willing to rely on advice from others. With no foreign policy specialists on his personal staff, he usually deferred to recommendations from his cabinet officers on issues of national security: Secretary of State Dean Acheson and Secretaries of Defense Louis Johnson (until September 1950), George C. Marshall (from September 1950 to September 1951) and Robert A. Lovett (from September 1951 to January 1953). As a result, these officials had great influence over policy (Stueck 1995: 137; Truman 1955: 13–17; Acheson 1969: 730–31; Crabb and Mulcahy 1991: 80; Donovan 1984: 81–83; Leffler 1992: 363; Sterling 1998: 349). Truman’s decision-making style reinforced this distinctive departmental ap-

the united states, part 1



proach to problems. He organized the policymaking process in terms of functional expertise, with himself playing the role of “chairman of the board, hearing sundry expert opinions on each aspect of the problem, then making a synthesis of them and announcing the decision” (George 1980: 151). His widely quoted motto—“The buck stops here”—reflected his firm conviction that he was in charge of American foreign policy and the ultimate decision-maker. Yet his decisiveness sometimes bordered on impulsive, often leading to premature cognitive closure—as the POW policy discussion in Chapter 6 will show (Crabb and Mulcahy 1991: 78–79). As Henry Wallace has noted, Truman “seemed eager to make decisions of every kind with the greatest promptness. Everything he said was decisive. It almost seemed as though he was eager to decide in advance of thinking” (quoted in Larson 1985: 146–47). Truman’s decisiveness, paired with his desire to administer the government through the cabinet, had important implications for information processing in his governing coalition. As Richard Neustadt argues, [Truman] did not seek as avidly . . . for bits of information unrelated to the day’s affairs. . . . Truman did not often try to fit the pieces into patterns, or to speculate about their implications, or to make connections between something learned in one context and something done in another. Characteristically, he concentrated on the here and now in terms befitting a decision machine, looking for what was to be decided, looking at what bore directly on the case, and tending to dismiss all else as interesting but irrelevant. Moreover, while he dealt with the decision next in line, the ones already processed and the ones that might come after were not often on his mind. . . . Just as in the case of information, Truman usually let other men’s initiatives time his decisions for him. The issues that existed were the issues [his advisors put] on his desk. (1990: 145, emphasis added)

Hence, unless specific information was brought to Truman as part of a particular decision “package,” he was unlikely to know and consider it. Moreover, those with the greatest access to Truman—Acheson and, during his tenure, Marshall—tended to dominate the policy process. Their preferences (and biases) significantly influenced which information Truman received and which policy options were considered (Neustadt 1990: 148). Among Truman’s subordinates, Acheson dominated Korean War policymaking and served as a powerful veto player in the coalition. Truman’s relationship with Acheson was extremely close, and Acheson’s views received priority in policy deliberations, usually winning out when advice was divided along department lines (Foot 1985: 28; James 1993: xi; Sterling 1998: 350). Nonetheless, Truman sometimes overruled him because of domestic political pressures. With many Congressional critics, Acheson was not well liked outside the executive branch, and McCarthyist

124

the korean war

attacks on his department decreased his ability to help sell Truman’s policies to the public (Hogan 1998: 276; Foot 1985: 30–31; Hamby 1990: 565–66). Acheson was part of the East Coast liberal internationalist and financial elite (Schulzinger 1984; Isaacson and Thomas 1986). Thus, not surprisingly, his major preoccupation was Europe, especially relations with Britain—perhaps because his major foreign policy conditioning had been in Western Europe. For example, during his tenure as secretary of state, Acheson spent more time with the British ambassador than with all other ambassadors combined. Acheson regarded Europe as the center of gravity in the struggle with the Soviets. Like Truman, he had little understanding of China and considered policymaking towards China an imposition (Smith 1972: 144–45; Sterling 1998: 350; Tucker 1983: 174). In addition to his European bias, Acheson believed Truman was under-spending on national security. He disliked Truman’s decision to subordinate the defense budget to his domestic economic agenda (Hogan 1998: 276). By early 1950, Acheson had assembled an impressive group of policymakers at State who shared his perspective, including Assistant Secretary of State for Far Eastern Affairs Dean Rusk and Paul Nitze, the new director of State’s Policy Planning Staff. If anything, Nitze and Rusk advocated a more globalist perspective for containing Communism than Acheson’s Europe-first viewpoint (Snyder 1991: 270–71). For example, Rusk’s experiences in Asia during WWII predisposed him “to believe the US had ‘an obligation to stand’ not just in one part of the world but everywhere it might be called upon to do so” (Foot 1985: 32). Similarly, Nitze—the architect of NSC-68 discussed below—believed the Soviets has aggressive intentions for global domination that required greatly increased foreign aid and military spending to contain (Fordham 1998: 43; Donovan 1982: 158–60; Pierpaoli 1999: 21; Pemberton 1989: 128–31; Rearden 1993: 193). As Hogan notes, Every bone in Nitze’s body ached with suspicion of the Soviet Union. . . . Given Acheson’s own geopolitical vision, it is clear that he picked the right man to join him on the front line . . . against the president and others who were trying to balance the needs of the welfare state against those of the warfare state without throwing the budget into permanent deficit. (1998: 294)

Clearly, Acheson assembled a team to make the case for a more aggressive, extensive (and expensive) plan to combat Communism globally. While Acheson’s plan for increased security spending mirrored the preferences of the uniformed Pentagon, their focus was more Asia-centric. This Asia-first perspective made sense, given their desire to extend the US military’s occupation of Japan into an “Asian strongpoint” and “a reliable set of bases that could be used for air attacks on the Soviet Union in a major war” (Snyder 1991: 269). The military members of the Truman coalition included (1) JCS Chairman General Omar

the united states, part 1



Bradley, (2) the service chiefs—General J. Lawton Collins (Army), General Hoyt S. Vandenberg (Air Force), and Admiral Forest Sherman (Navy)—and (3) the UN forces commander—General Douglas MacArthur (until April 1951), General Matthew Ridgway (April 1951–May 1952) and General Mark W. Clark (after May 1952). While Truman generally believed military leaders provided advice with the least partisan motivations, he had a long-standing fear of “political cliques” in the military. He also reportedly had strong antipathy towards the “flashy” MacArthur (Neustadt 1990: 148; Donovan 1977: 141–42; Truman 1956: 50–51, 59–60, 422–27; Nelson 1985). Two characteristics of these military officers are important for decisionmaking in the Truman governing coalition. First, until his April 1951 dismissal, MacArthur held a preeminent role in the coalition—albeit somewhat as a loose cannon. While ostensibly subordinate to the JCS, MacArthur’s seniority, public prestige and tremendous support (and idolization) from Congressional conservatives gave him an important political base from which to exert power. As a result, it was difficult for other coalition members to control and rein him in (James 1993: Chapter 2; Stueck 1983: 65). Second, because these military decisionmakers had formative experiences during the world wars, they tended to view the Korean conflict as a first step towards another global war. Not only did they try to duplicate the successful WWII bombing campaigns against Japan and Germany, but this worldview inclined them to treat Korea as a major test of the US’ ability to contain Communism everywhere. Most decision-makers in the Truman governing coalition viewed the Soviet Union with profound distrust, as a “wily adversary,” unpredictable, opportunistic and prone to taking risks. This outlook made it difficult for Truman and his advisors to acknowledge Soviet behavior that could be interpreted as moderate or cautious. Certainly, memories of events in the 1930s and 1940s shaped their belief in the danger of appeasing aggressors and the importance of America’s reputation for standing firm (Snyder 1991: 281; Blum 1982: 13; Foot 1985: 32–33; Sterling 1998: 349). As George Kennan later said, Truman, Acheson and other advisors “could not free themselves from the image of Hitler and his timetables. They viewed the Soviet leaders as absorbed with the pursuit of something called a ‘grand design’—a design for the early destruction of American power and for world conquest” (1972: 91–92). Leaders in the Truman governing coalition believed the Soviets had a grand design for world domination, with the United States as the primary obstacle on that path. Therefore, the United States and the Soviet Union were locked in a war of nerves, freedom fighting tyranny (Spalding 2006: 183–89). Despite his staunch anti-Communist worldview, Truman struggled to imple-

126

the korean war

ment liberal reforms in domestic policy—especially his Fair Deal agenda, which he unveiled in January 1949—and this drove his budget priorities before the Korean War. A legendary fiscal conservative, Truman was reluctant to increase military spending. He was more concerned with the national debt, which stood at $252 billion at the start of his second term, and by 1950, he came close to balancing the budget by downsizing the military. To reflect this priority, in March 1949, he replaced the hawkish James Forrestal with Louis Johnson as defense secretary, giving a big boost to the “economizers” in the governing coalition (Donovan 1984: 77; James 1993: 14; Sterling 1998: 349–50; Donovan 1982: 53–65). Like Truman but unlike the uniformed leaders in his department, Johnson was determined to hold the line on military growth and spending (Hogan 1998: 273; Tucker 1983: 201). Thus, in the year before the Korean War, Truman’s governing coalition was evenly matched with conflicting policy priorities. On the one hand, Truman, Johnson and some of the coalition’s economic advisors wanted to balance the federal budget. On the other, Acheson, his influential deputies and the uniformed Pentagon wanted to increase national security spending to fight the global Soviet threat. the broader d omest ic p olit ical context

Setting the stage for the wider domestic political environment requires some background about US foreign policy coalitions in the early Cold War period. Grand strategy in the early Cold War resulted from coalition-making that cut across the Democratic and Republican Party lines. Instead, the most important factional cleavage was between internationalist and nationalist/isolationist perspectives, complicated by the fact that different actors emphasized different global regions as most important to US strategic interests. Thus, while most internationalists focused on Europe as central in the global fight to contain Communism, some had a more global perspective. Similarly, many nationalists had an Asia-First mentality. By 1949, the Asia-Firsters often joined forces with traditional isolationists in criticizing Truman’s foreign policies, leading some scholars to group them together as “Asia-lationists” (Caridi 1968: 120; Snyder 1991: 264–73, 280–81). The Europe-first liberal internationalist faction drew from internationally oriented business and law communities and the government’s executive branch. Acheson and Marshall were key advocates of this perspective, as was Republican Senator John Foster Dulles, who eventually became Eisenhower’s secretary of state. Europe-first internationalists spanned both political parties, with Republicans hailing mostly from the East Coast. Internationalists mostly held a realpolitik perspective on American international commitments, with a tiered

the united states, part 1



approach of “core” and “peripheral” interests. Given their perspective, “core interests” resided in Europe, seen as central to containing the Soviet Union, while “peripheral” interests included much of Asia (except the major industrial centers in Japan). In the late 1940s, internationalists had favored a reduction in American tariffs, contingent on other states’ reciprocity, to boost trade with Europe and help rebuild European post-war economies (Frieden 1988; Schulzinger 1984; Isaacson and Thomas 1986; Ferguson 1983; Divine 1967: chapter 1; Snyder 1991: 264–65, 281–83). In contrast, the nationalists and isolationists resisted free trade, expensive foreign aid policies and global military commitments, considering these policies to be an expensive tax burden and excuse for big government. Before World War II, this perspective manifested as isolationism, but after World War II, it assumed an anti-UN, anti-Europe, Asia-first anticommunist unilateralist flavor, preferring stand-off air power to military ground commitments and alliances. During the Korean War, this perspective appeared as the Asia-lationists. Hailing from both sides of the aisle, well-known isolationists included Massachusetts Senator Joseph P. Kennedy and former President Herbert Hoover, who advocated building a Fortress America-style “Gibraltar in the Western Hemisphere.” But the most important isolationist was Ohio Senator Robert Taft, the leader of the Republican Party’s “Old Guard” right wing in Congress and Eisenhower’s main rival in the 1952 Republican presidential primary. More generally, this faction found support in import-competing industrial sectors, agriculture (except cotton), small business and the rural “heartland” of the country. Accordingly, many Congressional Republicans in this group hailed from Midwestern and Western states, while Democrats came from the South (Doenecke 1979; Snyder 1991: 266–67, 283–85; Rieselbach 1966: 106–20; Graebner 1956). This conservative coalition of Republicans and Southern Democrats also dominated domestic affairs in Congress. In their desire to preserve their region’s racial order, Southern Democrats aligned with Republicans on many issues—the civil rights agenda, opposition to strong national labor unions and policies related to housing, education, taxes, scientific research, economic planning and military manpower. Thus, this coalition strongly opposed Truman’s Fair Deal domestic agenda (Reinhard 1983: 65; Friedberg 2000: 43–45; Hamby 1990: 361–86, 488–508; Savage 1997: 165–67; Hartman 1971: 134–35, 155–56). The Europe-first internationalist perspective dominated early Cold War foreign policy, focusing especially on economic containment in Europe. The cornerstone policy was the 1947 Truman Doctrine, to provide economic assistance to Greece and Turkey so as to prevent them from being absorbed into the Soviet Communist camp. The Truman Doctrine became the primary building block

128

the korean war

of containment and post-war liberal internationalism, including the European Recovery Plan (Marshall Plan), tariff reductions to increase trade with European allies and the North Atlantic Treaty Organization (NATO). The Truman Administration not only exaggerated the Soviet threat to build Congressional and public support for these policies, but also used ideological, anti-Communist rhetoric suggesting the United States was the only nation capable of stopping tyranny’s encroachment on liberty and freedom everywhere (Freeland 1972; Spalding 2006: chapter 3). One important implication of this Europe-first priority was less attention to China. During the Chinese civil war, the Truman Administration began a gradual dissociation from the Chinese Nationalists. In 1949, after Chiang KaiShek and the Chinese Nationalists (KMT) fled to Taiwan, Acheson—mimicking Britain’s pragmatic perspective—pushed for US recognition of the new Chinese Communist government. Both Truman and Acheson saw the KMT’s fall as a potential solution to the political problem of recognizing China, because it would minimize accusations of abandoning an ally. Thus, from 1949 until the Korean War began, the Truman coalition asserted the KMT’s fall could not have been prevented. In August 1949, the State Department’s White Paper on China blamed Chiang’s downfall on internal failures in the KMT movement, not the inadequacy of American military and economic aid. By January 5, 1950, Truman announced the US government neither intended to acquire special concessions in Taiwan, nor would it intercede to prevent a Communist takeover from mainland China. The next week, in a speech at the National Press Club, Acheson excluded Taiwan and South Korea from America’s defense perimeter in Asia (Tucker 1983: 174–94; Reinhard 1983: 67–68; Blum 1982: 215). The decision to inflate the threat and rely on anti-Communist propaganda to sell its Europe-first internationalist agenda had longer-term implications that constrained the Truman governing coalition during the Korean War. In fact, by 1948, Truman had already been concerned about the anti-Communist rhetoric’s fever pitch and believed it needed to be tamped down (Freeland 1972: 335). Because the internationalist/isolationist cleavage crossed party lines, however, Truman could not count on party-line support for either his domestic or Europe-first internationalist agendas—even when the Democrats controlled both chambers of Congress after 1948. On most issues, he needed votes from moderate Republicans to offset the nationalist Southern Democrats. Hence, moderate Republicans, like Senator Arthur Vandenberg, had pressed Truman to justify his foreign policy in terms of an ideological struggle against Communism (Snyder 1991: 280). However, systematically exaggerating the Soviet threat to mobilize support in

the united states, part 1



Congress for authorizing containment commitments opened the door for the opposition—and ultimately constrained the Truman governing coalition—in two ways. First, it sowed the seeds for Wisconsin Senator Joseph McCarthy’s antiCommunist investigations of the US government bureaucracy, which exposed Alger Hiss and whipped up paranoia about domestic insecurity. Acheson’s public support of Hiss led to denunciations and accusations of treason.1 Second, it allowed Truman’s opponents to accept the most extreme view of the Communist threat and accuse the Truman Administration of “losing China” after Chiang KaiShek and the KMT fled to Taiwan in 1949. Having used global anti-Communist rhetoric to sell their policies towards Europe, it was almost impossible to keep that rhetoric from being directed globally. As Rosemary Foot argues, the Truman Doctrine sales job allowed the China lobby “to capitalize on the global rhetoric to promote their demands for military assistance” (1985: 44; also Hagan 2001: 32; Freeland 1972: 344–60; Reinhard 1983: 67–68; Snyder 1991: 288–90). In fact, these two constraints were related: throughout 1949, Republicans consistently charged that Truman’s China policies resulted from “subversive” influences in the State Department that were compromising US interests in Asia. During the Korean War, these attacks blossomed into calls for Acheson’s dismissal, while Democratic Senator Pat McCarran used his seat as chair of the Senate Judiciary Committee to investigate everyone connected with the China policy. These investigations gutted State of most of its China analysts during the Korean War (Hamby 1990: 565–66; Freeland 1972: 344). Both constraints—McCarthyism and conservative attacks about having “lost” China—provided fodder for opposition accusations that the Truman Administration was “soft on Communism.”

The NSC-68 Mindset: Information and Entrapment Obstacles at the Macro Level From the outset, the Truman governing coalition dealt with the Korean War within the wider Cold War context. Assessments of aggressive Soviet intentions colored the coalition’s decision-making process, while the conservative opposition’s attacks for having “lost” China significantly constrained its choices. The Korean War’s timing, three months after NSC-68, vindicated many of the policy document’s predictions and created distinct cognitive frames through which all subsequent information about the war was considered. NSC-68 also provided the impetus for a massive Cold War mobilization process, which became inextricably linked in the public’s and in Congress’ mind with the war effort. NSC-68 was a policy document used by Truman coalition members to mobilize the president’s and the public’s support for an expensive global policy of

130

the korean war

containing Communism. Behind its mobilizing prose, however, were myths, ideology and worldviews that framed the way its proponents interpreted events in the Korean War. To distinguish these beliefs from the document itself, I refer to them as the “NSC-68 mindset.” In fact, these beliefs may have had very little substance in the document itself. As the evidence in this section demonstrates, the NSC-68 mindset eventually trapped the Truman governing coalition in two interacting ways. First, it constrained how coalition members perceived information about the war and limited the range of strategic options they were willing to consider. Second, the NSC-68 mindset also affected how they pitched mobilization policies—for both the Korean War and the wider Cold War rearmament effort—to Congress and the American people. Thus, even when they avoided internalizing their own myths, they found themselves politically entrapped by their rhetoric (Snyder 1991: 41–42). nsc-68: containing the communists

NSC-68 was a systematic reappraisal of US military capabilities and a policy reassessment for how to counter the Soviet Union. Truman ordered the study after a series of international crises in 1949–50—including the September 1949 Soviet atomic detonation, the October 1949 establishment of the Communist People’s Republic of China and the German Democratic Republic, the February 1950 Sino-Soviet Treaty of Friendship, Alliance and Mutual Assistance and the beginning of McCarthyism during the Alger Hiss trial (FRUS 1950, 1: 141–42). Nitze and his team completed NSC-68 in April 1950, portraying Communism as a coordinated global movement and arguing that “the Soviet Union . . . seeks to impose its absolute authority over the rest of the world” (FRUS 1950, 1: 256). Significantly, the study predicted the Soviet Union might begin undertaking “piecemeal aggression” or “wars by proxy.” NSC-68 argued the Soviets would be able to launch a preemptive strike against the US—with a reasonable prospect of winning—by 1954, the “year of maximum danger.” As a result, it called for more foreign aid, greater military assistance to US allies, more investment in propaganda and information campaigns, increased intelligence capabilities and a massive expansion of the conventional and nuclear stockpile—with 1954 as the target date for completing the nuclear buildup. Such extensive measures were necessary, because as NSC-68 asserted, “a defeat of free institutions anywhere is a defeat everywhere” (FRUS 1950, 1: 240).2 Although Acheson took no direct part in writing NSC-68, his influence and thinking permeate the document (Rearden 1993: 166; McLellan 1976: 169). As Acheson later noted, “in dealing with the Soviet Union, the most useful negotiation was by acts rather than words. . . . From all this came insistence upon repair-

the united states, part 1



ing weaknesses and creating ‘situations of strength’ and, as a means to them, the NSC-68 program” (Acheson 1969: 378). The report not only accepted Acheson’s contention that “the Soviet Union has one purpose and that is world domination” (FRUS 1950, 1: 207), but it also reflected his belief that building “situations of strength” would require significant funding and this report needed to make that case. Moreover, Eastern internationalist elites shared Acheson’s perspective on the importance of increased foreign aid and military spending.3 Representatives of this group reviewed early drafts of NSC-68 and supported its efforts to increase the national security budget. Of the six consultants asked to review early drafts, five came from Eastern financial and academic institutions (Fordham 1998: 44–45; Nitze 1989: 98–99). NSC-68 may have intentionally simplified and overstated the threat in order to rally support for its recommendations (Snyder 1991: 271–72; Christensen 1996: 125–27). Nitze himself stated that the document was aimed at the general public and suggested parts of it be declassified and shared with the American people. Edward W. Barrett, a public relations expert working with the NSC-68 drafters, called for a “psychological ‘scare campaign’” to mobilize public opinion around the new national security imperatives (quoted in Gaddis 1982: 108; see also Hogan 1998: 300). As Acheson later admitted, “If we made our points clearer than the truth, we did not differ from most other educators and could hardly do otherwise” (1969: 374–75). The public effort by Acheson and others to mobilize the American public around higher foreign policy expenditures paid off. By May 1950, support for military assistance to Europe was at “an all-time high” (61 percent) and accompanied “by the high salience of international communism and a strongly heightened sense that America would soon be at war” (Christensen 1996: 127). But the most important sell-job was directed at Truman himself. As Acheson observed in his memoirs, “The purpose of NSC-68 was to so bludgeon the mass mind of ‘top government’ that not only could the President make a decision but that the decision could be carried out” (1969: 374). NSC-68 was “in essence, a battering ram with which its authors hoped to shatter the existing budget ceiling” (Friedberg 2000: 107–8; also Snyder 1991: 271–72). NSC-68 contained no cost estimates—an intentional omission, according to Acheson—but the suggested figure for defense alone ranged from $45 to $50 billion, which would triple the defense budget (Christensen 1996: 123; Acheson 1969: 377; Fordham 1998: 57; Gaddis 1997: 76). In effect, Acheson and the NSC-68 drafters were trying to change the nation’s fiscal ideology. Instead of making balanced budgets the priority, they were advocating a new rule: determine national security requirements first and then adjust fiscal policy to meet those demands (Pierpaoli 1999: 9). Acheson and the NSC-68

132

the korean war

drafters found common cause with Leon Keyserling, Truman’s chief economic advisor since late 1949. Pushing for an economic agenda that stimulated growth, Keyserling was willing to back NSC-68 and deficit spending in national security, because he thought budget deficits would benefit the US economy (Snyder 1991: 272; Friedberg 2000: 108–11; Pierpaoli 1999: 22). At least initially, however, Truman refused to accept the report’s premise. When Truman received it in April 1950, he passed it to the NSC, asking for further study of its assumptions and—most importantly—for cost estimates. In the meantime, NSC-68 was shelved. Fiscally conservative Truman knew its implementation would be prohibitively expensive and would thus force a retrenchment from his economic policies (Gaddis 1982: 91–93, 113; Pierpaoli 1999: 26; Wells 1979: 123–31; Christensen 1996: 126; Pollard 1985: 239). Paradoxically, the most vociferous critic of NSC-68 was Defense Secretary Louis Johnson, the economizer. In addition to philosophical differences about increasing military spending, Johnson believed he had been cut out of the NSC68 policymaking process (Etzold and Gaddis 1978: 384; Donovan 1982: 158–60; Rearden 1993: 164, 167; Acheson 1969: 373–74; Leffler 1992: 358). George Kennan and Charles Bohlen, leading Soviet experts at State, also disagreed with NSC-68’s militant line, mostly because they preferred to base prognoses of Soviet behavior on intentions rather than capabilities. Kennan especially disagreed with the report’s zero-sum worldview and the assumption the Soviets intended to pursue competition with the United States through war (Bowie and Immerman 1998: 20; Foot 1985: 39; Leffler 1992: 357–58; Kennan 1972: 92; May 1993b: 94–96). the nsc-68 mindset meets the north korean at tack

The North Korean attack vindicated many of the assumptions underlying NSC-68 (Rearden 1993: 159; Leffler 1992: 361; Weathersby 1993a). The outbreak of fighting just two months after NSC-68’s completion could not have coincided better to ensure its approval, given its prediction of greater Soviet bellicosity, including “piecemeal aggression” and “wars by proxy.” The North Korean attack confirmed these predictions and promoted NSC-68’s analysis and rationale for increased defense expenditures. As Ray Cline, then a CIA analyst who worked on NSC-68, later said, It was the Korean War that projected us into the NSC-68 world. We were not absolutely convinced as a matter of policy about the need for containment, deterrence, build-up of military forces, nation-building, alliance-building—those are the concepts that this paper stood behind. It was the Korean War that did it. (1998: 63)

From the Truman Administration’s perspective, the North Korean attack was started at Soviet behest. According to State’s intelligence office, since the “North

the united states, part 1



Korean government is completely under Kremlin control,” there was “no possibility that the North Koreans acted without prior instruction from Moscow.” This memorandum concluded: “By choosing Korea as the area of attack, the USSR was able to challenge the US specifically and test the firmness of US resistance to Communist expansion” (FRUS 1950, 7: 148–49, emphasis added). In effect, US decision-makers converted North Korean actions into Soviet actions, eliminating North Korea as an independent actor in the interaction (Milliken 2001: 50). The NSC-68 mindset had two implications for how the Truman coalition framed the attack and possible US responses. First, it led them to view the Soviet Union as seeking to “test the resolution of the US” in its defense guarantees and making “a direct challenge to the free world and US leadership thereof ” (FRUS 1950, 7: 148, 169–70). This framing suggested the Soviet Union was testing US leadership; if the United States failed to meet this “test,” the Soviets would make other aggressive moves. Thus, if the Soviets succeeded in Korea, it could not only encourage further Soviet aggression, but also damage US prestige, cast doubt on US leadership and perhaps even destroy the United Nations (Milliken 2001: 53; George 1955: 220). As Acheson put it, “To back away from this challenge, in view of our capacity for meeting it, would be highly destructive of the power and prestige of the United States” (quoted in McGlothlen 1993: 81). Likewise, the US ambassador in Moscow telegraphed that a failure to act in South Korea’s defense would lead many Asian states to undertake a “fundamental reconsideration of [their] orientation” in the Cold War (FRUS 1950, 7: 199). The second framing prompted by the NSC-68 mindset was that the Soviet Union was already fighting a war of conquest of Asia. In this view, the Soviets were already initiating a plan for regional or global military takeover, so other Soviet (or Soviet satellite) attacks could follow “at increased tempo regardless of Western interests and reactions” (quoted in Milliken 2001: 55; see also FRUS 1950, 7: 557–59). This framing projected potential harmful effects of a much greater magnitude: loss of control of Asia, weakening Western defenses of Europe and general global war. Understandably, this interpretation was more alarming to US decision-makers and compelled them to take actions to deter the wider war the Soviets were already perceived to be launching. the t ruman coalit ion ’ s resp onse to the north korean at tack

In retrospect, the Korean War served Acheson and the NSC-68 drafters quite well—after all, here was the proof of Soviet evil intentions to dominate the world. From the perspective of the schematic thinking represented by the NSC68 mindset, the massive military assault on the US client state in South Korea by

134

the korean war

the Soviet satellite state in North Korea clearly constituted a challenge the United States must answer. No matter that a few months ago South Korea had been excluded from the American defense perimeter. To some degree, as Cumings argues, “Truman and Acheson needed to interpret the North Korean attack as part of a global challenge to justify reversing their Asia policy” (1997: 31). Given the first NSC-68 mindset framing, there was a general consensus among the Truman governing coalition about the need to respond in force. Hedging against the second NSC-68 mindset framing, however, they agreed the initial response should be limited. Coalition members feared the North Korean attack was only a feint to test the willingness of the United States to confront Communist aggression, and a prelude to the main assault, expected to come in Europe (Trachtenberg 1988–89: 18–32). Thus, initially Truman’s war aims were to restore the status quo ante bellum. Acheson played a leading role in early war decision-making in June 1950: he advocated providing air and naval support to the Korean theater, increasing military aid to Indochina and the Philippines, and moving the US Seventh Fleet to the Taiwan Strait—a decision with huge consequences for Chinese decisionmaking, as we shall see in Chapter 7. Acheson was also behind Truman’s decision not to get a Congressional resolution in support of committing US troops to Korea, which influenced domestic battles later in the war. Military leaders backed Acheson’s proposals about the Taiwan Strait, because they wanted to keep the conflict localized on the Korean peninsula. They were less enthusiastic about committing overstretched US combat forces on the ground, but they were forced to concede as South Korean troops continued to fall back (Rose 1999: 191; Freeland 1972: 349; Foot 1985: 62; Tucker 1983: 196–97; Paige 1968; Thornton 2000: chapter 8; Milliken 2001: chapters 2–3). The “rally round the flag effect” was tremendous. After the intervention, Truman’s approval rating jumped 9 points from June to July, and 81 percent of Americans favored committing US troops to the war. The war instantly coalesced support around Truman who had been under attack for both his foreign and domestic policies. Republican conservatives, Asia-firsters and most isolationists all applauded Truman’s decision to go to war, noting that Truman was finally following advice in Asia that Republicans had been urging for years. Even Senator Taft endorsed the action, creating such surprise that Truman’s press secretary exclaimed, “My God! Bob Taft has joined the UN and the US” (quoted in Doenecke 1979: 191; also Reinhard 1983: 68–69; Tucker 1983: 198; Caridi 1968: 37–38). Within a few weeks, however, the Republican right wing began to undermine the united front that had been built in June. On July 6, Senator Taft began calling for attacks north of the 38th Parallel to occupy the southern part of North Korea,

the united states, part 1



or—better yet—to reunify the peninsula by force. Other Republicans criticized Truman for not asking Congress for a declaration of war. Moreover, Asia-Firsters asserted the Truman Administration’s previous Asia policies had rendered the nation ill-prepared for war. As Nevada Senator George Malone argued, “It all fits into a pattern—we deliberately lose Manchuria, China, Korea and Berlin. We follow the pattern of sometimes unrelated events—but it all adds up to losing strategic areas throughout the world” (quoted in Caridi 1968: 55). With the impending midterm elections in mind, the Republicans wanted to deny the Truman governing coalition’s willingness to aggressively handle the situation (Foot 1985: 69–70; Reinhard 1983: 69; Doenecke 1979: 192–93; Judd 1950: 54–55; Donovan 1984: 84–86; Caridi 1968: 73). In other words, as entrapment dynamics suggest, opposition leaders quickly started outbidding the Truman governing coalition with threat inflation and calls for more hawkish war policies. This external pressure was soon followed by calls for more aggressive action from within the governing coalition itself. By mid-July, shortly after being named UN forces commander, Gen. MacArthur began considering the possibility of crossing the 38th Parallel to occupy and unite all of Korea. He also began planning for the amphibious landing at Inchon, deep behind North Korean enemy lines. The Joint Chiefs agreed it made no military sense to halt at the 38th Parallel, and they advocated moving forward to a more defensible position. John Allison, the director of the Office of Northeast Asian Affairs, was decidedly hawkish, advocating as early as July 1 to advance to the Manchurian border and then hold UN-supervised elections. On July 22, he was placed in overall charge of studies on future Korean policy, where he was able to influence other coalition members, including Acheson, Nitze and Bohlen, for the northward advance. This internal and external pressure for escalation led to more hawkish war policies. By September 1, the Truman coalition agreed to authorize the use of force above the 38th Parallel as long as it could be carried out without Chinese or Soviet intervention (Collins 1969: 83; Foot 1985: 71–73). The decision to intervene in the Korean War had two wider policy effects that set the stage for later entrapment. The first relates to Taiwan. Earlier in the spring, before the North Korean invasion in June, the Republican right wing had opposed an economic aid package for South Korea. To get the measure passed, the Truman governing coalition agreed to placate the Congressional China lobby and extend an earlier aid appropriation to Chiang Kai-Shek through June 30. This Administration-sponsored decision implied a connection between South Korea and Taiwan, so that “when war broke out in Korea, protection of Taiwan appeared compulsory” (Tucker 1983: 197; also Snyder 1991: 292; Freeland 1972: 342). While Truman and Acheson supported using the Seventh Fleet to neutral-

136

the korean war

ize the Taiwan Strait, they never intended this emergency measure to become permanent (Tucker 1983: 200; Foot 1985: 64). Yet, after the war broke out, parts of the governing coalition and a variety of domestic constituencies wanted to link Taiwan policy more tightly to the Korean War. Within the coalition, Defense Secretary Johnson, the JCS and Gen. MacArthur all advocated coastal bombing to keep the Chinese Communists from invading Taiwan, while Rusk at State collaborated with Dulles to create a “two Chinas” policy. Most significantly, in July MacArthur took an unauthorized trip to meet with Chiang in Taiwan, during which they signed a joint communiqué about common security concerns in Asia. Chiang also offered to contribute troops to the UN forces fighting in Korea. In August, MacArthur wrote a public letter— again without clearance from the JCS or any civilian leaders—to the Veterans of Foreign Wars, arguing again for the strategic necessity of preserving Taiwan. (Truman ordered the letter to be officially withdrawn.) More broadly, the China lobby and Asia-lationists called for increased ties with Chiang as an indispensable ally in this global war against Communism. They also pointed to MacArthur’s unilateral efforts to assert the Truman Administration was not doing all it could to fight Communists in Asia (Schnabel 1972: 68; Doenecke 1979: 194; Tucker 1983: 198–201; Freeland 1972: 351–52; Blum 1982: 215; Foot 1985: 65–66). As a result of these pressures from within and outside the coalition, Truman and Acheson had to abandon their goal of recognizing the PRC diplomatically. As Nancy Tucker argues, “If the American people must resist communist expansion and fear communist subversion, they could not be expected to condone policies flexible enough to initiate relations with a communist giant like the People’s Republic of China” (1983: 198). Moreover, by August, Truman felt compelled to approve a $14 million military assistance package to strengthen KMT defenses. Thus, because of its globalist agenda, the Truman coalition became increasingly involved with preserving Chiang’s regime (Tucker 1983: 201; Freeland 1972: 349– 50; Caridi 1968: 11). The second wider effect of the Korean intervention was that Truman reluctantly authorized the Cold War rearmament and mobilization effort that Acheson and the NSC-68 authors had been advocating. On July 19, Truman sent Congress his first message about the war, requesting $11.6 billion in supplemental military funding for the rest of the fiscal year, signaling his determination to strengthen the NATO alliance, and requesting authority to ensure adequate supplies for military and civilian use were available. This request, which became the Defense Production Act of 1950, was approved—although Republicans resented the greatly increased executive control over the economy. Congress also authorized a substantial across-the-board tax increase for an additional $5 billion in

the united states, part 1



annual revenues. Moreover, Congress granted additional standby control powers over prices and wages, which Truman later invoked in January 1951 (Donovan 1984: 78; Pierpaoli 1999: 33–35; Friedberg 2000: 116, 120; Caridi 1968: 65–68). By mid-September 1950, Truman pushed out the last of the economizers, Defense Secretary Johnson, and replaced him with George C. Marshall. As the domestic coalition theory would predict, this hawkish coalition shift early in the war helped the Truman government mobilize to wage the war more effectively. Three other events about this time also signaled the hawkish strategy shift. First, Truman formally approved NSC-68. Second, on September 11, Truman approved NSC 81/1, a new strategy which recommended reunifying Korea if it could be accomplished without risk of war with China or the Soviet Union—in effect increasing US war aims. Finally, on September 15, MacArthur successfully conducted the amphibious landing at Inchon, turning the tide of the war (FRUS 1950, 7: 712–21; Stueck 1995: 63; Pierpaoli 1999: 30; Hamby 1990: 547; Pollard 1985: 240; Leffler 1992: 377; Schnabel 1972: 200; Truman 1956: 359). With this hawkish coalition shift, Truman accomplished two things. First, he removed the last veto player who opposed mobilization. Second, Truman hoped Marshall’s presence as a non-political elder statesman would defuse partisan attacks on his foreign policies. Unfortunately, Republicans used Marshall’s nomination hearings to air long-held grievances about Truman’s Asia policy, and over the next year until his resignation, Marshall was subjected to the same McCarthyist denunciations as Acheson. Indeed, for the upcoming midterm elections, the Republicans had settled on a campaign message of “Liberty against Socialism,” focusing on domestic security and anti-Communism in Asia. In September, Congress overwhelmingly passed the Internal Security Act of 1950, clearly a ploy to exploit the popularity of McCarthy’s anti-Communist paranoia. The strategy paid off, as Republicans gained 28 seats in the House and five in the Senate, leaving the Democrats with narrow majorities in both chambers (Freeland 1972: 352–53; Reinhard 1983: 65–67; Caridi 1968: 64, 90–103). the nsc-68 mindset meets chinese intervent ion

The United States adopted NSC-68 as the UN Command was improving its battlefield fortune. After MacArthur’s successful amphibious landing, Asialationist calls for Korean reunification intensified, while the press also supported an expanded war policy. Moreover, a Gallup poll in early October showed 64 percent of Americans would view halting at the 38th Parallel as appeasing the Communists (Whelan 1990: 223; Sterling 1998: 369; Foot 1985: 70; Caridi 1968: 73). Crossing the 38th Parallel raised the possibility China could intervene, but the temptation to exploit the great victory at Inchon, plus the desire to accomplish

138

the korean war

the long-standing UN political goal of reunifying Korea, caused policymakers to underestimate that possibility (Caridi 1968: 80; George 1980: 71; Neustadt 1990: 114–22). Thus, at the United States’ urging, on October 7 the UN General Assembly passed a resolution for the UN Command to reunify Korea militarily, as UN forces crossed north of the 38th Parallel. Because of the NSC-68 mindset, the Truman coalition initially thought the North Korean attack might be a feint to lure US forces away from a more vital area like Europe, where a follow-on Soviet attack was planned. Now, months later, there were no indications of a Soviet attack, so some officials argued direct Soviet entry into the hostilities was unlikely. Moreover, NSC 81/1’s analysis predicted some Soviet reaction—use of military force, attempts to reach settlement, or a hands-off policy—by the time US forces crossed the 38th Parallel (Milliken 2001: 123; Foot 1985: 74–75). Because no Soviet action occurred, coalition members assumed the Soviet Union had adopted a hands-off policy, as NSC 81/1 had suggested it might. Thus, as Neustadt argues, “With the military opportunity before them and with diplomatic dangers out of sight, the men [Truman] leaned on for advice saw little risk of any sort” (1990: 114–15). In this way, the NSC-68 mind-set predisposed them to interpret information according to their pre-existing beliefs, as the information obstacle would suggest. Thus, Truman’s coalition went along with MacArthur’s plans to continue north, even as some began to worry about Chinese intervention. One major problem is that, given the NSC-68 mindset, coalition members did not seriously separate Chinese intentions from those of the Soviets, despite Chinese warnings and available intelligence information to the contrary. As Chapter 7 will explain, starting on September 26, China provided at least three warnings about US intervention north of the 38th Parallel, in diplomatic meetings, public speeches and domestic newspapers. Moreover, US intelligence reports indicated Chinese troops were massing in Manchuria near the Yalu River. Cables from American embassies abroad, including London and Moscow, also reported an impending Chinese intervention (Sartori 2005: chapter 2; Heichal 1984: 94, 97; FRUS 1950, 7: 1061–96; Cohen and Gooch 1990: 169; Stueck 1983: 219, 224). Despite these clues, there were several reasons US decision-makers misperceived Chinese intentions to intervene. First, CIA analysts believed China had limited interests in the region. Moreover, many analysts agreed with MacArthur’s intelligence officer Gen. Willoughby that the auspicious time for intervention had passed. Thus, China was unlikely to intervene or at most would try to protect its hydroelectric complexes along the Yalu River and industry further north. The JCS concurred with the CIA’s conclusions and ordered MacArthur to refrain from bombing within five miles of the Manchurian border (FRUS 1950, 7: 117–21, 915, 933–34, 1025–26, 1101; Stueck 1995: 106, 111; Schnabel and Watson 1979: 262).

the united states, part 1



Edmund Clubb, head of State’s Office of Chinese Affairs, dissented from this consensus; he thought China might intervene because of broader ideological concerns (Whiting 1991: 106–8; Foot 1985: 80). Second, US decision-makers assumed the Chinese military was not very effective. Extending their assumption about a Soviet hands-off policy in Korea, they assumed Stalin would provide no Soviet air or naval support—without which they believed the Chinese could not decisively intervene (FRUS 1950, 7: 933–34; Schnabel 1972: 201; Stueck 1983: 219). As Acheson asserted, “it would be sheer madness for the Chinese to enter the Korean War” (quoted in Zhang 1992: 84). More importantly, while intelligence analysts had a fairly complete picture of the organization and equipment in a Chinese division, they explicitly compared the Chinese to the North Koreans and noted their lack of artillery, heavy weapons and transport. As a result, they came to view the Chinese forces as a second-rate North Korean army, instead of a different kind of enemy altogether (Cohen and Gooch 1990: 176–77; Zhang 1992: 102). Third, and perhaps most importantly, MacArthur considered Chinese intervention remote, as he told Truman during their October 15 meeting at Wake Island. As Truman recounts, MacArthur told him flatly “that if any Chinese were to enter Korea they would face certain disaster but he did not expect them to try anything that foolish” (1956: 373; also Leffler 1992: 378; FRUS 1950, 7: 948–60). Truman appears to have tied himself to MacArthur’s interpretation of events. As he remarked, “you pick your man, you’ve got to back him up” (quoted in Neustadt 1990: 122). After the Inchon landing, MacArthur’s influence was at its peak. Indeed, because the JCS had doubted the Inchon operation right up to its execution, they were even less inclined after its brilliant success to question MacArthur’s judgment than before (Collins 1969: 141). Thus, given the predominant view of China’s narrow interests and ineffective capabilities, tactical intelligence was forced to fit MacArthur’s belief that he could defeat the Chinese. MacArthur did not want the Chinese to enter the war, so his intelligence officer made it his job to underestimate the Chinese presence and convince everyone they would not intervene. Willoughby “was accustomed to telling MacArthur what the general wanted to hear, and he sympathized with his boss’s determination to march on to complete victory in Korea” (Stueck 1995: 112). For example, Willoughby may have knowingly falsified intelligence reports, setting an upper estimate of 60,000 Chinese soldiers in Korea when in fact more than 200,000 were already there. Because CIA’s intelligence work in Korea occurred in cooperation with Willoughby, “all the information from China was one and the same” (Heichal 1984: 94; also Marshall 1953: 213; Spanier 1959: 105; Edwards 2000: 81; Sterling 1998: 379; Andrew 1995: 189–90). In addition to these information obstacle dynamics, domestic political entrap-

140

the korean war

ment was starting to have an effect as well. Polls at the time revealed proponents for escalation had increased to more than 30 percent, three times the number who wanted to pull out (Sterling 1998: 389). Similarly, press coverage of the war remained glowing and editorials advocated uncompromising anti-Communist positions; almost every major newspaper had endorsed the decision to cross the 38th Parallel (Foot 1990: 70). MacArthur was promising the war would be over by Christmas; any alteration to existing policy—including a pull-back to the “waist” of the peninsula advocated by the Brits—would not permit the war’s quick end and would have a “catastrophic effect” on public opinion (Foot 1990: 27–28; Sterling 1998: 388; James 1993: 199–200). The result was a classic case of leaders becoming “too invested to quit” and thus interpreting information so as to justify a course of action despite evidence against it. The Truman governing coalition chose not to know that its northward advance was about to turn sour. Several scholars describe the Truman coalition at this juncture as desperate to reestablish its prestige in the Far East and to rectify its domestic reputation as “soft on Communism”—making it insensitive to intelligence suggesting Korean unification was unrealistic. Despite China’s repeated warnings, the Truman coalition was inclined to dismiss them as bluff. In short, “the administration’s wishful thinking encouraged grossly defective information processing” (George 1980: 75; also Huo 1989: 192; Lichterman 1963: 596; Tsou 1963: 567–69; de Weerd 1962: 451–52; Rees 1964: 110). As Alexander George and Richard Smoke argue, A reversal of policy was so unpalatable to the administration that it was unwilling to face up to the necessity for modifying its war objective before it was too late. . . . it reduced US receptivity to indicators of the growing threat of Chinese intervention. Washington’s reluctance to accept information that challenged the premises and wisdom of its policy strongly encouraged its tendencies to misread the frame of reference with which Peking perceived events in Korea, to misinterpret available information and to underestimate Peking’s motivation and willingness to take risks. (1974: 190–92, emphasis added)

In sum, facing a military opportunity and the chance for domestic political benefits, the Truman coalition assumed diplomatic dangers away, avoiding available information that the UN’s advance to the Yalu River could have nasty consequences. The NSC-68 mindset simply did not allow for the possibility that China and the Soviet Union might have different interests and goals in the war. Therefore, once convinced the Soviets would not intervene, the Truman coalition assumed it was free to unify the peninsula militarily. By publicly committing itself to the goal of Korean unification, it also set the stage for further entrapment by hawkish domestic constituencies. As Richard Barnet bluntly argues, it was “one

the united states, part 1



of the clearest cases where fear of a domestic political backlash drove a critical strategic decision” (1990: 313). Thus, the Truman governing coalition chose not to know that it needed to stop its offensive and start taking actions to end the war. the nsc-68 mindset meets “ an ent irely new war ”

Given such misperception, it is not surprising that the Chinese intervention was a massive strategic surprise. Upon comprehending the scale of the attack, MacArthur famously cabled Washington that “we face an entirely new war” (FRUS 1950, 7: 1237). UN troops suffered 11,000 casualties in the first two days and were in full-fledged retreat (Foot 1985: 101; Sterling 1998: 397). After having written off the potential for Soviet intervention entirely, now the pendulum swung abruptly the other way: the Truman coalition immediately assumed the Soviets were behind the Chinese counterattack (Donaldson 1996: 61). The Chinese intervention became yet another vindication for the assumptions and analysis of NSC-68; in this view, the Soviet Union so clearly controlled China that Rusk decried China’s sacrifice for its “foreign masters” as the “ambitions of the Communist conspiracy” (quoted in West 1993: 22). Because the NSC-68 mindset could not conceive that China and the Soviet Union might have different interests and objectives in the war, the Chinese intervention had to be an indication of a new phase in the Soviet-directed campaign. For example, State’s Office of Chinese Affairs now asserted China “was definitely on the Moscow line and the possibility of reorienting Peking away from Moscow was ‘negligible’” (quoted in Foot 1985: 102). His earlier prediction of Chinese intervention confirmed, Clubb now emphasized Chinese actions as part of a Soviet strategic design. China’s intervention in Korea was “largely meaningless,” he argued, unless it was “regarded as a component of a global Communist plan” (FRUS 1950, 7: 1292). Returning to the NSC-68 mindset’s second frame, CIA Director Walter Bedell Smith argued the Soviets’ presumed backing of Chinese intervention indicated a greater Soviet readiness to initiate a third world war, especially “if they can bog us down in Asia” (quoted in Merrill 1997: 107). Or, in Acheson’s words, “this aggression was ordered by the Kremlin” (quoted in Chang 1990: 76). In such an environment, the Truman coalition decided the best way to show US steadfastness was to “hold the line” in Korea. If the Soviet Union had a global master plan and could begin a global war when it wanted, then the United States had no alternative but to stand firm and protect its allies. Gen. Charles L. Bolte, the Army’s operations chief (G3) and the only major official in Washington to visit Korea since the Chinese intervention, argued firmness would deter further “aggression,” whereas timidity or withdrawal would produce the opposite result

142

the korean war

(Stueck 1995: 118). Rusk argued that abandoning South Korea would set a poor “example of what it means to be a friend of the US” (quoted in Foot 1985: 109). Acheson agreed if the US left Korea, “all Asians would hurry to make the best deals they could” with the Soviet Union and Communist China (FRUS 1950, 7: 1383). Moreover, Marshall and the JCS emphasized that US prestige would suffer if Korea had to be abandoned (FRUS 1950, 7: 1293–94, 1311, 1324, 1335–36, 1345– 46). If “holding the line” was given, there was considerable debate within the governing coalition over the next two months about whether to widen the war further into China. The most extreme proponent of this course of action was MacArthur, who argued that the US should either evacuate completely from Korea or widen the war into China. He proposed moving north of the 38th Parallel again, blockading the Chinese coast, bombing China’s industrial centers, supply bases and communications network, accepting Chiang Kai-Shek’s offer to send 30,000 troops from Taiwan and backing a KMT counterattack onto the Chinese mainland. In other words, MacArthur rejected the premise that the United States was still waging a limited war and suggested using every means at its disposal to win. MacArthur’s views found vocal support from the Republican right wing and Asia-lationists in Congress, whose pressures for escalation had only intensified since China’s intervention (MacArthur 1964: 379; Freeland 1972: 356; Caridi 1968: 116). While not as extreme as MacArthur, the three service secretaries and the JCS all supported retaliation against China to some degree (Merrill 1997: 379–81). Even Marshall agreed there was “heavy popular pressure for the United States to do something” and the United States could not afford to wait for the United Nations (FRUS 1951, 7: 93–94). Despite that consensus, Acheson proved a very powerful veto player during these debates and swung Truman to his position. Invoking the NSC-68 mindset’s second frame, Acheson believed a wider war with China would provoke an armed response from the Soviet Union, probably on a global scale. Moreover, he blocked wider dissemination of subordinates’ views that diverged from his interpretation that Chinese intervention was part of a Soviet master plan. These officials—Kennan, Rusk, Phillip Jessup and Walworth Barbour at the US embassy in Moscow— argued the Soviets might gain more from non-involvement and noted there was little evidence of Soviet intention to intervene (FRUS 1950, 7: 1603–4; Foot 1985: 125–26). Acheson suppressed these views, because they conflicted with his Europe-first preferences: fighting a wider war with China would not only diminish US ability to fight the Soviets, but also divert US resources from Europe (Leffler 1992: 383–90). Instead, Acheson argued a general lack of international confidence

the united states, part 1



in US leaders was already slowing European defense plans and promoting neutralist sentiments there. As he asserted, “We are fighting the wrong nation. . . . We are fighting the second team, whereas the real enemy is the Soviet Union” (FRUS 1950, 7: 1326). Acheson’s views were also supported by Nitze, Clubb and the CIA (Foot 1985: 117–23). Acheson almost single-handedly blocked any real discussion of retaliation against China, with his most important obstruction occurring at a NSC meeting on January 17, 1951. As a result, Truman delayed a decision on recommendations from the JCS and MacArthur and returned them to the NSC staff for further consideration—effectively stopping all momentum that had been building in the Truman coalition for retaliation against China. Instead, records from the meeting highlight how much Truman adopted Acheson’s perspective. The minutes noted “any actions undertaken by the US should be in furtherance of this [NSC68] concept.” Moreover, “because of the present military unpreparedness, action which might precipitate global war should be avoided. Further, those actions which would alienate our supporters in our over-all effort against the USSR should also be avoided” (Merrill 1997: 392–96, emphasis added). In sum, the NSC-68 mindset melded with Acheson’s Europe-first preferences to effectively stop all momentum within the governing coalition for escalating the war against China. Influenced again by the two frames of the NSC-68 mindset, the governing coalition believed, on the one hand, that it had to “hold the line” in Korea. On the other hand, powerful coalition members also believed any major increase in military pressure could turn the Korean conflict into global war; thus, it needed to focus on rearming and preparing for the war that might be coming in Europe. In other words, after the Chinese intervention, the NSC-68 mindset created twin constraints on the Truman coalition’s strategic choices— holding the line in Korea while avoiding global war—that remained in place for the rest of his administration. Thus the NSC-68 mindset served as a filter on the information that guided policy formulation. In turn, the policies interacted with political dynamics to entrap Truman. the coalit ion ’ s resp onse to the “ ent irely new war ”

Two weeks after China’s intervention, the National Security Council approved an amendment to NSC-68 to accelerate US and NATO rearmament efforts: the expected “year of maximum danger” was moved forward to 1952, and thus rearmament efforts to meet that threat needed to be completed by that date. The next day, Truman declared a State of National Emergency—something that did not happen in either World War II or the Vietnam War—and announced he was

144

the korean war

beginning “total mobilization” and establishing a new rearmament agency, the Office of Defense Mobilization (ODM). This move had the psychological and political effect of demonstrating that Truman was doing something about the situation in Korea. By publicly announcing the stepped-up rearmament efforts right after the devastating Chinese counterattack two weeks before, Truman indelibly linked the Korean War and the NSC-68 rearmament efforts in the minds of Congress and the public (Donovan 1984: 132; Friedberg 2000: 120–22; Pollard 1985: 241; Merrill 1997: 287–94; Pierpaoli 1999: 46, 117). The mobilization effort was impressive, but quite interventionist. Truman’s proposed budget for 1951, unveiled in January, revealed that nearly 69 percent of $72 billion in federal expenditures was earmarked for national security. Moreover, he proposed increasing the annual rate of defense spending from $20 billion to $50 billion by the end of 1951. To manage this massive rearmament effort, between December 15 and February 1, the Truman Administration established 19 new government agencies. The rapid mobilization and rearmament schedules badly distorted the economy, sparking a firestorm of inflation that ultimately forced Truman to issue mandatory wage and price controls in January 1951. Moreover, it eventually unbalanced the federal budget, led to across-the-board tax increases, raw material shortages and economic dislocations and encouraged the growth of the military industrial complex. By the spring, relations between Truman and Congress grew strained as Congress tried to limit his emergency and mobilization powers and stall tax increases. All of this, in turn, created acrimonious debates between liberals and conservatives over how best to distribute the nation’s resources, with the public and Congress both fretting openly about the long-term consequences of mobilization (Hogan 1998: 315–65; Pierpaoli 1999: 43–55, 65–68, 71, 83; Friedberg 2000: 117–22; Hamby 1990: 577). Opposition to mobilization received voice through the “Great Debate” in Congress (December 1950 through April 1951)—fueled by Asia-lationist critique of Truman’s December 19 decision to deploy troops to NATO without Congressional consent. Republicans were outraged by the choice to beef up security in Europe “while American boys were on the run in Korea” (Reinhard 1983: 71; also Caridi 1968: 126; Donovan 1984: 146). Fiscal conservatives joined in, arguing Truman’s commitments to NATO combined with the Korean War could bankrupt the nation. The debate was a last-ditch effort by Congressional conservatives to reverse the decision to implement the military buildup proposed in NSC-68, and the Korean War only stoked their latent hostility to a Europe-first containment approach (Pierpaoli 1999: 57–58). To be sure, the Republicans were being contrarian, at the same time advocating for an all-out escalation against China and a complete withdrawal from

the united states, part 1



Korea. The root of this inconsistency was their belief that, as Senator William Jenner put it, Truman “blundered, tricked and betrayed us into a war” (Reinhard 1983: 69; also Caridi 1968: 114–16). For example, in January 1951 Taft argued that stationing four US Army divisions in Europe would goad the Soviet Union into war; three months later, during the MacArthur hearings, he and many followers believed that the Soviets would tolerate US bombing near its Asian borders in an expanded war (Doenecke 1979: 205). These inconsistencies in the Republican right wing and Asia-lationist opposition to Truman’s war policies—simultaneously calling for American withdrawal and passionately associating themselves with MacArthur’s “no substitute for victory” philosophy—continued through the 1952 presidential elections (Caridi 1968: 175–82). Despite these partisan attacks, the Truman governing coalition held to its policy. The need to strengthen the NATO alliance and thus prepare for the “real” war ahead compelled it to focus on containing the Korean War. Holding the line in Korea not only demonstrated resolve to the Soviets, but as Truman and Acheson explained to British Prime Minister Clement Atlee, “any public compromise or conciliation with Beijing would destroy the momentum” of the rearmament and mobilization under NSC-68 and thus undercut American assistance for its European allies (Christensen 1996: 185). In other words, preparing for a feared potential Soviet attack took precedence over responding to the actual Chinese intervention that was pushing the UN forces back south across the 38th Parallel. In the Truman coalition’s view, a limited war was the only way to juggle the Korean challenge with the risk of global war with the Soviets. As Army Chief Gen. Collins summarized, “Since the United States is not now prepared to engage in global war, and will not be ready before July 1, 1952, we should take all honorable means to avoid any action that is likely to bring Russia into open conflict with the United States prior to that date” (quoted in Trachtenberg 1988–89: 26). Therefore, in March 1951, the State Department distributed a draft text on UN policies in Korea, including a statement on its willingness to settle. State sent MacArthur a copy of the statement before releasing it to the press; he preemptively issued a cocky statement of his own, telling the Chinese to surrender or be destroyed. This insubordination was the direct cause for MacArthur’s dismissal on April 11 (Truman 1956: 467–69; Caridi 1968: 143–45). However, MacArthur was no longer supporting the governing coalition’s priority to prepare for the general war thought to be coming in Europe—and thus was out of step with its policies more generally. As the coalition member with preferences furthest away from the majority’s, changing war policy would be easier without his constraints on the coalition.4 As his successor at the UN Command, Gen. Matthew Ridgway, later argued,

146

the korean war When MacArthur spoke of victory he did not mean merely victory in Korea. . . . What he envisaged was no less than the global defeat of Communism. . . . His plan therefore entailed the very considerable risk of igniting World War III and the consequent overrunning of Western Europe, with the loss of our oldest and staunchest allies sure to follow. It should be remembered too that we would have been running those risks when our own nation was lamentably unprepared, with our general reserve reduced to a single combat-ready army division. (Ridgway 1967: 145, emphasis added)

By the time the MacArthur hearings before Congress started in May 1951, the Truman coalition had passed NSC 48/5, a major review of its Asia policy, which returned to Truman’s initial status quo ante bellum war aims and advocated Korean reunification by political rather than military means. War strategy was modified to confine and terminate the conflict: the United States should “seek to avoid the extension of hostilities in Korea into a general war with the Soviet Union” (FRUS 1951, 7: 439–42). This emphasis on the risk of precipitating general war, if the United States tried anything other than the current restricted course of action, was also a major plank in the Truman Administration’s arguments at the MacArthur hearings. The NSC-68 mindset echoed throughout this testimony. Administration spokesmen argued that (1) limiting the scope of fighting in Korea would give the United States time to build up its own strength and (2) MacArthur’s proposed war escalation not only could not guarantee the war’s end but would also likely draw the Soviets directly into the war, while simultaneously estranging the United States from its allies (Rees 1964: Chapter 15; Foot 1985: 136). It was an “impressive show of solidarity”—all the governing coalition’s former proponents of a wider war with China toed Truman’s line (Acheson 1968: 529). The public remained more hawkish than the Truman coalition, in part reflecting the political fall-out from MacArthur’s removal. The general had returned to the United States as a great national hero and presented a dramatic address to Congress: “The ovation was akin to an eruption; the Congressional eulogies that followed were as overheated as lava from Vesuvius” (Hamby 1990: 562). Threequarters of the American public favored continuing operations, compared to only 14 percent advocating the United States “pull out now.” US casualties now exceeded 60,000, and MacArthur’s contention that limited war would only bring stalemate resonated with a war-weary public. As a result, roughly 45 percent also favored escalatory measures, such as bombing Manchuria and assisting Chiang Kai-shek’s Nationalists in returning to mainland China. Fully 60 percent wanted to give Chiang’s forces all the help they needed to attack Chinese Communists, and one-third even favored launching a general war against China (Schwarz 1994: 10, 14; Stueck 1995: 182–84; Foot 1985: 156; 1990: 139; Sterling 1998: 431). Paired with

the united states, part 1



the calls for escalation from Congressional Asia-Firsters, the domestic political environment could have supported the Truman governing coalition taking more escalatory measures against China, but it no longer actively considered them. The Truman governing coalition tried to respond to these hawkish domestic constituencies, combating the partisan attacks for being “soft on Communism” in Asia while remaining true to its more European-centric NSC-68 rearmament effort. For example, in March 1951, Truman agreed with a JCS recommendation to create a military assistance advisory group to aid Chiang in Taiwan. Later, he approved $300 million in military aid to the KMT and announced that Taiwan was sharing equal priority with Western Europe for arms shipments. Finally, on May 18, Rusk announced diplomatic recognition of Taiwan, asserting the United States “believe[s] it more authentically represents the views of the great body of the people of China, particularly the historic demand for independence from foreign control” (quoted in Caridi 1968: 121). These policy choices—a far cry from the previous spring’s plan to cut ties with the KMT and recognize the PRC—reflect just how trapped the Truman governing coalition had become. Unfortunately, the complete lack of understanding about Chinese interests or capabilities that the Truman coalition exhibited in its race to the Yalu River continued to haunt the United States for the next two years. Before the Chinese intervention, most US analysts had scoffed at the Chinese forces as a “second rate North Korean Army” rather than a force with its own capabilities, strengths and weaknesses. Now, after facing “an entirely new war,” the Truman coalition assumed the Soviets were behind the Chinese counterattack, if only to tie down US forces while preparing to launch the “real” war in Europe. Given their monolithic understanding of Communism, US leaders focused more intensely on Moscow’s motivations for escalating the war indirectly than on possible Chinese motivations as a separate actor. Trapped by the two frames of the NSC-68 mindset, the Truman governing coalition believed, on the one hand, that it had to “hold the line” in Korea. On the other hand, it believed that any major increase in military pressure could turn the Korean conflict into global war; thus, it needed to focus on rearming and preparing for the war that might be coming in Europe. These rearmament efforts became indelibly linked in Congress’ and the public’s mind with the Korean War, which is not surprising, since the NSC-68 mindset drove the logic for both. As a result—and quite ironically—the Truman coalition stopped pressing militarily just at the point when the Chinese offensive campaigns in Korea were stalling out and the Chinese military forces were at their weakest.

148

the korean war

Conclusions This chapter has examined the interacting dynamics of information and entrapment obstacles in US decisions about prosecuting and ending the Korean War. While Truman had emphasized the economic aspects of containment before June 1950, the Korean War and the resulting mobilization instigated the militarization of Truman’s containment policy. It also led to the reversal of many aspects of the Truman Administration’s pre-war Asia policy, in order to placate domestic constituencies to ensure support for its Europe-focused mobilization. Having used global anti-Communist rhetoric to sell its policies towards Europe since 1947, it was almost impossible to keep that rhetoric from being directed towards containing Communism in Asia during the Korean War. As Richard Freeland argues, “In the short period between the Truman Doctrine speech and early 1951 . . . the Administration had been led, against its will and policy, but according to the logic of its rhetoric and politics, to a full application of the Truman Doctrine in Asia” (1972: 358). At the macro level, the “stickiness” of the Truman coalition’s war policies resulted from its NSC-68 mindset and the policies pursued in response. Although the NSC-68 mindset predated the war, the North Korean attack vindicated its assumptions and created two cognitive frames through which all subsequent information about the war was interpreted. On the one hand, the NSC-68 mindset led Truman’s coalition to view the Soviet Union as seeking to test US resolve. Thus, it was unable to disengage completely from Korea because it believed the war was a major test of its ability to contain Communism everywhere. On the other hand, the NSC-68 mindset led Truman’s coalition to view the Soviets as already pursuing global domination, creating a belief that any military escalation could turn the Korean War into World War III. Thus, the NSC-68 mindset not only blinded the governing coalition to strategic alternatives that may have helped it end the war sooner, but it also drove a domestic mobilization effort that ultimately became very unpopular, leading to accusations of the “imperial presidency” (Schlesinger 1973; Donovan 1984: 167; Pierpaoli 1999: 160–94; Donovan 1982: 382–91). This chapter has explained the chronological story of US decision-making in the Korean War through the American decision to seek a negotiated settlement in the spring of 1951. Within the broader war policy narrative, it has highlighted many domestic challenges the beleaguered Truman faced—including the political irritants of mobilization, McCarthy’s assaults and the Republican right wing’s politically expedient strategy of simultaneously calling for all-out war against China and immediate withdrawal from Korea. All of these challenges contin-

the united states, part 1



ued throughout Truman’s tenure. For example, the mobilization’s wage controls eventually led to labor strikes and in March 1952, Truman even seized the nation’s steel mills. McCarthyist attacks eventually led Secretary of Defense Marshall to resign in September 1951, after being investigated for his role in “losing China” during his tenure as secretary of state. The Republicans honed their oppositional strategy into a foreign policy platform for the 1952 presidential election. Thus, it is not surprising that Alonso Hamby, Truman’s biographer, writes, “Few presidents have been as harried as Harry Truman during his last two and half years in office” (1990: 583). Chapter 6 continues this narrative, starting in July 1951 with the armistice negotiations. While this chapter has focused on the interacting dynamics of the information and entrapment obstacles at the macro (strategic) level, Chapter 6 delves into them at the micro (policy) level with the US decision to adopt the POW voluntary repatriation policy.

6

The United States, Part 2 Trapped by Voluntary Prisoner Repatriation

The previous chapter laid out the interacting dynamics of the entrapment and information obstacles in US decision-making in the Korean War at the macro level, using the framework of the “NSC-68 mindset.” Vindicating its fears about an imminent Soviet drive for world domination, the North Korean attack prompted the Truman coalition to embark on a massive Cold War mobilization effort while fighting a limited war in Korea. However, building political support for this strategy required winning over moderate Republicans and Asia-lationists through expanded war aims—increasing military aid to Taiwan, shelving plans to recognize China diplomatically and launching a ground offensive to reunify the Korean peninsula—all policies which directly drew China into the war. Subsequently, facing “an entirely new war,” the Truman coalition was trapped in two ways. First, it was trapped by its beliefs—embodied in the NSC-68 mindset—that holding the ground in Korea was critical while rearming to prepare for the “year of maximum danger.” These beliefs not only blinded it to differing interests and objectives among its Communist adversaries, but eclipsed strategic alternatives that may have helped it end the war sooner. More importantly, the Truman coalition was trapped by domestic constituencies. Because it believed the only course was conducting a limited war in Korea while implementing a massive mobilization at home, the Truman coalition was caught in a dilemma. On the one hand, the Korean War helped it secure necessary Congressional and public support for its massive Cold War mobilization plan. On the other hand, the threat inflation that resulted from selling mobilization to the American people stoked hawkish constituencies for all-out victory in Korea. In the hey-day of

the united states, part 2



McCarthyism, any efforts to move towards a negotiated settlement renewed calls of the Truman governing coalition being “soft on Communism.” Nestled within this broader entrapment was the US decision to adopt a policy of voluntary repatriation of prisoners of war (POWs), which exemplified information and entrapment dynamics at the micro (policy) level. The Truman Administration adopted the POW policy prematurely, without considering the effects such a position would create. Subsequently, the Truman coalition became “too invested to quit” and thus it systematically ignored unpalatable information suggesting this policy would conflict with other, more important, policy preferences. By the time it became aware of the policy’s full implications, it was too late to back down without signaling “weakness” to the Communists and incompetence to the domestic audience. Beleaguered by partisan attacks and domestic political crises—many resulting from the Cold War mobilization effort—Truman could neither abandon the POW policy nor escalate the war to bring the Communists to terms. It was only with a domestic governing coalition shift—the election of President Eisenhower—that entrapment and information obstacles at both levels were lifted and the United States could change course towards ending the war. While Eisenhower retained the same war aims that Truman had espoused, his governing coalition brought a different worldview—and a different conceptualization of the war—that left it freer to consider an expanded palate of strategic options. Notably, as a Republican, Eisenhower could more easily accept terms Truman could not while placating the hawkish Republican right wing in other issue areas. This chapter completes the analysis of US decision-making in the Korean War, in three sections and a conclusion. The first section examines the micro-level dynamics of the information and entrapment obstacles in the US decision to adopt the voluntary POW repatriation policy. The second section returns to the macro-level perspective to explain how the new Eisenhower coalition was able to overcome the obstacles to peace that hindered the Truman Administration to end the war. Finally, the third section examines some alternative explanations for US behavior during the Korean War.

Voluntary Repatriation: Information and Entrapment Obstacles at the Micro Level The Korean War armistice negotiations began in July 1951, and after a short recess that autumn, resumed and quickly resolved many issues. The talks were so productive that by February 8, 1952, Acheson presented Truman with a memo-

152

the korean war

randum that opened: “It now appears likely that the prisoners of war issue will shortly become the sole remaining fundamental issue in the Korean armistice negotiations” (FRUS 1952–54, 15: 44). From that point on, disagreement over POW repatriation took center stage and single-handedly held up negotiations for 15 months. What caused POW repatriation to become such a hindrance to the process? As Chapter 5 argued, after the Chinese intervention, the NSC-68 mindset created twin constraints on the Truman coalition’s strategic choices—holding the line in Korea while avoiding global war—that remained in place for the rest of Truman’s time in office. This macro-level world-view filtered down into its micro-level policy decisions, like the POW voluntary repatriation policy. This section argues that the Truman coalition did not want to compromise on the POW policy, because it believed that standing up to the Soviets on the issue of voluntary repatriation would fight the moral crusade against Communism and help deter a future war. At the same time, however, it did not believe it could escalate militarily to pressure the Communists in the negotiating tent, for fear of sparking World War III. In other words, these two constraints were a policy-level version of the NSC-68 mindset dilemma. voluntary repat r iat ion and the infor mat ion obstacle

The United States began armistice negotiations with a one-for-one prisoner exchange policy. This policy was strongly supported by Acheson, who argued against the Department of Defense’s preferred voluntary repatriation policy. The Army’s chief of psychological warfare, Brig. Gen. Robert McClure, first raised the issue of voluntary repatriation, arguing for both its humanitarian and propaganda benefits. McClure predicted many Chinese and North Korean POWs would be severely punished, sentenced to slave labor or killed on their return, as Soviet citizens had been after World War II. Thus, McClure argued voluntary repatriation and the prospect of asylum might induce Communist soldiers’ defections in this and future wars, while repatriating all Communists now could hinder future psychological warfare operations. After a Defense Department July 1951 study endorsed McClure’s ideas, the JCS and new Defense Secretary Robert Lovett agreed (Foot 1990: 87–88; Hermes 1966: 136–37; Stueck 1995: 244–45; Rose 1994: 307; Bernstein 1983: 276). Acheson unequivocally rejected the idea in September, arguing the only exceptions to forcible repatriation should be prisoners who helped the UN Command (UNC) and South Koreans who had been forcibly impressed into the North Korean Army. Acheson suggested the UNC could release these prisoners preemptively before an armistice and thus still abide by the Geneva Convention (FRUS

the united states, part 2



1951, 7: 857–58; Bernstein 1983: 277).1 Acheson’s response led Lovett and the JCS to revert to their former position of a one-for-one prisoner exchange, until all UN prisoners were home, and then the return of remaining Communist POWs in the UN Command’s possession. (In effect, they argued for a staggered all-for-all exchange.) This policy would probably have remained in effect if not for Truman’s sudden interest in the problem. As Chapter 5 explained, Truman placed the fighting in Korea in the context of the “struggle between freedom and Communist slavery.” Focusing on Communism’s moral and spiritual dangers, Truman argued faith and ethical action were integral to any victory in the Cold War. By September 1951, he explicitly called for an international religious campaign against Communism—the Campaign of Truth. Linking America’s founding and its religious heritage, Truman argued that “at this time of international peril and uncertainty we should look back to these beginnings and rededicate ourselves to those ideals.” Thus, the same month Acheson rejected the voluntary repatriation policy, Truman was reinforcing his interest in defending “the spiritual values—the moral code—against the vast forces of evil that seek to destroy them” (quoted in Spalding 2006: 213, 215, 217). In line with these ideals, on October 29 Truman told Acting Secretary of State James Webb that he was against the one-for-one POW policy, because it was not “equitable” and might lead to harm for those prisoners who had surrendered and cooperated with the United States. Webb pointed out that the time might come when voluntary repatriation would be the last outstanding issue and thus hold up the settlement. Truman acknowledged this fact, but said he would only agree to an all-for-all prisoner swap if the UNC received “some major concession that could be obtained no other way” (FRUS 1951, 7: 1073; Ferrell 1980: 250–51). There is little indication from this conversation that Truman had reflected on the POW policy’s ramifications—both its first-order effects on the Korean War armistice negotiations and its second- and third-order effects on other policies. The conversation highlights Truman’s natural inclination to make decisions rapidly, leading to premature cognitive closure (Larson 1985: 146–47; Sigal 1988: 211; Foot 1990: 89). Truman’s “strong personal interest” in POW policy reopened the debate in Truman’s government.2 During this debate, most uniformed and civilian defense officials argued for keeping the all-for-all POW exchange; notable exceptions were Lovett, his deputy and the air force secretary, who were uncertain about the policy. In contrast, most of the State Department—including Acheson, who reversed his position after hearing about Truman’s October comments—argued for voluntary repatriation. However, there were four State officials who bucked their

154

the korean war

colleagues’ support for voluntary repatriation. Presenting their views to Acheson, these officials raised some serious objections, including that the Administration (1) did not know the full size of the POW problem and that the Communists might break off negotiations over it,3 (2) had not considered the impact of additional casualties on allied and domestic opinion if negotiations broke down over the issue,4 (3) could not tell which POWs would want to be returned, because of internal camp politics5 and (4) had refused to consider whether the Communists would reprise against US prisoners if the United States refused to return Communist POWs.6 Acheson, knowing Truman’s preference, disregarded his subordinates’ objections to voluntary repatriation. Instead, he worded his February 8 policy proposal to Truman so that it would resonate and have the maximum appeal: Any agreement in the Korean armistice which would require United States troops to use force to turn over to the Communists, prisoners who believe that they would face death if returned, would be repugnant to our most fundamental moral and humanitarian principles . . . (FRUS 1952–54, 15: 43–45)

In this proposal, Acheson did not present his subordinates’ objections, so Truman never fully considered the policy’s implications and possible trade-offs. Instead, by approving this memo, Truman and Acheson effectively defined voluntary repatriation as an irrevocable moral principle. As Acheson later put it, “Thus, we took our basic position, from which we afterward never wavered. The execution of the decision did not proceed as planned. Circumstances unforeseen and embarrassing intruded” (Acheson 1969: 653). Truman’s intervention and Acheson’s reversal on POW policy catalyzed a consensus; all key members of the Truman governing coalition except Chief of Naval Operations Admiral Fechteler fell in line behind it. In the United States, all major newspapers and most Congressmen supported the policy (Bernstein 1983: 283; Caridi 1968: 179–80; Rose 1994: 291; Foot 1990: 92). At the February 27 meeting to approve the policy, the Truman coalition swept aside possible evidence that its decision could mean stalemate and significantly prolong the war. For example, Acheson told Truman that he consulted key allies and “none . . . indicated any disagreement with our position on this question” (quoted in Bernstein 1983: 281). Actually, most Western allies had mixed feelings about the policy but lacked enough confidence in opposing arguments to argue strongly against Truman (Foot 1990: 92). Even more importantly, two days before the decision, two officials returned from a trip to estimate the number of Communist prisoners unlikely to elect to return home. On February 25, Gen. Hull (Vice Chief of Staff of the Army) and U. Alexis Johnson (Deputy Assistant Secretary of State for Far Eastern Affairs)

the united states, part 2



told Truman that the UNC expected only 15,900 of 132,194 Communist prisoners would violently resist repatriation. The resisting POWs included 11,500 Chinese. Accordingly, Hull and Johnson acknowledged that “qualitatively as well as possibly quantitatively, the problem of CCF [Chinese] POWs and possible Communist reactions thereto was much more difficult than that of Korean POWs” (FRUS 1952–54, 15: 58–59; also Bernstein 1983: 275; Goodman 1978: 252). It is unclear from notes of the February 27 meeting whether these numbers played any role in the voluntary repatriation policy decision. In March 1952, as POWs became the last contentious issue in armistice negotiations, the US negotiators used Hull and Johnson’s February estimates to propose that POWs unwilling to return home be removed from the lists and all others repatriated. This proposal would have allowed the United States to achieve its aims while protecting the Communists from having to endorse the principle of voluntary repatriation.7 Unfortunately, when US negotiators made this offer, the February estimates were no longer accurate and more recent information was ignored. This more recent information suggested those numbers were at best wildly inaccurate and at worst too optimistic, because it would be virtually impossible to get an accurate screening and allow prisoners to make a genuine choice about repatriation. It also suggested that many more than 15,900 POWs would resist repatriation. Yet this information—from at least four sources, cited below—was ignored. First, more than a month before Truman adopted the voluntary repatriation policy, Frank Stelle of State’s Policy Planning Staff argued the UN camps embodied a “reign of terror.” Stelle argued Chinese prisoners especially believed the camps were “violently totalitarian,” implying that getting an accurate count among them would be impossible (quoted in Foot 1990: 125). Second, since early February, the US embassy in Korea had reported to Johnson that it would be impossible to count POWs accurately, for organizational and political reasons. These reports noted fear among prisoners intimidated by the brutal pro-KMT quasi-police who ran the camps and implied the number of non-repatriates would be very large (FRUS 1952–54, 15: 98–99; Foot 1990: 125). Third, the UN camps had already observed much violence, starting with disturbances in September 1951 and escalating to full-scale riots in the compound on Koje-do island by February 1952.8 Although these riots were not publicized, the UN Command, JCS and Johnson knew about them (Hermes 1966: 237–40; Ridgeway 1967: 208–9; Foot 1990: 118). This escalating pattern of violence reinforced the embassy reports that Communist prisoners were afraid and would probably not respond accurately in the repatriation screenings. There were many reasons for this violence. The overcrowded camps—some held five times the

156

the korean war

number they were designed to contain—lacked adequate sanitation, lodging and heating. Moreover, medical problems—especially digestive diseases, dysentery, malaria and tuberculosis—were rampant, often caused by the “inferior grains” that replaced two-thirds of the rice ration (Foot 1990: 110; MacDonald 1986: 135; Kaufman 1999: 63–64). Because of these conditions and shortfalls in guard capacity,9 the UNC authorities also sanctioned quasi-official police organizations among the prisoners, which operated like private armies (Foot 1990: 112). As one general involved with UNC camp administration understated, there were “many unsatisfactory aspects to the situation on the [Koje-do] island” (FRUS 1952–54, 15: 98–99). Fourth, and most importantly, UN Commander Matthew Ridgway informed the Joint Chiefs in mid-March that the number violently resisting repatriation “while admittedly highly conjectural” would be at least 73,000—not 15,900 as had been estimated in February (FRUS 1952–54, 15: 92). Hearing Ridgway’s numbers, the Joint Chiefs proposed reopening the whole question of voluntary repatriation. In a State-JCS meeting on March 19, State officials disagreed and argued instead that Truman considered the February 27 decision to be the “final position on the POW question” (FRUS 1952–54, 15: 97, 100). Then, in another State-JCS meeting two days later, Air Force Chief of Staff Gen. Vandenberg specifically asked about using Ridgway’s newer estimate in the negotiations. In reply, Johnson ignored the newer information and repeated the February numbers: “between 3,000 and 4,000 Koreans and 11,000 Chinese who would not want to be repatriated. [Johnson] personally, although he was obviously in no position to question the estimates, felt that the Chinese figure might be high” (FRUS 1952–54, 15: 113). Johnson directed the JCS to use the February numbers, which they instructed UN negotiators to do. In Johnson’s memoirs, he admitted receiving information in March about camp conditions from the US embassy in Korea and revised repatriation figures from Ridgway. However, he did not acknowledge having any opportunity to reopen discussions about the POW policy. Instead, Johnson argued that while the military largely ignored the embassy reports “because they came from outside the chain of command,” he was “able to make good use” of them in “private talks with Chiefs” (Johnson 1984: 148). It is unclear from the documentary evidence whether Acheson and Truman knew at this time about Ridgway’s numbers or Johnson’s unilateral decision to use the February numbers instead.10 On April 1, UN negotiators used the February estimates to tell the Communists they expected 116,000 of the 132,000 Communist prisoners would return home. The Communists assented, and polling began. As Ridgway had predicted in March, the 116,000 estimate was too high. Instead, on April 12, Ridgway reported that of the roughly 104,000 polled, about 75,000 (including 14,126 Chinese)

the united states, part 2



would resist repatriation. Moreover, there were an additional 44,000 who refused to be screened (FRUS 1952–54, 15: 144; Goodman 1978: 343, 346). The screening process also reinforced earlier information that most POWs refused repatriation because of the violent tactics of the pro-Chiang quasi-police before and during the screening. The Truman governing coalition was shocked, not least because it had never talked about what to do if many Communists, and especially Chinese, refused repatriation. Everyone agreed that “the Chinese Communists present our most difficult problem,” because by now the UNC understood Chinese delegates controlled the Communist side in the negotiations. Admiral Fechteler suggested repatriating the 44,000 who refused screening, to plump up the Communist numbers, but State’s Policy Planning Staff Director Nitze argued “we might have lost our moral position and be compelled to forcibly repatriate many POWs” (FRUS 1952–54, 15: 146–47). UN negotiators suggested rescreening POWs, especially the Chinese, but they were overruled. Immediate re-screening would admit that something had been remiss in the first poll (FRUS 1952–54, 15: 156; Goodman 1978: 354). On April 19, when UN delegates informed the Communists that only 70,000 prisoners could be returned, the Communists were understandably outraged and moved for a recess. The next day, the chief Communist negotiator charged the United States with a breach of faith: “You flagrantly repudiated what you said before” (quoted in Bernstein 1983: 284; also Goodman 1978: 367–69; Rose 1994: 292). Recognizing the screening results made an armistice unlikely in the near future, yet unprepared to back away from its position, the Truman coalition instructed UN negotiators to present a “package proposal” that bundled the prisoner question with two other disputed issues. This was a blatant ploy to save face; as the Joint Chiefs had cabled Ridgway on March 20, If it becomes necessary to recognize that total impasse has been reached, it would be to our advantage that Commies be forced to reject our proposals on several grounds and thereby emphasize their intransigence. . . . The package deal seems to us to have the advantage of having such recess occur with 3 items rather than merely the 1 issue of POWs. (FRUS 1952–54, 15: 107)

The UNC put forward the package proposal on April 28, but on May 2 the Communists agreed to trade concessions on the two other disputed items: The Communists agreed to stop insisting the Soviet Union serve on the neutral armistice supervisory commission, in exchange for building new airfields in North Korea after the armistice. As a result, prisoner repatriation formally became the last remaining issue, and on May 7, the talks broke down. That day, Truman told the American public that

158

the korean war To agree to forced repatriation would be unthinkable. It would be repugnant to the fundamental moral and humanitarian principles which underlie our action in Korea. To return these POWs in our hands by force would result in misery and bloodshed to the eternal dishonor of the United States and the United Nations. We will not buy an armistice by turning over human beings for slaughter or slavery. (Truman 1956: 489, emphasis added)

Reiterating the moral stance he expressed with the Campaign of Truth and using language suggesting he was already trapped by effort justification, Truman drew a line with this speech for the policy of voluntary repatriation. voluntary repat r iat ion and the ent r apment obstacle

It is important to remember that during this time, Truman was in the midst of the domestic steel crisis. Facing election year pressures not to abandon one of the Democrats’ most important constituencies—organized labor—Truman seized the nation’s steel mills in April 1952 and now faced angry opposition on the Hill and among the wider business community.11 After Truman’s all-time low approval rating in January 1952 (23 percent), the steel crisis eroded most of the remaining public support for the war and the economic stabilization policies associated with the Cold War rearmament effort.12 Adding insult to injury, the US Supreme Court ruled the seizure unconstitutional on June 2—with only two of Truman’s four court appointees supporting him—leading to a massive 53-day strike that Truman had hoped to avert in the middle of an election year. The steel crisis provided Republican critics with further evidence that Truman’s mobilization policies were askew, while they continued to argue the Communists were using armistice negotiations as a screen for their own massive military buildup behind the 38th Parallel (Caridi 1968: 182, 191–92; Pierpaoli 1999: 160–77; 192; Hamby 1990: 594–97). As a result, the public and Congress continued to link economic stabilization to the Korean War—a misperception that the Truman coalition could not dispel. The same day as Truman’s righteous speech about the morality of voluntary repatriation, the Koje-do POW camp commander, Brig. Gen. Francis Dodd, was kidnapped by prisoners when he entered a compound to discuss complaints about the screening process. This kidnapping was the beginning of wider public knowledge about POW camp conditions, prisoner riots and the UN’s screening problems. In an effort to save Dodd’s life, the general who took over made a statement acceding to most of the prisoners’ demands, legitimizing reports of brutality and intimidation that surrounded the polling. During this time, information about the Koje-do riots was leaked to the press (Hermes 1966: 245–50; Ridgway 1967: 211; Foot 1990: 119).

the united states, part 2



The unexpected poll result—and the particularly large number of Chinese who refused to go home—prompted the State Department to send two officers to interrogate Chinese prisoners. These officers concluded most POWs refused repatriation because of the pro-KMT quasi-police’s violent tactics before and during the screening (FRUS 1952–54, 15: 360). For example, two interpreters in the Chinese camps apprised UNC leaders about mock screenings the quasi-police conducted: “Those [who asked to be returned] were either beaten black and blue or killed . . . the majority of the POWs were too terrified to frankly express their real choice. All they could say in answer to the question was ‘Taiwan’ repeated over and over again” (Goodman 1978: 355; Bernstein 1983: 285). All of this information, with additional embassy reports about camp conditions, was passed to Acheson and Nitze in a memorandum on May 20—which Acheson marked as having seen (Rose 1994: 321; Foot 1990: 126). Thus, two weeks after Truman’s speech about the morality of the POW policy, it was clear—up to the level of Acheson—that the official POW screening was severely compromised and possibly even conducted unethically. Yet the POW policy remained unquestioned. Not only did it fit with Truman’s larger view of the US moral crusade against Communism, but in the midst of the domestic steel crisis, the beleaguered Truman governing coalition was saving its deflated political capital for mobilization concerns. Thus, the remaining eight months of Truman’s tenure saw little progress in the armistice negotiations. After some limited re-screening during the summer, the UN increased its estimate of potential returnees from 70,000 to 82,000. Shortly thereafter, the Communists dropped their demand from returning all POWs to returning 100,000—as long as that figure included all Chinese prisoners. But the deadlock persisted, because both sides stuck by their new demands and refused to change their positions on the principle of voluntary repatriation (Bernstein 1983: 293). After Truman’s May 7 speech, a rough consensus developed domestically about the necessity and rectitude of the voluntary repatriation policy. No mainline US politician—Republican or Democrat—objected and very few newspapers raised doubts (Bernstein 1983: 287). Instead, Chiang Kai-Shek’s Republican supporters immediately moved to ensure Truman would not waver. As Senator Knowland stated on the Senate floor, “I hope that the Government . . . is not going to consider turning over a single captive soldier . . . who has surrendered under those guarantees” (quoted in Foot 1990: 100). Congressmen waved copies of the UNC’s surrender leaflets (dropped by the millions over North Korea) and exhorted the “promises” the United States could not repudiate if it wished to maintain its moral standing around the world. The POW policy received further support after

160

the korean war

the Communists released names of UN prisoners under their control, listing only about 30 percent of the 11,200 Americans who had been captured and leading to fears the others had died or been murdered in the Communist camps (Foot 1990: 140–41; Caridi 1968: 180). While some domestic dissent of voluntary repatriation existed in 1952 among wives of American POWs and soldiers who had fought in Korea, it never achieved much traction. The dissent was uncoordinated and always counteracted by Administration officials and Congressional supporters of the policy. Moreover, those parts of the public who were against the war in general—radicals, pacifists and pragmatists—did not generally raise the POW issue (Mantell 1973: 210–11, 269–71). However, more significant dissent about voluntary repatriation existed abroad. Even before the US adopted its policy, many allies had been uneasy about pursuing it, although their objections did not reach the American public (Bernstein 1983: 287; Foot 1990: 100). Some of the allies’ objections were muted after Britain supported the policy, but the screening results and Dodd’s kidnapping after the Koje-do riots suggested to them the screening had been illegitimate. During the summer, allied objections to voluntary repatriation became more prevalent. As the Australian prime minister noted, “If the allied choice lay simply between a continuation of the war and acceptance of forcible repatriation . . . [I] would have no hesitation choosing the latter” (quoted in MacDonald 1986: 145). Allied objections to the POW policy were soon followed by their negative reactions to the new UN punishment bombing campaign, started in May 1952, which focused on attacking North Korean civilian morale. This Douhet-style “air pressure” strategy consisted of three operations: (1) a concerted attack on North Korea’s five hydroelectric power facilities, (2) incendiary raids on North Korea’s capital, Pyongyang and (3) firebombing attacks on 78 North Korean towns with Communist supply and maintenance depots. These large-scale bombing operations occurred on a regular basis over the summer (Foot 1985: 177–78; 1990: 136–37; Pape 1996: 161; Futrell 1956: 98–99; 1961: 449–51; Schnabel and Watson 1979: 843–44, 847; Crane 2000: 117–22). While these punishment bombing raids were extremely popular in the United States, they produced negative reactions in friendly and neutral countries. European critics feared these attacks, especially those against the power plants on the Yalu, might drive China to break off negotiations and widen the war. Britain was especially upset about not being informed beforehand. Despite Acheson’s apologies, Britain believed that consultations had been deliberately omitted for fear that the allies would raise objections (Acheson 1969: 657; Foot 1985: 179). The Republican opposition seized on this small crisis with the allies as further proof that they were “ungrateful” and “uncooperative” (Caridi 1968: 181).

the united states, part 2



Allied objections to the bombing campaigns, voluntary repatriation and POW screening all contributed to State Department officials assessing US war policy by this point as unsuccessful. Complicating matters, the allies objected to ROK President Syngman Rhee’s ongoing constitutional crisis (see Chapter 3). Allied and world opinion served as State’s policy assessment metric, and the metric suggested policy performance was poor. As a result, State officials began to search for ways to meet allied objections and move closer to settlement—even if this meant compromising on the principle of voluntary repatriation. Thus, Assistant Secretary of State for UN Affairs John Hickerson suggested the UNC “cautiously but promptly” follow up on two Communist proposals to conclude the armistice on the basis of the Geneva Convention, with prisoners not wishing repatriation handled afterwards (FRUS 1952–54, 15: 345).13 Nitze and the Policy Planning Staff advocated seizing any opportunity to exchange prisoners “with or without formal agreements” (FRUS 1952–54, 15: 262–63). Most ironically, Johnson—who withheld key information in order to achieve the voluntary repatriation policy in the first place—argued the POW question should be decided after the armistice (Foot 1990: 132). Other proposals discussed by State officials included presenting the case to the International Court of Justice, capturing more prisoners, juggling the figures or working for a settlement outside the current negotiations (Foot 1985: 175). In September 1952, two proposals for changing the POW policy—one from the Progressive Party’s presidential candidate and the other from the president of Mexico—reopened a debate about war policy within the Truman governing coalition.14 Predictably, as the indicator approach to war policy would suggest, most State officials found these proposals attractive, while military officials did not. State officials pointed to allies’ reactions as a reason for backing these proposals and deferring the question of non-repatriates until after the armistice was signed (FRUS 1952–54, 15: 463–65, 482–89). In contrast, civilian and uniformed Defense officials pointed to new UNC assessments that bombing was “having a material effect on the civilian morale in North Korea. . . . [and] apparently undermining . . . their ability to wage and support a war” (FRUS 1952–54, 15: 466; also 469–70, 506–7). Thus, meeting with Truman on September 15, Lovett argued that responding to those proposals would be a “sign of weakness” that “would un-do much of the advantage currently being gained by increased military pressure.” DoD opposed deferring the POW question until after the armistice and instead favored advancing a final negotiation offer that, if rejected, would lead to an indefinite recess of the talks (FRUS 1952–54, 15: 513). Given Lovett’s framing, Truman quickly gave his support to the military position—without even hearing the detailed views of the State officials (who were

162

the korean war

not present at the meeting)—because he feared appearing weak and indecisive (Foot 1990: 145). The cumulative effect of the Republican right wing’s attacks for being “soft on Communism” had taken its toll. Hence, Truman said he “saw no real prospect of getting an armistice other than to persist in our course of action and increase the military pressure in so far as possible” (FRUS 1952–54, 15: 513–14). While Acheson had not weighed in on this debate, he agreed with his subordinates that the United States had to respond to the Mexican proposal to show other UN members it “had exhausted every reasonable effort to obtain an armistice.” If the Communists rejected the Mexican proposal—which Acheson thought likely—the United States at least would have built “the best possible platform” for unilaterally recessing the negotiations (FRUS 1952–54, 15: 524; Foot 1985: 182; Bernstein 1983: 300). However, by the time Acheson discussed this with Truman on September 24, he must have learned about Truman’s preferences expressed to Lovett. As a result, Acheson pushed neither for a presidential statement responding to the Mexican proposal nor for leaving non-repatriates for negotiations after an armistice. Instead, he tilted towards the DoD’s preference for increasing military pressure and argued “the time has come to recess the armistice negotiations.” Perhaps Acheson saw the writing on the wall. In any case, Truman reiterated his support for increased military pressure and a continued hard line on the POW policy: “unless we wipe the slate clean, an armistice will do no good . . . [because] we cannot give in on principles” (FRUS 1952–54, 15: 534, 537). As a result, on September 28 the UN negotiators offered three final variations on the voluntary repatriation policy, which the Communists rejected—as expected—ten days later. UN negotiators recessed the talks indefinitely. At a press conference that same day, Acheson reiterated that the United States would “not compromise on the principle that a prisoner should be forced to return home against his will” and any weakened resolve on this issue “would constitute an abandonment of the principles fundamental to this country and the United Nations” (quoted in Foot 1990: 150). Hence, Acheson again linked the prestige and reputation of the Truman Administration with voluntary repatriation. In the process, the Truman governing coalition lost any remaining maneuvering room: there was no option left but to recess negotiations, step up bombing and hope for a settlement. In response, on October 6, Chinese forces began one of their largest attacks since armistice talks began in July 1951; October’s fighting produced the highest UNC casualties in a year—just in time for the US presidential elections (Stueck 1995: 291). Although this was definitely “too little, too late,” the Truman coalition had no other options. Public support for continuing negotiations had been declin-

the united states, part 2



ing steadily since late 1951, with most Americans wanting “stronger steps” if the Communists did not agree to an armistice on UN terms (Foot 1985: 220; 1990: 172). On the one hand, because the Administration was so wholly identified with a limited war policy, extending hostilities too much beyond bombing would have led to the charge that Truman was simply “electioneering.” While not running for reelection himself, Truman had personally orchestrated selection of the Democratic candidate, Adlai Stevenson. Although Truman campaigned vigorously for Democrats throughout October, he deliberately stayed away from Stevenson to allow him to distance himself as much as possible from “Truman’s War” (Hamby 1990: 600–614). At the very least, escalating now would have caused Republican critics to lambaste him for failing to follow MacArthur’s lead nearly two years earlier. On the other hand, after so publicly tying his governing coalition to voluntary repatriation, Truman saw no possibility for retreat and little room for compromise on that issue, either (Foot 1985: 220; Stueck 1995: 264). While the record suggests Truman had strong normative commitments to the voluntary repatriation policy, others in his coalition advocated the policy without adequately considering its full ramifications. We cannot know whether Truman could have been convinced to change his mind on voluntary repatriation, but knowing how much influence his advisors had on other policies suggests he might have been swayed had they made strong arguments for a different policy. The entrapment seems all the more tragic, because information about the difficulties associated with voluntary repatriation had been available before Truman had made his decision and publicized it. Ironically, the same State officials who fought initially for voluntary repatriation were the very ones willing to search for a compromise during the last eight months of Truman’s tenure. Finally, the Truman coalition’s stance about voluntary repatriation became irrevocably wedded to the NSC-68 mindset it had internalized. The Truman coalition did not want to compromise on POW policy, because it believed standing up to the Soviets on the issue of voluntary repatriation would fight the moral crusade against Communism and help deter a future war (Bernstein 1983: 288). As Truman argued, voluntary repatriation was “vital as a symbol of the strength and determination of the West. Firmness now would be the only way to deter new actions in other portions of the world” (Truman 1956: 358). At the same time, the Truman coalition could not escalate militarily too much, because it believed the Communists were capable of challenging us in a similar way in many places, and . . . if they chose, plunge us and the world into another and far more terrible war. Every decision I made in connection with the Korean conflict had this one aim in mind: To prevent a third

164

the korean war World War. . . . Every one of these steps had to be taken without losing sight of the many other places where trouble might break out or the danger might befall us if we hazarded too much in any one place. (Truman 1956: 364–65, emphasis added)

With such boundaries on the spectrum of possible policies, increasing conventional bombing was the only option the lame-duck Truman governing coalition felt it had. In sum, the US decision to adopt the POW voluntary repatriation policy demonstrates at a micro level the interacting dynamics of the information and entrapment obstacles. Truman adopted the voluntary repatriation policy prematurely, based on his beliefs about the moral fight against Communism but without fully considering its effects. Subsequently, information suggesting the policy would lead to an impasse in armistice negotiations and undermine the United States’ ability to “hold the moral high ground” was systematically ignored. In other words, the coalition chose not to know this policy would significantly prolong the war. The end result was an information obstacle in which Truman staked his—and the nation’s—reputation on voluntary repatriation, which led to entrapment dynamics, as well.

The New US Governing Coalition: Eisenhower’s Election While Truman stood firm on the repatriation policy—simultaneously juggling the domestic political fall-out from the steel crisis and the international effects of the prisoner screenings, POW camp riots and Rhee’s constitutional crisis—the Republican presidential primary was heating up. General Dwight D. Eisenhower faced off against his right-wing opponent, Ohio Senator Robert Taft, the leader of the “Old Guard” isolationists and Asia-Firsters in Congress. When Massachusetts Senator Henry Cabot Lodge filed Eisenhower’s name in the New Hampshire primary in January 1952, Eisenhower had neither Republican credentials nor embodied the isolationist perspective of the Republican majority. In fact, he epitomized a continuation of the Truman Administration’s Europefirst internationalist strategy. Eisenhower’s candidacy—supported by the Eastern internationalist Republicans who had been defeated in three successive presidential elections—infuriated Taft and Taft’s right-wing base of Asia-lationists. The antipathy ran both ways. As Eisenhower said bluntly during the primaries: “I’m running because Taft is an isolationist. His election would be a disaster” (quoted in Doenecke 1979: 218). Not surprisingly, the Republican National Convention was bitter, with Taft supporters packing the agenda and controlling the foreign policy platform and Lodge unveiling a last-minute “fair play” amendment to block disputed delegates from voting (Reinhard 1983: 84–89; Doenecke 1979: 221;

the united states, part 2



Caridi 1968: 217–21). Regardless, Eisenhower easily won the nomination, and most isolationists eventually supported his candidacy. Truman’s failure to end the Korean War contributed significantly to Eisenhower’s election. Although Truman announced in March 1952 he would not seek reelection, “Truman’s War” proved to be too big a hurdle for Stevenson to overcome.15 While Eisenhower did not offer an explicit war-ending strategy during the campaign, voters believed he would find a way to end the war successfully.16 In fact, the only Republican alternative proposed replacing US soldiers on the line with ROK troops (Caridi 1968: 230; Immerman 1999: 68). Nevertheless, Eisenhower steadily built his campaign so that by October, he sold himself as the “man who had virtually ended one war [World War II] and could surely terminate another” (Caridi 1968: 224). Americans believed him, delivering him a landslide election, with 55 percent of the popular vote (as opposed to Stevenson’s 44.3 percent) and an electoral count of 442 to 89. The Republicans rode his coattails into narrow majorities in both chambers of Congress, their first time controlling the executive and legislature in 20 years (Caridi 1968: 243; Reinhard 1983: 96). Eisenhower campaigned that his first priority would be to end the war. After the vicious Republican primary, Eisenhower slowly co-opted much of Taft’s hawkish foreign policy platform, including “massive retaliation” and “liberation” (discussed below). Thus, while in June Eisenhower upheld containment and backed armistice negotiations, by October he accused the Democrats of having “lost” China and favored bombing along the Yalu River and “liberating captive nations” (Doenecke 1979: 224). The climax of Eisenhower’s increasing focus on Korea was an October 24 speech, repeated twice in the next week, in which he pledged to go to Korea. Building on his military reputation, Eisenhower carefully cultivated the idea that only a new perspective, embodied in a new administration, could win the war. In his own words, he argued that a coalition shift was necessary: The first task of a new Administration will be to review and re-examine every course of action open to us . . . to bring the Korean War to an early and honorable end. . . . For this task a wholly new Administration is necessary. . . . The old Administration cannot be expected to repair what it failed to prevent. [The new Administration will] forego the divisions of politics and . . . concentrate on the job of ending the Korean War—until that job is honorably done. That job requires a personal trip to Korea. . . . I shall make that trip. . . . I shall go to Korea. (New York Times, October 25, 1952: 8, emphasis added)

Without supplying details, Eisenhower nurtured the impression that only a Republican Administration and a Republican Congress could honorably end the war (Zhang 1992: 12–21; Caridi 1968: 209–45).

166

the korean war

how the eisenhower coalit ion shift overcame t ruman ’ s obstacles to peace

Until resigning his post as NATO commander in June 1952, Eisenhower served as part of the Truman governing coalition and personally agreed with its Europecentric, internationalist strategy for containing Communism. Indeed, the major impetus for his candidacy was to block Taft’s isolationist vision from becoming a reality. Given his similar strategic perspective, how was Eisenhower able to end “Truman’s War”? Although Eisenhower did not shift US war aims, he did shift the strategy for achieving them. Like Truman, Eisenhower did not intend to reunify the peninsula, and he maintained the policy of voluntary repatriation (FRUS 1952–54, 15: 1062; Bernstein 1983: 303).17 The Eisenhower governing coalition agreed with Truman that voluntary repatriation could help prevent future conflict with the Soviets because it could encourage defections elsewhere in the Communist bloc. As Dulles argued in 1953, “soldiers in these armies who want freedom will be more apt to desert and surrender . . . the Red Armies [will] become less dependable and there is far less risk that the Communists will be tempted to use these armies for aggression” (quoted in Gaddis 1987: 176; also Foot 1990: 161, 198–99). However, while Eisenhower retained the POW policy, he implemented a macrolevel (strategic) shift in how to contain Communism. This shift helped him avoid the dilemma of the NSC-68 mindset—how to hold the line on POW voluntary repatriation while avoiding military escalation and global war. Thus, as the domestic coalition theory would predict, Eisenhower’s coalition could overcome Truman’s obstacles to peace, in four ways. First, as a Republican, Eisenhower was freed from the worst of Truman’s entrapment from the Republican right wing. Soon after the election, he took steps to placate Taft’s supporters, including (1) a publicized meeting with MacArthur to hear the former commander’s hawkish plan for ending the war and (2) his promised trip to Korea to meet with UN Commander Clark, as well as the commander of the US military advisors in Taiwan, who joined them there (Ambrose 1983b: 30–35; Stueck 1995: 305). More importantly, on February 2, he announced in his State of the Union address that he was withdrawing the Seventh Fleet from the Taiwan Strait. This opened the possibility for Chinese Nationalists to attack mainland China and—more importantly—mollified the Republican right wing. After the speech, Eisenhower left the chamber arm-in-arm with now Senate Majority Leader Taft. More generally, Eisenhower actively built bridges with this constituency, telling his subordinates to treat Taft with “presidential deference” (Doenecke 1979: 231; Reinhard 1983: 102, 108; Caridi 1968: 259).

the united states, part 2



Second and relatedly, Eisenhower decoupled the Korean War effort from Cold War mobilization and rapidly set about dismantling Truman’s highly unpopular mobilization infrastructure and economic controls. In the State of the Union address, Eisenhower announced the end to all price and wage controls and his plan to liquidate immediately most mobilization agencies Truman had created to rearm for the Cold War. The decision to eliminate the economic controls clearly was “bowing to pressure from business and other Republican-oriented groups that had helped him get elected,” because by that point 61 percent of the public favored continuing them (Pierpaoli 1999: 219). Thus, while Truman had intended the mobilization to be permanent, Eisenhower took it apart. To be sure, Eisenhower could do this because of the impressive rearmament legacy that Truman left him—with a defense budget roughly three times and a military buildup roughly five times as large as they were in June 1950 (Friedberg 2000: 125). Nonetheless, dismantling Truman’s “garrison state” allowed Eisenhower to decouple Cold War mobilization from the Korean War, thereby meeting political constituents’ needs and buying himself leeway for ending the war. Third, the public and the press gave the Eisenhower coalition significant leeway for ending the war. In February 1953, 70 percent of Americans gave Eisenhower a favorable foreign policy rating, compared with 30 percent for Truman the previous September. Polls also showed 62 percent supported “stronger steps” should the Communists not agree with the UN’s armistice terms. As Sterling argues, “with the ‘stronger steps’ undefined in the poll, US officials could consider public support for a wide array of options, including nuclear weapons” (1998: 459; also Foot 1990: 160, 172; Mueller 1985: 80). In other words, American public opinion was sufficiently mixed to allow Eisenhower freedom of action. Likewise, the press gave Eisenhower maneuvering room to solve the stalemate, with some papers questioning the risks of an expanded war while others argued escalation would break the stalemate (Foot 1990: 172; Sterling 1998: 459–60). In short, Eisenhower believed he had a “mandate from the American people to stop this fighting” (quoted in Foot 1985: 219). As long as he settled the war relatively quickly, it would remain “Truman’s War.” Because Eisenhower had built up the expectation during his campaign that he would end the war soon, the public both looked for and expected decisive action. Finally, the new Eisenhower coalition came to power with a different conceptualization of the war’s costs and benefits. Although the Eisenhower coalition shared the Truman coalition’s views about the Soviet Union and its expansionist intentions, they parted ways significantly in terms of strategy for dealing with this threat (Foot 1985: 33; Holsti 1970; Smith 1974; Mayers 1986: 111–12; Divine 1981: 22; LaFeber 1991: 148). While the Truman coalition retained faith in its NSC-68

168

the korean war

mindset until the end, the Eisenhower coalition advocated a policy of nuclear “deterrent power” as a more economical alternative to Truman’s “treadmill policies” of containment. Eisenhower’s secretary of state, John Dulles convinced Eisenhower that “only reliance on nuclear weapons would make containment work over the long haul at a reasonable cost” (Gaddis 1997: 107). As Dulles argued in his famous Life magazine article, instead of trying to defend the “25,000 miles . . . of the free world’s frontier with the Soviet world” on the ground, the United States needed to develop the will and organize the means to retaliate instantly against open aggression by Red armies, so that, if it occurred anywhere, we could and would strike back where it hurts, by the means of our own choosing. (Dulles 1952: 148–49, emphasis in original)

Thus, the Eisenhower coalition did not believe Communism must be faced on the ground in every limited war—like Korea—that promised to crop up. Rather than relying on limited wars and the expensive ground forces of containment, the Eisenhower Administration chose to rely on “massive retaliation”—threatening to respond with nuclear weapons to deter any Soviet-sponsored aggression (Donaldson 1997: 92; Bowie and Immerman 1998: 67). As Eisenhower told the public, “We have adopted a new principle. . . . Rather than let the Communists nibble us to death all over the world in little wars, we will rely in [the] future on [our] massive mobile retaliatory powers” (quoted in Ambrose 1983b: 336). Eisenhower and Dulles had a qualitatively different view of nuclear weapons than their predecessors. They frequently discussed nuclear weapons, both publicly and privately, as though they were “just another weapon in the arsenal” (FRUS 1952–54, 15: 977; also Keefer 1986: 277; Pape 1996: 159; Foot 1985: 204–5). Dulles argued the Soviets had convinced much of the world that atomic weapons must be set apart from all others as a “special category,” but this was a “false distinction”; the sooner it was exposed as such and the taboo on atomic weapons removed, the better (FRUS 1952–54, 15: 769–70). In the Eisenhower coalition’s view, nuclear weapons were not seen as the means of last resort, but readily usable instruments of American policy. For them, relying on nuclear weapons was not only economical, but quite necessary, if the United States were to pursue its policy objectives without expending too many resources (Calingaert 1988: 197). The new policy perspectives of “deterrent power” and “liberation” were matched by the increasingly hardline views of the military members of Eisenhower’s coalition.18 As already noted, UN Commander Clark and the JCS had increasingly argued for military escalation in Korea over the summer and fall of 1952. To a large degree, this greater willingness to escalate resulted from

the united states, part 2



the resurgence of American military power in late 1952 and 1953. The military mobilization begun in 1950 under the auspices of NSC-68 was finally showing results—especially in the nuclear weapons area. As a result, by January 1953, the same JCS that had been so cautious in 1950 now advocated nuclear escalation if that was what was needed to end the Korean War (Trachtenberg 1988–89: 28, 30). This shift in views resulted from the United States’ successful military buildup— not simply because Eisenhower was in charge. Regardless, the JCS’ more hawkish position helped Eisenhower create a coalition that supported escalation to end the war. This willingness in principle to use nuclear weapons had two major implications for ending the Korean War. First, unlike Truman’s coalition, Eisenhower did not believe the Soviets needed to be countered on the ground everywhere around the globe. Nuclear deterrent power obviated the need to take a direct stand in limited wars. Second, the Eisenhower coalition had a much more detached attitude about the European allies (Trachtenberg 1988–89: 38). After all, allies are not as important if you do not plan to fight every war on the ground and are less concerned with basing troops on allies’ soil. For both reasons, the Eisenhower coalition was freer to consider expanded military options. Without the partisan attacks from the Republican right wing, the politically unpopular mobilization campaign and the constraints of the NSC-68 mindset, the Eisenhower coalition had more leeway to end the war. In sum, while Eisenhower retained Truman’s war aims, the domestic coalition shift brought different leaders—with different constituencies, different worldviews and different conceptualizations of the war—that left them freer to consider an expanded palate of strategic options. the eisenhower coalit ion ’ s st r ateg y for ending the korean war

Although Eisenhower did not shift US war aims, he did shift the strategy for achieving them. The Eisenhower coalition believed it would need to escalate militarily in order to convince the Communists that US objectives were not negotiable. Although Eisenhower preferred to end the war peacefully, he was willing—as he details in his memoirs—to consider a major offensive, including “strikes against the supporting Chinese airfields in Manchuria, a blockade of the Chinese coast and similar measures.” In addition, he would build up US and ROK force levels and ammunition stocks in Korea. Finally, in an effort to prevent an expanded war from becoming “overly costly,” he was prepared to use nuclear weapons on “strategic targets in North Korea, Manchuria and on the Chinese coast” (Eisenhower 1963: 179–80). This section summarizes Eisenhower’s hawk-

170

the korean war

ish strategy shift to set the stage for Chapter 8, which will explore the strategic interaction between this shift and Communist behavior. The Eisenhower coalition did not switch its strategy until May 20. In April 1953, the National Security Council presented a study, commissioned in October 1952, which outlined six alternative courses of action that the United States could pursue in Korea. The study did not offer any new alternatives but reexamined all policy options—including those rejected by Truman. The first three options maintained Truman’s “limited war” restrictions on US forces, while the last three options lifted those restrictions and proposed action beyond the Korean peninsula and against Chinese territory. Five of the six options permitted the use of nuclear weapons. The report concluded that if nuclear weapons were used, “the Communists would recognize the employment of these weapons as indicative of Western determination to carry the Korean War to a successful conclusion” (FRUS 1952–54, 15: 845–46; also Foot 1985: 203–4, 208–9; Foot 1990: 163–64; Sterling 1998: 455–57). The Joint Chiefs endorsed a combination of the three options lifting current restrictions, including a UN advance to the “waist” of Korea and a reduction of China’s future aggressive capabilities in Korea and the Far East more generally (FRUS 1952–54, 15: 1062; Foot 1985: 206–7; Foot 1990: 163–64). In the JCS’ view, none of these courses of action could be successfully carried out without using nuclear weapons, “in considerable numbers in order to be truly effective” (FRUS 1952–54, 15: 1014). Eisenhower agreed with the Joint Chiefs’ recommendations. He believed that in order “to keep the attack from becoming overly costly, it was clear that we would have to use atomic weapons” (Eisenhower 1963: 180; see also Zhang 1992: 124–25). In May, he announced it was time for the United States “to consider the atomic bomb as simply another weapon in our arsenal” (FRUS 1952–54, 15: 977). Therefore, at a May 20 NSC meeting—despite expressed concerns about possible Soviet retaliation against Japan, local retaliation against UN forces in Korea and the “disinclination” of allies to agree—Eisenhower stated that “he wished the rec­ ord to show that if circumstances arose which would force the United States to an expanded effort in Korea, the plan selected by the Joint Chiefs was most likely to achieve the objective we sought” (FRUS 1952–54, 15: 1067–68).19 In other words, on May 20, the Eisenhower coalition endorsed a hawkish shift in American strategy for ending the war. There were four ways this change in strategy manifested. First, at that meeting, the Eisenhower coalition formally approved contingency plans for using nuclear weapons, something Truman never did. If the war could not be ended soon, they agreed that it would be necessary to “expand the war outside of Korea” and “to

the united states, part 2



use the atomic bomb” (FRUS 1952–54, 15: 1065). Second, Eisenhower granted UN Commander Clark—for the first time—the authority to terminate, not recess, the talks if the Communists refused the UN “final offer” to be presented on May 25 (Foot 1985: 204). Third, the United States resumed its punitive conventional bombing of the previous summer, but escalated to attack North Korean irrigation dams and destroy the country’s rice crop (Blechman and Powell 1982–83: 591; Foot 1990: 166; Pape 1996: 162–64; Futrell 1961: 626–27; Rees 1964: 381–82; Hermes 1966: 461; Jackson 1973: 156–57). Finally, the Eisenhower coalition tried two diplomatic channels to signal to the Communists a growing willingness to escalate. These included conversations between (1) Dulles and Indian Prime Minister Nehru, in which Dulles said “if armistice negotiations collapse, the United States would probably make a stronger . . . military exertion and that this might well extend the area of conflict,”20 and (2) US Ambassador to Moscow Charles Bohlen and Soviet Foreign Minister Molotov, in which Bohlen emphasized “the extreme importance and seriousness of the latest UNC proposals” and made it “clear these represent the limit to which we can go” (FRUS 1952–54, 15: 1103; Zhang 1992: 125). Chapter 8 will examine the strategic interaction between these threats of escalation and Communist behavior in the war termination process. Accordingly, one of the most prominent alternative explanations for the war’s ending—that Eisenhower’s nuclear threats compelled the Communists to settle—will be taken up in Chapter 8.

Alternative Explanations This section briefly examines six alternative explanations for US behavior during the Korean War. First, Bayesian bargaining models about war termination do not explain US behavior in prosecuting and ending the Korean War. The Korean War belligerents ultimately settled for terms essentially the same as those they could have obtained two years earlier, while the battlefield remained relatively static. With neither a significant change in the battlefield situation nor the belligerents’ war aims, why the delay? And, given the delay, why then the resolution? The Bayesian updating process discussed in Chapter 2 misses the rich complexity of US domestic politics during those intervening two years, which raised a relatively trivial issue—the POW repatriation policy—to a 15-month debate. The US case study provides ample evidence of leaders’ reluctance to confront painful truths after committing to particular policies—from advancing to the Yalu River to committing to POW voluntary repatriation—so that they did not know or chose not to know they should change those policies. Once they became aware of their policies’ full implications, they felt trapped and unable to back down

172

the korean war

without admitting failure. These dynamics are not compatible with the relatively linear relationship between exposure to new information and updated expectations that the Bayesian model predicts. Second, Hein Goemans’ (2000a) approach hardly fares better in explaining US behavior. As Chapter 2 explained, Goemans argues war duration varies by regime type; in democracies, following the logic of Bayesian updating, leaders should end a war once costs outweigh the benefits. Yet, as the first alternative explanation showed, that logic does not explain the Truman governing coalition’s behavior. Third, building on the Bayesian model, Dan Reiter (2009: chapter 5) argues the United States pursued absolute victory in its advance to the Yalu River because it feared any limited settlement would lead to another North Korean attack on South Korea. Reiter argues the United States pursued absolute victory before China intervened in the war, because it wanted to solve the Communist “commitment problem” and expected absolute victory would be possible at low cost. However, after China entered the war, the expected costs of absolute victory were now much higher—potentially fighting a nuclear World War III—and US leaders did not see this as a price worth paying for long-term South Korean security. Thus, the United States abandoned its high war aims and, presumably, its strategy to solve these commitment fears. Reiter acknowledges the last two years of stalemated war demonstrate the “inefficiency of combat” for providing information for Bayesian updating, but he argues the commitment dynamics—together with Bayesian information dynamics—provide a comprehensive account of US war termination behavior. Reiter’s explanation, while parsimonious, does not completely capture the dynamics of American behavior in the war, for two reasons. First, Reiter correctly notes the United States gradually moved from its initial (June) war aims of status quo ante bellum to its high (October) war aims of reunifying the Korean peninsula by force under non-Communist rule. But his argument misses the major driver for this change. As Chapters 5 and 6 have argued, the Truman governing coalition feared Soviet aggression from the beginning of the war, not just after China intervened. In other words, the potential costs of fighting a nuclear World War III existed in both phases of the war. If this was the case, then why was this price acceptable before the Chinese intervention but not after? Certainly, China’s intervention and the “entirely new war” highlighted an actual—rather than merely an expected—increase in costs. But the account is incomplete. The missing piece, of course, is that the Korean War vindicated NSC-68’s predictions of Soviet “grand designs” for global domination through “proxy wars.” Fearing Soviet aggression, the Truman coalition used the Korean War to embark on a massive Cold War rearmament effort and justify its globalist

the united states, part 2



agenda to Congressional and public constituencies. However, winning support for this mobilization required building a coalition with moderate Republicans and Asia-lationists, leading Truman to overturn previous Asia policies—thereby returning South Korea to the US security perimeter, increasing military aid to Taiwan, shelving plans to recognize China diplomatically and launching a ground offensive to reunify the Korean peninsula. Indeed, Reiter admits that the first public call after the June 1950 invasion for crossing the 38th Parallel and seizing part of North Korea came from Senator Robert Taft—the Republican right-wing Asia-lationist who had attacked Truman for “losing” China and being “soft on Communism” (Foot 1985: 69). In other words, the causal mechanism that best explains the gradual increase in war aims from status quo ante bellum to absolute victory is domestic coalition-building dynamics, trying to win support for broader Cold War rearmament before “the year of maximum danger.” Significantly, Reiter’s account contains no discussion of NSC-68 and its profound effect on the framing of the Korean War in the wider Cold War context. Second, as Reiter himself argues, information dynamics (Bayesian updating) does not explain the two years of stalemate while the armistice negotiations dragged on. Yet, neither do the commitment dynamics central to his argument, at least not fully. Reiter offers two ways that commitment concerns shaped the UN negotiating position from July 1951 to July 1953. First, he argues commitment concerns affected UN demands for the location of the border at the end of the war. However, the location of the demarcation line was determined by February 1952, fully 17 months before the war ended. Second, Reiter argues the UN was concerned an open POW exchange might exacerbate the commitment problem by shifting the balance of power in favor of the Communists, leading the United States to hold out for voluntary repatriation. Certainly, POW voluntary repatriation single-handedly held up an armistice settlement for 15 months. However, as the evidence in this chapter shows, besides McClure’s initial argument that the prospect of asylum might induce Communist soldiers to defect in future conflicts, commitment concerns played virtually no role in the US decision to adopt and maintain that policy. In sum, Reiter’s argument fails to capture the full complexity of US behavior in the war. A fourth alternative explanation focuses on regime type, specifically that democratic leaders are more sensitive to the costs of war than other regime types (Bennett and Stam 1996, 1998; Bueno de Mesquita and Lalman 1992; Doyle 1983; Filson and Werner 2004; Gaubatz 1991, 1999; Levi 1997; Maoz and Russett 1993; Reiter and Stam 1998, 2002; Schultz 1999; Stam 1996; Wagner 2000). Public opinion research has suggested that in democracies, public support for war declines with battle-deaths (Gartner and Myers 1995; Gartner, Segura and Wilkening 1997;

174

the korean war

Mueller 1985, 1994), and specifically in the Korean War, some scholars have found a statistically significant relationship between casualties and decline of US public support (Gartner and Segura 1998). However, the pattern of Korean War public opinion seems more complex than this. After an initial spike of public support, as the “rally round the flag” thesis might suggest, the American public and press supported the advance to the Yalu River (Christensen 1996: 152). Predictably, after the Chinese intervention, there was a steep drop in public support. However, it stabilized thereafter at a moderate level until the end of the war—even as casualties continued to escalate (Mueller 1985: 50–51). As John Mueller, the first scholar to examine the public opinion argument rigorously, notes, More striking than the drop in support caused by the Chinese entry is the near-absence of further decline for the remaining 2½ years of the war. From early 1951 until the end of the war in the summer of 1953, basic support for the war . . . remained largely constant—this despite the continually mounting casualties and despite a number of important events. (1985: 51, italics in original)

Indeed, as these chapters have shown, sections of the public remained more hawkish than the Truman governing coalition, consistently supporting “stronger steps”—such as escalating militarily to “roll-back” China—than the leaders themselves. Thus, to argue democracies are more cost-sensitive than other regime types does not adequately explain US behavior during the Korean War. The evidence suggests Truman was trapped more by a hawkish constituency than a dovish one. Gideon Rose (1994) advances a fifth alternative argument, that Truman was not truly entrapped by his voluntary repatriation policy. Had the Truman coalition wanted to, Rose argues, it could have generated support for forcible repatriation instead, because public opinion was “confused, fluid and manipulable” (Rose 1994: 330–31). For example, State’s Office of Public Opinion Studies had July 1952 data demonstrating there was “virtually no sentiment in favor of making ‘concessions’” in the armistice negotiations. However, when the public was “faced with the emotionally persuasive argument of returning all POWs for the sake of getting US prisoners back,” 31 percent favored returning all Communist prisoners immediately and an additional 40 percent were in favor if it were necessary to bring US soldiers home. Only 20 percent said they would never return the non-repatriates (Foot 1990: 140–41). Thus, according to data at the time, public opinion permitted the Administration’s stand on POWs but did not require it. Moreover, as this chapter discussed, there was some domestic dissent over the voluntary repatriation policy—albeit nebulous and marginalized. US allies never fully supported the policy, either. Hence, had Truman wanted to reverse himself,

the united states, part 2



and had domestic political entrapment been the only dynamic at play, he might have successfully maneuvered to co-opt his critics and change his policy. However, the Truman governing coalition did not try these tactics—largely because these tactics did not conform to the macro-level beliefs that drove the coalition’s understanding of the war. In the end, the voluntary repatriation policy became irrevocably wedded to the NSC-68 mindset the Truman coalition had internalized. On the one hand, it did not want to compromise on the POW policy, because it believed that standing up to the Soviets on this issue would help deter a future war. On the other hand, it did not want to escalate militarily too much to compel the Communists to accept voluntary repatriation, for fear of provoking World War III. Whether or not the Truman governing coalition was politically entrapped by the POW policy, it was certainly entrapped by the NSC-68 mindset. In other words, the voluntary repatriation policy alone cannot explain the stickiness in the Truman coalition’s decisions about prosecuting and ending the war. Finally, Barton Bernstein (1983) argues the Korean War continued so long because Truman, acting alone, was unable or unwilling to change his position on voluntary repatriation. Bernstein suggests had Truman changed his mind or been replaced, the war could have ended. To some degree, Bernstein’s argument is similar to the domestic coalition argument—that changing key leaders can change a “sticky” policy—but it differs in focusing completely on one individual, Truman. As this chapter has shown, Truman’s personal views about POW policy certainly affected how that policy was initially selected and then how information calling the policy’s wisdom into question was systematically ignored. Yet Truman was not the one ignoring this information, because he never received it. We cannot know what Truman might have done had he received the adverse information. The important point is that Truman never received it, because other coalition members believed he did not need to. Thus, other coalition members’ obstacles are important in understanding the policy’s “stickiness.” Moreover, the need to remain firm on the POW policy was reinforced by the entire coalition’s NSC-68 mindset, which exhorted remaining firm to contain the Communists everywhere, in every way. Notably, when the war began, Acheson and Defense officials (both civilian and military) embodied the NSC-68 mindset much more than Truman—eventually convincing him to prioritize the Korean War and mobilization above his domestic policy agenda. To a large degree, then, other coalition members were responsible for reinforcing Truman’s bias towards standing firm, no matter the cost. Once that stand was taken, it was very difficult to back down. Most importantly, Bernstein’s argument misses the crucial fact that Eisenhower retained Truman’s position of voluntary repatriation, yet was able to end the war.

176

the korean war

Thus, as with Rose’s (1994) argument, Bernstein’s narrow focus on the voluntary repatriation policy alone is not enough to explain the stickiness in the Truman coalition’s war policy. While Eisenhower retained the POW policy, he changed the US policy on nuclear deterrent power and thus felt free to escalate the war, if necessary. This decision allowed him to overcome the twin constraints of the NSC68 mindset—holding firm in Korea while avoiding global war—that trapped Truman for so long.

Conclusions This chapter, paired with Chapter 5, has examined the interacting dynamics of information and entrapment obstacles in US decisions about prosecuting and ending the war. While the previous chapter examined these dynamics at a macro (strategic) level, this chapter explored them at the micro (policy) level, in the US decision to adopt the POW voluntary repatriation policy. This policy single-handedly held up armistice negotiations for 15 months, and thus was an important part of the “stickiness” in ending the war. When Truman adopted the POW policy, it embodied his view that ethical action was integral for combating Communism’s moral and spiritual dangers. Yet, he adopted the policy without fully considering its second- and third-order effects. Subsequently, entrapment dynamics led his governing coalition to become so invested in the policy that some coalition members systematically ignored information suggesting it would conflict with other policy preferences—such as ending the war. By the time the Truman governing coalition became aware of the POW policy’s full implications, it was too late to back down. Beleaguered by partisan attacks and domestic political crises—many resulting from the Cold War mobilization effort—Truman could neither abandon the POW policy nor escalate the war to bring the Communists to terms. It was only with a domestic governing coalition shift—the election of President Eisenhower—that entrapment and information obstacles were lifted and the United States could change course towards ending the war. While Eisenhower retained the same war aims that Truman had espoused, his governing coalition brought different actors—with different constituencies, different worldviews and different calculations of the war’s costs and benefits—that left them freer to consider an expanded range of strategic options. Without the partisan attacks from the Republican right wing, the politically unpopular mobilization campaign and the constraints of the NSC-68 mindset, the Eisenhower coalition had more leeway to end the war. Thus, the peace Eisenhower eventually won was “indistinguishable” from

the united states, part 2



what Truman could have probably garnered, “except that it was achieved without the accompanying howls of the GOP” (Reinhard 1983: 107). Eisenhower could agree to terms that Truman could not, because he was less vulnerable to attacks from the Republican right wing and could placate hardliner critics of the armistice in other issue areas, by eliminating economic controls and dismantling Truman’s mobilization infrastructure. Finally, because Eisenhower had built up the expectation during his campaign that he would end the war soon, the public both looked for and expected decisive action. As long as he settled the war relatively quickly, it would remain “Truman’s War.”

7

China Trapped by a Hawkish Ally

This chapter returns to the Communist side to trace Chinese decision-making through the lens of entrapment. As the US chapters demonstrated, an entrapment obstacle occurs when leaders want to end the war but cannot. Unlike the internal entrapment in the US case, however, the Chinese governing coalition was trapped by an external ally—Stalin. While the Soviet Union and the People’s Republic of China (PRC) shared Communist roots, their relationship was hardly intimate as both tried to maneuver to their advantage. As Chapter 4 detailed, Soviet behavior towards the Chinese Communists before the Korean War certainly left reason for the Chinese to be wary, and thus it is not surprising they were looking to avoid dependence on the Soviet Union. Yet, to some degree, it was precisely because of the Korean War that such dependence developed and ultimately trapped China in a war it no longer wished to fight. This chapter argues that although opponents to the Korean War initially existed in the Chinese governing coalition, ultimately the Chinese leaders reached a consensus to intervene. But the war effort was very costly, both in economic and human terms. As the war’s domestic costs became apparent, the Chinese governing coalition experienced a dovish shift in the summer of 1952, which led to a policy shift towards wanting to end the war. Although it appears China was willing to settle in August 1952, it was prevented from settling by its ally and patron, Stalin. The new governing coalition tried to convince Stalin to support crucial concessions about voluntary repatriation to achieve a negotiated settlement, but Stalin was unwilling. In large part because of the war, China was now extremely dependent upon Soviet military and economic assistance for both fighting the

china



war and developing its own industrial base. As a result, China was trapped: unable to pursue a real war-ending strategy until the Soviets gave it permission. It was not until after Stalin’s death that China was finally able to make the POW policy concessions it had been advocating since the previous summer. Admittedly, we do not yet have any direct evidence of Mao’s views in 1952. Where the process tracing is inconclusive, because data is unavailable, we can only make provisional conclusions. Nonetheless, evidence shows the economic elites whose influence increased with the summer 1952 coalition shift clearly supported an early end to the war. Similarly, Foreign Minister Zhou Enlai evidently discussed war-terminating options with Stalin that fall. While some scholars argue Mao wanted to continue fighting as late as 1953, these same scholars argue foreign policy outcomes never diverged from Mao’s preferences. Therefore, while we lack a “smoking gun,” it is reasonable to argue—given the evidence available today and presented in this chapter—the dovish coalition shift in the summer of 1952 helped shift Chinese war policy towards ending the war. This chapter is divided into five sections. The first section explains the Chinese governing coalition. The second section explains its decision to intervene in the Korean War. The third section describes Chinese mobilization efforts and outlines the staggering domestic costs of its intervention, as well as the shift in the governing coalition that spurred China to want to end the war. The fourth section details China’s entrapment obstacle as Chinese leaders unsuccessfully tried to get Stalin to support concessions that would bring the armistice in reach. Finally, as in the previous case studies, the fifth section examines some alternative explanations for Chinese behavior.

The Chinese Governing Coalition From 1950 to 1953, the Chinese Communist Party (CCP) shared a “basic leadership consensus” to consolidate power and recover economically, thereby leading to a socialist transformation and planned economy. Led by Mao Zedong, the regime had a predominantly stable and unified leadership; during the war there were no leadership purges.1 From 1945 to 1953, the five-man Secretariat was clearly ranked Mao, Liu Shaoqi, Zhou Enlai, Zhu De and by 1950, Chen Yun (Teiwes 1990: 18; Chang 1975: 204). This leader stability created consistent expectations lower down the hierarchy. There were four reasons for this essential unity in leadership and decision-making. First, there was substantial elite consensus about the need to consolidate the new regime and the desire to follow the Soviet model. When the CCP created the People’s Republic in October 1949, it faced some serious challenges. The Chinese economy

180

the korean war

and financial structure were in terrible disrepair from the 12-year civil war and a rash of floods and droughts that affected about 40 million people. The CCP leadership’s first efforts were therefore directed at relieving famine, stabilizing commodity prices, checking inflation, balancing the budget and restoring basic transportation and communications (Gurtov and Hwang 1980: 27; Teiwes 1993: 86). Second, the leadership shared a generational worldview about foreign and domestic policy. The elites believed China was unique and important in international relations, but worried that foreign and internal enemies were trying to subvert the fragile regime. Thus, they wanted to exert strong state control, limit partnerships and maintain distance in foreign relationships. In short, they viewed foreign and domestic threats—and foreign and domestic policy—as indelibly intertwined (Hamrin 1994: 80; Kirby 1994: 13–29; Gurtov and Hwang 1980: 17). Third, Mao’s leadership of the CCP was undisputed. Mao’s prestige was enormous, in large part because of his crucial role in leading the Chinese Communists to win the civil war and establish the PRC. This status was amplified by the authoritarian traditions of the international Communist movement and Confucian beliefs, which translated into a political culture of submissiveness at all levels of the CCP. Many cadres—particularly from rural areas—adopted a deeply reverential attitude of “loyalty to the monarch” (Teiwes 1993: lxiii). Finally, because of his preeminent position, Mao could afford to govern through the principle of collective leadership. He used his prestige to persuade others to support his views and build a consensus. While reserving the right to insist on his own way, Mao’s general approach was to encourage broad discussion in order to reach a consensus position. It is true that group norms to avoid conflict encouraged leaders to take leaves of absence (i.e. illness, holidays) rather than formally dissent. Nonetheless, Mao avoided harsh rebukes of dissenters and—especially in areas such as industrial policy, where he acknowledged his own inadequacies—he would concede to contrary views favored by a majority, despite his own reservations. In turn, this allowed him to play the final arbiter of policy disputes: by pursuing consensus politics, Mao could secure the majorities required by the principle of collective leadership (Teiwes 1990: 17–20; 1993: lxv, 86–87; Chang 1975: 188–89).2 For these reasons, the CCP Politburo and its Standing Committee were more involved in decision-making than has been popularly understood. These bodies approved all decisions, at least formally. When policy disputes arose, they usually appeared because different leaders with different portfolios maintained divergent perspectives. As Parris Chang argues, “Despite Mao’s enormous power, policy was significantly influenced by debates and conflicts among leaders” (1975: 2). When top leaders appeared divided, Mao would use alternative bodies—enlarged

china



Politburo sessions, conferences and enlarged Central Committee plenums with members and nonmembers—to make major policy decisions (Chang 1975: 184; Jung and Halliday 2006: 356). For example, Mao resorted to such fora during the debates before China’s intervention in the Korean War. Under Mao’s influence, the CCP apparatus and generalists (rather than “ex­perts”—i.e. knowledgeable officials and professionals) took charge of the policy process. They implemented policies through mass campaigns rather than through functional bureaucratic agencies, using state institutions as a “transmission belt” to cloak decisions with legality. This “campaign approach” allowed CCP leaders to mobilize and concentrate efforts on achieving a defined objective at a given time, breaking through bureaucratization tendencies. However, this approach inadvertently created problems, because top leaders could not always control the forces they unleashed. As the domestic mobilization campaigns during the Korean War demonstrated, lower-level cadres tended towards excess in implementing policies, often taking a policy line too far (Chang 1975: 180–85; Jung and Halliday 2006: 318–23). Although the CCP worked through collective leadership, power was concentrated among a few people and the regime was run in a highly centralized fashion. Therefore, “ultimate power and decision-making were concentrated in the hands of a small group of leaders and were subordinated to the changing results of factional struggles. Political issues, although they did exist, often were obscured by factional and personal alignments” (Bergere 1997: 268). These personal alignments influenced the timing and shape of policy implementation. For example, they were important in the summer of 1952 in helping the new coalition to win over Mao to change China’s Korean War policy. There were two factional groups in the CCP decision-making structure. The first group was military elites, who during the regime’s formative years enjoyed great power and a high degree of regional autonomy. During the civil war, when People’s Liberation Army (PLA) troops took over an area, they abolished the Chinese Nationalist (Kuomintang, or KMT) government institutions and set up military control councils—usually led by local PLA commanders—as provisional organs of the new regime. In early 1950, these councils were synthesized into five Great Regions with parallel party, military and administrative structures run by the same leaders.3 PLA generals also controlled many of the regional and municipal administrations. In addition, the military elites comprised half of the CCP Central Committee in 1950 and had three seats on the Politburo—significantly higher than Red Army representation in the early Soviet regime.4 Finally, PLA leaders were highly represented in party and state bureaucracies (Fang 1998: 60–61; Thornton 1982: 231–32).

182

the korean war

The second group was civilian elites, who had been leaders of the underground party organizations in the formerly KMT- or Japanese-controlled “white areas.”5 Civilian elites were generally better educated, and therefore they assumed high-level administrative positions when the Communist regime came to power. For example, Liu Shaoqi was appointed vice chairman of the Central People’s Government Council, and Zhou Enlai became foreign minister and the premier of the Government Administration Council, or GAC (Fang 1998: 62–63). Because Mao had been part of the military faction, he deliberately populated the CCP Central Committee with civilian elites in 1945, to reflect the diverse constituencies of the party (Teiwes 1990: 15). Nonetheless, these factions served to reinforce the centralized, unified leadership, for three reasons. First, because of the long civil war, the CCP’s party and military organizations were quite intertwined, so that there was a blurring between civil and military bureaucratic interests. Second, because the regime was not yet institutionalized, informal ties were very important. Most leaders had been soldiers during the civil war and did not become civilians until after the PRC’s creation in 1949. Consequently, their informal networks cut across bureaucratic divisions and fragmented bureaucratic authority. They reached consensus through bargaining and personal ties (guanxi). Third, as with the Soviets, CCP elites feared political purges and persecutions, although no purges actually occurred during this period. Thus, while factional support helped ensure political and career survival in a system without reliable legal procedures, even the factions would only dissent from the group consensus so much (Fang 1998: 11–12; Bergere 1997: 268). Turning more specifically to decision-making during the Korean War, foreign policy decisions were highly centralized and compartmentalized from other functional areas. Mao was actively involved in guiding foreign policy and reunification policy (involving Taiwan and Hong Kong), in directing relations with the United States and the Soviet Union and in addressing regional crises. Mao alone received foreign ministry papers from Zhou Enlai for decisions, although Liu Shaoqi, Deng Xiaoping and Peng Zhen received informational copies. Likewise, Mao remained the final arbiter of Chinese strategy in Korea. Peng Dehuai, the Chinese military commander in Korea, regularly submitted his battle plans to Mao for approval. In some cases, Mao himself wrote directives dictating the scope and objectives of operations (Chen 1992: 28; Pollack 1989: 223; Hamrin 1994: 80, 83; Li and Tracy 1992; Li et al. 1992: 63–86). Mao also placed his son, Anying—killed in the war—on Peng’s staff in Korea as a Russian interpreter, in part to watch over Peng’s actions (Zhang 1993: 15).

china



The Chinese Decision to Intervene in the Korean War Why did the Chinese decide to intervene in the Korean War? China was still recovering from the civil war, faced a long list of domestic challenges and was not well situated for war with the most powerful country in the world. Moreover, at least initially, most Chinese leaders were against intervention. However, the US decision to cross the 38th Parallel, coupled with Stalin’s threat to let North Korea completely collapse (right on China’s border), triggered Chinese fears about its internal security. Ultimately, the Chinese governing coalition decided the Korean War was inextricably linked to China’s internal security and thus intervention was necessary. As its first priority, the Chinese governing coalition wanted to protect the PRC from external threats to concentrate on establishing the new regime. Most importantly, it wanted to end the civil war by capturing Taiwan, where the KMT had escaped, and it needed Soviet support to do it. Therefore, in June 1949, Mao reluctantly argued for aligning with the Soviet Union. In his words, “we are keenly aware that we must lean to one side in order to achieve and consolidate victory” (quoted in Hao and Zhai 1993: 145). Meetings between Chinese and Soviet leaders that fall and winter, including Mao’s two-month trip to Moscow, culminated on February 14, 1950, with the “Sino-Soviet Treaty of Friendship, Alliance and Mutual Assistance” (Goncharev et al. 1993: 76–129; Chen 1989: 187). Although Stalin snubbed Mao during the tense visit,6 and the agreement’s terms disproportionately benefited the Soviets,7 the Chinese Communists were willing to “secure their flank” to focus on internal issues—the first one being a Taiwan invasion. While Mao planned the Taiwan invasion in early 1950, Stalin tried to preempt Mao’s attack by hastening Kim Il-Sung’s invasion. As Chapter 4 explained, if Kim invaded South Korea before Mao started his attack, Mao would be unable to invade Taiwan. Thus, Chinese leaders were “deliberately kept in the dark on the details of the North’s preparations, including the timing of the attack” (Goncharev et al. 1993: 213; also Thornton 2000: 110–13, 229; Wolff 2000: 25; Kim 1996: 264; Hao and Zhai 1993: 148; Spence 1990: 526; Zhai 1993: 178; Gaddis 1997: 75).8 The outbreak of the Korean War was a fatal blow to Mao’s plans for Taiwan. The Chinese leaders backed the North Korean invasion in principle and knew they could not refuse to support Kim in light of their treaty with the Soviets. However, Mao explicitly warned Kim about possible US or Japanese invasion (Gaddis 1997: 78; Zhang 1992: 93; Weathersby 2002: 13; Goncharev et al. 1993: 214). US ground intervention in South Korea, increased military aid to Indochina and the Philippines, and the US Seventh Fleet’s deployment into the Taiwan

184

the korean war

Strait triggered Chinese fears that the United States would intervene in China’s internal affairs. Given their beliefs equating internal and external threats, it is not surprising the Chinese leaders linked American “imperialism” in Asia with support for the “reactionary” KMT that still threatened the PRC’s internal stability. Facing potential conflict now in two directions at once—Taiwan and Korea—the Chinese were worried that North Korea would be defeated, allowing the United States to control the entire peninsula and then attack China (Chen 1994: chapter 4; Milliken 2001: 114–17; Christensen 1996: 157–60; Goncharev et al. 1993: 215; Hao and Zhai 1993: 101, 149). Thus, a few weeks after the war began, China prepared—on its own initiative—for the possibility of intervention on the Korean peninsula. They dispatched Chinese diplomats and intelligence officials to North Korea to establish better communications with Kim Il-Sung and to observe the fighting directly. They redeployed and reflagged more than 250,000 troops, who became the backbone of the “Chinese People’s Volunteers” (CPV)—the Chinese army in Korea. Chinese military commanders stationed air defense units south of the Yalu River to defend major bridges against possible US air attack. Finally, the CCP launched the first of its mass mobilization campaigns during the war, the “Resist America, Aid Korea” campaign, with “beating American arrogance” as its central slogan (Chen 2001: 11, 88; Pollack 1989: 216; Nie 1988: 748–49; Stueck 1995: 52, 65; Jung and Halliday 2006: 318). chinese reluctance to intervene

Despite these preparations, it was not a foregone conclusion that China would inevitably intervene. Foreign Minister Zhou Enlai told Stalin in July that Chinese forces would have to enter the war if the United States attacked north of the 38th Parallel. Chinese officials gave three warnings between September 26 and October 3 to try to forestall this event.9 China genuinely sought through diplomatic means to prevent the United States from invading North Korean territory (Milliken 2001: 144–45, 150; Sartori 2005: 22–25; Zhang 1992: 94–95; Goncharev et al. 1993: 175).10 Indeed, at least initially, the majority of the Chinese governing coalition was against intervention. Between September 30, when the North Koreans approached the CCP for assistance,11 and China’s final decision to intervene on October 18, there was a flurry of deliberations among the Chinese leadership. The CCP Politburo held several meetings to debate the issue,12 and even Mao vacillated.13 The reasons cited against intervention need describing in detail, as these factors influenced how coalition members assessed the war’s costs and benefits and assumed greater importance when the governing coalition shifted in the summer of 1952.

china



Most coalition members who participated in the early October debates believed it was unwise to send troops to Korea.14 Many leaders were logical opponents to intervention. For example, GAC Premier and Foreign Minister Zhou Enlai and Finance Minister Chen Yun had designed a policy to balance the budget and control inflation by curtailing military and administrative expenses; Zhou even led the PLA demobilization effort. Rao Shushi presided over the tentatively recovering East China region that owed its improved economic circumstances to stable relations with Western industrialists and merchants. Similarly, Northeast region leader Gao Gang did not want war, because his region would become the staging base for the Chinese military intervention (Weiss 1981: 82; Huo 1989: 224). (Conversely, as will be seen below, some scholars argue Mao supported the war precisely for this reason, as it would break up Gao’s political monopoly in Manchuria.) The opponents to intervention listed four reasons. First, intervening in Korea would damage Chinese reconstruction. After decades of destructive war, China needed rehabilitation to deal with its huge fiscal deficit, high unemployment and agricultural and industrial recovery. Moreover, sending troops into Korea might cause discontent at home (Milliken 2001: 151; Keith 1989: 46; Gurtov and Hwang 1980: 55; Weiss 1981: 81; Huo 1989: 218). As Central-South region leader Lin Biao argued, the consequences of going to Korea after the debilitating civil war could be even worse than “igniting a fire to burn oneself ” (quoted in Zhang 1993: 7). Second, many outlying areas and offshore islands (including Taiwan) were not yet liberated, with one million KMT troops and bandits still moving freely about. These forces harassed and raided local councils, sabotaged economic projects, assassinated officials and resorted to guerilla warfare. In addition, some leaders worried a Korean intervention would divert scarce resources flagged for the Taiwan invasion and thereby postpone it (Paul 1994: 103; Gurtov and Hwang 1980: 31, 55; Hao and Zhai 1993: 154; Zhai 1993: 189; Milliken 2001: 151; Chen 1994: 182). Third, American war resources seemed inexhaustible, in terms of both military power and industrial reserves. Opponents pointed to US air supremacy and superior weaponry, and they argued the United States might bomb China’s cities and destroy its industrial base. Moreover, because the Chinese lacked air cover or control of the seas, they would have to fight without protected logistical supply trains. Even if the PLA was numerically superior, most of its forces were pinned down with domestic counter-insurgency efforts (Huo 1989: 215; Zhai 1993: 188; Zhang 1995: 80–81; Hao and Zhai 1993: 154; Jung and Halliday 2006: 356; Chen 1994: 182, 185; Milliken 2001: 152). Finally, after decades of war, many PLA cadres and soldiers were longing for peace and unwilling to fight in another major operation. The PLA had not taken

186

the korean war

“proper leave” or conducted training exercises in more than four years, and commanders were worried about getting bogged down in another protracted war (Zhang 1993: 11; Hsiung 1970: 172; Gurtov and Hwang 1980: 55; Zhai 1993: 189; Nie 1988: 735; Huo 1989: 214–15). In contrast, Mao was inclined to enter the war. Mao’s “People’s War” strategy for protracted war, with “man” beating “weapons,” had been very successful against the KMT, and he believed weak forces could defeat materially stronger enemies. In the policy debates, Mao emphasized conflict with the United States was inevitable and of the possible battlegrounds—Indochina, Taiwan or Korea—Korea was geographically and logistically best suited. Moreover, China’s entry into the war would not only strengthen the CCP’s control over the Chinese state and society, but allow China to reassert itself internationally (Zhang 1992: 89; 1995: 23, 29, 85; Sheng 1995: 301; Hao and Zhai 1993: 155; Paul 1994: 94; Stueck 1995: 99; Milliken 2001: 160–61; Chen 1994: 211–23).15 More parochially, while not a firstorder issue for Mao, other scholars suggest Mao favored military intervention to undermine Gao’s “independent kingdom” in Manchuria. Given his long-standing ties to Stalin, Gao was one of Mao’s rivals; the Korean War would allow Mao to move large numbers of troops through the region and thus break Gao’s military monopoly there (Petrov 1994a: 10; Fang 1998: 64; Gittings 1967: chapter 6; Schur­ mann 1966: 267–72; Whitson and Chen-Hsia 1973: 94).16 While Chen (2001: 90) characterizes Mao as unshakably determined to intervene in Korea, other scholars and participants depict him in a far less decisive light. In most accounts, Mao nervously deliberated for three days and nights before deciding there was no alternative to Chinese intervention (Noriyuki 1993: 164; Hao and Zhai 1993: 161; Gittings 1974: 183; Pollack 1989: 219; Zhai 1993: 199; Usov 1993: 251; Gurtov and Hwang 1980: 55; Zhang 1995: 78; Wang 1975: 206–7; Weiss 1981: 83; Sheng 1995: 304; Li et al. 1992: 73–74). The triggering events appear to have been the US decision to cross the 38th Parallel on October 7 and Stalin’s threat to order North Koreans to evacuate into Manchuria (Christensen 1996: 162, 164; Mansourov 1995–96: 114–15). The day after the UN Command crossed north, Chinese leaders decided to enter the war, and Mao established the CPV force (Li et al. 1992: 68–69; Zhang 1993: 17). However, in a meeting with Zhou on October 10, Stalin reneged on his July promise to provide air support. Stalin believed China was delaying intervention deliberately to get more Soviet assistance. Thus, he called China’s bluff by threatening to give the Korean peninsula entirely over to US control and order the North Korean People’s Army (NKPA) to evacuate into Manchuria (Thornton 2000: 342; Goncharev et al. 1993: 189–92; Gaddis 1997: 80–81; Jung and Halliday 2006: 358).17 After Stalin withdrew Soviet air support, Mao reversed the decision to inter-

china



vene and hosted another round of debate. Nonetheless, Chinese leaders finally agreed that even without Soviet air support, intervention was necessary. Linking foreign and domestic threats, they worried about the physical security of the new regime. In their view, the Korean War was inextricably linked to China’s internal security, because it would increase “reactionary” uprisings at home and encourage the KMT to fight back to the mainland. In other words, they worried about perceptions of China’s weakness and believed that China needed to appear strong to deter these threats. Moreover, should China stand by and let North Korea be defeated, China would face a hostile great power on its border and thousands of North Korean troops and their weapons streaming into Manchuria (Thornton 2000: 343; Zhang 1993: 23; Hao and Zhai 1993: 152; Zhai 1993: 185–86; Paul 1994: 92–93; Milliken 2001: 151). As a result, they “had to proceed even without Soviet air cover in order to save, not North Korea, but China” (Goncharev et al. 1993: 216). Thus, on October 18, Mao finally ordered Chinese troops to cross the Yalu River and intervene.

The New Chinese Governing Coalition and China’s Willingness to Settle The previous section outlined four reasons why many Chinese leaders were against intervening in Korea: (1) intervention would harm Chinese reconstruction efforts and might incite domestic discontent, (2) the homeland was still not liberated, with roughly a million KMT troops and bandits still moving freely about and the KMT still controlling Taiwan, (3) the Chinese military was technologically inferior to the US military, and the United States had more industrial resources to mobilize for a long war and (4) the PLA was exhausted from decades of fighting and unwilling to fight another major war. These four factors did not magically go away after Chinese troops crossed the Yalu River. Instead, as this section demonstrates, these factors persisted and assumed greater importance when the composition of the governing coalition shifted in the summer of 1952. chinese intervent ion

While the initial Chinese counterattack was devastating and UN forces fell back quickly, within a few weeks the Chinese forces were feeling the toll of heavy fighting, inferior equipment and logistical problems. By December 1950, CPV Commander Peng Dehuai wanted to halt for lengthy preparations before renewing the offensive. Peng was concerned about logistical supply lines stretching down the peninsula and lack of air cover for Communist troops. For example, by the time the CPV reached the 38th Parallel, its closest logistical depot was more

188

the korean war

than 190 miles away. These depots had no radio communication capabilities and were staffed primarily with soldiers lacking logistical experience. Moreover, transportation was inefficient and cumbersome. Rail was mostly destroyed, so Chinese trucks carried supplies after dark, often with headlights off, traveling on average 20 miles per night. CPV forces on the western front had fewer than 300 trucks for almost 300,000 troops. Chinese soldiers also lacked basic supplies, like shoes, ammunition and food (Stueck 1995: 358; Jung and Halliday 2006: 360; Zhang 1995: 167–68; Yu 2001: 17).18 The Chinese fought with “human-wave tactics,” to exploit their superior numbers, but casualties were extremely heavy. Prolonged and severe combat weakened CPV morale, while the number of political cadres to enforce discipline was waning. Chinese front-line disintegration manifested itself in uncontrolled retreat, desertions and group surrenders. Despite a trip back to Beijing, to explain the situation in person, Peng was overruled in his desire to slow the offensive. While Mao revised his estimate of the war’s length upwards to two years and developed a new policy of “rotational warfare,” he wanted to press on (George 1969b: 10, 167; Chen 1992: 34; Jung and Halliday 2006: 360; Zhang 1992: 112; Chen 1992: 34–35; Mastny 1996: 121; Stueck 1995: 177; Sheng 1995: 312). The Communists prepared for a large-scale offensive in April-May 1951 with twelve armies, which they believed would be the campaign to determine the war’s fate. How right they were. The “fifth campaign,” as the Chinese called it, lasted 50 days; the CPV suffered 85,000 casualties (wounded and killed). Moreover, during the campaign’s last days, UN forces encircled the CPV’s 180th Division, 60th Army, and the majority of the unit—about 5,000 soldiers—was captured (Yue 1998: sections 16–17; Zhang 1995: 152). CPV commanders called the 180th Division’s capture “the most serious loss of the CPV on the Korean battlefield” (quoted in Chen 1992: 38). When the fifth campaign failed to push UN forces further south, the Communists decided to pursue a negotiated settlement. As a result, by June 1951, the Chinese coalition’s new objective for the war became restoring Korea’s physical and political prewar status quo. The Chinese governing coalition formally agreed to a new policy called “negotiating while fighting,” which Peng described as “painting a picture with two brushes at the same time” (Zhang 1992: 129; Nie 1988: 641; Mastny 1996: 124; Chen 1992: 39; Paul 1994: 95).19 At the same time, the CCP republished Mao’s famous 1938 article “On Protracted War” (Zhang 1992: 112). Thus, until the summer of 1952, the Chinese viewed armistice negotiations in Korea as one tactic in the ongoing struggle with the United States that would long outlast the Korean conflict. In other words, armistice negotiations were a tactical maneuver, not a strategic policy shift, to alleviate the burden of fighting in Korea.

china



mobilizat ion for war

When China intervened in Korea, the new Chinese regime had accomplished little with its reconstruction programs. The nation desperately needed land reform, industrial and agricultural growth, a functioning currency, legal policies and dependable social programs. The CCP wanted to respond, but a number of the gradual reforms planned in 1949 were either scratched or rushed into implementation because of conflicting priorities with the war. The five campaigns of the Chinese offensive until June 1951 and the subsequent attrition strategy of “negotiating while fighting” required great sacrifice for a nation in need of reconstruction and rest. Initially, the Chinese governing coalition took several actions to mobilize the populace and spin up support for the expected hardships ahead. As Chapter 2 suggested, during the initial months of war, leaders will often take steps—such as demonizing the enemy and mobilizing the armed forces and industry—to ease the process of waging war. The Chinese governing coalition took several such actions to wage war more effectively. In the process, they also “sought to solidify CCP control by eliminating independent power and prestige of old elites” (Teiwes 1993: 85). First, Mao and Peng very carefully managed information reaching the wider public. For example, the day after issuing orders for CPV troops to cross the Yalu River, Mao instructed the five regions’ political commissars that “no open propaganda about what we are doing should appear in our newspapers; only highranking cadres of the party will be notified of the actions undertaken, so that they can be prepared in planning their work arrangements.”20 Similarly, in a telegram to the director of the General Information Bureau and Xinhua News Agency, Mao ordered news articles to be based on Xinhua’s own field reporting or on “internal information in Beijing.” Reports received from the Soviet news agency “should not be transmitted at all.”21 In addition, reports from foreign news agencies could not be published in Cankao Xiaoxi—an internally circulated paper carrying foreign news reporting—for four to five days.22 Mao also drafted press releases himself, praising the Communist troops and demonizing the enemy (Stueck 1995: 144). For example, in a People’s Daily report published December 5, Mao wrote that the NKPA and the CPV have struck devastating blows at the enemy forces. . . . The enemy forces on both the eastern and western fronts are running for their life with extreme anxiety and in great panic. The enemy forces inside the city of Pyongyang are committing monstrous crimes, slaughtering the people and burning down the city electric power plants and supply warehouses. (quoted Li and Tracy 1992: 79–80)

Second, the Chinese government launched a comprehensive effort to mobilize war-weary CPV troops and prepare them for combat. For two months before the

190

the korean war

intervention, they primed troops with ideological education, focusing on China’s Communist kinship with North Korea, China’s just cause and the “paper tiger” nature of the American imperialists. As General Du Ping, who served as the director of the CPV’s political departments, explained, they taught the troops that China and North Korea were as closely related as lips and teeth. If the lips [China’s neighbor] were gone, the teeth [China itself] would get cold. To help the neighbor was to help ourselves. We did not set fire to ourselves, but it was the enemy who had already set flames of war to our doorway. This war must be fought; the American imperialists were forcing us. (2001: 68)

Soldiers were also briefed by former KMT soldiers who had fought with Americans in Burma during World War II, to prove US troops were not invincible and could be defeated. Du argues this indoctrination campaign “established a ‘dare-to-fight’ and ‘must-win’ faith” among the troops (2001: 68–69). While initial indoctrination did not focus enough on the “unfavorable conditions” the Chinese soldiers found in Korea, the ideological education was continually refined throughout the war. For example, starting in January 1951, Du’s organization published Zhiyuanjun Bao [Volunteers Newspaper], to “strengthen both patriotism and internationalism” and “disseminate ideas of hate, spite and condescension toward the American imperialists so as to strengthen our soldiers’ confidence in winning the war” (Du 2001: 91–92, emphasis added). Third, the Chinese government set the economy on a wartime footing in order to collect—sometimes by force—necessary war resources. Waging the war was expensive: the Chinese force consumed 5.6 million tons of goods and supplies during its intervention, including the loss of 399 airplanes and 13,000 vehicles. To gather the necessary revenue and supplies, the CCP implemented several mass mobilization campaigns. In January 1951, the CCP launched a donation campaign to purchase “gifts” for the CPV, but it only collected 5 trillion yuan—onetenth of actual Korean War costs for that year. It also stimulated savings through government bonds by deliberately keeping consumer goods in short supply. A few months later, the CCP started a “patriotic bumper crops” movement that required all peasants to increase agricultural production. They also took emergency measures to monopolize grain purchase and marketing—a first step towards a central planning system. The middle class was hardest hit by these taxation and donation campaigns. For example, almost 70 percent of the total collected in the Shanghai area was given by “industry and commerce,” the euphemism for bourgeois groups (Li et al. 2001: 6; Hamrin 1994: 85; Weiss 1981: 180; Spence 1990: 533–35, 545; Zhang 1995: 200–202).

china



Much of this mobilization occurred through mass campaigns. Since October 1950, the Chinese “Resist America—Aid Korea” campaign had blanketed urban centers with its “beat American arrogance” slogan. In November 1951, this campaign spread to rural areas, calling for “patriotic pacts” to raise production and establishing special efforts to care for families of CPV servicemen. During that winter, the CCP also launched two campaigns to ameliorate economic problems and internal disorder: the Three Anti Campaign (opposing corruption, waste and obstructionist bureaucracy) and the Five Anti Campaign (opposing bribery, tax evasion, theft of state property, cheating on government contracts and stealing state economic information).23 Rearmament efforts increased further in January 1952, when the CCP proclaimed a policy of “total mobilization” with the goal of “winning victory, resisting the United States and aiding Korea.” This New Year’s proclamation called for agricultural collectivization, industrial production increases, an intelligentsia remolding campaign and a widening of the Three Anti-Five Anti campaigns (Teiwes 1993: 85; Chang 1975: 209; Zhang 1992: 131; 1995: 200–202, 223; Spence 1990: 533–37; Weiss 1981: 94–95, 101–3, 149; Stueck 1995: 275). To spur the populace on, Chinese leaders blatantly resorted to demonizing the enemy beyond its “beat American arrogance” slogan. In October 1950, to whip up anti-American sentiment, the People’s Daily alleged the United States had ordered a foiled assassination plot against Mao.24 More significantly, in February 1952, the CCP announced the United States had used bacteriological weapons over North Korea and Northeastern China—a propaganda campaign that recently declassified Russian archival documents demonstrate was a Communist hoax (Leitenberg 1999; Weathersby 1999; Jung and Halliday 2006: 369–70). This movement against “germ warfare” was the most intensely xenophobic campaign of the war. The CCP staged mass demonstrations against the “US criminal acts of spreading disease germs” and organized a “people’s investigation group” to collect evidence in North Korea. They also launched a “patriotic health and epidemic prevention campaign” (Zhang 1995: 183–85).25 As Shu Guang Zhang argues about these mass campaigns, CCP leaders “worried that they would not be able to sustain such a patriotic fever through the end of the war, given that KMT remnants were not yet eliminated, the future of Tibet was not settled, and pro-American elements within and outside the Party were not totally suppressed” (1992: 131, emphasis added). In sum, given how costly the war effort was and how much the Chinese people needed a rest after decades of war, it is not surprising the Chinese governing coalition took such actions to mobilize the nation to wage war more effectively.

192

the korean war

the d omest ic costs of war

At their height, these domestic mobilization campaigns were ruthless and extreme; roughly three million people died as a result. In Beijing alone, some 30,000 sentencing and execution rallies were held, attended by nearly 3.4 million people. Likewise, factory workers held mass tribunals to denounce and punish their corrupt capitalist employers, some of whom were killed for their crimes, and peasants attacked ordinary landlords when genuine despots did not exist (Schram 1976: 247; Jung and Halliday 2006: 318–19; Weiss 1981: 125; Spence 1990: 535–39). While excessive violence and killing were not the official policy, “there was tolerance for such expressions of the people’s outrage in view of their own long suffering, and a reluctance to ‘blame them too much’ if they went too far” (Shue 1975: 17). The governing coalition preferred to let the campaign go too far and then rein it in. As a result, throughout most of 1951, a real climate of terror reigned throughout the country. Certainly, these mobilization campaigns were only partly to extract materiel support for the Korean War. They were also a systematic attempt to consolidate the Communist regime by undermining armed resistance in the countryside and inducing long-term social order. Yet, at least through the “rumor” mill—how most Chinese expressed their real sentiments, given the preference falsification that was rampant after such terror—the campaigns were actively linked to a perception of Russian extortion for the war effort (Jung and Halliday 2006: 320). The other human cost to the war’s mobilization was within the military, although away at the front, they were less visible to the Chinese people. More than three million Chinese participated in the Korean War, the largest foreign war in Chinese military history. According to Chinese statistics, a total of 2.3 million Chinese soldiers fought in the war. As many as 25 infantry armies (73 percent of the PLA’s infantry forces), 16 artillery divisions (67 percent), 10 armored divisions (100 percent), 12 air force divisions (52 percent) and six security guard divisions fought in Korea. In addition, more than 600,000 Chinese laborers entered Korea, as part of the CPV’s logistical support services, to operate supply depots, transport troops and supplies and construct highways and railroads. Chinese casualty statistics are also sobering: 1.1 million casualties, including 152,000 dead and 4,000 missing in action. UN intelligence statisticians put Chinese losses far higher: 1.5 million casualties, with between 300,000 and 400,000 dead. An official Russian document even puts Chinese dead at 1 million (Li et al. 2001: 6, 246, FN 6; Jung and Halliday 2006: 372). Jung and Halliday argue that Mao consciously embarked on the Korean War with such a quid pro quo in mind, sacrificing Chinese lives in battle to industri-

china



alize at home. “What Mao had in mind boiled down to a deal: Chinese soldiers would fight the Americans for Stalin in exchange for Soviet technology and equipment” (2006: 352). They argue Chinese leaders repeatedly used the war to get Soviet support for building a domestic arms industry. For example, shortly after the Chinese decision to intervene, China’s navy chief was sent to ask for assistance in building the navy. In December 1950, China sent a high-level mission with a similar request for its air force, and in February 1951, Moscow agreed to build factories in China to repair and service planes. In May 1951, Mao asked Stalin to provide blueprints for weapons being used in Korea and assistance in building factories to produce them. While Stalin initially stonewalled, in October he finally agreed to transfer technology for seven kinds of small arms (Jung and Halliday 2006: 361). Although Stalin offered military aid to support the Chinese intervention, he initially stipulated the equipment could only be used for training and not on the frontlines. Thus, China entered the war using old weapons captured from the Japanese and KMT. Former CPV troops, in interviews years later, expressed resentment over how slowly modern arms arrived from the Soviet Union. They recalled using Soviet-made guns and weapons during the war but said these weapons did not arrive until late 1951. Moreover, except for artillery and combat aircraft, the CPV’s weapon supply “mainly depended upon China itself ” (Huo 1989: 78–79). In addition to military supply and logistics to support the CPV in Korea, China also provided support for the NKPA and the North Korean population. Given these many needs, the financial costs of the Korean War were staggering. In 1950, military expenses related to the war constituted 44 percent of the government’s budget; in 1951, 52 percent; and in 1952, 28 percent.26 Although military expenditures consumed a smaller fraction of the 1952 budget, that budget was larger, as economic development finally took root. Thus, the absolute magnitude of war expenses increased. Military spending and arms-related industries beyond the war effort accounted for 61 percent of the budget during this time, while education, culture and health shared only 8.2 percent (Jung and Halliday 2006: 374). Although groaning under the weight of these consistently high costs, China handled these expenses predominantly on its own. For example, while China spent 93 trillion yuan ($4.1 billion) on the Korean War during the 1950–53 fiscal years, it only received 21.5 trillion yuan ($950 million) in Soviet loans during that period.27 Moreover, the Chinese people were told that equipment from the USSR used in China’s industrialization was “Soviet aid,” implying it was a gift, but in fact it was paid for with Chinese food exports (Jung and Halliday 2006: 374).28 Although the CCP hiked taxes and found other ways to extract money from the

194

the korean war

economy, the war effort consumed whatever increases the government could squeeze from its rudimentary industrial base. For example, government revenue grew in 1951 by 56.4 trillion yuan, through a 13.6 percent increase in agricultural taxes, a 95.6 percent increase in industrial taxes and a 250 percent increase in state enterprise profits. Sadly, however, the Korean War consumed 50 trillion of it (Weiss 1981: 65). After the war, Mao admitted some of these difficulties. In September 1953, he argued that Chinese troops had gained military experience and showed that “US imperialism is not terrifying, nothing to make a fuss about.” However, he also confirmed that the “agricultural tax” to support the war effort had been “a shade on the heavy side” and that “some people raised an outcry” as a result and “wanted us to give up the war” (Mao 1977: 115–20). the new chinese gover ning coalit ion

By the spring of 1952, Chinese forces possessed sufficient arms and equipment to sustain the stalemate, but economic distortions, missed industrial production targets, domestic unrest and a famine demonstrated the inherent weaknesses in the Chinese mobilization strategy. The previous summer, CCP Vice Chairman Liu Shaoqi had balked at the consequences of industrialization and the war effort: “We cannot develop heavy industry first,” because it “consumes a tremendous amount of money with no returns . . . and the only way available to us is to raise the money by depriving our people. . . . Now people’s life is very miserable” (quoted in Jung and Halliday 2006: 377). In June 1952, Vice Premier and Finance Minister Chen Yun argued industrial stagnation at the height of the Three Anti Campaign was having adverse effects. Chen even predicted that “we will, one day, be unable to provide what we need in Korea” (quoted in Zhang 1992: 131). Regional leaders made similar comments. For example, Li Fuchun and Gao Gang argued the Three Anti-Five Anti campaigns pushed production drives far behind schedule; by June 1952, only 22.4 percent of the year’s target for capital construction in the Northeast, the most industrialized region, had actually been accomplished (Weiss 1981: 104, 127–28). In short, while the mobilization campaigns helped sustain the war, they cut away most of the economic support for domestic industrialization. Given so many opportunity costs from waging the Korean War, the Chinese governing coalition realized it was time to focus on economic reconstruction. The principal task was to build a national government to spearhead industrial development after decades of war. The CCP needed to build new economic ministries and staff them with senior leaders to direct and manage the increasingly heavy administrative workload. Thus, in the summer of 1952, a number of PLA

china



leaders were asked to relinquish their military posts, relocate to Beijing and serve full-time as central party or government officials (Fang 1998: 61). Starting that summer, five regional military leaders moved to Beijing: • Gao Gang, from the Northeast, was appointed chairman of the State Plan-

ning Commission (SPC). He also ran the Ministries of Heavy Industry, Number-One and Number-Two Machinery, Fuel Industry, Construction Engineering, Geology, Light Industry and Textile Industry. • Rao Shushi, from East China, was appointed director of the Central Committee’s Organization Department. He also ran the Labor Ministry. • Deng Zihui, from Central-South, was made first vice-chairman of three major committees—the Finance Committee, the Economic Committee and the SPC. He also ran the Ministries of Agriculture, Forestry and Water Conservancy and, later, the Central Committee’s Rural Work Department. • Deng Xiaoping, from the Southwest, became vice premier of the Government Administration Council (GAC). He also ran the Ministries of Railways, Transportation and Postal and Telecommunications. • Xi Zhongxun, from the Northwest, was appointed GAC secretary general and vice chairman of the Culture and Education Committee (Fang 1998: 62; Teiwes 1990: 21). All of these men were powerful military leaders who were serving, at the time of their transfers, as either political commissars or military commanders or both. As the list suggests, each of China’s regions was involved with the new postings. As a result, key CCP members moved to Beijing and became important members of the decision-making group.29 The new emerging leadership was concentrated institutionally and organizationally in the sphere of economic management. These rising leaders—together with Finance Minister Chen Yun, Chen’s deputy Li Fuchun and GAC leader and Foreign Minister Zhou Enlai—were a closely cooperating group of economic decision-makers.30 This group of economic decision-makers, including well-established Secretariat leaders and rising stars in the party leadership, formed the core of China’s new coalition making decisions about the war. Assembled together in Beijing, with a new mandate for economic growth and government centralization, they became increasingly influential in the Chinese governing coalition. Some—Deng Xiaoping, Deng Zihui and Xi—had not participated in the original decision to enter the war, and thus were new members of the decision-making coalition. Others—Zhou, Chen, Gao and Rao—had been included in the initial debates, but they had all opposed intervention. These leaders brought an interesting mix of personal perspectives and profes-

196

the korean war

sional expertise to the problem of the Korean War. On the one hand, having spent time in key regional leadership posts, they were very aware of the human and societal costs imposed by prolonged participation in the war. On the other, their expertise in managing economic recovery and industrialization clearly showed them the opportunity costs of waging war instead of rebuilding and modernizing the nation. By valuing centralized economic planning and industrialization more than the war effort, these leaders had a different vision of the war’s costs and benefits. Economic centralization created “new bases of power for key CCP leaders and new patterns of interaction within the top echelons of the Party and state” (Teiwes 1990: 20). Their increasing influence in Mao’s decision-making coalition helped spur China towards ending the war. Certainly, Mao remained at the helm of the government. However, having decided to shift priority to economic development, he assembled a new governing coalition with expertise in this issue area. Together, these leaders articulated arguments against fighting the war and thereby influenced Mao to change the war policy. These leaders had different judgments about the war, due to the different nature of their expertise and bureaucratic portfolios. Since the summer of 1951, Zhou Enlai had asserted the two fundamental problems facing China—“enormous financial tension” and “the acute lack of technical cadres”—were exacerbated by the Korean War. As Zhou argued, “If the war ends, then China will quickly escape from this situation . . . however, so long as the war goes on, it will be hard to solve this problem and it will be a fundamental constraint on Chinese industrialization.”31 Zhou’s perspective was shared by Agricultural Minister Deng Zihui and Finance Minister Chen Yun. As economic officials, they knew China’s economic realities better than anyone and “became more sensitive to objective limitations.” As a result, they “tended to be more cautious” about China’s ability to fight the war while building the economy (Chang 1975: 177). During this time, the CCP began planning for its first Five-Year Plan along the lines of Soviet-style centralized planning. There has been some speculation that the plan was actually written in Moscow, because it was first publicly announced there in October 1952, and only two months later mentioned publicly in Beijing (Weiss 1981: 130; Schurmann 1966: 171–72). Certainly, the Chinese followed their Soviet mentors’ techniques for rapid industrial growth: a focus on heavy industry, high rates of savings and investment, a bias towards capital-intensive methods and institutional transformations in agriculture. Moreover, thousands of Soviet technical advisors came to China to help with factory building, industrial planning, hydroelectric power development, railway network extension and urban architecture (Spence 1990: 544). As Lawrence Weiss argues, the Korean War

china



played a major role in restricting China’s strategic alternatives to reliance on the Stalinist model of industrialization and also encouraged the Chinese to move down the road paved by the USSR at a far greater speed than had been envisioned by all but the most ardent advocates of Sovietization prior to October 1950. (1981: 136–37, emphasis added)

Paradoxically, Soviet behavior towards the Chinese Communists before the war gave Chinese leaders reasons to want to avoid depending on the Soviet Union, yet to a large degree it was precisely because of the war that such dependence developed. As Robert Simmons suggests, China was caught in a paradox: she wished to end her involvement in a war that had become stalemated militarily and was a tremendous drain on her economy; yet, to exit from that war, she had to win the assent of her powerful neighbor, who was the only country equipped to help her in the about-to-be announced First Five Year Plan. (1975: 223–24)

In sum, by the summer of 1952, Chinese leaders not only had become aware of the domestic costs of the prolonged war, but they were ready to shift their focus towards industrializing and creating a centralized, planned economy. A new group of leaders—with economic expertise and a different conceptualization of the war’s costs and benefits—assembled in Beijing. Their increasing influence in Mao’s decision-making coalition helped spur China towards wanting to end the war. As a result, confronted with the pressing need to achieve dramatic progress with industrialization, Mao and his colleagues approached Stalin for massive, long-term assistance and relief from the war.

The Entrapment Obstacle: Constrained by a Hawkish Ally During August-September 1952, China’s top economic officials trekked to the Soviet Union to negotiate massive Soviet support for Chinese industrial development, while simultaneously seeking to end the costly Korean War. An armistice would naturally liberate vast amounts of domestically generated revenue for industrial development, and China was ready to divert its funds in that direction. Virtually all of China’s first-rank industrial planners—including Zhou Enlai, Chen Yun, Li Fuchun and Liu Shaoqi—spent extended periods in Moscow during the fall of 1952.32 During these visits, Zhou met with Stalin three times to discuss China’s proposed strategy for ending the war and support for industrialization. Russian transcripts of these discussions between Zhou and Stalin reveal that China was more eager to reach an armistice agreement than it had been before. Zhou’s dual

198

the korean war

aims were to achieve an armistice in Korea as quickly as possible while maximizing Soviet economic and military assistance to his ravaged country. In contrast, Stalin continued to press for extending the war, implicitly linking Soviet economic assistance to China’s continuing the war. Providing Soviet arms and economic assistance were a means to manipulate the Chinese into advancing Soviet interests (Shrader 1995: 66; Gittings 1967: 126–27). Yet both leaders never stated these goals so clearly and sometimes even seemed to be contradicting them. Warily circling the issue of a negotiated settlement, both leaders tried to avoid open disagreement without compromising their individual aims. During their first conversation on August 20, Zhou opened by noting North Korea had been willing to settle the war for some time.33 In response, Stalin lectured Zhou about the United States’ supposed military flabbiness. As Stalin sneered, “It’s been already two years and USA still has not subdued little Korea. What kind of strength is that?”34 Zhou then suggested that if the United States makes some sort of compromises, even if they are small, then they [the Communists] should accept. If America does not agree to return all POWs and proposes a smaller number, then they should accept the offer, under the condition that the question of the remaining POWs will be resolved under mediation by some neutral country, like India, or the remaining POWs transferred to this neutral country until the question is resolved.35

Stalin countered by proposing if the US were to retain a percentage of Communist POWs, the Communists should do likewise until a final solution was accepted. If this proposal was unsuccessful, then they could resort to mediation. Zhou replied by repeating Stalin’s suggestions as the first two of three alternative strategies: (1) “announce from the beginning that they will hold the same percentage of South Korean and American POWs as the percentage of North Koreans and Chinese held back by America,” (2) “resort to mediation by a neutral country” or (3) “sign an armistice agreement by putting off the POW question and resuming its discussion afterwards.” Since the first two alternatives were parroting Stalin, it seems likely that the third option—signing an armistice first and resolving the POW policy afterwards—was Zhou’s real preference. However, without reaching any decision about which of the three strategies to pursue, the discussion moved towards Soviet military assistance to China. During that discussion, Zhou argued inconsistently that “China is preparing for the possibility of another 2–3 years of war” in order to secure Stalin’s support for weapons and materiel.36 The same dynamic appeared again in their next conversation about the war on September 19.37 Stalin informed Zhou that the Soviet UN delegation would reject a Mexican proposal concerning POW exchange, since it conformed to the

china



American negotiating position. With characteristic crudeness, Stalin dismissed Chinese and Korean casualties, noting the North Koreans “lose nothing, except for their men.” In response, Zhou returned to his third strategy from August 20—delaying the POW question until after the armistice—by asking “whether comrade Stalin considers it possible to support this position.” Stalin retorted that he wanted “the return of all POWs. This also concurs with the Chinese position.” When Zhou then suggested sending the POWs to India, Stalin tried to dissuade him by cautioning that “the Americans will not want to deliver all the POWs, that they will want to keep some captives, with the intention to recruit them” as spies. Zhou persisted, trying to get a specific strategy nailed down, and suggested they “cease fire and resolve the issue of POWs later.” Zhou reminded Stalin that he also agreed with this, as long as both sides did not agree about the percentage of POWs to be withheld. But Stalin would not commit. As he ambiguously answered: “this can be considered as one of possible scenarios, but America is not likely to agree to it.”38 In other words, Zhou was unable to gain Stalin’s support for China’s interest in compromising on the POW question, by transferring non-repatriates to a neutral country. Nor did Zhou succeed with his proposal to conclude the armistice first and settle the POW question later. The inconclusive outcome of both conversations was a victory for Stalin, which may be why he agreed to equip 60 more CPV divisions and provide ammunition, artillery and new fighter jets. Zhou’s visit ended without clear agreement on a war-ending strategy. Ultimately, China’s need for Soviet economic assistance to prosecute the war and develop China’s industrial base outweighed China’s willingness to implement a warending strategy without its ally’s support. As Zhu argues, “it was Stalin’s desire to continue the war and Mao felt pressured to comply with it” (1998: 209). Or as Mastny argues, “despite the appearance of Beijing’s agreeing with Moscow in taking an intransigent position against [an armistice], Stalin was the one who stalled, while the Chinese were seeking ways to achieve it” (1996: 148). In short, China was trapped. Zhou’s negotiations with Stalin came at a time when China’s domestic situation was desperate, which certainly hindered China’s bargaining leverage. After Zhou’s return to Beijing, the full extent of the concessions that junior partner China was willing to make for Soviet assistance became clear.39 For example, while Stalin had previously agreed to return Port Arthur to China in 1952, during Zhou’s visit the terms were altered to stipulate that the Soviets would remain at Port Arthur until they signed a peace treaty with Japan (an unlikely event in the short term). Stalin even orchestrated the Chinese concession so Zhou was required to request continued Soviet presence in the port (Gurtov and Hwang 1980:

200

the korean war

56; Simmons 1975: 219). Moreover, Stalin may have required that Gao Gang—the Northeast region leader who had a special relationship with Stalin—be named head of the new State Planning Commission to supervise and implement the First Five Year Plan. Significantly, both decisions—Soviet occupation of Port Arthur and Gao Gang’s role as SPC chief—were reversed by the new Soviet governing coalition shortly after Stalin’s death (Weiss 1981: 215). the wait ing game: o ctober 1952–march 1953

After the Zhou-Stalin meetings, China tried once more to end the war. On October 31, 1952, Zhou publicly signaled Chinese interest in transferring prisoners to India or some other “neutral zone.” In a meeting with the Indian ambassador, to discuss a war-terminating Indian resolution under UN consideration, Zhou admitted some of the Chinese POWs in the UN camps “might not want to go back to China and since they were under orders of [KMT leader] Chiang Kai-Shek they could go back to him” (FRUS 1952–54, 15: 590). Moreover, Zhou expressed interest in the Indian resolution, which resembled one of the formulae he had allegedly put forward to the previous Indian ambassador in June 1952.40 Unfortunately, the Soviet UN delegation “preventively torpedoed” the idea (Mastny 1996–97: 23; 1996: 148). As it became clear the United States and other nations were coalescing around the Indian resolution, which China tentatively supported, the Soviet UN delegate stepped in and rejected the plan because it had become the “camouflage for horrible American policy” and was now “nothing but a slightly veiled American draft.” Soon after, China told India it agreed with Moscow that the resolution was unacceptable. Indian diplomats believed the Chinese had been subjected to Soviet pressure and that, had they been left to themselves, they would have endorsed the resolution (Foot 1990: 156; Stueck 1995: 303). On December 17, Mao warned Stalin that US President-elect Eisenhower was “currently carrying out preparations for military actions that will take place after he comes to power.” He informed Stalin of a new PRC effort to mobilize an additional 250,000 troops to compensate for China’s terrible war losses, and in exchange, he requested more Soviet armaments. Ten days later, Stalin replied, “it is fully possible that these plans will be changed by the Eisenhower government in the direction of less tension on the front in Korea,” but “you are acting correctly when you count on the worst and proceed from the probability of an attack by the Americans.” Then, “taking into account the situation you speak of in your telegram,” Stalin offered: “with great difficulty we can deliver to you in 1953, besides the arms and ammunitions for 20 divisions [already] earmarked . . . approximately one-fourth of the quantity you stated in your telegram.”41 By

china



advising the Chinese to prepare for a new US attack, Stalin again argued for prolonging the war—while only offering to provide one quarter of the arms Mao had requested. There was no movement in the Communist negotiating position until after Stalin’s death on March 5, 1953. His death prompted a radical change in the Soviet approach to the Korean War, to China’s and North Korea’s profound relief. On March 19—less than ten days after Stalin’s funeral—the Soviet Council of Ministers adopted a resolution to end the war, with attached letters to Mao and Kim IlSung. The resolution provided guidance about statements the three Communist governments should make to indicate their willingness to resolve outstanding issues and reach an armistice agreement. Chinese leaders saw Stalin’s death as the opportunity finally to end the war. Zhou traveled to Moscow for Stalin’s funeral and for meetings with the new Soviet leadership.42 According to a 1966 Soviet Foreign Ministry internal history of the Korean War, when Zhou discussed plans to end the war, he “urgently proposed that the Soviet side assist the speeding up of the negotiations and the conclusion of an armistice.”43 On March 21, Zhou met with the new Soviet leadership to discuss war-terminating strategy, producing a Communist consensus that “the Chinese and North Korean side was now in a position to conclude the war on the basis of reasonable compromises with the enemy.”44 On March 27, the Communists agreed to the UN suggestion that sick and wounded prisoners be exchanged. Three days later, Zhou publicly proposed that the POWs unwilling to be repatriated should be transferred to a neutral state “as to ensure a just solution to the question of their repatriation.”45 This statement mirrored the proposals Zhou had put forward to Stalin in August 1952. In making this substantial concession, Zhou expressed Chinese willingness to “take steps to eliminate differences on this question so as to bring about an armistice in Korea” (quoted in Zhu 1998: 234; Stueck 1995: 309). Following the same script, on April 2, Soviet Foreign Minister Vyacheslav Molotov talked about the need for peace in Korea and associated the Soviet Union with its allies’ actions (Mastny 1996: 172– 73). Significantly, the Soviet Union, “which had previously always been the state to announce important decisions from the communist side,” now followed Zhou’s declaration with Molotov’s endorsement (Simmons 1975: 234). Zhou’s statement was a major Communist concession—the largest the Communists made during the war—and it reopened the door to an armistice. Negotiations on resolving the POW issue resumed in late April, and the armistice was signed in July. Given its decision to “lean to one side” while consolidating its rule after the long Chinese civil war—on top of the staggering costs of its Korean War intervention—the Chinese governing coalition was extremely dependent upon Soviet

202

the korean war

military and economic assistance. China needed this assistance both for fighting the war and developing its own industrial base. Thus, despite its preference to end the war and concentrate on internal economic development, it could not end the war until the Soviets gave it permission. It was only after Stalin’s death that the Chinese were finally free to act as they wanted and make the crucial concession on POW repatriation that led to the war’s end.

Alternative Explanations As in the other case studies, this section briefly examines alternative explanations for Chinese behavior in the Korean War. I have argued the Chinese governing coalition wanted to end the war starting in the summer of 1952, but could not because it was trapped by its hawkish ally, the Soviet Union. There are four potential alternatives to this argument. The first alternative explanation, most different from the one advanced here, is that Mao alone was responsible for decisions about prosecuting and ending the war. A version of this argument is advanced by Jung and Halliday—in a chapter entitled “Mao Milks the Korean War”—who assert Mao went to war specifically to become a military superpower by extorting a domestic arms industry from Stalin. They argue Mao wanted to use his “millions of expendable Chinese” to get Stalin to provide “his twin key demands: turf [control over the Southeast Asian Communists] and arms industries” (2006: 363). In their view, Mao wanted to keep the war going to get “the Bomb,” but Stalin did not want to provide nuclear technology (2006: 368). Mao only ordered the end of the war on May 12, 1953, after the new Soviet leaders “recommended” Beijing drop its accusations about the United States having engaged in biological warfare. In their words, Mao could see that getting the Bomb from Russia was out of the question for now, as the new Kremlin was bent on lowering tension with America. So he recalled his nuclear delegation from Moscow, and settled for the arms projects [91 new factories] that the new Kremlin leaders had offered. He ordered his negotiators in Korea to accept voluntary repatriation of POWs, which had been on the table for over eighteen months. (2006: 371)

Jung and Halliday’s account is somewhat misleading. While Mao may have “ordered his negotiators in Korea to accept voluntary repatriation of POWs” on May 12—they cite Zhou’s papers for this date—the crucial concession about POW repatriation had already happened, on March 30. Although they concede the new Soviet leaders had “told [Z]hou they had decided to end the war” on March 21 (2006: 369), nowhere do they mention the critical Communist POW concession on March 30. On May 7, the Communist negotiator Nam Il had al-

china



ready outlined the concrete proposals on which the final armistice was based. By then, both sides had already narrowed the issues down to who would serve as the neutral power holding the non-repatriates, and where and for how long they would hold them (Stueck 1995: 314). Therefore, this May 12th document they cite refers instead to a tactical-level decision about the wording of the armistice agreement. Moreover, Jung and Halliday’s account of the August-September 1952 conversations between Zhou and Stalin is not balanced. They argue Zhou’s visit was “transparently aimed at enabling Mao to become a major power and a rival to Stalin,” and they cite those parts of the transcripts that focus on Chinese requests for Soviet assistance with armaments and industrialization (2006: 363–66). However, they never mention the POW discussions, which comprise the majority of the transcripts of the first and third meetings. They also never mention that the Communist concession on March 30 mirrored the proposals Zhou had put forward to Stalin in these meetings. Finally, as the empirical evidence in this chapter demonstrates, Mao made decisions in a consensual manner, even if he may have had the final word. There is ample evidence, from a variety of sources, about the coalition politics at play in China’s decision to enter the Korean War. Not only did the Chinese governing coalition engage in a vigorous debate about the intervention over many days, but Zhou may have even falsely characterized his October 10 discussion with Stalin about Soviet military support to tip the decision towards his preference for non-intervention. Coalition politics continued in the summer of 1952, when the coalition changed its composition to include members with expertise in industrialization and centralized economic planning. Once assembled in Beijing, the new governing coalition—including many leaders who had opposed intervention from the start—had enough collective clout to articulate arguments against fighting on and thereby to influence Mao to change the war policy. The second alternative argument comes from Goemans (2000a) who analyzes the role of regime type in war termination. In Goemans’ argument, dictators and democrats should engage in a rational cost-benefit calculation to end wars they are losing, while oligarchic leaders will choose high-risk strategies in wars they are losing to gamble for their own political redemption. Because China is classified as a dictatorship during the war,46 Goemans’ theory would predict China to act as a unitary rational actor and end the war as soon as the costs outweighed the benefits. Given the evidence presented here, that point probably occurred sometime in the summer of 1952. However, China remained at war for another year, which suggests Goemans’ argument does not hold. The only way Goemans’ argument could work is if we include China’s alliance with Stalin—and the mate-

204

the korean war

rial support the Soviets provided for the war and industrialization—as benefits in China’s cost-benefit calculations. However, Zhou’s trying to get Stalin’s support to end the war suggests the Chinese leaders did not believe these benefits outweighed the costs, but instead felt trapped. That Zhou had to make further concessions to Stalin to assure continued assistance—allowing the Soviets to keep Port Arthur and appointing Gao Gang as SPC chief—implies the costs were outweighing the benefits. Giving Goemans’ theory the benefit of the doubt, let us assume the difficulty is with China’s classification rather than the theory itself. Changing China’s classification from a dictatorship to an oligarchy brings different predictions. Once the war stalemated, Goemans’ theory would predict the oligarchy to undertake high-risk strategies to end the war successfully so as to protect the leaders’ personal political fortunes. Yet this did not occur, either. After the war stalemated, the Chinese did not embark on high-risk military strategies in the war; instead, they settled down for a protracted conflict of “negotiating while fighting.” They used most of their military resources to build extensive tunnels and other defensive systems to protect their soldiers. And in the summer of 1952, when a new group of leaders took charge to enact policies for economic development and industrialization, they acknowledged the war’s costs were impinging on this new policy focus and that it was time to shift priorities to end the war. A group of leaders worried about how defeat in the war would hurt their own political fortunes would not have done this. In short, regardless of which regime type we use to classify China during the war, Goemans’ argument does not seem to explain the Chinese case. A third alternative explanation would argue China was not trapped by its hawkish ally; rather, completely independent of the Soviet Union, China did not shift its war policy until it announced the POW repatriation concession in March 1953. One variant of this argument could be that China shifted its war policy in response to President Eisenhower’s election and alleged nuclear threats; this alternative explanation will be addressed in Chapter 8. A second variant of this argument, made by Chen Jian (2001), is that China was not willing to concede on the POW policy until late March 1953. Chen admits that “Stalin’s death might have contributed to this reversal, but it was more an outgrowth of Beijing’s existing policies based on Chinese leaders’ assessment of the changing situation than a reflection of altering Soviet directives” (2001: 115). Chen points to two documents to assert that “Beijing alone was responsible for its unyielding attitude over the POW issue” (2001: 114). The first document was Mao’s telegram to Stalin in December 1952, which I cited above. In Chen’s reading of this telegram, the Chinese leaders “expressed that ‘the armistice ne-

china



gotiations had encountered a deadlock’ and that ‘operations in Korea could be intensified in a given period in the future (say, a year).’” As a result, the Chinese leaders “were willing to wait to see what would happen after Eisenhower” became president, because they believed Eisenhower would probably launch a new offensive. Chen then makes a leap in logic to argue that “the implication was that Chinese policy could change after that time because the military situation in Korea would by then certainly have turned in [China’s] favor” (2001: 114). Even more importantly, Chen does not mention the second half of Mao’s telegram, which requested more military assistance from Stalin. Rather than implying an unyielding Chinese negotiating stance, it seems much more likely that Mao highlighted the negotiation deadlock and Eisenhower’s likely military escalation to build a case for more military assistance from his hawkish ally. The second document Chen cites to assert China’s unyielding stance on the POW issue was a February 19, 1953, cable from Qiao Guanhua to Zhou, which summarized the Chinese negotiation team’s analysis of the battlefield situation at that moment. In Chen’s account, the report observed that “there existed little possibility that the United States would initiate major military offensives on the Korean battlefield.” Moreover, Eisenhower’s decision to “unleash” KMT leader Chiang Kai-Shek in the Taiwan Strait “was designed to place more pressure on China rather than to allow [Chiang] to attempt amphibious operations against the Chinese mainland.” As Chen parenthetically notes, “(In other words, the Chinese negotiators believed that the Chinese/North Korean forces had achieved a relatively favorable position on the Korean battlefield.)” (2001: 115). Qiao argued if China proposed an unconditional reopening of negotiations under these circumstances, the United States might take it as an indication of Chinese weakness. Therefore, the report suggested China should do nothing and wait for the Americans to take the next initiative. Chen argues Mao agreed with Qiao’s recommendation (2001: 115). This second document certainly shows an unyielding stance at the tactical level. However, hawkish tactical moves are not necessarily incompatible with a more dovish strategic preference of wanting to end the war. Notably, at this point, Stalin was still in power and had not yet given China the green light to seek a negotiated settlement. If Chen is correct, that China adopted a more conciliatory policy about the POWs in late March “based on Chinese leaders’ assessment of the changing situation [rather] than a reflecting of altering Soviet directives,” then something must have changed between Qiao’s February 19 telegram and Zhou’s two concessions in late March. But the battlefield was not changing, and only one change occurred on the negotiating front: on February 22, UN Commander Mark Clark proposed an exchange of sick and wounded POWs. Conceivably, this proposal

206

the korean war

could be interpreted as the Americans making the first move to restart negotiations for which Qiao wanted to wait (Stueck 1995: 306). This leaves us with two choices for Chen’s account to hold. The first choice is that this UN proposal was enough of an American initiative to allow China to respond without “appearing weak” to the Americans, in which case why did they wait five weeks? The second choice is that China wanted to respond sooner but could not, because they were still being trapped by Stalin. In short, Chen’s explanation for the March 1953 change in Chinese policy, independent of its Soviet ally’s behavior, is lacking. On the whole, Chen’s entire recounting of Chinese decisions to initiate, prosecute and end the war is more “militant” than mine. As explained earlier, Chen’s portrait of Mao in the Chinese decision to intervene in Korea is more decisive and imposing than that of other scholars and underplays the role of the coalition debates about intervention. Moreover, Chen minimizes the August-September 1952 discussions between Zhou and Stalin; in his view, the two leaders agreed completely on continuing the war. Finally, Chen’s account does not discuss the trade-offs that the Chinese leaders were facing on the home front. He devotes very little space to the mobilization campaigns and no space at all to debates about industrial and agricultural development and internal security issues. As a result, his interpretation of Chinese decision-making in the Korean War appears to take fewer decision criteria into account. By focusing so heavily on Mao, it undersells the importance of coalitional constraints in the decision-making process. Finally, a fourth alternative explanation would argue China was trapped by its ally, but this entrapment occurred at a different time. Specifically, Lim (2002) argues that China did not adjust its war aims downward until after Stalin had died. In her estimation, Stalin’s death “removed Soviet support for any Chinese attempt to continue the stalemate in order to gain better terms at the armistice” and “pressured Chinese leaders to agree to return to negotiations” (2002: 122, 11). In many ways, this argument is similar to the entrapment obstacle explained in this chapter: China’s war policy was determined in large part by calculations about Soviet financial, military and political support. However, it differs in its understanding of why China was trapped. In Lim’s view, China needed to bow to pressure from the new Soviet leaders, even though it wanted to continue the war, because otherwise it would lose Soviet support. To make this argument, Lim cites a telegram from Mao to Zhou on March 19, quoted in Chinese sources, in which Mao intended to refute UN Commander Mark Clark’s proposal to exchange sick and wounded POWs.47 After Zhou’s March 21 meeting with Soviet leaders, however, Zhou agreed to Clark’s offer. Her conclusion is that the new Soviet leaders pressured the more hawkish Chinese to lower their aims and return to negotiations.

china



Mao may have intended to refute the UN’s proposal to exchange sick and wounded POWs, but this telegram alone does not seem to be enough evidence to assert China’s war position at this point was more hawkish than the Soviet Union’s. To look at Mao’s intention another way: because Mao sent this telegram the same day that the Soviet Council of Ministers issued its resolution to change Soviet policy towards ending war, it is possible the Chinese leaders were not yet aware that their Soviet entrapment was over. If Mao still believed the Soviets wanted to continue the war, then the tactical decision to refute the proposal to exchange sick and wounded prisoners makes sense—given the Soviets had not allowed the Chinese to propose similar moves earlier. More importantly, however, the telegram explicitly referred to the tactical-level decision in the context of stalled armistice negotiations. As with the February 19 cable, Mao’s intention for this hawkish tactical move—refuting the proposal to exchange the sick and wounded—is not necessarily incompatible with a more dovish strategic policy— wanting to end the war. The more important evidence against both Lim’s and Chen’s accounts seems to be Zhou’s March 30 concession, when he publicly proposed POWs unwilling to be repatriated should be transferred to a neutral state. This strategic-level concession is exactly what Zhou tried to convince Stalin to support the previous fall. Thus, while Lim may be correct that Chinese calculations changed after Zhou met with the Soviet leaders on March 21, I would argue the information revealed in that meeting helped convince the Chinese that the Soviets would support them in making the larger dovish policy shift they preferred. Admittedly, we do not yet have any direct evidence of Mao’s views in 1952, and thus the coalition argument laid out in this chapter is somewhat provisional. Nonetheless, the evidence shows the economic elites whose influence increased with the summer 1952 coalition shift clearly supported an early end to the war. Similarly, Foreign Minister Zhou Enlai evidently discussed war-terminating options with Stalin that fall. While this section has highlighted how some scholars believe Mao wanted to continue fighting as late as 1953, these same scholars argue foreign policy outcomes never diverged from Mao’s preferences. Therefore, although we lack a “smoking gun,” it is reasonable to argue—given the evidence available today and presented in this chapter—the dovish coalition shift in the summer of 1952 helped shift Chinese war policy towards ending the war.

Conclusions This chapter has provided another look at the entrapment obstacle, in which leaders want to end the war but cannot. Unlike the internal entrapment in the

208

the korean war

US case, however, the Chinese governing coalition was trapped by an external ally—Stalin. The role of allies in changing belligerents’ cost-benefit calculations has not received much attention in the war termination literature. This chapter takes a first step towards filling that gap by tracing how the Soviet Union constrained China’s decisions about prosecuting and ending the war. This chapter illustrated four key points. First, the PRC entered the war because its leaders perceived threats to its internal security. Second, China’s domestic situation was dire, because it was still recovering from its civil war and faced a long list of domestic challenges. In addition, the war effort was very costly. About three million Chinese troops and laborers entered Korea to fight and provide logistical support for the war—suffering as many as 1.5 million casualties, including more than 300,000 dead. At home, about three million people died during the ruthless domestic mobilization campaigns, initiated in large part to support the war effort. The Korean War’s financial costs were staggering as well, with military expenses related to the war comprising 44 percent of the Chinese government’s budget in 1950, 52 percent in 1951 and 28 percent in 1952. While the CCP hiked taxes and found other ways to extract money from the economy, the war effort consumed whatever increases the government could squeeze from its rudimentary industrial base. Third, given these tremendous war costs, the Chinese governing coalition decided in the summer of 1952 to focus on economic reconstruction. As a result, key CCP members moved to Beijing and became important members of the decision-making group. These rising leaders—together with economic policymakers already in the governing coalition—brought an interesting mix of personal perspectives and professional expertise to the problem of the Korean War. Having spent time in key regional leadership posts, they were very aware of the human and societal costs imposed by prolonged participation in the war. Furthermore, their expertise in managing economic recovery and industrialization clearly showed them the opportunity cost of waging war instead of rebuilding and modernizing the nation. By valuing centralized economic planning and industrialization more than the war, these leaders had a different vision of the war’s costs and benefits. Their increasing influence in Mao’s decision-making coalition helped spur China towards wanting to end the war. Fourth, although China was willing to settle in August 1952, Stalin prevented it from doing so. While the new governing coalition tried to convince Stalin to support crucial concessions about POW repatriation to achieve an armistice, Stalin was unwilling. Paradoxically, in large part because of the war, China was now extremely dependent upon Soviet military and economic assistance both for fighting the war and developing its own industrial base. As a result, China was

china



trapped. It was not until after Stalin’s death that China was finally able to make the POW concessions it had been advocating since the previous summer. This chapter concludes the three detailed case studies of decision-making in the Korean War’s primary belligerents. These four chapters have focused on the domestic coalition shift theory’s effects at the domestic level. The next chapter turns to the strategic interaction between the belligerents and explores the effect of domestic coalition shifts on the negotiating process at the international level.

8

Interacting Domestic Coalitions in Bargains for Peace

Now that we have examined decision-making within the individual belligerents of the Korean War, in this chapter we return to the strategic interaction between them. While the analysis up to this point may seem as if the only domestic coalition shift that mattered was Joseph Stalin’s death in March 1953, coalition shifts in the other belligerent states also played a role in the strategic interaction to end the war. Certainly, had the Chinese shift not occurred the previous summer, it is unlikely Chinese Foreign Minister Zhou Enlai would have appealed to the new Soviet leaders, right after Stalin’s funeral, for their support about the crucial POW concession that led to the war’s end. Less apparent, however, is how much President Dwight D. Eisenhower’s inauguration in January 1953 affected Communist cost-benefit calculations about the war. After successfully campaigning on a platform to end the war, Eisenhower came to office less constrained about the possibility of escalating than his predecessor. Thus, one important alternative explanation for the timing of the war’s end that still needs investigation is whether Eisenhower’s nuclear threats affected Communist calculations and caused the war to end. This chapter addresses these issues by examining the strategic interaction of the domestic coalition shifts in the Korean War, because actions taken by one side are directly linked to the other’s response. This discussion not only completes our understanding of the timing and terms of the Korean War armistice, but it also allows us to theorize more clearly about the role that domestic coalition shifts on opposing sides play in creating an overlapping bargaining space between belligerents. To do this, I return to the direction of the coalition shifts (i.e. hawkish,

interact ing d omest ic coalit ions



dovish or neutral) and look at the two mechanisms by which a domestic coalition shift can affect the creation of an overlapping bargaining space. This chapter is divided into three sections and a conclusion. The first section extends the theoretical discussion of coalition shift directions from Chapter 2 and theorizes about the strategic interaction of coalition shifts in the Korean War. The second section begins to examine this strategic interaction with one of the most famous alternative explanations for the war’s end—that Eisenhower’s nuclear threats caused the war to end. While the evidence does not support this particular argument, it does suggest domestic coalition shifts can provide information for decision-makers on the opposing side, thereby influencing when the war ends. Therefore, the third section examines the opposing side’s reactions after each belligerent’s domestic coalition shift. This analysis suggests that the Communist side may have been more sensitive to the signaling inherent in US coalition shifts than vice versa. However, such signaling seems to have the most explanatory leverage when coalition shifts on opposing sides are viewed through the lens of strategic interaction. Specifically, I argue the hawkish shift that brought Eisenhower to power was given more weight in Communist calculations because of the dovish shifts that occurred on the Communist side. Simultaneously, Eisenhower’s decision not to expand US war aims may have been the result of the Communist dovish shifts as well. Finally, the conclusion draws generalizations about the role that interactions between domestic coalition shifts play in war termination. While the findings from this one case are certainly not conclusive, they suggest that analyzing the interactions between shifts on opposing sides are a fruitful line for future research on war termination.

Coalition Shift Directions and the Interaction Between Shifts on Opposing Sides As Chapter 2 explained, there are two mechanisms by which domestic coalition shifts can lead to the creation of an overlapping bargaining space between belligerents. The first mechanism occurs when the new governing coalition modifies its own bargaining position vis-à-vis other belligerents. This mechanism is similar to unilateral escalation or de-escalation in interstate rivalries (Mor 1997; Colaresi 2004, 2005; Schultz 2005). It is important to note that this first mechanism operates unilaterally, independent of any strategic interaction with the adversary. In contrast, the second mechanism is clearly related to the strategic interaction between the opposing sides, because actions that one side takes are directly linked to the other side’s response. The second mechanism occurs when the coalition shift direction serves as an “index” for decision-makers

212

the korean war

on the opposing side to update their expectations about the war and thus change their bargaining position. It is important to remember that for the purposes of coding, the direction of coalition shifts is dependent on the ex ante intent of the new coalition, rather than any subsequent policies pursued by it.1 first mechanism: a coalit ion shift leads to a new bargaining p osit ion

The first mechanism occurs when the new governing coalition modifies its own bargaining position vis-à-vis other belligerents. While the new bargaining position may be affected by beliefs about the adversary, it may also be completely independent of any consideration of the adversary’s behavior. Therefore, this mechanism operates unilaterally and independently of strategic interaction with the adversary. In the first mechanism, hawkish coalition shifts may lead a state to take actions that move away from an overlapping bargaining space and ending the war. Hawkish shifts, therefore, are more likely to occur early in the war and, when they occur alone, be associated with long wars (H hawk). Similarly, dovish coalition shifts may lead a state to take actions that move towards an overlapping bargaining space and ending the war. Thus, we would expect them to be associated with hastening the war’s end (H dove). Neutral shifts are not intended to lead to a change in war policy, but they do bring new leaders, who may have a new way of looking at the war’s costs and benefits. Since the new leaders by definition came to power for reasons not related to the war policy, they could view the war as a distraction. Thus, they may end up taking actions that move towards an overlapping bargaining space and ending the war in order to focus on other things.2 As a result, we might expect neutral shifts to be associated with ending the war, but not as quickly as dovish shifts (H neutral). Unlike hawkish and dovish shifts, we would not expect neutral shifts to be clustered at any particular time in the war. With this first mechanism, the number of coalition shifts that will be necessary to create the overlapping bargaining space depends entirely on how much distance exists between the demands of the two sides. Until their demands become compatible, belligerents will not create an overlapping bargaining space. If the two sides’ demands are not very far apart, then a dovish coalition shift on one side may lead the new coalition to lower its aims enough to bring an overlapping bargaining space within reach. However, if the distance between the two sides is greater, coalition shifts in more than one belligerent may be necessary. During sticky wars that endure for many months or years, multiple coalition shifts may be necessary to create the overlapping bargaining space and dislodge the obstacles to peace. These multiple shifts could theoretically occur in multiple bel-

interact ing d omest ic coalit ions



ligerents or within one belligerent over time, each shift moving that belligerent’s bargaining position closer to the others’. As the previous chapters have suggested, primary belligerents are most important in their side’s decisions about prosecuting and ending war. Thus, we can imagine primary shifts will be likely to work through the first mechanism, changing their own side’s bargaining position. In contrast, secondary belligerents may not have as much influence in changing their side’s war policy, and thus secondary shifts may not operate through the first mechanism.3 second mechanism: a coalit ion shift leads to a new bargaining p osit ion on the opp osing side

The second mechanism through which coalition shifts can lead to an overlapping bargaining space between belligerents is through signaling. This mechanism is clearly related to the strategic interaction between the opposing sides, because actions that one side takes are directly linked to the other side’s response. Here, the direction of the coalition shift serves as an index to belligerents on the other side, conveying information about the first side’s future intentions and likely bargaining position. Remember that indices are behavior believed to be inextricably linked to an actor’s capabilities or intentions and “beyond the ability of an actor to control for purposes of projecting a misleading image” (Jervis 1989: 18). Thus, while international signaling is often ambiguous, domestic coalition shifts are less likely to be discounted as manipulative or deceptive than other kinds of signals. Why do domestic coalition shifts operate as indices? After a leader spends political capital campaigning or otherwise building support among his constituency for an intended war policy, he is likely to pay a domestic political price if he does not follow through on those intentions. According to the theoretical logic of domestic audience costs (e.g. Fearon 1994; Schultz 1998; Smith 1998), this vulnerability should enhance the leader’s credibility with the adversary. Certainly, leaders may misunderstand indices because they want to, because “basic strategy suggests they come out better that way” (Quester 1970: 30; see also Jervis 1989). Nonetheless, because of domestic audience costs, domestic coalition shifts are “costly signals” for leaders to send (Kydd 2000). As long as the adversary can “overhear” the policy statements used to build constituency support, the new coalition’s behavior before taking office—or immediately afterwards—reveals information about its underlying political incentives and intentions. Certainly, for the index to provide information to the other side, political competition and policy debate must take place at least somewhat in public, so that the positions of contending parties can be observable to foreign

214

the korean war

states (Schultz 1998). However, while domestic audience costs are most common in democracies, they exist in any regime where a leader can be removed from power through constituency retaliation (Smith 1998). Indeed, empirical research shows that foreign policy performance affects domestic political survival in all regime types (Bueno de Mesquita and Siverson 1995; Bueno de Mesquita, Siverson and Woller 1992). As Alastair Smith notes, “Leaders are able to send credible messages because they mortgage their domestic political survival on their ability to honor their commitments” (1998: 632). The direction of the coalition shift helps to signal information about the new governing coalition’s “type” (Morrow 1994: 221). This information may lead decision-makers on the opposing side to update their expectations about the war and change their bargaining position as a result. Thus, with the second mechanism, hawkish shifts may signal to the opposing side that the war will escalate and its costs will dramatically increase, causing the opposing side to sue for peace and thus hastening the war’s end (H hawk*). Dovish shifts may indicate weakness, so the opposing belligerents will push for even more concessions, thereby prolonging the war (H dove*). Because neutral shifts by definition are not related to the war, they are less likely to provide information to the other side. Thus, they should not affect the war’s duration through the second mechanism (H neutral*). There are three reasons to expect the second mechanism will be less powerful than the first in creating an overlapping bargaining space. First, the second mechanism can only work when the opposing side’s decision-makers are capable of perceiving and learning from the information conveyed in the coalition shift direction. Whether a message is perceived as intended depends not only on its clarity and plausibility, but also on how it fits with the recipient’s prior beliefs. Messages that are inconsistent with the recipient’s beliefs are less likely to be perceived as the sender intended (Snyder and Diesing 1977: 286; Jervis 1988: 692). Thus, the second mechanism cannot work if the information obstacle is blocking the governing coalition’s updating process. Second, if domestic coalition shifts are to serve as an index, belligerents must be paying attention to the domestic political processes in states on the opposing side. Given all of the other things that leaders must track during war, they might not give the other side’s domestic politics much attention. Finally, although domestic coalition shifts are “costly signals,” international signaling still has the potential to be garbled in transmission (Snyder and Diesing 1977: 323). Therefore, given the ambiguities inherent in international signaling, the second mechanism’s effect is expected to be less powerful than the first’s. Because primary belligerents are most important in their side’s decisions

interact ing d omest ic coalit ions



about prosecuting and ending war, primary shifts may send the most noticeable signals to the other side. Therefore, we can expect primary shifts will operate through the second mechanism as well as the first. In contrast, because secondary belligerents may not have as much influence in changing their side’s war policy, secondary shifts may be more likely to work through the secondary mechanism than the first. Dovish secondary shifts, in particular, may convey weakness, vulnerability or internal strife in that side’s international coalition, which may lead the opposing side to press for more concessions. On the other hand, the opposing side may disregard hawkish and neutral secondary shifts, supposing they will not have much effect on that side’s future intentions and bargaining position. In general, however, secondary shifts are less likely to work through either mechanism than primary shifts. inter act ions bet ween the t wo mechanisms and bet ween coalit ion shifts on opp osing sides

How can we be sure that these two mechanisms will not cancel each other out? In other words, if Side A experiences a dovish coalition shift and thus lowers its demands (first mechanism), what prevents Side B from gleaning information from A’s dovish shift and thus raising its demands (second mechanism), thereby keeping an overlapping bargaining space out of reach? As just explained, the second mechanism is likely to be less powerful than the first, which suggests that they will not cancel each other out. Moreover, as bargaining models of war explain, agreements will be self-enforcing at terms where both sides perceive the marginal benefits of additional demands as less than the marginal costs of fighting to achieve those demands. Agreements can be self-enforcing even for the stronger state, because it may perceive additional benefits are not worth the costs of fighting necessary to achieve them. Therefore, whether B will raise its demands in response to A’s dovish shift depends on two things: (1) whether it identifies and learns from the information inherent in A’s dovish shift and (2) how its governing coalition perceives the war’s costs and benefits, as well as the marginal costs and benefits of fighting on for additional demands. Although it seems likely that the first mechanism will be more important than the second, there is one situation where the two mechanisms are likely to operate in tandem. This occurs when opposing sides both experience coalition shifts at about the same time. This strategic interaction between opposing sides’ coalition shifts requires an additional level of theorizing. For example, what happens if Side A experiences a dovish shift while Side B experiences a hawkish one—as occurred in the Korean War? The interaction between opposing sides’ hawkish and dovish shifts is one of six possible permutations of the interaction of opposing

216

the korean war

sides’ coalition shifts.4 Because this particular interaction occurred in the Korean War, this chapter will theorize about its dynamics and then observe how they played out throughout this case study. Before turning to the hawkish-dovish interaction in more detail, it is important to include a few preliminary ideas about the other permutations. We can expect wars with dovish coalition shifts in belligerents on both sides to be more likely to end quickly, as both sides are more likely to lower their war aims and want to settle. Likewise, wars with hawkish shifts in belligerents on both sides will be more likely to continue, as both sides are more likely to expand their war aims and escalate fighting. The effects of dovish-dovish and hawkish-hawkish interactions are especially likely if the coalition shifts occur in both sides’ primary belligerents. Less clear are the effects of various permutations with neutral shifts (i.e. neutral-hawkish, neutral-dovish and neutral-neutral). On the one hand, neutral shifts bring to power new governing coalitions that do not intend to change war policy. Thus, we might suppose the other side’s shift would drive any potential changes to the bargaining space through the first mechanism. So, for example, if Side A experiences a neutral shift while Side B a hawkish one, we might expect A would not change its bargaining position while B moves further away from a bargaining space (first mechanism). On the other hand, since neutral shift coalitions by definition come to power for reasons not related to war policy, they might view the war as a distraction. As a result, they might be particularly open to information inherent in the other side’s shift and thus change the bargaining space through the second mechanism. Continuing the example, Side B’s hawkish shift might signal to Side A that the war is about to become very costly and lead A to lower its aims in order to focus on other things (second mechanism). Because of such complexities, further theorizing about these other permutations is needed. While these other permutations are beyond the scope of this book, the fact that the hawkish-dovish plausibility probe tested here works suggests this is a fruitful area for future research. inter act ions bet ween haw kish and d ov ish coalit ion shifts on opp osing sides

This section theorizes more carefully about the interaction between hawkish and dovish coalition shifts on opposing sides—as occurred in the Korean War. Because of the first mechanism, we would expect A to lower its demands after its dovish shift, while B would raise its demands after its hawkish shift. However, if B does not perceive additional benefits to be worth the costs of fighting to achieve them, it may decide not to raise its demands. Instead, it may decide to risk co-

interact ing d omest ic coalit ions



operation, prompted by the information inherent in the other side’s dovish shift. To explain this prediction, I draw on two-level game models of enduring rivalry termination and cooperation (Mor 1997; Kydd 2000; Schultz 2005). Kenneth Schultz (2005) argues that when trust is low but continued conflict is costly, cooperation in interstate rivalries is most likely to be initiated by a government with a hawkish reputation but moderate preferences. Schultz argues that moderate hawks are likely to initiate rivalry termination successfully for two reasons. First, domestic constituents are more likely to approve of a peace deal endorsed by a hawk than one endorsed by a dove. This logic accords with Bayesian arguments about the value of biased information sources: if a biased source gives advice against his bias, you are more apt to believe him (Morrow 1994: 170; Calvert 1985). Second, a moderate hawk can signal his moderation to the adversary by cooperating, since a hardline hawk would not cooperate. In contrast, Ben Mor (1997) argues that a well-timed diplomatic initiative or symbolic action from the other side may be enough to overcome rivalry deadlock caused by a hawkish constituency and bring the rivalry closer to reconciliation. Specifically, he argues that a peace initiative “can affect leadership and public preferences by providing evidence of the opponent’s peaceful intentions”—convincing them that settlement is feasible and reinforcing moderate views within the state (1997: 205). For example, Mor points to Egyptian President Anwar Sadat’s 1977 visit to Jerusalem as sparking a reversal in Israeli public opinion to eventually support withdrawal from the Sinai. Similarly, Andrew Kydd (2000) argues the dovish side can signal its peaceful intentions with “costly signals” designed to persuade the other side of its trustworthiness, since an untrustworthy opponent would find them too costly to send. For example, he points to Mikhail Gorbachev and the Soviet “new thinkers” that wanted to deescalate superpower rivalry to concentrate on domestic reforms. As Kydd argues, “This difference in preferences between them and previous leaders enabled them to send signals that previous Soviet leaders would have been unwilling to send, and hence these signals were persuasive to Western publics and leaders” (2000: 326). In the terminology of this book, the Soviet coalition shift that brought Gorbachev to power—as well as the new coalition’s behavior—provided reassurance to the other side, eventually leading to full cooperation. Applying both of these logics to interstate war termination after Side A experiences a dovish shift and Side B a hawkish shift, we might expect B to be willing to settle—rather than raise its war aims—if its shift brought to power a moderately hawkish governing coalition. Moderates prefer to cooperate only if they expect the other side will as well; thus, whether a moderate prefers cooperation or defection depends strongly on its beliefs about the adversary (Schultz 2005: 9). Because

218

the korean war

A has experienced a dovish coalition shift, B leaders may be more willing to move towards ending the war because they believe A wants to settle, as well. This is even more likely when A’s behavior after its dovish shift includes “costly signals” of peaceful intentions. Thus, if B’s new governing coalition is moderately hawkish and expects that continued conflict will be costly, it may be willing to forgo fighting for additional demands to take advantage of A’s dovish shift. However, even if B initiates settlement, it would still be in its interests to convey the “good” information inherent in its hawkish shift—i.e. a willingness to escalate if its demands are not met. After all, credible transmission of this information would work to B’s advantage (Dixit and Skeath 1999: 416): A is likely to see B’s settlement offer as a sincere “final offer”—one that, if refused, is likely to lead B to escalate. As Schultz argues, “A hawk that cooperates reveals itself to be moderate, which means it can be trusted to reciprocate cooperation but it cannot be exploited” (2005: 5). To summarize this special case of interacting coalition shifts: the first mechanism would predict that Side A would lower its demands after its dovish shift, while Side B would raise its demands after its hawkish shift—unless it does not perceive the costs of fighting for those additional demands to be worth their benefit. In that case, B is more likely to take advantage of A’s dovish shift to initiate settlement or accept an offer from the other side if (1) continuing the conflict is expected to be costly, (2) its new governing coalition is only moderately hawkish or (3) A undertakes some well-timed initiatives to signal its peaceful intentions. Interestingly, therefore, part of B’s calculation depends upon the second mechanism and its expectation that A’s dovish shift will increase A’s willingness to settle. If B initiates settlement and an overlapping bargaining space gets created, it is likely to be to B’s advantage, since the “good” information inherent in its hawkish shift should lead A to take its offer seriously. Indeed, the second mechanism would predict that Side A will accept B’s offer, knowing that if this offer were refused, B’s hawkish shift could lead to war escalation and a dramatic increase in war costs. Thus, in the special case of interacting coalition shifts on opposing sides, it appears that the two mechanisms for creating an overlapping bargaining space work in tandem. The next two sections examine this argument in the case of the interacting coalition shifts in the Korean War.

The Role of Eisenhower’s Nuclear Threats in Ending the War Casual observers might point to the hawkish US coalition shift in January 1953 and the Eisenhower coalition’s subsequent nuclear threats as a perfect example of the signaling described in the previous section. This explanation for the end of

interact ing d omest ic coalit ions



the Korean War—made by the Eisenhower Administration itself—is that Eisenhower’s nuclear threats caused the Communists to agree to the armistice terms. There has been great debate about the effectiveness of Eisenhower’s nuclear coercion in bringing about the end of the Korean War.5 While the evidence does not support this particular argument, it does suggest domestic coalition shifts can provide information for decision-makers on the opposing side, thereby influencing when the war ends. Shortly after the armistice was signed, members of the Administration (including Eisenhower and his secretary of state, John Foster Dulles) claimed that “a series of subtle and calculated moves” had been taken to convince the Communists that the United States might go nuclear if an armistice was not reached (Foot 1990: 176–77; Eisenhower 1963: 179–81; Joy 1955: 161–62; Freedman 1983: 85). For example, when Eisenhower’s chief of staff, Sherman Adams, asked the president how an armistice was finally reached, he answered, “Danger of an atomic war. . . . We told them we could not hold it to a limited war any longer if the communists welched on a treaty of truce” (Adams 1961: 102). Despite such contentions, Eisenhower and others have overstated the efficacy of these nuclear threats. As Chapter 6 explained, the Eisenhower Administration did discuss the possibility of using nuclear weapons in Korea, as part of its endorsing a shift in the American strategy for ending the war. In a meeting on May 20, the National Security Council formally approved contingency plans for using nuclear weapons (FRUS 1952–54, 15: 977). Moreover, the Eisenhower coalition tried two diplomatic channels to signal to the Communists a growing willingness to escalate the level of violence. These included conversations in late May between (1) Dulles and Indian Prime Minister Nehru, in which Dulles said “if armistice negotiations collapse, the United States would probably make a stronger . . . military exertion and that this might well extend the area of conflict,”6 and (2) US Ambassador to Moscow Charles Bohlen and Soviet Foreign Minister Molotov, in which Bohlen emphasized “the extreme importance and seriousness of the latest UNC proposals” and made it “clear these represent the limit to which we can go” (FRUS 1952–54, 15: 1103; Zhang 1992: 125). However, in retrospect, the nuclear threat did not cause the crucial Communist concessions about the POW voluntary repatriation policy that paved the way for an agreement. On March 27—long before the Eisenhower coalition changed its strategy and began to send nuclear coercion signals—the Communist team agreed to an exchange of sick and wounded POWs. Three days later, Zhou conceded that non-repatriates should be handed over to a neutral third party after the armistice. This latter concession was by far the biggest the Communists made

220

the korean war

throughout the settlement negotiations. In fact, the only signal that Eisenhower had sent by this date was his State of the Union announcement to move the Seventh Fleet from the Taiwan Strait. Indeed, as Gaddis argues, the Eisenhower coalition’s “most explicit [nuclear] warnings came after the armistice, and in the context of how it would respond to a violation” (1997: 108, emphasis in original). Indeed, it seems very unlikely the Communists were coerced by nuclear threats, for three reasons. First, although Chinese analysts carefully monitored and analyzed American press reports and debates about nuclear weapons, Chinese leaders believed that the United States had greatly exaggerated nuclear weapons’ destructive capabilities and that properly designed underground shelters would be an effective defense. Chinese leaders also denied that a nuclear strike could succeed against such an undeveloped and vastly populated country as China (Ryan 1989: 156; Foot 1990: 178–79; Chen 2001: 111). Second, Chinese documents reveal that China considered four options for the Eisenhower Administration to escalate militarily: (1) a large-scale offensive at the front, (2) intensive air raids on strategic areas in North Korea and Manchuria, with a naval blockade of China’s coastline, (3) the use of atomic weapons and (4) amphibious attacks on both sides of the Korean peninsula behind Communist front lines. The Chinese leaders rejected the first two options and argued the US would not launch nuclear weapons due to world opinion and the deterrent of Soviet retaliation. Instead, they concentrated their energies on extensive preparations for possible amphibious attacks (Zhang 1992: 132–37; 1995: 233–36; Zhai 1994: 126–28; Lim 2002: 125; Gaddis 1997: 108; Lewis and Xue 1988: 15). Thus, these first two reasons suggest that the nuclear threats were not credible. Most importantly, when asked years later about their reaction to American nuclear threats in Korea, “Chinese officials denied even having heard of them” (Gaddis 1997: 108; see also Zhang 1992: 133, 150; Holloway 1994: 335). Rosemary Foot argues it is highly unlikely that Dulles’ ambiguous threats ever made it via Nehru to the Chinese government. Nehru denied his supposed role as a conduit for these messages (Foot 1990: 177–78). As William Stueck argues, “Nothing has surfaced in China or the Soviet Union to confirm that its leaders received any threatening messages of this kind” (1995: 329). However, even if nuclear threats were neither credible nor clearly communicated, the record clearly shows the Eisenhower coalition’s threats to escalate did have some effect on ending the war. A careful tally of events suggests the Communist leaders were paying attention to the hawkish US domestic coalition shift and the possible escalation that might result. Chinese documents clearly suggest that authorities in Beijing monitored Eisenhower’s campaign statements for

interact ing d omest ic coalit ions



more aggressive strategies in Korea and started making military preparations in response. As will be seen below, these hawkish intentions played into the Communists’ March 30 concession on POW repatriation and paved the way for armistice negotiations to resume under the general framework of the Communist proposal. After the Communist concession, negotiations narrowed down to who would serve as the neutral power, where and for how long they would hold the nonrepatriates­, and what would happen to the ultimate hold-outs (Hermes 1966: 423).7 On May 7, Communist negotiator Nam Il outlined the concrete proposals on which the final armistice agreement was based. During the bickering about these last details, the Eisenhower coalition adopted its new strategy on May 20 and presented its “final offer” at Panmunjom on May 25. Shortly thereafter, Dulles met with Nehru, and Bohlen with Molotov, to signal the finality of the UN offer. The evidence suggests Bohlen’s conversation with Molotov did play into Communist calculations: on June 3, Molotov stated that although the outcome of the negotiations did not depend on Moscow, he could say that “the path to the successful conclusion of the armistice agreement has been mapped out” (quoted in Foot 1990: 177). The next day, the Communists accepted the bulk of the UN “final offer,” and on June 8, initialed a final agreement on the POW issue. In sum, the evidence does not support the argument that Eisenhower’s nuclear threats played an important role in ending the war. In retrospect, it seems more likely this argument was advanced by the Eisenhower Administration to prove the success of its “massive retaliation” strategy. After all, if nuclear threats ended the Korean War, there would be no need for a national security strategy that involved fighting limited wars—a very valuable “fact” to set the stage for Eisenhower’s “New Look” strategy adopted later that year. What the evidence does suggest, however, is that domestic coalition shifts can provide information for decision-makers on the opposing side and thereby influence when a war ends. It is clear that Communist leaders were paying attention to the hawkish US domestic coalition shift and the possible war escalation that might result. Therefore, the next section will examine this argument more explicitly.

Coalition Shift Directions as Signals in the Korean War The first section of this chapter suggested two ways that domestic coalition shifts can help create an overlapping bargaining space between belligerents. The first mechanism occurs when the new governing coalition changes its own bargaining position, while the second mechanism occurs when the coalition shift provides information for the opposing side to update its expectations and thus

222

the korean war table 8.1 Domestic Coalition Shifts in the Korean War

Country

Date

North Korea

6/26/1950

Hawk

Direction

Kim Il-Sung creates 7-man Military Committee that directs all policy decisions during war

Details

USA

9/19/1950

Hawk

George C. Marshall replaces Secretary of Defense Louis A. Johnson, the last economizer to be pushed out of Truman’s Administration

South Korea

11/23/1950

Neutral

Acting PM Shin Song Mo replaced by Chang Myon, Shin remains Defense Minister

North Korea

12/21/1950

Neutral

Mu-Chong, Kim Il, 7 others expelled from KWP at Third Joint Plenum

North Korea

1/31/1951

Neutral

Vice Premier and battlefield commander Kim Chaek dies

South Korea

5/16/1951

Neutral

Vice Pres. Yi Si Yong, DM Shin Song Mo, Interior Min. Cho Pyung Ok, Justice Min. Kim Jun-Yun resigned/relieved because of KoChang massacre and National Defense Corps embezzlement scandal

USA

9/17/1951

Neutral

Robert A. Lovett replaces Secretary of Defense Marshall

South Korea

4/24/1952

Neutral

PM Chang Myon resigns, with plans to run for president, replaced by Chang Taek-Sang

South Korea

5/25/1952

Neutral

President Rhee declares martial law, arrests 45 National Assembly opposition members

China

8/1/1952

Dove

Institutional leadership directly below Secretariat of the CCP Politburo is greatly expanded

South Korea

8/5/1952

Neutral

Rhee restructures gov’t with constitutional amendments to consolidate power, reelected in first direct presidential election, drops Interior Min. Yi Pom Sok

South Korea

10/6/1952

Neutral

PM Chang Taek-Sang replaced by Paik Too-Chin

USSR

10/16/1952

Neutral

At the 19th Party Congress, Stalin abolished the 12-member Politburo and created a Central Committee Presidium with 25 members and 11 candidate members, to minimize influence of his top deputies

USA

1/20/1953

Hawk

President Eisenhower sworn into office

USSR

3/5/1953

Dove

Stalin dies, replaced by troika of Lavrenty Beria, Georgi Malenkov, Vyacheslav Molotov

change its bargaining position. One way to test this second mechanism is to observe whether the opposing side does, in fact, change its behavior in response to a coalition shift. Certainly, the clearest way to test the mechanism’s effect would be to observe if the opposing side’s leaders refer to the shift in their deliberations and subsequent behavior change. Therefore, this section examines the opposing side’s reactions to each coalition shift in the Korean War, observing whether the actors explained their behavior change in light of that shift.

interact ing d omest ic coalit ions



As we shall see, not every domestic coalition shift brought a discernable reaction from the belligerents on the other side. The evidence suggests that the Communist side may have been more sensitive to the signaling inherent in US coalition shifts than vice versa. It also suggests that coalition shifts mattered most in terms of creating overlapping bargaining space when they interacted with shifts on the other side. Specifically, the hawkish shift that brought Eisenhower to power seems to have been given more weight in Communist calculations because of the dovish shifts that occurred on the Communist side. Similarly, Eisenhower’s decision not to raise US war aims—despite the hawkish strategy change—may have been caused by the dovish shifts on the other side. Given how many domestic coalition shifts occurred among the belligerents in the Korean War, Table 8.1 provides a reminder of the timing of the shifts. For the sake of completeness, the domestic coalition shifts in North and South Korea are included in this table. However, as the evidence in Chapter 3 suggested, their coalition shifts do not appear to have had much effect on the primary belligerents’ calculations. This supposition is reinforced with the discussion below, which highlights how primary belligerents gave little notice to the secondary shifts—either on their own side or the opposing side. The primary belligerents’ shifts are in bold. us react ions to communist shifts

Given its NSC-68 mindset, the Truman governing coalition was not particularly adept at reading domestic politics signals from the opposing side. As Chapters 5 and 6 demonstrated, for much of the war, the Truman coalition interpreted all Communist action as originating from Soviet directives. As a result, they missed crucial warnings from China prior to its intervention and then treated evidence from the “entirely new war” after China’s intervention as vindication of a Soviet “grand design” for world domination. Similarly, in the negotiating tent, the UN team was not only less informed than its Communist counterparts but also unable to disaggregate different interests and objectives within the Communist camp. As Foot argues, “relatively little effort had been put into understanding the objectives of these Communist states [China and North Korea] and even less time given to a consideration of their primary goals in the armistice process” (1990: 70, emphasis added; also Goldhamer 1952: 263). Because the information obstacle was so prevalent among the Truman coalition, it is not surprising that the United States appears to have given little attention to the domestic coalition shifts among the Communist belligerents before Stalin’s death. Indeed, there is very little discussion of North Korean or Chinese domestic politics in US government analysis about the Korean War, let alone a mention of the shifts in these states’ domestic governing coalitions. The Truman

224

the korean war

governing coalition did observe and speculate about the Soviet Union’s 19th Party Congress in October 1952, but the results of this neutral coalition shift were not directly interpreted in light of the ongoing conflict in Korea (Stueck 1995: 290). To be fair, none of these shifts was as obvious or historic as Stalin’s death. Moreover, the internal processes of the autocratic Communist regimes were more opaque than US democratic domestic politics. As noted above, for the index to provide information to the other side, political competition and policy debate must take place at least somewhat in public to be observable to the other side. More importantly, given the NSC-68 mindset, Stalin was the leader that really mattered, since he was presumed to control the monolithic Communist camp. Because the Truman coalition was unable to perceive the three Communist states as independent actors, each with their own interests and objectives, it missed the opportunity to learn from cues that the North Korean and Chinese coalition shifts were providing. Thus, the opposing side’s domestic coalition shifts did not lead the United States to update its expectations about the war or change its own bargaining position as a result. However, Stalin’s death—and the resulting dovish shift in the Soviet governing coalition—did receive attention from Eisenhower’s Administration. After Stalin’s death, the Soviet emphasis on peaceful coexistence markedly increased, with front page coverage in elite US media (Stueck 1995: 308–9). As Chapter 4 explained, in a speech before the Supreme Soviet on March 15, the new Soviet leader Georgii Malenkov stated, “there is not one disputed or unsettled question that cannot be decided by peaceful means on the basis of mutual understanding of interested countries”—singling out the United States as among the countries to which this statement applied (quoted in Folliot 1956: 12–13). On March 21, a Moscow radio broadcast departed from the Soviet’s customary characterization of Germany’s defeat in World War II, framing it instead as the result of cooperation among the three allies—the Soviet Union, the United States and Great Britain. Making these verbal gestures more concrete, Soviet Foreign Minister Molotov responded to a British request to help obtain the release of British and French civilians held in North Korea, and in early April agreed to help free 13 American civilians as well. By early May, Pyongyang released seven American civilians (Stueck 1995: 309; Foot 1990: 183). Moreover, in early April, a group of US newspaper editors—the first to be permitted to visit Moscow since World War II—received a warm welcome, guaranteeing front page coverage of these friendly gestures in the New York Times (April 5, 1953: 1). At the end of March, after Zhou proposed the Communists’ major concession on non-repatriates, the Soviets approached a Norwegian official, Hans Engen, for his response to Zhou’s POW concession and for Norwegian views on Soviet-

interact ing d omest ic coalit ions



American negotiations. The Soviet official suggested topics to be discussed at a US-Soviet summit, including the control of atomic energy and disarmament, and explicitly linked these broader questions to the end of the Korean conflict. Choosing to approach Engen was significant, because the Soviets had spoken with him in October 1950 about seeking an end to the Korean conflict after the Inchon amphibious landing. At the same time, the US ambassador to the United Nations reported to Dulles that he had received an indirect request for talks between Malenkov and Eisenhower (Foot 1990: 182). In response to these overtures, on April 14 the Eisenhower Administration leaked to the press that it favored reopening talks at Panmunjom and designating Switzerland as the neutral guardian of non-repatriate POWs (New York Times, April 15, 1953: 1). Two days later, Eisenhower delivered his “Chance for Peace” speech, in which he publicly acknowledged the shift in the Soviet governing coalition (and its resulting policy changes) and signaled back. As Eisenhower pointedly said, “recent statements and gestures of Soviet leaders” could only be “attested by deeds,” the first of which “must be the conclusion of an honorable peace in Korea.” Afterwards, both Pravda and Izvestiia devoted their front pages to an editorial reply and reprinted the entire speech inside (Stueck 1995: 312). How did the Eisenhower Administration perceive Communist behavior after the new Soviet governing coalition took power? The available evidence suggests it observed these gestures and concluded the Communists were eager to bargain (Stueck 1995: 311). For example, Soviet specialist George Kennan said the Soviet press coverage of Eisenhower’s speech revealed Soviet interest in reducing tensions with the United States. Kennan argued that if the United States wanted to take advantage of this conciliatory attitude to make progress on the armistice, it would need to help the Soviets offer inducements to China. Similarly, US Ambassador Bohlen reported from Moscow that the Soviets were “eager to support [Bohlen’s] statement regarding the importance of an armistice in Korea” (quoted in Foot 1990: 182). Finally, the US consul-general in Hong Kong, as well as the British foreign office, argued that the Communists’ March 30 concession about the POW policy had “arisen out of the need to consolidate the Communist bloc after Stalin’s death” (Foot 1990: 183). They recognized that this desire to consolidate the Communist bloc would require redirecting political energy and resources away from the war. Thus, the dovish shift in the Soviet governing coalition did provide information to the UN side about the Communists’ future intentions and bargaining position. The Eisenhower Administration was clearly aware that the Soviet dovish shift was leading the Communists to lower their demands in the armistice negotiations. What is interesting to see, however, is that the United States did

226

the korean war

not raise its demands in response. It is true that during this time the Eisenhower government began reviewing options for military escalation, eventually adopting a more hawkish military strategy and signaling to the Communists its willingness to escalate the war. Nonetheless, the Eisenhower governing coalition did not raise its war aims and move its bargaining position further away from an overlapping bargaining space. Thus, as least on the US side, the second mechanism did not have its predicted effect: the Soviet dovish shift did not lead the United States to press for more concessions and move further away from a negotiated settlement. What explains this curious outcome? Why did the Soviet dovish shift not lead the United States to raise its war aims in response? There appear to be three reasons why the United States chose not to profit from the bargaining “advantage” of the Soviet dovish shift. First, the Eisenhower coalition clearly recognized the Soviet dovish coalition shift and subsequent Communist behavior as signaling peaceful intentions. Obviously, Zhou’s concession on POW repatriation represented a major concession in the armistice negotiations. Moreover, the new Soviet leaders sent “costly signals” that Stalin would have been unwilling to send. Hence, these signals persuaded the United States that the Communists were anxious to bargain. Second, the Eisenhower coalition expected that continuing the conflict was likely to be costly, and it no longer wanted to bear these costs. A fiscal conservative, Eisenhower had largely been drawn to the nuclear deterrent strategy of “massive retaliation” precisely because it would allow him to cut defense spending. As Eisenhower notes in his memoirs, on the first day of his presidency he met with Budget Director Joseph Dodge to revise the budget downward to put the country on a “stronger financial footing” (Eisenhower 1963: 109–10). Expanding US war aims would require not only military escalation but also a prolonged war, which would drain US resources. Thus, as Treasury Secretary George Humphrey noted on May 1, if Eisenhower really wanted to reduce the budget deficit, “you have to get Korea out of the way” (quoted in Foot 1985: 223, emphasis in original). Finally, the Eisenhower coalition was only moderately hawkish. While the Republican right wing included hardliners, the majority of the coalition’s broader constituencies would not oppose cooperative movement to end the war. Internationally, the Eisenhower coalition recognized that neither neutral states in the United Nations nor US Western allies would accept overturning the armistice terms already agreed on or delaying the resumption of negotiations for much longer (FRUS 1952–54, 15: 894). Domestically, although American public opinion was sufficiently mixed to allow Eisenhower freedom of action, much of the press was already questioning the risks of an expanded war (Foot 1990: 172; Sterling

interact ing d omest ic coalit ions



1998: 459–60). Because Eisenhower had built up the expectation during his campaign that he would end the war soon, the public both looked for and expected decisive action. Eisenhower believed he had a “mandate from the American people to stop this fighting,” but he also knew he did not have a blank check. Thus, when Dulles proposed the United States attempt to secure a better agreement than the one based on the 38th Parallel, the president said he thought it “impossible to call off the armistice now and to go to war in Korea. The American people would never stand for such a move” (both quoted in Foot 1985: 219). In sum, because the information obstacle was so prevalent in the Truman governing coalition and domestic politics in the Communist states were more opaque, the United States did not give much attention to the domestic coalition shifts among the Communist belligerents before Stalin’s death. However, the Eisenhower Administration was clearly aware that the Soviet dovish shift was leading the Communists to lower their demands in the armistice negotiations. Nonetheless, the United States did not raise its demands in response, because—as bargaining models of war predict—the marginal benefits of additional US demands were less than the marginal costs of fighting to achieve them. As a result, neither the first nor the second mechanism from a US-only perspective explains US behavior: the United States neither raised its demands after its own hawkish coalition shift, nor raised its demands after receiving the signal inherent in the Soviet dovish shift. Instead, understanding how the overlapping bargaining space changed requires considering the strategic interaction of the domestic coalition shifts on both sides. communist react ions to us shifts

The Communists appear to have watched US domestic politics carefully during the war. Certainly, democratic domestic politics are more transparent for outsiders to observe and analyze. Only one of the US coalition shifts appears to have received attention and prompted reactions from the Communists. The Communists did not respond directly to the US hawkish shift in September 1950 or the US neutral shift in September 1951. (They also did not respond directly to any of the South Korean neutral shifts, given that President Syngman Rhee’s government remained consistently hawkish about its desire to unify the peninsula under Rhee’s leadership.) However, the Communists did respond to the US hawkish shift with Eisenhower’s inauguration in January 1953, leading the Communists to lower their bargaining position (second mechanism). This fact suggests the Communists were more able to apply the information learned from US coalition shifts after their own dovish shifts. One reason why the Communist side may have been more sensitive to the

228

the korean war

information inherent in the opposing side’s coalition shifts is that China placed great importance on the bargaining aspect of the war. Of the three primary belligerents, China seems to have most internalized the notion of war as a continuation of politics by other means. Moreover, Chinese strategic culture has historically placed great focus on information advantage and diplomatic action to defeat the adversary in war (e.g. Sun Tzu’s Art of War; Johnston 1995: chapter 3). Thus, China may have been most sensitive to signals from the other side that would affect the negotiating/fighting continuum. As Chapter 7 explained, the Chinese governing coalition viewed armistice negotiations as simply one tactic in the ongoing struggle with the United States, a policy it called “negotiating while fighting” (Nie 1988: 641). For example, the Chinese governing coalition provided both high-level attention and significant expertise to the negotiating process. Once armistice negotiations began in July 1951, Deng Hua, at that time still the Chinese People’s Volunteers (CPV) vice-commander, and CPV Chief of Staff Xie Fang represented the Chinese in the armistice negotiations. Yet unlike the UN Command representatives, who lacked special training or diplomatic advisors to help, Deng and Xie were joined by political advisors and diplomatic specialists in the negotiating tent. These included two of Foreign Minister Zhou Enlai’s best aides: Li Kenong, first deputy foreign minister and the Central Military Commission’s intelligence chief, and Qiao Guanhua, director of the Foreign Ministry’s press bureau. Deng and Xie met every night with Li and Qiao, who would brief them on UN propositions. Li and Qiao would report to Zhou, who would immediately relay the information to Mao. Mao’s instructions for the next step would go back through Zhou, Li and Qiao to Deng and Xie (Zhang 1995: 218–19; Chen 2001: 99; Wu 1985: Chapter 3; Foot 1993: 196). This intensive process, with Mao involved on a daily basis, implies that the Chinese governing coalition placed great importance on the bargaining process. As such, it is likely China would be most sensitive to signaling and cues from the other side—which is what the evidence suggests. In addition, the Communists appear to have been savvier about differentiating among the various UN belligerents’ interests than the United States was about differentiating among the Communists’. The United States mostly treated the three Communist states as a monolithic bloc under direct Soviet control and could not even perceive that these actors might have different stakes and objectives in the war. In contrast, the Communists frequently noted differences in war aims and preferred strategy between the United States and its more dovish Western allies, on the one hand, and between the United States and more hawkish South Korea, on the other. For example, after Rhee unilaterally released more than 25,000 anti-Commu-

interact ing d omest ic coalit ions



nist North Korean prisoners on June 18, 1953, Chinese leaders believed it unlikely that the United States would support Rhee’s attempt to delay the final agreement, because the United States was under “tremendous pressures” at home and from Western allies. Thus, Mao argued the Chinese strategy should focus on “taking advantage of the contradictions between the Americans and the South Koreans.” CPV Commander Peng was happy to oblige, suggesting that China postpone the armistice for a month, “deepen the contractions among the enemies, [and] give Syngman Rhee’s puppet forces another strike by annihilating 15,000 puppet troops.”8 Thus, over Kim Il-Sung’s objection, Mao approved an offensive targeted specifically at punishing the ROK army on June 21 (Chen 2001: 116).9 After this decision, an editorial on June 23 in the Beijing People’s Daily stated simply that the incident “has delayed the signing of the Korean armistice agreement” (quoted in Stueck 1995: 339). Moreover, Beijing propaganda broadcasts during this time were directed almost solely at Rhee and his government and not at the United States (Clark 1954: 284). And although the Communist negotiators at Panmunjom had initially demanded the recapture of the released POWs, on June 29 the Soviet ambassador in Beijing told the Swedish ambassador there that this demand “should not be taken literally” (quoted in Stueck 1995: 339–40). On July 7, the Communists even agreed to resume talks, which reconvened on July 10. During this time, the Chinese launched their punishment campaign directed at ROK troops, who were pushed back six miles and suffered the bulk of the UN Command’s 14,000 casualties from this campaign (Foot 1990: 186–87). The Communists stopped the campaign unilaterally on July 20, and the armistice was signed seven days later. As this example demonstrates, the Communists were more skilled at differentiating among the various UN belligerents and their interests and objectives. When considered with the resources they committed to the negotiation process, the Chinese especially appear to have been very sensitive to receiving signals from the other side, as well as to sending targeted signals back. Using this logic, we might even suppose the Communists would be more likely to manifest behavior in line with the second mechanism, changing their bargaining position in response to all US domestic coalition shifts. Indeed, there was even another hawkish shift on the US side in September 1950, and so we might expect the Communists would have lowered their reservation price in response. But they did not. The second mechanism occurred after only one shift—Eisenhower’s hawkish shift. Why did the Eisenhower hawkish shift provoke a change in the Communist bargaining position, when the previous Truman hawkish shift did not? Some other dynamic must be present, and I argue the lens of the Communist dovish shifts made a difference.

230

the korean war

There was no discussion in the Communist states about the US hawkish shift in September 1950. Right after this shift, MacArthur successfully conducted his amphibious landing at Inchon and turned the tide in the war. Within one month, the UNC had crossed the 38th Parallel and was rapidly approaching the Yalu River. As UN forces marched north, the Soviet Union did take actions that suggested a willingness to settle, but it seems clear these actions were prompted by the UN military successes and rapidly shifting battlefield, not the previous month’s US domestic coalition shift. In October 1950, just before the United States crossed the 38th Parallel, the Soviets presented a cease-fire proposal at the United Nations and approached a Norwegian diplomat, Hans Engen, for negotiations outside UN channels. Engen believed the Soviets were seeking to avoid US occupation of North Korea—the Soviets told him so—in exchange for the possibility of eliminating an independent North Korean government. Thus, realpolitik factors, not the US domestic coalition shift, explain this change in war aims (Stueck 1995: 103). In contrast, the signaling inherent in the US hawkish shift with Eisenhower’s inauguration received preferred treatment after each Communist state experienced its own dovish coalition shift. Remember that the Chinese governing coalition experienced its dovish shift in the summer of 1952, when a new group of leaders took power with a mandate for industrialization and economic centralization. The dovish shift led to a change in the Chinese coalition’s conceptualization of the war’s costs and benefits—with a greater focus on its costs and a concomitant desire to end the war. As a result, the Chinese coalition was particularly receptive to information about the upcoming US hawkish shift and the expectation such a shift would increase war costs further. In contrast, the Soviets did not have the same reaction to the signaling from the impending US hawkish shift. Stalin was still in power and still wanted to continue the war, regardless of its costs, because the Soviet Union was not directly bearing them. After the Chinese dovish shift in the summer of 1952, Zhou Enlai directed his Foreign Ministry analysts to follow closely each public speech of US Republican presidential candidate Eisenhower. It seemed clear to the Chinese that “Eisenhower would be more aggressive in Korea” (Zhang 1992: 132; 1995: 317–18). Zhang argues Chinese leaders especially noticed Eisenhower’s promise of an “honorable” peace in Korea through “deeds” not “words,” and they believed that Eisenhower would definitely intensify the battlefield situation. Chinese leaders, particularly military commanders, predicted that Eisenhower would probably conduct amphibious operations, and Chinese intelligence indicated that UN forces were prepared for such maneuvers (Zhang 1992: 132–37; 1995: 233–36; Zhai 1994:

interact ing d omest ic coalit ions



126–28; Lim 2002: 125; Gaddis 1997: 108).10 In contrast, Chinese leaders considered a ground offensive and strategic bombing unlikely, because such measures had already been tried, and threats to use nuclear weapons not credible, because of US public and world opinion (Zhang 1992: 133). Eisenhower’s election led Chinese leaders to conclude that US war policy would become more aggressive. They carefully watched his post-election activities, including (1) his publicized meeting with MacArthur to hear the former commander’s hawkish plan for ending the war and (2) his trip to Korea in early December to meet with UN Commander Mark Clark as well as the commander of US military advisors in Taiwan, who joined them (Ambrose 1983b: 30–35; Stueck 1995: 305). Around this time, the United States was also conducting stepped-up amphibious training maneuvers off the North Korean and Chinese coasts (Clark 1954: 97–99).11 Thus, the Chinese military increased its preparations to counter possible amphibious attacks, which were anticipated for the spring (Zhang 1992: 132–37; 1995: 233–36; Zhai 1994: 126–28; Lim 2002: 125; Gaddis 1997: 108; Stueck 1995: 306–7).12 More importantly, on December 17, Mao cabled Stalin that US President-elect Eisenhower was “currently carrying out preparations for military actions that will take place after he comes to power.” Knowing the Soviet leader would only advise prolonging the war, Mao informed Stalin of a new Chinese effort to mobilize an additional 250,000 troops to compensate for China’s terrible war losses, and he requested more Soviet armaments.13 Mao’s telegram clearly shows he expected that an escalated war would mean increased war costs, in terms of additional troop mobilization and more combat deaths. Stalin waited ten days to reply to Mao’s expressed concerns, and his reply was diffident. He advised the Chinese to “count on the worst and proceed from the probability of an attack by the Americans” and only offered one-quarter of the aid that Mao had requested.14 While Mao’s message highlighted the US hawkish shift’s potential for an escalated war and increased costs, Stalin in effect shrugged. Unlike the Chinese leaders, who were viewing the war through the cost-focused lens of their recent dovish shift, Stalin did not care about the possible effects of escalation. After all, he had orchestrated the war so that he was gaining benefits while pushing the costs onto the Chinese and North Koreans. Thus, until Stalin’s death on March 5, 1953, this evidence suggests that the Chinese were more sensitive than the Soviets to the likely escalation and increased costs that would result from Eisenhower’s election. However, after Stalin’s death and the resulting Soviet dovish shift that brought the troika to power, the new Soviet coalition acknowledged the hawkish US shift and made a number of gestures towards conciliation.15 These gestures can be

232

the korean war

viewed both as “costly signals” resulting from the Soviets’ dovish shift as well as reactions to the US hawkish shift. For example, the Soviet press editorials after Eisenhower’s “Chance for Peace” speech not only responded to the new American president’s views but provided an assurance that the Soviets were willing to help do their part to resolve controversial issues like the Korean War. The most important Communist concession during the war was Zhou’s March 30 statement suggesting the POW issue could be resolved by turning over all non-repatriates to a neutral state. This major concession came a few weeks after Stalin’s funeral and the resulting Soviet dovish shift. Just as the Chinese dovish shift the previous summer made the Chinese coalition more likely to expect Eisenhower would escalate the war, the new Soviet troika now had the same frame of reference. The new Soviet leaders worried about war escalation and increased costs, and they looked differently at information about potential US use of atomic weapons in Korea (Zubok and Pleshakov 1996: 155; see also Mastny 1996: 164). Moreover, the new Soviet leaders were ready to shift their economic priorities away from the Korean War to help China implement its First Five-Year Plan, deal with unrest and food shortages in Eastern Europe, and free resources for neglected domestic programs and agriculture (Mayers 1986: 113; Zubok and Pleshakov 1996: 148; Stueck 1995: 327; Richter 1994: 35, 37). Given that Eisenhower’s expected escalation would not only increase war costs but also tie up resources that they wanted to shift to other areas, the new Soviet leaders perceived the signal inherent in the US hawkish shift differently than Stalin had. Thus, it seems both Communist states needed to be viewing Eisenhower’s hawkish shift through dovish lenses before they could lower their demands and concede on the POW policy. Both Communist states needed dovish lenses for the second mechanism to have an effect in creating an overlapping bargaining space. On balance, it appears that the change in the Communist bargaining position is mostly explained by the dovish coalition shifts in China and the Soviet Union. In other words, the first mechanism had a greater effect on creating an overlapping bargaining space. However, the US hawkish shift did increase the effect of these dovish shifts by influencing Communist calculations about the war’s costs and benefits. The Communists were more able to learn from the information conveyed by the US hawkish shift—and change their bargaining position accordingly—because of their own dovish shifts. The Chinese and Soviet dovish shifts made these governing coalitions not only more conscious of the war’s costs in general, but also of how those costs would likely increase in the future. Both coalitions talked about their expectation that Eisenhower would escalate the war, but such comments only happened after each coalition experienced its respective dovish shift. The dovish frame of reference led them to project an adversary

interact ing d omest ic coalit ions



with a greater willingness to commit resources and political capital to win. Thus, the Communists responded to Eisenhower’s hawkish shift by lowering their demands—the only time the second mechanism affected the creation of the overlapping bargaining space during the war.

Conclusions Ending a war is a two-level game and any account of the war termination process must look at both levels and the interaction between them to answer the questions of how, when and why wars end. Building on the domestic-level analysis of the Korean War belligerents, this chapter completes our explanation of the timing and terms of the war’s ending. In the process, it presents a more nuanced theoretical understanding of the role that domestic coalition shifts play in international negotiations to end war. It has also offered an explanation for the surprising finding that the country doing better tends not to increase its demands in ways that cancel out the war-terminating effects of a dovish coalition shift in the loser, as some rationalist theories might expect. This chapter outlined two mechanisms by which domestic coalition shifts can help create an overlapping bargaining space between belligerents. The first mechanism occurs when the new governing coalition changes its own bargaining position. The second occurs when the coalition shift provides information for the opposing side to update its expectations about the war and thus change its bargaining position. Given the ambiguities inherent in international signaling and the possibility of an information obstacle blocking information from the opposing side’s shift, we would expect the second mechanism’s effect to be less powerful than the first’s. However, the chapter suggested the two mechanisms are likely to operate in tandem in situations when opposing sides both experience coalition shifts at about the same time. Because the Communist primary belligerents experienced dovish shifts and the United States experienced a hawkish shift, the chapter theorized explicitly about this particular interaction between coalition shifts on opposing sides. Assuming that Side A experiences a dovish shift while Side B a hawkish one, the first mechanism would predict that A would lower its demands, while B would raise its demands. However, if B does not perceive the costs of fighting for those additional demands to be worth their benefit, B is more likely to take advantage of A’s dovish shift to initiate settlement or accept an offer from the other side if (1) continuing the conflict is expected to be costly, (2) its new governing coalition is only moderately hawkish or (3) A makes some well-timed initiatives to signal its peaceful intentions. If B initiates settlement and an overlapping bargaining space

234

the korean war

gets created, it is likely to be to B’s advantage, since its hawkish shift should lead A to take its offer seriously. Interestingly, therefore, part of both sides’ calculations depend upon the second mechanism and their respective expectations that A’s dovish shift will increase A’s willingness to settle and B’s hawkish shift will increase B’s willingness to escalate. Thus, in the case of interacting hawkish and dovish coalition shifts, it appears the two mechanisms work in tandem to create an overlapping bargaining space. This chapter’s plausibility probe examining the behavior of the Korean War’s primary belligerents provides support for this proposition. On the Communist side, the belligerents did lower their reservation price with a major concession about POW repatriation. This change in the Communist bargaining position was mostly caused by the two Communist dovish shifts (first mechanism) but also influenced by the expectation that Eisenhower’s hawkish shift would lead to war escalation and increased costs (second mechanism). It appears the hawkish shift mattered most because it was being viewed through the lens of the dovish shifts. Indeed, the Communists had not responded to the previous US hawkish shift in September 1950 by lowering their demands. In contrast, the Communist dovish shifts created a framing focused on costs—not only the costs of prolonging or escalating the war, but also the opportunity costs of not being able to pursue other policy priorities, given their projections about an adversary with a greater willingness to commit resources and political capital to win. In other words, the hawkish shift that brought Eisenhower to power was given more weight in Communist calculations because of the dovish shifts that occurred on the Communist side. This analysis accords with other scholarship about how doves look at escalation and risk (Leng 2000; also Snyder and Diesing 1977). In his study of recurring crises, Russell Leng (2000) suggests that how policymakers view the costs and risks associated with war “turns partly on their perceptions of the two sides’ comparative war-fighting capabilities . . . and partly on estimates of the costs of war.” While a hawkish worldview focuses on the risk of appearing weak, a dovish worldview is “more likely to focus on the risks of the escalation spiraling out of control” and the costs that would result (2000: 12). Given this focus on the costs of escalation, Leng argues it is not surprising that leaders with a dovish worldview will advocate conciliatory gestures, open lines of communication with the adversary and even unilateral concessions—such as the one about POW repatriation policy the Communists made in March 1953. While Leng argues the dovish worldview is fixed as part of leaders’ belief systems, this chapter has argued the dovish worldview can result from a dovish coalition shift, when leaders with more dovish preferences take power. But the effects of the dovish lens are the same.

interact ing d omest ic coalit ions



In contrast, the UN side’s behavior provides evidence of neither mechanism—when viewed monadically—because the United States did not change its bargaining position. It neither raised its demands after its own hawkish shift (first mechanism), nor raised them in response to the Communist dovish shifts (second mechanism). However, US behavior does support the argument about interacting coalition shifts on opposing sides. As Eisenhower’s calculations demonstrate, the marginal benefits of additional US demands were less than the marginal costs of fighting to achieve them. The Eisenhower coalition was only moderately hawkish and it expected continued conflict to be costly. Thus, it chose not to profit from its bargaining “advantage” by raising its demands—despite the Communists’ dovish shifts and signals of peaceful intentions. Instead, the United States held its demands constant but sent credible signals (albeit not nuclear ones) that unless these demands were met, it would escalate the war. As a result, the final armistice terms were much closer to the US bargaining position. This empirical finding accords with other research about hawkish constituencies in two-level games (Schelling 1980: 28–29; Milner and Rosendorff 1997). Leaders can use the constraint of a hawkish domestic legislature to extract an international agreement more in line with their interests. As Helen Milner and Peter Rosendorff explain, “The need for the hawkish legislature to ratify the agreement means that its hawkishness can be used by the executive to extract an offer it might not have gotten otherwise” (1997: 131). Milner and Rosendorff find that such a strategy only works for moderate ranges of hawkishness. When the domestic divisions grow too large, the executive loses in the international negotiation. While they advance this argument to explain outcomes in international political economy, it also appears to hold in the US case of the Korean War. Although Eisenhower himself was a moderate hawk, he was able to signal credibly to the Communists that he was constrained by hardliner hawks, at home and in South Korea. Eisenhower sent these signals—and the Chinese clearly received them—before the inauguration through his publicized meeting with Douglas MacArthur and his trip to Korea to consult with US commanders about military escalation in Korea and Taiwan. These moves not only placated the Republican right wing at home, but they signaled to the Communists that Eisenhower would not be able to back down from his negotiating position very easily. Likewise, as Chapter 3 explained, the Eisenhower coalition walked a fine line in trying to move the armistice agreement forward while pulling its resistant South Korean ally along. Although the Communists were well aware of the different interests and goals on the UN side before, they became especially cognizant of Eisenhower’s constraints when Syngman Rhee unilaterally released 25,000 POWs. Seeing these hawkish constraints on Eisenhower helped influence Communist expecta-

236

the korean war

tions about increased war costs should Eisenhower escalate. In other words, if the Communist concessions occurred in part because of Eisenhower’s hawkish shift, then these even-more-hawkish constraints on Eisenhower—the Republican right wing and Rhee—helped to motivate those concessions. On balance, the greatest explanatory power for the timing and terms of the Korean War armistice can be attributed to the Communist dovish shifts (first mechanism). However, the US hawkish shift did increase the effect of these dovish shifts by influencing Communist calculations about the war’s costs and benefits (second mechanism). The Communists were more able to learn from the information conveyed by the US hawkish shift—and change their bargaining position accordingly—because of their own dovish shifts. In addition, Eisenhower’s decision not to raise US war aims may have been a result of the peaceful signals after the Communist dovish shifts (second mechanism). Thus, the historical rec­ ord from the Korean War supports the proposition that in the case of interacting hawkish and dovish shifts, the two mechanisms can work in tandem to create an overlapping bargaining space. This chapter theorized explicitly about the interaction between hawkish and dovish shifts on opposing sides because this permutation existed in the Korean War. However, the hawkish-dovish interaction is just one of six possible permutations of the interaction of opposing sides’ coalition shifts. Moreover, one case study is certainly not conclusive. Further theorizing and empirical analysis of the other permutations is certainly needed. As mentioned previously, analyzing these other permutations is beyond the scope of this book. However, the fact that the hawkish-dovish plausibility probe accords with the empirical record of the Korean War suggests this is a fruitful area for future research. Finally, it is very difficult to disentangle the relative explanatory power of the two mechanisms. As this chapter pointed out, one way to begin to weigh their relative effects might be to differentiate between domestic coalition shifts in primary and secondary belligerents. Because primary belligerents are most important in their side’s decisions about prosecuting and ending war, we can imagine primary belligerents’ coalition shifts will be more likely to work through the first mechanism. In contrast, secondary belligerents may not have as much influence in changing their side’s war policy, and thus their coalition shifts may have more effect through the second mechanism. The next chapter allows us to begin to tease these two effects apart by presenting a quantitative analysis of interstate wars since 1862.

9

Domestic Coalition Shifts in War Termination since 1862

Up to this point, Part I presented a new theory about shifts in domestic governing coalitions and their role in ending interstate war. Part II traced the theory’s causal mechanisms through detailed case studies of the primary belligerents in the Korean War. In this chapter, I test whether this argument is generalizable beyond the “Forgotten War.” As the first part of this book suggested, there are many empirical examples of domestic governing shifts leading to war termination, and many scholars from different theoretical perspectives have noted this tendency. However, no one has previously attempted a broad quantitative analysis of this important argument, until now. This chapter tests the domestic coalition shift theory with a quantitative analysis of all interstate wars since 1862, using existing datasets (Bennett and Stam 1996; Goemans 2000a; Goemans et al. 2009) and a new dataset of post–World War II wars specifically constructed for this purpose. The analysis in this chapter draws on research conducted with my colleague John P. Sawyer (e.g. Stanley and Sawyer 2008); I am indebted to him for allowing me to use it here. This chapter’s findings corroborate and extend upon the empirical analysis of two similar arguments about the effect of leadership change on ending rivalries and conflict. First, in a parallel argument about interstate rivalry, Scott Bennett has argued that interstate “rivalry termination is expected when there is a change in leadership reflecting new preferences and new groups backing a leader” (1998: 1203). To test this argument, Bennett (1997, 1998) examined enduring rivalries between states from 1816 to 1992 and found empirical support for the hypothesis that a change in the leadership or political system of one of the rivals leads to a

240

test ing and extending the argument

decrease in the duration of the rivalry. Second, Sarah Croco (2008) has argued that “culpable” leaders—those whom the public can clearly link to the decision to involve the state in an interstate war—will feel a greater pressure to secure a favorable outcome than “non-culpable” leaders. While culpable leaders face a greater likelihood of domestic punishment should they lose, new non-culpable leaders can distance themselves from the original war decision and thus gain more options for ending it. Croco examined all leaders who participated in interstate wars from 1816 to 1991 and found support for these hypotheses: non-culpable leaders are more likely to preside over poor outcomes and are less likely to be punished for these outcomes than culpable leaders. Taken together with the findings I present in this chapter, Bennett’s (1997, 1998) and Croco’s (2008) empirical results provide powerful support for the argument that leadership changes play an important role in ending a variety of interstate conflicts. This chapter tests my domestic coalition shift theory with quantitative analysis of all interstate wars since 1862.1 To briefly recap the coalition shift argument in terms of the testable hypotheses presented in Chapter 2: H realpolitik: The greater the power imbalance between the two sides, the more quickly the war will end. H negative duration dependence: The longer the war has already lasted, the harder it is to end. H regime type: Regime type should have less explanatory power when coalition shifts are included in models of war duration. H shift: The longer the war, the more likely a domestic coalition shift is needed to end it. H loser: Domestic coalition shifts are most likely among “losing” and “stalemated” states. H primary shift: Domestic coalition shifts in primary belligerents will have a greater effect towards hastening the end of the war than shifts in other belligerents. Extending the basic model to examine the effect of the direction of coalition shifts on war termination yielded some additional hypotheses. The analysis in Chapter 8 suggested there are two mechanisms by which coalition shifts can affect the creation of an overlapping bargaining space. The first mechanism, which is captured by the first three hypotheses, occurs when the new governing coalition changes its own bargaining position. H hawk: Hawkish coalition shifts will occur early in the war and be associated with prolonging the war.

d omest ic coalit ion shifts



H dove: Dovish coalition shifts will occur late in the war and be associated with shortening the war. H neutral: Neutral shifts will shorten the war, but not as quickly as dovish shifts. The second mechanism occurs when the domestic coalition shift provides information for the opposing side to update its expectations about the war and thus change its bargaining position. This mechanism is captured by three alternative hypotheses: H hawk*: Hawkish coalition shifts will cause the other side to sue for peace and thus shorten the war. H dove*: Dovish coalition shifts will cause the other side to push for more concessions and thus lengthen the war. H neutral*: Neutral coalition shifts are unlikely to provide information about the war to the opposing side and thus should not affect the war’s duration through the second mechanism. As Chapter 8 explained, given the ambiguities inherent in international signaling, we would expect the second mechanism’s effect to be less powerful than the first’s. Moreover, the second mechanism can only work when the information obstacle is not blocking that side’s rational updating process. This chapter allows me to weigh the relative importance of these two mechanisms’ effects. This chapter is divided into five sections. The first section explains the methodology and operationalization of the model. The second section examines the hypotheses about domestic coalition shifts at the belligerent level with descriptive statistics. The third section tests the hypotheses with hazard analysis at the belligerent and war levels, while comparing the findings with existing empirical research on war duration and war termination. The fourth section provides a robustness check by examining each of the wars qualitatively. Then, after demonstrating the merit of the coalition shift argument, the fifth section examines what effect coalition shift directions have on the theory.

Methodology and Operationalization To test these hypotheses, I initially used war-level datasets from Bennett and Stam (1996) and Archigos (Goemans et al. 2009), to establish the theory’s viability over a wider sample of wars—interstate wars from 1862 to 1990. Then, I compiled a dataset of 78 belligerents in all but one of the 28 interstate wars between 1945 and 2006.2 In this analysis, because my argument pertains to domestic governing

242

test ing and extending the argument

coalitions, each belligerent serves as the subject, rather than the war itself.3 Like Ireland and Gartner (2001), I use a monadic approach—looking at states individually rather than in dyads—to predict war hazard functions. Later, to account for the strategic interaction between belligerents and the lack of independence between observations in the same war, I examine the wars qualitatively with war as the unit of analysis. The data largely were derived from the Correlates of War Inter-State War Data v3.0 (Sarkees 2000), although I made a number of corrections to improve historical accuracy, limit belligerents to major combatants and bring it up to date.4 The data about domestic coalition shifts were culled from primary and secondary sources about the domestic politics in each belligerent state during the wars. I limit my analysis to major combatants, defined as all belligerent states which either suffered at least 5 percent of the battle-deaths on its side in the war or served as the patron supplying the majority of the war material resources consumed by its side.5 In addition, consistent with the qualitative case studies of the Korean War, I define a primary belligerent as the main decision-maker involved in that side’s decisions about prosecuting and ending the war. Most wars have only one primary belligerent on each side, but some wars have more than one on each side—especially when there are great powers fighting together.6 This role can theoretically change over the course of the war. For example, the United States was the primary belligerent on its side in the Vietnam War until it left in 1973, at which time South Vietnam took over. Given the high degree of empirical detail necessary to code domestic governing coalition shifts, the data needed to test this argument empirically are quite extensive. Because information about domestic governing coalition composition and the timing of coalition shifts in wars after World War II is significantly more accessible, I selected this sub-sample of all interstate wars to conduct a more detailed probe of the argument. Theoretically, there should be no difference in the fundamental decision-making processes during war before that time, as the findings using the Archigos dataset demonstrate. However, there are some features of the post–World War II environment that make these wars stand out as a cohesive sub-sample for the more nuanced analysis of the domestic coalition shift theory. Recent scholarship on war termination documents a marked increase in the number of interstate wars ending in stalemate in the post–World War II period (Dunnigan and Martel 1987; Fortna 2004b; Pillar 1983; Smith 1995). While most other empirical studies of interstate war duration and termination have focused only on wars that ended conclusively, my model acknowledges the increasing numbers of stalemates since WWII and broadens the findings of the current literature by including them explicitly. This

d omest ic coalit ion shifts



increase in stalemates may be driven by some structural factors in the international system—the presence of nuclear weapons, the strength of the United Nations and its role in bringing interstate wars to settlement before a decisive victory and the dampening effect of bipolarity during the Cold War. In addition, Fortna (2004b) suggests that the increase in democracies in this era may be driving more willingness among belligerents to settle for stalemates. model specificat ion

I follow the tradition set out by Bennett and Stam (1996) of using survival (or hazard) analysis techniques to test hypotheses about war duration and termination. One important feature of hazard analysis is that it can model the “duration dependence” of a process under consideration (Bennett 1999). In the context of war duration and termination, duration dependence measures whether the prospect of a war terminating at any given duration increases or decreases over the course of its lifetime. Hazard models focus on the hazard rate, which is the instantaneous rate at which events terminate at time t, given that they have survived until t. Hazard rates may vary over time, and thus one important use of hazard models is correctly identifying the duration dependence of a process. I evaluated different survival models and chose the Weibull model for theoretical and statistical reasons.7 I use the Weibull distribution for the baseline hazard, for three reasons. First, the Weibull model allows me to compare my results to previous models (e.g. Bennett and Stam 1996; Goemans 2000a, b). Second, it appears to be statistically appropriate for the data. 8 Finally, and most importantly, it allows me to test my theory that wars are negatively duration dependent. Because the domestic coalition shift theory tells us what kind of duration dependence to expect (negative duration dependence), there are strong theoretical reasons for choosing the Weibull based on the theory itself (Box-Steffensmeier and Zorn 1998: FN 5; Bennett 1999: 261; Cleves et al. 2004: 246). As Scott Bennett argues, “Because duration dependence has clear substantive implications, parametric duration models may actually be chosen specifically to test particular hypotheses” (1999: 261). A Weibull model approximates duration dependence through the duration parameter p. When p=1, there is no duration dependence. If the hazard rate is positive (p> 1), the hazard rate increases over time and an event becomes more likely. In such cases, termination accelerates over time, and an event is more likely to end in each subsequent interval. If the hazard rate is negative (0