Über die Autorin bzw. den Autor

H. E. Goemans is Assistant Professor of Political Science at Duke University.

Von der hinteren Coverseite

"War and Punishmentis a refreshing combination of rigorous reasoning and elegant historical analysis on one of the most important, and least understood, aspects of war: what causes wars to end. Anyone interested in the interplay between leadership, domestic politics, and international conflict must read this book. It is a superb and enlightening analysis."--Bruce Bueno de Mesquita, Senior Fellow, Hoover Institution, Stanford University

"A fresh and thought-provoking analysis of why the First World War lasted so long and of what caused it to end. Goemans bridges the gap between international history and security studies, offering challenging conclusions for experts in both disciplines. His book deserves a wide readership."--David Stevenson, Professor of International History, London School of Economics and Political Science

"This may very well be the best book written in the last twenty years by a political scientist on war termination, war aims, and World War I. Goemans presents a formidable challenge to the way that we think about the prosecution of war, war duration, and the links between these and regime type. Not only does the book contain tremendously valuable theoretical insights, but it also provides compelling empirical evidence."--Allan C. Stam, Yale University

Auszug. © Genehmigter Nachdruck. Alle Rechte vorbehalten.

WAR AND PUNISHMENT

PRINCETON UNIVERSITY PRESS

Contents

List of Tables and Figures...................................................................viiAcknowledgments..............................................................................ixCHAPTER ONE Introduction.....................................................................3CHAPTER TWO A Theory of War Termination......................................................19CHAPTER THREE Large N: The Fate of Leaders and the Duration of War...........................53CHAPTER FOUR Germany.........................................................................72CHAPTER FIVE Russia..........................................................................122CHAPTER SIX France...........................................................................134CHAPTER SEVEN Great Britain..................................................................181CHAPTER EIGHT The Termination of the First World War, I: 1914–1917.....................232CHAPTER NINE The Termination of the First World War, II: 1918................................270CHAPTER TEN Conclusion.......................................................................310Bibliography.................................................................................325Index........................................................................................351

Chapter One

Of war men will ask its outcome, not its causes (Seneca)

The Central Question

Now, as always, states fight wars. As one of the most destructive forms of human behavior, war and its study lie at the heart of the discipline of international relations. It is not surprising, therefore, that much theoretical work has been done to explain the causes of war initiation. What is surprising, however, is the paucity of effort to understand and explain the causes of war termination. While it is interesting and important to know why wars break out in the first place, it is no less interesting or important to know why it often takes so much time and such enormous costs before wars end. Why, for example, did the First World War not end after the Schlieffen Plan failed? This book therefore seeks to answer the question, asked more than twenty years ago by William Fox: "What keeps wars going and what finally makes them stop?"

The Central Argument

The answer to this question, the central argument of the book, is laid out in chapter 2. Wars can end only when the minimum terms of settlement of both sides become compatible, when both are asking no more than the other side is willing to give up. Therefore, the fundamental cause of war termination is a change in the minimal demands, that is, the minimum terms of settlement of the combatants. Starting with the rational unitary actor approach, the theory developed in chapter 2 first argues that changes in a nation's minimum terms of settlement are a function of new information about the outcome and the costs of war. But not all states change their war aims according to this logic because domestic politics can fundamentally affect how different regimes change their war aims. The second half of chapter 2 therefore builds on the rational unitary actor theory and offers a theory that explains why some regimes refuse to lower their war aims—indeed, sometimes increase their war aims—even when they learn they will probably lose. Specifically, semirepressive, moderately exclusionary regimes will refuse to lower their war aims and keep fighting where nonrepressive, nonexclusionary and repressive, exclusionary regimes would prefer to lower their aims and settle because leaders of semirepressive, moderately exclusionary regimes anticipate severe domestic punishment for even a moderate defeat. Thus, the main contribution of this book is the theoretical argument and empirical evidence that the decision to continue fighting or settle depends on the nature of the domestic political regime.

Competing Explanations

An exhaustive search through the existing literature on war termination yielded no well-articulated and testable competing theories. However, individual, disconnected, and sometimes contradictory propositions and generalizations abound. The existing propositions in the literature on war termination can be divided into two main strands. The first strand argues that war ends when the "loser" recognizes his position and accedes to the "winner's" demands. In other words, war ends when one side, the "loser," cries uncle. The crucial assumption in this strand of the literature is that war has only two outcomes: victory or defeat. War ends when the loser gives up his demands and accepts the demands of the winner.

The second strand argues that war ends when the combatants agree to a settlement that both prefer over continued fighting. In other words, war is seen as a form of bargaining in which states make rational cost/ benefit calculations about war and its termination and maximize their expected value. This strand of the literature maintains that demands or war aims of both sides fluctuate during war, and that the final terms of settlement may be totally different from the combatants' original aims. In other words, the terms of settlement are endogenous. Thus, this strand of the literature maintains that there exist an infinite number of possible outcomes of war because there exist infinite combinations of potential terms of settlement.

One-Sided Termination: "The Vanquished Makes Peace"

At first glance the argument that war ends when the loser cries uncle seems plausible. Its plausibility stems from the commonly held conviction that it takes only one side to start a war. It is assumed, therefore, that it also takes only one side to end it. This view is still widely held in and outside of academia.

If one aggressor can start a war by a sudden attack, this argument goes, war will simply stop when the aggressor recants or the loser submits. Calahan, for example, wrote in 1944 that "war is pressed by the victor, but peace is made by the vanquished. Therefore, to determine the causes of peace, it is always necessary to take the vanquished's point of view. Until the vanquished quits, the war goes on."

In a similar vein Francis A. Beer and Thomas F. Mayer argued in 1986 that "[t]he major proposition of rational choice theory on how wars end is ...: Wars terminate because one or more combatants decide they have more to gain, or less to lose, by making peace than by continuing to fight." Beer and Mayer's proposition implies that if one side decides it has more to gain by making peace than by continuing to fight—in other words, when the vanquished quits—war will end. Many other students of war termination, such as Robert Randle, Paul Kecskemeti, Lewis Coser, Nicholas S. Timasheff, and Hobbs, all either implicitly or explicitly argue that war ends when the vanquished quits.

More recently Allan Stam has offered a theory of war termination that argues that "war ends when one or both sides see costs as exceeding benefits." Stam argues that war has three possible outcomes: one side quits, the other side quits, or both quit—as he puts it in the title of his book, Win, Lose, or Draw. Among the scholars who have tried to operationalize their version of the "vanquished makes peace" argument, such as Frank Klingberg and Lewis Richardson, Stam's work stands out. While Klingberg and Richardson have tried to predict when the loser will quit by focusing on battle deaths, Stam and his coauthor Bennett offer a far richer model. In their article Bennett and Stam do not try to offer a model or a test of the causes of war termination but focus on the closely related topic of war duration. Overcoming difficult data problems and incorporating tremendous amounts of research, Bennett and Stam present a rich model that includes variables such as strategy, terrain, capabilities, and government type and that generates strong and surprising statistical support.

A crucial assumption in their model is that

wars occur over some issue, with some fundamental benefit at stake, such as a piece of territory, economic influence over an area, or some government policy. We also assume these issues to be fixed and exogenous, despite the possible manipulation of the apparent stakes by political leaders. Unlike benefits, the costs of continuing a war rise over time.

This assumption that the stakes of the war are fixed also defines the arguments of the scholars who propose variants of the vanquished makes peace argument. Richardson even went so far as to argue that the "possible terms of peace" do not significantly affect the course of the war and the prospects of its termination. The assumption that war aims are fixed poses a fundamental puzzle.

Why the Argument Is Flawed

The above argument claims that war ends when the "loser" decides to submit to the "winner's" demands. This argument is fundamentally flawed because it simply assumes that the winner does not raise his demands once he realizes his advantage. This model of war termination describes only the extreme and relatively rare case of unconditional surrender or "absolute war," where the winner cannot possibly ask for more. In most cases, however, a winner can raise his demands if he discovers an unexpected advantage.

During the First World War the negotiations at Brest-Litovsk provide a classic example of the flaw in this argument. On February 10, 1918, some among the new Bolshevik leaders, in particular Trotsky, thought Russia could end the war on the Eastern Front simply by refusing to fight; literally the vanquished quit. The Germans, however, thought otherwise and on the 16th announced they would continue the war. Facing little opposition, they quickly and easily advanced. After the Bolsheviks formally announced on February 21 that they would accept Germany's terms, the Germans issued new and harsher terms on the 23d. On the 24th the Bolsheviks signaled their acceptance of the new terms.

The argument that the vanquished makes peace, thus, leaves unanswered this theoretical puzzle: When a "winner" realizes his advantage, why would he not raise his demands? While empirical examples where states do not raise their war aims exist, for example, the coalition war against Iraq in 1990, they beg the question. Why did the United States not raise its demands after its brilliantly successful and relatively costless military victory in the war against Iraq? Note that this is also a clear case where it was the winner who decided to end the war. Iraq would have been happy for the fighting to have stopped earlier.

Aside from its conceptual shortcomings, the argument that it takes only one side to end a war also suffers from poor implementation and testing. The proposition is theoretically underspecified because it provides no mechanism that causes a "loser" to cry uncle. This makes it very difficult to empirically test the claim. Klingberg and Richardson have both tried to provide mechanisms, mainly by focusing on the number of battle deaths, but no general relationship between fatalities and war termination has been found. Moreover, the argument is also difficult to test since it is problematic to identify the loser ex ante, that is, independent of the outcome.

Two-Sided Termination: Both Must Prefer Peace over War

The second current explanation of war termination is that it takes both sides to end a war. Intuitively, and with the example of Brest-Litovsk in mind, this argument seems to be much more compelling. However, as the argument currently stands in the literature, it is incomplete at best. The best statement and elaboration of this argument can be found in an excellent article by Donald Wittman. He argues that "An agreement (either explicit or implicit) to end a war cannot be reached unless the agreement makes both sides better off; for each country the expected utility of continuing the war must be less than the expected utility of the settlement."

Some authors, such as Geoffrey Blainey, Paul Pillar, and R. Harrison Wagner, most likely endorse this statement in toto. Even some wary of the rational unitary actor approach, such as Janice Gross Stein and Charles Fred Ikle, would probably largely agree with Wittman's statement.

Why the Argument Is Flawed

This second argument claims that states end a war because they reach an agreement that makes both sides better off than continued fighting. The argument is incomplete for two reasons. First, it provides no mechanism that brings such an agreement in reach for both sides and fails to explain why a bargaining space opens up. Second, it completely ignores the enforcement problems of any agreement to end war. Thus, the argument overlooks the fundamental anarchic nature of international relations. How can agreements that change the relative power of the antagonists be enforced? The argument truncates the problem by omitting considerations of the possibility of future war and other potentially dangerous consequences of changes in relative power.

Focusing on the first problem, most authors assume (implicitly or explicitly) that when the belligerents agree more about the outcome on the battlefield, an agreement to end war becomes more likely. In other words, it is assumed that when belligerents agree more on their estimates of their probability of winning it becomes more likely that a bargaining space opens up. However, Wittman's analysis suggests that even when both sides agree more on the outcome of the war, settlement need not become more likely. The combatants adjust their estimates of their probability of winning as a result of new information and unexpected events on the battlefield. When one side raises its estimate of its probability of winning, its reservation value increases, and it will therefore raise its demands. When the other side lowers his estimate of his probability of winning, he will lower his demands. Both sides now agree more on their probability of winning; however, it is not necessarily true that the difference in their demands diminishes. Agreement becomes more likely and a bargaining space is more likely to open if and only if one side lowers his demands more than the other side increases his demands. Thus, more agreement about the outcome does not automatically make an agreement to terminate the war more likely.

The second theoretical problem is a familiar one from the literature on international cooperation: How can two states reach an agreement in an anarchic realm with no central authority to enforce its terms? This problem should be particularly severe in agreements to end a war because the settlement of a war almost always entails some shift in relative power between the belligerents. Simply put, the "winner" has more power (which is the reason he is winning), and will get still more power in the settlement. Therefore, the winner is in a better position to demand further concessions from the loser later on, whereas the loser will be in an even worse position to resist. This argument implies that no state will accept any agreement that diminishes its relative power, without some sort of insurance that the enemy will make no further claims. In other words, the stronger side must credibly commit itself not to exploit its bargaining advantage in the future. Where can such insurance come from?

A historical example serves to illustrate the problem. When Hitler demanded the Sudetenland from Czechoslovakia in the 1938 Munich crisis, he promised this would be his last demand. When the Czechoslovakians conceded, they ceded territory to Germany that contained the major defensive works along the German-Czechoslovakian border. The ensuing shift in relative power not only made it easier for the Germans to invade the "rump" of Czechoslovakia in March 1939 when Hitler's promise could not be enforced but also made it easier to fight the Battle of France in 1940.

Although the argument that war termination requires an agreement that leaves both sides better off than continued war has much appeal, it fails to specify a mechanism that creates the necessary preconditions for any such agreement. These necessary preconditions are, first, the creation of a bargaining space (which produces the possibility for an agreement that leaves both better off) and, second, a credible commitment from the stronger side that it will not raise further demands in the future. In the next chapter I show how these puzzles can be solved.

(Continues...)

Excerpted from WAR AND PUNISHMENTby H. E. Goemans Copyright © 2000 by PRINCETON UNIVERSITY PRESS. Excerpted by permission of PRINCETON UNIVERSITY PRESS. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.