Acknowledgments
I FIRST STARTED THINKING about how wars are fought and ended in the mid-1990s. Allan Stam and I were hard at work on our project explaining why democracies win wars, and in 1997 we presented one of the early papers from that project at a special conference on the conduct of war organized by Scott Gartner at the University of California, Davis. Several important papers on the prosecution and termination of war appeared at that conference, including a paper by Hein Goemans (a precursor to his major war-termination book War and Punishment), the first in a string of papers connecting wartime casualties to public opinion by Scott Gartner and Gary Segura, a quantitative paper by Suzanne Werner explaining why wars end in absolute or limited outcomes, a quantitative paper by Allan Stam and Scott Bennett on the duration of wars, and one of the earliest theoretical papers explaining bargaining during war by Alastair Smith, among others. Donald Wittman, author of a critical and often underappreciated 1979 Journal of Conflict article on war termination, appropriately served as discussant. This collection of papers helped establish and advance the developing theoretical and empirical agenda for understanding how wars are fought, won, and ended.
Around this time, the field was just beginning to realize how powerful the bargaining metaphor is for describing war. Harrison Wagner’s 2000 American Journal of Political Science article, which built on James Fearon’s seminal 1995 International Organization article, really opened my eyes to the power of thinking about war as a bargaining process. I was invited to a 2001 conference organized by Alastair Smith at Yale on the bargaining model of war, for which I wrote a paper that reviewed the bargaining model of war literature. That conference included a number of important papers by scholars that pushed the boundaries in this area. Some of the same people who were at the 1997 Davis conference attended, plus others such as Robert Powell, Barbara Walter, and William Reed. The 2001 Yale conference produced a number of important papers that went on to be published in top journals and to shape the field. My review paper was probably the least ambitious in the group, and eventually was published in Perspectives on Politics in 2003.
It struck me around this time that although the bargaining model of war was enjoying extensive theoretical development, the empirical side was lagging. Do wars actually play out as the bargaining model portrays? My first inclination was to construct a gigantic quantitative dataset of intrawar behavior, collecting data in particular on the outcomes of battles within wars. I knew that such a project would likely consume hundreds of thousands of dollars in grant money and take several years. However, two problems soon appeared. First, my initial grant requests were rejected. Second, the more I thought about applying quantitative/statistical techniques to tackle this issue, the more I concluded that the applications of such techniques faced serious problems. Most importantly, the data were too messy and wars were too heterogeneous. I recount some of my thinking on these issues in chapter 4 of this book.
I also started to think about the theoretical structure of the bargaining model of war. I scratched out one or two ideas in a paper I presented at a 2003 special conference on military effectiveness hosted by Risa Brooks at Northwestern University. That paper received what can kindly be described as a lukewarm reception, and I switched topics for the next conference in the Northwestern project (I ending up writing a paper on Japanese military effectiveness in World War II; working on that paper helped prepare me to write chapter 10 in this book).
But I kept thinking about the theoretical side of the bargaining model. I went back to the drawing board on theory, and sketched out a few more ideas. I wrote another paper laying out those ideas and then illustrating them with a case study of the American Civil War. I presented that paper at the 2005 meeting of the Peace Science Society (that paper provided some of the foundations for chapter 8 in this book). This paper also attracted little interest, and I began to have doubts about what to do next.
After the Peace Science meeting, I decided that another way to go might be to develop at least one of my theoretical insights within the structure of a formal model. I selected one of the half dozen or so theoretical ideas I had been batting around, and roped a colleague and a graduate student (Cliff Carrubba and Scott Wolford) into coauthoring a paper with me presenting a formal model of the basic theoretical point that fear of a belligerent reneging on a war-ending agreement may affect wartime bargaining behavior (this insight is presented informally in chapter 3 of this book). We presented this paper at the 2006 Midwest Political Science Association meeting and a few other places and spent the next couple of years getting feedback and revising it (as of this writing, the paper is under review for publication). I began to have faith there was an important theoretical point to be made.
However, I could not let go of the empirical side. I was torn. I had long believed that quantitative analysis was the most productive way to test scientific propositions, but my doubts about executing quantitative tests in this area remained substantial. To my great luck, I had coincidentally agreed to teach summer school at Dartmouth College in 2006. While in Hanover, I had several discussions with various members of the faculty in Dartmouth’s government department about the project, both the theoretical and empirical sides. From my Dartmouth colleagues, I received useful feedback and general encouragement that a case study approach would make a contribution. As is often the case, the most useful advice came here and there, while shooting the breeze in someone’s office, grabbing a hamburger for lunch, or strolling to Fenway Park in Boston. Reinvigorated, I returned to Atlanta that August ready to tear into the project.
The bulk of the book gushed out in the 2006–7 academic year. Once I was ready to let go of quantitative tests and really embrace case studies, the pieces quickly fell together. My colleagues at Dartmouth did me another great favor, agreeing to attend a small conference on the first draft of the book manuscript, offering me feedback. The conference took place in summer 2007, and I was again blessed with considerable encouragement, constructive feedback, and ideas for new directions. Around this time, I finished a somewhat related quantitative empirical paper, coauthored with two Emory graduate students, Nigel Lo and Barry Hashimoto. This paper provided quantitative empirical support for the model’s empirical assumption that the imposition of foreign-imposed regime change would significantly decrease the likelihood of war between two former belligerents. We presented this paper at the 2007 annual meeting of the American Political Science Association. It was published in International Organization in 2008.
As is likely evident, I have accumulated substantial personal debts in writing this book. Thanks to Virginia Lewick of the Roosevelt Presidential Library in Hyde Park, New York. David Clark and the rest of the staff at the Truman Presidential Library in Independence, Missouri were extremely helpful. For research assistance of all types, my thanks to Rochdi Alloui, Barry Hashimoto, Aryeh Reiter, Chana Reiter, John Reuter, Emily Hencken Ritter, Phil Fuerst, Jakub Kakietek, Maryann Gallagher, Matthew Payne, Michiyo Sakai, Goran Peic, and Terry Chapman. I received helpful comments when portions of this project were presented at the 2005 Peace Science Society (International) meeting, the 2006 Midwest Political Science Association meeting, the University of Chicago, the University of South Carolina, Emory University, Yale University, Duke University, and Princeton University. For comments on various stages of this project, I’d also like to thank Ashley Leeds, James Morrow, Hein Goemans, and Bumba Mukherjee. Thanks as well to Bill Wohlforth and the Dickey Center at Dartmouth and the Institute for Comparative and International Studies at Emory for supporting a small conference at Dartmouth on the book project. Participants at the book conference, including Allan Stam, Ben Valentino, Bridget Coggins, Daryl Press, Jenny Lind, and Stephen Brooks all provided tremendously useful feedback. Thanks to Amy Benson Brown for her feedback on the writing and style of the manuscript. Thanks to Randy Strahan, Kyle Beardsley, Jack Snyder, and Tanisha Fazal for graciously reading the entire manuscript and providing feedback. Thanks to my coauthors, Cliff Carrubba, Barry Hashimoto, Nigel Lo, and Scott Wolford, for working with me on papers that served as pieces for various parts of the book. Thanks to Shane Kelley for drawing such useful maps. Thanks once again to Chuck Myers of Princeton University Press for his always helpful editorial guidance. This research was supported in part by the University Research Committee and the Bill and Carol Fox Center for Humanistic Inquiry, both of Emory University. All opinions and any remaining errors are of course exclusively my own. My deepest gratitude, however, is saved, as always, for my wife Carolann, for all her patient support.
I left out various exhortations, notions, original and hortatory allusions, notes, denials; zeal exceeded volume. Such omissions notwithstanding, I dedicate this book with love to my sons, Noah Charen Reiter and Zev Franklin Reiter. When writing this book, I often thought of them and John Adams’ famous remark, “I must study politics and war that my sons may have liberty to study mathematics and philosophy. My sons ought to study mathematics and philosophy, geography, natural history, naval architecture, navigation, commerce, and agriculture in order to give their children a right to study painting, poetry, music, architecture, statuary, tapestry, and porcelain.” I must confess that unlike John Adams I find the study of politics and war as joyful as the study of poetry, music, or anything else. I wish my sons great luck in finding their own paths to joy.
How Wars End
CHAPTER ONE
Love is like war; easy to begin but very hard to stop.
—H. L. Mencken
HOW, WHEN, AND WHY do belligerents end wars? Why do some losing belligerents, such as the United States in the early months of the Korean War, the Confederacy in the twilight of the American Civil War, Britain during the dark night of May 1940, and the United States in the first months of World War II, refuse to consider negotiating to end their wars on acceptable terms and instead fight on in pursuit of victory? Why do some winning belligerents, such as the Soviet Union in the latter months of its 1939–40 and 1941–44 wars against Finland, elect to stop fighting and accepted limited gains rather than fight onward in pursuit of the total defeat of the adversary?
This book seeks to solve these and other puzzles regarding the termination of wars. Why are some belligerents willing to end wars on limited terms? Why do some belligerents refuse to end war short of total victory? What factors push belligerents to demand more or less of the adversary at the negotiating table while war is raging? We know relatively little about how wars end, in contrast to the mountain ranges of ideas and scholarship we have about how wars start. Indeed, there has been something of an aversion to the study of war termination over the past several decades. The total nature of World War II seemed a denial of the political significance of war termination since in an era of total war the belligerents fight with all their resources until one side is utterly crushed. This neglect of war termination persisted through the Cold War, when most assumed that any major war would quickly escalate to nuclear attacks and Armageddon, making the topic of war termination a grim joke. Interest in war termination received little boost by the Vietnam War, as the bad taste left by that conflict encouraged thinking about stopping wars from happening rather than stopping wars once they have started. Beyond work of purely historical interest, such as the voluminous literature on why Japan surrendered in 1945, relatively few works exist that consider the question of war termination more generally.1
War termination must receive closer attention. The end of the Cold War did not bring the “end of history” and the end of war as some had forecast and fervently hoped for. In the twenty-first century, policymakers must understand how to end the wars that their nations have become involved in, and they must have better tools to help end wars being fought by other nations. On the scholarly side, a full understanding of war, if not international relations more broadly, requires developing theoretical structures that provide integrated accounts of all aspects of war, its duration, outcome, termination, and post-war phase, as well as initiation. The theory of war termination provided here is developed in the context of comprehensive ideas about the nature of war and politics, and offers progress towards developing a grand unified theory of war that would provide an integrated account of all of war’s phases.
THE NATURE OF WAR, THE NATURE OF WAR TERMINATION
War termination, war, and even international relations itself can be usefully conceived as a bargaining process, following Thomas Schelling’s famous insight that “most conflict situations are essentially bargaining situations.”2 States clash over issues such as the placement of international borders and the composition of national governments, and bargain with each over how these issues should be settled (for example, where an international border should lie). War is part of this bargaining process, since states start wars to get more of what they want on disputed issues, and the ends of wars are literally bargains struck to create new settlements of the disputed issues. Underlying this notion of war as bargaining is the assumption that war is fought for political goals. As the Prussian military thinker Carl von Clausewitz memorably put it, “war is nothing but the continuation of policy with other means.”3
This book presents a bargaining theory of war termination. A central advantage of using a bargaining approach to war termination is that it nicely incorporates two of our most profound and powerful insights about international relations: that uncertainty about the power and intentions of states pervades the international system, and that states cannot make binding commitments to each other. The bargaining approach proposes that the problems of uncertainty and unenforceable commitments cause war, fighting war serves to alleviate these problems, and war ends when these problems have been reduced sufficiently or eliminated.
Uncertainty and incomplete information comprise a long-standing account of how wars break out. Specifically, when two states in dispute disagree about the balance of power or the relative steadfastness of each side to prevail, war may result, especially if each side is confident it can prevail in a clash of arms. The outcomes of combat and the decisions of states to make concessions or hold fast help reduce disagreement between the two sides about the balances of power and resolve. Eventually, when enough disagreement has been reduced and the two sides are in sufficient accord about the true balances of power and resolve, then war-ending agreement becomes possible. In short, uncertainty causes war, combat provides information and reduces uncertainty, and war ends when enough information has been provided. A central information-based wartermination hypothesis is that a belligerent will be more likely to make concessions to its adversary in pursuit of war termination following combat defeats, and following combat successes it is less likely to make concessions and may even present new demands of its adversary. This well-established insight is described and developed in chapter 2.
The exposition of a commitment-based explanation of war termination, the heart of chapter 3, is this book’s greatest theoretical contribution. The chapter considers a core problem with war-termination settlements. Even if two combatants sign an agreement to end their war, the commitment of each not to attack the other after war’s end is not enforceable. There is no world government with the authority or power to compel two former belligerents to comply with the terms of a war-ending agreement. The lack of such a government is the central difference between international politics and national politics. Of course, scholars as far back as the ancient Greek Thucydides have used the commitment insight to explain the causes of war, proposing in particular that a changing balance of power between two states can make war more likely, because the nation growing in relative power cannot credibly commit not to attack in the future.4 The commitment insight has been fruitfully applied to explain the dynamics of civil wars, because civil war belligerents, governments, and rebel groups gravely fear that the other side will violate a warending agreement, especially if the deal calls for one side to lay down its arms. This commitment fear explains why civil wars last so long, and also why third-party intervention is so important for helping achieve a negotiated settlement.5
Although the existing commitments scholarship has provided some very useful insights and advances, it leaves essentially unasked questions of how war (and specifically war outcomes) can solve commitment compliance problems. This book pushes these core insights about commitments and war termination in new directions, examining interstate as well as civil wars and moving beyond viewing commitment noncompliance fears as simply barriers to war termination. Specifically, this book explores how these noncompliance fears shape war-termination decisionmaking during wartime, and relatedly how states pursue certain war outcomes in order to solve these noncompliance problems.
One central point is that total victory can solve commitment problems. Sometimes a belligerent can achieve what Clausewitz called an “absolute” war outcome, utterly vanquishing the adversary. In application, this can mean the victor installing a new leadership in the defeated state, occupying or annexing the adversary’s territory, or at worst annihilating the adversary’s entire population. All of these outcomes permit the belligerent to impose essentially whatever war outcome it wishes. These outcomes also permit the belligerent to remove in effect or in fact the adversary’s ability to violate the agreement, thereby solving the compliance commitment problem. Liquidating the adversary through annexation or genocide, of course, directly prevents breaking the commitment to peace. Foreign imposed regime change, a more frequent war outcome, substantially decreases the likelihood that the war-ending agreement will be broken, since an installed puppet is generally unlikely to deviate from the demands of the puppet-master victor. The central proposition in chapter 3 is that the more a belligerent fears its adversary may violate the war-ending terms of an agreement, the more likely the belligerent will be to pursue an absolute victory.
War can also solve the commitment problem with more limited outcomes that fall short of absolute victory. Sometimes the terms of a limited war outcome directly alter the balance of power, as when a peace treaty transfers strategic territory from one side to another, or when a peace treaty demands that one side abandon or eschew a powerful weapon. A belligerent, fearful that its adversary may reattack in the future, may be willing to accept such a limited war outcome that shifts the balance of power and makes a future attack less appealing to the adversary.
This book integrates the information and commitment perspectives on war termination into a single theory, towards understanding how information and commitment dynamics interact in affecting war-termination decision-making. Importantly, these two dynamics sometimes operate in tension with each other. For example, credible commitment fears may delay war termination, even as uncertainty and disagreement are being reduced. Specifically, under some conditions severe credible commitment fears may cause a state to seek absolute victory over its opponent even in the face of battlefield defeat. This prediction runs counter to an information-only view of war termination, that battlefield defeats should always make a belligerent consider making diplomatic concessions as a means of hastening war’s end. Conversely, the reduction of uncertainty may make war termination possible, despite a belligerent’s enduring concerns about commitment credibility. As a belligerent comes to learn that continuing the war will pose escalating costs and dangers, it may become willing to accept a peace settlement, even if such a settlement leaves standing an untrustworthy adversary.
The theoretical portion of this book provides a structure for understanding when information dynamics are likely to determine war-termination behavior, and when commitment dynamics are likely to determine war-termination behavior. Specifically, belligerents fearful of a credible commitment problem are more likely to fight on to solve the commitment problem when the costs of continuing to fight are acceptable, when the dangers of a broken war-ending agreement grow, and/or when there is at least some hope of eventual victory. Conversely, belligerents fearful of a credible commitment problem are more likely to accept war termination short of total victory and a completely satisfying solution to the commitment problem when the costs of continuing to fight threaten to escalate significantly, when the dangers one side will break a war-ending agreement are lower, when the chances of eventual victory approach zero, and/ or when the belligerent is able to reduce the commitment problem through limited means, such as acquiring strategic territory.
THE TERMINATION OF ACTUAL WARS
Any understanding of war must be grounded in the empirical. How have wars actually ended? How well does actual behavior square with our theoretical expectations?
I provide some of the first empirical evidence that directly assesses how belligerents try to end wars, and whether and how war-termination behavior is shaped by information and commitment dynamics. Specifically, a belligerent who loses battles is supposed to downgrade its estimates of its own military power, and be more willing to offer concessions. Is this what actually happens? Does war-termination diplomacy tend to follow battlefield outcomes in this manner? If not, are there conditions under which this pattern is observed, and conditions under which it is not? While there is a small body of research that looks for empirical patterns indirectly implied by the growing theoretical scholarship on bargaining and war, almost no work exists that attempts to assess more directly whether the exact processes forecast by these and other models, such as battlefield defeat causing pessimism and thus diplomatic concessions, actually occur during wartime.
This book examines war-termination decisions and dynamics in an array of wars, including the Korean War, the Allies in World War II, Japan during World War II, Germany during World War I, the Union and Confederacy during the American Civil War, and Finland and the Soviet Union during both the Winter War and the Continuation War. These cases are neither a complete sample of all belligerents during wars nor are they a random sample. However, they do represent a wide range of historical/ political/military contexts since they include long wars and short wars, both civil and international wars, wars between equal powers and between unequal powers, wars fought to the finish of unconditional surrender and wars fought to more limited outcomes, wars fought in a variety of regions including North America, Europe, and East Asia, and wars fought across nearly a century of time, from the 1860s to the 1950s.
Analysis of these wars permits exploration of a number of specific historical war-termination puzzles, beyond those enumerated at the beginning of this chapter. Some of these puzzles include:
• How did the Soviet Union react to the December 1939 battlefield disasters it faced against Finland during the Winter War?
• How did the Soviet Union react in the first few months after the June 1941 German invasion, when the capture of Moscow appeared likely and imminent?
• How did President Lincoln react to the apparent collapse of support for the Union war effort in summer 1864, when continuation of the war seemed to ensure his electoral defeat that November?
• Why did Lincoln embrace the emancipation of the slaves, an act that raised the Union’s war aims, in autumn 1862?
• Why did Germany, as it was defeating Russia in the East, reject Allied peace overtures in the winter of 1917–18 and instead choose to continue the war by renewing its offensive in the West?
• How did Japan react to the steady slide in its military fortunes from 1943 to 1945, and why did it eventually accept near unconditional surrender in August 1945?
The empirical results in this book are complex and nuanced. Both commitment and information dynamics play important roles in determining war-termination behavior. The results provide support for the most novel theoretical proposition of the model, that credible commitment fears help determine war-termination behavior. Belligerents sometimes press for absolute victory because they fear this is the only true solution to an enduring credible commitment problem. The results also indicate that information dynamics play an important role in determining war-termination behavior, although perhaps in contrast to some conventional thinking, information cannot provide a complete account of how states end wars. Sometimes belligerents pursue absolute victory to solve commitment problems, even in the face of combat setbacks. Conversely, though, not all belligerents who fear that the adversary will break a war-ending agreement decide to pursue absolute victory. Belligerents are less likely to pursue absolute victory if they fear the costs of war will escalate gravely (perhaps because of third-party intervention) if such an outcome is pursued, if breaking the war-ending commitment has less than catastrophic consequences, if they see the chances of eventual victory as approaching zero, and/or if the vagaries of geography or military technology make a limited war outcome an acceptable if partial solution to a credible commitment problem.
There are other findings as well. The fog of war and the patience of leaders also sometimes sever the connection between battle outcomes and war-termination behavior. Additionally, the cases indicate the curious insignificance of domestic politics in war-termination decision-making, in contrast to existing theoretical and empirical research on how democracies fight wars. There is little evidence that democratic leaders are especially casualty-sensitive, and hence are more likely than other kinds of leaders to consider concessions as casualties mount. There is also little evidence for the specialized hypothesis that leaders of semirepressive, moderately exclusionary regimes are likely to raise their war aims when their states are losing in order to avoid facing the severe personal punishment that awaits them in the event of defeat in war. That being said, there is some evidence for a perhaps related pattern of behavior, that belligerents with weak civilian control of the military may fight longer wars.
OUTLINE OF THE BOOK
The structure of the book is as follows. The theory of war termination is developed in chapters 2 and 3. Chapter 2 lays out the main ideas and propositions of the information approach. Chapter 3 develops a number of ideas about how commitment concerns shape war-termination behavior. Chapter 4 addresses a number of issues regarding research design and methodology. Chapters 5 through 10 present the historical case studies.
Chapter 11 seeks to accomplish two tasks. First, it offers a summary of the book’s empirical findings, relating them back to the theoretical ideas laid out in chapters 2 and 3, and to the study of international relations more broadly. Second, it applies the empirical findings to problems of American foreign policy in the twenty-first century. After 9/11, the George W. Bush Administration thought about wars in the context of commitment problems, that rogue states like Iraq could not be trusted to adhere to their international commitments, meaning that war culminating in absolute victory, such as foreign imposed regime change, may be the only way of assuring American national security. Some have also suggested lesser means of solving commitment problems faced by the United States, including launching air strikes against rogue states’ weapons of mass destruction production facilities. Chapter 11 discusses and evaluates both the regime change and airstrikes strategies, framing them within the book’s general theory.