The War Ledger 9780226351841

Citation preview

The War Ledger

A. F. K. Organski Jacek Kugler

The War Ledger

The University of Chicago Press Chicago and London

The University of Chicago Press, Chicago 60637 The University of Chicago Press, Ltd., London © 1980 by the University of Chicago All rights reserved. Published 1980 Paperback edition 1981 Printed in the United States of America 00 99 98 97 96 95 94 93 92 91

5678910

Library of Congress Cataloging in Publication Data

Organski, A F K, 1923The war ledger. Includes bibliographical references and index. 1. War. 2. International relations. I. Kugler, Jacek, joint author. II. Title. U21.2.07 355.02 79-23366 ISBN 0-226-63279-2 (cloth) 0-226-63280-6 (paper)

For Christian, Elizabeth and Eric

Contents

Acknowledgments ix Introduction 1 Of Power 4 Of Size and Growth 8 Of Nuclear Weapons 9 Plan of the Book 10 One

Causes, Beginnings and Predictions: The Power Transition 13 Three Models 14 Comparison of the Three Models 22 Preparation for the Testing of a Model 28 Empirical Tests of the Power-Distribution Models Conclusion 61

49

Two

Davids and Goliaths: Predicting the Outcomes of International Wars 64 Power Indicators: Existing Measures 66 The Missing Measure of Political Development 68 Construction of a Measure of Political Development 71 An Index of Governmental Extraction 74 A New Measure of National Capabilities 85 Tests, Hypotheses, and Findings 86 Conclusion 101

Three

The Costs of Major Wars: The Phoenix Factor 104 Theoretical Propositions 105 Indexing National Capabilities or Power Resources Estimating Consequences of War 108 Choice of Test Cases 119 Actors 123 Empirical Propositions 130 Findings 132 The Phoenix Factor 142 Conclusion 144

vii

107

viii

Contents

Four

Nuclear Arms Races and Deterrence 147 Deterrents and Deterrence 150 Testing Deterrence: Outcomes of Crises 158 Testing Mutual Deterrence: The Nuclear Arms Race 180 Conclusion 199

Five

Conclusion 203 A Note on Architecture 203 Major Wars: Beginnings 206 Predictions of War Outcomes 208 The Phoenix Factor 210 Deterrence and Arms Races 213 Beyond the Data 217 Appendix 1: Index of Political Development 227 Appendix 2: Postwar American Aid 234 Appendix 3: Analysis of Models 236 Notes 243 Bibliography 271 Index 283

Acknowledgments

ix

Almost ten years have passed from the time this book was conceived. This has permitted us, the authors, to understand better the problem we were exploring, and, we hope, to improve the manuscript as well. Research collaborations are never easy, but, when successful, they permit some distinct pleasures to the participants. And so in this instance. This book began as a common enterprise between a teacher and a student, and has permitted the authors to become true collaborators and fast friends. We are both thankful for the opportunity and would like to acknowledge publicly our debt to each other. But we also have a joint debt to many people who deserve our thanks for having helped us. We cannot list all of them. We must pick and choose. Scholars, by necessity of academic life, are mendicants. To do their work they must beg for money, data, and help. All who know the economics of research in the academic setting also know what researchers owe to patrons. We wish to acknowledge publicly our debt to the Earhart and Ford foundations and to DARPA. The officers of these institutions who, over the years, were directly involved in decisions to support pieces of this research-Drs. Stephen Andriole, Judith Daly, Kalman Silvert, Antony Sullivan, Robert Young, and Mr. Richard Ware-truly deserve our thanks. Their understanding of what we were about and their encouragement were as important to us as the financial support provided by their institutions. Those who have done cross-national empirical work with census and national account data understand full well our dependence on people who have special knowledge of the meaning of the numbers they have put together. Gaining access to them and to additional necessary unpublished sources was a major problem. We cannot list all who generously helped us with their data and with information on what

x

Acknowledgments

these data meant. There are too many of them. We should like, however, to acknowledge the help of Drs. Ronald Tammen and Arthur House, legislative assistants of Senators William Proxmire and Abraham Ribicoff, and to Dr. Richard Solomon, director of the social sciences at the Rand Corporation, who worked mightily to obtain such access for us. Without the efforts of these persons this study could not have been completed. The extensive data-sets we obtained that made possible the calculations necessary to test the hypotheses underpinning this book will be deposited in the ICPSR at the Center for Political Studies and the Crisis Management Center of DARPA for all to use. We should like to acknowledge special assistance from Drs. Raja Chelliah, formerly of the International Monetary Fund, Roy Bahl of Syracuse University, and Elliott Morss with the United States government, who helped us at a critical juncture in our efforts to measure political capacity. That work is still going on. And we also wish to thank Mr. Daniel Fox of the Statistical Bureau of the University of Michigan, an unerring guide through the uncertainties of mathematical statistics. A number of other people merit special thanks. Professors James Caporaso, Bruce Bueno de Mesquita, and Robert North read the entire manuscript and made very valuable comments. Their efforts improved this book. Dr. William Domke, Mr. Michael Hom, and Mr. Steven Rood labored unceasingly to put our data in order and helped with the analysis. Mr. Christopher Braider and Mrs. Barbara Skala edited the manuscript in its entirety and their advice and literary skills helped us to pare and improve our prose. Miss Deborah Eddy kept control of all the pieces of a vast project, typed, and did endless copy editing on the manuscript. Chapter 2 of this book is a revised version of an article by A. F. K. Organski and Jacek Kugler, "David and Goliaths: Predicting the Outcomes of International Wars," that was published in Comparative Political Studies 11, no. 2 (July 1978): 141-80, and is used by permission of Sage Publications, Inc. We also wish to thank the editors of American Political Science Review and International Security for permission to use material that originally appeared in those journals.

xi

Acknowledgments

Both authors, moreover, wish to acknowledge their indebtedness to Cheryl Kugler, who cheerfuliy put up with deprivations inevitable in the lives of those tied with bonds of family and affection to those who do research and write books. One person who had always helped could not this time. Katherine Fox Organ ski died on February 15, 1973, when the first findings of this research were just beginning to be put on paper. Throughout, we missed her knowledge of the matter at hand, her judgment, her critical bent of mind, and her unparalleled editorial skills. Yet her influence on this work remained very marked. Her high standards in research and writing were made our goals. We need hardly add that all who helped us most did not always agree with us. The responsibility for all we have written is ours alone. A. F. K. Organ ski Ann Arbor, 1980 J acek Kugler Boston, 1980

Introduction

The story we are about to tell is a tale of conflict among nations. The wars we will be concerned with are few in number, but they are the fiercest and most lethal ever fought. Our book is not a series of case studies, however. A clinician's exploration of a single war is a very different task from the one undertaken in this volume, and the deep probing of individual cases requires skills we do not possess. The deadliness of wars in which the combatants fight with all-out efforts was a major criterion for our selection of a tiny sample of conflicts, but not for the obvious reasons. The vast size, scope, and ferocity of the fighting are important because they help to create the conditions of high stress essential to test our notions about the causes and consequences of armed conflicts. In our examination of these wars, then, we will be primarily concerned less with the wars themselves than with four general kinds of questions that can be posed about them. First, why do major wars begin? What are the conditions that provoke the most powerful nations in the world to fight with one another? Second, why exactly does one side win and the other lose? The obvious explanations of clever generalship, the size and self-sacrifice of armies, the quality and quantity of the weapons used, or combinations of all these, did not seem to us entirely persuasive. Third, we were interested in the rules that govern behavior of the contestants after the actual fighting has ceased. Some countries obviously recover faster than others after a war. How do some of the obvious factors such as victory and defeat influence the recovery of the combatants? Is there a predictable pattern in the behavior of the winners and losers? Fourth, and finally, have the rules governing conflict behavior between nations been drastically altered since the

2

Introduction

advent of the nuclear era? Popular credence argues that they have been largely, if not entirely, changed. The explosive power and sheer horror of the new weapons make this conception seem plausible. But is it true? We had our doubts. There is a charmed circle of ideas about war that has been passed on from one generation of war students to the next, in spite of evidence. There is the belief, for example, that men who wish peace should prepare for war. The Roman motto si vis pacem para bellum is often quoted. The Romans, however, fought all the time. There is the belief that love, understanding, and turning the other cheek can avert war; but there are many cases within memory of unsuccessful appeasement. There is the belief that hunger, popUlation pressure, and the like lead to war; but the evidence is conclusive that the weak, the hungry, and the overpopulated nations are the meek of the world and have never attacked anybody. On war everyone plays expert, and it is a bore. But experts do not seem to know much more. It is appalling how little is really known or, at least, how little is known by those who have to make decisions affecting peace and war. Think a moment about the questions to be treated in this book-beginnings, outcomes, and consequences of war-and think about the performance of leaders in recent military conflicts. For example, the leaders of the major powers at the beginning of World War I did not realize that a war was coming or the nature of the war their nations were going to have to fight. The comment made by one German general on the behavior of British soldiers, "they fight like lions but they are led by asses," should not, in justice, be restricted to the British alone. Did French, Italian, or heaven help us, Russian leaders perform any better in World Wars I or II? Stalin, even after being told by both Roosevelt and Churchill that the USSR was about to be invaded, refused to believe that Hitler would violate the 1939 pact and was immensely surprised when he did. And what about the problem of the factors determining victory or defeat? Think, for example, of the Japanese attack on Pearl Harbor. How incredible it is today that most of

3

Introduction

the Japanese leaders believed they could win a war against the United States. Even those who worried-their most "far-sighted" leaders (who were in the minority)calculated that Japan could win, although only if the war were of short duration. But what did these leaders expect? That after they had destroyed a portion of the American fleet and captured the Philippines from the United States, but without even denting American potential strength, the United States would simply quit and tum the other cheek? And again, how incredible that the Germans, and much of the world along with them, could seriously think that World War II would end differently than had World War I, despite the fact that Germany had fought the same countries just twenty years earlier and been badly defeated for her pains. Nor are American leaders immune to deadly blunders. Did our own "best and brightest" know how the war in Vietnam had to end? Evidently not, even after the stalemate the United States was forced to accept in Korea, fighting underdeveloped China. Should not that signal have been picked up? Finally, did anyone at the end of World War II guess what the ultimate consequences of that war would be? Did anyone think at the time that Germany and Japan would recover as rapidly as they have after the devastation they suffered at Allied hands? Would anyone in 1945 have believed that Germany would pass France in the production of steel by 1950? Yet the picture of leaders at their job is not as dismal as we have just implied. Some leaders seem to have shown better sense in regard to war than others. French generals at Versailles in 1919 feared that, after her defeat in World War I, Germany would attack France again within twenty years. It is amazing how right they were. Italian generals feared that Italy would lose if she entered World War II. They, too, were right. Franco was not impressed by Germany's defeat of France, guessed the strength of England, and refused to let Hitler strong-arm him into joining the Axis powers as a belligerent. Hitler was furious at Franco's pusillanimity and ingratitude, but it was Franco who died in his bed. What

4

Introduction

made these leaders guess right? What exactly did they sense about the outcome of the war that made them properly wary? There is much that cannot be known about war because the information is lost. Our ledger has no entry for the human pain caused by war. One can measure the physical havoc wrought: the cities destroyed, the houses demolished, the forests burned, the roads torn up. Such damage can be calculated with estimates filling gaps in the real data. But the suffering and the diminution in the quality of life cannot be measured: the data base is gone. For lives disorganized, for the displaced and the dispirited, for the suffering of persons wrenched from roots and routine, for the countless number of those who tried to "fit in" again after the fighting stopped, there is no adequate information. How many times in interviews with survivors has one heard, "I was doing this or that, but then the war came." or, "What can I do now?" No account of war can be considered complete which does not include such things, but no such data can possibly be made available, and we can only note their omission from our ledger. But though there is much that cannot be known, yet there is also much that could and should be known, and is not. It is to this area that we will devote ourselves in the pages that follow. The reader should be aware that the kind of things about which no positive knowledge can be obtained are only remotely relevant to the questions we seek to answer here. For these, quantifiable data may be gathered. Of Power

A close and complex connection exists between war and power. Shifts in the international distribution of power are often believed to create the conditions likely to lead to at least the most important wars, and power is the most important determinant of whether a war will be won or lost. And power, again, is the resource that leaders hope to preserve or to increase by resorting to armed conflict. Inevitably, this book will pay a great deal of attention to questions of power: the differences between strength and

5

Introduction

military might, the limits of power, the resources of power, the problem of measuring power. It is from the answers to all these questions that we have derived the substance of this book. Nevertheless, we will be chiefly concerned only with the last two, for reasons that will become apparent after a brief consideration of the differences between national power, national capability, and military might. These three concepts, despite the close relations by virtue of which they tend to be confused, are not identical. National power is the broadest of the three and can be defined simply as the ability of one nation to control the behavior of another for its own ends.! Such control may result from quite intangible factors: for example, from the fact that the arguments advanced by one nation in defense of its position may be thought persuasive by other nations; or from the circumstance that some nations are simply willing to follow one nation's lead because its requests are also in their own interests--or because they simply don't care, or because that nation enjoys great esteem and is viewed as a model by the other nations that deal with it. For a long honeymoon period after World War II the United States wielded a great deal of power simply because it was held in the highest esteem by a majority of the nations outside the Russian sphere. To take another example, Russia has been able to exercise a great deal of power for some fifty years now because millions of non-Russians have espoused the Communist ideology. Of course, fear can be persuasive as well. It is another intangible but extremely effective resource for international control. The United States, the USSR, China, and Japan, to name but a few nations, have all exercised a considerable degree of power over their neighbors for various periods of their history simply because the smaller states were frightened, and with good reason, of what the giants might do to them if they didn't comply with their requests. An attractive aspect of exercising power through persuasion is the cheapness of the method. But persuasion is not always effective, and it is certainly not the only means available. One nation can influence the behavior of another by directly rewarding or punishing it for what it does. This

6

Introduction

is the most frequent and the most stable way of exercising power. Each nation has needs that only other nations can supply. Wherever a nation controls something another nation wishes, and can give or withhold its prize at will, it commands the means of influencing the behavior of that nation. There is an endless variety of ways by which one nation can reward another: by providing it with goods for consumption and for defense, with funds and technology for the building up of its economy, with markets for its products. The rich industrial nations employ such means continuously to have other nations do their bidding. England was the most powerful country in the nineteenth century not only because of the might of the British Navy but also and, indeed, primarily because, as the first industrial power, she generated a large fraction of all that went into the international trade of the time and because her home market consumed much of what the rest of the world produced. Similarly, the United States today looms large in international politics because her economy produces so much of what the rest of the world wants, and her large and prosperous population represents a huge market for other peoples' products. Even the poorer nations may have rewards to give. Because they do not produce what other people want, they often give as rewards the riches of their subsoil, or pieces of their territory on which richer and more powerful nations can establish bases for their ships, their planes, their missiles, their soldiers, and their businesses. The capacity to reward implies, of course, capacity to punish. This means of exercising power requires no lengthy explanation. The record of international politics is replete with instances where the haves have withheld sales or aid, or threatened to do so, as signals to the have-nots to mend their ways. And recipients of foreign help, displeased with their benefactors, have, ironically, on occasion turned the tables on their patrons and changed suppliers. In the late seventies, in Africa alone, the Egyptians and the Somalis, dissatisfied with Russian support, or incensed with Russian interference, kicked the Russians out and sought help from the United States, while Ethiopia went the other way, switching from the American to the Russian side. Some-

7

Introduction

times punishments backfire. Our chastisement of Argentina and Brazil on human rights, or the Russian attempt to punish the Chinese, did not compel these countries to alter their behavior but simply induced them to become even more independent or more hostile than they were. Most of the time, however, the withholding of goods or assistance that other nations really need does work. The last and most obvious way in which nations exercise control over the behavior of others consists in the use of military force. This is the face of power with which we shall be most concerned because war obviously constitutes the extreme use of force in the exercise of control. But force can also be employed as punishment. Force used as punishment differs at least theoretically from force deployed in war because, in the case of the former, the wielder of force hopes that the punished nation will change its behavior of its own accord. Since World War II, the Arab-Israeli, Cambodian-Vietnamese, EthiopianEritrean, Saudi-Yemani, American-Vietnamese conflicts, and the frequent acts of terror and counterterror witnessed everywhere in the world, all furnish examples of force being employed in order to persuade adversaries to change their behavior. Such acts, successful or not, have that as their purpose. In the case of all-out war, however, disagreement between the combatants is of such a nature and degree that the goal each sets itself is no longer just to induce the other party to change its mind and course of action but to crush the other's resistance and control its behavior regardless of its wishes. At the stage of punishment, an element of choice remains; the object of war is the elimination of the power to choose. Two additional general points should be made. The use of force to control behavior is the most demanding and infrequent of all the ways in which power is exercised in international relations. The second point is most important for much that is to follow in this book. National capability and military might are two distinct quantities. A nation's strength goes far beyond its military might. To confuse the two can be disastrous, as the colossal blunder committed by the Japanese in attacking the United States at Pearl Harbor

8

Introduction

amply testifies. To know a nation's strength, one must look at its capacity to generate the resources that represent the major source of any nation's might. 2 Three extremely large and complex factors are primarily involved: (1) the number of people in a nation who can work and fight, (2) the skills and productivity of the active population, and (3) the capacity of the governmental system to mobilize the human and material resources at its disposal and devote them to national goals. Of Size and Growth

The sources of strength and power just mentioned are not constants. They vary in slow, intricate, and, in the long run, largely predictable ways, with changes in one master variable reinforcing changes that have already taken place and creating opportunities for new changes in each of the other two. It is this interactive process that is at the core of what has come to be called national development or growth. And it is this process of development that determines the power available to a country, that is, increases the pools of national capabilities (or power resources) available to central elites in their dealings with other countries. Power and development go together, and as one changes so does the other. The order of the international system rests on such a connection. It is the thesis of this book that the manner and the speed of national growth and development change the pools of resources available to nations and that such changes create the conditions in which international conflicts occur. Such changes also determine the outcomes of wars and, further still, so rigorously shape the politico-economic future that nations driven to fight in order to preserve or alter an existing distribution of power do so to no avail. Indeed, nations electing to fight in order to accelerate trends favorable to them eventually come to enjoy the superiority they seek, even if they lose. Critical to any understanding of the way the system of international power works is the realization that the developmental process is not uniform across countries. There are major differences in the timing, the sequences of growth, and the speed with which changes take

9

Introduction

place. There is no single road of development most nations must follow, and the different combinations of forces that determine the different ways in which development occurs exert an important influence on the level of power available to any given nation, because alternate developmental patterns produce different kinds of resources for the elites to draw upon in their dealings with other nations. One or two examples will clarify what we mean. The high level of political capacity of the system of North Vietnam gave that country's elite the kind of control over human resources that made it possible to repulse the American assault, but the lack of economic development left the same elite in a weak position in their dealings with other countries once the war was over. Again, the wealth of the Saudis will make them privileged customers in any industrial country but will not enable them to defend themselves against attack. If, then, one wishes to make reasonable estimates of the strength of nations, and of the kinds of effort they can make when pressed (and this is precisely the information one requires if, as here, one studies international military conflicts), one must first understand the patterns of national growth that provide the pool of critical resources necessary for a war and the capacity to deploy them usefully. It is in the relative size and the patterns of growth of the members of a system that the rules governing behavior in international military conflicts are to be found. Of Nuclear Weapons

The last portion of this book will deal with one final set of questions. Are the principal mechanisms of international conflict the same now as they were before the advent of nuclear weapons? Does the connection between growth and power still hold in the nuclear age? Have not nuclear weapons reduced to secondary importance the need for a large population, increased productivity, and highly developed political capacity? In view of the immense strength of these new and terrible weapons and the swiftness of their means of delivery, are not strength and military might finally the same, as military thinkers for the last hundred years

10

Introduction

have said was the case? Have events caught up with Mao's assertion that power grows out of the barrel of a gun-or, to update it, a missile launcher-and made it true? If this has happened, the international pecking order is being radically reordered. If the strength of nations does in fact, solely, or at least primarily, depend on nuclear weapons, then small, unproductive, and weakly governed systems can be as strong as the largest, most productive, and well-run systems because nations with all these deficits have, or are said to have. nuclear weapons. The unkempt giant, India, has nuclear weapons, and she is both poor and badly governed; a favorite lilliputian of international politics like Israel is rumored to have nuclear weapons; an economically unproductive country like China has nuclear weapons. Have, then, nuclear weapons made such a difference in the way conflict is carried out in international politics? If not, what difference have they made? Plan of the Book

We have ordered our inquiry to obtain maximum information about two cardinal interests that underpin this volume: national growth and international war. Because the two are so closely connected, it is possible to use one as an instrument in the study of the other. In this book, we consider the evolution and nature of the influence exerted by size and growth on war. It is our hunch that, at the inception of the modern international system, war and the preparation for it had much greater effect on development than the other way around. But in modern times growth is really far more influential on war than war on growth, hence this is the way that war should be studied. But precisely because in an examination of the intersection between development and conflictual behavior, it is inevitable that. in a review of one set of data, much will be learned about the other, the problem we encountered in properly designing the work was that of asking questions about, and exploring data on, war in such a fashion as to lose the least possible information on growth. The reader will find that we have exploited every opportunity for doing so. In each chapter of this book, some of the

11

Introduction

critical questions regarding the growth of nations has been explicitly or at least implicitly discussed. The angle of vision adopted in the first chapter permits us to regard international politics, and specifically military conflict, as an externality created by differential rates of growth in the countries who are members of the system. The first chapter, then, asks the question: "What happens to distribution of power in the international system during the years preceding the occurrence of a war?" Because of the nature of the problem, this chapter deals only with major wars. The second chapter contains, inter alia, a major attempt to put together a procedure that would permit rigorous and systematic evaluation of that portion of political development that is political capacity. This attempt is lodged in a broader effort to formulate a more accurate overall measure of national capability. Chapter 2 raises a second fundamental question: "What accounts for victory or defeat in war?" It is our contention that politico-economic and demographic development is the source of power, and the measure of political capability presented in this chapter gives powerful support to this thesis. The third chapter can be seen as a search for an answer to a further critical question concerning development: How resistant are the patterns of national growth to external attempts to redirect them? We ask: "What are the ultimate consequences of the outcomes of war in terms of power?" The reader should find in the chapter some answer to the related question, "Is war an instrument of policy or merely behavior expressive of economic forces?" We grant that it cannot be entirely one or the other. One can still begin to answer the question for the majority of cases in regard to the central tendency of the phenomena. The last chapter carries implicit in it yet another important developmental theme: "Is national growth the source of the behaviors of countries in dealing with one another in peacetime?" Specifically, when nations arm, do they respond to international exigencies or are they really responding to the internal pressures generated by patterns of growth which provide ever larger pools of resources to be

12

Introduction

used for the build-up of arms? We attempt to evaluate the influence of nuclear weapons on international politics, and particularly, on military behavior. Does nuclear deterrence take place? And, are nuclear weapons constructed because of external danger or simply because the necessary resources are available? We shall tackle all of these questions in the following pages. Such questions can not be answered in their entirety or definitively at this time. But we hope to leave them at the end of this book forward of the point at which we found them when we began. Let us tum now to our first question: what are the conditions favoring the outbreak of major international war?

One

13

Causes, Beginnings and Predictions

The Power Transition

Despite the vast literature devoted to war, little is known on the subject that is of practical value. Theories about the origins of wars remain tentative. One need only recall the extraordinary exchange between German Chancellor von Bulow and his successor, in the early hours of World War I. "How did it happen?" asked von Bulow. "Ah, if we only knew," was the reply.\ John F. Kennedy, recounting this episode many times, expressed horror at it. He would know; he would do better. Yet, even as he voiced his dismay, he was taking those very steps that led to American intervention in Vietnam. The record of ignorance is depressing. Our interest lies in explaining major wars. Theories about why wars begin are in a highly embryonic state. 2 Attempts at solutions of the problem require the joining of information collected from two fundamental sources. In the first place, accurate observations are needed on the power possessed by all nations in the system. It has long been believed that the outbreak of major hostilities is connected to changes in the power structure of the international order. The core of this first argument is as follows: If one nation gains significantly in power, its improved position relative to that of other nations frightens them and induces them to try to reverse this gain by war. Or, vice versa, a nation gaining on an adversary will try to make its advantage permanent by reducing its opponent by force of arms. Either way, changes in power are considered causae belli. It is also clear that "structural" changes can explain only a portion (though a critical portion) of the problem of why wars occur. What if the leaders of the nations affected do not perceive a threat in what has taken place and therefore do not choose to fight? There can be little doubt that some of the incendiary factors essential to the outbreak of wars are

14

Chapter One

lodged in the culture of elites, their belief systems, their skill in negotiation, their ability to decipher signals from other leaders, as well as in the constraints and opportunities imposed on and provided for all elites by the institutions in which they must operate. Our second source of information on the beginnings of wars is, then, the process whereby elites elect either to go to war or to keep the peace. It is difficult to say which of these two types of information is more important. Intelligence agencies quarrel endlessly over the question of whether estimates of an adversary's capacities are more important than information on the intentions of its leaders. Clearly, neither estimates of a country's changes in power nor of the pugnacity of its elites can alone account for the entire process that leads one nation to war against another. But, taken together, perhaps they may enable us to answer two fundamental questions of international politics: What causes major powers to enter into major wars? Can one reliably predict the approach of wars? For a long time, three models have been deployed in an effort to relate specifically different distributions of power to the coming of war or the preservation of peace, and to tie such estimates to assumptions as to when and why elites choose to fight. These models will serve as points of departure for our own analysis. We shall first explore in what ways the distribution of power is linked with the beginning of war. Second, we shall examine the problem of how to index decisions by elites to fight a war. Third, we shall try to test the models to see which one is correct. Three Models

The Balance of Power One model, respectably ancient, has served specialists and practitioners of international relations for centuries. The balance-of-power model suggests that when power is more or less equally distributed among great powers or members of major alliances peace will ensue. Conversely, as large asymmetries become discernible in the distribution of power resources, the probability of war increases markedly. Ac-

15

The Power Transition

cording to this theory, the country whose power is increasing will take advantage of its superior strength to attack its now weaker adversaries. Hence, the trinity of beliefs that constitutes the balance-of-power model: equality of power is conducive to peace; an imbalance of power leads to war; the stronger party is the likely aggressor. 3 This is how the power system is supposed to work: ... Given large numbers of nations with varying amounts of power each one striving to maximize its own power, there is a tendency of the entire system to be in balance. That is to say the various nations group themselves together in such a way that no single nation or group of nations is strong enough to overwhelm the others, for its power is balanced by that of some opposing group. As long as the balance can be maintained, there is peace and the independence of small nations is maintained. 4 The quotation adumbrates the central mechanism on which the balance of power operates in determining either international stability or the outbreak of war. It describes the nature of the motivating force that impels the actors to arrange themselves in such a way that a balance of power may result, and it explains why at least a roughly equal distribution of power is necessary in order to keep the peace. Just as theoretical economists explain the behavior of "economic man" as motivated by a desire to maximize his profits, so specialists in international politics who accept the balance-of-power model postulate the political motives of nations as motivated by their desire to maximize their power. The motive purportedly inspiring all actors in the system to behave as they do also implies the fundamental rule governing all decisions in the field of foreign policy. Since nations that have an advantage will maximize their power positions by attacking the weak, these weaker nations, in turn, will gain strength by allying themselves with other countries in comparable positions. All nations, of course, can also increase their strength by breaking up the alliances of their opponents, or even by fighting in order to protect the distribution of power that, in the long run, will protect their well-being and their existence.

16

Chapter One

It should be clear that the major mechanism through which the balance-of-power system is maintained is the making and unmaking of alliances. The reason for this dependence on coalitions in order to change the distribution of power is that the power resources of each member of the system are viewed as inelastic. There is no way a nation can increase its own strength very much except by adding its allies' strength to its own, or by decreasing its adversaries' strength by separating it, through persuasion, bribery, or subversion, from its allies. Finally, the cited passage discloses why it is that the "balance" must represent an equal distribution of power. Proponents of the balance-of-power system often fudge on the question of what kind of power distribution is necessary to assure the security of all its members. Some writers have made the point that by a "balance" one means simply an equilibrium, and that such a state can result from almost any distribution of power. 5 But this, of course, contradicts the first commandment of the system, that all nations will try to maximize their power. Any gain of a decisive advantage by one nation or faction will represent the beginning of the pyramiding of power by the stronger until its hegemony is established or until, somehow, its power is matched by the equal power of its opponents. One feature of the balance-of-power system needs additional comment: the system is homeostatic. Indeed, it is ultrastable. In maximizing their own power positions, nations group themselves in the kind of balances that tend to keep the system stable, peaceful, and secure. If the equilibrium is disturbed, the system favors adjustments that will return it to equilibrium. But if this jockeying process in which members of the system engage cannot reallocate the power loads sufficiently to obtain a roughly equal distribution among the major actors in the system, then one nation, possessing decisive strength and uninvolved in existing coalitions, will step in on the weaker side and redress the balance, thus rendering the system ultrastable. In the past, this role of balancer was associated with Great Britain, which is held by many historians to have acted in this fashion in Europe and the world at large, at least during the eighteenth and nineteenth centuries.

17

The Power Transition

Those who espouse the balance-of-power model do not clearly explain why one nation should be exempt from the otherwise universal rule of wanting to take advantage of its superiority to expand its power at the expense of others. In view of the way all nations are supposed to behave, according to the rules of the model, it would not seem implausible to argue that a balancer, given its superior strength, would seek to maximize its power by attacking one coalition with the help of the other until it could reduce all other nations into submission. One cannot be quite comfortable with any other assumption. Do all nations really wish to maximize their power? One cannot help noticing variations, over time, in the degree to which they have wished to do so. Sweden is a model of a peace-loving nation today, but it was once a feared aggressor. The United States was a peaceful, indeed an isolationist nation, in the past, but it certainly defies those definitions today. We lack, in short, the kind of universal behavior that would have to prevail for the first law of the balance-ofpower system to be, as it is held to be, immutable. There are further uncertainties on this and other points. But to us the most important questions are those related to the validity of the model, and what must interest us most is whether or not the equal distribution of power does in fact, keep the peace; or, conversely, whether or not an unequal distribution of power produces war. The other two models argue against both of these possibilities. Collective Security A second model, based on the notion of collective security, gained both its name and its renown from its role in the formulation ofthe peace after World War I. Collective security presented a different set of power requirements for the maintenance of international peace. The distribution of power resources between opposing factions had to be extremely lopsided; collective security required that all members of the system move against the aggressor. "All against one" was the order of the day. If a peaceful nation failed to do its duty because of un interest in the immediate quarrel or in the fate of the victim, or if an aggressor were able to win over potential defenders of the victim by playing on their

18

Chapter One

fears or on their greed for booty, the chances of war would grow with each such defection. Even a rough equality of resources among the members of the coalition defending the victim would fail to prevent war. But if the prescription of collective security, "all against one," were obeyed and war still came, the defeat of the aggressor would be inevitable. Collective security would provide security, if not peace. Thus, the fundamental tenets of this model are as follows. A lopsided distribution of power (with defenders much stronger than the aggressor) will support peace; an equal or approximately equal distribution of power will mean war, but the aggressor will be weaker than the coalition. 6 This second model makes three additional assumptions. First, when a serious international dispute threatens an outbreak of hostilities, the identity of the aggressor will be clear to all. This seems very uncertain, however. There are many illustrations of one country's claiming another to be the aggressor, with every such accusation being widely credited. The wars in Vietnam, the conflict between India and Pakistan, and the recurrent struggles in the Middle East are cases in point of the difficulty of definitively identifying the aggressor. A second assumption fundamental to the collectivesecurity model is that all nations will be equally interested in preventing aggression and thus can be expected to regulate their political and military behavior to that end. While peace is the prized but largely unintended consequence of the ways in which the balance-of-power system works, in the collective-security system, peace is the direct and explicit aim of all its members (aside from the aggressor). 7 A third assumption is that alliances are the major method by which the necessary imbalance of power between aggressive and peaceful nations is to be effected. In this, the collective-security and the balance-of-power models are as one. However, only in the former is the commitment to resist aggression made a priori: the necessary coalition is to follow automatically once the need arises. The validity of some of these assumptions will be further discussed below. At this point, we should present a third model, developed since World War II, which has some ob-

19

The Power Transition

vious differences from, and some important similarities with, the propositions of the two models already discussed. The Power Transition Our third model, evolved from the conception of the power transition, was formulated in the fifties. 8 Some of its conclusions are much the same as those we have just described: an even distribution of political, economic, and military capabilities between contending groups of nations is likely to increase the probability of war; peace is preserved best when there is an imbalance of national capabilities between disadvantaged and advantaged nations; the aggressor will come from a small group of dissatisfied strong countries; and it is the weaker, rather than the stronger, power that is most likely to be the aggressor. The following passage summarizes the major mechanics of this model: At the very apex of the pyramid is the most powerful nation in the world, currently the United States, previously England, perhaps tomorrow Russia or China.... Just below the apex of the pyramid are the great powers. The difference between them and the dominant nation is to be found not only in their different abilities to influence the behavior of others, but also in the differential benefits they receive from the international order to which they belong. Great powers are, as their name indicates, very powerful nations, but they are less powerful than the dominant nation .... As we have seen ... the powerful and dissatisfied nations are usually those that have grown to full power after the existing international order was fully established and the benefits already allocated. These parvenus had no share in the creation of the international order, and the dominant nation and its supporters are not usually willing to grant the newcomers more than a small part of the advantages they receive .... The challengers, for their part, are seeking to establish a new place for themselves in international society, a place to which they feel their increasing power entitles them. Often these nations have grown rapidly in power and expect to continue to grow. They have reason to believe that they can rival or sur-

20

Chapter One

pass in power the dominant nation, and they are unwilling to accept a subordinate position in international affairs when dominance would give them much greater benefits and privileges. 9 This model insists that the significant differences in the distribution of international power are rooted in the different capacities of member states to utilize their own human and material resources. The model argues that the source of war is to be found in the differences in size and rates of growth of the members of the international system. If one introduces controls for size of nation-states-and differences here are truly spectacular-the rest of the differences in power can be accounted for by differences in levels of development of key sectors of national life. Most important are economic productivity and the efficiency of the political system in extracting and aggregating human and material resources into pools available for national purposes. 10 The claim that national power stems from national development has important implications for the way this model works. It is often stressed that the development revolution is worldwide. And so it is. But the changes that make up development are not spread evenly across all countries and all regions of the world. Even today, only one-third of the earth's nations are developed and at the stage of power maturity. Roughly one-third are still developing and are at some lower point of the power transition. The remainder have barely begun the long trek toward wealth and power. What is particularly important for international peace and security is the fact that the big nations-those with the largest populations and, consequently, with the largest potential for power-are spread out along the development continuum. Some, like India or Indonesia, are extremely low on the economic scale and still in the stage of potential power; others, like China and Brazil, are at different points in the transitional stage; finally, such countries as the United States, Japan, Germany, the United Kingdom, and the Soviet Union are in various advanced degrees of power maturity. It is obvious that capacity to disturb the equilibrium ofthe system is largely dependent on the base from which the

21

The Power Transition

country begins. The full development of Guatemala, Costa Rica, or Albania will pass unnoticed, for these countries are small; but if India or Indonesia begins to modernize in earnest, the effects of such events will inevitably shake up the international power distributions. This is what happened when China began to evidence major gains in the accumulation of power resources. The power-transition model postulates that the speed with which modernization occurs in big countries is also quite important in disturbing the equilibrium that existed theretofore. For if development is slow, the problems arising from one nation's catching up with the dominant one may have a greater chance of being resolved. On the other hand, if growth takes place rapidly, both parties will be unprepared for the resulting shift. The challenger may not have had the opportunity to develop a realistic evaluation of its position because its elites will be strangers to power, and the sources of new-found strength are almost entirely the result of internal changes. It seems plausible to think that the chances for miscalculation consequently increase. The developmental sequence should also be considered. It makes a difference to a nation's power whether development begins with a sharp rise in economic productivity, or with rapid political mobilization, or with dramatic increases in social and geographic mobility. Changes in these different sectors yield different power resources. Two examples will help to make this point clear. In most cases in the Western European experience, the powerful propellant of overall national development was economic change, and it was the steep rise in economic productivity in Western Europe in the nineteenth century that afforded the Europeans advantages in trade, weapons, and large armies and navies that enabled them to subjugate a backward world. However, in the case of some of the Communist countries, the motivation for national development has proved to be political mobilization and organization. The emergence of high levels of political mobilization, through the creation of political networks penetrating deeply into the mass of the population, has been largely independent of the socioeconomic changes that reinforced and propelled political changes in

22

Chapter One

the Western nations. The persuasiveness of Chinese ideology to peoples beyond China's frontiers and the astonishing effectiveness of Chinese, North Korean, Vietnamese, and Khmer Rouge armies against vastly superior forces are rooted in the success of political organization in those countries in mobilizing major fractions of their populations and creating the necessary structure to sustain a successful military effort. Political "development" in advance of economic development is a critical problem in the construction of measures capable of predicting the outcomes of wars and is given a complete airing in the next chapter. The population of a nation, the speed, timing, and sequence of its political and socioeconomic development, have important consequences for the power of a country at any given time; all factors are critical to the operation of the power-transition model. Comparison of the Three Models

The models differ in fundamental ways but also share a number of important features. To locate what we wish to know, we should ask three questions of each model. First, what rules the decisions of the actors in the system to keep the peace and to keep their places? Second, how are the power distributions essential for peace, and how do they bring about war? Third, and to us the most important, what are the power distributions that each model associates with peace or with the outbreak of major conflicts? The Goals of Elites It seems clear that the motives of decision-makers in maneuvering their nations away from or towards conflicts differ fundamentally in every model. In the balance-of-power model, the leaders of a nation seek to maximize its power. Strong nations try to expand, while their potential victims, seeking to protect themselves from aggression, band together to augment their offensive and defensive capabilities. Decision-makers in the collective-security system are moved by a rational desire to prevent (or to defeat) aggression. The power-transition model differs from the other two in basic ways. It provides no general rule to explain and pre-

23

The Power Transition

dict the circumstances in which elites will move toward war. On the other hand, it warns that changes in the power structure will not, in and of themselves, bring war about. Satisfied great powers are not likely to interpret advantages gained by satisfied lesser powers as threatening. Moreover, the powerful and satisfied do not start wars. Only if the great powers think that the changing system challenges their positions, or if they no longer like the way benefits are divided, should the shifts be deemed dangerous. All three models, then, ascribe predictable behavior to nations. The first two models differ in that decision-makers are moved by a desire to maximize the utility of the system in the collective-security model, while in the balance-ofpower model the motivations of the actors are to maximize individual utility, with the benefits accruing to the members of the system as a whole being simply a consequence of the selfish behavior of its members as individuals. The reader acquainted with the school of laissez-faire economics will, of course, find all this material most familiar. One should also note that both models assume that the currents leading to war and peace and to the preservation of the system are manipulable, that they can and must be managed, and that foreign-policy elites are key actors in the play. These are models suitable for the action-oriented. According to the power-transition model, on the other hand, it is not a desire to maximize power or a single-minded urge to guarantee security in the narrow sense that leads nations to start major wars, though the latter is often the excuse furnished. In this model, it is a general dissatisfaction with its position in the system, and a desire to redraft the rules by which relations among nations work, that move a country to begin a major war. We should note one point. The power-transition model does not require that the dissatisfaction felt by the challenger be judged valid by an "objective" observer. Dissatisfied powers mayor may not have "good reason" for feeling aggrieved. While their positions may be somewhat disadvantageous when compared with those of a few other nations, their advantages over the rest of the world are substantial. Indeed, according to the model, the truly disadvantaged nations are by definition too weak to disturb the

24

Chapter One

peace. Valid or not, however, the choice of methods of the significant actors in the power-transition model remains predictable, and in this sense at least this model is at one with the other two. A final point of difference needs to be cited. The trajectories that lead nations to collide with one another are not easily manipulable. Some fine-tuning of their movement is often tried, with very uncertain results. But the fundamental evolution of power distribution is set and cannot be manipulated. (The evidence of the consequences of war in Chapter 3 supports this view.) The power-transition model, therefore, may be of little comfort to activists interested in international engineering to preserve the peace. The Mechanisms That Redistribute Power What causes the pernicious distributions of power that lead to war? The models account for such power changes differently. The balance-of-power and collective-security models argue that changes are the results of alliances. The units of the system do not change (at least not much); they simply combine in different ways, and different distributions are the result of such combinations. The rule is simple: a nation can influence the balance of power in its own favor by allying itself with other nations and by adding to its own capabilities those of its allies. Other means are available if a nation wishes to improve its power position. It can also arm and even fight for this purpose, or redress its weaknesses. But the least costly and most certain way for a nation to improve its power position is to combine its strength with that of friends or to break the coalitions of adversaries. There is, it should be noted, substantial support for this view. 11 The model based on the concept of the power transition is at odds with such conclusions because it assumes that the major source of power for a nation is its own socioeconomic and political development. How else can one explain the rise of the Soviet Union and the United States, or the decline of the United Kingdom and France? These major changes in the international distribution of power occurred

25

The Power Transition

outside the normal pattern of alliances and have affected the stability and viability of the system far more than the alignments and realignments of coalitions. Most of the time alliances are simply not a realistic method of preventing threatening changes in the distribution of world power, given the skewness of relations between the great and the lesser nations, and also among the half-dozen great powers themselves. In times of peace, the dominant nation is substantially stronger than the remaining great powers. Consider the present worldwide distribution of power among the international giants from the angle of vision imposed by the power-transition model. The difference in power between the present leading nation, the United States, and its nearest challenger, the Soviet Union, is still very large, although the gap between them has been closing. And it is, precisely, the relationship between the challenger and the dominant country that, in the transition model, is likely to occasion a major war. Equally large is the interval separating the Soviet Union from Germany and Japan on the next tier of great powers, and finally, the gap separating those nations from France and the United Kingdom. It is clear that, if the intervals separating the nations in question are as large as we suggest, more probable alliances could affect only the size of the intervals between the strata, but could not alter the fundamental ranking of the great powers dominating the international system. Moreover, alliances cannot easily be made or unmade. For the six or seven nations that represent the major powers of the international system, there are a large number of possible but wildly implausible combinations. The plausible ones are very few; most are not plausible because, precisely, the socioeconomic and politico-ideological ties that bind nations together in the modem era resist yielding solely to considerations of power advantage. Witness how difficult it has been for the United States to consent even to diplomatic civilities with the People's Republic of China, or how difficult it was for Germany to change the nature of its relations with France, even though a French connection would have significantly improved German chances to outstrip Great Britain in the decades immediately preceding World Var I.

26

Chapter One

The assumption that underlies much of the balance-ofpower model, namely, that the dictates of power considerations are sufficiently strong to guide the behavior of countries in making and breaking alliances, is not true. However, countries do change sides. For example, Italy and Japan moved from the side of the Allies to that of Germany between the two world wars and, together with Germany, changed sides once more after they were defeated in World War II. We can only guess at the complexities that playa role in such shifts. One can perhaps advance the notion that changes in alliances are connected with changes in the combinations of elites that have access to power in countries that change sides. Such changes are precipitated either by the socioeconomic and political shifts which occur as a result of the developmental process that all nations undergo once they begin to modernize, or are forced upon countries by defeats in major wars. The former cause was probably at work in Italy's change between the two wars and the latter in the passage of the Axis powers to the side of the Allies after World War II. There are, however, some situations covered by the power-transition model in which alliances could make a difference in the power distribution of a system. But, as stated, such occasions are few, even though they could be of long duration. Obviously, if the intervals separating the great powers are very large and the units themselves are unchanging, there is virtually no possibility of the power rankings being altered. But according to the model of power transition, the units in the system are not immutable. National growth will cause a small number of major countries to overtake rivals who were far ahead of them at earlier points in time. The period of passage (which may require several decades) witnesses an equal distribution of power among the major contenders, with accompanying perils to peace. It is also a period when different alliances among major nations could effect a change in the distribution of international power. Two examples will help make the point. There was a period in the last third of the nineteenth century when Germany had overtaken France and was catching up on Great

27

The Power Transition

Britain; an alliance of the French and Germans during this period could have gained dominance for the latter, at least briefly. One may speculate about the future, in terms of the power-transition model, and foresee a relatively long period in which China will have surpassed the Soviet Union but not the United States, and when an alliance between the Russians and the Chinese could spell an earlier demise of American dominance in the international system than is currently anticipated. A solid alliance between the United States and the Soviet Union, on the other hand, would delay the moment at which China becomes the dominant power. These examples make it clear that, insofar as distributions of power are concerned, whether or not alliances should be considered as an important mechanism in the short-term redistribution of power in the system depends on the observer's interest in those periods when alliance behavior can affect such distributions. Alliances, however, cannot in the long run alter secular trends. Power Distributions The final and most critical difference one finds in comparing the three models appears when one asks what kind of power distribution is to be associated with the preservation of peace and the outbreak of hostilities. Here, the models divide differently than they do on the questions we have already discussed. To this question, the power-transition and collective-security models respond that, to preserve peace and security, the power distribution must be lopsided in favor of the defenders of the system and against the nations that wish to attack it. This is the very opposite of the prescriptions contained in the balance-of-power model. Its recipe for peace is an equal distribution of power between the major contesting sides, because the danger of war increases dramatically when one side begins to gain a substantial advantage over the other. Moreover, balance-of-power predicts that the stronger will attack, collective-security posits that the aggressor will be weaker than the coalition, while power-transition argues that the attacker will be the weaker party. One final point. In the case of the power-transition model,

28

Chapter One

there is a period during which both dominant and challenging nations are roughly equal in power. The challenger has finally caught up with the dominant country, passage is a reality, and the elites on both sides view the shifts in power as threatening. The model insists that it is an attempt to hasten this passage that leads the faster-growing nation to attack. At the same time it is a desperate attempt on the part of the still-dominant nation to intercept the challenger's progress that leads to war. Moreover, the passage may not be quick-it may take several decades-and the period may thus be punctuated by a number of armed conflicts. In addition, the model insists that attempts to arrest the gains ofthe faster-growing nation will fail. Whatever the fortunes of war, the challenger will probably "win" sooner or later. Preparation for the Testing of a Model It is clear from our discussion that in some ways these models are complementary and that in others they contradict one another. They can all be wrong, but they cannot all be right. Each of the models seems plausible enough, providing interesting explanations, for different circumstances and periods, of the way peace is maintained or war breaks out; but there is, so far, no way to tell which of the models is correct. And each of our explanations has its partisans. Which of the models describes accurately (or, at least, more accurately than the others) how the international political system works? The models are interesting one and all. But are they valid? We can never know, of course, unless we find some way of testing them, and this is precisely what we shall attempt to do hereY We should be careful not to claim too much. The models cannot be tested in their entirety. They are far too complex and contain too many implicit and explicit propositions to be exhaustively examined here. Such a test would be a monumental work; indeed, it may not even be possible at this stage. What we wish to test is only one of the propositions contained in the models. The keystone of each of them is the distribution of power that it argues is associated with war and peace. In other words, what happens to the power dis-

29

The Power Transition

tribution, at least among the great-power systems, when wars occur? It is this question that we shall try to answer. Even such a test requires extensive preparation. The question "Which distribution of power leads to war?" inevitably involves a certain amount of conceptual looseness and ambiguity in its theoretical formulation that is no longer tolerable when one turns to the business of an empirical test. Things need to be clarified. How are we to measure national power or national capabilities? How are we to index the conception of power equality and inequality? Which wars are to be considered major? Our first chore, therefore, if our tests are to be made possible, is to develop readily usable measures of national capabilities. This means, inter alia, that only measures for which data are available across time and major countries can be considered. A second task is to define explictly the manner in which the changes in the distribution are to be indexed. If we argue that a certain type of relationship between the power possessed by each of two countries will lock the countries into a course that eventually leads to war, how is that specific configuration to be rendered explicit? Our third task is to make clear which powers represent the actors in a given system. The matter, as we shall see, goes beyond simply defining which power is a "great" one. In any event, the identification of the actors constitutes a critical preparatory step before our experiment can be executed. Fourth, we have to identify which wars the measured changes in power are supposed to explain. Clearly, the models do not pretend to explain all wars, although in the rarified atmosphere in which such ideas are discussed, it often appears that way. What they aim at explaining are major wars. It follows, then, that the problem is one of defining what constitutes a major war. One final task remains. Implicitly or explicitly, all three models suggest that changes in the power distribution are not coded in the same way by different elites guiding major powers in their international dealings. A friend's power gains are not disturbing, but the newly won power of an

30

Chapter One

adversary may be seen by many as serious business. In some cases, it may be seen as casus belli. Thus we must have an acceptable measure for showing whether the elites of our national actors have interpreted the changes in the power structure preceding wars as not threatening and, therefore, as requiring no action; or whether, conversely, they have deemed them threatening and been spurred by them to gear their nations for war. The Measurement of Power Resources Power has long been considered to be the capacity of an individual, group, or nation to control the behavior of others in accordance with its own ends. 13 It is an element of every relationship, with each party in possession of resources, tangible and intangible, likely to alter the conduct of the other. Power becomes apparent only when a disagreement arises between the parties, in which case the desire of the more powerful will prevail. The measurement of power, therefore, is vital to the prediction and explanation of joint behavior. There have been attempts to study the actual degree of control which one nation has exercised over another. For the most part, however, specialists in international politics have retreated to a fallback position and have contented themselves with measuring the resources that generate power.14 This procedure presents problems. Power resources may not necessarily reflect the exercise of their potentialities, nor do reliable estimates of them include essential components (e.g., diplomatic skills, charismatic leadership, and internationally appealing belief systems) that are not susceptible to easy or dependable measurement. Moreover, hard estimates do not indicate the power a particular nation may ostensibly possess simply because other nations may mistakenly assume that it is more or less powerful than is actually the case. The semblance of power often passes for its reality. 15 Because our concern is with the connection between power and war, the last problem may not loom so large. A discrepancy between the possession of real power and the external perception of it is more likely to occur in times of

31

The Power Transition

peace than in wartime. During and after war, however, the two views tend to merge, for perceptions are then put to the test. Thus, Mussolini's threats and bluster won for Italy a degree of international deference which proved greatly out of proportion with that country's performance when its actual power was demonstrated in World War II, the period during which perceptions of Italian power came to collide with its reality. Similarly, the perception and the reality of Japanese military power were far apart until this country's gradual commitment in World War II brought the two together again. Procedures for the measurement of national capabilities comprise three steps: (1) listing all the factors that may serve as indicators of what influences the exercise of national power; (2) selecting the number of such indicators considered important; (3) determining a way of aggregating the components thus identified in such a way as to obtain a single measure of national capability. The first two steps have been performed repeatedly. A list of indicators often thought essential factors were composed and reduced to manageable proportions, so that measurable elements could be intelligently combined. 16 Among scholars interested in the construction of such empirical estimates of national power there has long been agreement that measures of economic, technological, political, military, and demographic capabilities suffice to furnish a reasonably accurate overall indication. 17 Quantitative indicators are available. The economic capacity of a nation, for example, may be reliably suggested by data disclosing per capita, total, or disposable output. Demographic capabilities are grossly reflected in any calculation of total population or, more accurately, by the fraction in working and/or fighting age-groups. Military preparedness may be inferred from ascertainable expenditures on arms and the size of military forces. Only political capabilities are difficult to measure. The necessary data have not been available until recently, and are not easy to interpret. This problem of satisfactorily measuring political capability has been a major defect in the construction of measures of power and has important implications for some of the

32

Chapter One

questions we raise in this chapter and in our study of the consequences and outcomes of war. Our initial steps toward a possible solution of this problem have enabled us to have a try at forecasting who will win and who will lose when nations fight. We deal fully with the problem in our second chapter. All we intend to do here is to present the measures (which exclude direct measurement of political capabilities) that were used to test whether the power distributions and the changes in them, which the models claim accompany the preservation of peace or the outbreak of war, actually obtained before conflicts occurred. The lack of a measure of political development, however, is not debilitating for the kind of test we are doing in this chapter and in Chapter 3. Direct measures of political capacity become essential only in the estimation of the national capabilities of deVeloping countries. The reason for this should be made plain. Countries that were industrialized before World War II followed a pattern in which entrance of the mass of the population into the political system was a response to socioeconomic change. Thus, the expansion of the political system lurched forward roughly in step with the expansion of economic productivity and urbanization. Therefore, it is possible in those cases to deduce the level of political development from the measurement of key socioeconomic variables. As we have pointed out, however, this is not possible where this pattern of development has been violated, as indeed it has been by countries that are developing today. In these cases, such estimates, as measures of national capabilities, are seriously defective and should not be used. We shall have a good deal more to say about this problem when we deal with predicting the outcomes of conflict. Fortunately for our purposes here the major powers in the international system whose development in the political sphere coincided with development in the economic one have been the Western nations. The selection of critical indicators is only a first step. Also required is a method of combining these indicators to form a single measure. IS Until recently, no such aggregation has been attempted. Frequently, values of all the indicators were presented, and the reader was left to bring them into

33

The Power Transition

some sort of focus. One could derive an intuitive or impressionistic measure if a nation scored equally well on all the elements measured. However, if a country scored well in some areas and poorly in others, impressionistic estimates became fanciful. It is essential, if one is to judge the effects of international power distribution on major conflicts, to establish a single, reliable measurement. In the main, three types of aggregation have been suggested. The first one simply adds the values of the indicators together. A second suggestion is to multiply together, rather than add, the elements of the equation. A third suggestion goes beyond the problem of the form of the aggregation and points out that not all the elements of the power equation are of equal value and that, therefore, components should be weighted. While many measures have been proposed, only a handful have been developed to the point of genuine utility. We have chosen to compare the two whose theoretical and empirical development is most advanced and at the same time represent the conceptual extremes in the debate over the best method of aggregation. Using total output as a measure of national capabilities. A. F. K. Organ ski and Kingsley Davis argued early that gross national product and/or national income could serve as good yardsticks of national capabilities. The utility of measures of total output for the estimation of national capabilities should not be surprising. Estimates of gross national product closely reflect the movement of the underlying variables crucial to the generation of national resources-the fraction of the population of working and fighting ages, and the level of productivity. Measurements of productivity are particularly informative, for the contributions of individuals to the gross national product accurately parallel the levels of available technology, education, capital intensity, and many other attributes crucial to the establishment and maintenance of national power. Moreover, high levels of productivity also denote the capacity of a society to pay for external security, because military expenditures are approximately related to levels of national wealth.

34

Chapter One

Because total output is at the core (the result of the interaction between the size of the productive population and its level of productivity), the national power equation can be expressed in the following fashion. Power

=

· PopuIatlon x PGNP l. opu ation

=

GNP

In this formulation, total population implies the size of the fraction of members of working age, and per capita product implies the productivity level. 19 The interaction of components presumes an implicit weighting system. Productivity and population are proportionately related. One population twice as productive but half as large as another implies that two individual workers in the less productive economy are required to perform the labor of one in the more productive, but the power contribution of both is the same. This weighting system, while arbitrary, seems theoretically justifiable. More important, it reflects the realities of international politics. 20 As we have noted, the major defect of GNP as an overall yardstick of national capabilities is that it does not measure directly the capability of the political system to do its job, but the problem is not acute in the case of developed countries which became industrialized in the nineteenth century. In their case it is possible to deduce the level of political development from the measures of key socioeconomic variables. Let us now tum to the second measure of national power resources that could be used in our evaluations of power distributions. The J. D. Singer-S. Bremer-J. Stuckey (SBS) measure of national capabilities. This measure is important both because the procedure devised by the researchers is interesting and also because they collected the data required in a way that makes their indices usable and not merely admirable. 21 It is important to describe in some detail how this measure

35

The Power Transition

of power was devised by the Singer group. The core of the procedure was as follows. a. The authors argued that three major variables are sufficient to give an indication of overall national capabilities: military, industrial, and demographic capacities. Other variables are considered much less important, or are so closely related to the major variables that they are well represented by them and the indicators chosen to denote them. h. The indicators chosen to measure each of the three factors are: industrial capacity (represented by figures for energy consumption), military capacity (measured by expenditures and the number of men under arms), the demographic component (expressed by total popUlation and the number of inhabitants in cities of twenty thousand and more). c. Having selected the countries judged to be critical, the authors proceeded to gather data for each. With these in hand, they added up the values of each indicator for all the countries in the system. This total was considered to be 100 percent of the values for that capability for all the countries in the system; each country was apportioned its appropriate percentage share. d. The percentage share for each indicator for each nation was added up across all indicators, and the result was divided by the number of indicators (six). The percentage result was taken as the share for the country of all the national capabilities available to it in the international system as a whole. Table 1.1 will help the reader to understand the procedure that Singer and his colleagues followed. 22 This procedure has a number of advantages. It permits standardization of the different components of the index prior to their aggregation into a single indicator. The national capabilities of different nations can be compared without regard to the fluctuations of real capabilities in the system, and the number of nations in the sample can be increased at will and national comparisons can still be drawn, since each evaluation results in a different scale.

36

Chapter One

Table 1.1

Computation of National Capabilities Using the Singer-Bremer-Stuckey Military Dimension

Nation

A B C Totals

Military Expenditures

Industrial

Military Personnel

Iron-Steel Production

Real Units

%

Real Units

%

Real Units

%

1,000,000 1,000,000 I ,000,000 3,000 ,000

33.4 33.3 33.3 100.0

10,000 30,000 160,000 200,000

5.0 15.0 80.0 100.0

100,000 100,000 100,000 300,000

33.4 33.3 33.3 100.0

NOTE: Table prepared to illustrate how capabilities are derived. All data are imaginary.

Nevertheless, the procedure has some disadvantages that are particularly severe when one tries to make cross-time comparisons, for these can only be made so long as the nations composing the system remain the same. If there are alterations in membership of the system, comparisons become meaningless. This is an especially grievous handicap if one seeks to evaluate the merits of a dynamic model, such as the power-transition model, which demands a comparison over time. There is another problem. The measure that Singer and his colleagues have produced is a relative measure, in which the capabilities of one nation depend not only on its own performance but also on that of the sample as a whole and of every other nation in the sample. When the relative power of one nation declines, one cannot determine whether this is because that particular nation is doing worse or whether the average growth ofthe sample as a whole is improving; or, in the latter case, whether the overall improvement of the sample is due to a general increase in performance or to the increase in performance of one nation in particular. One cannot make a satisfactory deduction unless one goes back to the original data from which the percentage shares were computed. And such questions-which nation is doing better, which is catching up with its rivals, which is being outstripped by which of its rivals-are the keys to our tests of the three models.

37

The Power Transition

Model

Demographic Dimension

Dimension Energy Consumption

Total Population

Relative Capabilities

Urbanized

Real Units

%

Real Units

%

Real Units

%

100,000 50,000 50,000 200,000

50.0 25.0 25.0 100.0

20,000 20,000 160,000 200,000

10.0 10.0 80.0 100.0

15,000 10,000 25,000 50,000

30.0 20.0 50.0 100.0

Total (Indexl6) % Adjusted All Dimensions %

161.8 136.6 301.6 600.0

27.0 22.8 50.2 100.0

Comparison of the SBS and totaL output as measures of nationaL capability. Both measures have advantages and drawbacks. One major advantage of total output is the parsimoniousness of the index and, perhaps, the better quality of the data used in its compilation. On the other hand, the SBS measure, though more cumbersome and more inhibiting to over-time comparison, has the attraction of incorporating some direct measures of the social structure of countries analyzed and of the investments made on defense. It comes down to this question: How do the two measures perform? If one performs more satisfactorily, it should be chosen. If they perform equally well, then theoretical considerations, or considerations as to the utility of the indices for future research, or questions of the resources saved in gathering data, ought to determine the selection. The question of performance is central and cannot be resolved without a rigorous and systematic comparison of the two measures. This comparison was made and is fully reported elsewhere. 23 Here is a brief summary. The data used in the total output measure for the period 1870--1965 were transformed to make them entirely comparable to those used for the SBS scale. At every point of comparison (for every five years of this time segment), the same countries were selected for the GNP index as had been chosen by Singer and his colleagues for their own. Their GNPs were added and percentage shares calculated to

38

Chapter One

obtain a relative scale similar to the one developed by Singer. The two series were then compared by means of regression techniques. Because the more recent data were better than earlier information, two comparisons were made: one for the entire period (1870-1965) and the entire sample of countries, and another for the period 1895-1965. The results of the comparisons show that the two measures are similar and arrive at much the same scaling. When we used the full sample of countries, the two measures, while not identical, were highly correlated, with a coefficient of determination of .86. When we restricted the tests to the shorter period, we obtained a smaller standard error, in spite of the reduced number of cases, and a better overall fit in the regression line. The finding is reflected in the coefficient of determination, which moved from .86 to .95. Finally, in a country-by-country analysis that constituted our second test of reliability, we found again strong support for the view that the two measures make substantially the same evaluation of the behavior of the countries involved. The small differences observed between the two measures can be attributed to the unreliability of the data, unreliability that increases sharply as one goes back in time. We concluded, therefore, that so far as performance is concerned, there is no particular advantage in choosing one measure over the other. Doubts often expressed about the advisability of using a single economic indicator of overall national capabilities are' not warranted. We also concluded that the measure of GNP was to be preferred for three reasons. One is that the data available are probably more reliable for that measure than the several series gathered to construct the Singer index. Second, and perhaps more important, the GNP index is evidently more parsimonious from the user's point of view. Third, and most important, it was a theoretically more attractive measure. We therefore chose to use GNP in spite of our awareness of the inevitable weaknesses attending the utilization of a single series. 24 Alliance Behavior and Measurement of Threat-Perception We said at the beginning of this chapter that all theories suggest that the outbreak of major war is a result both of

39

The Power Transition

changes in the power structure of the international system and of the willingness of elites to fight in order to prevent or hasten the changes in question. The power-transition model, for example, argues that wars occur only when a dissatisfied great power catches up with the dominant nation. Satisfied powers do not fight. The balance-of-power advocates, on the other hand, contend that all nations will seek to attack their fellows whenever they gain a power advantage over them. Clearly, then, before this and other models can be tested, we must develop a measure of the willingness of elites to fight. We have dipped heavily into the recent work of Bruce Bueno de Mesquita to satisfy this purpose. 2S The indicators he provides are measures of changes in alliance behavior. The argument runs as follows. If alliances tighten, and interaction among alliance groups decreases, such behavior may be taken as an indication that those who have responsibility for guiding their countries in their international dealings perceive the environment as presenting a threat to the security and/or the power positions of their countries, and are preparing to fight. The opposite behavior-the loosening of alliances~an be taken as an indication that similar responsible elites have judged the danger to have passed or to have been a false alarm. As a consequence, peace should continue. Wars, of course, are not excluded if alliances loosen; but the frequency of their occurrence should be low and they should be presumed to be very largely the function of miscalculation. The last point is an important one, and encourages a reexamination of the theoretical structure underpinning the measurements. First, we should stress again that it is not simply the degree of tightness or discreteness in the alliance system but the shifts in these arrangements that are critical. Changes toward greater tightness or discreteness make clearer to elites which states are likely to fight with them and which against them and, therefore, make possible more accurate estimations of what resources will be available to them and to their opponents in the event of war. Obviously, this also clarifies the probabilities of winning and losing. If one assumes uniform rational behavior, this information becomes critical in a decision to go to war. Hence, the connection between (a) the elite's perceptions of threat, (b) the

40

Chapter One

tightening of alliances, and (c) the decision whether or not to fight. It is this interrelationship that underpins the measurements. In our scheme, therefore, alliance behavior is taken to measure threat. The assumption of uniform rational behavior requires comment. As Bueno de Mesquita uses it, this assumption screens out an important set of variations in power redistributions that obviously play an important role in the kind of decision-making we are discussing. The alliance measure cannot evaluate the tendency or propensity of different elites to take risks. Plainly, some elites may be more willing than others to take chances and begin wars, if the benefits of the conflicts are seductive enough and if the possibility of winning seems imaginable. On the other hand, more cautious leaders would require a different ratio between benefits and risk before launching a fight. 26 The assumption of uniform rational behavior also raises important questions in that it precludes consideration of irrational elites. In view of the historical record of elites in societies, stretching from the traditional to the modem, from the democratic to the authoritarian, it is plain that this is a significant omission. We have raised questions about the inclusiveness of the measure but not about its validity. Any set of measurements dealing with elite perceptions of danger and with their motives in deciding to go to war will inescapably rest on assumptions as to the nature of the paths those elites must traverse in determining whether an environmental change is a threat to the integrity or power position of their respective nations, and whether to fight in response to such threats. A definitive resolution of the puzzle represented by the nature of the mechanisms which influence elites in decision-making would be invaluable, but it is not yet in sight. Besides, such a resolution is not strictly necessary to meet the basic requirements of our effort here. Let us sketch how the threat-perception indicator was rendered operational. First, measures of alliance behavior were developed from an original scale built from four types of alliance: defense pacts, mutual nonaggression pacts, ententes, and no alliances at all. Defense pacts were consid-

41

The Power Transition

ered the greatest commitment between nations, while no alliance represented the least. These relationships were then scaled to reflect the clusters of nations with the greatest similarities and dissimilarities in their commitments. Measures of associations (using tau) were used to estimate the degree of tightness within each cluster in relation to the others. These measures were computed for every year and every nation in the period covered in the analysis. Using the tightness and looseness of alliances, we developed a simple eight-point scale that reflects both degree of commitment and the direction of change in commitment (see fig. 1.1). A positive position on the scale means that there has been a change in the tightness of alliances between the two actors in the pairs under consideration and, in addition, that each of the actors has increased its alliance commitments with other nations with whom the second nation in the pair also has alliances. And a position on the negative end of the scale means, in effect, the opposite of what we have just described-a cutting of ties with the opposite number in the pair and with its allies. A score at the extreme negative pole of the scale indicates that no ties have been maintained by either of the nations and their allies with the opposite number and its allies. Eachjudgement on degree of commitment and direction of alliances was made by observing the movements of coalitions over a period of twenty years. And there was still another contingency to cover to avoid

High positive

Positive

Low positive

Nonaligned

II Indifferent

Fig. I.l

Threat-perception scale.

Low negative

Negative

High negative

42

Chapter One

misunderstanding. There were situations where changes in alliance behavior occurred without moving the relationship of the pair being considered from one side of the scale to the other. For example, the relationship between France and the United Kingdom moved in the period between 1885 and World War I from a high position to a nonaligned position, but the movement was not sufficient to carry the relationship into the negative portion of the scale. In an absolute sense the countries remained friends, but they were less firmly committed than they had been earlier. And to eliminate any misunderstanding, we coded points on the scale stretching from nonaligned to positive as nonhostile, and all of the points stretching from the center of the scale to the negative pole as hostile. The "indifferent" position on the scale also requires a word of explanation. Nations to be categorized as "indifferent" are nations which do not have and never have had any ties with any nation in the system and have no record on which judgments or predictions can be made. The indifferent position, situated outside the scale, is occupied by nations outside the system. The United States and Japan were precisely in this situation in the nineteenth century. Nations entering the international system are also inevitably in this position. The Actors Our third task is to identify the nations to be classified as great powers, because this is a step essential to the selection of the wars whose outbreaks we wish to explain. Since major wars can only occur if great powers fight them, to know which nations are to be classified as great powers is a prerequisite for the identification of major conflicts. Moreover, we shall here explain the reasons behind our selection of the particular great powers that we consider the actors best able to test our propositions. The elite nations are few enough to stand out clearly from the rest of the members of the international system on such critical dimensions as population, economic productivity, and military might; international relations specialists have long agreed on their identity. The entire list includes the United

43

The Power Transition

States and the USSR/Russia, the United Kingdom and France, Japan and Germany, China and Italy, and Austria-Hungary.27 If we are to test fairly any connections between power changes and the outbreak of war, we need to select different countries from our master list at different times, because not all of the nations we have listed were great powers during the whole of the period covered in our study. The United States and Japan joined the ranks of the great powers in 1900. Austria and Hungary dropped out of the great-power class with the breakup of their unit after World War I. We also need to distinguish whether the nations involved are members of central or peripheral international systems, and whether the actors are major powers or contenders. The latter distinction is quite important. Contenders alone are strong enough to determine the direction the politics of the world order are to take. To account for such distinctions, we devised two different sets of criteria. The distinction between center and periphery is indicated by alliances among the relevant actors. (The reader will recall that the behavior of uninvolved nations cannot be expected to follow the rules of the powerdistribution models and, so, is not predictable.) Table 1.2 shows which nations belonged in which system in which periods. The table should make clear that in modem times European hegemony in international politics was complete. Up Table 1.2

Major Powers in Central and Peripheral Systems, 1860-1975

Nations Italy

France Austria-Hungary* Prussia-Germany-West Germany United Kingdom Russia-USSR Japan United States China*

* Data not available.

Years in System Center Periphery 1860-1975 1860-1975 1860-1918 1860-1975 1860-1975 1860-1975 1900-1975 1940-1975 1950-1975

1860-1900 1860-1940 1860-1950

44

Chapter One

through the nineteenth century, great powers were exclusively European, the substance of international politics was European politics, and world politics consisted of European quarrels, often over other portions of the globe. Although the United States and Japan began to be considered as great powers at the turn of the century, they kept their distance and were really not part of the central system. Only with World War II, when the United States (and to a much lesser degree Japan) became clearly recognized as having sharply outdistanced all of the European powers and had become willing participants in the central system itself, was the system inevitably expanded to include first the United States and then Japan. Most recently, Communist China has become the system's newest important member. This expansion of the central system from Europe to the world is the most critical change in international politics since World War II. And this distinction between the center and the periphery will prove important in one of the analytic steps we plan to take. Our second distinction, that between major powers and contenders, is made operational in a simple way. We have already argued that the most powerful nation in the world at any given time is always a member of the contending class. Any other nation whose score is at least as high as 80 percent of the capabilities of the strongest nation would also be considered a contender. When no other nation in a given period met this criterion, we considered as contenders the three strongest nations in the system. It is only in the central system that one needs to define the three most powerful nations, because it is only here that the power-distribution models apply and because, as one would expect, different nations at different times compose the triumvirate of the most powerful. Table 1.3 lists the most powerful countries. Two points should be noted. Italy and Austria-Hungary, which were on our list of great powers, never make our list of the three most powerful nations of the central system. The United States, on the other hand, surfaces as the most powerful nation in the system only during and after World War II. But the reader should bear in mind that the United

45

The Power Transition

Table 1.3

Contenders in the Central System Contenders The USSR/Russia United Kingdom France Germany United States Japan

Years

1860-1975 1860-1945 1860-1890 1890-1945 1945-1975 1950--1975

States passed all countries in potential power by the end of the nineteenth century and has maintained her lead from that time to ours. She appears on this list only with World War II because it was not until then that she had come to view herself as part of the central system. Test Cases: Total and Major Wars We know which nations will be the actors that will perform in our tests. Now let us tum to the selection ofthe conflicts that will serve as our test cases. The models we are comparing, it should be recalled, do not claim to establish connections between changes in the international power structure and the outbreak of wars among small nations, or among large and small nations; nor do the models explain colonial wars. Such conflicts (according to the models) may occur unrelated to fundamental changes in the power structure of the system, and, therefore, power distributions in the period preceding such wars cannot be used to disprove the propositions advanced by any of the models we are considering. The hypotheses in question can be tested fairly only if we locate conflicts whose outcomes will affect the very structure and operation of the international system. In short, what we need are major international military struggles. Our selection of wars, then, is based on three criteria. We thought that a conflict in which a major power actively participated on each side would escalate to proportions we would consider those of a major war. Thus, major-power participation in each opposing coalition became our first criterion. In order to insure that our selection would include

46

Chapter One

only wars in which both major powers involved made an all-out effort to win, we imposed a second condition: the conflicts selected would be those in which the number of battle deaths reached higher levels than in any previous war. The third criterion, particularly designed to insure that the contestants were really trying to win, was to choose struggles which would result in the loss of territory or population for the vanquished. It seemed reasonable to assume that if the elites of a country viewed, as a consequence of defeat, a threat to the integrity of the nation, they would prosecute the war with all available resources. The theoretical constraints we imposed in the selection of our sample reduced the number of conflicts available for analysis to five: the Napoleonic Wars, the Franco-Prussian War of 1870-71, the Russo-Japanese War of 19~5, and World Wars I and II. And of these the Napoleonic Wars had to be excluded because we do not have data series that go back that far. Clearly, four observations are insufficient for any attempt to generalize from regularities that might be found. The number of conflicts available for analysis could be increased if one ceased to treat collectivities involved on each side of the conflict as if they were one unit. But such increases would still not be enough, if one wished to carry on a number of tests. However, if we tested the behavior of individual nations rather than that of groups of nations fighting on each side in the conflict we would gain greater insight into the way the system works. Hence we decided in our analysis to investigate what happens when the alliances are disaggregated. We followed a two-step procedure. First, we paired each nation on our list of actors chosen as relevant for our tests with every other actor on the same list. Second, we located the periods in which, according to the models, the ratios of national capabilities of the pair of nations would make for war between the two members of the pair and then we determined whether the expected state of affairs actually took place. Whether war occurred was indexed by coding as "0" any year in which no conflict developed and coding as "1" every year in which one did. The best one can say for such a rigid

47

The Power Transition

dichotomization of our dependent variable is that this is obviously a wasteful way of measuring relations that, in reality, can range subtly in degree from full cooperation to armed conflict. But we simply had to bite the bullet; less gross measures of levels of cooperation and conflict among nations over time do not exist. One consequence of defining operationally, in the binary code "war/no war," the behavior we wished to explain was the necessity of making appropriate adjustments in our continuous index of power to match the new nature of our dependent variable. The procedure we used is described in the next section. Test Periods and Power Distributions Two major questions remained to be answered. The first question was how much power needs to shift from one actor to the other before war is likely to break out. Realistically, we could assume that only substantial shifts in the power loads among major powers would trigger the beginning of major wars. Since we had settled on an index for our dependent variable that discriminates only whether war actually occurred, we were not interested in an independent variable that would permit us to measure the inching of countries toward war, since the overall transformation is inevitably affected only by means of small yearly changes in power. It certainly would defeat the rules of a fair test to setup an experiment that could establish, nineteen times out of twenty, that the inevitably small yearly changes in power did not trigger off major wars, and to conclude therefrom that no real connection between power-distribution changes and major conflict could be shown. We solved the problem by turning the question of how far power needed to shift from one side to the other in the system before a war would begin, into a question of how much time needed to elapse either previous to or following the intersection of the power trajectories of the nations involved before a conflict could reasonably be expected to occur, thus arguing that the power changes and the outbreak of conflict were indeed connected. The possibility of turning the question of "how much power" into one of "how much time" also permitted us to reach another major goal of this

48

Chapter One

research. The proposItIOns set forth by the balance-ofpower and the power-transition models appear to contradict one another, and although both models may be incorrect only one of them can be right. To collect evidence enabling us to choose between them, we had to decide how long a period of time was needed to track the movement of all our pairs of warring nations in order to see whether the patterns of growth in power they established moved away from or toward one another during the periods preceding the conflicts. In posing the question of how much time needed to elapse before changes in power could be expected to trigger off a war, we were also asking how long the periods covered by our test would need to be. There is nothing in any of the theories we have discussed to indicate how long after the power changes that are alleged by the models to trigger a war adversaries are to be expected to initiate their fight. There is no guide to follow in the establishment of "reasonable length." Should one anticipate war or peace a year, ten years, or twenty years before or after the point when two countries become equal? Rates of growth prevalent in the system give a clue to a possible answer. Because such growth is slow, a relatively long period should be required to produce sufficient change in the power distributions between possible adversaries for war to break out. We thought that a period of roughly twenty years preceding each war would be sufficient time. We felt that the years of the actual fighting should be excluded from the analysis, so our estimates leave them out. The wars under consideration fall within each of the segments of time listed in table 1.4. One task remained before completion of our preparation for the, analysis. We had to compute the distribution of power positions between our actors and the rate of change among them over the six periods of table 1.4. We settled on two simple procedures. We first computed the power relationship by taking the ratios of the GNPs for the entire period for each pair of countries. The country that was less powerful at the beginning of the period was placed for the entire period in the denominator of our fraction. And we took the mean of this ratio to be our indicator of the relative standing of each actor for each full period.

49

The Power Transition

Table 1.4

Test Periods and the Onset of Wars War

Test Periodst 1860-1880* 1880-1900 1900-1913 1920-1939 1945-1955 1955-1975

Franco-Prussian, 1870 Russo-Japanese, 1904 World War I, 1913 World War II, 1939

* Lack of data prevented a start in 1850. t War years 1914-1918 and 1940-1944 were excluded from test periods.

With this measure as our point of departure, we next distinguished whether an equal distribution of power existed. In view of the imprecise nature of GNP as a measure of power, a ratio of the means larger than 80 percent was taken as evidence that equality existed between the powers. Smaller ratios were taken as evidence of inequality. A country was seen to have passed another when the nation that was less powerful at the beginning grew more powerful than the other member before the period ended. Empirical Tests of the Power-Distribution Models

Now, at last, we come to the analysis of our data. The first question we wish to have resolved is whether an equal or unequal distribution of power between the members of all the pairs of our sample is associated with the members fighting each other. The answer is instructive. When we match our two variables of war and power-distribution we obtain the results in table 1.5. Were we to go no further, the findings of this table would be quite disturbing. All of our cases are distributed in almost even proportion across the four cells of the table. Wars seem to occur both when adversaries are equal and unequal in power. In this initial step, then, power distributions are obviously not a predictor of the coming of war. The introduction of the concept of one nation surpassing another in power as the independent variable brings us an important new piece of information. Table 1.6 suggests the distribution of cases obtained when this is done.

50

Chapter One

Table 1.5

Power Distributions and Incidence of Conflicts Power Distributions Unequal Equal

No

81 (86.2%)

26 (81.3%)

13 (13.8%)

6 (18.8%)

126 pairs Tau B = .06 Not significant

N =

War Yes

Both equal and unequal shares of power between adversaries are associated with war. Now we see that if the distribution of power is equal, it is so simply because one nation is passing another and is abreast of it precisely at the time for which we have taken our sounding. But table 1.6 tells us something more. There is no case of military conflict among the most powerful nations of the world when power is shared equally by both members of each pair and when one member is not in the process of overtaking the other. In other words, at the level of great powers, wars occur if the Table 1.6

Power Distributions and Incidence of Conflict When Nations Overtake One Another in Power Power Distributions Equal, No Equal and Unequal Overtaking Overtaking

No

81 (86.2%)

11 (100%)

15 (71.0%)

13 (13.8%)

0

6 (29.0%)

War Yes

126 Tau C = .05 Not significant

N =

51

The Power Transition

balance of power is not stable-if, and only if, one member ofthe pair is in the process of overtaking the other in power. This represents an important clue as to the manner in which the international system works. Nevertheless the evidence presented in the table remains inconclusive in regard to the central question we posed: Which of the power distributions suggested by our models actually obtains before an outbreak of hostilities? The stubborn reluctance of the data to disconfirm one of the hypotheses begins to weaken if we separate major powers from contenders and peripheral countries from full members of the central system. Only then does the distribution of cases across our table point clearly in the direction to be followed for a solution. Table 1.7 presents the distribution of cases after application of the new controls. The table displays quite clearly the fundamental relationship we wished to trace and does this despite the paucity of the data in several of the table's cells. The major point obtained from the information displayed in the final portion of the table is that, if conflicts occur among contenders, they do so only if one of the contenders is in the process of passing the other. This process is clearly, however, a necessary but not a sufficient condition for conflict, because it can also take place without conflict. This finding was obscured in the previous table because we did not account in it for the possible differences in behavior between the major power and the contenders in this regard. There can be little doubt left of the correctness of the powertransition explanation of conflict behavior in the contenders' class. Among major powers, on the other hand, our specific power distributions are not good predictors of oncoming conflict. In their case, the table shows that the same proportion of conflicts occur when the two combatants are equal in strength as when they are not. The major powers seem to fight, whether they are weaker, as strong as, or stronger than their opponents. Such data closely resemble the inconclusive findings in table 1.6. Finally, if one looks at the information on conflict behavior in the periphery in the first portion of table 1.7, one finds that all conflicts occur when combatants are all unequal in strength. In this portion of

52

Chapter One

Table 1.7

System Membership, Actors' Rank in Power, Power Distributions, and Incidence of Conflict Power Distributions Periphery

No

Unequal

Equal, No Overtaking

Equal and Overtaking

26 (86.7%)

I (100%)

4 (100%)

4 (13.3%)

0

0

35 Tau C = -.07 Not significant

N =

War

Yes

Center: Major Powers Unequal No

51 (100%)

Equal, No Overtaking

Equal and Overtaking

4 (85.0%)

6 (85.7%)

0

I (14.3%)

N = 71 Tau C = -.03 Not significant

War

Yes

9 (15.0%)

Center: Contenders Equal, No Equal and Unequal Overtaking Overtaking

No

4 (100%)

6 (100%)

5 (50.0%)

War

0

Yes

0

5 (50.0%)

N = 20 Tau C = .50 Significance = .01

53

The Power Transition

international relations, the balance-of-power system seems to work (i.e., whenever a balance of power is in effect there is no war). But, ironically, this finding seems to confirm the most serious doubts as to the utility of the model, when one recalls our conclusion (and the reasons for it) that we cannot predict anything about the behavior in the peripheral portion of the system. We have examined the power distributions as a possible cause of war. Now let us gauge the possible effects of alliances. We should recall that the movement of alliances is our operational indicator of leaders' perceptions of threats. We argued that, if leaders of nations tighten their alliances, this may be taken as a sign that they are frightened of their environment and that they are preparing to fight to protect their nations. On the other hand, if the same leaders loosen their ties with other nations, this may be taken as a sign that they see no danger to their nations from movements in their environment. Of course, we hypothesized that perceptions are as important as capabilities in explaining the oncoming of wars, and this, too, is a hypothesis we will wish to test. When we look at the role of alliances as factors in creating conflicts among our sample of pairs we find, not at all surprisingly, that alliances do make a difference. 28 Table 1.8 displays the data on alliances and war. In the table we have used only a three-point scale of threat perception: negative, positive, and neutral. Alliances taken alone are associated Table 1.8

Alliance Formation aDd the Incidence of War

No

Positive

Neutral

Negative

20 (95.2%)

14 (93.3%)

17 (65.4%)

I

1 ( 6.7%)

9 (34.6%)

War Yes

* No

( 4.5%)

Italian data; periphery not applicable.

N = 62* Tau C = .29 Significance = .01

54

Chapter One

strongly with the occurrence of war. In only one case where alliances loosened before the conflict did the pair of nations move into armed conflict. On the other hand, in nine out of a total of eleven cases where war occurred, alliances measurably tightened before the conflict. We should emphasize that the table does not include data on behavior of nations on the periphery because by definition such nations have no alliances with any of the nations in the central system, and it is the lack of interaction on their part that prevents one from predicting their behavior in initiating or joining conflicts. There is an obvious contradiction in the results of our tests up to this point. We have found the power-transition model to be a good predictor of the coming of war, yet the model discounts alliances as a factor. On the other hand, our tests clearly find alliances to be an important factor in the initiation of major conflict in the central system. To resolve this matter, we shall attempt to approximate as closely as possible a full specification of the model of the power transition. Three elements seem to be critical. 1. The overtaking of one nation by the other will lead to conflict. 2. Rates of growth will influence the probabilities of conflict. The faster the challenger overtakes the dominant nation, the greater the chances that the two will fight. A slower overtaking by the challenger of the dominant nation should diminish the likelihood of war. 3. Alliances should not playa major role in the initiation of war because they are presumed by the protagonists to be reasonably permanent. The test of this complex proposition required the utilization of every bit of the data we had generated. We used the continuous variable developed to determine the power ratios between nations to indicate the degree of closeness between the power capabilities of the members of each pair of nations. We multiplied it by the indicator of convergence and divergence between the two countries to obtain an accurate picture of whether or not one country passed another and how rapidly this process had been completed. We further used the full alliance scale to indicate the degree of threat the leaders felt. The sample was divided between

55

The Power Transition

contenders and major powers, but those in the periphery were again left out because our theory asserts that the interaction of these countries cannot be predicted correctly. 29 The results of probit analysis are shown in table 1.9. 30 Table 1.9

Test of Power-Transition Model Major Powers War = 1.0 + .18 Relative Power x Growth + .71 Alliance Structure Std. Error: (1.03) (.12) a proportion (.27) explained: (0.0) Proportion explained: .27 Significance: .004 N =

44

Contenders War = 12.4 + 7.0 Relative Power x Growth + .7 Alliance Structure Std. Error: (2.16) (.43) a proportion (.57) (.06) explained: Proportion explained: .63 Significance: .0008 N = 17

As a result of this analysis, all of the pieces fall into place, and we can at last describe which factors in which cases play an important role in pushing different actors to fight with one anotherY The conditions for war appear as follows. Consider the contenders. Two factors predominate in bringing about any conflict where dominant nations and challengers contend for first place: the power position of the two nations relative to each other, and the speed with which the challenger is passing the dominant nation. The interaction of these two factors accounts for 57 percent of the proportion to be explained. On the other hand, only 6 percent is accounted for by the movement in the structure of the coalitions that make up the international system. We have further demonstrated that the contributions of relative power and of the speed with which the overtaking occurs are approximately equal. Clearly, the contenders are so

56

Chapter One

strong and dominate the international scene so completely that, in deciding whether to plunge the whole system into war, they are almost impervious to the claims of other nations or to their plight. Consider now the major powers. In their case the two factors that explain most of the difference in the behavior of the contenders account for nothing at all. Alliances, which in the case of contending major powers accounted for almost no change in the incidence of armed conflict, are precisely the factor that now accounts for the behavior of the contenders. The proportion explained by the behavior of alliances accounts for all but 1 percent of the 27 percent explained by the model in these cases. Clearly, in the decision processes of major powers, the most important common factor in determining whether or not they fight is the ties they have with other nations. Their experience is, thus, exactly the opposite of that of the contenders. What is one to make of the" dependence" of great powers in the past? It is certain that a finding of such dependence would not surprise us in the nuclear age, but one has the impression that great powers acted under far less constraint in the past. The obvious inference is that even before the nuclear age the great powers, although far stronger than the middle and small powers and very much involved in the diplomatic interchanges that preceded wars, were not the initiators of major wars. They fought when others decided to fight. The principal difference in their role then was that, when they fought, their intervention was of critical importance in deciding the outcome of the conflict; now they can make no major contribution to victory in a nuclear war. In the nuclear age, the preeminence of the superpowers seems complete. Our findings confirm most of the major tenets of the model based on the power transition. 32 To be sure, the model does not predict war with certainty but it outlines with great clarity the necessary but not sufficient conditions for war. War is associated with shifts where one contender passes another in power. The speed with which the challenger overtakes the dominant nation is an important variable in governing the chances that the passage will be carried out in peace. Finally, alliance commitments to other nations do

57

The Power Transition

not count for much with those that have the final say in initiating major wars, but they are critical in the case of major powers. Our comments on the role of alliances represent a major part, but still only a part, ofthe story of the manner in which alliance behavior can inform our understanding of the determinants of the outbreak of war. One can gain glimpses of other ways in which distributions of power, including those within and across alliances, may have an effect on destabilizing the system. Now statistical analysis (because there are too few cases) must be set aside. But one can explore for possible hints in two of the four cases of major wars in which alliances playa major role and for which data are available. What one needs to do is to view the behavior of national capabilities of original combatants-whether they are individual nations or coalitions-in the light of the familiar hypotheses in our model connecting power structures to conflict. Such an approach can be fruitful of new insights and hypotheses, which, of course, should not be confused with evidence from which generalizations can be drawn. Of the four conflicts for which we have data, the FrancoPrussian War of 1870 and the Russo-Japanese conflict of 1904-5 are not useful to us because the two principals concerned in each case fought unaided by allies. 33 World Wars I and II, on the other hand, are two conflicts in which large coalitions were involved on each side of the military struggle. Our estimate of the power resources (measured in GNP) available to the two Central Powers (Austria-Hungary and Germany) immediately before World War I amounts to 62 percent of the total resources available to the Allied side (Russia, France, and the United Kingdom). The estimate of resources available to the Allies is in our view inflated by a Russian GNP value which reflects the substantial but largely unmobilizable population of the country at the beginning of the First World War. On the other side, if one adds Italy to the coalition of the Central Powers (and Italy was indeed a member of that coalition almost up to the beginning of hostilities), the pool of resources available to the Central Powers was 77 percent of the resources available to the Allies. Roughly, then, the power positions of the two coalitions just

58

Chapter One

before the world conflict began was roughly what the model based on the power transition would lead us to expect. Even more important, if one looks at the vectors of power of the two coalitions, one sees that the sides close the gap separating them over the twenty years preceding the coming of the war. In World War II, one sees a repetition of the behavior we have just described. If one considers the three major powers that entered the war in the first days after hostilities began, and whose differences transformed the German attack on Poland in 1939 into a world war, changes in the levels of national capabilities move very much in the way the power-transition model leads one to expect. Immediately before the war the value of the pool of capabilities available to Germany alone is roughly 90 percent of the pool available to the Allies (France and the United Kingdom). If one adds Italian resources to the German pool, although Italy entered the war almost a year after it began, the Axis is somewhat stronger still than the Allies. Either way, the two sides are approximately equal. If one looks back twenty and ten years before the war began, the two sides are far apart, with the countries that later become the Axis powers gaining ground rapidly. Again, the evidence seems to support the hypothesis prescribed by the power-transition model. But when coalitions are involved, the change in the levels of the pools of resources available to the combating sides are only part of the story. If one continues to probe and singles out in one coalition the dominant nation and in the other the challenger, and compares the resources available to each of them, one finds that shifts in the amounts of resources available to the two nations leading the stronger and the weaker coalitions may be an important element in moving the two sides toward war. In the case of World War I, the United Kingdom begins with an advantage over Germany some twenty years before the war, with Germany catching up by 1905. By 1913, Germany has clearly surpassed the United Kingdom. Mter losing the war, Germany drops behind the United Kingdom in 1919 (Germany has at this point 84 percent of the power of the United Kingdom),

59

The Power Transition

catching up with it in the early twenties and retaining a minimum advantage for the ensuing decade. Then, the "scissors" begin to open. By the time World War II breaks out, Germany has a significant advantage over the United Kingdom. The pattern one can establish by considering first the power of each of the contenders taken alone and then that of the entire coalition they form seems to be as follows (see fig. 1.2). When considered separately, the challenger has overtaken its major rival prior to the eruption of the conflict. So the temptation on the part of the dissatisfied challenging power to risk trying to break the opposing dominant nation by force of arms is understandable, as are the hopes of the challenger for a quick victory. After all, the challenger is stronger than any single nation clearly ranged against it. Force of arms appears the only way by which it can, in the short run, undertake to impose its will. Equally understandable, however, is the outcome of the war that the challenger

ii:' z

1.60

GERMANY UNITED KINGDOM 1939

Cl

ii: z

~ i!?

1.50

""c

1.40

8 go

1.30

~

YEAR OF POWER OVERTAKING

•

0

E ~

~

1.20

~ 0

c.

'0 0

:; 0:

1.10

tOO*

USSR - GERMANY 1941 PRUSSIA-FRANCE 1870

•

GER~NY-UNITEDKINGDOM

1914

.GERMANY-RUSSIA 1914

-5 Time (years)

Fig. 1.2

Power ratios between challengers and dominant nations, and timing of overtaking. NOTE: Challengers are listed first, dominant nations second. * Contenders are equal in power.

60

Chapter One

thus initiates. The coalition which the hitherto dominant nation has put together cannot be overtaken because the challenger has fewer and weaker friends than the dominant nation, and cannot muster a coalition capable of overcoming the combination of powers ranged against it. Our analysis leads to a possible insight on yet another point. If one looks at the point at which conflicts begin in relation to the point at which the trajectories of the challenger and the dominant nation intersect, it is clear that conflicts occur after the intersection when the two nations fight alone (which is contrary to what the power-transition model leads us to expect), but before the coalition of the challenger overtakes the coalition of the dominant country. With such small numbers one has, at best, traces of trends, but this seems a curious pattern. Conclusions are obviously impossible. One could at most hazard some plausible explanation that might constitute the beginning of the formulation of a hypothesis. When two nations fight alone, there can be little doubt in the defender's and attacker's minds what their respective positions are and what will be the prospects for each if things are left to drift. On the other hand, when alliances are present the challenger may be in a position to afford to hesitate longer, for there is always hope that some important country will be separated from the rest of the defending coalition, thus tipping the balance. The dominant nation, secure in the support of the stronger coalition, also may tend to procrastinate before it faces up to the necessity of trying to tum back the foe. Our data ever so indirectly suggest why some theorists believed the balance of power brought about wars. We must hasten to add that, while much of our scenario drawn is based on evidence, a good deal of it is still based on conjecture. Our data reveal that the largest number of conflicts have occurred after the point of intersection of the power trajectories of the competing countries. The sightings that led to the refurbishing of the balance-of-power theory just before the tum of the century were taken when the aggressor seemed to tower over the dominant nation and the major powers of the defending coalition, and the challenger's trucu-

61

The Power Transition

lence and superior strength seemed plain for all to see. An imbalance of power seemed clearly to bring about war. To assume that a balance would bring about peace seemed sensible. What was missed was critical but not plainly visible: the fact that the challenger had been the weaker party only a generation earlier and had leapfrogged over the dominant nation. If this is so, the theorists of the balance of power of the time committed the cardinal sin frequently indulged in by social scientists of building dynamic models with no longitudinal data at their disposal and with observations drawn from, at most, one point in time. Inevitably, they made assumptions and inferred a large number of behaviors that fit the data they could see. Their guesses were entirely plausible but also entirely wrong. This long footnote to the possible effects of alliances on the conflict behavior in the system concludes the analysis of our data. Conclusion Our probes point firmly to the fact that the basic propositions in the balance-of-power model miss most if not all of the critical behaviors our data show to be responsible for moving a whole system of nations toward major war. It is not only details in the model that are in error. The conception of the system that underpins this model seems to be wrong. It is the model based on the concept of the power transition that specifies correctly the behaviors, and the connections between behaviors, that our data show to be the necessary conditions for major wars to break out. The mechanisms that make for major wars can be simply summed up. The fundamental problem that sets the whole system sliding almost irretrievably toward war is the differences in rates of growth among the great powers and, of particular importance, the differences in rates between the dominant nation and the challenger that permit the latter to overtake the former in power.34 It is this leapfrogging that destabilizes the system.

62

Chapter One

The relative speed with which both countries travel on their power trajectories is also important. The faster one nation overtakes the other, the greater the chances for war. Finally, this destabilization and the ensuing conflict between giants act as a magnet, bringing into war all the major powers in the system, dependent as they are on the order established by their leaders for what they already have, or for what they hope to gain in the future if they upset the existing order. And alliances are important as a cause of war in yet another way. While it is true that the challenger overtakes the dominant nation and that at the outset it is the challenger who is the stronger, it is equally true in the two cases tested that the coalition with the dominant nation is stronger than the coalition shaped by the challenger to unseat the leader and recast the international order. When there are changes in the levels of power of the two leaders and in those of the two coalitions both sets of changes are responsible, in different ways, for bringing the two sides to the point at which they fight. One final point. Anyone who probes for regularities in the conflict behavior of the tiny set of elite nations cannot but feel uncomfortable about the thinness of the evidence on which one hazards generalizations. Given the dearth of cases available for comparison and the few observations generated from them, how valid can such generalizations be? How can one ever be sure of not having over-stepped the bounds of the evidence? We can only repeat our warning: the results we have presented must be treated not as definitive answers but as tentative findings. Nevertheless, results must still be taken seriously. Just how serious such findings may prove to be and how powerful the theory which led to their unearthing really is, can perhaps be suggested by the following simple test. If we were to assume that, in each case we have studied, the leapfrogging process between contenders and the war that followed were totally independent events, what would be the likelihood of those events occurring simultaneously, as they actually did? Such a probability can be calculated in two ways. If one takes into account the entire period of our study, year by year, there is one chance in ten thousand that the two phenomena would

63

The Power Transition

have coincided had they been independent. And there is one chance in six thousand that the two events would have occurred together, if we use the period of twenty years as a temporal frame of observation. With such results, it seems reasonable to assert that the theory backing our findings must have a good deal of explanatory power. As we suspected, then, power and power changes are some of the fundamental reasons why wars occur. But the sources of national power of any nation are but the patterns of socioeconomic and political development; and it is differences in the rates of change inscribed in these patterns that we think are responsible in the end for the fact that wars break out. The reader will consider this finding depressing. And, indeed, so do we. The trends in question are not reversible, nor when they act in our favor would we choose them to be if we had the choice. Industrialization, urbanization, political mobilization, and the drafting of the population into political structures cannot be easily controlled. Such trends are simply not manipulable in response to foreign-policy needs. Nor are their political consequences at the international level. International political engineering, especially among great powers, has more myth than reality to it. Let us now tum to a consideration of whether the patterns of growth, and the consequent responses of power, determine the outcomes of conflicts they seem to have caused.

Two

Davids and Goliaths

64

We have now explored the conditions that make for war and are ready to explore how such conditions help to bring about victory or defeat once the fighting has started. In this chapter we shall try to see whether there is a way to predict which of the combatants will win and which will lose in a conflict. Caution suggests that we avoid prediction. Yet repeated and accurate forecasting of events is the most useful validation of scientific theory. Predicting the outcome of wars is exciting business. When war looms or is in progress, its probable outcome is vital information to all concerned. In this chapter we shall not be predicting-not quite-but rather forecasting the outcomes of conflicts that have already occurred. That is, in social science jargon, we shall "post-dict" some conflicts. Scientifically, there is no difference between the two procedures; in practice, the step we are taking here is equally essential for the forecasting of events that have not yet occurred. Prediction is more difficult than any of the problems we have so far discussed. The careful reader may have guessed that the crucial issue in this regard will be the measures of national power we have used. While the defects of our measures have not been ruinous in the first chapter, they will become critical in any attempt to deal with conflicts among developing nations or between developing and developed nations. Our incapacity in the past to predict winners and losers in such international conflicts is clearly rooted in our lack of knowledge with respect to the proper measurement of power. A major source of past error has been our inability to estimate in a rigorous and systematic manner the "political development" of a nation, that is, the capacity or the

Predicting the Outcomes of International Wars

65

Predicting the Outcomes of International Wars

effectiveness of a government in doing its job. 1 What procedures there are for the appraisal of the performance of national economies and social structures have not yet been applied successfully in the evaluation ofthe performance and/ or the capacity of governments and political systems. As a consequence, estimates of national power have been and still are based in very large part solely on socioeconomic and/or military data; direct measures of political performance have had to be left out. But if we could measure the political capacity of a nation and combine such estimates with other data relating to social and economic performance, our evaluation of the total strength of each combatant should be sufficiently improved to allow us to predict with reasonable success the outcome of any total but nonnuclear war. This assumption is our point of departure for this chapter. 2 We chose with the greatest care the conflicts whose outcomes we tried to predict. First, we sought wars in which the major combatants tried wholeheartedly to win the conflict to which they had committed their resources. Also, we selected examples in which, had earlier methods of measurement been used, the results would have been in error. 3 Four recent conflicts satisfied our criteria. Three of them had outcomes different from those one would have expected at the onset. The four conflicts are: the wars between the Arabs and the Israelis; the war between North and South Vietnam, with the assistance of the Soviet Union, the People's Republic of China, and the United States; the very brief struggle between India and China in the Himalayas in 1962; the Korean War, in which North and South Korea were the combatants of record but where the major portion of the fighting was done by the United States and the People's Republic of China. If the results of our measurements are congruent with what actually occurred, we shall have made an important first step toward predicting the outcomes of conflicts and toward validating our overall measure of national capabilities. If that measure is valid, then we shall also have evolved a measure of political development. 4 The information in this chapter is presented in the following fashion. We begin by explaining our suspicions of the

66

Chapter Two

inadequacy of existing measures and by pointing out why habitual procedures for estimating national capabilities have proved deficient. We move next to the proposal of an index of political development and a demonstration of how this measure has been combined with existing socioeconomic indices to obtain a comprehensive estimate of the strengths of nations. The final section tests our new measures of national capabilities. As noted, we intend to determine whether or not the model we have constructed permits the estimation of the relative strengths of combatants in recent wars with sufficient accuracy to enable us to predict the actual results. Power Indicators: Existing Measures In many respects, the measures of national capabilities based on hard data from national accounts distort reality in the same way as do intuitive procedures. A particular weakness is that such measures perform unevenly in comparisons across time and across countries. Although socioeconomic indicators in the case of developed countries can generate some fairly reliable estimates of national capabilities, the same measures, applied to other systems, lead to substantial errors. Thus, time series of socioeconomic data provide very uncertain footing for the researcher interested in indexing national capabilities. As we have indicated in Chapter 1, a half-dozen methods of evaluating national capabilities have been proposed, but only two are developed to the point where they may be readily used in cross-national and cross-temporal comparisons. We have already reviewed extensively and compared systematicaily the two indices in question: gross national product as proposed by Organski, and the measure developed by Singer and his colleagues. We shall not go over again the arguments presented. It should suffice to remind the reader that we chose gross national product because we felt that it was theoretically the more attractive and more parsimonious model and because the data on which it was based inspired us with greater confidence. Our analysis in all we have written to this point was based on estimates of total product.

67

Predicting the Outcomes of International Wars

One issue not previously discussed in the comparison of the two indices is that of their validity in estimating the national strength of actors at different points of development. Because that issue is critical for our attempt to forecast the outcomes of nonnuclear conflicts, we designed a test to evaluate the performance of these two measures of national power in this regard. If one uses an estimate of the relative strength of two combatants (both of whom, it is assumed, are exerting themselves to the utmost), to predict accurately the eventual winner and loser ofthe struggle, one then can be reassured about the validity of the measure employed to make that prediction. To subject gross national product and the Singer et al. index to this kind of test we needed first to render the two measures comparable. We followed for the GNP measure the same two steps Singer et al. used to construct their measure. The GNP values for the countries we wished to compare were aggregated to provide a total value of capabilities for the system. Then each country was assigned the percentage share that each GNP represented of the system's total. We used the Arab-Israeli war of 1967 as a test case for the performance of the two measures. What we wished to determine was how well the GNP and Singer et al. measures predicted the outcome of a conflict whose result was known. Table 2.1 presents the results of our comparison. 5 Table 2.1

Comparison of Power Shares of Egypt and Israel before 1967 War

Israel Egypt Total

Organski-Davis GNP Index

Singer-Bremer-Stuckey Power Index

39%

27% 73% 100%

61% 100%

Two points are evident. The Singer et al. measure distorts reality far more than does the GNP measure. Although for our purposes here this is the less important of the two results, it does suggest that the addition of more indicators does not improve the overall estimate. The second finding, on the other hand, is critical. Regardless of the differences

68

Chapter Two

between them, both measures widely miss the mark. It is impossible to consider indices trustworthy that distort the strength of combatants and give rise to completely erroneous expectations of the outcome of a conflict. The results we have obtained are not the consequence of simply comparing the two countries at a single point in time. More extensive comparisons would not provide significantly different results. If, for example, we use GNP as a measure of national capabilities and compare the United Arab Republic with Israel for the entire period from 1955 to 1975, the results are similar to those indicated in table 2.1. By both methods, the Arab side is made to appear much stronger than the Israeli, and we know this appreciation to be at odds with reality. Even more drastic distortions are evident in the case of Vietnam. 6 Any measure of national capabilities that is to be considered acceptable must perform far better than these in forecasting winners and losers in military conflicts. The Missing Measure of Political Development

The measures of national capabilities we have tested are not entirely wrong. They are adequate measures of the power of developed countries. Such measures fail mainly in cases in which a developing and a developed nation, or two developing nations, go to war with each other. In these instances, a measure of "political capacity" is essential. Yet so far none has been available. This may seem incredible, for the essential meaning of "the development of a political system" has been a major concern of political scientists since the middle 1950s; but their efforts have provided little help to our work here. With some exceptions, the work done under the rubric of political development has been largely theoretical. By and large, however, the body of material developed in the I%Os, when interest in the exploration of the problem was at a high pitch, does not provide much direct assistance to the question we pose. 7 Most important, literature in this area does not offer measures of political change that can be applied across time and nations. There were some notable exceptions to the theoretical

69

Predicting the Outcomes of International Wars

orientations of the early wave of studies, and among these stand out the studies by Adelman, Morris, Inkeles, Deutsch, Rokkan, Gurr, and Cutright. 8 Our dissatisfaction with some of the measures these authors have proposed centers on their almost universal choice of additive procedures for the aggregation of the indicators they deploy, the excessively large number of such indicators, and their crosssectional rather than cross-temperal approach. Most important of all, most of these measures seem to draw on behavior already well monitored by measures of economic performance. 9 Beyond the work mentioned, other materials should be cited. There is a very large body of research in the field of political behavior (particularly comparative political behavior) that has been concerned mainly with the degree of representativeness and the levels of participation in political systems. This work is both rigorous and systematic, and addresses itself to an important aspect of political development. With the behavioral revolution, it became possible for research to treat empirically some fundamental propositions of democratic theory, and the democratic bias in the research carried out was accentuated. Substantive concerns with participation, representation, socialization, electoral behavior, identification with political parties and political institutions generally, the development and transmission of political ideas, preferences, attitudes, and belief systems are, however, in large part questions of the "quality" of political life. They probe only very indirectly into the question of how things get done. Yet how things get done is precisely what we seek to find out. The question we need to pose is not, "Are elites representative of mass publics?" Rather we want to know whether elites have the tools to generate resources to be used for national purposes. We ask questions diametrically opposed to those that Western and, especially, American scholars in the field of politics have been posing. The question to which we require an answer should be phrased as follows. Do elites have the tools to extract human and material resources from their societies, aggregate the many contributions each citizen makes into national pools, and use

70

Chapter Two

them for national purposes? In this form, it seems to us, the question relates directly to the capacity of a political system to perform the tasks required of it. In short, then, the reason why we do not have the answers we need is that, over the last decade and a half, the proper questions have not been asked. Ironically enough, this omission is principally the result of the democratic bias of reasearchers. There are other reasons why the need for direct measurement of political development has finally emerged. In the past, it was taken for granted that economic and political development go hand in hand. The failure to measure political mobilization separately was not perceived as a serious lacuna, for it could be assumed that if a nation possessed a high level of economic productivity it also possessed a high level of political capacity to mobilize the human and material resources in the national society. 10 This seemed reasonable. In the Western experience, the expansion of political networks penetrating mass publics, and of political systems capable of sustaining their societies through major efforts, grew in response to a desire on the part of the masses for governmental protection against the vicissitudes of industrial life. Because in Western Europe, in European enclaves overseas, and in Japan, the effectiveness of the political systems kept pace roughly with socioeconomic changes, it was possible to infer political development from the scores of key socioeconomic variables. If the latter were relatively high, one could assert with confidence that, were such measures available, scores would be as high for the political variables. The error lies in assuming that low levels of economic productivity and low effectiveness of the political system are similarly associated. The last three decades have witnessed repeated examples of nations experiencing a degree of political mobilization far in advance of substantial economic development. This is particularly so in the case of non-European Communist countries, but it is also true, in less vivid form, of other developing countries as well. Because of this abrogation of the pattern of development along the lines of the Western model, economic and social performance ceased abruptly to be uniformly a good predictor

71

Predicting the Outcomes of International Wars

of political performance, and the need for separate measures became imperative. The change in the sequence of development had immediate, massive effects on international behavior. The incredible began to occur. The weak, not content to inherit the kingdom of heaven, seemed bent on conquering the earth as well. The Viet Cong and North Vietnamese mauled and finally fought the Americans to a standstill, as the Chinese had done fifteen years earlier in Korea. The Israelis, in their new desert site, defeated the forces of countries collectively larger than Israel by factors of fifty to one hundred. The root causes of what was happening did not immediately become apparent to policy-makers, and the implications were not clear to scholars, either. But even when the problem was identified, no one produced a satisfactory answer, and the matter remained unresolved. Such are the circumstances that have contributed to the absence of the required measure of political development and the reasons why the need for such a measure became visible. Construction of a Measure of Political Development

Before considering a measure of political development, we must state succinctly what it is we seek to index. We begin with a definition, for the expression "political development" has become thoroughly muddled. Some writers define it as the transformation of a population from subjects into citizens; still others describe it variously as the spread of political participation, as the building-up of a state structure (e.g., the civil administration and the army), as the mobilization of a population, as the development of mass political parties, as the increasing capacity of the political system to direct the socioeconomic subsystems, as the development and dissemination of nationalistic feelings, as the differentiation of institutional structures, or as the increasing capacity of the political system to handle the load imposed on it by its environment. I I Each such definiton appears defensible and we do not intend to dispute any of them. We must, however, elaborate what we mean by "political

72

Chapter Two

development. " This is the capacity of the political system to carry out the tasks imposed upon it by its own political elite, by other important national actors, or by the pressures of the international environment. It is evident to us that a highly capable political system need not be free, democratic, stable, orderly, representative, participatory, or endowed with any of the other desiderata alluded to by laymen and experts as bases for evaluating the political life of a nation. One may well argue that nondemocratic, nonparticipatory, or nonrepresentative systems could not be regarded as developed in any "normative" sense. But we do not consider normative criteria relevant to the information we seek. It is also clear that, at the level of the individual, political development is mostly a matter of individual behavior and attitude. For individuals, "development" in the field of politics means awareness, political participation, a feeling of efficacy, and a realistic appreciation of whether or not problems can be disposed of by political means. But at the level of the system, political development is only in small part a matter of individual attitudes. To assert that one polity is more developed than another means that it can generate more human and material resources, ceteris paribus, to accomplish necessary ends. Political development means capacity, and capacity is dependent on political performance in two areas: penetration of the national society by central governmental elites to control as many subjects/citizens as possible within the political jurisdiction of the state; and the capability of the government to extract resources from its society. Penetration and extraction are the behaviors we need to index in order to measure the extent of internal development of the national polity. The power to penetrate and the power to extract are obviously related to each other. If government has successfully penetrated a society, one should expect it to perform well in extracting resources. The reasons are not hard to find. Historically, central governmental elites, buttressing the dynastic or colonial regime, sought incessantly to penetrate the mass of the population to extract resources. Louis XII, Louis XIII, and Louis XIV of France, the Tudors of Great Britain, Peter the Great of Russia, Frederick the Great of

73

Predicting the Outcomes of International Wars

Prussia are examples from a long list of monarchs who sought ever greater control of their nobles and the mass of their subjects in order to obtain increasing pools of resources to support their armies, which, in tum, were the indispensable support of the dynasties' "right" to rule. Only by compelling their "estates" to work harder, produce more, and collect more could the dynasts be sure of the security of their own power, and even of their persons. The motive power and mechanisms that gave rise to the nation state were brutally simple. Though penetration and extraction should be highly correlated in most cross-national comparisons of developed nations, they should behave autonomously in the very initial stages of political development. In such cases one cannot be certain by direct observation of one factor that the other is always present or is operative to the same degree. We should also note that penetration and extraction may spread very unevenly across a national society. One sector may be highly penetrated and the central government may be able to extract from it a large portion of the wealth it produces, while another fraction is scarcely affected by central power, and is permitted to live by its own rules, contributing little or nothing to the central pool of resources. The tribal peoples of the Philippines, the Maoris of New Zealand, the Kurds of Iraq are but three extreme instances among dozens of others from which to choose. Although penetration and extraction are both essential components in the process of political development, extraction is often the more complicated, costly, and dangerous process because, at least initially, populations may resist governments' intrusion into their lives. Governments, especially weak ones, have sought to have their cake and eat it too, by obtaining resources and avoiding face-to-face confrontations with the people who pay the bill. Indirect taxes, tariffs on imports and exports, value-added taxes, and taxes on foreign enterprises are ways to raise revenue while minimizing the need to deal directly with the national population. To summarize, we define political development as the capacity of the political system to fulfill tasks imposed by both

74

Chapter Two

its domestic and international environment. This capacity of the political system rests in tum on its performance in penetrating the society and in extracting resources from it. An Index of Governmental Extraction It should not surprise the reader that we should tum to the field of taxation in order to transform our theoretical concerns into operational measures of penetration and extraction. Taxes are exact indicators of governmental presence. Few operations of governments depend so heavily on popular support-or on fear of punishment. Few affect so directly the lives of most individuals in a society, and few are avoided so vigorously. Without some form of tax revenue, there is no national unity, and no control. Failure to impose and extract taxes is one of the essential indicators of governmental incapacity to obtain and maintain support. Gabriel Ardant has put the matter well: "The fiscal system is the 'Transformer' of the economic infrastructure into political structure." 12 We were successful in constructing a promising measure of governmental capability to extract resources. This work was greatly aided by the development of a procedure for determining governmental success in the collection of revenue. The index of tax effort permits us to express the degree of governmental effort and performance in tax collection and, although developed for purposes entirely different from our own, seems tailored to our need for a measure of governmental performance in the extraction of resources. The index is obviously more than a device to estimate the performance of the tax system. It is an important measure of the capacity of governments to extract resources, and its variation cross-nationally is a significant measure of political differences. In this role, the index can be used as an indicator of political development. For our purposes, it can be used as multiplier of the other factors that form an overall measure of national capabilities, thus completing the specification of that model. Attempts to estimate tax efforts represented a considerable departure from earlier research directions. In the past,

75

Predicting the Outcomes of International Wars

comparisons of intercountry performance in the collection of revenue consisted of taking simple ratios of collected revenues to total product. Such ratios had little utility for the evaluation of political performance. Without adjustments, they could not be expected to perform better than other economic indicators, serving as proxy measures for political development, had done in the past. If one compares degrees of governmental success in raising tax revenue, one must control for factors that affect the tax bases to the advantage of some countries and to the disadvantage of others. Otherwise, the results obscure precisely the performance one wishes to isolate. Indeed, in extreme cases, one might obtain ludicrously distorted conclusions. The major oilproducing nations provide the best illustration. Saudi Arabia, Venezuela, Kuwait, or Iran have high government revenues because of royalties received from the sale of their crude oil. It would be a major error, however, to give high marks to the governmental performance of these countries in extracting resources from their societies. If one controls for mineral exports, all of those governments show up as doing quite poorly in taxing their populations. The opposite situation obtains for North Vietnam and the People's Republic of China. In these systems, economic conditions are not good because the nations are economically underdeveloped and the tax bases are consequently meager. Thus, the absolute amounts of revenue that the two governments can collect are low by comparison with those of some richer countries, but the political systems are doing all they can, and should be given high marks for the success of their efforts. If we wish to evaluate governmental performance, we must control for the unequal distribution of economic resources across countries. The attempt to evaluate governmental effort in collecting revenue requires an estimate of the taxable capacity of the country. One can formulate the problem in this way: tax ratios depend on the economic resources of the system and on the effort the government makes in extracting these resources. If one is to solve the equation for tax effort, one must know the economic tax capacity of the country. Indeed, the model that Raja Chelliah, Roy Bahl,B and their

76

Chapter Two

colleagues have constructed assumes knowledge of two important elements: an estimate of the amount governments tax on the average, given similar tax bases, and the amount of revenue individual governments actually collect. The comparison between the two enables one to say whether or not a government is effective. It is evident that if the actual and estimated revenues are the same, the nation is operating with normal effectiveness, given its tax base. If actual revenues exceed the estimated level the government could be collecting, it is doing better than could be reasonably predicted. If actual revenues fall below the estimated level, the government is performing less effectively than it should. One can obtain evidence of the level of collected taxes from national accounts. But how is taxable capacity to be estimated? Taxable capacity should reflect the differences in potentially taxable resources available to nations in the system. One must first specify what these factors are thought to be. Second, one must determine how such adjustments are used to derive the estimates of tax capacity for the countries in the system. The developers of this model experimented with a number of factors and constructed a number of equations (see Appendix 1). Some comments are required on the choices of factors used in formulating the equation we selected to measure tax effort and, consequently, the capacity of a government to extract resources from the national society. In the final equation, we used three factors to adjust for differences in tax capacity. One, measuring the openness of the economy, was the fraction of GNP derived from the export sector. Exports-for economic and administrative reasons -were found to offer systematic advantages in the collection of revenue. Exports are more readily taxable than other wealth, not only because their value can be easily estimated but also because it is administratively easy to collect taxes on them, due to the centralization of the channels through which exports flow. A second factor, measuring the level of economic development, was found to require control. We used as an indicator the fraction of the total gross product originating in the agricultural sector, because it reflects the measure of development of the countryside.

77

Predicting the Outcomes of International Wars

This argument will not surprise the reader versed in development problems. It would be startling if this fraction of GNP were not negatively related to the collection of revenue. Resources available in a subsistence farming economy are difficult to evaluate and therefore difficult for governments to tax. Moreover, governments have trouble extracting resources from farmers who can be expected to resist surrendering even a small portion of the very little they have and who generally themselves consume all they produce. A third factor, reflecting the composition oftotal product, was measured by the fraction of gross product resulting from mining activities. The reasons why mineral production should be selected for adjustment have been enumerated in our earlier example of the revenue advantages of oilproducing nations that obtain their governmental revenues largely in the form of royalties from the sale of petroleum products rather than from taxation of the populations. Mineral wealth is a matter of luck rather than productivity. Again, governments can estimate accurately the value of the wealth produced and mineral resources are administratively easy to tax because, usually, only a few large firms are engaged in mining activities. 14 The next step was to use multiple regression that allows controls for the above differences among all members of the sample and produces estimates of how much each factor added to or subtracted from tax totals in each country for every year. The regression yielded predicted values that are estimates of capacity. The final step, the measurement of tax efforts, is obtained by calculating the ratios of real taxes to expected tax capacity. 15 The index can be written as follows: T ax

1'&

ellO

rt

=

Real Tax Ratio . Tax Capacity

Figure 2.1 furnishes an example of an intercountry comparison of tax efforts of all the countries in our sample for the year 1970. In the case of most countries, the graph shows that tax capacity and actual levels of taxation are very similar. However, some countries deviate from the predicted standard. The Israelis and the North Vietnamese,

78

Chapter Two

50

Averoge Tal( Capacity ('Normal')

_ N. VIETNAM

40

1L z

~

_ISRAEL

30

--

!

0

~ ~ larply

122

Chapter Three

restricted. It is preferable, at this stage, to proceed with conflicts where such problems do not arise. The availability of data is a final and indispensable criterion in the selection of wars suitable for the testing of our proposals. Lack of data has been a problem besetting most rigorous studies of international politics. Only very recently has research made available reliable time-series, in a small number of areas, going back to the end of the nineteenth century. These data have been rightly considered providential for the study of the movement of national capabilities. 17 Such series, however, are still very rare; quality and availability drop virtually to the vanishing point when material relevant to the period prior to 1900 is at issue. ls We have mentioned some of our data requirements: the number of casualties, the magnitude of losses of territory and/or of population groups. Even more important, however, are the data from which to derive indicators of national capabilities. We have reiterated in the previous section the limited availability of such data. These stringent theoretical requirements have reduced to a very small number the sample of wars suitable for the purposes of our study. Although the Napoleonic Wars and World Wars I and II meet the theoretical criteria we have established, only for the latter pair have we time-series data at frequent enough intervals and of high enough quality to permit the analyses we wished to undertake. Here, therefore, we must raise the question of whether or not the analysis of two wars can be more generally representative than the examination simply of two case studies, with all the limitations this approach implies for the drawing of broad inferences beyond the cases themselves. Can one justify generalizations based on these two instances as being applicable to wars not reviewed here? If not, the value of the findings of this study is greatly diminished. We think that the results of this study should be considered sources of significant generalization as far as the effects of major war on power distribution are concerned. Although the number of wars we analyze is extremely small, the sample of thirty-one cases we observe in these conflicts is not small. We shall discuss more fully in the next section the

123

The Costs of Major Wars

cases we observe throughout the two most terrible wars in history, which include most of the countries of the central international system. Moreover, the two great wars we examine are those which offer the only opportunity for testing our propositions. The reduction of our area of study to two wars has undoubtedly sharply restricted the nature of the inferences that one might wish to make. On the other hand, our findings here certainly demonstrate that the hypotheses are solid and should be tested on a much enlarged sample as soon as improved theory, methodology, and data permit it. Actors Some would argue that in the study of war one does not learn much of value by attending only to the behavior of nations, and that other levels or kinds of analysis are more appropriate. For certain types of research on war, this is true. In some cases, one might justifiably prefer to concentrate on units of analysis other than nations. Were one interested in questions related to causes and outcomes, particularly the former, it might be meet to observe the views and behavior of military leaders, politicians, diplomats, industrial and labor officials, and mass publics. However, if one is interested in the effect of victory and/or defeat in war on the distribution of international power in postwar periods, this is not the case. Here, decision-making and the behavior of individuals play an inconsequential role; nations are the only actors. Casting nations as the actors uniquely relevant poses problems. In order to compare the performance of a country before and after war, one must be certain that one is comparing the behavior ofthe same entity over time; otherwise, differences in behavior would legitimately be ascribed to the fact that the entity itself is different. The problem inherent in this kind of comparison is that nations are units not so stable, over time, as one could wish. Indeed, there is considerable debate about the nature of vital signs in the life of a state. Ideally, one would expect, as a minimum: a fixed territory and population; governmental sovereignty over

124

Chapter Three

both; the formal diplomatic recognition of most of the international community. Reality is different from the ideal. Nations shrink and expand, established countries disappear, and new ones appear in their places. Besides, some nations delegate control over their foreign affairs to other powers. Some are administered, at least in part, by international bodies, while still others make a token effort to supervise populations and territories that are within their legal jurisdiction but are in fact regulated by rebel groups prompted by aUthorities operating abroad. 19 Increasingly, this reality has been recognized;20 the most exhaustive list of nations compiled to date has been established on the premise that not all of the essential elements-sovereignty, territorial continuity, and international recognition-need to be present all the time. 21 In seeking the cast of nations whose performance we should watch and measure, it was clear to us that it should incorporate all belligerents in World Wars I and II, with a group of nonbelligerents serving as a control group. The need for such a control group is obvious: we must be able to show that the behavior of belligerents is peculiar to them, not shared by nonbelligerents. In casting the combatants, we began with the comprehensive list of belligerents compiled by Singer and Small in their classic data-gathering effort, The Wages of War, to which we added, as an initial massing for a control group, all of the nations in existence at th~ time of the conflict under study that had not participated in it. 22 The first list had to be pared. We immediately eliminated from our sample all nations that were not fully developed. There was no way to estimate satisfactorily the national capabilities of underdeveloped or developing countries. 23 A number of nations listed as combatants by Singer and Small in one or the other of the two wars had to be dropped from consideration for this reason. This category is, unfortunately, rather large, including Mongolia, Bulgaria, Poland, Rumania, Ethiopia, Greece, Brazil, Turkey, and Finland. Some of these nations, had they survived the first culling, would have been deleted for another reason: lack of relevant data. South Africa and New Zealand were elimi-

125

The Costs of Major Wars

nated from the study because data concerning their capabilities are inadequate for the kind of analysis we sought. We also had to take into account the fact that South Mrica, New Zealand, Australia, Norway, Germany, Japan, and Austria were not totally free and independent for part of the period important to this study-though none of them was eliminated from consideration solely on this ground. For instance, Australia, New Zealand, and South Africa had not achieved dominion status, and Norway had not been separated from Sweden. De jure, therefore, they were not free. But since they were free in fact and in action, disregarding their legal status, we treated them in the end as free and independent units. Germany and Japan were not free throughout the period under analysis because of their occupation after World War II, and Austria was annexed by Germany before the onset of that war. For our purposes, this inhibition could be devastating. Loss of freedom of action was not important if it occurred during a war, but a nation under the control of one or more other nations either before or afterwards might lack independence in the allocation of its resources; this, in tum, could be a factor influencing its prewar pattern of growth or its recovery period. As we decided that there was no reasonable way to establish control for the possible data-distortions due to foreign occupation, we determined simply to consider as continuous units all nations which recovered their identities after occupation. Germany and Japan were included without our giving weight to their occupation periods. However, the dismemberment of the Austro-Hungarian empire in 1919 made necessary the elimination of Austria. 24 Another reason for eliminating Austria from our study was the serious prewar disruption created in that country's economy, and thus reflected in its national data. The annexation by Germany was merely the last blow to a nation already traumatized. We could find no satisfactory way of dealing with the resulting distortions. Also eliminated was Spain, a neutral in both wars. In the period between the two great wars, Spain suffered a revolution and civil war which laid waste to the land and its people. The effect on the

126

Chapter Three

Spanish system was disastrous, and rendered the estimation of its national capabilities so difficult, for the critical interwar period, that deletion appeared the only course. A similar but less aggravated problem arose with other countries and ought to be mentioned here, although, as noted earlier, they were retained for consideration. There were terrible difficulties with the Soviet Union and with Russian data generally, but we managed to retain that country. The United States and a number of other industrialized countries, central to our research, were seriously affected by the Great Depression, which made a shambles of all sectors of national life that contribute to the formation of power resources. These economic dislocations were sufficiently severe to create major difficulties in our attempt to establish estimates of normal levels of national capabilities during the interwar period, and therefore affected our measurements of what those capabilities should have been had no war occurred. The Analytic Groups Having assembled our dramatis personae, we must now arrange the countries into groups central to our analysis. We must divide our cast into belligerents and nonbelligerents, winners and losers, active and occupied. First, a distinction must be made between nations which were belligerent and those which were merely hangers-on, always present in major coalitions. This determination is not always easy to make. Argentina, for example, declared war on the Axis in the waning days of World War II. She never dispatched a soldier to the war area, and for most of the war her sympathies were clearly with the opposite side. Was Argentina a belligerent? Brazil is another case in point. She sent only a miniscule force to fight, and her principal contribution was permission to use an air base for refueling planes destined for Mrica. The Soviet Union entered the war against Japan in the last months of World War II, after German surrender had become a certainty. Soviet participation in that aspect of the conflict, therefore, was for the purposes of bootygathering, to stake out a new sphere of influence for the postwar period. How is one to separate the belligerents from

127

The Costs of Major Wars

the countries whose behavior ranged from total neutrality to one of merely symbolic participation? Much the same kind of difficulty presents itself within the category of winners. The position of France in World War II is an example of this problem. For much of the war, France was wholly occupied by Germany. Yet France began and ended that war on the side of the winning coalition. Was France a winner, in the same sense as England, the Soviet Union, or the United States can be thought winners? One would be forced to reject such a notion. The same distinction separates losers. Italy began the war on the side of the Axis, switched allegiance in the middle, and ended as a member of the Allied forces. Is Italy to be construed a winner or a loser? Operational definitions must clarify ambiguities that arise from the formulation of theoretical conceptions. Distinctions require definition: of belligerents and nonbelligerents, of active and occupied nations, of winners and losers. Some definitions follow. 1. Belligerent nations are those whose participation in the conflict resulted in military losses of at least five thousand troops. 2. Nations whose strict neutrality or symbolic participation resulted in losses of less than five thousand troops were considered nonbelligerents. Moreover, four further theoretical and operational distinctions were used to categorize belligerents. 3. Belligerents still fighting in the final third of the conflict on the same side as at its beginning were classified as active belligerents. 4. Belligerents that, in the last third of the conflict, were (because of occupation) no longer members of the coalition they had joined at the beginning were classified as occupied belligerents. Thus, France, England, and Italy were all active belligerents in World War I; Italy and France were occupied belligerents in World War II. There were, additionally, problems of providing operational definitions for winners and losers, so that one could resolve uncertainties at the theoretical level. These definitions follow.

128

Chapter Three

5. Nations that retained all their territories or extended them, immediately after a conflict and as a direct result of that conflict, were considered active belligerent winners. The rational support for such a definition is plain: no victor in a major war would tolerate the loss of any territory under its jurisdiction. 6. The reverse of this proposition was used to define a loser: loss of territory would be construed as the attribute of a loser. Even transfer of territory, with full compensation made for that loss, should be considered an overt sign of defeat; for no victor would submit to such terms, and only a loser would have no alternative. We have coded such countries as active belligerent losers. 25 7. An additional category should be mentioned: Nations that lost territories after a conflict in the course of which they had been occupied were coded as occupied belligerent losers. Instances are Poland and Czechoslovakia, both losing territory for which they were compensated at German expense. This grouping is almost unused in our analysis because data in support of it are wanting. The analytical groupings that result from different combinations of our three fundamental dichotomies (belligerent! nonbelligerent, active/occupied, and winner/loser) do not exhaust all the logical combinations one could explore. They do, however, satisfy the theoretically interesting possibilities for the analysis of war, and can be represented simply in a Venn diagram (fig. 3.3).26 We are also interested in an additional distinction that crosses all the analytic groups we have established. We should like to explore the consequences of war for major powers as well as for the entire group of nations. We should also like to analyze, for each of the two wars, the consequences, individually and taken together. The reason for this attempt to dis aggregate is that our sample of cases is so small that it is imperative to establish whether or not the patterns of behavior we seem to find are simply a by-product of our aggregation, and, as such, are not present when disaggregation is effected. Two points need emphasis, however. We do not wish to dis aggregate at the level of individual nations. Second and equally important, it should be clear

129

The Costs of Major Wars

that the results of disaggregation are not independent findings, for they are based on the same sample. 27

NON - BELLIGERENTS

o

o

SET OF ALL NATIONS I \L) SET OF ALL BELLIGERENT NATIONS I B )

[(\{) SET

~ Fig. 3.3

OF ALL OCCUPIED

NATIONS

(0)

SET OF ALL LOSING NATIONS IT)

Venn representation of analytical groups.

The distribution of our total cast of nations into analytic groups discussed above can be seen in table 3.2. The sample of eighteen nations and thirty-one cases thus composed was the very best we could assemble given our theoretical constraints and the lack of data. It is scant, but sufficient for the research we wished to undertake.

130

Chapter Three

Table 3.2

Final Sample of Belligerents and Non-Belligerents in World Wars I and II World War I *United States *United Kingdom *Italy *France Australia Canada

World War II ABW ABW ABW ABW ABW ABW

*United States *United Kingdom *USSR Canada Australia

ABW ABW ABW ABW ABW Total: 5

Total: 6 *Germany (West) *USSR

ABL ABL Total: 2

*Japan *West Germany *Italy Hungary

ABL ABL ABL ABL Total: 4

*Japan Denmark Sweden Netherlands Norway

NB NB NB NB NB

*France Belgium Denmark Netherlands Norway Yugoslavia

OBW OBW OBW OBW OBW OBW

Total: 5 Total: 6 Czechoslovakia

Total: I

* = Major Powers ABW=Active Belligerent Winner ABL=Active Belligerent Loser OBW=Occupied Belligerent Winner OBL=Occupied Belligerent Loser NB = N on-Belligerent

OBL

Sweden Switzerland

NB NB Total: 2

Empirical Propositions

We are now ready to reduce the theoretical propositions we presented at the beginning of this chapter into empirical propositions phrased in the precise operational language of this study. We have eight hypotheses: the first four represent our short-range expectations, the second four our long-range expectations.

131

The Costs of Major Wars

Our propositions for short-range effects are: 1. Belligerent countries that emerge as winners of major wars gain in power; nonbelligerents retain their antebellum power patterns; belligerent losers suffer substantial power losses. 2. In the period immediately following a war, belligerent winners and nonbelligerents retain antebellum power patterns; belligerent losers suffer substantial power losses. 3. After major wars, all belligerent countries suffer major losses in their power capabilities; nonbelligerents retain antebellum power patterns. Our null hypothesis is stated thus: 4. After major wars, the power patterns of belligerents and nonbelligerents are not affected in a systematic manner. Our propositions for long-range effects are: 5. All groups involved in war suffer long-range losses of power and do not regain power patterns at levels established before the onset of major conflict. 6. Belligerent winners and nonbelligerents retain antebellum levels of growth; power cleavages between belligerent winners and nonbelligerents on the one hand and belligerent losers on the other are rapidly erased by the greater acceleration of recovery rate evidenced by losers. 7. Differences which result from victories and defeats in major wars are maintained or even slowly increased; thus, the immediate postwar gap between belligerent winners and nonbelligerents on the one hand and losers on the other is maintained or possibly enlarged. Our null hypothesis for long-range expectations is: 8. The postwar patterns of all groups considered are not affected in a systematic manner as a result of war. Two final points. First, it should be clear that in order to test our propositions about long-range effects, the null hypothesis for short-range expectations must be disproved. Second, the two sets of hypotheses, as well as the theoretical propositions that precede them, do not represent all the logical possibilities but do summarize the views expressed in the literature regarding the outcomes of war and the effects of these outcomes on power distribution.

132

Chapter Three

Findings

We first wish to examine the behavior of the entire sample for both wars; then the behavior of the subset of great powers for both wars; and then the behavior of the entire sample and of the great powers for each war taken separately. Had our sample of countries and wars been larger, the performance of all countries taken together would have offered evidence of the kind of behavior we could expect from winners, losers, and neutrals as a consequence of even a major war. However, due to the smallness of the sample, it seemed wise to explore, as deeply and broadly as possible, how important subsets of the total sample behave. It is important to note that while each of the partitions of the total sample is separately and discretely observed, they are not independent, since the sample is identical for all. Our first partition permits us to see the behavior of all nations in both wars. Only three of our analytic groups (active belligerent winners, active belligerent losers, and nonbelligerents) are considered. Occupied belligerent winners and losers are excluded because, by our definition, there were no occupied countries during World War I. If we wished to test the behavior of the system in both wars taken together, we had thus to eliminate such categories in this partition. The results (fig. 3.4) show that in the first two postwar years belligerent winners and nonbelligerents lose from 1.5 to 3.5 years. The deviation of both groups from expected performance is minimal. Active belligerent losers, however, suffer losses in a comparable period of 20.4 to 21.6 years. The difference between the two groups of close to nineteen years is substantial. We must now consider the long-term implications of the outcomes of war on the power distribution among the three analytic groups. The nonbelligerent nations retain growth rates characteristic of their prewar performances. Active winners incur an average deviation of about three years from the zero line, indicating that losses suffered after postwar demobilization are maintained. Among active belligerent losers, however, the Phoenix phenomenon manifests itself. Losers begin and maintain a steadily accelerating re-

133

The Costs of Major Wars

covery rate after the war, and overtake the winners in the eighteenth year of the postwar period. At the conclusion of that period, differences in power distribution among all groups have been eradicated; their levels of power return to points one would have anticipated had no war occurred. Thus, the results we obtained strongly support proposition 2 (in the short run, major power differences will occur between groups) and proposition 6 (in the long run, these differences are eliminated). LEGEND

10 -

ACTIVE BELLIGERENTWINNERS

- - - - ACTIVE BELLI~ENT LOSERS ••••••• NON- BELUGERENTS

END OF RECOVERY PERIOD

-5

-10

-15

~

-20

- 25

END OF WARS

20 Recovery period (years)

Fig. 3.4

Partition I: general consequences. Sample: all nations, both wars.

134

Chapter Three

In testing the second subset of actors in our breakdown, the behavior of the great-power system over two wars, the results are very much the same in their general outlines (see fig. 3.5). Only active belligerent winners and losers are considered, for none of the great powers was neutral in World War II. Active winners begin with a two-year loss immediately after the war, sliding away from expected normal growth; then they recover briefly, only to slide away once more. Overall, their performance is below expected levels and is heavily influenced by depressions in the interwar LEGEND

10 -

ACTIVE BELLIGERENT WINNERS

- - - - ACT I VE BELLIGERENT LOSERS

o

END OF RECOVERY PERIOD

"NORMAL" GROWTH PROJECTION

-I~

~ -20

S

- 25

END OF WARS

5

10

15

Recovery period (years)

Fig. 3.5

Partition II: major-power consequences. Sample: major powers, both wars.

135

The Costs of Major Wars

period. However, the trajectory as a whole reveals no loss in relation to prewar capabilities. The active belligerent losers, on the other hand, suffer a twenty-year loss immediately after the war, but recover rapidly, overtaking the winners in the fifteenth year. After this point, both groups resume previous patterns. Propositions 2 and 6 are confirmed again by a study of the composite data of the great powers for both wars. In our next four operations, we partitioned our sample to show the performance of the entire system and the subsystem of great powers for each world war. Our third partition permits us to observe the behavior of the entire system after World War I (see fig. 3.6). In the short run, nonbelligerents appear slightly affected during the first year after the war but remain within one year of the zero line during the second. The active belligerent winners incur losses of from five to seven years right after the conflict. Active belligerent losers suffer losses of from twentyone to twenty-five years. Winners lose, but losers suffer four times more severely. There can be no doubt that, in the short run, there are serious power consequences both to winners and losers of major wars. Again, the evidence in large part supports proposition 2, although active belligerent winners do suffer markedly. The evaluation of long-term consequences is intimately related to the effects of economic depression. Consider first the initial twelve-year period after World War I. The characteristics described in proposition 6 are supported. Active belligerent winners recover very slowly from war effects; nonbelligerents retain previous growth patterns; active belligerent losers, after the immediate postwar period of heavy loss, display a substantially faster rate of recovery than winners. Then the Great Depression strikes and, within two years, performance of all groups are diminished. Belligerents, however, seem to suffer more than nonbelligerents. After 1933, predepression trends reestablish themselves, but the subsequent period is too short for adequate evaluation.

136

Chapter Three

LEGEND

10 -

ACTIVE BELLIGERENT WINNERS

----ACTiVE BELLIGERENT LOSERS ........ NON- BELLIGERENTS

END OF RECOVERY PERIOD

"NORMAL" GROWTH PROJECTION

o

......

......... -Ci

-10

-115

-20

-215

Recovery period (years)

Fig. 3.6

Partition III: World War I consequences. Sample: all nations, World War I.

The great-power system during World War I behaves in much the same way described for the sample as a whole (see fig. 3.7). Active belligerent winners fare slightly worse than the whole sample, but differences are so marginal that we feel secure in concluding again that the results support propositions 2 and 6. We draw this conclusion despite the disruptions caused by the depression. The figure shows the depression to be a major factor in distorting the recovery patterns of all countries taken together, not merely for

137

The Costs of Major Wars

LEGEND

10 -

ACTIVE BELLIGERENT WINNERS

- - ACTIVE BELLIGERENT LOSERS ::

~

c

"ij

15

(!)

'.

•••••••• NON- BELLIGERENTS

END OF RECOVERY PERIOD

0

"C C

~ iii

-15

~

"au; E

lii

£ ~ ~

-10

t:

"c""

-Ie

~ -20

.3

-215

1918 1920

19215

1930

19315

Recovery period (years)

Fig. 3.7

Partition IV: World War I major-power consequences. Sample: major powers, World War I.

World War I but for both conflicts. Moreover, the depression is also a major reason why our projections for the period following World War II are so weak, underestimating egregiously the growth trends (see fig. 3.8). In our fifth partition, all major analytic groups are represented for World War II. Some of these representations, however, are so tentative that one needs to take the information they convey with caution. The consequences of World War II on power distribution,

138

Chapter Three

in the immediate aftermath of the war, are much as expected. In the first year, active belligerent winners are slightly ahead, but move toward the zero line in the second year. Nonbelligerents are slightly below the zero line but move to points within tolerance limits (two years, for this partition) in the second year. Active belligerent losers suffer substantial losses, between sixteen and seventeen years, in the first twenty-four months after the war, and the occupied belligerent winners lose from eleven to fourteen years in the 10

LEGEND ACTIVE BELLIGERENT WINNERS - - - - ACTIVE BELLIGERENT LOSERS - - OCCUPIED BELLIGERENT WINN ERS OCCUPIED BELLIGERENT ~~::",//:~i:l LOSERS

o

-15

-20

-25 1945

1950

1955

1960

Recovery period (years)

Fig. 3.8

Partition V: World War II consequences. Sample: all nations, World War II.

139

The Costs of Major Wars

same period. The differences ofloss between active winners and active losers range from twenty-one to twenty-three years in the initial period. In the long-run analysis we find, for the first time, evidence to support the proposition that the gap between winners and losers can continue instead of closing. 28 A number of points should be made. The logarithmic projections seriously underestimate the growth of the system. An indication of this is to be seen in the fact that nonbelligerents post increasing gains over time, indicating that the prewar patterns, distorted by the depression, are not a good indicator of the behavior of the group as a whole in the period after World War II. In some part, however, as Kuznets suggests, economic growth since 1945 is due to a liberalization of trade in the industrial world and is therefore unexpected. 29 It is our impression that a more accurate projection of growth trends would place the zero line approximately where one finds the trajectory for the nonbelligerents. In any event, our main concern is with active winners and losers, and we must first establish that distortions in the pattern do not, in that respect, affect relative calculations. Over the entire period, the active belligerent winners maintain a constant but slight edge over nonbelligerents of about three years. Absolute differences between winners and losers are not distorted by the acceleration of recovery rates. One should note that here, too, active losers enjoy a sharply accelerated recovery pattern and regain the prewar level of growth within the stipulated period. They do not actually close the gap between themselves and the winners for two reasons: the acceleration of the entire system; the last five years decided deceleration in recovery, which may cause an absolute loss of from five to eight years. Nevertheless, a gap of roughly nine years remains at the end of the recovery period, and one might argue that we have, as a result, evidence supporting proposition 7, that over the long run winners keep the advantage they gain from victory. The remaining analytic groups behave in interesting ways. The performance of occupied belligerent winners is somewhere between that of active winners and losers and follows

140

Chapter Three

closely the performance of the former. Occupied belligerent winners regain prewar rates in fifteen years, surpass them, and come close to convergence with winners at the end of the recovery period. Since only Czechoslovakia falls into the category of occupied belligerent losers, we obviously cannot refer to "findings" in observing the behavior of one country. But we should note that the matter will merit investigation when sufficient data become available; for if other occupied belligerent losers behave as Czechoslovakia does, we may have identified the real losers in major wars-the nations that do not recover. It is possible that we have also identified the conditions necessary to support an alternative hypothesis, different from those advanced here. This may be significant. In all the analyses so far offered, we have discerned only marginal differences in the long-range consequences of victory and defeat on the power distribution of the system. However, if the case of Czechoslovakia (in the period for which we have data) is a true indication of what obtains for other occupied belligerent losers, it would then be clear that, had the victors insisted on occupation, exploitation, and repression of defeated populations, our findings would be dramatically different. It may be that victors can delay the recovery of the vanquished by occupation and repression. For example, if Hitler, with his plans to depopulate and exploit his victims, had won the war, the vanquished might not have recovered. Had Hitler been victorious, proposition 5 would have been sustained. But Hitler could not have won. Be that as it may, the results of the fifth partition should not be viewed separately but rather must be compared with those of partition six (see fig. 3.9), for the deviations disappear when one observes only the subsystem of great powers in World War II, which behaves entirely in consonance with the expectations of propositions 2 and 6. In the short term, active belligerent winners suffer no loss, and occupied belligerent winners and active belligerent losers suffer a loss of twenty years of "normal" growth in the first two years after the war's end. The relative differences between active winners and the latter two categories in this period are 20.9

141

The Costs of Major Wars

10

LEGEND ACTIVE BELLIGERENT WINNERS - - - - ACTIVE BELLIGERENT LOSERS - - - - - OCCUPIED BELLIGERENT WINNERS

~

~

o

~~------~--~~~

END OF RECOVERY PERIOD

-!!

-10

-I!!

-20

-2!!

END OF WORLD WAR II

194!!

19!!0

19!!!!

1960

1964

Recovery period (years)

Fig. 3.9

Partition VI: World War II major-power consequences. Sample: major powers, World War II.

and 18.1 years respectively. Evidence on long-term effects fully supports proposition 6. Active winners maintain expected growth patterns, while active losers accelerate their recovery rates and overtake winners in the sixteenth postwar'year. How is one to reconcile these results with those obtained from an examination of the behavior of the total sample? Some explanatory points should be considered. The differences noted across the two partitions are not rooted in

142

Chapter Three

the behavior of all our analytic groups. The behavior of the active losers and occupied winners changes only slightly from one partition to the other. It is also plain that the logarithmic model we chose projected accurately the performance of great powers in the period after World War II. Not projected accurately, therefore, are the behaviors ofthe smaller winners. Their unprecedentedly rapid growth inflates the performance of the entire sample. The Phoenix Factor Most unexpected and interesting is the discovery that, after wars, the active losers catch up with winners in comparatively short order, and that the system of international power begins to behave as one would have anticipated had no war occurred. We cannot explain the phenomenon; we do not know why losers rise from the ashes as they appear to do. We can, however, make some surmises. It is plausible to believe that structural elements playa part. For example, favorable occupational distributions may help to accelerate recovery rates, as may the destruction of obsolescent plants and industrial equipment. It is probable that attitudinal factors also play a significant role in increasing the pace of recovery. A defeated but economically developed population, living in the midst of destruction, will recall the status quo antebellum and be motivated to rebuild. A populace motivated in this way would have the technology to make an economic system function well. It is also plausible to assume that the defeated population would exert a greater effort to recover than would the population of a victorious country-the latter more intent on enjoying the spoils of war. The necessity for work and sacrifice is evident to all members of a vanquished society. Charles Tilly found, for instance, substantially fewer strikes in Italy and Germany for a time after World War II than in England and France. 3o These reasons are credible, but we have no assurance that any of them is accurately to be judged responsible for hastening the recovery rates of defeated nations. We do know, however, that one aspect, widely believed to be influential, had little if any effect. Losers do not rise from the ashes

143

The Costs of Major Wars

because winners pick them up and help them to their feet. Were this true, it would completely overturn the results we have obtained. For if losers could not recoup their losses without aid, the gap between them and winners would remain if the active winner refused this kind of assistance. This would support the continuing-gap proposition, one of the possibilities previously hypothesized. The point was of major concern, and we tested it in the form of two propositions. First, we were interested to discover whether or not aid was positively associated with the recovery of the recipient and, second, whether or not a large-scale foreignassistance effort does start recovery on its way, as some economists believe. Our test was simple. Since the United States was the source of such aid after World War II, and since such help was dispensed annually between 1948 and 1961, we compared the amounts of aid given by the United States, in totals and per capita, with the relative growth rates of the recipients.3 t Had there been a direct relationship between aid and recovery, and if one controlled for population, growth rates would show increases as a result of aid. Had aid intensity been a factor one would also expect that growth rates would show strong gains after those years when recipients received particularly large gifts. 32 Because we are dealing with time-series data we repeated each evaluation controlling for the possible linear, delaying influences of time. 33 From a consideration of the figures in table 3.3, we can determine only the weakest association between external aid and recovery. The variance explained by the coefficient Table 3.3

Correlation of u.s. Aid with Recovery of Recipients, 1948--61

Aid Years 1948--61 Recovery Rate

Aid - .01 = 63)

(N

- .33 = 23)

(N

(N

1948--53

Aid Partialed on Time

(N

Aid per Capita

Aid per Capita Partialed on Time

- .01 = 63)

(N

-.18 = 23)

(N

-.17 = 63)

(N

-.21 = 23)

(N

-.15 = 63)

- .30 = 23)

144

Chapter Three

of determination is always below .1, indicating that growth and foreign aid, totals or per capita, are almost wholly independent of each other. Such relationship as may exist is negative: the countries that received most of the aid for the longest period performed worst. The United Kingdom received much more aid, on a total and per capita basis, than France; France received much more than Italy, Italy much more than Germany, and Germany much more than Japan. Yet it was Japan that enjoyed the most rapid rate of recovery, followed by Germany, Italy, and France, with the United Kingdom bringing up the rear. It is, therefore, very hard to credit the conviction that foreign assistance and recovery are closely associated. These particular findings are not completely unexpected. 34 Many economists have questioned the efficacy of this kind of aid. What the figures underscore is that foreign assistance, as a form of investment in the economy of another country intended to incline it toward faster recovery, is not very effective. The variables truly important to recovery lie within the devastated nations themselves. Previous patterns of performance are far more significant than external aid. Conclusions We began this inquiry with a number of questions. Do the outcomes of major wars reshape the distribution of international power? Does it make a real difference-in power terms-whether a country wins or loses a major war? How long can winners hold on to their advantages? How long do losers stay behind? Let us begin with a simple list of what we have found. 1. Systematic patterns in the distribution of power (as measured by gross national product) are registered after major conflicts. 2. The power levels of winners and neutrals are affected only marginally by the conflict. 3. Nations defeated in war suffer intense short-term losses; the outcomes make much difference to them in the short run, especially in terms of power levels.

145

The Costs of Major Wars

4. In the long run (from fifteen to twenty years), the effects of war are dissipated, because losers accelerate their recovery and resume antebellum rates. They may even overtake winners. Soon, the power distribution in the system returns to levels anticipated had wars not occurred. We have evidence that this happens and we can speculate about the explanation, but we have no definitive solution. There is substantial research remaining to be done. If one wishes to forecast the behavior of a country fifteen or twenty years after the end of a war, one should not refer to the outcome of that war, whether the country in question participated in it or not, whether it was a winner or a loser. The best indicator of the power posture of a nation less than a generation after the conclusion of a war is its performance before that conflict. One other finding should be mentioned. It is clear that the assistance offered by winners to losers is not a significant factor in the losers' recovery rate. We are tempted to suggest that the outcome of war, insofar as international power is concerned, makes no difference. We cannot forget, however, that we have found traces which indicate that we may have heard only part of the story. Although winners cannot help losers to recovery by contributing aid, whatever the quantity, winners may be able to prevent or delay the recovery of losers. If the behavior of Czechoslovakia is an accurate indication, the victor may retard the recovery of the vanquished by occupation and exploitation. Adequate information on this aspect of the question is not available; there are merely hints on which to base speculation. Such findings are clearly tentative. In this case, however, the tentative nature of the conclusions should be stressed once more because we have been plagued with data problems. If our findings are confirmed through comparable and more exhaustive researches, the implications for research strategies could be substantial. If the distribution of international power, and changes in that distribution, are shaped by differential rates of growth across critical sectors of a nation's life and across the nations of that system, and if such rates cannot be altered even

146

Chapter Three

by the most violent forms of international interactions, such as major wars, then what must be studied are the causes of such alterations and not the interaction of countries. Thus, the origins of the independent variables in evidence in international relations are not found in international relations but in the growth of the units that constitute the system. Some scholars already study international politics in this fashion, but this is still a very different conception of the field from that traditionally held.

Four

147

Nuclear Arms Races and Deterrence

A reader may well be tempted to note that the arguments and concepts put forward in the preceding pages may seem in some measure obsolete, valid only in the context of an international order that no longer exists, one that was governed by laws that no longer apply. In the nuclear age are the forces that lead to war, to victory or defeat, and that generate recovery, still wholly resistant to manipulation and redirection? What about nuclear weapons? Have they not entirely changed the rules of international conflict? President Kennedy thought so; he once stated that the possession of nuclear weapons by the United States and the Soviet Union "Changes the problem .... It changes all the answers and all the questions." 1 But is this really true? One point should be made at the outset. What has emerged in our previous chapters is that major wars among the great and superpowers were not so much the result of conflicting goals and policies of the different countries as they were the consequence of the patterns of economic and political development and demographic change of the actors in the system, and particularly those of the superpowers. Armament programs were also, in large part, the consequence of such internal processes. Nations bought arms because, as they developed, they became richer and obtained additional resources that could be dedicated to defense. In tum, of course, the acquisition of arms accelerated the development of the country. The connection between development and military strength should be kept in mind. We have talked of war throughout this book but we have not talked directly of military power. We have had no need to. We have considered the connection between politicoeconomic power on the one hand and military strength on the other to be so close

148

Chapter Four

that one could legitimately infer one from the other. If the overall scores on the socioeconomic and political variables were found to be high, one could assume in the vast majority of cases that scores on an index of military power would be high as well. It is worth restating here the reasons why this should be so. As a rule, after the industrial revolution, the correlation between industrial and military power was very close. The destructive capability of individual weapons was quite small, and the upper limits of the size of armament and armed forces were set by productivity of the economy and the availability of manpower to work and fight. No nation could overcome any serious disadvantage due to levels of development and population size simply by arming. No nation was, in fact, much stronger than her economic productivity indicated. Before the advent of nuclear weapons, even the most determined armament policies could improve the overall power position of any nation only at the margins. Such policies could only make a difference in cases where rivals were not too different in demographic terms, economic productivity, or political effectiveness. This, for example, was often the case with the major European powers. Hence the legitimacy of inferring from the direct measure of politicoeconomic power to military strength. The international system rested on this connection. But it is precisely this connection which is claimed to have been severely weakened, if not completely severed, by the advent of nuclear weapons. As a result, the rules presiding over the life of the system seem to have been completely turned around. What counts today is no longer the social, economic, and political capacity that has always made the difference before. What counts now is pure military strength, specifically nuclear strength. In this new view of the world, military power appears finally freed from the constraints imposed on it in the past by the demographic base and by the productive and political capacity. What consequences are believed to flow from this new state of affairs? In summary two claims are made. 1. It has been contended that these new and terrible weapons would be largely effective without being used. The

149

Nuclear Arms Races and Deterrence

threat of their use would suffice to frighten other nations away from behavior unacceptable to the nuclear power. In short, the function of nuclear weapons is held to be deterrent. 2 Fundamentally, the peace of the world rests on a rough balance in the distribution of nuclear strength. 2. The new weapons are so powerful, it is believed, that technological or quantitative breakthroughs can make an appreciable difference in the relative distribution of nuclear strength. It is useless to argue that the distribution of nuclear weapons can fluctuate a good deal so long as both sides retain second-strike capabilities (the capability of striking back even after absorbing a nuclear attack by the enemy). The safety of the second-strike force of a nuclear power is at the mercy ofthe strength and accuracy of the aggressor. For a nation to fall behind in nuclear weapons might be disastrous for her own security and, since the peace of the world is connected with the balance of nuclear forces, for international peace as welI.3 For a nuclear power there is no other course but to pay very careful attention to what other nuclear powers are doing, in order not to be caught at a disadvantage. Competition is the rule of life for a nation in the nuclear age. It is clear that if such claims are even largely true much of what we have written earlier is open to revision in any explanation of war in the present era. It is therefore, a major task of this chapter to explore these claims. Our strategy will be to test the above propositions by looking systematically at the behavior of the participants in every case where nuclear confrontation has taken place (i.e., where at least one of the contestants has had nuclear weapons), and trying to see whether the behavior of the participants would have been any different had nuclear weapons not been in the picture. Where both contestants had nuclear weapons, our strategy will be to explore whether the contestants competed with each other to forestall falling behind. The race by competitors to stockpile weapons is the essential link in nuclear deterrence theory between fear and deterrence; this in turn is important evidence that armaments programs are the result of external danger. If, on the other hand, the reasons for armaments programs are not related to

150

Chapter Four

external threats but are, rather, the result of internal pressures, then the development and production of nuclear arms will not be determined by interaction between competing nations. In such a case one should also find that decisions to invest more and more resources in the manufacture and purchase of arms are made, as in other programs, incrementally. This explanation of the building and stockpiling of arms is, incidentally, more in accord with our previous findings in this book. And were the data to demonstrate that such explanations are, in fact, correct, they would gain even more force in the special context of the nuclear age. We shall examine first the whole conception of nuclear deterrence. We shall then examine whether, in fact, as one would expect if one accepts the theory of deterrence, the major nuclear powers are shaping their arsenals in direct response to what their adversaries are doing. Finally, we shall present some ideas about the most vital of questions: granted that nuclear war has not occurred, under what conditions would nuclear weapons be used? Deterrents and Deterrence The concept of nuclear deterrence is quite simple. According to it, the presence of nuclear weapons so terrifies any potential victim that the latter would never dare attack or cross one armed with such weapons. The model of deterrence (see fig. 4.1) contains two distinct expectations: the presence of nuclear weapons will inspire great fear in those threatened with nuclear attack; and terror of nuclear punishment will provide a sufficient corrective to deflect the aggressive behavior of a potential aggressor. In our view the

Nuclear weapons: deterrents

Fig. 4.1

-..

The process of deterrence.

Credible threats: terror

-..

Change in behavior: deterrence

151

Nuclear Arms Races and Deterrence

first expectation is essentially correct, but the second may not be. Nuclear weapons are, indeed, incredibly destructive and they do inspire terror in potential victims. But does this terror have the effect the model of deterrence alleges it to have? Leaders ofnatlons menaced by nuclear retaliation are doubtlessly terrified, yet do they show evidence of having acquired that respect which would prevent them from doing battle with those very countries alleged to have been made inviolable by their possession of nuclear arms? And if the second expectation is false, then the concept of deterrence does not hold true. True or false, however, deterrence holds such an important place in current explanations of the behavior of great powers in their conflicts with one another that it deserves careful consideration. There are three components in the conception of deterrence: the terrible destructiveness, the incredible speed, and the long reach of nuclear weapons; the fear that such weapons inspire; and the claim that nations, once threatened with nuclear destruction, will abandon their aggressive moves. We are not really vitally concerned with the first two. The facts, in each case, have been discussed many times and we are ready to stipulate them as true. We shall only briefly recall them and no more. We are vitally concerned with the validity of the third component. There are major flaws in the way in which, according to the theory of deterrence, the three components interact to produce deterrence. In deterrence theory it is terror that deters but nuclear weapons that are the source of that terror. Appropriately, therefore, nuclear weapons are called "deterrents." The development of nuclear arms is the easiest portion of the story we have to tell, for here we have hard data in abundance. Indeed, so much of the history of the evolution of nuclear weapons is common knowledge that we need only remind the reader of its principal moments. 4 Since World War II the technology of weaponry has undergone three successive revolutions. The first massive change occurred with the appearance of atomic bombs, operating on the principle of fission. Atomic bombs of this sort had an explosive power one thousand times greater

152

Chapter Four

than that of conventional explosives that had been used before them. On the heels of the first revolution came a second massive change that ushered in thermonuclear bombs, operating on the fusion principle, which in tum increased a thousandfold the explosive power obtainable from atomic devices. In less than a decade after World War II, then, the destructive capacity of the bomb increased a million times. The third revolution concerned the addition of missiles to the already existing means of delivery of the new explosives. As long as airplanes were the major mode of transporting explosives to their targets, the intended victims had at least sufficient time in which to launch a counterstrike (if they too possessed nuclear weapons) and to mount some sort of defense to reduce the full measure of the blow they were about to receive. But with the new delivery vehicles capable of generating speeds several times greater than that of sound the victims of attack are left only minutes in which to ascertain that an attack is coming and to activate a counter-attack (if they too possess a nuclear force) before the enemy's nuclear missiles strike their targets. Moreover, one cannot really hope to interpose an effective defense against ICBMs carrying multimegaton warheads programmed to hit different targets simultaneously. And, certainly, thermonuclear bombs plus guided missiles preclude any expectation that a potential victim may somehow avoid devastation if attacked with nuclear weapons. Nuclear weapons are, then, fully as terrible as the proponents of deterrence theory say they are. Where only one of two parties in a conflict possesses nuclear weapons, the operation of deterrence seems relatively simple and straightforward. But in situations where both the parties are nuclear powers, deterrence becomes infinitely more complicated. How can a nuclear power make a nuclear adversary understand in advance that its attack will be resisted and that nuclear weapons will be used? How can a nuclear power convince another that it would place itself in mortal danger in pursuing its course of action when the threatener would itself be placed in mortal danger by its intended victim? Would such nuclear threats be credible?

153

Nuclear Arms Races and Deterrence

It should be noted that different nuclear strategies handle such claims differently. It is impossible to summarize all of the complicated processes, models, and ideas that deal with actors' behavior, intentions, and expectations in nuclear confrontation. One need not, for they are all readily available should one care to review them. The sketch below will touch only certain core problems that were chosen for discussion because they must figure in any theory of nuclear war based on the notion of deterrence. In view of the fact that, should a nuclear power decide to attack another with nuclear weapons, the victim cannot escape devastation, the victim's deterrent capability is no greater than that portion of its nuclear force that can be supposed to survive the strike the aggressor has been able to land before a retaliatory strike could be gotten under way. The recognition of this fact has led to the distinction between first- and second-strike capability. The term "first-strike capability" refers, of course, to the nuclear force available to a nation that might decide to strike first. Presumably, this might include its total nuclear force. Second-strike capability, on the other hand, is that portion of the total nuclear arsenal that has survived the first blow of an enemy and remains available to strike back. In all major strategies of deterrence, invulnerable second-strike capability is the guarantor of peace, an important tool in the regulation of nuclear war, or the assurance of revenge. Most of the theoretical works on nuclear strategy have been written by American strategists who, obviously, had the specific case of the United States in mind. But some of the questions they have asked reflect the problems faced by any nuclear power. Nuclear strategists have had to pose two major questions. First, how large and varied an arsenal of nuclear weapons must a nuclear power possess so that a sufficient fraction of its total force may survive the initial enemy attack and will in tum be able to penetrate all enemy defenses in such a way as to insure that the aggressor. will suffer unacceptable devastation under any and all circumstances? Second, how is the nuclear force to be employed? Snould the bulk on one's nuclear-tipped missiles and other dt ivery

154

Chapter Four

vehicles be targeted at the adversary's nuclear arsenal or at its cities? Because nuclear war would necessarily entail an attempt at the total destruction of each side by the other, an attempt which, given the nature of the weapons used, would inevitably succeed, one group of strategies makes no provision for a subsequent period of whatever sort after the first blow and counterblow have been struck. Such strategies also make a number of assumptions. In view of the finality of the consequences, nuclear weapons are not to be used first, and the purpose of using them would be the extinction of the enemy and nothing less. Given the nature of the weapons, nuclear war will either be total from the outset, or will not occur at all. A second group of strategies proposes almost diametrically opposite expectations. First, nuclear war need not immediately and inevitably take the form of an exchange of massive nuclear strikes, such an exchange being far more likely to occur at the very end of a long continuum of terror than at the beginning and as the result of the failure of the mechanism of deterrence at lower levels. Second, the threat of the use of nuclear weapons may operate as a form of blackmail to dissuade an adversary from committing aggressive acts of a limited nuclear and nonnuclear nature that do not immediately endanger the existence of the United States. (The eventuality of a limited nuclear war, then, excluded from the calculations of one set of strategies, is clearly possible under the provisions of the second.) Third, the preferred targets of nuclear strikes would be the opponent's nuclear forces and not primarily his cities, although different strategies suggest different proportions of military and nonmilitary targets. Fourth, any escalation of the conflict should be gradual. According to the proponents of this view, a highly discriminate selection of targets and the gradualness of escalation would cause deterrence to take hold at some point in the continuum of conflict below-and, it is hoped, well below-the point of exchange of massive retaliatory strikes. In any event, the enemy's population centers should be the principal target of nuclear attack only in the last resort-when the mechanism of deterrence fails

155

Nuclear Arms Races and Deterrence

to take hold earlier and nuclear war becomes total, with the complete obliteration ofthe enemy as the goal. At this point, of course, all strategies prescribe roughly the same course of action and formulate more or less identical expectations of the outcome. 5 Two examples might help in understanding how the proponents of the second set of strategies expect their schemes to work. The first portrays their image of the sequence of events expected to occur had the Soviet Union chosen to overrun Western Europe during the fifties and early sixties. In such an event the United States would have compelled the Russians to back down by means of a judicious use of the vast nuclear superiority she enjoyed at that time. America would counterpunch not by engaging the Russians in Western Europe but by hitting and destroying her nuclear offensive force at home. Left defenseless, the Russians would have no option but to retreat. They would then rebuild their nuclear force and mutual deterrence would be reestablished. And here, as our second example, is an eyewitness description of a proponent of the same strategic views giving advice in the Cuban missile crisis as to what might happen should the Americans decide to bomb the missile sites in Cuba . . . . A Soviet military response was likely. "What will the Soviets do in response?" He replied, "I think they'll knock out our missile bases in Turkey." "What do we do then?" "Under our NATO Treaty, we'd be obligated to knock out a base inside the Soviet Union." "What will they do then?" "Why, then we hope everyone will cool down and want to talk. " It seemed rather cool in the conference room as he spoke. 6 The key to the deterrence mechanism, then, is terror. Etymologically, the root of the word "deterrence" derives, appropriately enough, from the Latin terrere, which means to terrify. "To deter" means, simply, to frighten someone in such a way as to discourage him from doing something he might otherwise do. It is the fear of nuclear punishment that acts as a deterrent.

156

Chapter Four

It certainly makes good sense to assert that both elites and mass publics are in fact frightened by the presence of nuclear weapons in the hands of hostile powers. The constant admonishments addressed by elites to one another and to their respective mass publics as to the need to control arms races, to avoid direct nuclear confrontations and involvement in the quarrels of third parties, and the sighs of relief heard when such dangers are avoided, would certainly lend support to such an interpretation. But such information is not precisely what we need to know to solve the problems we have posed. What we wish to know is whether national leaders of the potential victims of nuclear powers are more frightened of such adversaries than they would be if the latter did not possess any nuclear arms. Or, even more precisely, in what manner does the fear of nuclear weapons influence the making of decisions by that handful of elites who have final choice in decisions of peace and war; and is that influence, in practice as it is in theory, of such a kind as to alter in any fundamental way the decisions that might otherwise be made if nuclear weapons did not exist? Direct evidence on such questions is hard to come by because, in the first place, the periods of maximum tension involving nuclear powers are mercifully rare, and also because the relevant decisions are made in secret. Whatever the reasons, we are all but reduced for our evidence to the reading of tea leaves. For example, there were rumors that the Soviet Politburo discussed the possibilities of using nuclear weapons against the Chinese when the two countries tilted over the Siberian border and that the decision not to use them won by one vote. 7 Again we have a rare glimpse of the kind oftorment undergone by President Kennedy during the Cuban missile crisis in this brief passage written by his brother Robert:

I think these few minutes were the time of gravest concern for the President. Was the world on the brink of a holocaust? Was it our error? A mistake? Was there something further that should have been done? Or not done? His hand went up to his face and covered his mouth. He opened and closed his fist. His face seemed

157

Nuclear Arms Races and Deterrence

drawn, his eyes pained, almost gray. We stared at each other across the table. For a few fleeting seconds, it was almost as though no one else was there and he was no longer the President. 8 We have, incredibly, a comparable account of the torment Nikita Khrushchev suffered on the other side. One can sense the Russian leader's despair in a candid passage in his memoir: I remember a period of six or seven days when the danger was particularly acute. Seeking to take the heat off the situation somehow, I suggested to the other members of the government: "Comrades, let's go to the Bolshoi Theater this evening. Our own people as well as foreign eyes will notice, and perhaps, it will calm them down. They'll say to themselves if Khrushchev and our other leaders are able to go to the opera, at a time like this, then at least tonight we can sleep peacefully." We were trying to disguise our own anxiety, which was intense. 9 But from such documentary evidence one can only detect the anguish of the two leaders; on other important points we still have no evidence. What was it precisely that stayed the Russian and American hands in these cases? How did the fear of nuclear retaliation influence the decisions? It also seems plausible to argue that the effects on governmental actions of the fear felt by governmental leaders will be related to the willingness to accept damage at the hands of the enemy. And it is at that point at which the fear of attack surpasses the willingness to accept damage at the enemy's hands that deterrence is supposed to take hold and brake the slide toward aggression. The difficulty is locating a priori where that point is. One may reasonably expect that its position will vary if the nation is democratic or not, is defending itself or aiding allies, has a large population or only a few million people, and so on. But we should also note that some evidence contradicts such propositions. Some American respondents indicated, early in the nuclear era, that between 10 million and 60 million American lives would be an appropriate price to pay for helping Western

158

Chapter Four

Europe against Russian aggression, which would seem to indicate that, contrary to expectations, the defense of allies may not greatly diminish a nation's readiness to suffer the consequences of nuclear war. Mao is reported to have suggested that China could survive the loss of 300 million Chinese because there would still be another 300 million left to carry on.10 When the numbers run into tens and hundreds of millions, there is an obvious air of unreality in calculations that try to pin down when it is that nations would consider the loss of lives to have reached an unacceptable level. One final point. Fear of nuclear weapons flows two ways. Such fear is not confined to the victim alone; the possessor is also frightened by the possible consequences of using the arsenal at his disposal. Little attention has been paid to this problem, and it constitutes an important lacuna in the theory of deterrence. We shall pursue it with determination in later discussion. Let us summarize. Nuclear weapons ought to put the governing elites and mass publics of nations which are potential victims in fear for their lives. And we shall assume that they do. But the reader should be aware that we have by way of proof for this assertion very little and often contradictory evidence. We know much less than we should. We do not know what nuclear weapons will generate sufficient fear to deflect the aggressor from his course. We do not know how to establish a priori the point at which damage will be thought by the potential victim to have reached unacceptable proportions. We do not have firm evidence that terrorizing potential victims does deflect them, as it is alleged to, from their aggression; very different conclusions can be drawn from the evidence than those proposed by the proponents of deterrence theory. This last matter is particularly serious, and the answer to the question it raises strikes at the heart of the belief that nuclear blackmail works. Testing Deterrence: Outcomes of Crises The last component of deterrence is the one we must examine with the greatest care. It contains a series of claims

159

Nuclear Arms Races and Deterrence

central to deterrence theory: that the nuclear threat will change the behavior of the threatened nation; that a nation threatened with nuclear retaliation will change her mind about committing aggression; that leaders have become more cautious in their dealings with one another because of the danger of a nuclear holocaust-in short, that nuclear blackmail works. These propositions are widely and firmly believed. They serve as a point of departure in the current interpretations of the behavior of great powers. They also are a point of departure for justifications of nuclear armament programs. But are such claims true? Do the claims fit the facts of international conflicts? It should be clear that if such claims are not true then the theory of unilateral or mutual deterrence is without foundation. If the claims are not true then one has to come to the conclusion that nuclear weapons are useful to fight wars, not to prevent them. The emperor has no clothes. Preparations to fight wars would become the sole explanation for the endless and expensive undertaking of some nations in developing and stockpiling nuclear weapons. But how is one to determine whether the recipients of nuclear threats actually change their minds and their behavior? Three research strategies seem indicated. The first consists of comparing two sets of conflicts-one in which nations possessing nuclear weapons are involved, and one in which they are not-in order to see whether there is at least a greater tendency in the second set of conflicts to solve disputes by war. The second strategy consists of a systematic examination of the first of these two sets of cases-those in which nuclear powers either were or could have become involved-to see at close range what bearings nuclear weapons had on outcomes. A third strategy is the exploration of a small number of the most important cases to observe at the individual case level the effects of nuclear blackmail. One could use either of two sets of data for crises in which nations possessing nuclear weapons are not involved. One could either use the crises before World War II, for it is with the end of World War II that nuclear weapons first appear,

160

Chapter Four

or conflicts since World War II where nuclear weapons were not and could not possibly have become a factor. We chose the latter. We compared the behavior of the contestants involved in all conflicts between 1945 and 1975 where nuclear power played no role with that of the behavior of contestants involved in all incidents in which the possible use of nuclear weapons may have influenced the outcome. I I One had to have a rigorous evaluation of another important factor: how badly each of the contestants in the crisis wanted whatever values were at issue. Such an estimate is essential. Otherwise how is one to tell whether the winner won because the loser was not trying; or, in the language of deterrence, how is one to tell that the deterred was deterred because he did not care enough about the issue to see the conflict through? The reader will recall that this problem posed itself again and again in the previous chapters.12 In our first comparison we were interested fundamentally in the simple question: were participants in conflicts where nuclear weapons were not and could not be involved more willing to take the risk of going to war than were the parties in cases where nuclear weapons were at least an important background factor? (Obviously, involvement of nuclear superpowers was a necessary condition for nuclear weapons being used.) The answer to our question emerges very neatly from a simple cross-tabulation (see table 4.1).13 Table 4.1

Conflicts in the International System, 1945-77

None No war War Total

143 (69.5%) 63 (30.5%) 206

Likelihood of Superpower War Low Possibleto-Likely 16 (59.3%) 11 (40.7%) 27

4 (28.6%) 10 (71.4%) 14

Actual

o o o

Note that variations in the degree of likelihood of a dispute being settled by military means, in situations in which nuclear powers are not involved, moved in the opposite direction from that which one would have expected. In the 206 cases where there was no possibility that nuclear

161

Nuclear Arms Races and Deterrence

weapons would ever have been used, because the countries possessing them could not have become involved, 69.5 percent of the disputes were solved without military conflicts and 30.5 percent were solved by war. But in the 27 cases where the possiblility of superpower involvement was low, only 59.3 percent were resolved peaceably while 40.7 percent were settled only after the outbreak of armed conflict. Finally, among the 14 cases where the probability of nuclear-power confrontation passes from low to possible or likely, the tendency to solve dispute by force of arms rises further still, peaceful solutions being found in 28.6 percent of the cases, military ones in 71.4 percent. These data convey their message loud and clear: the tendency to go to war increases as the likelihood of great-power involvement increases and as the possibility that nuclear weapons may be used becomes more real. Why, then, is the opposite believed? It can only be from wishful thinking. If, as is alleged, leaders have become more cautious when the danger of the use of nuclear weapons increases, these data do not show it. And deterrence is defined as the tendency to avoid risks among those who might become the victims of attack. The argument that leaders have become more aggressive in their dealings with one another since the advent of nuclear weapons may appear startling and almost incredible to all those who have lived through the reckless aggressions of a Hitler and a Tojo before World War II. But such readers should remember that a comparison from which a generalization is drawn must be systematic. There is so much contradiction in the real world that one can always prove almost anything one wishes with biased reporting. In the pre nuclear age one could certainly find instances of reckless risk-taking by such as Hitler and Tojo, but one could also find the excessive caution of a Stalin, a Franco, a Daladier, and a Chamberlain. And much the same thing can be said of conflicts in the nuclear age. One need only to recall the acts of some of the major participants in the Cuban missile crisis, and in Korea and Vietnam, to mention but three of the most important conflicts after World War II. If one were to compare Hitler and Eisenhower, and generalize solely on that

162

Chapter Four

basis, one would certainly say that leaders are more cautious in the nuclear age. But if one compares Franco and Giap, or Chamberlain and Khrushchev, would one reach the same conclusion? One simply cannot be allowed to pick and choose one's observations. And assertions that deterrence works because the Russians did not attack Western Europe when we had a nuclear advantage, or because the disputes that have arisen between the United States and the Soviet Union did not escalate into nuclear conflicts, or because there has been no World War III, are unproven though plausible. How do we know that the Russians wished to attack Western Europe and were deflected from this course solely by the American superiority in nuclear force? Viewed with the benefit of hindsight, it seems more likely that they were in no mind or position to attack Western Europe and risk war with the United States, even if the latter had possessed no nuclear weapons. And how can we be sure that, had nuclear weapons not been a factor, the chances of escalation of disputes into full-scale wars would have been greater? We cannot be sure. It is the forty-one recorded cases (see table 4.1) in which there was at least some likelihood of superpower involvement and nuclear confrontation which are the most important for our purposes. And among these forty-one cases, it seemed simplest to examine only the fourteen cases in which escalation to nuclear war was at least thought likely-and in which the workings of deterrence should be at their most visible-and to omit examination of the category in which such likelihood was low. If there had been evidence that the discarded conflicts might not have escalated precisely because the governing elites wished to avoid the risks of nuclear war, the decision to omit them might have been unwise. But no such evidence existed. Certainly such conflicts were not resolved in favor of nuclear powers more frequently than the conflicts we are discussing. Moreover, precisely because the fourteen cases we chose to examine were those serious enough to have escalated to the point at which it was reasonable to fear that nuclear weapons might be used, they offered the best chance of testing whether

163

Nuclear Arms Races and Deterrence

participants sought to avoid increasing the risk of nuclear conflict. But how is one to isolate the effect that shows the avoidance of such risks? The best and most direct way is to observe whether nuclear powers in confrontation with other countries get what they want. If both sides want in equal measure whatever chances to be at issue, the following results should be expected. The nonnuclear power must give way; it can do so grudgingly or willingly, surlily or with grace, but give way it must. The nuclear power should get its way. It can be truculent, bullying, self-righteous, understanding, or dignified. But in the end, after the bluster, the meetings, the compromises, the headlines, it should have its way. That is, if deterrence works. And the nuclear power should have its way in the central issue in the dispute. Compensation elsewhere for having lost in the main arena should not suffice to qualify the loss. In every case the loser is thrown a bone, a side benefit, a device with which it may save face. Experts argue endlessly over what weights to assign such things. But all this does not matter in our coding system: only the major issue counts. All this is relatively easy to capture when one party to the dispute has nuclear weapons and the other does not. But it becomes far more complicated in instances of mutual deterrence, in which both parties possess nuclear arms and both are thus possible attackers and potential victims. The fourteen cases we shall examine are as follows: the Chinese Civil War in 1945-49; the Berlin blockade in 1948; the Czechoslovakian coup in 1948; the Korean War in 1950-53; the Hungarian revolt in 1956; the Suez crisis in 1956; the Berlin Wall in 1961; the Cuban crisis, 1962; the Vietnam War in 1964-73; the Arab-Israeli War in 1967; the second Czechoslovakian coup in 1968; the Sino-Soviet dispute in 1969; the Arab-Israeli War in 1973; the SinoVietnamese-Soviet dispute in 1979. This list includes seven cases in which one side had nuclear weapons while its adversary either did not have any nuclear capability at all or had an ineffective one; four cases of mutual deterrence; and three cases which were so complicated that they could fit in

164

Chapter Four

one or the other category. We shall deal with each of the groups separately and within each group we shall deal with each case in its chronological order. Unilateral Deterrence Cases The seven cases where a nuclear power was opposed by a nonnuclear power or by one with an ineffective nuclear force were the Chinese Civil War, the first Czech coup, the Berlin blockade, the Korean War, the Hungarian revolt, Vietnam, and the Sino-Soviet dispute in 1969. The nonnuclear opponent was at different times another superpower, a great power, or a small country. The seven cases we have are not many, but they furnish precisely the ideal conditions in which the mechanism of deterrence should operate and in which, therefore, the theory of deterrence can be tested. Evidence of failure of deterrence in such cases would make it extremely unlikely that the results would be different in any other test. The results of the confrontations are as follows. The Chinese Civil War. The occupation of China by the Communist forces of Mao Tse-tung was a major setback for the United States. In the whole period from 1945 to 1949 the United States had hegemony in nuclear weapons, and the memory of American use of atomic power against Japan must have been very vivid. Almost to the end, the American commitment to oppose the Communist takeover was strong. The American government helped Chiang Kai-shek with diplomatic, economic, and military aid. When Chiang began to sink under Communist pressure, and Communist forces chased him out of China to Formosa, all of America's help went down the drain. In years to come this loss was to shake American political society to its roots and shape the nature of American international participation. And yet the possibility of American intervention did not deter the Chinese. They took whatever risk there was. H could be argued that the Chinese Civil war should not be included among our cases; that it was an entirely internal conflict and that nuclear weapons were not involved. Such a

165

Nuclear Arms Races and Deterrence

view is not entirely implausible, but, of course, one could argue the other way. The Berlin Blockade, 1948. In the Berlin blockade the bare bones of the dispute were as follows. The USSR had closed the access road to Berlin in violation of agreements the Soviets had signed with the Allies, and the only access to the city was by air. The United States resupplied Berlin through an airlift for the entire year the dispute continued. The USSR finally gave way and reopened the surface approaches to the city. There can be no doubt what the central issue was in this dispute. There can be no doubt that the United States won and Russia lost. The USSR had overwhelming superiority on the spot but the United States had nuclear hegemony. Under our coding rules deterrence worked. Czechoslovakia I, 1948. In the Czech coup of 1948 Czech Communists, relying on the threat of the use of the Soviet military forces, maneuvered themselves into control of the Czech government. Their action and Russian support of the takeover were incredible reminders of Hitler's tactics. The American government and people were deeply shocked. It can be argued that the conflict in which the Communists unseated the Czech government and replaced it with one subservient to Moscow was entirely an internal conflict, not an international one, and that therefore the conditions for a test of deterrence do not obtain. But this view does not seem justified. The Russo-American conflict was only very thinly veiled. The Russians should have considered it risky to have their dependents act as they did in Czechoslovakia. How large the risk was could be seen from the international consequence ofthe case; it was the beginning of the cold war. In any event, we judge that the Russians won this conflict and that deterrence did not work. The Korean War. In the Korean War deterrence failed. The central issue was the Chinese pushing the American forces from the Yalu River, the border between China and North Korea, back to the middle of the country, not far from

166

Chapter Four

where the war had started when North Korea had attacked the South. The provocation for the United States was very high. America was directly involved. The Chinese had attacked and routed the American army in Korea and had then stalemated it in the field (the Americans never got back up north). Because of Chinese intervention, American casualties and losses of equipment were high and American prestige suffered greatly. Important elements of the American civil, political, and military elites believed that the United States should have escalated the war by attacking China. The asymmetries in ultimate capabilities were awesome. Moreover, during the whole of the Chinese intervention, Chinese industries and most of the Chinese military forces in the staging areas for Korea were within easy reach of American air power. America still had hegemony in atomic weapons, she was invulnerable to Chinese attack, and the memories of Nagasaki and Hiroshima were still vivid. How, then, is one to explain the willingness of the Chinese to take the risk? Two explanations have been advanced for the Chinese behavior. The first is that the Chinese counted on the protection of the atomic arsenal of the USSR. Although the Soviet Union had, in fact, exploded a device, it is doubtful that she had atomic weapons in any quantity. She certainly did not have the means to reach the United States with ease, and she \yas herself exposed to American retaliation. This first exphination seems groundless. Second, it has been argued that the Chinese knew that the United States would not use nuclear weapons against them. To be sure, there was no certainty that the United States would use nuclear weapons to repel the attack and stave off the initial rout and the eventual stalemate. But one could have reasonably expected a repetition of what the United States had done five years earlier in Nagasaki and Hiroshima. Korea was the first major American war since World War II, and the United States was handed what for her was a damaging setback. It seems strange that under the circumstances the Chinese did not think it very risky to rely on American restraint. The argument that America's opponents knew that the United States would not use her nuclear

167

Nuclear Arms Races and Deterrence

power is an explanation that has been used often to account for the choice by those opponents of a course that otherwise might have appeared extremely hazardous. But such an argument gained credibility only as the memory of Hiroshima and Nagasaki faded and the evidence built up that the United States would accept grave defeats and refrain from using nuclear power. Such evidence, however, was lacking in 1950 and the argument is totally incredible for that time. The Hungarian Revolt, 1956. The Hungarian revolt is also a good test of deterrence. This revolt occurred in the last years of the period when the United States, because of the inadequacy of the Russian delivery systems, could still devastate the Soviet Union without fear of effective retaliation. The target of Russian nuclear forces was Western Europe. Russia then, held an immediate advantage on the spot but the United States had a critical edge in destructive capability. The central issue was whether the Soviet Union would be able to reestablish her rule over the Hungarians; they had revolted and toppled from power their government, which was subservient to the Russians, and had then withdrawn from the Warsaw Pact and declared their independence between East and West. When the Hungarians revolted there was strong expectation that America would intervene on their behalf. The Hungarians pleaded with the West, and particularly with the United States, to help. The Eisenhower administration, and the Republicans then in power, had made public commitments to help liberate Eastern Europe, and the commitments were believed to be serious ones. The United States lost, the USSR won. Why was Russia not afraid enough? Why did deterrence fail? The Vietnam War, 1964-73. Throughout this war North Vietnam defied the United States at every tum. Knowing full well that the United States was the major nuclear power in the world and that she herself possessed no nuclear weapons, why was North Vietnam willing to take the risk of annihilation? Or was it that the risk of nuclear retaliation was not very high? Two explanations are usually given for North Vietnamese behavior. First, North Vietnam could

168

Chapter Four

count on Russian help; second, the United States had assured the world that she would not use her nuclear power in Vietnam. Let us review each of these explanations. There can be no question of the close tie between North Vietnam and the Soviet Union, and of the fact that the USSR was up to fighting a nuclear war with the United States. Whether the USSR thought North Vietnam worth the risk of her own nation's life cannot be known. The other reason can also be taken seriously. But should American assurances have been entirely credible to North Vietnam? Consider. The provocation for America was the most acute imaginable. Her armies were frustrated in the field. Conventional air power had failed utterly to stop the flow of arms and replacements from North Vietnam to the South. The war was fomenting revolts within the United States and was ruining the American economy. American international prestige had fallen to a low it had never reached before. Under the circumstances should American assurances of self-restraint have been entirely credible to a potential victim? Should not North Vietnam have been more than a little anxious? Should not North Vietnam's anxiety over nuclear retaliation have become more serious as the Vietnamese handed the Americans military defeat after defeat, as American losses of men and materials grew, as American desire to get out of the mire grew more frantic? The central issue of the Vietnamese conflict was clearly the attempt of the United States to prevent North Vietnam from gaining control of South Vietnam. The American loss is clear and so is the failure of deterrence. The Sino-Soviet Dispute, 1969. This was the first of two crises between the Russians and the Chinese. In this case the Russians and the Chinese clashed over their rights on the Damansky/Chenpao Island in the middle of the Ussuri River. The Chinese ambushed a Russian patrol, wounding and killing a number of Russian soldiers. The Russians renewed the duel with the Chinese ten days later, by which time both sides had built up considerable local forces. Reports suggest that the Chinese ended in control of the area. The dispute was part of a larger continuing disagreement

169

Nuclear Arms Races and Deterrence

about where and how the border between the two countries should be drawn, and that dispute was in tum part of a bitter and fundamental fight between the Communist super powers. Although in this wider perspective the issue is a stalemate, we code the Chinese as having won the fight. We find that the distribution of nuclear power in this case permits an excellent test of the deterrence concept. Both the Chinese and the Russians could mobilize substantial forces on the ground. But Russia was a major nuclear power and the Chinese simply did not have, at the time of the Ussuri incident, sufficient nuclear capability to hurt the USSR.14 The only Chinese delivery vehicles were bombers that were unlikely to be able to penetrate Russian airspace. The initial provocation for the Damansky/Chenpao Island incident came from the Chinese, and the Chinese continued to act in the toughest manner toward the Russians. If deterrence works, it certainly did not work for the USSR against the Chinese, just as it did not work for the United States against the Chinese. In the second incident, to be discussed below, one witnesses a repetition of the story. Why have not the Chinese ever proven sufficiently afraid of countries they know to be bitterly hostile and who happen to be the two major nuclear powers in the world? Why do they invariably brave the risk involved? Mutual Deterrence Cases In four cases-the building of the Berlin Wall in 1961, the Cuban missile crisis in 1962, the Czechoslovakian coup in 1968, and the Sino-Soviet confrontation in 1979-we have situations of mutual deterrence. Because mutual deterrence is involved, a discussion exploring the behavior of the contestants case by case can go only so far in answering the question of whether deterrence has worked. Yet under mutual deterrence conditions, unless the contest results in a draw, one of the contestants still wins the central issue of the dispute. But what role does mutual deterrence play in the victory of one party and the defeat of another? The Berlin Wall, 1961. The central issue of this dispute was clearly the action ofthe USSR and the East Germans in

170

Chapter Four

building the wall that split the city. The reason for erecting the wall was the large-scale defection of skilled East Germans; dissatisfied with the poor quality of life in East Germany, they passed over to West Berlin, and then to the West. The defection of a critical fraction of her population was ruinous for East Germany's economy and morale, and it was essential for the government to stem the tide. The wall did just that: the flood turned to a trickle. The United States raged, but did not act. The USSR won the conflict, but did she do so because she deterred the United States? The distribution of force was roughly as follows. The Russians had superiority on the ground and by 1961 had a nuclear capability that could reach American territory. Would the United States have gone in and leveled the wall but for Russian nuclear power and the danger of nuclear war? There is really no way to tell with any degree of certainty. But one should remember that, in the blockade some dozen years previously, when the United States was not in danger of nuclear retaliation, American actions were equally cautious. The building of a wall intended to imprison one's own people was a barbaric act, but the barbarism was inflicted by Communist authorities on a Communist population, within the Communist sphere of influence. It was, in a sense, perfectly "legitimate" for the USSR to try to consolidate her power in her own world. The Cuban Missile Crisis, 1962. The central issue of the Cuban missile crisis was the missiles the USSR had sneaked into Cuba and that the United States had discovered and wished removed. After some awful days of indecision, the USSR dismantled the missiles, crated them, and shipped them home. We consider the Cuban missile crisis a clear win for the United States and a clear loss for the USSR. There are those who disagree with this evaluation. They argue that the USSR also won. The reason they put forward is that, in the negotiations, the USSR asked the United States to make a public declaration that she would not again try to invade Cuba as she had done just a year before. And the United States agreed. Should not the USSR, then, also be coded a winner? In our judgment, this would be a com-

171

Nuclear Arms Races and Deterrence

pletely wrong evaluation of the substance of the case and a clear violation of our coding rules. In the Cuban missile crisis the central issue was the presence of the missiles. The Russians, incidentally, coded the outcome as we have. One of the reasons Khrushchev was removed from power was the Cuban fiasco. One could doubt that the government and party in the Soviet Union would have rewarded a success by retiring the leader responsible for it. But what role did deterrence play? The missile crisis was the most dangerous and the most famous of all the cases where deterrence is alleged to have played a part. In this case the confrontation was direct. Both countries had nuclear forces sufficient to obliterate the other, but the United States, due to the proximity of Cuba to American shores, could commit a superior conventional force to the site of the dispute. This advantage proved crucial. Russian ships carrying missiles to Cuba turned back, on orders from Moscow, when they met the line of the American blockade at sea; and Russian missiles already in Cuba were dismantled and shipped home when the United States told the Russians that, unless the missiles were removed, the United States would send her troops and planes to take them out. When the Russians backed down, everyone credited the nuclear deterrent with the victory. But if it was the nuclear deterrent that determined victory for the Americans, how can one explain the fact that the United States was not herself deterred by the Russian nuclear force and did not pursue any of the many avenues open to her to avoid a nuclear confrontation? The Czechoslovakian Coup, 1968. The second Czech crisis was in large part a replay of what had occurred in Hungary a dozen years earlier. The Czech population revolted against a Communist government subservient to Moscow and replaced it with one that was much more independent. The revolt was less violent than the one in Hungary had been, and the transfer of power was far more legal, but popular support and street violence were critical in moving the Russian-supported forces out of power and putting a more nationalistic coalition in. The Russians intervened

172

Chapter Four

only when they were convinced that the independence of the national Communists in Czechoslovakia might become a threat to Soviet rule over the country and over Eastern Europe. They reconquered Czechoslovakia, had the rebel leaders executed or imprisoned, and reestablished control. Americans wrung their hands, but did not move. After all, Eastern Europe was part of the Russian world. Schemes of liberating Eastern Europe had long been dissipated. There is one way in which the Czech case of 1968 differed from the Hungarian case that had occurred a dozen years earlier. This time the USSR had a nuclear force that threatened the United States. Whether the Russian invasion of the country constitutes a failure of deterrence depends on whether the United States wanted to stop the Russians from reducing Czech independence a second time. In 1968 the United States was very busy in Vietnam. One also has the impression that, by 1968, international spheres of influence were becoming thoroughly accepted and that each of the leaders of the competing international orders had gained a free hand to keep the smaller members of her system in line. What the Soviet Union did in Czechoslovakia was, then, "legitimate." What was not tolerable was for the Soviet Union to try to expand her influence beyond her allotted sphere. This was precisely what she had tried to do in Cuba.

The Sino-Vietnamese-Soviet Confrontation. 1979. This conflict contains in reality two cases. A confrontation between a nuclear and nonnuclear power, and a case of mutual deterrence. The conflict was precipitated when Russia's ally, Vietnam, attacked and conquered China's ally, Cambodia, and replaced the government in power with one tied to Hanoi. The Chinese invaded Vietnam to "punish" it for its "mischief." Before and during the weeks-long fighting the Russians threatened the Chinese with retaliation, but the Chinese persevered and the fighting continued in accordance with a time table the Chinese had announced publicly. After having administered "punishment" to their own satisfaction, in behavior reminiscent of what they had done

173

Nuclear Arms Races and Deterrence

in the Sino-Indian dispute in 1961, the Chinese disengaged and withdrew from Vietnam. It is clear that Vietnam in attacking and conquering Cambodia was defying a nuclear power, and that this defiance escalated when the Vietnamese refused to withdraw and fought a number of pitched battles when the Chinese attacked Vietnam, with hundreds of thousands of troops on both sides. The defiance paid off. The Vietnamese resisted successfully giving up their kill and the Chinese accepted what was a fait accompli in Cambodia. A second confrontation in the conflict was going on between the Chinese and the Russians. As China began threatening Vietnam with retaliation and then began its attack, the Russians began threatening the Chinese with dire consequences if they attacked Vietnam. The Chinese obviously were concerned and took the Russians seriously enough to evacuate civilians from some border towns. But China defied the Russians, and continued doing so until it felt it had achieved its purpose. It can be argued, of course, that the Vietnamese felt they could defy the Chinese with impunity because the Chinese were explicit that they only wanted to teach the Vietnamese a lesson, and because the Vietnamese were under Soviet protection. And this may indeed be so. But if Chinese nuclear power was sufficient to deter the Russians, as everyone argues, then it follows that the Chinese had a free hand in Vietnam, and the behavior of the Vietnamese is again inexplicable. And again, why was Russian nuclear power, immense as it was, insufficient to deter the Chinese? One cannot answer these questions. But it does seem certain that this conflict violates every expectation advanced by the theory of deterrence. A nonnuclear power defies a weak nuclear power and gets away with it, and a weak nuclear power in tum defies a nuclear superpower and gets away with it. What happened to deterrence? Also, what happened to the assumption that the Soviet Union was inherently aggressive and would take advantage of a nonnuclear, or a weaker nuclear, adversary whenever it could? Why did the Russians stay their hand?

174

Chapter Four

Defective Tests of Deterrence In the final three cases the conditions for a sufficiently clear test of deterrence do not exist. We shall discuss them briefly, nevertheless, but we shall not use them in our calculations of the number of cases where deterrence worked. The cases in question are the Suez crisis in 1956 and the Arab-Israeli conflicts of 1967 and 1973. The Suez Crisis, 1956. Suez was a bizarre affair. The crisis was precipitated by the Franco-British forces, aided by the Israelis, trying to recapture the Suez Canal Nasser had nationalized only a short time earlier. Both the United States and the USSR threatened the British and the French with retaliation unless they withdrew. The USSR threatened nuclear retaliation. The United States responded to the USSR that she would protect the British and the French, but she threatened those countries with economic sanctions unless they complied. The invaders complied with the demands of the United States and the USSR. One can argue that deterrence appeared to work. If the British and the French withdrew in response to Russian threats, the judgment is justified. But such an interpretation of Franco-British behavior is outlandish. The Russian threat was hardly credible. The USSR at that very moment was doing to Hungary what the British and the French were doing to Egypt and was scarcely in a position to start a major conflict. Moreover, the American counterthreat of nuclear retaliation against the Russians, at a time when the Russians could not reach American targets but the United States could reach with ease the Soviet Union, was sufficient to make Russian nuclear rattling simply posturing. It does seem that the threat of economic sanctions, not Russian nuclear retaliation, brought the French and English to heel. On the other hand it could be argued, if one accepts the Russian threats as serious, that American counterthreats deterred the USSR and that, therefore, deterrence worked. This view seems totally unsubstantiated. The Russians clearly were posturing, and so was the United States. But of course we cannot be absolutely sure.

175

Nuclear Arms Races and Deterrence

The Arab-Israeli Conflicts, 1967, 1973. The two conflicts can be treated together. From the point of view of deterrence the two are perplexing cases. The bare facts are as follows. In 1967, menaced by the actions of the Arab states, the Israelis decided to attack first; the lightning victory that followed brought major territorial gains to the Israelis and crushing, humiliating defeat to the Arab states. In the second war the Arabs attacked first and the central issue was the reconquest by the Arabs of the lands lost in 1967. The Arabs failed in this, but they had a chance to show that they could fight. Their new performance gained them what everybody called a moral victory. In neither case did the superpowers intervene directly, but the United States provisioned the Israelis and the USSR the Arabs. The fact that the Israelis won both times strengthened immensely the American presence in the Middle East. On the other hand, the improved Arab performance in the field did not hurt the Soviet Union. It is clear that neither the United States nor the USSR tried to deter the fighting by the combatants. The USSR made no move to stop Israel's victory in 1967. Again, neither the United States nor the Soviet Union interfered with the development of the battle in the 1973 war until the Arabs had been beaten in the field. The United States did not make threats of nuclear retaliation to counter Arab initial successes (though it placed its forces on alert at the news that the Czechs were going to ship arms to Syria), and the Soviet Union did not try to deter the Israelis from regaining the upper hand. The rule that both superpowers followed seems to have been that they would help each side to help itself but would not do any of the fighting for either side. If deterrence played a role, it could only have been in inhibiting the United States and the USSR from interfering in the fighting. But there is no way of knowing whether that happened. Naturally, the United States did not want to interfere, since her side was winning. Whether the USSR would have interfered directly but for the nuclear power of the United States will never be known. It seems clear that victory or defeat was determined in major part by conventional strength on the ground.

176

Chapter Four

We code the United States as winning and Russia as losing in the 1967 war, and both the superpowers as winning in the 1973 conflict. Whether, and how, deterrence played a role in the conflicts cannot be determined at this point. Conclusions Let us now pull together some of the conclusions that flow from this portion of the discussion. Several points need to be made. 1. When we compared the behavior of countries in conflicts where nuclear weapons were available with that of countries in conflicts where nuclear powers could not possibly have been involved, we found no evidence at all that countries are more cautious when conflicts have the potential to escalate to nuclear war. In fact the evidence pointed the other way. 2. When we narrowed our focus to an examination of the conflicts where the possibility of escalation to nuclear war was a serious matter, we found no support at all for the way the theory of deterrence purports to account for the behavior of countries involved in potentially nuclear conflicts. There were seven of the fourteen cases in which a nuclear power confronted nonnuclear adversaries-the most powerful natural test of deterrence one could find-and in only one ofthe seven, that of the Berlin blockade, did the nuclear power win, so that one could logically argue that the necessary conditions for successful deterrence were present. In the remaining cases deterrence did not work. Nonnuclear powers defied, attacked, and defeated nuclear powers and got away with it. There is simply no way in the world given these findings, one can support the theory of deterrence as founded. New ways to account for behavior in nuclear conflicts need to be found. 3. Four of the cases we have examined have been cases of mutual deterrence. They are the Czech coup of 1968, the Cuban missile crisis in 1962, the Berlin Wall of 1961, and the Sino-Soviet dispute in 1979. Did deterrence work in these cases? At this point the answer is both yes and no. In all four confrontations it is clear that the judgment for or against deterrence rests on the angle of vision of the observer. One

177

Nuclear Arms Races and Deterrence

can say that the victor in all these cases deterred the loser. But, then, one is still duty bound to explain why it was that, given mutual deterrence, the loser did not deter the winner. We shall turn to a further test of the validity of mutual deterrence in the remainder of this chapter. 4. There were three cases-Suez and the Arab-Israeli Wars-where the superpowers had a major interest in the dispute but were not directly involved. We have discussed these cases fully and concluded that the arguments one could make for or against any connection between deterrence and the outcomes are so tenuous that it would be prudent to omit their use in our estimate of the workings of deterrence. 5. Can one identify what factors go into determining victories or defeats in nuclear confrontations, and whether a nation threatened with nuclear retaliation will take the risk and defy the nuclear power? It is clear that the evidence we have reviewed is meager, that one can at most wonder about possibilities. But one point comes to mind. In all the cases we have reviewed (with the exception of the United States and the USSR in the Suez crisis and the Berlin blockade) winners have had conventional military superiority on the site of the dispute. This factor seems critical. If a contestant is weak on the spot where fighting is to be done, chances are that, whatever its status in terms of nuclear armaments, it will lose the fight. In all of our cases the winners have been the countries that could win without turning to nuclear weapons. Those who could not win without resorting to nuclear weapons did not use them even if they had them, and lost the conflicts. In other words, the victor was the country that could win without escalating the dispute. And this seems equally true when only one side to the dispute had an effective nuclear arsenal, as in the Korean and Hungarian cases, or when effective mutual deterrence was operating, as in the cases of Cuba and the Berlin Wall. It is telling that in both types of cases the results of the conflicts should be the same. And since forces on the ground seem to have controlled outcomes, it is not likely that the results would have been different if any of the conflicts we have reviewed had taken place in a world without nuclear weapons. The

178

Chapter Four

implications of this hypothesis will be taken up in the last section of this chapter. We have not stressed the other half of the information contained in our findings. One should also ask why nuclear weapons were not used, or were not more often threatened to be used, when deterrence failed. Such weapons were not seriously considered even when the nuclear powers were very sorely tried. If one takes into account only actual behavior and disregards the pronouncements that are made, one is led to the bizarre conclusion that nuclear weapons do not deter the opponents of nuclear powers, they deter only their possessors. One is tempted to be malicious and suggest that, on the basis of the record, nuclear proliferation is the best guarantee against nuclear war. 15 These observations lead to another point. Precisely because America was for so long the only major nuclear power, Americans have done most of, if not the best, theorizing on deterrence; and confronted with this phenomenon of nuclear self-restraint they chose to explain it by asserting that it was the unavoidable result of their own international political culture of civility and humaneness. For the United States, they suggested, no other course was possible. Their explanation implied that if, unfortunately for the world, the Soviet Union had eIijoyed the crushing American nuclear superiority, Russia would have taken every possible advantage of countries at their nuclear mercy. This view appealed to American biases and was widely believed. More recent evidence casts serious doubt on it, however. In her dispute with the Chinese the USSR never fully used threats of nuclear retaliation as an instrument to bring her impossible neighbor to her senses, and Russian self-restraint could not possibly have been motivated by fear of nuclear retaliation. 16 One could argue, of course, that Russian restraint against the Chinese is no guarantee of Russian restraint against Western powers; but if the feelings of the USSR for her competitor were what she said they were, China was in far greater danger than the Western powers. In any event, Russian restraint toward the Chinese does represent some evidence that the estimates that Western European tranquility and security would have

179

Nuclear Arms Races and Deterrence

been in extreme danger if only the Russians and not the Americans had had nuclear superiority, should no longer be accepted without reservation. It does seem then that, regardless of risk, a nation will fight if she feels her action to be legitimate. We should warn the reader that we have used the term "legitimacy" in quite a restricted sense. No other term, however, would fit better what we wish to express. What a nation may be doing may be an act of barbarism, but the moral quality of the act itself is not under discussion. Rather, what is at issue is whether an international actor has the right to do what it is doing, repellent though the act may be. For example, the world thought it legitimate for the Soviet Union to invade Czechoslovakia and Hungary and stamp out their freedom; and similar rules were applied in our own interference in Cuba. For a world divided into spheres of influence it is "legitimate" for the powers that head the competing international orders to deal harshly with rebellions in the ranks of nations recognized to be subject to their control. On the other hand, it was not thought legitimate for the Soviet Union to try to extend her area of control in a portion of the world that the United States had considered within her sphere. Hence American readiness to fight in the Cuban missile crisis and Russian willingness to give way. Legitimacy, then, is of maximum importance in nuclear confrontations. One final point. It seems clear from the evidence we have reviewed that nuclear missiles are not the miracle weapons they were thought to be at the beginning of the nuclear era. All attempts to isolate behavior that would constitute deterrence fail. Terror does not provide security. Nuclear weapons do not produce peace without being used, and they do not protect their possessors; if peace is really the business of those who are charged with the responsibility of using nuclear weapons, then clearly a very great deal of money and a very great deal of work have gone down the drain. In short, nuclear weapons do not deter confrontations at all levels. To believe they do is to believe in magic. 17

180

Chapter Four

Testing Mutual Deterrence: The Nuclear Arms Race But we cannot be absolutely sure; in cases where both sides have nuclear weapons it is very difficult to isolate the role that those weapons allegedly play in the conclusion of a conflict. And one cannot prove one way or the other whether deterrence plays the role that the doctrine of deterrence says it does. Thus the case of mutual deterrence remains open, and the most important part of our problem is unresolved. Even after assuming that nuclear weapons in the hands of some countries do not deter nonnuclear powers, one can argue that it is an illicit inferential leap of major proportions to conclude that deterrence also plays no role in deciding disputes between the United States and the Soviet Union as well. Indeed many of the deterrence theorists would argue that the power of nuclear weapons is so great that it is reserved solely to forestall the possibility of nuclear attack. But is this claim true? Unfortunately, when both litigants have nuclear arms, it is not a simple matter to establish the role that deterrence plays in the settlement of a dispute. The problem should be obvious. When both countries have nuclear forces, and the result of the conflict is not a draw, how is one to know whether nuclear weapons have had anything to do with the victory of one party and the loss of the other? How is one to explain the behavior of the opponent that stands its ground and is not deterred? One can say that the side that gave in did so because the winner had nuclear weapons, and happily conclude that the theory of mutual deterrence was validated. This is precisely the way commentators (but interestingly enough not President John F. Kennedy) treated the results of the Cuban missile crisis. We ourselves coded the United States as the winner in Cuba because she obtained what she wanted in that dispute, and we let it go at that. Nor could we have done more with the evidence we were using. But the reader is fully aware that we did not answer the really important question. Why, in the Cuban missile crisis, was Russia frightened off the island but the United States was not deterred from pushing Russia so near the brink of nuclear war? Would the United States have been deterred

181

Nuclear Arms Races and Deterrence

had the Russians held their ground? These are precisely the questions we should like to answer. The Action-Reaction Model It should be clear why the ways we used to explore the cases where nuclear powers confronted nonnuclear ones cannot be employed in the examination of mutual nuclear deterrence. We cannot set up natural experiments in the case of mutual deterrence; we must come at the question another way, with different data and methods, if we are to get off dead center. What we need to do is to identify a condition essential to the operation of mutual deterrence which lends itself to empirical control and thus enables us to test for its presence. A necessary condition for the theory of deterrence to be valid, in cases where both sides have nuclear weapons, is that the competing nations react to one another. Because, in deterrence doctrine, fear of the nuclear arsenal of an opponent and of his threat to use it deters a potential aggressor from attacking, the occurrence of nuclear deterrence is heavily dependent on the credibility, in the eyes of the attacker, of the retaliatory power of the defendant. But the credibility of threats and the perceptions of each contestant are rooted in the realities of the nuclear capabilities of an opponent. Given the intelligence at their disposal, each of the superpowers cannot but have a realistic appreciation of what it has to fear at the hands of the other. This fear drives each country not to fall behind in its nuclear capabilities and sets off nuclear arms races. The reasons for all this should be plain. If each country must rely on its ability to threaten the adversary with nuclear retaliation in order to guarantee its own safety, then each member of the dyad must keep pace with the development of its adversary's nuclear arsenal. If the credibility of the deterrent lies in the fear of retaliation, a major tool available to each nuclear power to stoke the fear of its opponent to desired levels is the judicious increase of its nuclear force to cover every contingency. It should be noted that in the context of deterrence theory it is not necessary to assume that both sides in nuclear arms races must make the same amount of effort or have the same level of capabilities. One

182

Chapter Four

must assume, however, that each of the contestants will allocate substantial portions of the resources scheduled to be used in the improvement of nuclear capabilities in direct response to the other's allocations. Hence one must compete and even race with one's opponent. And the race continues even after both contestants reach a second-strike capability. One must always keep in mind that the invulnerability of the defendant's deterrent depends on the power of the aggressor's initial attack. The logic of nuclear arms races, balances of terror, and deterrence are very close. ls If effective arms control agreements are introduced in the model, the situation described above changes. Fear is still a critical variable, spurring both the management of nuclear arsenals and deterrence, but each side is restrained from amassing nuclear arms beyond the established limits. Both sides are still competing, still fearful, and still deterring each other. The positions we have summarized have long been known. Much has been written about them. 19 Out of this work the relevant alternatives consistent with deterrence theory can be represented in a deterrence model (see fig. 4.2).

The model contains two propositions: 1. Fear of an opponent's weapons, in the absence of any agreement to limit arms, will lead to an arms race to insure stable deterrence.

-

if

/

~ race

Fear of nuclear conflict

if

Fig. 4.2

"Stable deterrence

"

Mutual arms limitation

The deterrence model.

-

Controlled competition

/

183

Nuclear Arms Races and Deterrence

2. Fear of the nuclear capabilities of an opponent, in the presence of nuclear arms control agreements, should lead to limited arms competition and to stable deterrence. These propositions have the following implications. (1) For each of the contestants fear of nuclear retaliation is a necessary condition for arms races, arms competition, arms limitation agreements, and stable deterrence to occur. (2) The expectation that nuclear weapons will deter is a necessary condition for nuclear nations wishing to increase or limit their nuclear arsenals. (3) In the absence of agreements to limit arms buildups, each side's efforts to match the arms-building efforts of the other side is a necessary condition for stable deterrence. 20 The alleged causal link between fear, nuclear arms races, and deterrence offers us a major opportunity to test the validity of the theory of deterrence. The interaction between contestants is the condition essential for deterrence to occur. Arms competition is the obvious component of the model that lends itself most easily to empirical control, and we intend to test for its presence. If an arms race can be shown to exist and one assumes that it reflects the mutual dread consistent with the notion of deterrence, then we may also assume that deterrence is taking place. Were we to find that the contestants were not competing, the proposition underpinning the theory of nuclear deterrence would be invalid. The Incremental Model All we have said up to this point simply assumes that nations build armaments almost in a Pavlovian response to the building of arms by their adversaries. We have noted at the beginning of this chapter, however, that there exists a completely different way to explain investment of resources in defense. Many have argued that to view defense budgeting from an international angle of vision-intuitively the sensible thing to do-is to overlook critical portions of the reality-how in truth military budgets are put together; and that, in fact, internal stimuli, rather than international determinants, are, year in and year out, the major factors in the allocation of resources to military power. The central

184

Chapter Four

proposition of this approach can be presented simply. The last allocation for defense is always the point of departure for decisions to allocate new money for arms. "Organizational politics" --organizational requirements, bureaucratic coalitions, factions, powerful clienteles, and independent allied decision-makers-push in the direction of increases. On the other hand, competition among bureaucracies for strictly finite resources tend to keep increases low. Changes in budgets are, therefore, incremental. In short, budget behavior depends largely on bureaucratic and organizational competition. 21 We are not thinking of the problem at hand as being reducible to a clear choice between the arms race and the incremental model of budgetary expenditures. It may tum out, of course, that both internal and external pressures are responsible for decisions on defense expenditures. Most people probably hold this view and would consider the "either-or" proposition as a red herring. The important question would really be, in this view, not whether international or domestic variables influence the pattern of expenditures but, rather, how much each of the internal and external influences contribute to the total result. We ourselves shall follow this approach. There have been a number of disparate attempts to probe the question of whether arms races or internal pressures shape decisions on defense allocation. Each attempt has produced suggestions but no final answers.22 What follows is an attempt to move the problem forward. Because we cannot observe directly mutual deterrence, we shall test which of these explanations is correct by monitoring whether conditions essential to deterrence are met. We shall test for the presence of the action-reaction process between the United States and the USSR, which we can observe directly and which we know to be essential for mutual deterrence to operate over a period of time. If this necessary condition for deterrence is not present, we can assume that the explanation given by mutual deterrence of the conflict behavior of nuclear powers is also inoperative.

185

Nuclear Arms Races and Deterrence

Indicators of Arms Competition We have selected as our indicator of nuclear capability the portion of each country's defense budget devoted to offensive strategic nuclear systems. Strategic nuclear weapons were considered those systems designed for use in total nuclear war. The indicator of nuclear capability used in the U.S'/USSR comparison to follow reflects only costs for the offensive strategic force: (a) the cost of delivery capacity of missiles, submarines, and bombers, (b) spending on development and procurement of nuclear explosives, and (c) the expenditure for installation and operation of nuclear strategic systems. 23 The objective is to exclude the costs of weapons which could be used in conventional wars or in tactical nuclear confrontation. As always, there is no problem at the extremes. The ICBMs and SLBMs are strategic weapons; the tank and the cannon are conventional; the neutron bomb is tactical. But clearly there also exists a class of weapons that can be used for both strategic and nonstrategic purposes-the B52 bomber, for example. The solution to the problem here has been to include in the strategic category all weapons systems that have a strategic function regardless of any additional conventional uses. We have also concluded that the offensive strategic capability of both nations reflects most accurately the deterrence portion of the budget. We excluded the costs of antiballistic missile systems because their military effectiveness has always been questionable; the systems have been only thinly deployed in the last decade and are now largely abandoned. We have also decided, as a measure of nuclear capability, to compare the costs of strategic weapons systems. This indicator is often used by intelligence agencies, not only because it simplifies the task of comparison (otherwise of unmanageable proportions) but also because, in all countries, costs are an important factor in determining the acquisition of nuclear arms. Indeed, top decision-makers generally have an imperfect knowledge of the technical merits of individual weapons but can effectively compare the efforts they are making with those of the opposition by

186

Chapter Four

using as a measure the amounts of money spent on specific weapons systems. For this reason, intelligence agencies provide summary cost figures rather than complex weapon evaluations to political decision-makers who want to compare weapon availability prior to decisions on overall weapons allocations. We chose to use constant dollar figures because they reflect consistently the amount of weapons capacity obtained over time, and already have a control for inflation. One should make explicit the assumption underpinning any attempt to estimate the nuclear capabilities of two competitors by looking at the budgetary allocations made by the two sides over time. The assumption is that anything done by one competitor can be matched by the other provided the latter is given a few years to make the required adjustments. Technological parity, then, is not an issue. What is required is that any increase in nuclear capacity realized on one side be matched by the other either through the development of similar technologies or by substituting for them other weapons systems with equivalent destructive power. For example, many observers argue that the United States is ahead in the accuracy, reliability and multiple targeting, of delivery vehicles, and in submarine technology, yet the USSR can offset these advantages with larger nuclear payloads delivered by a greater number of strategic launchers.24 As the reader is probably aware, estimates of expenditures for strategic weapons for either the United States or the Soviet Union are most difficult to come by and hard to compare. Distortions in such comparisons are inevitable. The consequences of such distortions, however, have been reduced as much as possible because of several related factors. Confining ourselves, as we do, to the component made up mostly of hardware and technology, and by using constant rather than current prices to avoid distortion produced by inflation, we were able to minimize distortions produced by differential pricing. This is an important point. The estimates for the Soviet Union were reached by the device of estimating what the USSR weapons systems would have cost if they had been produced in the United States. This

187

Nuclear Arms Races and Deterrence

procedure is in large part necessary to avoid the problem produced by the shroud of deep secrecy surrounding all arms-building activities in the USSR. In addition, because the economic systems and the accounting procedures of the two countries are so different it would be incredibly difficult to compare defense expenditures in any other way. 25 Moreover, the cost estimates that we used reflect directly the nuclear capability deployed rather than what might have been spent on weapons systems that were found wanting and had to be abandoned. We are aware that distortions are only mitigated. 26 It should be kept in mind, however, that it is precisely the distorted estimates of Russian allocations that the American decision-makers consider when planning United States nuclear defense spending (and that Russian leaders do the same); and that what we are after is to see whether there is a consistent systematic action-reaction process going on. Therefore, minor distortions in expenditures, so long as such distortions are consistent, do not affect our work. Comparison of Russo-American Investments in Strategic Systems The first question we need to pose is whether changes in Soviet outlays for offensive strategic systems in the years 1952-76 are directly related to changes in outlays for the same purpose in the same period by the United States. There are, logically, only two possible sources of pressures that can account for increases in the expenditures of the two countries for strategic weapons: pressures from outside and pressures from within each nation. Thus we constructed a simple model in an effort to probe the existence of an action-reaction process between the two countries and the presence of internal pressures within each of them. One critical issue required solution. How long an interval could be allowed to elapse before the expenditures of each contestant should no longer be considered a response to the previous expenditures of its adversary? One could argue that such intervals should be kept short. During this interval the retaliatory power of the defendants or deterrence, is compromised and a potential aggressor is tempted to strike

188

Chapter Four

first, finishing off at one blow its competitor and the competition. The whole concept of mutual deterrence is based on the need to keep an adversary from gaining such an advantage. Ideally then, in testing for the presence of competition or arms races, one should set a very short interval between action by one contestant in its arms-building program and reaction by the other. One should be justified in stipulating that whatever effort one competitor made in one year the other competitor should try to match in the next. Proponents of deterrence would certainly argue, however, that any expectation of year-by-year responses envisions too short an interval and may not be a fair test of the presence of competition, given the time necessary for detection, evaluation, development, and deployment of nuclear weapons. But if immediate year-by-year responses are not a fair test, how long should the interval be? We thought that all intervals up to a maximum of five years should be fair, indeed generous. It seems reasonable to argue that in the context of nuclear arms any alleged responses delayed for more than half a decade after the initial move of the opponent could be taken to mean that competition, or at least effective competition, is not taking place. Thus we provide that our analysis would scan the interactions, if any, for lags in the reaction of each contestant to the expenditures of the other at any time between the first and the fifth year. We shall return to this point when we discuss the results we obtained. We set down two postulates. We argued that decisionmakers react in fear of the total strategic arsenal of the opponent. This postulate differs from Lewis Richardson's in that he tested simply to see whether changes in strategic capabilities of one country were followed by changes in the other nation. Our assumption appears more in keeping with the notion of deterrence because fear is fueled by the possible use of all the adversary's stockpile, not only by the new weapons added periodically to existing stocks. We still assume, however, that the decision-makers in both countries are aware of the decisions taken by their counterparts and that they can adjust budgets up or down to match. Our second postulate deals with the effects of internal pressures on expenditures for strategic arms. Two very dif-

189

Nuclear Arms Races and Deterrence

ferent types of elements may be involved. First, a number of factors have been suspected, over the years, to be accountable for increases in strategic budgets: the process of research and development that creates new weapon technology-when new weapon technologies are developed it is hard to resist putting them to use; the politicoeconomic complex responsible for decisions to acquire new weapons; changes in national leadership; the annual scramble of the military services for resources; and so on. One cannot be certain, however, precisely which factors really influence allocations for defense, and why their pressures work. The second element of the internal component is depreciation. The stock of offensive strategic weapons, like any other commodity, depreciates over time and needs to be refurbished, updated, or replaced to maintain efficiency and effectiveness. We will assume that this depreciation can be represented at a constant rate over time and across the two nations and that the current value of all the stock of arms is equivalent to the present expenditures plus the value of all the previous years after depreciation has been taken into account. This assumption is critical for what we are going to do, but it should be clear that the assumption can distort reality in two ways. It fails to take into account the sharply uneven way in which technological breakthroughs render obsolete one system while hardly touching another. For example, the Minuteman has remained since the sixties, with modifications, the major missile delivery system of the nuclear force of the United States. On the other hand, the technology of the fighter plane has undergone drastic revisions, and no standard aircraft has emerged. Second, the depreciation of the American and Russian stocks of strategic weapons is not necessarily the same. Obviously the composition of the two strategic arsenals is different, and components differ in cost, longevity, and utility. In the United States, for example, the bomber fleet is more expensive to run and to replace than that of the Soviet Union. The USSR has deployed many more types of ICBMs than the United States, but some Soviet missiles are more or less obsolete.27 These distortions are real enough. We assumed, however, that equal depreciation rates for the two countries

190

Chapter Four

are a reasonable reflection of the overall reality because both nations maintain such large and varied systems that the distortions of individual systems should tend to cancel each other out. The main effect of this assumption is that the nation with the larger stockpile of strategic weapons is affected more drastically than its competitor. Formal Representation of Arms Allocations Having described the components, we can now put forward a dynamic model that seeks to account simultaneously for the influence of internal and external pressures. Consider the following general equation for the Soviet Union which, of course, can be replicated exactly for the United States: (1.0) USSR Off. Exp. (t)

+ al USSR Off. Stock (t - 1) + f31 U.S. Off. Stock(t - 1) + e(t)

= "Yl

The first component represents the influence the United States stock of arms may have on the allocation of arms by the Soviet Union, and the second monitors the impact of the Soviet's own stock of arms on the Soviet new allocations. Since the American decisions on strategic expenditures are announced in July, while the Soviet Union makes all such decisions in January, the expectation is that the Soviet Union, in allocating resources to the strategic system, would respond to the previous year's allocations of the United States, while the allocations by the United States would react to the current year's decisions made in the Soviet Union. The two equations have, therefore, a slightly different time frame to compensate for the half-year's difference in budgetary periods. Our model cannot be estimated as it stands, but it can be reduced to a statistically tractable form. The resulting equations, which account for the behavior of expenditures for strategic arms of the United States and the Soviet Union, can be written in the following manner: 28 (2.0) U.S. Off. Exp. (t)

= "Yl(1 - A) + (al + A) U.S. Off. Exp. (t - 1) + f31 USSR Off. Exp. (t) + e(t)

191

Nuclear Arms Races and Deterrence

(2.1) USSR Off. Exp. (t) =12(1 - >-) + (a2 + >-) USSR Off. Exp. (t - 1) + (32 U.S. Off. Exp. (t - 1) + e(t) The decomposed parameters are: >-

= depreciation parameter assumed constant for both nations

11,12 = constant effects parameter for U.S. and USSR

a Io a2 = internal effects parameter for U.S. and USSR (31' (32 = external effects parameter for U. S. and USSR A number of comments are in order. The external interactive effects (f3) can be estimated directly and are not confounded by internal factors. However, the effects from internal and external sources cannot be disentangled from one another without some manipulation, because, in the models we use, they have to be estimated simultaneously. The constant term (1) can be estimated only in conjunction with the depreciation rate (1 - >-).29 Moreover, the composition of the internal component gives us some difficulties. We obtain an internal component that combines the effects of depreciation and the pressures from the politicoeconomic coalition (a + >-), but we will attempt to estimate the value of each of these components in our analysis. Results Let us now turn to what we have found. We estimated first the reduced equation for all the years of nuclear competition between the United States and the Soviet Union; because the first successful explosion of a nuclear device by the USSR occurred in 1950, we started our analysis with 1952. The results are presented in table 4.2.30 The most dramatic finding is obviously the fact that the expected interaction between the two countries does not account for much of the variance explained. The external component is never significant in the case of the Soviet Union, nor is it significant for the United States. Moreover, even if we dismiss the problems of significance and strength of association, the impact of the external component is

192

Chapter Four

Table 4.2

Effects of External and Internal Pressures on U.S. and USSR Strategic Budgets, 1952-76 U.S. Off. Exp.m = 7.095 +.639 U.S. Off. Exp.et-ll - .345 USSR Off. Exp.(tl Standard error (3.550) (.165) (.196) Significance (0.058) (.001) (.094) R2 = .72 Standard error = 2.82 Significance = .000 N = 25 USSR Off. Exp.(tl = 1.835 + .933 USSR Off. Exp.et-Il - .057U.S. Off. EXp.et-11 Standard error (1.242) (.071) (.059) Significance (0.154) (.000) (.347) R2 = .95 Standard error = 1.03 Significance = .000 N = 25 The models are stable, and no significant autocorrelation of residuals appear in either version; the independent variables are not multicolinear. 31

NOTE:

always negative. In effect, one country actually reduces its expenditures for strategic weapons when the other nation increases theirs. And this is precisely the inverse behavior one would expect if the two countries were competing. Because the internal component accounts for almost all of the variance explained by the model, we must conclude that internal factors determine expenditures on nuclear weapons. We conclude therefore, that no arms race was waged, that the two nations were scarcely competing. Our general finding supports the view expressed recently by Bernard Brodie, the originator of the conception of deterrence: The numbers of those [strategic] forces, incidentally, grew during the nineteen sixties like the British Empire was said to have grown-in a series of fits and absentmindedness. There are reasons why the number 1,000 was chosen rather than a lesser number of Minuteman missiles, in addition to our 56 Titans, and also why we chose to build 41 Polaris-Poseidon submarines capable of firing 16 missiles each, in addition to the 400 plus B-52s we had at the time, not to mention the quick reaction alert forces we had in Europe. But whatever these reasons were, they were not in response to Soviet figures. 32 It is very important to note that the results we report, the best we have obtained, occur when the interval between action and reaction is shortest, that is, when we find out

193

Nuclear Arms Races and Deterrence

whether allocation of American resources one year are affected by the Russian allocation in the preceding year, and vice versa. In short, one-year lags produced our best fits. As we indicated we would, we tested for delayed reactions of two through five years and all such results were even less significant than the one we reported. The argument that the United States and the USSR are competing but that we could have missed the pattern of action and reaction because it occurs at a slower tempo than our model can detect is not correct. We should warn the reader that it is possible that we cannot detect interaction between the two nations because the data are available for too short a period or because the United States, secure in an early, strong advantage, has waited to compete until this advantage has been erased. But if America and Russia were bitter enemies during this period why did the United States not try to keep the nuclear advantage it had? Why would a race start when levels of nuclear destructive power had passed the point of overkill many times? Such questions cannot be answered with the data we have. Our second most important finding is that strategic offensive expenditures are strongly related to each country's stock of weapons. Note that in the sample the internal component, that is, depreciation and the politicoeconomic coalition pressing for increases in the allocation to the strategic arm, accounts for 95 percent of the variance in the case of the Soviet Union and 72 percent of the variance for the United States despite the disturbance produced by a few unusual fluctuations of the data. In a restricted sample (excluding 1952-1954 and 1965) the variance explained for the Soviet Union remains virtually unchanged, but increases to 91 percent for the United States. 33 Recall that the outliers are removed for purely statistical reasons (and we do not want to overemphasize the results), but the reader should note that the levels of variance explained with or without the outliers are remarkably high. Internal factors then are the overwhelming source of pressures shaping decisions on strategic budgets, and it is most important to pin down, as much as possible, the

194

Chapter Four

strength of the two elements making up the internal component. Recall that the internal coefficients contain internal factors and the depreciation rate. If we assume that depreciation is constant, we calculate the difference between the level of internal pressures of the United States and USSR to be approximately .2, or 20 percent of the strength of this coefficient. It is of interest that the overall internal component is more influential in the USSR than in the United States. But can we go further? Can we try to estimate the value of each of the two elements making up the internal component of our equation? If we succeed in estimating one of the two components, we will know both. Depreciation gives us our opportunity, although estimating depreciation without access to information about individual weapon systems, their maintenance, their durability, their numbers, and so on, is uncertain business. Very rough and tentative estimates might be possible however. The American experience with its own arsenal suggests that strategic arms are completely replaced or abandoned in fifteen or twenty years. 34 If this is true, an approximate and crude estimate of depreciation would be .75. We can now go on to evaluate the relative levels of influence in the two countries of the political-economic-bureaucratic coalitions by subtracting the depreciation from the overall internal coefficient. In the Soviet case the strength of the coefficient is .92. And since depreciation is .75, and accounts for 81 percent of the total, the internal pressures contribute .17 and must account for the remaining 19 percent of the coefficient's value. For the United States the internal component is .72 and internal pressures equal -.3 and account for 2.5 percent in excess of the value of the coefficient. The findings are clear. The USSR trend is accelerating, the American trend is decelerating ever so slightly, but the United States is still ahead due to her original advantage. This finding, too, is arresting. The Russian coalition is more effective than its American counterpart in influencing the allocation of resources to the strategic arm. Perhaps in view of the extremely high level of bureaucratization in the Russian economic and political systems, such "better" Soviet performance should not surprise us.

195

Nuclear Arms Races and Deterrence

As we have already noted, the results ofthe analysis point clearly to the conclusion that the Russians do not change their allocations to strategic weapons systems because of what the Americans have decided to do and that the Americans similarly disregard the Russians. On the evidence so far, one must reject the prevailing notion that there is an arms race or a direct competition between the two nations. However, could it be that the cyclical trend in the American data obscured interaction between the United States and the Soviet Union expenditures and that our results were affected by our insistence on fitting data with cyclical characteristics onto a linear model? What would happen if we detrended and reanalyzed results? We explored the problem through Fourier analysis specifically designed to pick up systematic cycles in a series. 35 In the case of the Soviet Union not much help from the Fourier synthesis could be expected (fig. 4.3). The Soviet pattern is still linear with a slight but consistent increase over time. The case of the United States is different (fig. 4.4). A systematic and consistent cyclical pattern is detected, and it signals that something other than a simple desire to replace depreciated hardware and concurrently to increase allocation is at work on the American scene driving - - Standardized Observed Values: Mean =100, SO= 20 - - - Trend Component: FF=I, 100% Prewhitening. RSQ=.95 140r------ - - 120

_._- - - -

- - - - . - --::--

--------------~-

100

80 -

~--...- - - . - - - - - - - - j

_

~---------.-

60e-- - - - - - - - - - - - - - - - - . - - -

1950

Fig. 4.3

I

I

I

I

I

1955

1960

1965

1970

1975

Fourier synthesis: USSR Strategic. The estimates for the Soviet Union are relatively unsophisticated compared to those for the United States. The lack of variation from year to year may, in effect, be due to the method of estimation. When more accurate estimates will be available, it is possible that year-to-year variations will appear in the Soviet series.

NOTE:

196

Chapter Four

expenditures up and down. 36 The pattern also forecast accurately the increases in weapons allocation after 1974 that had remained undetected by the linear model. Most important of all, the analysis matching the detrended patterns against one another yielded an R2 of .09, and the direction of the interaction was negative. Results were proved lower when time lags were applied, and there was no sign of interaction. The two trends had once again clearly proved to be almost independent of one another, and the slight relationship encountered was in the inverse of the direction predicted. A final question was left in regard to the cycles in the American expenditures. Did the high and low points in the single cycle shown coincide, precede, or follow in a systematic manner the kind of international events which could serve to account for the fluctuations in American expenditures for nuclear weapons? For example, did expenditures on strategic weapons decrease when defense resources were being absorbed by conventional capabilities in conventional wars, and did they increase in relation to the danger of nuclear war? The fact is that they did not. The rise and fall in American expenditures had little to do with external events. Systematic analysis is difficult, but a few examples will help make the point. During the Korean War (1950-53) the expenditure on strategic capabilities was extremely high but during the Vietnamese conflict (1965-74) it was very low. The rise in tensions produced by the Cuban crisis was fol---Standardized Observed Values: Mean =100, SO =20 ---Trend Companent: FF=I, No Filters RSQ=.75 140------- -- ---- -

-

-

-

-

::~~~~

----=====~

n~/_

- .. , ~

___

80

---- ---

'''0-

--

40r--- - -

--

I

1950

Fig. 4.4

1955

1960

Fourier synthesis: U.S. strategic.

1965

1970

1975

197

Nuclear Arms Races and Deterrence

lowed by a decline in allocations, while the decline in tension following the conflict in Vietnam was followed by an increase in strategic expenditures. The international actionreaction process, if there was one, played no role. Levels of Investment Efforts in the Strategic Force We have now tried twice to test the central hypothesis and both times our findings plainly indicated that the conditions essential for the existence of competition or for an arms race do not obtain. Such results are so unbelievable that we will try once more. The problem can be approached in still another way. One can monitor the levels of effort the two countries make in securing strategic arms, and if such levels rise and fall in some systematic relation to one another, one can argue with some justice that there is evidence of interaction between the United States and the USSR in the field of strategic nuclear weapons. We estimate levels of effort in three different ways: by considering expenditures on strategic systems as a percentage of the total defense expenditure, as a percentage of total product of the country, and as a fraction of the per capita product of each country. 37 The comparisons we made are summarized in table 4.3. Table 4.3

Levels of Effort by the u.s. and the USSR in Developing Strategic Offensive Capabilities Years 1951-55 1956-60 1961-65 1966-70 1971-75

Off. as% of Tot. Def. Exp. U.S. USSR 12.2 17.7 13.4 5.9 6.8

2.9 6.6 13.2 12.4 12.7

Off. as% of Total GNP U.S. USSR 1.7 1.7 1.2 0.5 0.4

1.5 2.1 3.1 2.6 2.5

Off. as % of GNP per Capita USSR U.S. 2.3 2.3 1.6 0.6 0.5

2.1 2.9 4.9 3.6 3.4

The first two columns in the table give percentages of the total defense budget allocated for offensive strategic capability. A glance up and down these columns tells the following story. The USSR and the United States pursue courses diametrically opposed to each other. The United

198

Chapter Four

States moved from a high investment level of between 12 and 17 percent in the first three quinquennia of the period studied to roughly half that amount in the last decade. The Soviet Union, on the other hand, began low in the first decade and then doubled the investment in the last fifteen years. The same inverse relationship is found in all other measures of effort presented in table 4.3. On the whole, it is clear that the countries move in opposite directions. Turning to offensive expenditure as a percentage of GNP, we see that the United States and the USSR began by investing the same fraction of the total GNP in offensive strategic weapons, but the first quinquennium represents the high point for the United States and the low point for the Soviet Union. From then on both nations proceed in linear fashion. In the seventies the USSR is spending 2.5 percent of total GNP on strategic weapons systems while the United States is spending less than half of 1 percent. In view of the pattern that emerges when one matches investment in strategic weapons with total GNP, one should not be surprised to discover that the percentages of per capita GNP invested in strategic systems should move apart as well. The United States and the Soviet Union began by extracting roughly the same fraction of per capita resources, the United States at 2.3 percent and the Soviet Union at 2.1 percent, and ended at 0.5 and 3.4 percent respectively. A number of interesting points are imbedded in the figures in our table. First, there is no question but that the Russians are trying harder. Table 4.3 demonstrates unequivocally that, whether one measures effort by resources invested in strategic weapons as a fraction of total GNP, or of per capita GNP, the Russians are clearly spending more of the smaller amounts of resources available to them on strategic arms. We should note, however, that although the United States is hardly trying at all, her effort in the 1950s, conceived as a percentage of her defense budget, equals the Russian effort in the 70s. Further, in the last decade the USSR has invested considerably below the level reached between 1961 and 1%5. Columns 1 and 2, on the other hand, give a somewhat different picture than that established by the data in the other portions of the table. Here we observe what the mili-

199

Nuclear Arms Races and Deterrence

tary establishment of the two countries decided to do with the resources allocated for defense. It should be kept in mind that these are percentage shares of different absolute totals. What is interesting is that the lows and the highs reached in the investments made by the two countries are very similar. One sees this more clearly if one compares the percentage of the defense budget spent on strategic allocation in the United States from 1951 to 1960, contained in the first two cells of column 1, with similar expenditures by the USSR from 1966 to 1975, contained in the last two cells in column 2. Would it not be fair to state that both countries have made about the same level of effort, but that they made it at different times in the period? The Americans decided to do it at the beginning; the Russians decided to do it at the end. There can be no doubt that the data are suggestive of subterranean processes which would only become distinctly visible if we could disaggregate the series we have. On the other hand, the data are explicit on one central point: there is no process of interaction visible from the patterns of efforts analyzed. Because of the absence of this necessary condition there can be no direct arms race. These findings are incredible and fly in the face of all existing assumptions, but we have merely followed where our data led. Conclusion

We have been concerned in this chapter with a question of great importance. According to our data, the presence of a nuclear arms race, far from constituting a given of international politics, proves to be a chimera. We have tried again and again to test for the presence of arms competition or an arms race and each time we have failed to find anything. It is obvious that the United States and the USSR are building nuclear arms, but are not doing so, as they allege, because they are racing or competing with one another. The absence of competition between the two countries leads to the startling finding that logical conditions for deterrence are absent, and by inference to the conclusion that mutual deterrence is not taking place. This is shocking. Our finding

200

Chapter Four

does not fit with notions of the way the two contenders function and how they interact. Is the behavior of great powers involved in international conflict in the nuclear era fundamentally different from what we have found it to be in our investigation of major-power war in the pre-nuclear age? Before nuclear weapons appeared on the scene every important aspect of war among major powers-i.e., its beginnings, its immediate outcomes, and its ultimate consequences-seemed to be tied to national development and the differential rates of growth of the major actors in the system. The question we asked was whether nuclear weapons had, in fact, changed all that, as they are alleged to have. We have determined that they have not. As in the past, armaments and the tensions they reflect are the consequences of internal processes. External threats may serve as the initial mechanism of legitimation that permits a nation to begin arming in earnest. Once the armament program has begun, however, external threats are of little importance in its continuance. Moreover, nuclear arms still cannot erase differences in socioeconomic and political capacity any more than their conventional predecessors could. They are not the means by which to redistribute the total strength in the system among competing nations as it was dreamed they would be. They are not the tools that give leaders the long-sought control over their nations' destinies and over international peace. They are not built in relation to what the opponent is doing. Finally, the miracle quality that nuclear weapons were alleged to have, turned out, as miracles often do, not to be real. Deterrents do not deter. Nuclear arsenals are no more than what they were supposed to be in the first place, and would have been clearly seen as being had we not had an interest in conceiving them in the perspective of deterrence (the one palatable justification for their construction) as terribly frightening weapons waiting to be used. Our findings disturb the common wisdom on nuclear international politics in yet another way. If one believes that nuclear mutual deterrence is operative, the horror with which the elites and mass publics of the United States have

201

Nuclear Arms Races and Deterrence

regarded the proliferation of nuclear capabilities to the USSR, then to France, then to China, and finally to India, with other nations waiting in the wings, represents the reaction of people whose anxieties have outpaced their capacities to analyze the problem. Although the possibility of accidental war increases with the spread of nuclear weapons, it is also true, if one believes the doctrine of deterrenee, that the spread of nuclear weapons to many countries should also spread stable deterrence. 38 It does seem a logical contradiction to believe in the doctrine of nuclear deterrence and at the same time to regard nuclear proliferation as having no benefit for peace. On the other hand, if our conclusions are correct and nuclear arsenals are built almost solely as a result of internal pressures rather than external factors, and deterrence is not operative, then nuclear proliferation is dangerous to peace, is rightly feared, and should be stopped. Finally, how is one to make sense of the finding that strategic arms races are really not taking place and that the explanation given by the theory of mutual deterrence for the building of arms and the keeping of the peace is invalid? No present answer is complete, but there is a plausible explanation. There is evidence that the various U.S'/USSR experiences of discovering each other as bitter competitors after World War II were never properly analyzed or understood by either nation and that the resulting anger, bewilderment, and suspicion have dominated the relations of the two countries from that day to ours. Each country identified the other as the enemy, and this coding amounted to a license for strong, continuing strategic arms buildups. How and why these original impressions have been passed on from one generation to another, how each new set of leaders was socialized into seeing the world with their predecessor's eyes, regardless of the way it really was, is not really known. Moreover, the American side provided a rationale that turned the purpose of strategic arms completely around. Strategic arms were there not to fight with but to keep the peace. Be that as it may, from the late 1940s to the present there has been no letup in the perceived necessity to acquire strategic weapons. No analysis has been done to see

202

Chapter Four

whether beginning or continuing to arm is or was really appropriate to the circumstances of the time, or whether strategic arms programs were or are really matching the strategic punch on the other side. The builders of strategic arms have operated in almost totally closed systems. Perhaps the discovery that they have not been interacting would not upset them. In the absence of competition, the furious stockpiling of nuclear arms is not easy to evaluate. While it may not be appropriate to think of the process whereby the two countries acquire strategic weapons as "neurotic," it is the term that comes to mind. "Neurotic," then, may have to do. lt is all very different from what one expected. Some readers will find these findings and conclusions provocative, even offensive. They were not meant to be. Our findings fly in the face of intuition and the major assumptions of the international political thinking of our time. Perhaps we are wholly wrong. But what data we had and what data there were spoke clearly. We followed them. One feels a sense of foreboding in view of the massive opinion opposing our conclusion. To such an opinion one is tempted to repeat the entreaty of Oliver Cromwell to the General Assembly of the Church of Scotland in 1650, "I beseech you, in the bowels of Christ, think it possible you may be mistaken."

Five

Conclusion

It is now time, before bringing our ledger to a close, to

review the findings of our previous chapters and weave together the most important threads spun out in the preceding pages. We had many goals in mind while composing this volume and all of them have, in different degrees, gone into shaping not simply each separate part but the whole they collectively describe. A second section of the conclusion will contain a brief review of our major findings. And the coda to this essay will also bring to the attention of the reader some new hypotheses which are only implied by our findings but which we consider intriguing enough to be made explicit. We offer them as points of departure for new research: drawing up a research agenda is always exciting business, and we consider it proper to conclude in this manner. These new ideas do not really deal either with international politics or with war as such but offer, rather, a fresh angle of vision on the operation of our own system and that of our adversaries, one in which the significance normally accorded politics and war is drastically altered. Ours cannot be regarded a final statement of the problem. It is, we hope, the best that can be done at present. It would be immensely flattering if any of our readers wished to pursue, amend, or even disagree with what we have written, still more so were someone to choose some of the new questions as a topic for future research. For it is a mark of special success when one book gives birth in this way to another. A Note on Architecture

203

A note on the way we have organized this book is now in order. The decisions we took on how to proceed at different times during its composition should be made explicit in order to reveal the underlying strategy of our presentation.

204

Chapter Five

Our substantive interests lay in determining the origins, outcomes, and consequences of major wars; in elucidating the role of national growth in such processes; and in defining more exactly the nature of national power. Past theory and research alleged power to be the critical variable in the way wars occurred. Moreover, it has always been thought that, inter alia, wars were fought in order to preserve power and, if possible, to increase it. We wished to test the propositions contained in such beliefs and try to account for crucial portions of the behavior of such conflicts at critical points in their evolution. The key to our research strategy proved to be the connection between national growth, power, and war. Growth is the source of power. The distribution of power and shifts in that distribution shape the evolution of conflicts. We began with a review of existing propositions concerned with the way in which the patterned changes occurring in the distribution of national capabilities across nations and over time generate and harden the conditions that lead to war and render the movement toward it irresistable. We then went on to consider traditional conceptions of the manner in which, having brought the war about, such changes prove decisive in terms of its immediate outcome and its longrange consequences once peace has been reestablished. We next advanced a number of rival hypotheses that we proceeded to put to an empirical verification. We also tried, as the reader will remember, to explore the possibility of using wars as tests to gauge national developmental progress, taking advantage of two opportunities for such analysis. Our final chapter was an attempt to test whether nuclear power had replaced growth as the most important variable in international conflict, thus transforming the nature of international politics; or whether, on the contrary, the role of nuclear weapons could only be correctly understood as a function of these very same patterns of growth. The sample of conflicts we used was very small. Our theoretical interests dictated that we concentrate for the most part on major wars, although not exclusively. Had it been possible, it would have been best to organize this book in such a way as to test all our propositions by using the

205

Conclusion

same (major) wars and the same power measures. But major wars and existing measures of national power, though sufficient for the purposes of studying the origins and the long-range consequences of war could not be used to answer our questions on the outcomes of war. Studies of the beginnings, outcomes, and consequences of war permitted different tolerances of error in the estimation of national capabilities. Such tolerance was highest when the question was of war consequences. Tolerance fell, on the other hand, when we turned to the consideration of the effects of the distribution of power on the outbreak of war. Indeed, available measures in this case proved barely adequate. They were completely inadequate, however, in any attempt to account for the outcomes of war. Better models, including a direct measure of political capacity, were essential. We were not unaware that a valid empirical measure of the capability of political systems, if one were to be developed, would have an application well beyond its decisive contribution to a model of national capabilities and would be of the greatest theoretical importance. The elaboration of such a measure may certainly be considered the most important contribution in this book. But as the reader will remember, the measure that was constructed could not be used for estimating the political capacity of developed countries; that task remains to be done. And although, as we found, political capacity plays a crucial part in determining who will win a war and who will lose, the absence of a measure of political capacity appropriate to developed nations made it impossible to employ the measure we did construct in the study of major wars, since the protagonists in major wars are the developed nations forming the central international system. We had no option, then, but to confine ourselves to lesser wars, that is to conflicts in which most of the principal actors were developing countries. These wars, as the reader also knows, provided perfect experimental cases for testing whether the new measures of political capacity and the model of national capabilities were in fact valid. Success here represented something of a breakthrough. In any event, the shift from major to lesser wars, and from one type of measure to

206

Chapter Five

another and then back again, met the requirements of the theoretical propositions we put forward. Finally, although we may have had some success in identifying some of the rules governing behavior in an important class of conflicts, a critical question remained. Were the rules still the same in the nuclear world? Was the connection of growth, power, and war still the same as before? Let us now tum to a summary of our findings. Major Wars: Beginnings There are marked regularities governing changes in the distribution of national capabilities across countries and over time and the inception of major wars. Two different patterns are involved, however, for the behaviors of the dominant nation and the challenger (the two strongest nations in the system) and the behavior of the major powers (that class of countries on the next rung down the ladder of power) are quite distinct. The dominant nation and the challenger are very likely to wage war on one another whenever the challenger overtakes in power the dominant nation. It is this shift that destabilizes the system and begins the slide toward war. The speed with which the challenger overtakes the dominant nation is also important: the faster one country overtakes the other, the greater the risk of war. These are necessary but not sufficient conditions, however, for unless the challenger actually surpasses the dominant nation, war will not break out. Moreover, the challenger may also surpass the dominant nation without fighting it. Wars, therefore, are not inevitable. The case of the major powers, on the other hand, is different from that of the contenders. Their behavior is not shaped by rates of growth. They find themselves drawn into wars, rather, by the network of legal and other ties that bind them (like all nations in coalitions) to one of the two leaders. Fighting begins, then, as a result of differential rates of growth between the contenders; and it subsequently assumes the proportions of a world war because of the obligations the major powers in each coalition have toward their respective leaders.

207

Conclusion

Long before the nuclear age, the contenders were the only sovereign countries in the system: they alone possessed freedom of choice whether to fight or not, and they alone considered only their own interests in making this choice. The major powers, who generally had been perceived to enjoy a good deal of latitude in making such decisions, appear instead quite bound by the wishes of the contenders. The pre nuclear international system, then, appears in the past to have been as tightly controlled by the superpowers of the time as the nuclear international system is today by the nuclear superpowers. The respective structures of international power obtaining in each of the two eras prove, consequently, not to be as different as is often alleged. All that seems really new in the nuclear age is the recognition and the acceptance of the realities that have already long existed. The role of alliances requires a further word of comment. Alliances not only playa decisive part in widening a narrow conflict between the challenger and the dominant nation into a world war, they also exert a decisive role in winning such wars once they are started. Even after the challenger has overtaken the leading nation in power, the challenger's coalition remains decidedly weaker than the alliance commanded by the dominant nation. It is precisely this situation that may, indeed, furnish the determining motivation pushing the challenger to resort to war in the first place. It is the stronger country, and it wants immediately what it believes is its due by right of strength, refusing to take time to seduce supporters away from the old leader. This is a strategic error. The outcome of the fight is settled by the greater strength of the coalition supporting the established leader, and the challenger and its friends will suffer defeat on the field of battle. We should repeat here the warning we gave the reader at the beginning of Chapter 1. Our data can only lead to conclusions about the necessary but not about the sufficient conditions of war, and necessary conditions alone cannot suffice in predicting whether wars will in fact take place. In order to make such predictions, one would also need good data on how those who have the power of deciding peace and war for their nations make their choice.

208

Chapter Five

Predictions of War Outcomes

But can the same behaviors responsible for bringing contestants within fighting range of one another be used to account for who will win once the fighting starts? It seems clear that the core of the problem lies in developing an accurate way of measuring the strength of nations. Accurate measures of the strength of nations would need to estimate not only the pool of human and material resources available to potential combatants but also the capacity of the combatants' respective political systems to utilize the resources available to them. The degree to which nations can mobilize what resources they possess is in large part a function of the way the political system performs in extracting these resources and directing them to their intended targets. The omission of a direct measure of political capacity may be tolerable in cases where the conflict involves economically developed countries and where economic development has universally moved in tandem with increases in political capacity. In such cases economic performance adequately reflects political capacity, a factor not itself susceptible to direct measure. But in the case of economically underdeveloped countries, low productivity must not be interpreted as an indication of equally ineffective political systems. In some developing countries, though the total pool of resources would be reduced by a fall in the productivity of the economy, the proportion of the total resources available to the government will vary a great deal depending on the efficiency of the political system. Thus, in developing countries, any estimate of overall power may miss the mark quite substantially unless it includes a direct measure of the capacity of the political system. Political capacity was thought to be composed of three highly interrelated elements: the level of penetration of governmental power into the national society, the capacity of the governmental system to extract resources from its national society, and, finally, the performance of the government in delivering such resources to their intended ends. Of the three, the second element was judged the most critical

209

Conclusion

for the purposes of the indicator we wished to construct, and we turned to revenues as our measure of political capacity. This measure cannot be applied to developed countries. We wanted to resolve a fairly simple problem: compared to all other countries in the sample, how well is this government doing in collecting resources from its people? One needed to be certain, however, that performance could legitimately be interpreted as a measure of capacity. In underdeveloped or developing countries, need far outruns the capacity to produce. In their case, therefore, the need of governments for resources is so great that the only limit on the extraction of resources is the capacity of the government to collect revenues. In developed systems, on the other hand, need and capacity are roughly in balance with one another and, consequently, performance in regard to collection of revenues can no longer be taken as prima facie evidence of capacity. This new index was then included with indicators of economic and demographic resources into a full model of national capabilities, with foreign help to a combatant being considered part of the resources available to the recipient, and, consequently, adjusted by the recipient's capacity to utilize the resources at its disposal. We chose as our test cases the Korean and Vietnamese wars, the Arab-Israeli wars of 1967 and 1973, and the SinoIndian conflict. The results of the first four conflicts could not have been predicted with existing measures. The new model, on the other hand, performed beyond every expectation. It traced the outline of the fighting with amazing accuracy in all four conflicts. In the Korean War, when both the Americans and the Chinese had joined the fighting, the new measure estimated both sides to have been roughly equal in strength and the Sino-American stalemate in the field of battle answers perfectly to this estimate. In the Vietnamese war, the North was found immeasurably superior to the South when the United States was not on the scene, and the easy victories won by the North Vietnamese both in the early and late phases of the war fit that estimate exactly. On the other hand, once American forces entered the fighting in Vietnam, the two sides emerged as being

210

Chapter Five

roughly equal, the South Vietnamese and the Americans on the one side and the North Vietnamese and the Vietcong on the other proving unable to dislodge each other. Once again we find exact confirmation in the actual results. Again in the case of the 1967 and 1973 Arab-Israeli wars, when the Arabs and the Israelis fought alone, our estimate of the distribution of strength between the two sides shows the Israelis and the Arabs to have been roughly equal in 1967 and the Israelis much stronger in 1973. If one recalls that, in 1967, the Israelis surprised the Arabs, while, in 1973, the Arabs caught the Israelis off guard, such estimates of the relative strength of the two sides, though contrary to what popular intuition would have suggested, make in fact a great deal of sense. In the Sino-Indian conflict, both the new and the old estimates show China to have been stronger than India at the time of hostilities, and this, of course, finds corroboration in the crushing Chinese victory over Indian forces. The new model, however, judges China to have been much stronger than estimated with the available measures. In view of the total superiority of the Chinese when the two clashed, the absolute control by the Chinese of the field of battle, their complete capacity to initiate and terminate the fight, it is our view that the more recent calculations are far more accurate than the old ones in estimating the fighting capacity of the two sides.

The Phoenix Factor The third facet of war we explored has to do with the consequences of war in terms of power. In power terms wars could "make sense" if, and only if, superior powers could tum back their challengers, or inferior powers reduce their superiors, by a military defeat. Our major theoretical concerns could be translated into the following questions. Does the outcome of a war determine its ultimate consequences? If a great power is on the victorious side in a major conflict, will this assure its superiority in the future? What will happen after the war to the pool of resources from which the victor will have to draw its power in relations with other

211

Conclusion

countries? Does the pool of resources go up, stay the same, or go down? And even more important, if a country loses a major armed struggle, will this put a decisive end to its pretensions"! More precisely, is this nation reduced in her power resources in any permanent way? As the reader will be aware, these questions go to the heart of all the suppositions underlying what has been written in the past, and what continues to be written in the present, on the subject of war, security, and the international distribution of power. The answers we needed required some complicated calculations. We again used GNP as our measure of power, as we already had done in our discussion of the beginnings of war, and for the same reasons. Most of the participants in the wars used as test cases were developed countries, and where they are concerned GNP is the most adequate and most readily available estimate of the pool of power resources at their disposal. The model we settled on for the estimate of changes in power as a consequence of war, permitted us to extrapolate the antebellum performance of a country into the postwar period. The extrapolation served as a trace of what would have happened to the distribution of power had no war occurred. The expected and the actual performances of each country could then be compared. Our choice of the wars used as tests and of the countries regarded as actors presented complications. We decided to consider only the most lethal wars because they were the ones having the most pronounced consequences. The best guarantee against the charge that the lack of visible losses was simply a result of the weakness of the stimulant was to select the most deadly wars possible. And so we chose the two most terrible wars in history: World War I and World War II. The two wars, and thirty-nine resulting test cases, provided us with our essential data. Such data were meager, but not so meager that we could not carry on the proper analyses which interested us. We also wished to ensure, in our assignment of roles to various actors, that the costs on the winning and losing side registered after the war actually had to do with the part each nation played in the fighting, and thus that the postbellum situation could be viewed as a result of the war. For this reason countries that had not

212

Chapter Five

taken part in the fighting served as controls. Otherwise one had no check that combatants' losses in resources were in fact a consequence of having participated in the war. Both short- and long-range consequences in war were considered. The results were startling. The short-run results proved much as one would have expected. The countries that managed to stay out of war did not have their power affected in any visible way by the conflict. Their power performance was found to be pretty much as it would have been had no conflict occurred. Winners underwent only very slight losses as a direct result of the war, while losers, as one would have anticipated, suffered heavily. On the other hand, in the long-run, while the other two groups of actors, that is, the neutral powers and the victors, continued to show little change from trends established during the prewar period, the effects on the losers were totally unexpected. Mter an initial heavy dip, the losers accelerated their recovery and, roughly fifteen to eighteen years after the guns stopped, were back up to levels they would have attained had no war occurred. In power terms their losses had been erased. In some cases they even surpassed the winners. It is this incredible recovery that we have called the Phoenix factor. We do not know why the Phoenix factor occurs, why the losers recover as quickly as they do. We do know, however, that the losers do not recover because the winners help them. We tested this proposition by looking at the recovery performance of all recipients of American Marshall Plan aid for all the years such aid was given. It is clear that the help the United States dispensed obeyed marked political considerations. The English, America's greatest allies, received more of the aid than any other recipient. The French received the next greatest amount. The Italians, who had begun the war as enemies and had switched sides in midstream, benefited less than any other ally but more than any of the defeated enemies who also received aid. The Germans followed the Italians in levels of aid, and the Japanese received the least amount of American Marshall Plan help. Our analysis reveals no clearcut relationship between the

213

Conclusion

aid and rates of recovery enjoyed by individual countries. But it is noteworthy that the nations in question recovered in exactly the inverse order to the amount of aid received: Japan performed best, then Germany, then Italy, then France, and the United Kingdom did the worst. It is very hard to argue with such results. Not only are shares of power not affected for any length of time by the terrible losses due to defeat in war, but, interestingly enough, they are also not increased by aid from abroad. External variables seem to be of minimal importance on a long-range basis so far as the power of a country is concerned. What matters is the way the system functions, that is, the political, economic, and social organization, and the values ofthe people. The reader will, of course, be aware that all we have said here about power applies to growth as well. The remnant of believers in the Protestant ethic should be pleased: God helps those who help themselves. Deterrence and Arms Races

We finally turned to military power. In our earlier chapters, all major modem wars-in their beginnings, outcomes, and consequences-appear to be a result of secular changes in productivity, population growth, and political development. Military power in the prenuclear age was largely a function of these three factors: as more resources became available through development, more arms were produced and stockpiled. And statesmen were left to play the role of earthquake detection units. But that was the pre nuclear world. It has been the common belief, strongly held by scholars and practitioners alike, that international power relations have been radically changed by the advent of nuclear weapons. More particularly, the power of nuclear arms has increased the relative influence of military might to such an extent and in such a way as to enhance the position of statesmen and free them to exercise direct control over events. Three fundamental elements make up such beliefs. First, due to the incredible explosive power of nuclear weapons, military strength no longer moves within the narrow limits

214

Chapter Five

imposed by economic productivity and governmental capacity to extract resources. Nuclear strength fluctuates independently of such constraints. It has been alleged that nuclear proliferation has already reordered the old international power hierarchy, and even larger changes are expected if nuclear proliferation continues. Second, nuclear weapons are so awesome that they will not even have to be used in order to obtain the desired effect; their mere presence will deter the adversaries of nuclear powers from attacking. Third, the sheer destructive force of the new weapons means that no nuclear nation could allow its competitor to gain a substantial advantage in explosive power, assuredness of delivery, or precision and speed of missiles. Matching as soon as possible any advantage won by an opponent,has been a critical safeguard of one's security, one's deterrent, and world peace. The connection between arms races or arms competition and deterrence should, therefore, be evident, for deterrence rests on knowledge on the part of both parties that each of them is as vulnerable to nuclear attack and total devastation as the other. We asked two questions. Do nuclear powers deter their opponents as they are alleged to do? And, are the major nuclear powers racing or competing with one another in the nuclear field? Our second question had a double purpose. The absence of nuclear arms races would constitute additional proof that nuclear deterrence did not take place. And if both deterrence and nuclear arms races were found to be absent, then the politics among nations have not changed. To answer the first of these two questions, we had to find out whether terror of nuclear retaliation deflects an adversary from attacking. Is it in fact true that, in confrontations where the disputants are free of the worry of nuclear war, they prove more likely to settle disputes by force, or that everyone "cools it" when the danger of escalation and nuclear confrontation stands in the background? If one takes all the international disputes since World War II and separates out those conflicts in which nuclear weapons were not a factor from those in which the involvement of a nuclear power was possible, if not likely, one finds, to one's great surprise, that exactly the opposite of the accepted proposi-

215

Conclusion

tion actually proves to be the case. There exists in reality a greater tendency to resort to force when there is risk of nuclear war, and not the other way around. When we explored very closely the instances in which the danger of nuclear weapons being used was highest, we again found the opposite of what one is led to expect by the deterrence model. Nuclear powers were often the target of extraordinary provocative attacks on the part of nonnuclear countries. Such acts can be interpreted as nothing less than a refutation of the deterrence model. The nonnuclear disputant held the field after the dispute was over. In more than half of these conflicts the nuclear power did not get its way, and getting its way was our definition of deterrence working. The availability of nuclear weapons, therefore, did not prevent nuclear powers from being attacked, nor did it prevent nonnuclear powers from defeating nuclear ones. The combined evidence clearly indicates, then, that it is a case of the purest magical thinking to believe that nuclear weapons, even though never used, are decisive in tipping the results of conflicts in favor of their possessors. But the matter could not be left to rest there. Midway through the nuclear age, the argument in support of deterrence theory was turned around. It was now claimed that nuclear weapons were never meant to stop nonnuclear attacks but were only supposed to deter nuclear aggression, hence both sides had to have nuclear weapons. Disproving mutual deterrence, however, proved far more difficult than was the case where one contestant had nuclear weapons and the other did not. When one of two nuclear quarrelers gave way to another in a dispute, one could never be sure that, at least in part, fear of nuclear retaliation was not involved in the nation beating a retreat. We had to approach the problem, consequently, in another, indirect, way. We argued that interaction was a necessary condition of mutual deterrence, and the very best indicator of interaction between two competing nuclear nations was the amount of money each spent on its own nuclear force in response to the nuclear allowance made by the other for the same purpose. If such interaction existed, the presence of deterrence could, at least, not be disproven.

216

Chapter Five

We then looked at the relationship between American and Russian investments in strategic nuclear weapons, trying to detect a systematic pattern of interaction in the expenditures of the two competitors. And the plain and startling fact of the matter is that hardly any trace of such a pattern could be found. The figures we came up with bear repeating. Ninety-six percent of the variance in Russian behavior was explained by factors other than American nuclear arms building. Between 76 and 91 percent of the variance in American behavior can be explained by their own stockpiling and development in the nuclear field. And none of the unexplained variance was attributable to interaction between the two nuclear powers. Neither country has been reacting to the other. We should make clear what it is we have found. We do not know what it is that leads nations to build nuclear weapons, but we see no logical alternative to the conclusion that internal factors determine budget decisions. We do not know with any precision, however, what these pressures are or how they function. If we did, we would be far ahead of our present position. We do know, on the other hand, that Russian behavior is not really responsible for the nuclear restocking and enlargement of the American arsenal, and American behavior is not really responsible for the concentrated effort made to the same end by the Russians. And though all current explanations of United States and Soviet behavJor in this regard argue a view contrary to this one, and though all current accounts of international politics are rooted in this contrary view, the data clearly indicate that such explanations are wrong, and present conceptions of international affairs must alter accordingly. In short, there is no deterrence, or nuclear arms races, or a balance of terror, and what peace we have has not been imposed by a nuclear deterrent. Nonnuclear powers have fought one another and nuclear countries as well. Nuclear countries may well fight each other with nuclear weapons one day, should their privileged position or the present international order be threatened. This is most likely to happen, as was always the case in past wars, when the challenger has surpassed the dominant nation in power.

217

Conclusion

If nuclear weapons have influenced the rules of international politics inherited from prenuclear times, the influence has not operated in the expected direction: the weapons have not modified these rules, they instead have reinforced them. Beyond the Data

We wish to raise by way of conclusion a number of questions which, though they cannot be answered now, suggest several themes for future research. They derive from some of the things we have found and they clearly lead beyond our data: (1) Do forms of a nation's political system have importance for the way things get done in government? For example, does it make a difference, issues relating to the quality of life apart, whether a country is totalitarian or democratic? (2) What exactly can be done about war? Is there room for political engineering at any point? (3) Finally, given our findings, what possible motive have the nuclear contenders for seeking security agreements with one another? Let us begin with the influence of political forms on the way the work of government gets done. It is widely held that these forms are important. Democracy's major advantage lies in the quality of life enjoyed by free men and women. But such advantages prove terribly costly in other ways. Democracy's major cost is alleged to be its lack of effectiveness owing to the difficulties democratic governments experience in getting free men and women to pull together in collective action, a problem said to be obviated in totalitarian systems. A second proposition contends that democratic and totalitarian systems differ in that the masses exercise some control over ruling elites in democratic countries, while in totalitarian nations control flows entirely the other way. Let us begin with the first of these propositions. Is it really true that democracies are less effective than dictatorships? We have only bits of evidence, barely enough even to question such accepted wisdom. But our own view is that there is no clear connection between effectiveness and forms of

218

Chapter Five

government, and that ceteris paribus, democratic and totalitarian systems could be equally effective. The evidence currently available is too scanty but seems to support the feasibility of a weak-to-no-connection proposition. For example, China and North Vietnam, both totalitarian countries, had the most effective political systems of any countries examined in this book. But Israel, a decidedly democratic state, was close behind. Again, a comparison by other scholars of the performance of the major actors in World War II in mobilizing resources for the war effort elicited results which feed our own suspicions. The data in question are presented in table 5.1. 1 Table 5.1

Proportions of GNPs Allocated for War Purposes in World War II by Selected Major Powers

Year 1938 1940 1941 1942 1943 1944

U.S. % Actual 2 11

33 45 46

(2.0)* (4.7) (29.8) (99.7) (149.7) (165.0)

Resources Allocated Japan Germany % Actual % Actual

U.K. % Actual

30 (28.2)* 37 (36.6) 41 (41.6) 44 (44.0) 45 (45.4) 50 (51.4)

7 (4.6)* 43 (32.7) 52 (41.7) 52 (42.8) 55 (46.6) 54 (45.0)

17 (9.5)*

17 23 30 42 51

(9.0) (12.3) (16.3) (22.8) (26.7)

* The actual portion of GNP, in 1965 U.S. constant dollars, allocated to the war effort was calculated from the Maddison collection. See n. 24, chap. 1.

Compare Britain's performance portrayed in the last column with that of Japan and Germany in columns 2 and 3. Britain, certainly one of the greatest of all democrcies, clearly mobilized more of her resources faster and for a longer period of time than did her totalitarian enemies. Albert Speer is clearly correct in suggesting that Germany was able to mobilize more in 1944, when she was already on her knees, than in 1940, when she dominated Europe. Even in their years of maximum need, however, when they had clearly begun to lose the war, the Germans and the Japanese could not really match the effort made by the British. Particularly impressive is the increase in British mobilization of

219

Conclusion

resources between the last year of peace and the first full year of war. The difference between the capacity of the Germans and British political systems in allocating resources to the war effort may go some length toward explaining the result of the Battle of Britain. The happy outcome of this most noble of British endeavors was not due either to the Channel, the climate, or the Royal Air Force, or even to Hitler's unstable strategic thinking. The major reason for British ability to repulse what appeared to the naked eye to be a vastly superior force lay simply in the success of Britain's democratic political system in directing a substantially greater fraction of its available resources to the war, thus making up for the fact that Britain's total pool of resources was considerably smaller than Germany's at the time. It was not then, as Churchill said, that so many owed so much to the few, but the other way around: The few owed so much to the many. The RAF won because the many pooled sufficient resources so that the few could do the job. Be that as it may, it seems legitimate to suggest, at least by way of hypothesis, that political forms and capacity are independent of one another. The proposition should be investigated further, and the study of international politics, and particularly of war, offers an ideal opportunity to undertake researches and perform vital tests in this critical area. We likewise harbor doubts about the second proposed difference between totalitarian and democratic systemsthat the elites in democratic systems are controlled from below, while in totalitarian systems it is the political elites themselves who exert control over their subordinates and mass pUblics. We are accustomed to thinking of the American and Russian systems as being poles apart. We are not concerned here with differences in the quality oflife: human rights, elections, freedom, mobility, self-expression, and so on, but of the way things get done in government. We think resource allocation for defense ought to reflect how things get done. Presumably things get done very differently in the United States and Russia. American presidents do not control large sectors of the American system, and often such

220

Chapter Five

elites can and do band together to impose their will in the struggles over the allocation of resources. It seems plausible to argue that an American president can indeed be looking for help to ease the pressure. In Russia, on the other hand, control is alleged to move in the opposite direction, from the top down. The chairman of the Russian Communist party is a dictator and the Russian productive machine is controlled by public enterprises owned by the government. Should not the system behave differently in the matter of allocating resources for strategic defense? How are we to interpret the fact that it does not? Our analysis of American and Soviet patterns of allocation of resources for the purposes of strategic nuclear defense, however, leads us to suspect that control in the two societies does not flow in different directions but answers in both cases to the same description. The established view that the two systems differ fundamentally in the way things get done and, particularly, in matters of control, cannot be completely true. Now let us tum to our second question: What can be done to stop world wars? Is there room for political engineering? Our data are not precisely those we would have collected had we had in mind the question, Is war an instrument of policy? But there is enough in what we have found to enable us to fashion a competing hypothesis to that favored by established military wisdom. We can, for example, take the data cited earlier on the actual distribution of power following major conflicts and see whether the combatants' expectations have been fulfilled. This simple approach does provide a clear indication that those who try to ride the tiger end up inside it. War is not the rational or effective instrument of national policy it is supposed to be. Indeed, it is no instrument at all. Consider as examples the cases of Germany and England in World Wars I and II. Germany began both wars. She fought them in order to force the other European countries to recognize her superiority over them all. She lost both wars badly, and her losses set her back momentarily; but she soon recovered her rate of climb and, today, after two terrible defeats, and with one-third of her population and territory amputated, Germany is, in terms of potential

221

Conclusion

power, roughly where she would have been had no wars occurred. This finding is appalling in its implications. Both world wars cost some tens of millions of lives. What for? Why did Germany start them? And take the case of England. England fought Germany in two world wars in order to maintain Germany's capabilities below her own. England won on both occasions and her victories reduced German power well below her own level for periods of between ten and fifteen years after each war. Today, England, in spite of her victories, is in precisely the situation she fought so strenuously to avoid. If power is the goal, how is it possible to believe, taking the long view, that wars are in reality the instruments of policy that those responsible for waging them sometimes say they are?2 One almost has the impression that "policy-makers" do not make policy in regard to war at all. They, along with the rest of us, are merely actors, speaking their lines on cue without being able to change the script. Suggestions that the problem can be solved simply by virtuous conduct-love, forbearance, understanding-are usually offered by people whose good intentions outstrip their knowledge of the problem. Our findings are clearly disturbing. There can be no question that major wars and armament races are intimately connected with secular trends in the growth patterns ofthe countries of the world. Even the biggest nations are inextricably caught in the grip of such currents. In the last decade, though for entirely different reasons from those which would motivate us here, the question has been raised seriously whether the industrial world should continue to organize its economy on the principle of growth. Though our own findings in this book could be used to promote a similar idea, they also cast doubt on its value. Other writers skeptical about the virtues of growth assume that a choice exists. Our analysis suggests that choice is largely illusory. Let us tum now to our final point. What will be the moments of maximum danger of a nuclear war? The reader may have found our conclusions about nuclear arms races and deterrence heartening. If there is no arms race and no arms competition, if there is no substance to the notion of the nuclear deterrent and to the notion of mutual deterrence,

222

Chapter Five

with all its chilling chances for miscalculation; then neither, surely, is there any danger of nuclear war, for there is no reason why nuclear weapons should be used. But the reader should not relax. Not quite. The absence of an arms race or arms competition does not mean at all that we are safe from the menace of nuclear war. Quite the contrary. But first an aside about this nonevent. The reasons we should continue to fear the worst are not those of classic theory on conventional arms competition. In that body of theory the link between the efforts of building up arms and the outbreak of war is, very briefly, as follows. Conventional arms races between major powers are very costly in resources. Owing to the limited destructive power of conventional weapons systems, a vast mass ofthose weapons is required to make any difference in a nation's total military strength. Moreover, in a competitive situation, each party is, by definition, trying to invest as much of its resources as it can, and is always seeking to escalate the effort required of itself and its adversary. Sooner or later one of the competitors tires and begins to fall behind. The loser of the race cannot, however, allow itself to be too far outdistanced, for in that case, should war break out, its own defeat would become inevitable. It would have little choice but to force the fight. But this detonating mechanism is absent in nuclear weapons competition. Nuclear arms are so efficient that it is possible to increase destructive power for very long periods of time without unduly straining resources. As technology progresses, weapons systems can be developed that do not rely heavily on human guidance or maintenance, and thus costs fall even more drastically. One example will illuminate the point. The United States has consistently decreased the allocation of resources to strategic arms, yet the effectiveness of her arsenal has by all reports increased. Consider. The average yearly operating cost of a B52 is twice that of a nuclear submarine. And the delivery capability of the former is obviously much smaller than that of the latter with its twenty multiple-headed missles which together can effectively destroy 148 separate targets. The costs of ground missile systems are even lower, and development of

223

Conclusion

technology seems to insure even larger savings. From this point of view, the buildup of stocks of nuclear strategic arms can be characterized as the accumulation of explosives without a fuse. Our conclusions have important implications. Even if nuclear arms races did exist, which they do not, such races would not be as destabilizing as conventional arms races are alleged to be. Then why in heaven's name do political leaders say that arms races are going on, that they are terribly dangerous, that they should be controlled; and why do such leaders often claim that they have controlled them? Why arms control? Why SALT? A number of explanations come to mind, but only one fits with our findings that nuclear weapons do not deter and that they are built because of internal pressures. In Russian and American decisions to acquire new weapons, external factors appear to playa negligible role. The argument runs something like this: at heart it is a matter of defense budgets. The costs of defense worries leaders caught continuously in conflicting demands for available resources. Their plight is similar to that of any individual in his private life. Consider any family breadwinner. His solvency is a prey to his boss, his union, his government, and, of course, to his wife and teenage children. In the case of armament programs, arms control agreements are important to national leaders because they permit them to resist more easily demands from many parts of the vast, heterogeneous coalition of forces that are essential to the governing of their own nations for ever greater allocations to the defense budget. These other "players" in the arms "game," unless they get what they want, are no less difficult to control by the American and Soviet leaders than is the enemy himself. It is a naive observer indeed who, judging by appearances, believes these people are easy to control because they are subordinates, because they speak to national leaders in respectful tones, because they render them military salutes, because they are cabinet or party officers, because they believe in the same general philosophy that the leaders believe, or because they have helped the leaders get elected. The way we account for the desire for arms control almost suggests that American

224

Chapter Five

presidents and the chairman of the Communist party of the Soviet Union are natural allies in league against coalitions of their own subordinates, and that such groups of subordinates may well be tacit allies against their superiors. It is a conception of the stratification of the international system not ordinarily considered. It is also a new view of the way the politics of each unit affects or tries to affect the international system. Arms control agreements constitute by their very nature, as seen in this new light, satisfactory compromises. A SALT agreement, for example, would offer all of the contenders their second choices. The establishment of an internationally agreed ceiling would help contain internal pressures for more resources for strategic arms. On the other hand, the coalitions in each country pushing for more investments in nuclear weapons are at least assured of an internationally agreed level of investment in the strategic forces. Both sides would not get all they might want, but each side would get something of importance. The reader will certainly see that this explanation does not really square with our earlier observation that nuclear weapons are relatively inexpensive, and that, therefore, preventing or limiting increases in the nuclear field does not save much money. If reducing costs is a major preoccupation of arms control, why try to control strategic weapons that cost so relatively little? Why not aim at more costly and less efficient conventional programs? The question is well taken. A plausible answer would be to suggest that arms control agreements in the nuclear field would, first of all, signal a major decrease in the extent of external danger, which is the main legitimizer of the defense budget as a whole. Once the diminution has been established, one could tum to paring other portions of the budget, less efficient and more costly than nuclear weapons, but also less vulnerable to cutbacks so long as an atmosphere of all-out effort in matters of security exists. Let us be clear. The pressures to arm do not end with arm control agreements, but they are eased a little. Let us now tum back to the main question. Is the fact that nuclear weapons have not deterred and have not been used

225

Conclusion

when deterrence failed an indication that they will never be used? It would be reassuring to think that this would be in fact the case, but, unfortunately, it is not. Let us review our findings. What do we think we know? First, we have substantial evidence that nuclear weapons, in advance of their being used, do not change the conflict behavior of the actors in international politics. Second, nuclear weapons do not deter. Third, they are not likely to be used against nonnuclear powers. Fourth, we have only one case where the superpowers confronted one another directy, and in that case the deterrent effect of nuclear weapons was muddled-and, of course, they were not used. It is certainly not possible to generalize from one case. One therefore has no evidence to go on. One can only speculate. It makes some sense to argue that. if nuclear weapons do not deter, the reason why we and the Soviet Union have continued to build them is that we want them handy if the need arises. They have not been used in the past because conditions for their use have not been there, that is, the nuclear powers did not care enough about the conflict. But what might be the conditions for their use? We suggest they are the same as the conditions that led the contenders-the challenger and the dominant nation-to fight world wars before nuclear weapons came on the scene. Let us circle back to our propositions and findings in the first chapter. In a nutshell, we found there that the contenders slide into world wars if the challenger overtakes the dominant nation in overall national capabilities. Let us underline that the national capabilities in question are economic, social, and political, and do not include a consideration of armed strength. Advantages in arms cannot make up for serious differences in the socioeconomic and political capacities of a nation. 3 If a challenger overtakes a dominant nation, a change in leadership and/or a change in system may be in the offing. At least it is likely to be tried. War results. The established leader fights to retain control, the challenger to obtain what is his by right of strength. We can now suggest that if nuclear weapons are available to both sides when such changes occur, there is maximum danger of their being used.

226

Chapter Five

We can make our attempts at forecasting a little more specific. We can at least enumerate the major actors in the possible drama. They are the same dramatis personae of major international conflicts today. These are: the present dominant nation, the United States; the present challenger, the Soviet Union, which is also the dominant nation in the international Communist order; and China, the present rival of the Soviet Union for possible leadership of the Communist world. Given the pools of human and material resources available to each of the three nations, and the probable rates of change of such pools in the coming decades, we guess that the USSR will approach the United States at some time in the future but without being able to surpass her. China, on the other hand, given her present base and the gains that will be hers if (and the "if" is real) she develops economically, is likely to overtake the Soviet Union and then, decades later the United States. The period after each passage will create the conditions for nuclear wars to occur. But will they? Will not the USSR and then the United States let China pass in peace? The reader will recall that peaceful passages by challengers have taken place. England did accommodate to the United States. War is not a certainty. It is, however, foolish to bank on future luck and the wisdom of the contenders. We will not know until the time has come. How long before the trajectories of the three countries intersect? These futures are below the present horizon. Attempts to extrapolate the recent past for fifty or a hundred years are plainly silly. Thus we cannot really tell. Perhaps it is better this way.

Appendix One

Index of Political Development

Studies of tax effort by fiscal economists, especially those by Raja Chelliah and Roy Bahl, provide the methodological background of this appendix.! The use of variations in tax ratios to account for meaningful intercountry differences in the size of the public sector, obscures the fact that rich nations can extract more because they have a larger base to tax. The estimation of tax capacity and control for the differences in available resources is a preliminary step in the direction of the estimation of tax effort. Real tax ratio (T/GNP) is the quotient of total taxes collected by the government over total output. Taxable capacity is the revenue collected given available resources relative to other nations in the system. It is calculated by the predicted values (T/GNP) obtained by regressing the real tax ratio against economic indicators of differences in resource base. Finally, the tax effort index is the quotient of real tax ratio to estimated tax capacity. This appendix is devoted almost exclusively to the methodological problems encountered in the derivation of the tax effort index. Tax Capacity: Alternative Models

Regression techniques have been used to control for differences in taxable resources among nations. Several equations have been proposed as the best representation of tax capacity. The two most widely employed and refined are:

227

1. Tax/GNP = A + Bl Exports/GNP + B2 Agricultural ProductionlGDP + B3 Mineral ProductionlGDP + Error 2. Tax/GNP = A + B! GNP/Total Population + B2 Nonmineral Exports/GDP + B3 Mineral ProductionlGDP + Error

228

Appendix One

We chose equation 1 for a number of reasons. The data for the independent variables are more complete and reliable for this equation. This is particularly true for nonmineral exports that fluctuate unexpectedly from year to year, due to major changes in reporting. Also, in the case of developing countries, agricultural production measures directly the proportion of total output that is extremely difficult for governments to tax; while GNP per capita used in equation 2 does not directly reflect this component. Finally, the results obtained from both equations are very similar, but generally equation 1 provides more stable coefficients and predicted values. This is congruent with our a priori expectation ofthe way a political development indicator would behave. Measurement Error Previous studies employed three-year moving averages to minimize the year-by-year fluctuations of the indicators caused by financial fluctuations and, more importantly, by inconsistent reporting. Because a complete time series is now available we controlled for fluctuations by pooling the time series and introducting directly a time component to detrend the equation. The resulting equation is as follows: Tax/GNP

= A + Bl

Time + B2 Export/GNP + B3 Agricultural ProductionlGDP + B4 Mineral productionlGDP + Error

where: Time

=

1,2,3,4, ... ,26 corresponding to 1950, 1951, 1952, ... 1975.

Pooling improves the consistency and efficiency of estimates. Comparisons of time series estimates with those obtained from year-by-year estimates produced coefficients that in most cases are within the confidence interval of those obtained in the time series (see table Al.l). Note that prior to 1973, very few coefficients obtained from the crosssectional regressions fall outside the 95 percent confidence interval obtained from the pooled model. Major deviations occur only for the exports coefficients in 1966 and for the

229

Index of Political Development

mining/GDP for 1957 and 1971. The increased instability for 1973 and 1974 results mainly from the differences in the sample of nations used in the estimate caused by drastic reductions in data availability that also prevent crosssectional estimates for 1950-53. Tax capacity values, measured by predicted values (Tax/GNP), are also very close to those obtained in the year-by-year estimates; but they do not fluctuate drastically in years when the sample of nations is reduced. The trend in the series is positive and large. This reflects the fact that most developing nations have increased governmental participation in all phases of economic life, partly as a result of the increasing complexity of their developing economies, but probably largely as a consequence of the process of socialization and centralization of governmental activities as governments take a direct role in the national economic life. Thus, the index of taxable capacity obtained from the pooled time series regression are stable over time. An added advantage-not exploited thus far-is the possibility of directly extrapolating future points for short time-periods. Centralized Economies

There are profound differences between the taxation systems of nations with centralized economies and nations with open economies. In open economies governments extract resources by taxing directly or indirectly income from industry, commerce, labor, and agriculture. On the other hand, in centrally controlled economies governments extract resources by taxing directly or indirectly profits from nationalized industries, by controlling labor salaries, and by centralizing the purchasing and selling of industrial and agricultural commodities. There is no need, therefore, in centralized economies to tax either mining operations or exports and imports. Thus, the inclusion of Communist nations in the sample (China, North Vietnam, and North Korea) drastically affected the coefficients for exports and mining, reversing the signs and rendering the estimate insignificant. Clearly two models are at work within the same equation. Given the number of available cases, the simplest

230

Appendix One

Table A1.1

Confidence Interval Table 95% Confidence Interval on Coefficients from Pooled Model

1954

1955

1956

1957

1958

1959

1960

1961

/3.: (.062, .194) /33: (- .189, - .098) /3.: (.021, .217)

.17 -.15 .07

.22 -.17 .15

.16 -.13 .13

.12 -.14 .01

.17 -.11 .80

.17 -.08

.19 -.07 .12

-.09

.09

.20

Where coefficients are for 2 export/GNP 3 agricultural production/GDP 4 mining production/GDP The confidence intervals in the left-hand column are derived on the basis of estimates taken from the pooled model. The numbers in each subsequent column to the right are the value of coefficients taken from the crosssection regression one year at a time. For example, the first value (.17) under the year 1954 is the regression coefficient /1. for the cross-section model for that year. It is contained in the confidence interval (.062, .194) of the parameter /3•.

solution was to incorporate a dummy variable for the different types of economies and determine its effects. We proceeded as follows: Tax/GNP

= A + Bl Time + B2 Type of Economy + B3 Exports/GNP + B4 Agricultural Production!GDP + B5 Mineral Production! GDP + Error

where: Type of economy

o = Open Economy 1 = Centralized Economy

With this dummy coefficient, we obtain stable, statistically significant, and theoretically acceptable results. To assure that the coefficients for all countries were stable, we added, in the case of each independent indicator, interactive terms, first for time, then for open and closed economies, and, finally, for both in combination. The more complicated formulations were not justified. Changes in slope over time did not materialize, and results were insignificant. Similarly, changes in slope for independent variables were absent in

.14

Index of Political Development

231

Coefficients from Cross-Sectional Model 1%3

1%4

1%5

1966

1%7

1968

1%9

1970

191

1972

1973

1974

.17 -.14 .07 .12

.15 -.16 .13

.11 -.16 .17

.04 -.19 .14

.07 -.15 .19

.09 -.16 .. 17

.07 -.18 .15

.05 -.18 .17

.08 -·.17 .32

.02 -.005 -.20 -.24 .17 .10

.07 -.30 -.02

1%2 .19

-.11

the case of centralized economies alone. The simplest equation was clearly the most appropriate given our data restrictions. Estimate of Tax Capacity

The final estimate replicates closely the findings of previous studies on this subject, despite our more limited sample. The results can be summarized as shown in table A1.2. Table Al.2

Statistical Estimation of Tax Capacity

Tax GNP

Agr Min Exp 11.66+ .23Time+ 23.87TypeEco. + 13 GNP -.14 GDP + .12 GDP

Stan- (.71) (.02) dard error (.30) Partial r

(.79)

(.02)

(.01)

(.03)

(.71)

(.21)

( -.32)

(.13)

Standard error = 5.05 R2 = .58 N = 909 of 988 Significance = .001, all coefficients

232

Appendix One

First, the 40 percent of the variance that is not explained we attribute to political rather than economic capabilities. These numerical results are very similar to those obtained in previous estimates. 2 Second, the magnitude of the coefficients indicate strong effect of all independent variables on the tax ratio, and affect the estimate in the predicted direction. Thus, the coefficient for centralized economies is strong and positive, indicating the pervasive intervention of government in economic activities. All results are significant. The application of ordinary least-squares estimation procedure to the pooled (time series of a cross-section) model is justified so long as one satisfies some basic assumptions about the error term Eit where i stands for country, and t for time. The error term may be encompassed by two statistically independent parts: a country-specific effect and a remainder. Eit

= Vi + V it

It is assumed that: EVitVi't'

= {O"~ ,i = i' and t = t' 0, otherwise

EV i V i' =

{

(T,2..,

u, I = I

0, otherwise This rules out autocorrelation among residuals for each country, covariation between residuals for different countries, and correlation between country-specific error components over different points in time. 3 We tested for autocorrelation among residuals for the sampled countries and discovered no serious problem (at worst, the DurbinWatson statistic showed indecisive tests for very few countries). The problem of covariation between residuals of different countries can be safely ruled out on theoretical grounds. If one assumes that the taxable capacity variables are exhaustively specified, then the effect of the tax effort variables, which are not specified in the model, would be absorbed by the error term. If the error terms for the dif-

233

Index of Political Development

ferent countries were correlated, the systematic element-tax effort factors-of each country's error term would have a similar effect on the tax performance of each country over time. This, however, is not the case; each country's tax effort over time is· affected primarily by domestic, idiosyncratic factors such as public demand or fiscal controls, which co-vary only by remote change with another country's tax effort patterns. Data

The data for the components employed in this study were collected from a variety of published and unpublished sources. Most of the economic indicators were obtained from the World Bank, Socio-Economic Data Bank, revised October 13, 1975; from the original collection by Raja Chelliah and Margaret Kelly of the International Monetary Fund used in the article on tax effort cited at the beginning of this appendix; and from the special collection supervised by Arthur House at the World Bank. For Communist nations data were obtained from documents provided by Senator Proxmire's Office, U.S. Senate; particularly important was the work "Notes on Statistics Provided for China, North Korea and North Vietnam" (1975). Data on Foreign Aid were obtained from the Agency for International Development, U.S. Foreign Assistance Annual (1960-1973), and from U.S. Department of State, Bureau of Intelligence, "Research Study," (annually, mimeograph).

Appendix Two

Table A2.1

Postwar American Aid

United States Aid to Major Powers, 1948-61

Year 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961

Japan Aid per Capita ($ per Capita)

West Germany Aid per Capita ($ per Capita)

Aid (Millions, U.S. $)

Aid (Millions, U.S. $)

4.4 5.4 1.5 1.0 1.7 2.2 2.1 1.6

400.5 367.3 392.9 147.5 97.3 97.1 36.3 14.0 10.7 10.5 6.9 8.7 0.7

0.05 0.06 0.02 0.01 0.02 0.02 0.02 0.02

8.50 7.68 8.12 3.03 1.98 1.95 0.72 0.28 0.21 0.20 0.13 0.16 0.01

SOURCES: Data on aid were obtained from the Agency for International Development, Office of Statistics and Reports, U.S. Economic Assistance

Programs Administered by the Agency for International Development and Predecessor Agencies, April 3, 1948-June 20, 1971 (Washington D.C.: AID, 1972), pp. 68-76, 46; population figures came from Arthur Banks, Cross Polity Time Series Data (Cambridge: MIT Press, 1971), segment 1, pp. 3-54. Our sample included only West Germany, Japan, France, Italy, and the United Kingdom, as we were only interested in the behavior of major powers. No data on Soviet aid were available. See Jacek Kugler, "The Consequences of War" (Ph.D. diss., University of Michigan, 1973), pp. 196-202. NOTE: Lack of aid after war was considered missing data because alternate sources of aid may have existed.

235

Postwar American Aid

Italy United Kingdom France Aid (Millions, Aid per Capita Aid (Millions, Aid per Capita Aid (Millions, Aid per Capita ($ per Capita) ($ per Capita) U.S. $) U.S. $) U.S. $) ($ per Capita) 8.3 459.8 286.9 301.7 216.8 191.9 113.4 47.1 18.6 4.0 1.0 0.7 0.7 0.0

0.18 7.00 6.20 6.42 4.58 4.04 2.38 0.98 0.39 0.08 0.02 0.01 0.01 0.00

102.3 1210.4 949.0 415.0 304.0 415.4 229.9 152.8 34.2 21.8 0.0 0.0 0.0 0.0

2.04 24.03 18.80 8.21 5.99 8.16 4.50 2.98 0.67 0.42 0.00 0.00 0.00 0.00

57.6 994.2 616.5 477.9 473.4 288.0 205.3 73.7 53.2 29.9 0.2 0.3 0.1 0.0

1.40 23.% 14.70 11.36 lJ.J8 6.75 4.77 1.70 1.22 0.68 0.01 0.01 0.00 0.00

Analysis of Models

Appendix Three

Models

Assuming that the stock of weapons depreciates at a ratio of (1 - A), the current value of a stock can be estimated for

nation X as:

Sit)

= X(t) + AX(t -

1) + A2 X(t - 2) + ...

where 0 < A< 1 X(t) = Expenditures by nation X for strategic weapons in year t. Sx(t)

= Stock of strategic weapons at time

t.

This series can be reduced to:

Using this formulation for the stock of weapons, consider the general model that included effects of internal and external sources simultaneously: (A1.0) Y(t) = y + aSy(t - 1) + j3Sit - 1) + e(t) (We reproduced this equation in its country-specific form in text model [1.0].) In this form the model cannot be estimated directly. Reducing by subtraction: AY(t -1)

= Ay + AaSy(t - 2) + Aj3Sx(t - 2) + Ae(t -

1)

yields (A1.1) Y(t) = y(1 - A) + (a + A)Y(t JL(t) where JL(t) = e(t) - Ae(t - 1).

236

1)

+ j3X(t -

1)

+

(We reproduced this equation in its country-specific form in the text models [2.0] [2.1].) Note that, empirically, this model cannot be distinguished from the model

237

Analysis of Models

(A2.0) Y(t)

= y + jiSx(t - 1) + e(t)

which again cannot be estimated directly. Subtracting AY(t - 1)

= 1\y + AjiSx(t - 2) + Ae(t - 1)

yields (A2.1) Y(t) = yO - 1\) + 1\ Y(t - 1) + jiX(t - 1) + lL(t) where lL(t)

= e(t) - 1\e(t - 1)

From this point, the Richardson formulation can be derived by assuming implicitly that 1\ = 1 and that the grievance parameter (y) is a trend parameter: 1-1

(A3.0) Y(t)

1-1

= 6 + y(t) + a2YG) +ji2X(j) + e(t) J=1

j=1

1-2

1-2

Again, subtracting Y(t - 1) =6 + y(t - 1) +a2 Y(j) + ji2X(j) + e(t- I) j=1

j=1

yields (A3.I) Y(t)

= y + (1 +

a)Y(t - I) + jiX(t - 1) + e(t)

- e(t - 1) which is the original Richardson equation. The difference between the internal estimates (a*) for the United States and the Soviet Union measures the difference in the level of internal pressures in the two countries: i.e., 6a = (al + 1\) - (a2 + 1\). If this difference turns out to be substantial, one would be persuaded that model Al.O, which specifies internal pressures a, is correct. If the two estimates are identical, one can choose between two equally plausible explanations. One can argue that depreciation is the only internal force generating pressures for investment in strategic goods. Or one can still prefer model Al.0 as theoretically more appealing, and argue that internal presures (a) other than depreciation are identical for both nations and that the contribution for both factors are absorbed by the depreciation rate (1\). Given our results, we are persuaded by model Al.O and believe that it is insufficient to specify external components alone.

238

Appendix Three

Analysis of Residuals

An examination of the residuals yielded the autocorrelations shown in table A3.1. Table A3.1

Autocorrelations of Residuals for U.S. and USSR in Arms Allocations Model

Lag

1952-1976 Sample* U.S. USSR (N = 25) (N = 25)

.11

1955-1%4, 1966-1976 Samplet U.S. USSR (N = 21) (N = 21)

1 2 3 4 5

-.29 -.17 -.01 .04

.27 -.30 -.19 .05 -.01

-.38 .30 -.50 .22 -.22

.18 -.29

Approximate 95% acceptance interval

±.39

±.39

±.43

±.43

-.09 .04 -.16

* See table 4.2. t See table in n. 33, chap. 4.

The Durbin-Watson test is inappropriate since the model contains the lagged value ofthe dependent variable as an explanatory variable. However, a different test suggested by Durbin (see Rao and Miller,Applied Econometrics, pp. 12326) yields the following test statistics (compare with the standard normal percentiles): -1.400 .746 Durbin's h 1.716 1.618 Only the first is (barely) significant at the .05 level. However, without the extreme values for the United States for 195254 and 1965, the tests are not significant. Hence, procedures for estimating the coefficients in the presence of autocorrelated disturbances, such as generalized least squares, Cochrane-Orcutt, etc., seem unnecessary. It is important to remember that these are derived models, and that the disturbances in the derived model have a different pattern of autocorrelation than in the original model (see Rao and Miller, Applied Econometrics, p. 168). If the disturbances in the original model (Et) follow a firstorder autoregressive scheme, with autocorrelations:

239

Analysis of Models

k

1,2, ...

=

and we apply the transformation yielding disturbances ILt = Et - AEt-I' then the disturbances in the derived model (f.Lt) have autocorrelations: _ (1

fk(lL) -

+ A2)pl k I -

Api

k-II -

Api

k+1 I

1 -2A+A2

_

k -I, 2, ...

If A is "close" to p (e.g., A = .75 and p = .90), then the autocorrelation function for the disturbances of the derived model will be considerably "damped." For example, with A = .75 and p = .90, the largest autocorrelation for the derived model disturbances is at lag 1 and is less than .23. Stability The stability of the action-reaction model depends upon the roots of the matrix ( a~ {32

{31 ) a1

lying inside the unit circle, where for model (AI.O) a~ = (al a1= (a2

+ A) + A)

and for model (A2.0) a~

= a1 = A.

The roots (z) are given by the solution of

Ia~

=0

- Z

a1- z

{32

or (a~

- z)(a1 - z) - {31{32

=0

or

~

z

=

a*I + a*2

±

2 ±

(a~

+ a1)2 + {3 {3 4

1

2

240

Appendix Three

Since both f3t and f32 are negative, we have real roots. For the models using data from 1951-76, we have: a~

f3t = - .34453

= .63928

a~ =

.93337 f32 = ~ z = .786325 ± .2031

=

.05697 .5832, .9894

The estimated system is therefore stable. For the models using data from 1955-64, 1966--76 (see n. 38, chap. 4) we have: a~ a~

f3t = -.45131 f32 = -.03848

= .72017 = .92016

~z =

.820165 ± .1654

=

.6547, .9856

This estimated system is also stable. Table A3.2

Estimated Lifespan for Selected Strategic Weapons Systems, 1950-80 Weapon System

Year Year Years Introduced Deactivated of Life*

Bombers B47 B52 C/D/E/F B52 G/H

1952 1956 1959

1965 1977 1980s

13 21 26

Avg. = 20 yrs.

ICBMs Atlas Titan II Minuteman I Minuteman II Minuteman III

1959 1962 1962 1966 1970

1964 1980s 1974 I980s 1980s

5 23 12 19 15

Avg. = 14.8 yrs.

Submarines George Washington Ethan Allen Lafayette

1960 1962 1964

1980s 1980s 1980s

25 23 21

Avg. = 23 yrs.

SLBMs Polaris A-2 Polaris A-3 Poseidon C-3

1962 1964 1970

1975 1980s 1980s

21 15

13

*Years of life computed to 1985 for systems still active

Avg. = 16.3 yrs.

241

Analysis of Models

Depreciation Rate

The depreciation rate for strategic weapons was calculated by approximating the life of different strategic elements in the United States arsenal. Table A3.2lists the weapons system involved, the year in which each system was introduced, the year in which it was removed, and the total number of service years. The data required for such calculations were obtained from a collection prepared in 1968-73 by Janet Burmester for the Nuclear Balance Project, University of Michigan. For later years data were obtained from the United States Air Force, Statistical Digest and Monthly Aircraft and Missile Digest (1946-72), and Sipri yearbooks (1974-77). This table distorts actual depreciation somewhat because some systems are only partly deployed, while others are fully operational; adjustments for such deviations appeared infeasible, however, because of the differences in costs involved. We also assume that all active systems will depreciate completely by 1985, and this may involve another distortion. Moreover, we do not consider the costs of refurbishing, updating, and improving present equipment. We conclude that the life of strategic equipment is approximately fifteen to twenty years. We can now insert this value in our original depreciation formula. Recall that Sit) = X(t)

+ ASx(t -

1)

The complete depreciation of existing stock can be simply calculated as follows: after 1 year: ASit); after 2 years: A2 Sit);

after n years: AnSit). We can calculate the depreciation rate (A), specifying at what value (e.g., 1 percent, 5 percent, etc.) the stock will be considered totally depreciated and how many years it will take for the stock of weapons to depreciate to that point. Thus, for example, if we assume 1 percent of value as total depreciation and we assume that it will take fifteen years for the

242

Appendix Three

stock to reach 1 percent of its original value, we can calculate that, e.g., 15 = = .736. Of course, different solutions will be obtained depending on the level that is set as total depreciation and the number of years assumed to be required to reach that level. Some reasonable ranges are:

v.m

Years

Value of the stock at end of period

.10% 1.0%

15 .631 .736

20 .708 .794

5.0%

.819

.861

In our analysis we use. 75, which is the approximate mean between fifteen and twenty years when total depreciation is considered to be reached at .01.

Notes

Introduction 1 A. F. K. Organski, World Politics, 2d ed. (New York: Alfred A. Knopf, 1968), p. 104. Alternate treatments of power can be found in Hans Morgenthau, Politics among Nations, 3d ed. (New York: Alfred A. Knopf, 1%0) p. 28; Karl Deutsch, The Analysis of International Relations (Englewood Cliffs, New Jersey: Prentice-Hall, 1967), p. 22. 2 Organski, World Politics, chaps. 7-8. See also David Singer, Stuart Bremer, and John Stuckey, "Capability Distribution, Uncertainty, and Major Power War, 1820--1%5," in Bruce Russett, ed., Peace, War, and Numbers (Beverly Hills: Sage Publications, 1972), pp. 21-27; and Nazli Choucri and Robert North, Nations in Conflict (San Francisco: W. H. Freeman, 1974), pp. 26-43. One

243

Causes, Beginnings and Predictions 1 Robert F. Kennedy, Thirteen Days (New York: W. W. Norton, 1969), pp. 127-28. 2 Notable efforts would include: Geoffrey Blainey, The Causes of War (New York: The Free Press, 1975); Quincy Wright,A Study of War, abridged ed. (Chicago: The University of Chicago Press, 1964); Russett, Peace, War, and Numbers; Manus Midlarsky, On War (New York: The Free Press, 1975); Choucri and North, Nations in Conflict; Organski, World Politics. 3 For a critical review of the assumption and prescriptions of the model, see E. H. Carr, The Twenty Years' Crisis, 1919-1939 (New York: Harper and Row, 1974), chap. 8; Morgenthau, Politics among Nations; Organski, World Politics, pp. 404:-27. For a recent, more formal specification of assumption, see Charles Lucier, "Power and the Balance" (Ph.D. diss., University of Rochester, 1974). 4 Organski, World Politics, p. 274. 5 E. Haas, "The Balance of Power: Prescription, Concept, or Propaganda," World Politics 5, no. 4 (1953). See also George Liska,

244

Notes to Pages 18-20

International Equilibrium: A Theoretical Essay on the Politics and Organization of Security (Cambridge, Mass.: Harvard University Press, 1957).

6 Inis Claude, Power and International Relations (New York: Random House, 1962), pp. 114-15, and chap. 4. 7 The collective-security model differs from the balance-of-power and power-transition models in that it is not directly concerned with explanations of the way the international system moves toward war but in prescribing what should be done to avoid war. See Organski, World Politics, pp. 404-27. 8 The power-transition model is not the only one that relies on internal dynamics to predict the likelihood of conflict. Choucri and North, Nations in Conflict, argue that lateral pressure generated by internal domestic growth creates friction among major competitors in the international system and may result in conflict, especially after expansion is halted by the expansion of other powers. They conclude that: "Our most important finding is that domestic growth (as measured by population density and national income per capita) is generally a strong determinant of national expansion. Our investigations have identified strong linkages from domestic growth and national expansion to military expenditures, to alliances, and to international interactions with relatively high potential for violence" (p. 278). 9 Organski, World Politics, pp. 364-67. During the transition from "nondeveloped" to "developed" status, nations undergo large and very rapid gains in their capacity to utilize resources and influence the behavior of other countries. Since any nation's power position is fixed entirely in relation to that of other countries, nations in development gain massive advantages by leaving behind those that are not developing and by closing the gap separating them from the leaders of the pack. This phase of a nation's life has been appropriately called that of "transitional growth in power." When a nation becomes developed and the gap separating it from other developed nations is closed, rates of increase in power inevitably fall sharply, although impressive absolute gains may continue. The reasons for this are plain. Advantages to be gained by catching up with more highly developed countries have already been absorbed in the previous stage. In relation to the advances of other developed countries-and power can only be relative-gains must be minimal. Power growth will slow down even though economic growth may continue to increase at a fast pace. At this point, the developed nation has reached the stage of power matur-

245

Notes to Pages

2~O

ity. See A. F. K. Organski, Stages of Political Development (New York: Alfred A. Knopf, 1965). 10 The significant changes in the power of any nation are related to describable and predictable progress in development. Three phases are involved: In the first, before industrialization and the accompanying increase in the capacity of the political system to extract and pool resources, nations find themselves capable of utilizing only the smallest fraction of their human and material wealth. Development is low and a nation is consequently weak. Change, if it occurs at all, is almost imperceptibly slow. All the power gains to be derived from development lie in the future. This period is called "the stage of potential growth in power." 11 William Riker, The Theory of Political Coalitions (New Haven: Yale University Press, 1962). 12 This is not the first attempt to test empirically competing power theories: see Richard Rosecrance, Alan Alexandroff, Frian Heal, and Arthur Stein, Power, Balance of Power, and Status in Nineteenth-Century International Relations (Beverly Hills: Sage Publications, 1974). See also Singer et al., "Capability Distribution," pp. 19-48.

13 Organski, World Politics, p. 104. In the extensive literature dealing with power there is a loose consensus how power should be defined. The range of disagreement is narrow and it is suggested by the differences among the definitions proposed by Karl Deutsch, Hans Morgenthau, and the one quoted in the text above. Karl Deutsch defines power "as the ability to prevail in conflict and overcome obstacles" (The Analysis of International Relations, p. 22); and Hans Morgenthau defines it as "man's control over the minds and actions of other men" (Politics among Nations, p. 28). 14 This distinction between national capabilities (or power resources) on the one hand and national power on the other is widely accepted among specialists in international politics. Throughout this chapter we are concerned solely with national capabilities and power resources. However, in order to relieve the tedium of repeating over and again" national capabilities" we shall use three terms-national capabilities, power resourceS, and national power-interchangeably. Thus the reader is advised that whenever such terms appear we mean national capabilities. 15 Why "perceptions" constitute an important ingredient of any attempt to estimate national power, is given an excellent discussion in Ole Holsti, "The Belief System and National Images: A Case Study," Journal of Conflict Resolution 6, no. 3 (1962): 244-52; Cf.

246

Notes to Pages 31-34

also K. J. Holsti, International Politics: A Framework for Analysis (Englewood Cliffs, New Jersey: Prentice-Hall, 1967), chap. 7. 16 For a review of the factors frequently considered as important components of national power, see Raymond Aron, Peace and War (New York: Doubleday, 1966); Ray Cline, World Power Assessment (Washington, D.C.: The Centerfor Strategic and International Studies, 1975); also William Coplin, Introduction to International Politics (Chicago: Markham Publishing Co., 1970); Organski, World Politics, chaps. 7-9. 170rganski, World Politics, chaps. 6-8; Singer et aI., "Capability Distribution," pp. 21-27; Nazli Choucri and Dennis Meadows, International Implications of Technological Development and Population Growth: A Simulated Model of International Conflict (Cambridge, Mass.: Center for International Studies, MIT, 1971) pp. 23-24. This presentation considers technology as an independent variable rather than as a component of economic output. 18 There was, in fact, not a wide choice of measures. There have been to date ony a few explicit attempts to estimate a single measure of national capabilities: Wilhelm Fuchs, Formein zur Macht (Stuttgart: Deutsche Varlage-Anstaldt, 1965); Clifford German, .. A Tentative Evaluation of World Power," Journal of Conflct Resolution (March 1960): 138-44; Organski, World Politics, pp. 208-9. One should note that Klaus Knorr has made a major contribution to the discussion of the problem of aggregation. See his Military Power and Potential (Cambridge, Mass.: D.C. Heath, 1970); see also Knorr, The War Potential of Nations (Princeton, N.J.: Princeton University Press, 1956) and Wayne Ferris, The Power Capabilities of Nation States (Lexington, Mass.: Lexington Books, 1973). 19 A very similar formulation has been developed by Wilhelm Fuchs, in Formein zur Macht. He suggests the following equation: Power = Populationa x Industrialization". Fuchs adds exponential weights to the components, and the implicit weighting scheme in our own formulation is made explicit and variable by him. In principle the addition of the weights to the components is, of course, a major step forward. However, in this case, the assumptions underpinning the assignment of such weights are highly debatable. There is another major flaw in the utility of Fuchs's interesting attempt at the measurement of national power. He begins by assuming the distribution of power and reduces the problem of estimation to a simple manipulation of components. But this, of course, is clearly

247

Notes to Pages 34-35

unacceptable. If the power relationship is known a priori there is no need to estimate components, and if it is not known the extrapolation from assumed values are themselves assumed values. For a review of Fuchs's work see Jacek Kugler, "The Consequences of War: Fluctuations in National Capabilities Following Major Wars, 1880-1970" (Ph.D. diss., University of Michigan, 1973), pp. 71-74; and cf. also Klaus P. Heiss, Klaus Knorr, and Oskar Morgenstern, Long-Term Projections of Political and Military Power (Princeton, N.J.: Mathematica Inc., 1973), pp. 335-43. 20 The fact that data on total output provide good estimates of the power distribution in the international system has been repeatedly noted in the writings of scholars concerned with the problem of evaluating the various attempts to measure national capabilities. Cr., for example, Steven Rosen, "War, Power and the Willingness to Suffer," in Russett, Peace, War, and Numbers, p. 171; and Charles Hitch and Roland McKean, The Economics of Defense in the Nuclear Age (Cambridge, Mass.: Harvard University Press, 1967), chap. 1; Normal Alcock and Alan Newcombe, "Perceptions of National Power," Journal of Conflict Resolution 14, no. 3 (1970): 335-43. Support for total output as the best overall measure of national power is not universal. Some have argued that estimates of energy consumption by a society are, perhaps, a better overall measure of its national capabilities because such a measure is more readily comparable across nations. The two measures are highly correlated. See Heiss et al., Long-Term Projections, pp. 59-62. Nevertheless, the conceptual arguments pushing for the selection of total output seem very strong. 21 Singer et aI., "Capability Distribution," pp. 19-26. See also James Ray and J. David Singer, "Measuring the Concentration of Power in the International System," Sociological Methods and Research 1, no. 4 (May 1973): 403-37. 22 The procedure can be formally written as follows: Let X' u be the measure, in real units, of country j on the power component i where j = I, ... , n nations and i = I, ... , r components of power dimensions. The final relative measure of capabilities Xu for any country in the system is derived in a two-step procedure: (1) X'ij are converted from level units to share units Xii as follows: X'\" X I j =--' where X't. = I X'ii X'. j

I.

(2) Then the capability measure for country j is derived from: 100 r

,.

I i

=I

Xii

248

Notes to Pages 37-55

23 Kugler, "Consequences of War," pp. 82-94. 24 Most of the gross national product figures in our study come from

Angus Maddison, "Trends in Output and Welfare," mimeograph (1975), recently published in Carlo M. Cipolla, ed., Contemporary

Economies. The Fontana Economic History of Europe, vol. 6, pt. 2 (New York: Barnes and Noble, 1977), that provides adjusted figures of Maddison's previous estimates published in Economic Growth in the West (New York: Twentieth Century Fund, 1964), pp. 194-%, and Economic Growth in Japan and the USSR (London: Allen and Unwin, 1969), pp. 154-56. In some countries adjustments had to be made or the series had to be completed. For a detailed description of these adjustments and the data-set, see Kugler, "Consequences of War," pp. 293-97. 25 Bruce Bueno de Mesquita, "Measuring Systemic Polarity," Journal of Conflict Resolution 19, no. 2 (1975): 187-216. Also Bueno de Mesquita, "The Effects of Systematic Polarization on the Probability of War" (Paper delivered at the annual meeting of the International Studies Association, Toronto, Canada, February 1975). 26 For a detailed analysis of risk, alliances, and war, see Bruce Bueno de Mesquita, The War Trap (forthcoming). 27 J. David Singer and Melvin Small, Wages of War 1816-1965: A Statistical Handbook (New York: John Wiley, 1972), pp. 58-90; see also J. David Singer and M. Small, "The Composition and Status Ordering of the International System, 1815-1940," World Politics 18, no. 2 (1966). 28 Other empirical works address this problem as well: see J. David

Singer and Melvin Small, "Alliance Aggregation and the Onset of War, 1815-1945," in J. D. Singer, ed., Quantitative International Politics (New York: The Free Press, 1968) pp. 247-86, and J. David Singer and Melvin Small, "War in History and the State of the World Message," in W. Coplin and C. Kegly, eds., Analyzing International Relations (New York: Praeger, 1975), pp. 220-49. See also Bueno de Mesquita, The War Trap. 29 We excluded from the sample three pairs where one nation overtook another, but at the same time they were fighting a third contender. For example USSR and U.K. pairs were excluded during World Wars I and II. 30 The results of the probit analysis were replicated using ordinary least squares, with no significant changes. For references regarding the controversy about the interpretation of probit results, see John Aldrich and Charles F. Cnudde, "Probing the Bounds of Conventional Wisdom: A Comparison of Regression, Probit, and Dis-

249

Notes to Pages 55-65

criminant Analysis," American Journal ofPolitical Science 19, no. 3 (August 1975); Richard McKelvey and William Zavoina, "A Statistical Model for the Analysis of Ordinal-Level Dependent Variables," mimeograph (University of Rochester, 1974); Martin Zechman, •• A Comparison of the Small Sample Properties of Probit and OLS Estimators with a Limited Dependent Variable," mimeograph (University of Rochester, April 1974); William H. DuMouchel, "The Regression of a Dichotomous Variable," mimeograph (Institute for Social Research, The University of Michigan, 1973). 31 When relative power and growth are run as independent rather than interactive components, the variance explained by each is approximately the same, with the first accountin~ for about 30 percent (R2 : .33) and the second about 20 percent (R2 : .23) of the variation in the model. But results are unstable due to multicolinearity among these independent variables. 32 Two recent studies seem to support the generality of these findings even though they utilize different measures and different samples. See David Garnham, "The Power Parity and Lethal International Violence, 1969-1973," Journal of Conflict Resolution 20, no. 3 (1976): 379-94, and E. Weede, "Overwhelming Preponderance as a Pacifying Condition among Contiguous Asian Dyads, 19501969," Journal of Conflict Resolution 20, no. 3 (1976): 395-412. 33 It is of course true that in every interaction among nations there may be present residual effects of alliances that are not registered here. Napoleon III may have believed that he would receive assistance from the British, Russian and Austro-Hungarians in his war with Prussia. This belief may have played a role in his planning for the war. Also, Britain had a defense pact with Japan and the British could have chosen to honor the commitment but did not. Such calculation may have nevertheless affected the strategies of both sides. 34 For an excellent discussion of the problem, see Choucri and North, Nations,in Conflict. Two

Davids and Goliaths 1 There is a good deal of agreement among writers on development that the dimension of capacity is a critical, if not the central, element in any definition of political development. See J. Coleman, "The Development Syndrome: Differentiation-EqualityCapacity," in L. Binder and Joseph LaPalombara, eds., Crises and Sequences in Political Development (Princeton, N.J.: Princeton University Press, 1974), chap. 2; S. Verba, "Sequences

250

Notes to Pages 65-67

and Development," in Binder and LaPalombara, eds., Crises and Sequences, chap. 8; R. Holt and J. Turner, "Crises and Sequences in Collective Theory of Development," in American Political Science Review (hereafter referred to as APSR) 69, n. 3 (September 1975): 979-94. 2 Viewed in this fashion, the construction of an index of political development is simply incidental to our attempt to predict the outcome of a major class of wars. But any attempt to measure •• political development" cannot be viewed simply as another exercise in index construction. The construction of a measure that will trace systematically, even in general outline, differences in development of political systems is so exciting and important an exercise in its own right that many readers may prefer to view that in itself as the main event, and the prediction of winners and losers in an international conflict simply as the laboratory exercise through which the performance of the measures of political capacity can be monitored. This angle of vision on the problem is, of course, entirely legitimate. In all cases where the connection between proposed measures and behaviors have not previously been tested, one is faced with a crucial question: how does one know that the measures one has come up with are really an index ofthe behavior one claims they measure? Validity cannot be assumed. And it is entirely legitimate to regard the prediction of the outcome of conflicts simply as the validation procedure of a first attempt to measure political capacity. In reading this research, the reader can choose either angle of vision at no extra cost. In either case, the procedure, the data, and the findings are exactly the same. 3 One could, of course, include a list of wars where traditional measures of national capabilities also predict the outcome accurately. Every one of the confrontations between developed nations considered in Chapter 1 falls in this category, but such inflation of cases would not aid in the task at hand. 4 Our view is not unorthodox. As we have noted already in Chapter 1, the strength of a nation rests in the basic institutions, and in the number of skills possessed by the populations making up the national society. Military organizations are, at best, an intervening variable whose ultimate influence, however, will correlate closely with the socioeconomic and political performances we measure directly. Mao's apothegm that "power grows out of the barrel of a gun," is partially true but totally misleading. 5 Data from United Nations, Demographic Yearbook, 1974 (New York: United Nations, 1976); Arms Control and Disarmament Agency, World Military Expenditures, 1970 (Washington, D.C.:

251

Notes to Pages 68-74

Government Printing Office, 1970), tables 2-5; and United Nations, Statistical Yearbook, 1976 (New York: United Nations, 1977). 6 See figures 2.7 and 2.10 above. 7 Samuel Huntington, Political Order in Changing Societies (Yale University Press, 1967), pp. 290-310. 8 Karl Deutsch, "Social Mobilization and Political Development," in Jason Finkle and Richard Gable, eds., Political Development and Social Change (New York: John Wiley and Sons, 1966); Ted R. GUIT, "Persistence and Change in Political Systems, 18001971," APSR 68, no. 4 (December 1974): 1482-1504; Phillips Cutright, "Political Structure, Economic Development, and National Social Security Programs," in J. V. Gillespie and B. A. Nesvold, eds., Macro-Quantitative Analysis: Conflict, Development, Democratization, vol. 1 (Beverly Hills: Sage Publications, 1971); Stein Rokkan et aI., Citizens, Elections, and Parties (New York: David McKay, 1970). 9 United Nations Research Institute for Social Development (UNRISD), Contents and Measurement of Socio-Economic Development (New York: Praeger Publications, 1972). 10 Organski, The Stages of Political Development; Organski, World Politics; A. F. K. Organski, Bruce Bueno de Mesquita, and Alan Lamborn, "The Effective Population in International Politics," in Keir Nash, ed., Governance and Population: The Governmental Implications of Population Change; Commission on Population and Growth and the American Future, Research Reports (Washington, D.C.: Government Printing Office, 1972). 11 Huntington, Political Order in Changing Societies; Lucien Pye, "The Concept of Political Development," The Annals of the American Academy of Political and Social Science (March 1965): 2-13; Binder and LaPalombara, Crises and Sequences. 12 Gabriel Ardant, "Financial Policy and Economic Infrastructure of Modem States and Nations," in Charles Tilly, ed., The Formation of National States in Western Europe (Princeton, N.J.: Princeton University Press), p. 220. Many have suggested that electoral re-

turns offer the right information one would need to estimate governmental penetration. And, at first glance, the use of electoral data appears to offer a promising solution. In addition, its relative accessibility would permit at least limited international comparisons. But electoral data can be very misleading. We are riot worried here by the degree to which a system is truly representative. Our problem resides in the fact that voting, per se, does not tell us whether the level of penetration is symbolic or genuine. In the

252

Note to page 75

Soviet Union and in most East European countries almost the entire population votes. This electoral procedure does not mean, of course, that those elected represent the real choice of the voting population. Although the form of government is not democratic, the degree of government is high. Conversely, in India, in Indonesia, and in non-Communist Southeast Asia generally, the size of the voting population is large (though not nearly so large as in the Communist countries), but penetration in most cases is more symbolic than real. Transporting villagers quadrennially or quinquennially to polling places so that they can put their marks on pieces of paper is often the limit of governmental influence on their behavior. Control and penetration are more imagined than actual. Electoral data used for the purpose we have in mind would be deceptive. And one should add that electoral data used for the purpose of indexing political development are almost equally unsatisfactory . In the case of penetration nothing would be more logical or elegant than to estimate the depth of penetration by determining the number of people taxed directly by the central government. What would serve us best as a direct measure of governmental penetration is an enumeration of taxes imposed directly on citizens by the central government. However, data are lacking or totally unreliable in the case of many-perhaps most-of the developing countries. Information on the number of income-tax payers is available for a limited number of Western European countries. In the more advanced nations, where such data exist, the problem of gathering the information and making the data comparable are overwhelming. Moreover, comparisons cannot be made with centrally controlled economies because the Communist countries, for example, have rejected personal income taxes as a source of revenue and, instead, tax their populations indirectly. Finally, even in the case of the free economies of the non-Communist world, possible comparison is relatively narrow. Perhaps these difficulties will be surmounted, but at present they remain in the category of work to be done. Currently, levels of political penetration can only be estimated indirectly through measures of governmental extraction of resources. The problem is not critical because the two are quite obviously highly correlated, but the fact remains that of the two behaviors we need to measure we have succeeded in estimating only one. 13 The basic research in this area was carried out by economists of the International Monetary Fund. The most relevant publications in this area are: Jorgen Lotz and Elliott Morss, "Measuring 'Tax Effort' in Developing Countries," International Monetary Fund

253

Notes to Pages 77-106

Staff Papers (hereafter referred to as IMF Staff Papers), (July 1, 1971), pp. 254-331; Roy Bahl, "A Regression Approach in Tax Effort and Tax Ratio Analysis," IMF Staff Papers (November 1971), pp. 57(}....610; and Raja Chelliah, Hassel Baas, and Margaret Kelly, "Tax Ratios and Tax Effort in Developing Countries, 1969-1971," IMF Staff Papers (May 1974). 14 Chelliah et aI., "Tax Ratios," p. 295; Bahl, "A Regression Approach," p. 590. 15 Ifthe resulting ratio that represents tax effort equals 1, the tested country is performing predictably or normally. If the ratio equals more than 1, the country has a performance better than the norm for the whole sample. And if the ratio is less than 1, the country is doing worse than could be expected given its economic resources. For results, see App. 1. 16 Bahl, "A Regression Approach," pp. 582-83. 17 As Klaus Knorr pointed out, levels of military preparedness in peacetime are not a good indicator of the military strength of a nation in times of war. A direct measure of political development reduces the problem of estimating the strength of a nation over time, because the capacity of the political system in peacetime indicates the level of mobilization likely to be reached in times of military conflict. Knorr, The War Potential of Nations. 18 It has been alleged that some air operations on the Communist side were actually carried out by the Soviet Union, but the charge has never been substantiated. The number of Chinese soldiers committed to the Korean War was estimated to be about 1,300,000. The United Nations forces numbered about 365,000 men. These figures were reported by Vincent J. Esposito, Head ofthe Department of Military Art and Engineering, United States Military Academy, in Encyclopedia Americana, 16 (1965): 527-28k. The estimated aid to North Korea over the war period (1950-53), is $14,815 millions. The Korean War cost the United States an estimated $49 billion (M. T. Haggard, "United States Expenditures in Indochina and in Korea," Congressional Research Record [Washington, D.C.: Library of Congress, April 1975]).

Three

The Costs of Major Wars 1 John Keynes, The Economic Consequences of the Peace (New York: Harcourt Brace and Rowe, 1920). Keynes makes the most authoritative statement in support of the belief that a gap between winners and losers developed as a consequence of the war will continue to increase incrementally, but that in the very long run,

254

Notes to Pages 106-10

the rot in the economic systems ofthe losers will spread to winners as well. For discussion of the short-term effects of wars on belligerents, see United States Strategic Bombing Survey, The Effects of Strategic Bombing on the German War Economy (Washington, D.C.: Government Printing Office, 1945); cf. also, idem, The Effects of Strategic Bombing on Japan's War Economy (Washington, D.C.: Government Printing Office, 1946). 2 Two works in particular argue at length that there are enormous permanent losses as a result of war. See Norman Angell, The Great Illusion (New York: Putnam, 1933); John U. Nef, War and Human Progress (Cambridge, Mass.: Harvard University Press, 1950). There are a number of excellent demographic studies exploring the question of the impact of war on the demographic structure and behavior of belligerent populations at the end of wars. See for example, Frank Notestein, Irene Taeuber, Dudley Kirk, Ansley Coale, and Louise Kiser, The Future Population of Europe and the Soviet Union (Geneva: League of Nations, 1944), chap. 3; and B. T. Urlanis, Wars and Population (Moscow: Progress Publishers, 1971). 3 In the demographic field, Gregory Frumkin, Population Changes in Europe since 1939 (New York: Augustus-M. Kelley, 1951) for one, has indicated that "baby booms" after a war may erase demographic losses sustained during a war. A recent United Nations study of the demographic consequences of all wars in recorded history estimates that global losses are not larger than the growth rate for thirty years. See United Nations, The World Population Situation in 1970, Population Studies No. 49 (New York: United Nations, 1971). The estimates of global losses, however, do not show the demographic effects of war on individual nations. 4 It is worthy of note that this procedure has been used by demographers in their attempts to estimate the impact of wars on participating populations. See, for example, Notestein et aI., The Future Population; see also Frumkin, Population Changes. 5 This procedure has been occasionally suggested in the study of politics. See, for example, Kenneth Boulding, A Primer of Social Dynamics (New York: The Free Press, 1970), pp. 99-107; cf. also Ted GUIT, Why Men Rebel (Princeton, N.J.: Princeton University Press, 1970), chaps. 2-4. The major empirical applications of this procedure have appeared in economic literature. 6 Lawrence Klein, "The Procedure of Economic Prediction: Standard Achievement, Potential" (Paper presented at Conference on the Economic Outlook for 1973, University of Michigan, November 16-17, 1972).

255

Notes to Pages 111-21

7 Heiss et aI., Long-Term Projections of Political and Military Power, p. 107. 8 Alternates to the simple estimation techniques utilized here are available, and a word about the reasons for our choice is necessary. Specifically, the Box and Jenkins models that extend regression analysis, as well as the Fourier analysis (see Chapter 4), are designed to provide accurate forecasts of a single time-series and to reflect the internal structures. In our data, the single cycle corresponding to the depression years is dominant and the less pronounced variations do not affect substantially the forecasts over the periods chosen. Moreover, given the quality of the data-set for some of the nations in our sample, the application of more sophisticated techniques would of necessity have to be restricted to a subsample of the cases employed. The tradeoffs did not, in our view, justify complicating the analysis even further. One consequence of this decision is that we underestimate the performance of some nations after World War II, but the reader can adjust with ease for the distortion by inspection. See George Box and Gwilym Jenkins, Time Series Analysis (San Francisco: Holden-Day, 1970). 9 Simon Kuznets demonstrated convincingly that growth rates of developed countries remained constant or increased in the last century. See his Economic Growth of Nations (Cambridge, Mass.: Harvard University Press, The Belknap Press, 1971), chap. 1. 10 Kugler, "The Consequences of War," pp. 116-7. 11 Frumkin, Population Changes, p. 17. 12 Maddison, Economic Growth in the West, pp. 193-96. 13 Thomas Sanders, Lutz Erbring of the Department of Political Science, and J. Landwehr and D. Fox of the Statistical Research Laboratory of the University of Michigan deserve our thanks for the ideas and advice they generously contributed to the solution of this difficult problem. They are not responsible, of course, for any weaknesses in the final product. That responsibility rests with us. 14 See Kugler, "The Consequences of Wars," chap. 4 and app. 3. 15 This procedure has been used by economists in the study of economic depressions. Economists have focused their attention on the most intense depressions and found correlates that could be used to forecast smaller recessions. See Arthur Burns, ed., The Business Cycle in a Changing World (New York: National Bureau of Economic Research, 1969), pp. 3-53. 16 We accept Quincy Wright's theoretical definition of war as "the legal condition which equally permits two or more hostile groups to carry on a conflict by armed force" (A Study of War, p. 7); and we

256

Notes to Pages 122-25

further accept the operational definition of Lewis Richardson that any conflict resulting in the death of approximately three hundred people can be considered a war. Since our concern is with international war, we employed the list of wars between 1815 and 1965 provided by Singer and Small in their Wages of War, pp. 17-19, 30-32, and 58-70. 17 The gratitude of the authors goes to Angus Maddison, who generously allowed us access to a data-set, unpublished at the time of writing, which made this study possible. The bulk of the capability data for our study comes from his "Trends in Output and Welfare." For a complete data-set see Kugler, "The Consequences of War," app. 2. 18 Cf. Maddison, Economic Growth in the West. 19 Even some ofthe most firmly established nations-e.g., Germany, France, Austria, Poland, Finland-lacked one or another of the characteristics of nationhood for protracted periods during the half century covered by this research. 20 There has been disagreement, of course, over which characteristics are most important. Karl Deutsch, for example, has suggested that de facto control over substantial portions of the governmental machinery, population, and territory would be sufficient to substantiate the claim that a nation existed (The Analysis of International Relations, p. 70); John Herz, on the other hand, has argued that the ability to administer policies free of outside interference is the important characteristic of a nation (International Politics in the Atomic Age [New York: Columbia University Press, 1959], p. 104). 21 Singer and Small, Wages of War, pp. 58-59, and Bruce Russett, J. David Singer, and Melvin Small, "National Political Units in the Twentieth Century: Standard List," APSR, 62, no. 3 (September 1968): 932-51. 22 This second set of countries was culled from the list of countries in Russett et al., "National Political Units." 23 We measured levels of economic development in an orthodox way, that is, by looking at the distribution of the labor force between the agricultural and nonagricultural sectors. Wherever more than 50 percent of the males of working ages were engaged in nonagricultural pursuits, the country was considered developed. 24 An important factor in this analysis is the issue of continuity in the territorial integrity of a country. Some nations in their history undergo drastic changes in territory, so much so that a successor state must be identified. In our study we chose no successor to

257

Notes to Pages 128-29

Austria-Hungary, the Soviet Union as a successor to Russia, and West Germany as a successor to Germany. The territorial adjustments in the second case were small, but in the others they were very significant. It is possible to argue that the consequences of war are not registered in power terms because we simply eliminated the territorial changes from consideration. Germany for example, may have ceased to be a great power because its territory was reduced, and that should be seen as a consequence of war. The same can be said of the fate of Austria-Hungary. But this argument is not convincing. Germany's loss of territory is the result of her weakness, not the other way around, and if one controls for territory the recovery path of the political jurisdiction should be properly considered an indication of the consequences of war. 25 Our theoretical and operational definition of "active belligerent loser" requires an additional word of comment. The definition was suggested by an early treatment of the problem by Charles Kindleberger in his "International Political Theory from the Outside," in William Fox, ed., Theoretical Aspects of International Relations (Notre Dame: Notre Dame University Press, 1959). It is clear that our operational indicator of defeat in war is based on the observation that in the past winners have extracted from losers some territory, while never giving up any of their own. We assume, but have no evidence in support of the assertion, that losers would have extracted territory from their adversaries had they won the war. We cannot rerun the conflicts to see what would have happened had the outcome been reversed. But it is reasonable to assume that the losers would have demanded territory if they had happened to win. There is a good deal of indicative evidence that suggests such an outcome to be very highly probable. Germany and Italy, for example, at the beginning of World War II had defeated France and had exacted territory from that vanquished nation. The number of other such illustrations would be very long indeed. 26 Such groups can be most simply and elegantly expressed in the theoretical language of set theory: Active Belligerent Winner = Bn(OUT)C Active Belligerent Loser = Bn Tn OC Occupied Belligerent Winner = B non T Occupied Belligerent Loser = B non T Non-Belligerent = BC 27 It should be emphasized, however, that these results are not simply the effects of the aggregations. Because of the limited size of our sample, the number of nations in each of our major analytic

258

Notes to Pages 139-43

categories is very small; however, the number of points used in our base period in making our calculations are quite adequate. Conscious that aggregations do inevitably distort results, we inspected the behavior of each of the countries to see if their individual performances deviated widely from the performances of the analytic groups. And we found that this was not the case. Moreover, the significant behaviors of the analytic groups are found in partition after partition, making clear that aggregation is not the origin of the results obtained. 28 For two works that claim there are great permanent losses as a result of war, see Angell, The Great Illusion, and Nef, War and Human Progress. For the effect of war on demographic structure and postwar behavior of belligerent populations, see Notestein et aI., The Future Population of Europe and the Soviet Union, and Urlanis, Wars and Population. 29 Simon Kuznets, Economic Growth of Nations, p. 43. 30 Charles Tilly, personal communication. 31 One would expect the growth rates of aid recipients to show effects of the foreign help received a year after aid had been received. We calculated the recovery rates of each recipient by subtracting the growth rate in the year aid was received from the growth rate posted in the following year. For example, if a nation posted -15 years of growth in 1948 and -13 years of growth in the following year, the recovery rate was adjudged 2 years. It should be emphasized that negative numbers do not necessarily indicate a lack of growth but rather a lack of recovery. Recovery rates were calculated as follows: Recovery Rate; = Relative Growthi+l - Relative Growth; where i = Recovery Years. For data for these calculations, see Appendix 2. 32 The literature on aid is immense. An excellent review is available in I. M. D. Little and J. M. Clifford, International Aid (London: Allen and Unwin, 1965). The "big push" proposition was derived from Paul Rosenstein-Rodan, "International Aid for Less Developed Countries," Review of Economics and Statistics 43 (May 1961): 107-38, and his classic arguments on the "big push" proposition in "Problems in Industrialization of Eastern and South Eastern Europe," Economic Journal (June 1943): 204-7, elaborated in "Notes on the Theory of the 'Big Push' " in Howard Willis, ed., Economic Development for Latin America (New York: St. Martin's Press, 1961), pp. 57-66.

259

Notes to Pages 143-51

33 The correlation coefficient R is used because we want to indicate whether the influence is positive or negative, but R2 is never above .1. We tested our assumptions as to the amount of time that would elapse before a response would appear by lagging aid and recovery rates for two and three years, and our results remained unchanged. 34 For a classic discussion of the uses of foreign aid, see Milton Friedman, "Foreign Economic Aid: Means and Objectives," and Charles Wolf, Jr., "Economic Aid Reconsidered," in Gustav Ranis, ed., The United States and the Developing Economies, rev. ed. (New York: W. W. Norton, 1973), pp. 25078.

Four

Nuclear Arms Races and Deterrence 1 Theodore C. Sorensen, Kennedy (New York: Harper and Row, 1965), p. 512. 2 Bernard Brodie introduced the classic notion of deterrence in 1946:

"Thus, the first and most vital step in any American security program for the age of atomic bombs is to take measures to guarantee to ourselves in case of attack the possibility of retaliation in kind. The writer in making that statement is not for a moment concerned about who will win the next war in which atomic bombs are used. Thus far the chief purpose of our military establishment has been to win wars. From now its chief purpose must be to avert them. It can have almost no other useful purpose" (Bernard Brodie, ed., The Absolute Weapon: Atomic Power and World Order [New York: Harcourt Brace, 1946], p. 76). See also Bernard Brodie, "The Development of Nuclear Strategy," Center for Arms Control and International Security, Working Paper 11 (University of California, Los Angeles, February 1978). 3 The classic statement of the arms race process was developed by Lewis F. Richardson, Arms and Insecurity (Chicago: Quadrangle Books, 1960); also Samuel Huntington, "Arms Races: Prerequisites and Results," Public Policy 8 (1958): 41-86. For an elegant discussion of current mathematical formalizations, see Dina Zinnes and John Gillespie, eds., Mathematical Models in International Relations (New York: Praeger, 1976), pt. 3. 4 For excellent summaries, see Samuel Huntington, The Common Defense (New York: Columbia University Press, 1961); Bernard Brodie, Strategy in the Missile Age (Princeton, N.J.: Princeton University Press, 1959); Arnold Horelick and Myron Rush, Strategic Power and Soviet Foreign Policy (Chicago: University of

260

Notes to Pages 155-58

Chicago Press, 1966); William Kaufman, Military Policy and National Security (Princeton, N.J.: Princeton University Press, 1956); William Kaufman, The McNamara Strategy (New York: Harper and Row, 1964); Henry Kissinger, Nuclear Weapons and Foreign Policy (New York: Harper and Row, 1957). 5 John Foster Dulles, "Massive Retaliation," in Robert Art and Kenneth Waltz, eds., The Use of Force (Boston: Little, Brown, 1971), pp. 128-32; Albert Wohlstetter, "The Delicate Balance of Terror," Foreign Affairs 37, no. 2 (January 1959): 211-56; Brodie, Strategy in the Missile Age, pp. 267-71; Kaufman, The McNamara Strategy, pp. 114-20; Glenn Snyder, Deterrents and Defense: Toward the Theory of National Security (Princeton, N.J.: Princeton University Press, 1961); Thomas Schelling, Arms and Influence (New Haven: Yale University Press, 1966); Brodie, "The Development of Nuclear Strategy," pp. 3-7. 6 Sorensen, Kennedy, p. 685. 7 Jonathan Pollack, "China as a Nuclear Power," in William H. Overholt, ed., Asia's Nuclear Future (Boulder, Colo.: World View Press, 1977), pp. 44-45; William Griffith, ed., The World and the Great Power Triangles (Cambridge: MIT Press, 1975), p. 30; H. R. Haldeman, The Ends of Power (New York: Quadrangle, 1978). 8 Kennedy, Thirteen Days, pp. 69-70. 9 Nikita Khrushchev, Khrushchev Remembers (Boston: Little, Brown, 1970), p. 497. Finally, we have direct evidence that the weaker side believed that its nuclelJr capability was sufficient to impose unacceptable damage on the United States: "I am emphasizing once more that we already possess so many nuclear weapons, both atomic and hydrogen, and the necessary rockets for sending these weapons to the territory of a potential aggressor, that should any madman launch an attack on our state or on other Socialist states we would be able literally to wipe the country or countries which attack us off the face of the earth" (From an address to the Supreme Soviet, January 14, 1960, in Art and Waltz, The Use of Force, p. 134). 10 In much ofthe deterrence literature, the absolute level of potential destruction is used to estimate the deterrence potential. One presumes that the higher the level of destruction the higher the fear. At one time this level seems to have been arbitrarily agreed upon, at least in the American strategic community: "Washington's criteria for defining assured destruction are arbitrary and conservative. Back in the mid-1960s, the Pentagon hit on roughly 25% of the Soviet population and 45% of Soviet industry as a cut-off

261

Note to page 160

point in targeting weapons. Beyond that point defense planners reckoned that more than double the number of weapons would be needed to gain even marginal increases in levels of destruction" (John Newhouse, Cold Dawn: The Story of SALT [New York: Holt, Rinehart and Winston, 1973], p. 18). This conclusion is supported by more recent official sources: "it is the clear and present ability to destroy the attacker as a viable 20th Century nation and an unwavering will to use these forces in retaliation to a nuclear attack upon ourselves or our allies that provides the deterrent, and not the ability partially to limit damage to ourselves. . .. the first quantitative question which presents itself is: What kind and amount of destruction must we be able to inflict upon the attacker in retaliation to ensure that he would indeed be deterred from initiating such an attack? As I have explained to the Committee in previous years, this question cannot be answered precisely .... In the case of the Soviet Union I would judge that a capability on our part to destroy, say, one-fifth to one-fourth of her population and one-half of her industrial capacity would serve as an effective deterrent. Such a level of destruction would certainly represent intolerable punishment to any 20th Century industrial nation" (Secretary of Defense Robert S. McNamara before the Senate Armed Services Committee on the Fiscal Year, 1%9-70, Defense Program and 1969 Defense Budget, January 22, 1968 [Washington, D.C.: Government Printing Office, 1968], pp. 47-50). And more recently: "According to one approach, planners could simply target major cities, assume that population and industry are strongly correlated with them, and measure effectiveness as a function of the number of people killed and cities destroyed. Thus, as one example, prompt Soviet fatalities of about 30 percent and 200 cities destroyed would constitute a level of retaliation sufficient to assure deterrence" (Donald Rumsfeld, Annual Defense Department Report, 1978 [Washington, D.C.: Government Printing Office, 1978], p. 68). The debate currently is not about the absolute levels needed, but about how to achieve them. 11 Our major interest is, of course, to relate nuclear capability with change of behavior. The limited sample of cases is sufficient for this purpose. An additional reason for our choice is that no complete collection that systematically evaluates the likelihood, prior to 1945, of the escalation of crises to major wars was available at the time of writing. This situation is now being remedied by the work of Charles Goshman, University of Pittsburgh, and Hayward Alker, Massachusetts Institute of Technology.

262

Notes to Pages 160-79

12 We obtained the complete collection of all confrontations from 1945 to 1975 in which the threat of nuclear weapons was and was not involved from the work of Robert Butterworth. We amended Butterworth's evaluation of the likelihood that the conflict would have escalated into a confrontation of great powers. We also decided that war was to be defined as conflict with more than a hundred fatalities. Robert Butterworth, Managing Interstate Conflict, 1945-1974: Data With Synopsis (Pittsburgh: University Center for International Studies, University of Pittsburgh, 1976). 13 Robert Butterworth coded the likelihood of superpower war by coding all conflict in the system in response to the question: How likely was Soviet-American war in this conflict? We made only slight adjustments to this coding. The dichotomy between war and no war during crises indicates whether the crises resulted in more than a thousand casualties or less. For details on these codings, see Butterworth, Managing Conflict, pp. 471-74, variables I, 4, 8. 14 William Griffith interprets the evidence as supporting deterrence. In effect he argues that because the Chinese had no effective nuclear force they gave in. "The first Ussuri incident was probably initiated by the Chinese. Sharp Soviet retaliatory attacks on the U ssuri and later in Sinkiang were followed by what amounted to a successful Soviet ultimatum to Peking either to cease border incidents and begin border negotiations or to face Soviet attack" (The World and the Great Power Triangle, p. 4). And of a later period he writes: "Another question is whether the Soviet will attack China. Launching an attack was considered in Moscow in 1969-70, but it is now less likely because the Chinese are acquiring a minimal nuclear deterrence capability aimed at the Soviet Union, including Moscow. It is their contemporary equivalent of China's Great Wall. The Soviets are more likely biding their time until Mao disappears in the hope they can interfere in the Chinese succession struggle" (ibid., p. 30). 15 Michael Intriligator and Dagobert Brito, "Nuclear Proliferation and the Probability of War," Working Paper, mimeograph (University of California, Los Angeles, 1979). 16 "China, which has been testing MRBM since the mid-1960s, has apparently deployed, mainly in north-western and north-eastern China, about 20 operational missiles with a range of up to 1,000 miles" (The International Institute for Strategic Studies, The Military Balance, 1971-1972 [London, 1971], p. 40). 17 For an earlier analysis cha\1enging the operation of deterrence, see Organski, World Politics, chap. 5.

263

Notes to Page 182

18 It should be noted that arms races are thought to be destabilizing to the international system for a very specific reason. The reason is not obvious and runs as follows: because large-scale arming is very costly, the economic strain of competition in which each of the competitors seek to outdo the others is very high. The weak nation is likely to tire and fall behind and is threatened with defeat if war should come. The only way out is to force a fight before it is too late, or accept the eventual ascendance of one's rival. The argument is faulty so far as nuclear weapons are concerned. We shall return to this important point in the conclusion. Cf. Paul Smoker, "Fear in the Arms Race: A Mathematical Study," Journal of Peace Research 1, no. 1 (1964): 55-64. 19 Over four decades ago Lewis Richardson, in his classic Arms and Insecurity, referred to above, in n. 3, stated in formal terms that increases in the allocations for arms in the budget of one side are the result of increases in the allocation for defense in the budget of its adversary. This insight was considerably extended in its application to the nuclear world by introducing game theoretical logic to arms races and elegant demonstrations that contenders caught in a "prisoner's dilemma," fearing the consequences of cooperation, evince behavior resembling arms races. Such demonstrations also indicated how such fear could be reduced and escalation reversed. See also Anatol Rapoport, "Lewis R. Richardson's Mathematical Theory of War," Journal of Conflict Resolution 1,3 (September 1957): 249-99, also his Fights, Games, and Debates (Ann Arbor, Mich.: The University of Michigan Press, 1961), chap. 1, and Anatol Rapoport and Albert M. Chammah, Prisoner's Dilemma (Ann Arbor, Mich.: The University of Michigan Press, 1965). For an informal specification, see the classic presentation referred to above in n. 3, Samuel Huntington, "Arms Races: Prerequisites and Results," pp. 41-86. For an early survey of the vast literature, see Peter Busch, "Mathematical Models of Arms Races," in Bruce Russett, What Price Vigilance? (New Haven: Yale University Press, 1970), pp. 193-233. For an important discussion of changes of outcomes due to variations in payoffs, see Robert Luce and Howard Raiffa, Games and Decisions (New York: John Wiley, 1957), pp. 94-101, and for a full exploration of changes resulting from varying assumptions see Dina Zinnes, John Gillespie, and Michael Rubinson, "A Reinterpretation of the Richardson Arms-Race Model," in Zinnes and Gillespie, Mathematic Models in International Relations, pp. 189-216. Finally, if one assumes cybernetic, rather than rational, processes, substantial changes occur, but the relationship between interaction and

264

Notes to Pages 183-84

deterrence is not affected. See John Steinbruner, "Beyond Rational Deterrence: The Struggle for New Conceptions," World Politics 28, 2 (January 1976): 223-45. 20 Joseph Pechman, ed., Setting National Priorities, The 1979 Budget (Washington, D.C.: The Brookings Institution, 1978), p. 259. 21 Incrementalism has become a dominant theme in many areas of domestic budgeting and the models proposed by Wildavsky and his collaborators have been applied with modifications to defense policy. The work ·of Crecine and Kanter using budgetary data, and that of Tammen, Allison, Morris, and Halperin analyzing how decisions are made in the development of particular weapon systems, offer impressive testimony that the overall fluctuations of the defense budget and decisions on the development of weapon systems are made largely in response to internal pressures. The most general treatment is Otto Davis, M. A. H. Dempster, and Aaron Wildavsky, "A Theory of the Budgetary Process," APSR 60, 3 (September 1966); Aaron Wildavsky, Budgeting, A Comparative Theory of Budgetary Processes (}3oston: Little, Brown, 1975), pp. 47-69, 344-58; see also Morton Halperin, Bureaucratic Politics and Foreign Policy (Washington, D.C.: The Brookings Institution, 1974), and Graham Allison and Morton Halperin, "Bureaucratic Politics: A Paradigm and Some Policy Implications," in R. Tanter and R. Ullman, eds., Theory and Policy in International Relations (Princeton, N.J.: Princeton University Press, 1972). For specific studies of defense budgeting, see John Crecine and Gregory Fisher, "On Resource Allocation Processes in the U.S. Department of Defense," Institute for Public Policy Studies Discussion Paper No. 31 (October 1971), and John Crecine, "Fiscal and Organizational Determinants ofthe Size and Shape of the U.S. Defense Budget," Institute for Public Policy Studies Discussion Paper No. 69 (April 1975); Arnold Kanter, "Congress and the Defense Budget: 196~70," APSR 66,1 (March 1972): 129-43; Graham Allison and Fredric Morris, "Armaments and Arms Control: Exploring the Determinants of Military Weapons," Daedalus 104, 3 (Summer 1975): 99-189; Ron Tammen, MIRV and the Arms Race: An Interpretation of Defense Strategy (New York: Praeger, 1973); Robert Art, The TFX Decision: McNamara and the Military (Boston: Little, Brown, 1968). 22 Albert Wohlstetter, seeking to test for the presence of a nuclear arms race between the United States and the USSR, examined American expenditures on nuclear weapons and found that, during the time for which the data were available, allocations went up in the first quinquennium but came down in the second. Albert

265

Note to Page 185

Wohlstetter, "Is There a Strategic Arms Race?" Foreign Policy 15 (Summer 1974): 2-21, and "Rivals, But No 'Race'," Foreign Policy 16 (Fall 1974): 48-81. His findings, however, could not rule out the possibility that the curvilinearity of American expenditure is part of a cyclical movement hidden from view by the inadequate length of the available series, nor the possibility that a relation between Soviet and American expenditures would have emerged had Russian data been available. Some scholars have attempted to test simultaneously the effects that internal and external factors may have on decisions to allocate resources to the building of arms, but have fallen short of their goal due to the seeming impossibility of constructing a model that, given the data, could disentangle "internal" and "external" influences from one another. Two excellent studies tried. One by Wagner, Perkins, and Taagepera used Richardson's own data to show that either an arms-race model or an incremental model could account equally well for the behavior of military budgets prior to World War II. Thus both models could be right, and there was no way to discriminate between them. David Wagner, Ronald Perkins, and Rein Taagepera, "Complete Solution to Richardson's Arms Race Equation," Journal of Peace Science 1, no. 2 (Spring 1975): 159-72. A second study, by C. Ostrom, used two series of total defense expenditures by the United States and the Soviet Union, for the period 1954--75. He confronted the two approaches and found that the incremental and arms-race models performed about as well in accounting for the observed pattern of defense spending in the two countries as a random model would have done. He rightly concluded on this basis that one could not choose between the two approaches. Again, one was back at the beginning. Charles Ostrom, Jr., "Evaluating Alternative Foreign Policy Decision-Making Models," Journal of Conflict Resolution 21, 2 (June 1977): 235-65. 23 The Defense Department uses similar breakdowns starting in 1962 in the annual reports. For example, Annual Defense Department Report. FY 1974 (Washington, D.C.: Government Printing Office, 1974), table 1, p. 118. The Central Intelligence Agency adopted a similar strategy, estimating in United States and Soviet currencies allocation of resources for nuclear arms. For example, see" A Dollar Comparison of Soviet and U. S. Defense Activities, 19651975," SR76-10053 (February 1976), fig. 3, p. 3 (there are a number of preceding pUblications also containing relevant estimates); "Estimating Soviet Defense Spendings in Rllbles, 1970--1975," SR76-101121U (May 1976). Many academic have

266

Notes to Pages 186-87

also adopted similar breakdowns for purposes of arms-race comparisons; see Michael Squires, "Three Models of Arms Races," in Zinnes and Gillespie, Mathematical Models in International Relations, pp. 260-1; and Albert Wohlstetter, "Rivals, But No 'Race'," p. 62. 24 Most observers agree that the gap in strategic weapons has been closing. See Fred Payne, "The Strategic Nuclear Balance," Survival (May-June 1957): 109-10; and Donald Rumsfeld, Annual Defense Department Report, 1978, pp. 60-61. The assumption of technological equivalence is common. "As you know, we have been able to observe a number of New Soviet systems which use highly advanced technology and production techniques: the Foxbat aircraft, nuclear-powered ballistic missile submarines, new types of attack submarines, new radars and missiles both for missile and for air defense, anti-ship missiles, new ASW ships equipped for helicopter operations, and smaller items such as the advanced rocket-launcher introduced effectively into Vietnam. The technology of many of these systems is comparable to the U.S. technology. In some cases, however, our current systems are clearly more advanced" (Defense Program and Budget, FY 1971, A Statement by Secretary of Defense Melvin R. Laird before the House Subcommittee on DOD Appropriations, p. 67 [Washington, D.C.: Government Printing Office, 1971]). Also, see William Baugh, "An Operations Analysis Model for the Study of Nuclear Missile System Policies," in Zinnes and Gillespie, Mathematical Models in International Relations, p. 277. 25 For an excellent review of problems of comparison, see Andrew Marshall, "Estimating Soviet Defense Spending," Survival 18, no. 2 (March-April 1976): 73-79; see also Alec Nove, "Soviet Defense Spending," Survival 13, no. 10 (October 1971): 328-32, and Michael Boretsky and Alec Nove, "The Growth of Soviet Arms Technology-A Debate," Survival 14, no. 4 (July-August 1972): 169-77; W. T. Lee, "Soviet Defense Expenditures for 19551975," Tempo, General Electric Company, mimeograph (July 31, 1975); William Colby (pp. 21-23) and Daniel Graham (pp. 92-95) in Joint Economic Committee, Allocation of Resources in the Soviet Union and China, 1975 (Washington, D.C.: Government Printing Office, 1975); and Paul Cockle, "Analysing Soviet Defense Spending: The Debate in Perspective," Survival 20, no. 5 (September-October 1978): 209-19. 26 Ideally, one should compare the nuclear capabilities of two countries by evaluating every individual weapon system across the whole range of nuclear weapons systems in the arsenal of each

267

Notes to Pages 189-93

country; compare such systems with their counterparts in the opposing country; and aggregate the comparisons to determine the degree and manner in which each competitor has reacted to the development of specific weapon systems by the other. One must also be able to aggregate the results of such comparisons to detect overall reaction patterns rather than individual systems responses. One should note that the detailed comparison cannot be undertaken without the development of the necessary criteria and full access to the characteristics of each weapons system. This type of work is impossible for scholars outside of the intelligence community. 27 This conclusion is supported by the difference in commitment of resources to strategic systems. The Central Intelligence Agency reports that, between 1967 and 1977, "Within the respective intercontinental attack forces, a substantial difference in emphasis on weapons is apparent: Almost 60% of the estimated dollar costs of Soviet activities during the period were for the ICBM force, compared to only about 20% for the U.S. On the other hand, outlays of the U.S. bomber force comprised about 40%, compared to a Soviet share of less than 5% (excluding the Backfire aircraft). While the Soviets exceeded the U.S. levels of activities for ICBMs in every year of the period and for submarines in all but two, U.S. outlays for bombers were higher every year" (Central Intelligence Agency, "A Dollar Cost Comparison of Soviet and U.S. Defense Activities, 1967-1977," SR78-10002m [January 1978], p. 6, and fig. 3, p. 9). 28 For details, see Appendix 3, model AI.O. 29 Given that depreciation is assumed to be constant across the two nations, the ratios and the absolute difference between the estimated constant coefficients can be obtained from the two equations (3.0 and 3.1) for the Soviet Union and the United States. 30 The strategic nuclear offensive series used in this study was constructed from data provided to the Joint Economic Committee, Subcommittee on the Priorities and Economy in Government, the United States Senate. Data was taken directly from declassified 1964, 1969, and 1978 intelligence documents currently retained in the committee files that string together Defense Department and Intelligence reports from 1945 to 1977. 31 See Appendix 3, the section on analysis of residuals. 32 Brodie, The Absolute Weapon. 33 An examination of the residuals showed outliers for the United States in 1952-54 and for the USSR in 1965. Without altering the

268

Notes to Pages 194-97

model, we reestimated the equation and excluded these observations, in the hope of obtaining more stable results that could be analyzed in conjunction with the original sample. The results are presented in table 4.4. In the case ofthe USSR we found no change in the coefficients as a result of the exclusion of selected years. For the United States, on the other hand, the coefficients were found to be more stable, had smaller standard errors, and yielded a stronger overall fit. Table 4.4

Effects of External and Internal Pressures on U.S. and USSR Strategic Budgets,

1955-64, 1966-76 U.S. Off. Exp.m = 7.827 + .720U.S. Off. Exp.(t·!) - .151 USSR Off. Exp.m Standard error (2.124) (.092) (.125) Significance (0.001) (.000) (.002) R2 = .92 Standard error = 1.50 Significance = .000 N == 21 USSR Off. Exp.m = 1.932 + .920 USSR Off. Exp.(I·!) - .038 U .S. Off. Exp.u-I) Standard error (1.289) (.790) (.059) Significance (0.151) (.000) (.520) R2 = .94 Standard error = .97 Significance = .000 N = 21

34 See Appendix 3. 35 We gratefully appreciate the assistance of Marcial Losada and the use of the MIFAS (Multivariate Interactive Fourier Analytic Synthesizer) program he and Keith Sentis developed at the University of Michigan. Fourier synthesis is used because it superimposes algebraic summations of a series of sinusoidal waves that represent a significant contribution to the variance of the phenomena expressed in the time series. Fourier synthesis translates back simplified information from the frequency domain to the time domain and allows easy access to the consistent and prominent features observed in the data. Fourier synthesis can be meaningfully applied only to stationary series. The transformation of original data to scores with zero mean and unit variance does not affect our tests for interaction since the graphs produced preserve the configuration of the original data intact and can be meaningfully evaluated. For details of the program, see Marcial Losada, "EventSystem Decomposition: Predicting Employee Turnover in Manufacturing" (Ph.D. diss., The University of Michigan, 1977). 36 With Fourier analysis the variance explained rose from an R 2 of .72 to an R2 of .75. 37 The distortions created by cross-national comparisons are magnified when we move away from direct comparison of weapon values to estimates of effort. The ratios we used are congruent with

269

Notes to Pages 201-25

similar ones obtained from estimates that use rubles rather than dollars. See Central Intelligence Agency, "Estimating Soviet Defense Spending in Rubles, 1970-1975," SR76-10121v (May 1976). The GNP data were obtained from the World Bank, World Tables, 1976 (Baltimore: Johns Hopkins University Press, 1976). Slight readjustments of constant values were required. 38 For a formal demonstration of the conditions under which nuclear proliferation insures stable deterrence rather than increases the probability of war, see Intriligator and Brito, "Nuclear Proliferation and the Probability of War."

Five

Conclusion 1 Percentages for the United States and Japan are from the United States Strategic Bombing Survey, The Effects of Strategic Bombing on Japan's War Economy, p. 16; percentages for Germany and the United Kingdom from Knorr, The War Potential of Nations, p.239. 2 In World War II, England also fought in order to eradicate Hitler's scourge from Europe, and many justifiably feel that World War II was the only modem war in which moral right was all on one side and evil all on the other. But one should not lose sight of the fact that eliminating Hitler was only one of many reasons why the Allies fought Germany. And one should not forget that the same countries fought Germany in World War I when Hitler was not a factor. 3 In such speculation any evidence is welcome. We have only a crumb but a fascinating crumb, in the Cuban missile crisis. It bears clearly on the question of why the United States was ready to go to nuclear war, if that were necessary, to stop the attempt of the Soviet Union to put missiles in Cuba. The issue has always posed the question of the missiles in Cuba as a military advantage to the Russians that the United States could not permit. But the problem was not the military balance at all. It is worth quoting Theodore Sorensen, President Kennedy's principal adviser on the point. "These Cuban missiles alone, in view of all the other megatonnage the Soviets were capable of unleashing upon us, did not substantially alter the strategic balance in fact . ... But that balance would have been substantially altered in appearance; and in matters of national will and world leadership, as the President said later, such appearances contribute to reality." And at another point: "The President ... was concerned less about the missiles' military implications than with their effect on the global political balance" (Kennedy, pp. 678, 683). Such comments fit hand in glove with what we are saying. What

270

Notes to Pages 227-32

the United States was ready to fight about was its position in the world, its dominance over the international system outside the Communist bloc. The political balance was threatened because the establishment of a Soviet base in America's backyard would, in the eyes of the world, put the United States and the Soviet Union on the same level. What America had done on Soviet frontiers the Soviet Union, now, would do on America's. This was intolerable to the United States. It was bad enough to have Cuba desert the American system; it was intolerable to have her become a Russian base. In short, the United States was ready to fight when her political position in the world was called into question and her domination of the international order was threatened. On the other hand, the Soviet Union was not ready to fight, because she knew she was not nearly as strong as the United States; her claims that she was were pure bombast and the rest of the world did not regard her as the dominant nation. The world certainly did not regard American missiles on Soviet borders and Soviet missiles on American borders in the same light.

Appendix 1 1 Raja Chelliah, "Trends in Taxation in Developing Countries," IMF Staff Papers (July 1, 1971), pp. 254-331; Bahl, "A Regression Approach," pp. 57~1O. 2 Chelliah, "Trends in Taxation," pp. 254-331. 3 See Pietro Balestra and Marc Nerlove, "Pooling Cross Section and Time Series Data in the Estimation of a Dynam'ic Model: The Demand for Natural Gas," Econometrica 34, no, 3 (July 1966).

Bibliography

271

Agency for International Development, Office of Statistics and Reports. u.s. Economic Assistance Programs Administered by the Agency for International Development and Predecessor Agencies, April 3, 1948-June 30, 1971. Washington, D.C.: AID, 1972. Alcock, Norman, and Newcombe, Alan. "Perceptions of National Power." Journal of Conflict Resolution 14 (1970): 335-43. Aldrich, John, and Cnudde, Charles F. "Probing the Bounds of Conventional Wisdom: A Comparison of Regression, Probit, and Discriminant Analysis." American Journal of Political Science 19 (1975). Allison, Graham, and Halperin, Morton. "Bureaucratic Politics: A Paradigm and Some Policy Implications." In Theory and Policy in International Relations, ed. R. Tanter and R. Ullman. Princeton: Princeton University Press, 1972. Allison, Graham, and Morris, Fredric. "Armaments and Arms Control: Exploring the Determinants of Military Weapons." Daedalus 104 (1975): 99-189. Angell, Norman. The Great Illusion. New York: Putnam, 1933. Ardant, Gabriel. "Financial Policy and Economic Infrastructure of Modem States and Nations." In The Formation of National States in Western Europe, ed. Charles Tilly. Princeton: Princeton University Press, 1975. Arms Control and Disarmament Agency. World Military Expenditures, 1970. Washington, D.C.: Government Printing Office, 1970. Aron, Raymond. Peace and War. New York: Doubleday, 1966. Art, Robert. The TFX Decision: McNamara and the Military. Boston: Little, Brown, 1968.

272

Bibliography

Bahl, Roy. "A Regression Approach in Tax Effort and Tax Ratio Analysis." International Monetary Fund Staff Papers, November 1971, pp. 57~1O. Banks, Arthur. Cross Polity Time and Series Data. Cambridge: MIT Press, 1971. Baugh, William. "An Operations Analysis Model for the Study of Nuclear Missile System Policies." In Mathematic Models in International Relations, ed. Dina Zinnes and John Gillespie. New York: Praeger, 1976. Binder, L., and LaPalombara, Joseph, eds. Crises and Sequences in Political Development. Princeton: Princeton University Press, 1974. Blainey, Geoffrey. The Causes of War. New York: The Free Press, 1975. Boulding, Kenneth. A Primer of Social Dynamics. New York: The Free Press, 1970. Box, George, and Jenkins, Gwilym. Time Series Analysis. San Francisco: Holden-Day, 1970. Brodie, Bernard. Strategy in the Missile Age. Princeton: Princeton University Press, 1959. - - - . "The Development of Nuclear Strategy." Center for Arms Control and International Security, Working Paper 11. Los Angeles: UCLA, 1978. - - - , ed. The Absolute Weapon: Atomic Power and World Order. New York: Harcourt Brace, 1946. Bueno de Mesquita, Bruce. "Measuring Systemic Polarity." Journal of Conflict Resolution 19 (1975): 187-216. - - - . "The Effects of Systematic Polarization on the Probability and Duration of War. " Paper delivered at the Annual Meeting of the International Studies Association. Toronto, Canada: February, 1975. - - - . The War Trap. New Haven: Yale University Press, forthcoming. Bums, Arthur, ed. The Business Cycle in a Changing World. New York: National Bureau of Economic Research, 1969. Busch, Peter. "Mathematical Models of Arms Races." In What Price Vigilance?, ed. Bruce Russett. New Haven: Yale University Press, 1970. Butterworth, Robert. Managing Interstate Conflict, 19451974: Data With Synopsis. Pittsburgh: University Center for International Studies, University of Pittsburgh, 1976. Carr, E. H. The Twenty Years' Crisis, 1919-1939. New York: Harper and Row, 1974.

273

Bibliography

CheUiah, Raja. "Trends in Taxation in Developing Countries." International Monetary Fund Staff Papers, 1 July 1971. Chelliah, Raja; Baas, Hassel; and Kelly, Margaret. "Tax Ratios and Tax Effort in Developing Countries, 19691971." International Monetary Fund Staff Papers, May 1974. Choucri, Nazli, and Meadows, Dennis. International Implications of Technological Development and Population Growth: A Simulated Model of International Conflict. Cambridge, Mass.: Center for International Studies, MIT Press, 1971. Choucri, Nazli, and North, Robert. Nations in Conflict. San Francisco: W. H. Freeman, 1974. Cipolla, Carlo M., ed. Contemporary Economics, The Fontana Economic History of Europe, vol. 6, pt. 2. New York: Barnes and Noble, 1977. Claude, Inis. Power and International Relations. New York: Random House, 1962. Cline, Ray. World Power Assessment. Washington, D.C.: The Centerfor Strategic and International Studies, 1975. Cockle, Paul. "Analysing Soviet Defense Spending: The Debate in Perspective." Survival 20 (1978). Colby, William. Allocation of Resources in the Soviet Union and China, 1975. Testimony in Hearings before the Subcommittee on Priorities and Economy in Goyernment of the Joint Economic Committee, 94th Congress, First Session, June 18, 1975. Washington, D.C.: Government Printing Office, 1975. Coleman, J. "The Development Syndrome: DifferentiationEquality-Capacity." In Crises and Sequences in Political Development, ed. L. Bind~r and Joseph LaPalombara. Princeton: Princeton University Press, 1974. Coplin, William. Introduction to International Politics. Chicago: Markham Publishing Co., 1970. Coplin, W., and Kegly, C. Analyzing International Relations. New York: Praeger, 1975. Crecine, John. "Fiscal and Organizational Determinants of the Size and Shape of the V.S. Defense Budget." Institute for Public Policy Studies Discussion Paper No. 69 (April 1975). Crecine, John, and Fisher, Gregory. "On Resource Allocation Processes in the V. S. Department of Defense." Institute for Public Policy Studies Discussion Paper No. 31 (October 1971).

274

Bibliography

Cutright, Phillips. "Political Structure, Economic Development, and National Social Security Programs." In Macro-Quantitative Analysis: Conflict: Development, Democratization, ed. J. V. Gillespie and B. A. Mesvold. Beverly Hills, Sage Publications, 1971. Davis, Otto; Dempster, M. A. H.; and Wildavsky, Aaron. "A Theory of the Budgetary Process." The American Political Science Review 60 (September 1966). Defense Department. Annual Defense Department Report FY 1974. Washington: UPSO, 1974. Deutsch, Karl. "Social Mobilization and Political Development." In Political Development and Social Change, ed. Jason Finkle and Richard Gable. New York: John Wiley and Sons, 1966. - - - . The Analysis of International Relations. Englewood Cliffs, N.J.: Prentice-Hall, 1967. Dulles, John Foster. "Massive Retaliation." In The Use of Force, ed. Robert Art and Kenneth Waltz. Boston: Little, Brown, 1971. DuMouchel, William H. "The Regression of a Dichotomous Variable." Mimeograph. Ann Arbor: Institute for Social Research, The University of Michigan, 1973. Ferris, Wayne. The Power Capabilities of Nation States. Lexington, Mass.: Lexington Books, 1973. Finkle, Jason, and Gable, Richard, eds. Political Development and Social Change. New York: John Wiley and Sons, 1966. Fox, William, ed. Theoretical Aspects of International Relations. Notre Dame: Notre Dame University Press, 1959. Friedman, Milton. "Foreign Economic Aid: Means and Objectives." In The United States and the Developing Economies, ed. Gustav Ranis. Rev. ed. New York: W. W. Norton, 1973. Frumkin, Gregory. Population Changes in Europe Since 1939. New York: Augustus M. Kelley, 1951. Fuchs, Wilhelm. Formein Zur Macht. Stuttgart: Deutsche Varlage-Anstaldt, 1965. Gamham, David. "The Power Parity and Lethal International Violence, 1969-1973." Journal of Conflict Resolution, 20 (1976): 379-394. German, Clifford. "A Tentative Evaluation of World Power." Journal of Conflict Resolution, March 1960, pp. 138-44.

275

Bibliography

Gillespie, J. Y., and Mesvold, B. A., eds. MacroQuantitative Analysis: Conflict, Development, Democratization, vol. 1. Beverly Hills: Sage Publications, 1971. Graham, Daniel. Allocation of Resources in the Soviet Union and China. 1975. Washington, D.C.: Government Printing Office, 1975. Griffith, William, ed. The World of the Great Power Triangles. Cambridge: MIT Press, 1975. Gurr, Ted R. Why Men Rebel. Princeton: Princeton University Press, 1970. - - - . "Persistence and Change in Political Systems, 1800-1971." American Political Science Review, December 1974. Haas, E. "The Balance of Power: Prescription, Concept or Propaganda." World Politics 5 (1953). Haggard, M. T. "United States Expenditures in Indochina and in Korea." Conl?ressional Research Record. Washington, D.C.: Library of Congress, April 1975. Haldeman, H. R. The Ends of Power. New York: Quadrangle, 1978. Halperin, Morton. Bureaucratic Politics and Foreign Policy. Washington, D.C.: Brookings Institution, 1974. Heiss, Klaus P.; Knorr, Klaus; and Morgenstern, Oskar. Long-Term Projections of Political and Military Power. Princeton: Mathematica Inc., 1973. Herz, John. International Politics in the Atomic Age. New York: Columbia University Press, 1959. Hitch, Charles, and McKean, Roland. The Economics of Defense in the Nuclear Age. Cambridge: Harvard University Press, 1967. Holsti, K. J. International Politics: A Framework for Analysis. Englewood Cliffs, N.J.: Prentice-Hall, 1967. Holsti, Ole. "The Belief System and National Images: A Case Study." Journal of Conflict Resolution 6 (1962): 244-52. Holt, R., and Turner, J. "Crises and Sequences in Collective Theory of Development." American Political Science Review 69 (September 1975): 979-94. Horelick, Arnold, and Rush, Myron. Strategic Power and Soviet Foreign Policy. Chicago: University of Chicago Press, 1966. Huntington, Samuel. "Arms Races: Prerequisites and Resuits." Public Policy (1958): 41-86.

276

Bibliography

- - - . The Common Defense. New York: Columbia University Press, 1961. - - - . Political Order in Changing Societies. New Haven: Yale University Press, 1967. International Institute for Strategic Studies. The Military Balance 1971-1972. London, 1971. Intriligator, Michael, and Brito, Dagobert. "Nuclear Proliferation and the Probability of War." Mimeographed. Los Angeles: University of California, 1979. Kanter, Arnold. "Congress and the Defense Budget: 196a-70." American Political Science Review 66 (March 1972): 129-43. Kaufman, William. Military Policy and National Security. Princeton: Princeton University Press, 1956. - - - . The McNamara Strategy. New York: Harper and Row, 1964. Kennedy, Robert F. Thirteen Days. New York: W. W. Norton, 1969. Keynes, John. The Economic Consequences of the Peace. New York: Harcourt Brace and Rowe, 1920. Khrushchev, Nikita. Khrushchev Remembers. Boston: Little, Brown, 1970. Kindleberger, Charles. "International Political Theory from the Outside." In Theoretical Aspects of International Relations. ed. William Fox. Notre Dame: Notre Dame University Press, 1959. Kissinger, Henry. Nuclear Weapons and Foreign Policy. New York: Harper and Row, 1957. Klein, Lawrence. "The Procedure of Economic Prediction: Standard Achievement, Potential." Paper presented at the Conference on the Economic Outlook for 1973, 16-17 November 1972, University of Michigan. Knorr, Klaus. The War Potential of Nations. Princeton: Princeton University Press, 1956. - - - . Military Power and Potential. Cambridge, Mass.: D. C. Heath, 1970. Kugler, Jacek. "The Consequences of War: Fluctuations in National Capabilities Following Major Wars, 188a-1970." Ph.D. dissertation, University of Michigan, 1973. Kuznets, Simon. Economic Growth of Nations. Cambridge: Harvard University Press, Belknap Press, 1971. Laird, Melvin R. Defense Program and Budget FY 1971. Statement before the House Subcommittee on DOD Ap-

277

Bibliography

propriations. Washington, D.C.: Government Printing Office, 1971. Lee, W. T. The Estimation of Soviet Defense Expenditures, 1955-1975. New York: Praeger, 1977. Liska, George. International Equilibrium: A Theoretical Essay on the Politics and Organization of Security. Cambridge, Mass.: Harvard University Press, 1957. Little, I. M. D., and Clifford, J. M. International Aid. London: Allen and Unwin, 1965. Losada, Marcial. "Event-System Decomposition: Predicting Employee Turnover in Manufacturing." Ph.D. dissertation. University of Michigan, 1977. Lotz, Jorgen, and Morss, Elliott. "Measuring 'Tax Effort' in Developing Countries." International Monetary Fund Staff Papers, July 1971, pp. 254-331. Luce, Robert, and Raiffa, Howard. Games and Decisions. New York: John Wiley and Sons, 1957. Lucier, Charles. "Power and the Balance." Ph.D. dissertation. University of Rochester, 1974. Maddison, Angus. Economic Growth in the West: Comparative Experience in Europe and North America. New York: The Twentieth Century Fund, 1964. - - - . Economic Growth in Japan and the USSR. London: Allen and Unwin, 1969. - - - . "Trends in Output and Welfare." In Contemporary Economics, The Fontana Economic History of Europe vol. 6, pt. 2, ed. Carlo M. Cipolla. 1977. New York: Barnes and Noble, 1977. Marshall, Andrew. "Estimating Soviet Defense Spending." Survival 18 (1976): 73-79. McKelvey, Richard, and Zavoina, William. "A Statistical Model for the Analysis of Ordinal Level Dependent Variables." Mimeographed. University of Rochester, 1974. McNamara, Robert S. Speech before the Senate Armed Services Committee on the Fiscal Year, 1969-70, Defense Program and 1969 Defense Budget, January 22, 196ft Washington, D.C.: Government Printing Office, 1968. Midlarsky, Manus. On War. New York: The Free Press, 1975. Morgenthau, Hans. Politics Among Nations. 3d. ed. New York: Alfred A. Knopf, 1960. Nash, Keir, ed. Governance and Population: The Governmental Implications of Population Change. Commission

278

Bibliography

on Population and Growth and the American Future, Research Reports, ed. Keir Nash. Washington, D.C.: Government Printing Office, 1972. Newhouse, John. Cold Dawn: The Story of SALT. New York: Holt, Rinehart and Winston, 1973. Nef, John. War and Human Progress. Cambridge: Harvard University Press, 1950. Notestein, Frank; Taeuber, Irene; Kirk, Dudley; Coale, Ansley; and Kiser, Louise. The Future Population of Europe and the Soviet Union. Geneva: League of Nations, 1944. Organski, A. F. K. Stages of Political Development. New York: Alfred A. Knopf, 1965. - - - . World Politics. 2d ed. New York: Alfred A. Knopf, 1968. Organski, A. F. K.; Bueno de Mesquita, Bruce; and Lamborn, Alan. "The Effective Population in International Politics." In Governance and Population: The Governmental Implications of Population Change. Commission on Population and Growth and the American Future, Research Reports, ed. Keir Nash. Washington, D.C.: Government Printing Office, 1972. Ostrom, Charles, Jr. "Evaluating Alternative Foreign Policy Decision-Making Models." Journal of Conflict Resolution 21 (June 1977): 235-65. Overholt, William H., ed. Asia's Nuclear Future. Boulder, Colo.: World View Press, 1977. Payne, Fred. "The Strategic Nuclear Balance." Survival (May-June 1957): 109-10. Pechman, Joseph, ed. Setting National Priorities, The 1979 Budget. Washington, D.C.: Brookings Institution, 1978. Pollack, Jonathan. "China as a Nuclear Power." In Asia's Nuclear Future, ed. William H. Overholt. Boulder, Colo.: World View Press, 1977. Pye, Lucian. "The Concept of Political Development." The Annals of the American Academy of Political and Social Science (March 1965). Ranis, Gustav, ed. The United States and the Developing Economies. Rev. ed. New York: W. W. Norton, 1973. Rao, PotIuri M., and Miller, Roger L. Applied Econometrics. Wadsworth Publishing, 1971. Rapoport, Anatol. "Lewis R. Richardson's Mathematical Theory of War." Journal of Conflict Resolution 1 (September 1957): 249-99.

279

Bibliography

- - - . Fights, Games, and Debates. Ann Arbor: University of Michigan Press, 1961. Rapoport, Anatol, and Chammah, Albert M. Prisoner's Dilemma. Ann Arbor: University of Michigan Press, 1965. Ray, James, and Singer, J. David. "Measuring the Concentration of Power in the International System." Sociological Methods and Research 1 (May 1973): 403-37. Richardson, Lewis F. Arms and Security. Chicago: Quadrangle Books, 1960. Riker, William. The Theory of Political Coalitions. New Haven: Yale University Press, 1962. Rosen, Steven. "War, Power and the Willingness to Suffer." In Peace, W(tr and Numbers, ed. Bruce Russett. Beverly Hills: Sage Publications, 1972. Rosecrance, Richard; Alexandroff, Alan; Healy, Brian; and Stein, Arthur. Power, Balance of Power, and Status in Nineteenth Century International Relations. Beverly Hills: Sage Publications, 1974. Rosenstein-Rodan, Paul. "Problems in Industrialization of Eastern and South Eastern Europe." Economic Journal (June 1943): 204-7. - - - . "International Aid for Less Developed Countries." Review of Economics and Statistics 43 (May 1961): 107-38. - - - . "Notes on the Theory of the 'Big Push'." In Economic Development for Latin America, ed. Howard Willis. New York: St. Martin's Press, 1961. Rumsfeld, Donald. Annual Defense Department Report, 1978. Washington, D.C.: Government Printing Office, 1978. Russett, Bruce. What Price Vigilance? New Haven: Yale University Press, 1970. - - - , editor. Peace, War, and Numbers. Beverly Hills: Sage Publications, 1972. Russett, Bruce; Singer, J. D.; and Small, Melvin. "National Political Units in the Twentieth Century: Standard List." The American Political Science Review 62 (1968): 932-51. Schelling, Thomas. Arms and Influence. New Haven: Yale University Press, 1966. Singer, J. D., ed. Quantitative International Politics. New York: The Free Press, 1968. Singer, J. David, and Small, Melvin. "The Composition and Status Ordering of the International System, 1815-1940." World Politics 18 (1966).

280

Bibliography

- - - . "Alliance Aggregation and the Onset of War, 1815-1945." In Quantitative International Politics. ed. J. D. Singer. New York: The Free Press. 1968. - - . Wages of War. 1816-1965: A Statistical Handbook. New York: John Wiley and Sons, 1972. - - - . "War in History and the State of the World Message." In Analyzing International Relations. ed. W. Coplin and C. Kegly. New York: Praeger, 1975. Singer, J. David; Bremer, Stuart; and Stuckey, John. "Capability Distribution, Uncertainty and Major Power War, 1820-1965." In Peace. War. and Numbers. ed. Bruce Russett. Beverly Hills: Sage Publications, 1972. Stockholm International Peace Research Institute. World Armament and Disarmament. SIPRI Yearbook, 1974-77. Cambridge, Mass.: MIT Press. Smoker, Paul. "Fear in the Arms Race: A Mathematical Study." Journal of Peace Research 1 (1964): 55-64. Snyder, Glenn. Deterrents and Defense: Toward the Theory of National Security. Princeton: Princeton University Press, 1961. Sorensen, Theodore. Kennedy. New York: Harper and Row, 1965. Squires, Michael. "Three Models of Arms Races." In Mathematical Models in International Relations. ed. Dina Zinnes and John Gillespie. New York: Praeger, 1976.

Steinbruner. John. "Beyond Rational Deterrence: The Struggle for New Conceptions." World Politics 28 (January 1976): 223-45. Tammen, Ron. MIRV and the Arms Race: An Interpretation of Defense Strategy. New York: Praeger, 1973. Tanter, R., and Ullman, R., eds. Theory and Policy in International Relations. Princeton: Princeton University Press, 1972. Tilly, Charles, ed. The Formation of National States in Western Europe. Princeton: Princeton University Press, 1975.

United Nations. The World Population Situation in 1970. Population Studies No. 49. New York: United Nations, 1971. - - - . Demographic Yearbook. 1974. New York: United Nations: 1976. United Nations Research Institute for Social Development.

281

Bibliography

Contents and Measurement of Socio-Economic Development. New York: Praeger, 1972. United States Central Intelligence Agency. "A Dollar Comparison of Soviet and U.S. Defense Activities, 19651975." Washington, D.C.: Government Printing Office. SR76-10053 (February 1976). - - - . "Estimating Soviet Defense Spendings in Rubles, 1970--1975." Washington, D.C.: Government Printing Office. SR76-101121U (May 1976). United States Strategic Bombing Survey. The Effects of Strategic Bombing on the German War Economy. Washington, D.C.: Government Printing Office, 1945. - - - . The Effects of Strategic Bombing on Japan's War Economy. Washington, D.C.: Government Printing Office, 1946. Urlanis, B. T. Wars and Population. Moscow: Progress Publishers, 1971. Verba, S. "Sequences and Development. " In Crises and S equences in Political Development, ed. L. Binder and J. LaPalombara. Princeton: Princeton University Press, 1974. Wagner, David; Perkins, Ronald; and Taagepera, Rein. "Complete Solution to Richardson's Arms Race Equation." Journal of Peace Science 1 (Spring 1975): 159-72. Weede, E. "Overwhelming Preponderance as a Pacifying Condition among Contiguous Asian Dyads, 1950--1969." Journal of Conflict Resolution 20 (1976): 395-412. Wildavsky, AaroJl. Budgeting, A Comparative Theory of Budgetary Processes. Boston: Little, Brown, 1975. Willis, Howard, ed. Economic Development for Latin America. New York: St. Martin's Press, 1961. Wohlstetter, Albert. "The Delicate Balance of Terror." Foreign Affairs 37 (January, 1959): 211-56. - - - . "Is There a Strategic Arms Race?" Foreign Policy 15 (Summer 1974): 2-21. - - - . "Rivals, But No 'Race'." Foreign Policy 16 (Fall 1974): 48--81. Wolf, Charles, Jr. "Economic Aid Reconsidered." In The United States and the Developing Economies, ed. Gustav Ranis. Rev. ed. New York: W. W. Norton, 1973. World Bank. World Tables 1976. Baltimore: Johns Hopkins University Press, 1976. Wright, Quincy. A Study of War. Abridged ed. Chicago: University of Chicago Press, 1964.

282

Bibliography

Zechman, Martin. "A Comparison of. the Small Sample Properties of Probit and OLS Estimators with a Limited Dependent Variable." Mimeographed. University of Rochester, April 1974. Zinnes, Dina, and Gillespie, John, eds. Mathematical Models in International Relations. New York: Praeger, 1976. Zinnes, Dina; Gillespie, John; and Rubinson, Michael. "A Reinterpretation of the Richardson Arms-Race Model." In Mathematical Models in International Relations, ed. Dina Zinnes and John Gillespie.

Index

283

Adelman, Irma, 69 Africa, 6 Aggressors, 48 Aid, foreign: as factor in capabilities, 82, 83-86; Marshall Plan, 144, 234-35; measurement of, 215-16,236--37,265 nn.23, 24; Vietnam, 89; to South Korea, 97; to South Vietnam, 69 Alcock, Norman, 247 n.20 Aldrich, John, 248, n.30 Alexandroff, Alan, 245 nn.2, 12 Alliances: and the balance of power, 15,26,27,38,53-57,207; collective security, 24; disaggregation for analysis, 46, 122; and dominant nations, 60; measure of threat perception, 39, 40-41, 249 n.33; and the power transition, 25-27, 38, 54-57, 60; and war, 14, 39, 53-55, 56-59, 60, 62 207 Allison, Graham, 264 n.21 Angell, Norman, 106, 254 n.2, 258 n.28 Antiballistic missile systems, 185 Arabia, Saudi, 7, 9 Arab states, wars with Israel, 7, 65, 67,89,92-93,210 Ardant, Gabriel, 74, 251 n.12 Argentina, 7, 126 Arms control, and stable deterrence, 182-83 Arms races, conventional: destabilizing mechanisms, 222; and development, 147-48; and organizational politics, 183-84; tests of, 184

Arms races, nuclear: 182; actionreaction model, 239; arms expenditures models, 190--91, 236--37; depreciation rate as factor in expenditures, 241-42; destabilizing mechanism, 222, 259 nn.3, 4, 263 nn.18, 19; and deterrence, 181-82, 215; disconfirming presence of, 192-93, 199-200; indicator of allocations, 216; interaction in budget allocations, 191-92, 215-16, 236--37, 265 nn.23, 24; internal factors source of arms buildups, 193-94,200,216-17, 236--37; internal processes causing, 260; and national growth, 221; and organizational politics, 18384; size of stocks related to increases, 193; and stability, 183, 214-15; tests of, 184, 190,216-17, 264 n.22; versus incremental models of expenditures, 184 Aron, Raymond, 246n.16 Art, Robert, 265 n.21 Australia, 125 Austria-Hungary, 43, 49, 113, 175 Axis, the, in World War II, 26, 127 Bahl, Roy, 75, 227, 253 nn.I3, 16 Balestra, Pietro, 270 n.2 Banks, Arthur, 267 n.14 Bass, Hassel, 253 n.13 Baugh, William, 264 n.24 Belligerents, in world wars, 126, 257 n.25 Binder, L., 249 n.I, 251 n.1I Blainey, Geoffrey, 243 n.2 Boretsky, Michael, 266 n. 25

284

Index

Boulding, Kenneth, 254 n.5 Box, George, 255 n.8 Brazil, 7, 20, 124 Bremer, Stuart, 5, 34, 243 n.2 Brito, Dagobert, 262, n.15, 269 n.37 Brodie, Bernard, 192; and the conception of deterrence, 259 n.2, 4, 260 n.3, 267 n.32 Bueno de Mesquita, Bruce, 39. 40, 248 nn.25, 26, 251 n.1O Bulgaria, 124 Burns, Arthur, 255 n.15 Busch, Peter, 263 n.19 Cambodia, 7 Carr, E., 243 n.3 Challengers, 25, 55, 58-60; overtaking dominant nation, 216; perceptions of, 60-61 Chamberlain, Neville, 161 Chammah, Albert N., 263 n.19 Chelliah, Raja, 75, 227, 252 n.l3, 253 n.l3 Chiang Kai-shek, 164 China, Peoples Republic of, 3, 5, 7,22,25,27; aid to North Korea, 97,. 99; alliances of, 55-56; Communist takeover, 99, 164; crosses the Yalu, 165-66; defiance of USSR, 173; and development, 20; great-power status, 43--44; help to Nprth Vietnam, 89; Korean War, 87, 97-99; national capabilities of, 97; political capacity of, 87, 97, 218; risk-prone in nuclear confrontations, 166, 168, 172-73; and Vietnam, 65, 71 Choucri, Nazli, 243 n.2, 244 n.8, 246 n.17, 249 n.34 Churchill, Winston, 2, 219 Cipolla, Carlo, 248 n.24 Colby, William, 266 n.25 Coleman, J., 249 n.1 Collective security, 17-19; aggressors, 27; alliances in, 18; assumptions in, 18; elite role in, 18, 2223; and power distributions, 22, 24, 27-28; tests of, 28-30; and the

power transition, 244 n.7; and war, 22 Communism, 5, 21, 70 Contenders, 43, 44; and major powers, 44, 54-55; and war, 51, 54-56, 58-60 Coplin, William, 246 n.l Crecine, John, 265 n.21 Cuban missile crisis: decisionmaking, 155, 161; illegitimacy of USSR act, 178, 269 n.3; U.S. deterrence, 180; USSR loss in, 170 Cutright, Philipps, 69, 251 n.8 Czechoslovakia, 140, 145; aid to Syria, Yom Kippur War, 175; Czech coup I, 165; Czech coup II, 171 Daladier, E., 161 Damansky/Chenpao Island, 169 Davis, Kingsley, 33 Davis, Otto, 264 n.21 Dempster, M. A. H., 264 n.21 Depression, economic, 112, 119. 126, 135, 139 Deterrence, mutual, 155, 157; action-reaction model, 187-88. 238-40; The Arab-Israeli conflicts, 175-76; comparisons of expenditures for, 186. 188-89, 266 n.26, 267 nn.27, 33, 268 n.35; and conventional strength, 169-74; the Cuban missile crisis, 170-71. 176; and democratic systems. 157; factors in expenditures on strategic weapons 191; and fear of proliferation, 201; lack of evidence of, 194-95, 216; models of allocations of expenditures for nuclear arms, 190-91, 236-42; Sino-Vietnamese-Soviet confrontation, 172-73; strategic weapons expenditures (1951-75), 185, 197-98; the Suez crisis, 174; test by Fourier analysis, 195-96; testing, 169, 176, 180. 181, 184, 186-87. 189-91, 236-42; un-

285

Index

answered questions in theory, 176-77; USSR victory in Berlin Wall conflict, 169-70, 176-77; U.S. acceptance of costs, 158 Deterrence, nuclear: change of behavior hypothesis, 159, 214-15; of China by USSR, 156; conventional strength and, 177; credibility of threat, 181, 214; definition of, 155, 214-15; destruction of the adversary, 154; determination of level of terror, 156-57; failure of, 176, 179,215; faults in theory, 216; importance of legitimacy, 178; nuclear versus nonnuclear nations, 159; possessors' fear of own weapons, 158; propositions in, 150-51; risk acceptance, 15960, 214-15; strategies of, 153-55; terror as cause of, 151, 158,21617; tests of, 150, 152-53, 159-73, 216; U.S. contribution to theory of, 153-55; of USSR by U.S., 155,216 Deterrence, unilateral: Berlin Blockade, 165, 176-77; Chinese Civil War, 164-65; Hungary, 177; Korean War, 165-66, 176-77; Sino-Soviet conflict (1%9), 16869; Vietnam War, 167-68, 176 Deutsch, Karl, 69, 243 n.2, 245 n.12, 251 n.8, 256 n.18 Development, national: and arms races, 221; definition of, 8-10; and economic development, 256 n.23; 1870-1970 rates of growth, 112, 114, 139; factors in, 9-10; and military forces, 147,213-17,221; and national capabilities, 9, 20, 48,54,67,68,147,204; and stratification, 20; sequences in, 32; and war, 11,48,54,57-59,61-63, 147,204. See also Political capability Dominant countries, 21, 25, 28-31, 38-39,58-59; legitimacy of, 179; likelihood of war if overtaken, 216; perceptions of, 60-61

Dulles, John Foster, 260 n.5 Du Mouchel, William H., 249 n.30 Egypt. See United Arab Republic Electoral data, as measure of political development, 251 n.12 Elites: balance-of-power role, 15, 22-23; decision on war, 22, 39; evaluation of nuclear capabilities by, 185; governing capacity, 69; and nuclear threats, 156, 158; power perceptions, 29-30, 40, 53; representativeness, 69; threat perceptions, 14,40 Erbring, Lutz, 255 n.13 Eritrea, 7 Esposito, Vincent J., 253 n.18 Ethiopia, 6, 7, 124 Europe, Western: and development, 70; and European international hegemony, 43-44; as a target of nuclear attack, 155 Ferris, Wayne, 246 n.18 Finkle, Jason, 251 n.8 Finland, 124 Fisher, Gregory, 264 n.21 Fourier analysis, 268 nn.35, 36 Fox, D., 255 n.13 Fox, William, 255 n.25 France: alliances, 42; attack on UAR, 174; decline of, 24, 26; defeat in World War 11,3; dynastic politics, 72; Marshall Plan aid to, 144, 212-13; participation in world wars, 127; threatened by USSR, 184; victory in World War I, 3; after World War II, 142 Franco, Francisco, 3, 161 Formosa, 164 Friedman, Milton, 259 n.34 Frumkin, Gregory, 114,254 nn.3,4, 255 n.ll Fuchs, Wilhelm, 246 nn. 18, 19 Gable, Richard, 251 n.8 Garnham, David, 249 n.32 German, Clifford, 246 n.18

286

Index

Germany, 125; aggression by, 58, 220, 269 n.2; development, 20; dynastic politics in, 72; greatpower status, 43; Marshall Plan aid to, 144, 212-13; miscalculations in World War II, 3, 26; mobilization of, 140; national capabilities of, 218, 220-21; overtaking France, 3, 26; power transition, 20, 58-59; victory in 1890, III; von Bulow, 13; after World War II, 126, 142 Germany, East, 169-70; exodus to West Berlin, 170 Gillespie, J. Y., 251 n.8, 259 n.3, 263 n.19, 265 n.23 GNP, as a measure of national capabilities, 33-35, 37-38,48, 57, 67-68, 85, 86, 89, 92, 94, 97, 99, 101, 108,211,246 n.19 Graham, Daniel, 266 n.25 Great Powers: alliance behavior, 57; definition of, 39; impotence in nuclear war, 56; performance in wars, 2 48, 51, 55-59. 62, 121 Greece, 124 Griffith, William, 260 n.7, 262 n.14 Gurr. Ted, 69, 251 n.8, 254 n.5 Haas, E .. 243 n.5 Haggard, M. T., 253 n.18 Haldeman, H. R., 260 n.7 Halperin, Morton, 265 n.21 Healy, Brian, 245 n.12 Heiss, Klaus P., 247 nn.19, 20, 255 n.7 Herz, John, 257 n.20 Hiroshima. 166 Hitch, Charles, 247 n.20 Hitler, Adolph, 2, 3, 140. 161. 165, 269 n.2 Holsti, J. K .. 246 n.15 Holsti, Ole, 245 n.15 Holt. R.. 250 n.1 Horelick, Arnold, 259 n.4 Hungarian Revolt, 167; failure of deterrence, 167 Huntington, Samuel, 251 nn.7, II, 259 nn.3, 4, 263 n.19

ICBMs: speed of, 152; strategic character, 185; U.S. and USSR deployment of, 189. See a/so Nuclear capabilities; Deterrence, nuclear India, 10, 65; and development. 20; and power transition, 20; war with China, 89, 210 Indonesia, 20 Inkeles, Alex, 69 International Institute of Strategic Studies, 262 n.16 International Monetary Fund, 252 n. 13 Intriligator, Michael, 262 n.15, 268 n.37 Iraq, 73 Israel, 65, 67, 68; political capability of, 77, 82, 218; war with Arab states, 7, 90, 92-93, 129, 175. 209-10; war with UAR, 89,92. 174,209-10 Italy: loss in World War II, 3, 26, 31; Marshall Plan aid, 144,21213; participation in World War II, 127; power status of, 43, 44, 58; after world war period, 142 Japan, 2-3, 7, 26; development in, 20, 78; Marshall Plan aid. 144. 212-13; pow.erof, 20, 43, 44, 21820 Jenkins, Gwilym, 256 n.8 Jordan, 120, 125 Kanter, Arnold, 264 n.21 Kaufman, William, 260 nn. 4, 5 Kegly. C. 248 n.28 Kelly, Margaret, 253 n.13 Kennedy, John F .. 13. 147, 156. 180,269 n.3 Kennedy. Robert, 156-57. 243 n.l, 260 n.8 Keynes, J. M., 106. 253 n.1 Khmer Rouge, 22 Khrushchev, Nikita, 157, 171,260 n.9 Kindleberger, Charles, 257 n.25 Kirk, Dudley, 254 n.2

287

Index

Kiser, Louise, 254 n.2 Kissinger, Henry, 260 n.4 Klein, Lawrence, 254 n.6 Knorr, Klaus, 246 n.18, 247 n.19, 253 n.17, 269 n.l Korea, North, 22, 65; aid to, 99100; attack of South, 166 Korea, South: aid to, 99-100; invasion of, 166 Korean War, 161; aid by China, USSR, and U.S., 97, 99-100, 253 n.18; failure of deterrence, 165, 166-67; initiation of, 166; stalemates in, 3; test of capabilities, 89-90, 209 Kugler, Jacek, 246 n.18, 248 nn.23, 24,255 nn. 10, 14,257 n.17 Kuznets, Simon, 139, 255 n.9, 258 n.29 Laird, Melvin, 265 n.24 Laissez-faire, power model, 23 Lamborn, Alan, 251 n.1O Landwehr, J., 255 n.13 Lebanon, 78 Lee, W. T., 266 n.25 Liska, George, 243 n.5 Little, I. M. D., 258 n.32 Losada, Marcial, 268 n.35 Lotz, Jorgen, 251 n.13 Luce, Robert, 263 n.19 Lucier, Charles, 243 n.3 McKean, Roland, 247 n.20 McKelvey, Richard, 249 n.30 McNamara, Robert S., 261 n.1O Maddison, Angus, 114, 248 n.24, 255 n.12, 256 nn.16, 18 Mao Tse-tung, 10, 158, 164 Marshall, Andrew, 266 n.25 Marshall Plan: as test of Phoenix factor, 142, 212-13; lack of success, 144 Mass publics, representation of, 69 Mesvold, B. A., 251 n.8 Mexico, 78 Middle East, 90, 120 Midlarsky, Manus, 243 n.2

Military forces, and development, 7, 147-48 Minuteman, 189 Mobilization. See Political capacity Mongolia, 124 Morgenstern, Oskar, 247 n.19 Morgenthau, Hans, 243 n.l, 3, 245 n.12 Morris, Fredric, 264 n.21 Morss, Elliott, 252 n.13 Mussolini, Benito, 31 Nagasaki, 166 Nash, Keir, 251 n.1O Nasser, GamaI Abdul, 174 National capabilities: components of, 8, 22, 33, 34, 107-8; estimates of, 9; and foreign aid, 86; oflosers in war, 144-45; measurement of, 22, 29, 30-38, 64-70, 78, 81-82, 83,84-86,101-2,105,107-8,110, 113-14, 118-19,205-6,208-9; methods of exercising, 5-7; and military might, 9, 148-49; and national development, 5 8, 9, 1923, 61, 107-9, 148, 204-5; new measure of, 85-86, 94; and nuclear weapons, 148-49, 199-201; perceptions of, 13, 30-31, 38-39; and political capacity, 22, 33, 87; and population, 8, 22, 32, 34, 35, 85, 107, 148; and power, 5, 7-9, 30-31,33,35; transformation of national capabilities into time units, 47-48; and war, 7,11,12, 22,24-25,28,32,48,50-53,144, 204; war as a test of, 65, 88-100, 102-3; of winners in war, 144-45 Nation-states, as actors in wars, 123-26 Nef, John U., 254 n.2 Nerlove, Marc, 270 n.3 Newcombe, Alan, 247 n.20 Newhouse, John, 261 n.1O New Zealand, 73, 124, 125 Nonbelligerents, in world wars, 126-27 North, Robert, 243 n.2, 244 n.8, 249 n.34

288

Index Norway, 125 Notestein, Frank, 254 nn. 2, 3, 258 n.28 Nove, Alec, 266 n.25 Nuclear capabilities, 245 n.14; change in distributions of national capabilities, 9, 200, 214; comparison of, 26, 236-40, 26667; contenders, 56; definition of strategic, 185; depreciation of, 189, 241-42, 267 n.29; destructiveness of, 9,151,214,260 n.IO; and deterrence model, 150, 151, 260 n.16; development of, 151-52; expenditures (1951-75), 197-99; external and internal pressures on production, 150, 184-85, 193; fission bombs, 151; and high risks in conflicts, 159; indicator of offensive strategic, 185; and the internationalorder, 148-49, 21314; the issue of self-restraint, 158, 178; limited use of, 154-55; and national capabilities, 9-11, 148, 213-14; and nuclear races, 149; propositions in theory of nuclear races, 149-50; states more riskprone in spite of, 196-97; thermonuclear bombs, 152; total destruction of adversaries, 154. See also Deterrence, nuclear Organizational politics, and arms races, 183-84 Organski, A. F. K., 33, 66,106,243 nn. 1-4, 244 nn. 7, 9, 245 n.12, 246 nn.I6-18, 251 n.l0, 262 n.17 Ostrom, Charles Jr., 265 n.22 Overholt, William H., 260 n.7 Payne, Fred, 266 n.24 Pearl Harbor, 2, 7 Pechman, Joseph, 264 n.20 Perkins, Ronald, 265 n.22 Philippines, 73 Phoenix factor, 107, 132-33, 135, 145-46, 210; and economic development, 142; lack of explanation of, 142-43, 210-11 ; test of aid

and, 142-44, 212-13 Poland, 124 Politburo, Soviet, and nuclear attack on China, 156 Political capacity: in China, 87; components of 82-83; definition, 71-74; democracy and stability, 69,72,217-20; economic development and, 32, 70, 72, 169; elite role in, 78-81; estimation of, 44, 64-65,68-69,77-81,209; factor in national capabilities, 9, 22, 64-65,86-87,88,205; forms of political systems and, 217-20; governmental penetration and extraction, 73, 251 n.12; measurement of, ll, 31, 32, 33, 64-69,71,74-80,80-82,85, 96-97, 101, 107,208-9,227-33; and nuclear strength, 148, 250 n.2, 251 n.12, 253 nn.15, 17; and taxation, 74, 227-33; war as a test of, 31, 32, 88, 94, 96, 100, 147, 205-6, 209 Pollack, Jonathan, 260 n.7 Population, factor in power, 8, 20, 32, 34, 35, 85, 147 Power, balance of, 14-17; the aggressorin, 16,27, 28,59-61; and alliances, 15, 16, 26, 27, 58-61; assumptions, 16; balancer concept in, 16-17; elite decision, 22-23; and equal distribution of power, 16,58-61; errors in model, 17,60-61; hypotheses, 48; major mechanisms of, 16,25,39; motives leading behavior, 15; perceptions of, 39; power distributions, 14, 16, 24-25, 27-28, 48,55-61; and the power transition, 48; test of, 28-30, 60-61 Power, national: definition of, 5, 107,245 n.14; methods of exercising, 5-7. See also National capabilities Powers, Great. See Great Powers Power transition, 19,22; aggression and, 19, 27; alliances and, 25, 26-27, 53-54, 60; balance of

289

Index

power and, 48,58; challenger, 19, 21, 25, 55; collective security and, 20; demonstrated correct, 61; development and, 21-22; dominant nations, 19, 21, 25, 55; elites' roles in, 22-23; great powers'behavior,56;hypotheses, 54; and lateral pressure, 244 n.8; mechanics of, 19-20, 27; population and, 20-21; power distributions in, 20, 23, 24-25, 27-28, 47, 48, 56-58,61,216-17; stages in, 244 n.9, 245 n.lO; satisfied and dissatisfied countries, 19, 23, 39, 59-60; test of, 28-30, 54-55; war and, 22, 23, 59 Raiffa, Howard, 263 n.19 Ranis, Gustav, 259 n.34 Rapoport, Anatol, 263 n.19 Ray, James, 247 n.21 Richardson, Lewis, 186,255 n.16, 259 n.3 Riker, William, 245 n.11 Robinson, Michael, 263 n.19 Rokkan, Stein, 69, 251 n.8 Roosevelt, Franklin, 2 Rosecrance, Richard, 245 n.11 Rosen, Steven, 247 n.20 Rosenstein-Rodan, Paul, 258 n.32 Rumania, 124 Rumsfeld, Donald, 261 n.l0, 265 n.24 Rush, Myron, 259 n.4 Russett, Bruce, 243 n.2, 247 n.20, 256 nn.21, 22, 263 n.19 Russia. See USSR Sanders, Thomas, 255 n.13 Schelling, Thomas, 260 n.5 Singer, J. David, 34, 35, 67, 124, 243 n.2, 245 n.12, 246 n.17, 247 n.21, 248 nn.27, 28, 255 n.16, 256 n.21 SLBMs, strategic character, 185 Small, M., 124,248 n.28, 255 n.16, 256 n.21 Snyder, Glenn, 260 n.5

Sorensen, Theodore, 259 n.l, 260 n.5, 269 n.3 South Africa, 124, 125 Soviet Union. See USSR Spain, 113, 125-26 Squires, Michael, 266 n.23 Stalin, J., 2, 161 Stein, Arthur, 245 n.12 Steinbruner, John, 264 n.19 Stuckey, J., 34 Sweden, 125 Syria, war with Israel, 78, 93, 120 Systems: central and peripheral, 44, 48, 51, 53, 54; international, 43; prediction in behavior in, 55 Taagepera, Rein, 265 n.22 Tammen, Ron, 265 n.22 Tanter, R., 265 n.21 Tauber, Irene, 254 n.2 Taxation: controls in model, 76-77; measure of power, 73, 74; taxable capacity, 76, 78, 80, 227-29, 231-33; tax effort, 74, tax effort in developed countries, 82; tax effort in developing countries, 75-80. See also Political capacity Terror, nuclear: and arms races or competition, 182, 216; the balance of, 182, 216; and nuclear capability, 181; and security, 179; and stable deterrence, 182; under arms control, 183 Tilly, Charles, 142,251 n.12, 258 n.30 Tojo, 161 Turkey, 124 Turner, J., 250 n.1 Ullman, R., 264 n.21 United Arab Republic, 6, 120; power of, 68, 89, 92; wars with Israel, 89, 92-94 United Kingdom: alliance of, 42; belligerence in World War I, 127; decline of, 24, 26; France and, 42, 44; great-power status, 43; Marshall Plan aid to, 144, 212-13, 236-37; national capabilities, 2,

290

Index

24-25, 58, 218-21; after world war period, 142 United Nations, 89, 250 n.5, 254 n.3; Research Institute of, 251 n.9 United States: advantage in nuclear capabilities, 186; aid to Israel in Arab-Israeli wars, 175; aid to South Korea, 89, 97; aid to South Vietnam, 94, 96; and Arab-Israeli conflict, 120; Central Intelligence Agency, estimate of Soviet strategic expenditures, 265 n.23, 268 n.37; competition with USSR, 201, 216-17; data weaknesses, 126; Department of Defense, estimate of Soviet strategic expenditures, 265 n.23; difference from Russian system, 219; dominant nation, 25, 44-45; expenditures on strategic weapons (1951-75), 197-99, 216-17; great-power status, 43, 44; Joint Economic Committee, 267 n.30; Korean War, 65, 88; loss in Czech coup, 165; Marshall Plan aid by, 144, 236-37; military-industrial complex weaker than USSR's, 194-95; miscalculation in Vietnam, 3; national capabilities of, 5, 6, 24-25, 27,96, 100,218-20; nuclear hegemony of, 164, 166; nuclear restraint in Vietnam, 168; opinion in, to escalate Korean War, 166; opposition to Communist China, 164; promise not to invade Cuba, 170; provocation in Korea, 166; recognized sphere of influence, 172; request for help from Hungary, 161; response to USSR as test of nuclear race, 184; restraint or deterrence at Berlin Wall, 170; strategic bombing survey, 254 n.1, 269 n.1; strategic weapons of, 189, 240; success in ArabIsraeli wars, 176; suspicions of USSR nuclear behavior, 178; and theory of deterrence, 178; threatens France and UK, 174;

victory in Berlin blockade, 165; victory in Cuban missile crisis, 170, 180-81 Urlanis, B. T., 254 n.2, 258 n.28 USSR, 6, 2, 72; aid to Arabs in wars with Israel, 93-94, 174; aid to North Korea, 89; aid to UAR, 94; Arab-Israeli conflict, 120; border conflicts with Chinese, 168-69; as challenger, 25; and China, 7, 27; competition with U.S., 25, 27, 201,216-17; data problems with, 113, 126; defied by Chinese. 173; difference from U.S. system, 219; expenditures on strategic weapons, 197-99,265 n.23, 268 n.37; failure in Arab-Israeli war (1973), 176; greater effort in building nuclear arms, 198; great-power status, 43; inferiority in Cuba, 171; invasion of Czechoslovakia, 171-72; invasion of Hungary, 167; "legitimacy" of international aggressions, 178; loss in the Berlin blockade, 165; loss in Cuban missile crisis, 170-71, 180; method of estimating nuclear costs, 186-87; military-industrial complex, 194; national capabilities of, 5, 24. 25, 219; nuclear restraint with Chinese, 168-69, 176; offsetting U.S. nuclear advantages, 186; participation in World War I, 113; participation in World War II. 2, 126-27; protector of China in Korea, 166; reason for Czech invasion, 172; recognized spheres of influence, 172; removal of Nikita Khrushchev, 171; response to U.S. as test of nuclear race, 184; strategic weapons of, 189; success in Arab-Israeli war (1973), 176; superiority in Berlin, 170; threatened by U.S., 174; threatens France and UK, 174; victory in Czech coup, 165; and Vietnam, 65; war with Japan, 126

291

Index

Verba, S., 250 n.1 Versailles, 3 Viet Cong, 71, 88, 89 Vietnam, defiance of Chinese nuclear power, 173; war with China, 172 Vietnam, North, 9, 22, 65, 71; foreign aid to, 89; governmental performance, 74; national capabilities of, 94, %, 209-10; political capacity of, 88, 218; risks provinces in war, 167; USSR protection of, 168; war with South Vietnam, 89 Vietnam, South, 65, 94; foreign aid to, 89; national capabilities of, 94, 96,209-10 Vietnam War, 94, 161; as test of capabilities, 209-10; as test case of deterrence, 167-68 Wagner, David, 265 n.22 Waltz, Kenneth, 260 n.5 War: aggressors in, 47-48, 51; and alliances, 14, 25-27, 40, 54-57; and the balance of power, 14-15, 22-23, 25-27; beginning points, 110; belligerents in samples, 124; causes, 1-4,8, 11, 13:-15, 19-23, 27-28,48,50-51,104,204-5; and collective security, 17-19; consequences, 8, 11; consequences for entire sample, both world wars, 132-33, 144-45, 212; consequences for losers, both world wars, 132-33, 144-45, 212; consequences for nonbelligerents, both wars, 132-33, 144, 212; consequences for winners, both wars, 132-33, 144, 212; consequences for entire sample, World War I, 135-37, 144,212; consequences for great powers, both wars, 134-35, 144, 212; contenders, 58-60; criteria for choice of sample, 45-46,120-22,211-13; definition of belligerency, 127-28, 255 n.16; and development, 147, 204; disaggregation of analytic

groups, 128-29; distribution of national capabilities, 13, 28-29, 47,48,50,51,54-55, 137; effects on power distributions, 104-5, 122, 132-33, 134-39, 144-45,204; extrapolation procedure, 108-9, 110-13, 255 n.8; fundamental questions in regard to, 1-2, 3; governmental performance in, 14; hypotheses on consequences, 97, 106-7, 131, 135, 136, 140, 210-13; limited wars, 121; losers' behavior, 106-7, 132, 212-13; major power and, 51, 55-59; outcomes, 1-2,4,8, 11,64-65; and perceptions ofthreat, 40; political engineering in regard to, 220-22; and population, 254 nn.2, 3; and the power transition, 19-23, 27-28; prediction of outcomes, 101,204-5,208-9; prediction of war, 64-65; predictors of consequences, 113, 205, 211-13; real loser in, 140, 145; research to be done, 145-46; sample of, 204-5, 211-13; sample of nations in analysis of consequences, 129-30; as tests of capabilities, 29,31-32,44-47,65,81,88-100, 119-23, 204-5, 208-13; threats, 13; winners' behavior, 106-7, 132.211-13; World Wars I and II, 2, 5, 26, 54-59. See also World War I; World War II Warsaw Pact, 166 Weede, E., 249 n.32 Wildavsky, Aaron, 264 n.21 Wohlstetter, Albert, 260 n.5, 264 n.22, 265 n.23 Wolf, Charles Jr., 259 n.34 World War!, Ill, 112, 113, 118-19; consequences for losers and winners, 135-37; participants in, 113; rates of growth before, 112 World War II, 110, Ill, 112, 113, 118, 269 n.2; consequences for entire sample, 137-42, 144-45, 211-13; consequences for losers, 137-42, 144-45; consequences for

292

Index

neutrals, 139, 144; consequences for winners, 137-42, 144-45; nuclear weapons, 152; projections of consequences, 118-19,211, 213; rates of growth after, 112; underestimation of growth after, 139, 141-42; USSR belligerence

with Japan, 126 Wright, Quincy, 243 n.2, 255 n.16 Zavoina, William, 249 n.30 Zechman, Martin, 249 n.30 Zinnes, Dina, 259 n.3, 263 n.19, 266 nn.23,24