Conceptual Models and the Cuban Missile Crisis...Conceptual Models and the Cuban Missile Crisis...

Conceptual Models and the Cuban Missile Crisis

Graham T. Allison

The American Political Science Review, Volume 63, Issue 3 (Sep., 1969), 689-718.

Your use of the JSTOR database indicates your acceptance of JSTOR’s Terms and Conditions of Use. A copy ofJSTOR’s Terms and Conditions of Use is available at http://www.jstor.org/about/terms.html, by contacting JSTORat [email protected], or by calling JSTOR at (888)388-3574, (734)998-9101 or (FAX) (734)998-9113. No partof a JSTOR transmission may be copied, downloaded, stored, further transmitted, transferred, distributed, altered, orotherwise used, in any form or by any means, except: (1) one stored electronic and one paper copy of any articlesolely for your personal, non-commercial use, or (2) with prior written permission of JSTOR and the publisher ofthe article or other text.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen orprinted page of such transmission.

The American Political Science Review is published by American Political Science Association. Please contactthe publisher for further permissions regarding the use of this work. Publisher contact information may be obtainedat http://www.jstor.org/joumals/apsa.html.

The American Political Science Review01969 American Political Science Association

JSTOR and the JSTOR logo are trademarks of JSTOR, and are Registered in the U.S. Patent and Trademark Office.For more information on JSTOR contact [email protected].

02001 JSTOR

http://www.jstor.org/Mon Sep 3 13:14:45 2001

The AmericanPolitical Science Review

VOL. LXIII SEPTEMBER, 1969 NO. 3

CONCEPTUAL MODELS AND ‘THE CUBAN MISSILE CRISIS”

G R A H A M T . ALLISONHarvard University

The Cuban missile crisis is a seminal event.For thirteen davs of October 1962, there was ahigher probability that more human lives wouldend suddenly than ever before in history. Hadthe worst occurred, the death of 100 millionAmericans, over 100 million Russians, and mil-lions of Europeans as well would make previousnatural calamities and inhumanities appear in-significant. Given the probability of disaster-which President Kennedy estimated as “be-tween 1 out of 3 and even”-our escape seemsawesome? This event symbolizes a central, ifonly partially thinkable, fact about our exis-tence. That such consequences could follow fromthe choices and actions of national governmentsobliges students of government as well as partici-pants in governance to think hard about theseproblems.

Improved understanding of this crisis de-pends in part on more information and moreprobing analyses of available evidence. To con-tribute to these efforts is part of the purpose ofthis study. But here the missile crisis serves pri-marily as grist for a more general investigation.

* A longer version of this paper was presented atthe Annual Meeting of the American Political Sci-ence Association, September, 1968 (reproduced bythe Rand Corporation, P-3919). The paper is partof a larger study, scheduled for publication in 1969under the title Bureaucracy and Policy: Concep-tual Modek and the Cuban Missile &isis. For sup-port in various stages of this work I am indebtedto the Institute of Politics in the John F. KennedySchool of Government and the Center for Inter-national Affairs, both at Harvard University, theRand Corporation, and the Council on ForeignRelations. For critical stimulation and advice I amespecially grateful to Richard E. Neustadt, ThomasC. &helling, Andrew W. Marshall, and ElisabethK. Allison.

’ Theodore Sorensen, h’ennedy (New York, 19651,p. 705.

This studv proceeds from the premise thatmarked improvement in our understanding ofsuch events depends critically on more self-con-sciousness about what observers bring to theanalysis. What each analyst sees and judges tobe important is a function not only of the evi-dence about what happened but’ also of the“conceptual lenses” through which he looks atthe evidence. The principarl purpose of this essayis to explore some of the fundamental assump-tions and categories employed by analysts inthinking about problems of governmental be-havior, especially in foreign and military affairs.

The general argument can be summarized inthree propositions :

1. Analysts think about problems of foreign andmilitary policy in terms of largely implicit concep-tual models that have significant consequences forthe content of their thought?

Though the present. product of foreign policyanalysis is neither systematic nor powerful, ifone carefully examines explanations produced byanalysts, a number of fundamental similaritiesemerge. Explanations produced by particularanalysts display quite regular, predictable fea-tures. This predictability suggests a substruc-ture. These regularities reflect an analyst’s as-sumptions about the cha.racter of puzzles, thecategories in which problems should be consid-ered, the types of evidence that are relevant,and the determinants of occurrences. The firstproposition is that clusters of such related as-sumptions constitute basic frames of referenceor conceptual models in terms of which analysts

*In attempting to understand problems of for-eign affairs, analysts engage in a number of related,but logically separable enterprises : (a) description,(b) explanation, (c) prediction, (d) evaluation, and(e) recommendation. This essa.y focuses primarilyon explanat’ion (and by implication, prediction).

689

690 THE AMERICAN POLITICAL SCIENCE REVIEW V O L. 63

both ask and answer the questions: What hap-pened? Why did the event happen? What willhappen?” Such assumptions are central to theactivities of explanation and prediction, for inattempting to explain a particular event, the an-alyst cannot simply describe the full state ofthe world leading up to that event. The logic ofexplanation requires that he single out the rele-vant, important determinants of the occur-rence.4 Moreover, as the logic of predictionunderscores, the analyst must summarize thevarious determinants as they bear on the eventin question. Conceptual models both fix themesh of the nets that the analyst drags throughtlhe material in order to explain a particular ac-tion or decision and direct him to cast his net inselect ponds, at certain depths, in order to catchthe fish he is after.

2. Most analysts explain (and predict) the be-havior of national governments in terms of variousforms of one basic conceptual model, here entitledthe Rational Policy Model (Model I) 7

In terms of this conceptual model, analystsatt,empt to understand happenings as the moreor less purposive acts of unified national govern-ments. For these analysts, the point of an expla-nation is to show how the nation or government

‘In arguing that explanations proceed in termsof implicit conceptual models, this essay makes noclaim that foreign policy analysts have developedany satisfactory, empirically tested theory. In thisessay, the use of the term “model” without quali-fiers should be read “conceptual scheme.”

4For the purpose of this argument we shall ac-cept Carl G. Hempel’s characterization of the logicof explanation: an explanation “answers the ques-tion, 9%~ did the explanadum-phenomenon oc-cur?’ by showing that the phenomenon resultedfrom particular circumstances, specified in C1, CL,

Ck, in accordance with laws L,, 212, . . . Lr. Bypbmting this out, the argument shows that, giventhe particular circumstances and the laws in ques-tion, the occurrence of the phenomenon was to beeqected; and it is in this sense that the explana-tion enables us to understand why the phenome-non occurred.” Aspects of Scientific Explanation(New York, 1965), p. 337. While various patternsof explanation can be distinguished, uiz., ErnesthTagel, The Structure of Science: Problems in thetigic of Scientific Explanation, New York, 1961),satisfactory scientific explanations exhibit this basiclogic. Consequently prediction is the converse ofexplanation.

’ Earlier drafts of this argument have arousedheated arguments concerning proper names forthese models. To choose names from ordinary lan-guage is to court confusion, as well as familiarity.Perhaps it is best to think of these models as I,II, and III.

could have chosen the action in question, giventhe strategic problem that it faced. For example,in confronting the problem posed by the Sovietinstallation of missiles in Cuba, rational policymodel analysts attempt to show how this was areasonable act from the point of view of the So-viet Union, given Soviet strategic objectives.

3. Two “alternative” conceptual models, here la-beled an Organizational Process Model (Model II)and a Bureaucratic Politics Model (Model III>provide a base for improved explanation and pre-diction.

Although the standard frame of reference hasproved useful for many purposes, there is pow-erful evidence that it must be supplemented, ifnot supplanted, by frames of reference whichfocus upon the large organizations and politicalactors involved in the policy process. Model I’simplication that important events have impor-tant causes, i.e., that monoliths perform largeactions for big reasons, must be balanced by anappreciation of the facts (a) that monoliths areblack boxes covering various gears and levers ina highly differentiated decision-making struc-ture, and (b) that large acts are the conse-quences of innumerable and often conflictingsmaller actions by individuals at various levelsof bureaucratic organizations in the service of avariety of only partially compatible conceptionsof national goals, organizational goals, and polit-ical objectives. Recent developments in the fieldof organization theory provide the foundationfor the second model. According to this organi-zational process model, what Model I catego-rizes as “acts” and “choices” are instead outputsof large organizations functioning according tocertain regular patterns of behavior. Faced withthe problem of Soviet missiles in Cuba, aModel II analyst identifies the relevant organi-zations and displays the patterns of organiza-tional behavior from which this action emerged.The third model focuses on the internal politicsof a government. Happenings in foreign affairsare understood, according to the bureaucraticpolitics model, neither as choices nor as outputs.Instead, what happens is categorized as o u t -comes of various overlapping bargaining gamesamong players arranged hierarchically in the na-tional government. In confronting the problemposed by Soviet missiles in Cuba, a Model IIIanalyst displays the perceptions, motivations,positions, power, and maneuvers of principalplayers from which the outcome emerged.6

61n strict terms, the “outcomes” which thesethree models attempt to explain are essentially ac-tions of national governments, i.e., the sum of ac-tivities of all individuals employed by a govern-ment relevant to an issue. These models focus noton a state of a.ffairs, i.e., a full description of the

1969 CONCEPTUAL MODELS AND THE CUBAN MISSILE CRISIS 691

A central metaphor illuminates differencesamong these models. Foreign policy has oftenbeen compared to moves, sequences of moves,and games of chess. If one were limited to ob-servations on a screen upon which moves in thechess game were projected without informationas to how the pieces came to be moved, he wouldassume---as Model I does-that an individualchess player was moving the pieces with refer-ence to plans and maneuvers toward the goal ofwinning the game, But a pattern of moves canbe imagined that would lead the serious ob-server, after watching several games, to considerthe hypothesis that the chess player was not asingle individual but rather a loose alliance ofsemi-independent organizations, each of whichmoved its set of pieces according to standardoperating procedures. For example, movement ofseparate sets of pieces might proceed in turn,each according to a routine, the king’s rook,bishop, and their pawns repeatedly attacking theopponent according to a fixed plan. Further-more, it is conceivable that the pattern of playwould suggest to an observer that a number ofdistinct players, with distinct objectives butshared power over the pieces, were determiningthe moves as the resultant of collegial bargain-ing. For example, the black rook’s move mightcontribute to the loss of a black knight with nocomparable gain for the black team, but withthe black rook becoming the principal guardianof the “palace” on that side of the board.

The space available does not permit full de-world, but upon national decision and implementa-tion. This distinction is stated clearly by Haroldand Margaret Sprout, “Environmental Factors onthe Study of International Politics,” in James Ros-enau (ed.), International Politics and ForeignPolicy (Glencoe, Illinois, 1961)) p. 116. This re-striction excludes explanations offered principallyin terms of international systems theories. Never-theless, this restriction is not severe, since few in-teresting explanations of occurrences in foreignpolicy have been produced at that level of anal-ysis. According to David Singer, “The nation state-our primary actor in international relations . . .is clearly the traditional focus among Western stu-dents and is the one which dominates all of thetexts employed in English-speaking colleges anduniversities.” David Singer, “The Level-of-AnalysisProblem in International Relations,” Klaus Knorrand Sidney Verba (eds.), The International Sys-tem (Princeton, 1961). Similarly, Richard Brody’sreview of contemporary trends in the study of in-ternational relations finds that “scholars have comeincreasingly to focus on acts of nations. That is,they all focus on the behavior of nations in some re-spect. Having an interest in accounting for thebehavior of nations in common, the prospects fora common frame of reference are enhanced.”

velopment and support of such a generalargument.7 Rather, the sections that follow sim-ply sketch each conceptual model, articulate itas an analytic paradigm, and apply it to producean explanation. But each model is applied to thesame event: the U.S. blockade of Cuba duringthe missile crisis. These “alternative explana-tions” of the same happening illustrate differ-ences among the models--at work? A crisis de-cision, by a small group of men in the context ofultimate threat, this is a case of the rational pol-icy model par excellence. The dimensions andfactors that Models II and III uncover in thiscase are therefore particularly suggestive, Theconcluding section of this paper suggests howthe three models may be related and how theycan be extended to generate predictions.

MODEL I: RATIONAL POLICY

RATIONAL POLICY MODEL ILLUSTRATED

Where is the pinch of the puzzle raised by theNew York Times over Soviet deployment of anantiballistic missile system?” The question, as theTimes states it, concerns the Soviet Union’s ob-jective in allocating such large sums of moneyfor this weapon system while at the same timeseeming to pursue a policy of increasing detente.In former President Johnson’s words, “the para-dox is that this [Soviet deployment of an anti-ballistic missile system] should be happening at atime when there is abundant evidence that ourmutual antagonism is beginning to ease?) Thisquestion troubles people primarily because Sovietantiballistic missile deployment, and evidence ofSoviet actions towards detente, when juxtaposedin our implicit model, produce a question. Withreference to what objective could the Soviet gov-ernment have rationally chosen the simultaneouspursuit of these two courses of actions? Thisquestion arises only when the analyst attempts tostructure events as purposive choices of consis-tent actors.

‘For further development and support of thesearguments see the author’s larger study, Bureauc-racy and Policy: Conceptual Models and the Cu-ban Missile O&is (forthcoming). In its abbrevi-ated form, the argument must, at some points, ap-pear overly stark. The limits of space have forcedthe omission of many reservations and refinements.

’ Each of the three “case snapshots” displays thework of a conceptual model as it is applied to ex-plain the U.S. blockade of Cuba. But these threecuts are primarily exercises in hypothesis genera-tion rather than hypothesis testing. Especiallywhen separated from the larger study, these ac-counts may be misleading. The sources for theseaccounts include the full public record plus a largenumber of interviews with participants in the crisis.

’ NeTu YorzC Times, February 18, 1967.lo Ibid.

692 THE AMEHICAX POLITICAL SCIENCE REVIEW V O L. 63

How do analysts attempt to explain the So-viet emplacement of missiles in Cuba? The mostwidely cited explanation of this occurrence hasbeen produced by two RAND Sovietologists,Arnold Horelick and Myron Rush? They con-clude that “the introduction of strategic missilesinto Cuba was motivated chiefly by the Sovietleaders’ desire to overcome . , . the existing largemargin of U.S. strategic superiority?2 How dothey reach this conclusion? In Sherlock Holmesstyle, they seize several salient characteristics ofthis action and use these features as criteriaagainst which to test alternative hypothesesabout Soviet objectives. For example, the size ofthe Soviet deployment, and the simultaneousemplacement of more expensive, more visible in-termediate range missiles as well as mediumrange missiles, it is argued, exclude an explana-tion of the action in terms of Cuban defense-since that objective could have been securedwith a much smaller number of medium rangemissiles alone. Their explanation presents an ar-gument for one objective that permits interpre-tation of the details of Soviet behavior as avalue-maximizing choice.

How do analysts account for the coming of theFirst World War? According to Hans Morgen-thau, “the first World War had its originexclusively in the fear of a disturbance ofthe European balance of power? In the pe-riod preceding World War I, the Triple Allianceprecariously balanced the Triple Entente. If ei-ther power combination could gain a decisiveadvantage in the Balkans, it would achieve a de-cisive advantage in the balance of power. “Itwas this fear,” Morgenthau asserts, “that moti-vated Austria in July 1914 to settle its accountswith Serbia once and for all, and that inducedGermany to support Austria unconditionally. Itwas t,he same fear that brought Russia to thesupport of Serbia, and France to the support ofRussia .“l* How is Morgenthau able to resolvethis problem so confidentlv? By imposing on thedata a “rational outline.“l” The value of thismethod, according to Morgenthau, is that “itprovides for rational discipline in action andcreates astounding continuity in foreign policywhich makes American, British, or Russian for-

llArnold Horelick and Myron Rush, StrategicPower and Soviet Foreign Policy (Chicago, 1965).Based on A. Horelick, “The Cuban Missile Crisis:An Analysis of Soviet Calculations and Behavior,”World Politics (April, 1964).

la Horelick and Rush, Strategic Power and SovietForeign Policy, p. 154.

ls Hans Morgenthau, Politics Am,ong Nations(3rd ed.; New York, 1960), p. 191.

l4 Ibid., p. 192.~5 Ibid., p. 5.

eign policy appear as an intelligent, rational con-tinuum . . . regardless of the different motives,preferences, and intellectual and moral qualitiesof successive statesmen?

Stanley Hoffmann’s essay, “Restraints andChoices in American Foreign Policy” concen-trates, characteristically, on “deep forces”: theinternational system, ideology, and nationalcharacter-which constitute restraints, limits,and blinders .I7 Only secondarily does he con-sider decisions. But when explaining particularoccurrences, though emphasizing relevant con-straints, he focuses on the choices of nations.American behavior in Southeast Asia is ex-plained as a reasonable choice of “downgradingthis particular alliance (SEATO) in favor ofdirect U.S. involvement,” given the constraint:“one is bound by one’s commitments; one iscommitted by one’s mistakes.“lg More fre-quently, Hoffmann uncovers confusion or contra-diction in the nation’s choice. For example, U.S.policy towards underdeveloped countries is ex-plained as “schizophrenic.“lg The method em-ployed by Hoffmann in producing these explana-tions as rational (or irrational) decisions, heterms “imaginative reconstruction.“20

Deterrence is the cardinal problem of thecontemporary strategic literature. ThomasSchelling’s Strategy of ConfEict formulates anumber of propositions focused upon the dy-namics of deterrence in the nuclear age. Oneof the major propositions concerns the stabilityof the balance of terror: in a situation of mu-tual deterrence, the probability of nuclear war isreduced not by the “balance” (the sheer equal-ity of the situation) but rather by the stabilityof the balance, i.e., the fact that neither oppo-nent in striking first can destroy the other’sability to strike back .2l How does Schelling sup-port this proposition? Confidence in the conten-tion stems not from an inductive canvass of alarge number of previous cases, but rather fromtwo calculations. In a situation of “balance” butvulnerability, there are values for which a ra-tional opponent could choose to strike first, e.g.,to destroy enemy capabilities to retaliate. In a

l6 Ibid., pp. 5-6.” Stanley Hoffmann, DaedaZus (Fall, 1962) ; re-

printed in The State of War (New York, 1965).l8 Ibid., p. 171.lg Ibid., p. 189.2o Following Robert MacIver; see Stanley Hoff- ’

mann, Contemporary Theory in International Re-lations (Englewood Cliffs, 1960), pp. 178-179.

21 Thomas Schelling, The Strategy of Conflict,(New York, 1960), p. 232. This proposition wasformulated earlier by A. Wohlstetter, “The Deli-cate Balance of Terror,” Foreign Aflairs (January,1959).

1969 CONCEPTUAL MODELS AND

“stable balance” where no matter who strikesfirst, each has an assured capability to retaliatewith unacceptable damage, no rational agentcould choose such a course of action (since thatchoice is effectively equivalent to choosing mu-tual homicide). Whereas most contemporarystrategic thinkin,0 is driven implicitly by themotor upon which this calculation depends,Schelling explicitly recognizes that strategictheory does assume a model. The foundation ofa theory of strategy is, he asserts: “the assump-tion of rational behavior-not just of intelligentbehavior, but of behavior motivated by con-scious calculation of advantages, calculation thaltin turn is based on an explicit and internallyconsistent value system.“22

What is striking about these examples fromthe literature of foreign policy and internationalrelations are the similarities among analysts ofvarious styles when they are called upon to pro-duce explanations. Each assumes that whatmust be explained is an action, i.e., the realiza-tion of some purpose or intention. Each assumesthat the actor is the national government. Eachassumes that the action is chosen as a caIcuIatedresponse to a strategic problem. For each, expla-nation consists of showing what goal the govern-ment was pursuing in committing the act andhow this action was a reasonable choice, giventhe nation’s objectives. This set of assumptionscharacterizes the rational policy model. The as-sertion that Model I is the standard frame ofreference implies no denial of highly visibledifferences among the interests of Sovietologists,diplomatic historians, international relationstheorists, and strategists. Indeed, in most re-spects, differences among the work of HansMorgenthau, Stanley Hoffmann, and ThomasSchelling could not be more pointed. Apprecia-tion of the extent to which each relies predomi-nantly on Model I, however, reveals basicsimilarities among Morgenthau’s method of“rational reenactment,” Hoffmann’s “imaginativereconstruction,” and &helling’s “vicarious prob-lem solving ; ” family resemblances among Mor-genthau’s “rational statesman,” Hoffmann’s “rou-let te player ,” and &helling’s “game theorist.“23

Most contemporary analysts (as well as lay-men) proceed predominantly-albeit most oftenimplicitly-in terms of this model when attempt-ing to explain happenings in foreign affairs. In-deed, that occurrences in foreign affairs are thenets of nations seems so fundamental to think-

23 Schelling, op. cit., p. 4.“‘See Morgenthau, op. ci~., p. 5; Hoffmann, Con-

temporary Theory, pp. 178-179; Hoffmann, “Rou-lette in the Cellar,” The State of War; Schelling,op. Gt.

THE CUBAN MISSILE CRISIS 693

ing about such problems that this underlyingmodel has rarely been recognized: to explain anoccurrence in foreign policy simply means toshow how the government could have rationallychosen that action .2* These brief examples illus-trate five uses of the model. To prove that mostanalysts think largely in terms of the rationalpolicy model is not possible. In this limitedspace it is not even possible to illustrate therange of employment of the framework. Rather,my purpose is to convey to the reader a grasp ofthe model and a challenge: let the reader exam-ine the literature with which he is most familiarand make his judgment.

The general characterization can be sharp-ened by articulating the rational policy model asan “analytic paradigm” in the technical sensedeveloped by Robert I<. Merton for sociologicalanalyses.25 Systematic statement of basic as-sumptions, concepts, and propositions employedby Model I ana1yst.s highlights the distinctivethrust of this style of analysis. To articulate alargely implicit framework is of necessity to car-icature. But caricature can be instructive.

RATIONAL POLICY PARADIGM

I. Basic Unit of Analysis: PolicyChoice

Happenings in foreign affairs areactions chosen b y the nation

as National

conceived asor national

1 I ’w The larger study examines several exceptionsto this generalization. Sidney Verba’s excellentessay “Assumptions of Rationality and Non-Ra-tionality in Models of the International System” isless an exception than it is an approach to a some-what different problem. Verba focuses upon modelsof rationality and irrationality of individuaZ st.ates-men: in Knorr and Verba, The International Sgs-tern.

%Rohert K. Merton, So&I Theory and Social&ruc tures (Revised and Enlarged Edition ; NewYork, 1957), pp. 12-16. Considerably weaker thana satisfact,ory theoretical model, paradigms never-theless represent a short step in that directionfrom looser, implicit conceptual models. Neitherthe concepts nor the relations among the variablesare sufficiently specified to yield propositionsr de-ductively. “Paradigmatic Analysis” neverthelesshas considerable promise for clarifying and codify-ing styles of analysis in political science. Each ofthe paradigms stated here can be represented rig- .orously in mathematical terms. For example,Model I lends itself to mathematical formulationalong the lines of Herbert Simon’s “BehavioralTheory of Rationality,” Models of Man (NewYork, 1957). But this does not solve the most dif-ficult problem of “measurement and estimation.”

694 THE AMERICAN POLITICAL SCIENCE REVIEW V O L. 6 3

government? Governments select the actionthat will maximize strategic goals and objec-tives. These “solutions” to strategic problemsare the fundamental categories in terms ofwhich the analyst perceives what is to be ex-plained.

II. Organizing ConceptsA. National Actor. The nation or govern-

ment, conceived as a rational, unitary decision-maker, is the agent. This actor has one set ofspecified goals (the equivalent of a consistentutility function), one set of perceived options,and a single estimate of the consequences thatfollow from each alternative.

B. The ProbZem. Action is chosen in responseto the strategic problem which the nation faces.Threats and opportunities arising in the “inter-national strategic market place” move the na-t,ion to act.

C. Static Selection. The sum of activity ofrepresentatives of the government relevant to aproblem constitutes what the nation has chosenas its “solution.” Thus the action is conceived asa steady-state choice among alternative out-comes (rather than, for example, a large numberof partial choices in a dynamic stream).

D. Action as Ratio& Choice. The componentsinclude :

1. Goals and Objectives. National securityand national interests are the principal cat-egories in which strategic goals are conceived.Nations seek security and a range of further ob-jectives. (Analysts rarely translate strategicgoals and objectives into an explicit utilityfunction; nevertheless, analysts do focus onmajor goals and objectives and trade off side ef-fects in an intuitive fashion.)

2. Options. Various courses of action rele-vant to a strategic problem provide the spec-trum of options.

3. Consequences. Enactment of each alterna-tive course of action will produce a series of

26 Though a variant of this model could easily bestochastic, this paradigm is stated in non-probabil-istic terms. In contemporary strategy, a stochasticversion of this model is sometimes used for predic-tions; but it is almost impossible to find an ex-planation of an occurrence in foreign affairs that isconsistently probabilistic.

Analogies between Model I and the concept ofexplanation developed by R. G. Collingwood, Wil-liam Dray, and other “revisionists” among philoso-phers concerned with the critical philosophy ofhistory are not accidental. For a summary of the“revisionist position” see Maurice Mandelbaum,“Historical Explanation : The Problem of CoveringLaws,” History and Theory (1960).

consequences. The relevant consequences consti-tute benefits and costs in terms of strategicgoals and objectives.

4. Choice. Rational choice is value-maximix-ing. The rational agent selects the alternativewhose consequences rank highest in terms of hisgoals and objectives.

111. Dominant Inference PatternThis paradigm leads analysts to rely on the

following pattern of inference: if a nation per-formed a particu1a.r action, that nation musthave had ends towards which the action consti-tuted an optimal means. The rational policymodel’s explanatory power stems from thisinference pattern. Puzzlement is relieved by re-vealing the purposive pattern within which theoccurrence can be located as a value-maximizingmeans.

IV. General PropositionsThe disgrace of political science is the infre-

quency with which propositions of any general-ity are formulated and tested. “Paradigmaticanalysis” argues for explicitness about the termsin which analysis proceeds, and seriousnessabout the logic of explanation. Simply to illus-trate the kind of propositions on which analystswho employ this model rely, the formulation in-cludes several.

The basic assumption of value-maximizingbehavior produces propositions central to mostexplanations. The general principle can be for-mulated as follows: the likelihood of any partic-ular action results from a combination of thenation’s (1) relevant values and objectives, (2)perceived alternative courses of action, (3) esti-mates of various sets of consequences (whichwill follow from each alternative), and (4) netvaluation of each set of consequences. Thisyields two propositions.

A. An increase in the cost of an alternative,i.e., a reduction in the value of the set of conse-quences which will follow from that action, or areduction in the proba,bility of attaining fixedconsequences, reduces the likelihood of that al-ternative being chosen.

B. A decrease in the costs of an alternative,i.e., an increase in the value of the set of conse-quences which will follow from that alternative,or an increase in the probability of attainingfixed consequences, increases the likelihood ofthat action being chosen.27

27 This model is an analogue of the theory of therational entrepreneur which has been developedextensively in economic theories of the firm andthe consumer. These two propositions specify the“substitution effect.” Refkement of this model and


V. Specific PropositionsA. Deterrence. The likelihood of any particu-

lar attack results from the factors specified inthe general proposition. Combined with factualassertions, this general proposition yields thepropositions of the sub-theory of deterrence.

(I) A stable nuclear balance reduces the like-lihood of nuclear attack. This proposition is de-rived from the genera1 proposition plus the as-sert.ed fact that a second-strike capability affectsthe potential attacker’s calculations by increas-ing the likelihood and the costs of one particu-lar set of consequences which might follow fromattack-namely, retaliation.

(2) A stable nuclear balance increases theprobability of limited war. This proposition isderived from the general proposition plus the as-serted fact that though increasing the costs of anuclear exchange, a stable nuclear balance nev-ertheless produces a more significant reductionin the probability that such consequences wouldbe chosen in response to a limited war. Thus thisset of consequences weighs less heavily in thecalculus.

B. Soviet Force Posture. The Soviet Unionchooses its force posture (i.e., its weapons andtheir deployment) as a value-maximizing meansof implementing Soviet strategic objectives andmilitary doctrine. A proposition of this sort un-derlies Secretary of Defense Laird’s inferencefrom the fact of 200 SS-9s (large interconti-nental missiles) to the assertion that, “the So-viets are going for a first-strike capability, andthere’s no question about it.“23

VARIANTS OF THE RATIONAL POLICY MODEL

This paradigm exhibits the characteristics ofthe most refined version of the rational model.The modern literature of strategy employs amodel of this sort. Problems and pressures inthe “international strategic marketplace” yieldprobabilities of occurrence. The internationalactor, which could be any national actor, is sim-ply a value-maximizing mechanism for get,tingfrom the strategic problem to the logical solu-tion. But the explanations and predictions pro-duced by most analysts of foreign affairs dependprimarily on variants of this “pure” model. Thepoint of each is the same: to place the actionwithin a value-maximizing framework, givencertain constraints. Nevertheless, it may behelpful to identify several variants, each ofwhich might be exhibited similarly as a para-digm. The first focuses upon the national actor

specification of additional general propositions bytranslating from the economic theory is straight-forward.

28 New York Times, &larch 22, 1969.

and his choice in a particuIar situation, leadinganalysts to further constrain the goals, alterna-tives, and consequences considered. Thus, (I)national propensities or personality traits re-flected in an “operational code,” (2) concernwith certain objectives, or (3) special principlesof action, narrow the “goills” or “alternatives”or “consequences” of the paradigm. For exam-ple, the Soviet deployment of ABMs is some-times explained by reference to the Soviet’s “de-fense-mindedness.” Or a particular Soviet ac-t$ion is explained as an instance of a special ruleof action in the Bolshevik operational code.2g Asecond, related, cluster of variants focuses onthe individual leader or leadership group as theactor whose preference function is maximizedand whose personal (or group) characteristicsare allowed to modify the alternatives, conse-quences, and rules of choice. Explanations of theU.S. involvement in Vietnam as a natural conse-quence of the Kennedy-Johnson Administra-tion’s axioms of foreign policy rely on this var-iant. A third, more complex variant of the basicmodel recognizes the existence of several actorswithin a government, for example, Hawks andDoves or military and civilians, but attempts toexplain (or predict) an occurrence by referenceto the objectives of the victorious actor. Thus,for example, some revisionist histories of the ColdWar recognize the forces of light and the forcesof darkness within the U.S. government, but ex-plain American actions as a result of goals andperceptions of the victorious forces of darkness.

Each of these forms of the basic paradigmconstitutes a formalization of what analysts typi-cally rely upon implicitly. In the transition fromimplicit conceptual model to explicit paradigmmuch of the richness of the best employments ofthis model has been lost. But the purpose inraising loose, implicit conceptual models to anexplicit level is to reveal the basic logic of ana-lysts’ activity. Perhaps some of the remainingartificiality that surrounds the statement of theparadigm can be erased by noting a number ofthe standard additions and modifications em-ployed by analysts who proceed predominantlywithin the rational policy model. First, in thecourse of a document, analysts shift from onevariant of the basic model to another, occasion-ally appropriating in an ad hoc fashion aspectsof a situation which are logically incompatiblewith the basic model. Second, in the course ofexplaining a number of occurrences, analystssometimes pause over a particular event aboutwhich they have a great deal of information andunfold it in such detail that an impression of

29 See Nathan Leites, A Studg of Bokkevkm(Glencoe, Illinois, 1953).

696 ?‘IIIE AMIi:RICAN POLITICAL SCIEKCE REVIEW V O L. 6 3

randomness is created. Third, having employedother assumptions and categories in deriving anexplanation or prediction, analysts will presenttheir product in a neat, convincing rational pol-icy model package. (This accommodation is afavorite of members of the intelligence commu-nity whose association with the details of a pro-cess is considerable, but who feel that by puttingan occurrence in a larger rational framework, itwill be more comprehen.sible to their audience.)Fourth, in attempting to offer an explanation-particularly in cases where a prediction derivedfrom the basic model has failed-the notion of a“mistake” is invoked. Thus, the failure in theprediction of a “missile gap” is written off as aSoviet mistake in not taking advantage of theiropport,unity. Both these and other modificationspermit Model I analysts considerably more vari-ety than the paradigm might suggest. But suchaccommodations are essentially appendages tothe basic logic of these analyses.

THE U.S. BLOCKADE OF CUBA: A FIRST CUT3’

The U.S. response to the Soviet Union’s em-placement of missiles in Cuba must be under-stood in strategic terms as simple value-maxi-mizing escalation. American nuclear superioritycould be counted on to paralyze Soviet nuclearpower ; Soviet transgression of the nuclearthreshold in response to an American use oflower levels of violence would be wildly irra-tional since it would mean virtual destruction ofthe Soviet Communist system and Russian na-tion. American local superiority was overwhelm-ing: it could be initiated at a low level whilethreatening with high credibility an ascendingsequence of steps short of the nuclear threshold.All tthat was required was for the United S-tatesto bring to bear its strategic and loca.1 superior-ity in such a way that American determinationto see the missiles removed would be demon-strated, while at the same time allowing Moscowtime and room to ret,reat without humiliation.The naval blackade-euphemistically named a“quarantine” in order to circumvent the nicetiesof international law-did just t,hat.

The U.S. government’s selection of the block-ade followed this logic. Apprised of the presenceof Soviet missiles in Cuba, the President assem-bled an Executive Committee (ExCom) of t,he

“As stated in the introduction, this “case snap-shot” presents, without editorial commentary, aModel I analyst’s explanation of the U.S. block-ade. The purpose is to illustrate a strong, charac-teristic rational policy model account. This accountis (roughly) consistent with prevailing explanationsof these events.

National Security Council and directed them t o“set aside all other tasks to make a prompt andintense survey of the dangers and all possiblecourses of acti0n.“31 This group functioned as“fifteen individuals on our own, representing t,hePresident and not different departments.“32 Asone of the participants recalls, “The remarkableaspect of those meetings was a sense of completeequality.“33 Most of the time during the weekthat followed was spent canvassing all the possi-ble tracks and weighing the arguments for andagainst each. Six major categories of action wereconsidered.

1. Do nothing. U.S. vulnerability to Sovietmissiles was no new thing. Since the U.S. al-ready lived under the gun of missiles based inRussia, a Soviet capability to strike from Cubatoo made little real difference. The real dangerstemmed from the possibility of U.S. over-reac-tion. The U.S. should announce the Soviet ac-tion in a calm, casual manner thereby deflatingwhatever political capital Khrushchev hoped tomake of the missiles.

This argument fails on two counts. First, itgrossly underestimates the military importanceof the Soviet move. Not only would the SovietUnion’s missile capability be doubled and theU.S. early warning system outflanked. The So-viet Union would have an opportunity to re-verse the strategic balance by further installa-tions, and indeed, in the longer run, to invest incheaper, shorter-range rather than more expen-sive longer-range missiles. Second, the politicalimportance of this move was undeniable. TheSoviet Union’s act challenged the AmericanPresident’s most solemn warning. If the U.S.failed to respond, no American commitmentwould be credible.

2. Diplomatic pressures. Several forms wereconsidered: an appeal to the U.N. or O.A.S.for an inspection team, a secret approach toKhrushchev, and a direct approach to Khru-shchev, perhaps at) a summit meeting. The UnitedStates would demand that the missiles be re-moved, but the final settlement might includeneutralization of Cuba, U.S. withdrawal fromthe Guantanamo base, and withdrawal of U.S.Jupiter missiles from Turkey or Italy.

Each form of the diplomatic approach had itsown drawbacks. To arraign the Soviet Unionbefore the U.N. Security Council held littlepromise since the Russians could veto any pro-posed action. While the diplomats argued, the ’missiles would become operational. To send a se-cret emissary to Khrushchev demanding that

31 Theodore Sorensen, op. cit., p. 675.” Ibid., p. 679.*Ibid., p. 679.

1969 CONCEPTUAL MODELS AND THE CUBAN MISSILE CRISIS G97

the missiles be withdrawn would be to pose un-tenable alternatives. On the one hand, thiswould invite Khrushchev to seize the diplomaticinitiative, perhaps committing himself to strate-gic retaliation in response to an attack on Cuba.On the other hand, this would tender an ulti-matum that no great power could accept. Toconfront Khrushchev at a summit would guar-antee demands for U.S. concessions, and theanalogy between U.S. missiles in Turkey andRussian missiles in Cuba could not be erased.

But why not trade U.S. Jupiters in Turkeyand Italy, which the President had previouslyordered withdrawn, for the missiles in Cuba?The U.S. had chosen to withdraw these missilesin order to replace them with superior, lessvulnerable Mediterranean Polaris submarines.But the middle of the crisis was no time forconcessions. The offer of such a deal might sug-gest to the Soviets that the West would yieldand thus tempt them to demand more. It wouldcertainly confirm European suspicions aboutAmerican willingness to sacrifice European in-terests when the chips were down. Finally, thebasic issue should be kept clear. As the Presi-dent stated in reply to Bertrand Russell, “Ithink your attent,ion might well be directed tothe burglars rat,her than to those who havecaught the burglars.“34

3. A secret approach to Castro. The crisisprovided an opportunity to separate Cuba andSoviet Communism by offering Castro the alter-natives, “split or fall.” But Soviet troops trans-ported, constructed, guarded, and controlled themissiles. Their removal would thus depend on aSoviet decision.

4. Invasion. The United States could takethis occasion not only to remove the missiles butalso to rid itself of Castro. A Navy exercise hadlong been scheduled in which Marines, ferriedfrom Florida in naval vessels, would liberate theimaginary island of Vieques.35 Why not simplyshift the point of disembarkment ? (The Penta-gon’s foresight in planning this operation wouldbe an appropriate antidote to the CIA’s Bay ofPigs ! )

Preparations were made for an invasion, butas a last resort. American troops would beforced to confront 20,000 Soviets in the firstCold War case of direct contact between thetroops of the super powers. Such brinksmanshipcourted nuclear disaster, practically guaran-teeing an equivalent Soviet move against Berlin.

5. Surgical air strike. The missile sites should

a Elie Abel, The Missile CrSs (New York ,1966), p. 144.

sIbid., p. 102.

be removed by a clean, swift conventional at-tack. This was the effective counter-actionwhich the attempted deception deserved. A sur-gical strike would remove the missiles and thuseliminate both the danger that the missilesmight become operational and the fear that theSoviets would discover the American discoveryand act first.

The initial attractiveness of this alternativewas dulled by several difficulties. First, couldthe strike really be “surgical”? The Air Forcecould not guarantee destruction of all themissiles?* Some might be fired during the at-tack; some might not have been identified. Inorder to assure destruction of Soviet and Cubanmeans of retaliating, what was required was nota surgical but rather a massive attack-of atleast 500 sorties. Second, a surprise air attackwould of course kill Russians at the missile sites.Pressures on the Soviet Union to retaliate wouldbe so strong that an attack on Berlin or Turkeywas highly probable. Third, the key problemwith this program was that of advance warning.Could the President of the United States, withhis memory of Pearl Harbor and his vision offuture U.S. responsibility, order a “Pearl Harborin reverse”? For 175 years, unannounced Sun-day morning attacks had been an anathema toour tlradition.3*

6. Blockade. Indirect military action inthe form of a blockade became more attractiveas the ExCom dissected the other alternatives.An embargo on military shipments to Cuba en-forced by a naval blockade was not withoutflaws, however. Could the U.S. blockade Cubawithout inviting Soviet reprisal in Berlin? Thelikely solution to joint blockades would be thelifting of both blockades, restoring the new sta-tus quo, and allowin,a the Soviets additionaltime to complete the missiles. Second, the possi-ble consequences of the blockade resembled thedrawbacks which disqualified the air strike. IfSoviet ships did noti stop, the United Stateswould be forced to fire the first shot, inviting re-taliation. Third, a blockade would deny the tradi-tional freedom of the seas demanded by severalof our close allies and might be held illegal, inviolation of the U.N. Charter and internationallaw, unless the United States could obt,ain atwo-thirds vote in the O.A.S. Finally, how

Vorensen, op. cit., p. 684.81 Ibid., p. 685. Though this was the formulation

of the argument, the facts are not strictly accurate.Our tradition against surprise attack was ratheryounger than 175 years. For example PresidentTheodore Roosevelt applauded Japan’s attack onRussia in 1904.

698 THE AMERICAN POLITICAL SCIE:NCE REVIEW VOL. 63

could a blockade ‘be related to the problem,namely, some 75 missiles on the island of Cuba,approaching operational readiness daily? Ablockade offered the Soviets a, spectrum of de-laying tactics with which to buy time to com-plete the missile installations. Was a fait accom-pli not required?

In spite of these enormous difficulties theblockade had comparative advantages: (1) Itwas a middle course between inaction and at-tack, aggressive enough to communicate firm-ness of intention, but nevertheless not so precip-itous as a strike. (2) It placed on Khrushchevthe burden of choice concerning the next step.He could avoid a direct military clash by keep-ing his ships away. His was the last clearchance. (3) No possible military confrontationcould be more acceptable to the U.S. than anaval engagement in the Caribbean. (4) Thismove permitted the U.S., by flexing its conven-tional muscle, to exploit the threat of subsequentnon-nuclear steps in each of which the U.S.would have significant superiority.

Particular arguments about advantages anddisadvantages were powerful. The explanation ofthe American choice of the blockade lies in amore general principle, however. As PresidentKennedy stated in drawing the moral of thecrisis :

Above all, while defending our own vital inter-ests, nuclear powers must avert those confronta-tions which bring an adversary to a choice of eithera humiliating retreat or a nuclear war. To adoptthat kind of course in the nuclear age would beevidence only of the bankruptcy of our policy-ofa collective death wish for the world?

The blockade was the United States’ only realoption.

MODEL II : ORGANIZATIONAL PROCESS

For some purposes, governmental behaviorcan be usefully summarized as action chosen bya unitary, rational decisionmaker: centrally con-trolled, completely informed, and value maxi-mizing. But this simplification must not be al-lowed to conceal the fact that a “government”consists of a conglomerate of semi-feudal,loosely allied organizations, each with a substan-tial life of its own. Government leaders do sitformally, and to some extent in fact, on top ofthis conglomerate. But. governments perceiveproblems through organizational sensors. Gov-ernments define alternatives and estimate conse-quences as organizations process information.Governments act as these organizations enactroutines. Government behavior can therefore be

35 New York Times,, June, 1963.

understood according to a second conceptualmodel, less as deliberate choices of leaders andmore as outputs of large organizations function-ing according to standard patterns of behavior.

To be responsive to a broad spectrum ofproblems, governments consist of large organiza-tions among which primary responsibility forparticular areas is divided. Each organizationattends to a special set of problems and acts inquasi-independence on these problems. But fewimportant problems fall exclusively within thedomain of a single organization. Thus govern-ment behavior relevant to any important prob-lem reflects the independent output of severalorganizations, partially coordinated by govern-ment leaders. Government leaders can substan-tially disturb, but not substa,ntially control, thebehavior of these organizations.

To perform complex routines, the behavior oflarge numbers of individuals must be coordi-nated. Coordination requires standard operatingprocedures: rules according to which things aredone. Assured capability for reliable perfor-mance of action that depends upon the behaviorof hundreds of persons requires established“programs.” Indeed, if the eleven members of afootball team are to perform adequately on anyparticular down, each player must not “do whathe thinks needs to be done” or “do what thequarterback tells him to do.” Rather, eachplayer must perform the maneuvers specified bya previously established play which the quarter-back has simply called in this situation.

At any given time, a government consists ofexisting organizations, each with a fixed set ofstandard operating procedures and programs.The behavior of these organizations-and conse-quently of the government-relevant to an issuein any particular instance is, therefore, deter-mined primaIrily by routines established in theseorganizations prior to that instance. But organi-zations do change. Learning occurs gradually,over time. Dra!matic organizational change oc-curs in response to major crises. Both learningand change are influenced by existing organiza-t ion al capabilities.

Borrowed from studies of organizations, theseloosely formulated propositions amount simplyto tendencies. Each must be hedged by modifierslike “other things being equal” and “under cer-tain conditions.” In particular instances, tenden-cies hold-more or less. In specific situations,the relevant question is: more or less? But this .is as it should be. For, on the one hand, “organi-zations” are no more homogeneous a class than“solids .” When scientists tried to generalizeabout “solids,P’ they achieved similar results.Solids tend to expand when heated, but some doand some don’t. More adequate categorization

1969 CONCEPTUAL MODELS AND THE CUBAN MISSILE CRISIS

of the various elements now lumped under therubric “organizations” is thus required. On theother hand, the behavior of particular organiza-tions seems considerably more complex than t!hebehavior of solids. Additional information abouta particular organization is required for furtherspecification of the tendency statements. In spiteof these two caveats, the characterization ofgovernment action as organizational output. dif-fers distinctly from Model I. Attempts to under-stand problems of foreign affairs in terms of thisframe of reference should produce quite differ-ent explanations.3g

ORGANIZATIONAL PROCESS PARL4DIGM40

I. Basic Unit of Analysis: Policy as Organiza-<tional Output

The happenings of interna,tional politics are,in three critical senses, outputs of organizationalprocesses. First, the actual occurrences are orga-nizational outputs. For example, Chinese entryinto the Korean War-that is, the fact that]Chinese soldiers were firing at U.N. soldierssouth of the Yalu in 19&--is an organizationalaction: the action of men who are soldiers inplatoons which are in companies, which in turnare in armies, responding as privates to lieuten-ants who are responsible to captains and so on

as The influence of organizational studies uponthe present literature of foreign affairs is minimal.Specialists in international politics are not studentsof organization theory. Organization theory hasonly recently begun to studycisionmakers and has not yet

organizationsproduced bell

as de-avioral

studies of national security organizations from adecision-making perspective. It seems unlikely,however, that these gaps will remain unfilled muchlonger. Considerable progress has been madein the study of the business firm as an organiza-tion. Scholars have begun applying these insightsto government organizations, and interest in an or-ganizational perspective is spreading among insti-tutions and individuals concerned with actual gov-ernment operations. The “decisionmaking” ap-proach represented by Richard Snyder, R. Bruck,and B. Sapin, Foreign PoIicy Decision-Making(Glencoe, Illinois, 1962)) incorporates a number ofinsights from organization theory.

‘O The formulation of this paradigm is indebtedboth to the orientation and insights of HerbertSimon and to the behavioral model of the firmstated by Richard Cyert and James March, A Be-havioral Theory of the Firm (Englewood Cliffs,1963). Here, however, one is forced to grapple withthe less routine, less quantified functions of theless differentiated elements in government organi-zations.

699

to the commander, moving into Korea, advanc-ing against enemy troops, and firing accordingto fixed routines of the Chinese Army. Govern-ment leaders’ decisions trigger organizationalroutines. G-overnment. leaders can trim the edgesof this output and exercise some choice in com-bining outputs. But the mass of behavior is de-termined by previously established procedures.Second, existing organizational routines for em-ploying present physical capabilities constitutethe effective options open to government leadersconfronted with any problem. Only the exis-tence of men, equipped and trained as armiesand capable of being transported to NorthKorea, made entry into the Korean War a liveoption for the Chinese leaders. The fact thatfixed programs (equipment, men, and routineswhich exist at the particular time) exhaust thera.nge of buttons that leaders can push is not al-ways perceived by these leaders. But in everycase it is critical for an understanding of what isactually done. Third, organizational outputsstructure the situation within the narrow con-straints of which leaders must contribute their“decision” concerning an issue. Outputs raisethe problem, provide the information, and maket.he initial moves that eolor the face of t,he issuethat is turned to the leaders. As Theodore So-rensen has observed : “Presidents rarely, if ever,make decisions-particularly in foreign affairs-in the sense of writing their conclusions on aclean slate . . . The basic decisions, which con-fine their choices, have all too often beenpreviously made.“dl If one understands the strut-ture of the situation and the face of the issue-which are determined by the organizational out-puts-the formal choice of the leaders is fre-quently anti-climactic.

II. Organizing ConceptsA. Organizational Actors. The actor is notI

a monolithic “nation” or “government” butrather a constellation of loosely allied organiza-tions on top of which government leaders sit.This constellation acts only as component orga-nizations perform routines.42

B. Factored Problems and FractionatedPower. Surveillance of the multiple facets of for-

41 Theodore Sorensen, “you Get to Walk toWork,” New York: Times Magaxine, March 19.1967.

42 Organizations are not monolithic. The properlevel of disaggregation depends upon the objectivesof a piece of analysis. This paradigm is formulatedwith reference to the major organizations that con-stitute the U.S. government. Generalization to themajor components of each department and agencyshould be relatively straightforward.

700 THE AMERICAN POLITICAL SCIENCE REVIEW VOL. 63

eign affairs requires that problems be cut upand parcelled out to various organizations. Toavoid paralysis, primary power must accompanyprimary responsibility. But if organizations arepermitted to do anything, a large part of whatthey do will be determined within the organiza-tion. Thus each organization perceives problems,processes information, and performs a range ofactions in quasi-independence (within broadguidelines of national policy). Factored prob-lems and fractionated power are two edges ofthe same sword. Factoring permits more special-ized attention to particular facets of problemstOhaln would be possible if government leaderstried to cope with these problems by themselves.But this additional attention must be paid for inthe coin of discretion for what an organizationattends to, and how organiza’tional responses areprogrammed.

C. Paroclzial Priorities, Perceptions, and Is-sues. Primary responsibility for a narrow set ofproblems encourages organizational parochial-ism. These tendencies are enhanced by a numberof additional factors : (1) selective informationavailable to the organization, (2) recruitment ofpersonnel into ‘the organization, (3) tenure ofindividuals in the organization, (4) small grouppressures within the organization, and (5) dis-tribution of rewards by the organization. Clients(e.g., interest groups), government allies (e.g.,Congressional committees), and extra-nationalcounterparts (e.g., the British Ministry of De-fense for the Department of Defense, ISA, orthe British Foreign Office for the Departmentof State, EUR) galvanize this parochialism.Thus organizations develop relatively stable pro-pensities concerning operational priorities, per-ceptions, and issues.

D. Action as Orga~2izational Output. The pre-eminent feature of organizational activity isits programmed character: the extent to whichbehavior in any particular case is an enactmentof preestablished routines. In producing outputs,t,he activity of each organization is characterizedbv:

U I. Goals: Constraints Defining AcceptablePerformance. The operational goals of an orga-nization are seldom revealed by formal man-d&es. Rather, each organization’s operationalgoals emerge as a set of constraints defining ac-ceptable performance. Central among these con-straints is organizational health, defined usuallyin terms of bodies assigned and dollars appro-priated. The set of constraints emerges from amix of expectations and demands of other orga-nizations in the government, statutory author-ity, demands from citizens and special interestgroups, and bargaining within the organization.These constraints represent a quasi-resolution of

conflict-the constraints are relatively stable, sothere is some resolution. But conflict among al-ternative goals is always latent; hence, it is aquasi-resolution. Typically, the constraints areformulated as imperatives to avoid roughlyspecified discomforts and disasters.“3

2. Sequential Attention to Goals. The exis-tence of conflict among operational constraintsis resolved by the device of sequential attention.As a problem arises, the subunits of the organi-zation most concerned with that problem dealwith it in terms of the constraints they take tobe most important. When the next problemarises, another cluster of subunits deals with it,focusing on a different set of constraints.

3. Standard Operating Procedures. O r g a -nizations perform their “higher” functions, suchas attending to problem areas, monitoring infor-mation, and preparing relevant responses forlikely contingencies, by doing “lower” tasks, forexample, preparing budgets, producing reports,and developing hardware. Reliable performanceof these tasks requires standard operating proce-dures (hereafter SOPS). Since procedures are“standard” they do not change quickly or easily.Without these standard procedures, it would notbe possible to perform certain concerted tasks.But because of standard procedures, organiza-tional behavior in particular instances often ap-pears unduly formalized, sluggish, or inappropri-ate.

4. Programs and Repertoires. Organizationsmust be capable of performing actions in whichthe behavior of large numbers of individuals iscarefully coordinated. Assured performancerequires clusters of rehearsed SOPS for produc-ing specific actions, e.g., fighting enemy units oranswering an embassy’s cable. Each cluster com-prises a “program” (in the terms both of dramaand computers) which the organization hasavailable for dealing with a situation. The list ofprograms relevant to a type of activity, e.g.,fighting, constitutes an organizational repert(oire.The number of programs in a repertoire is al-ways quite limited. When properly triggered, or-ganizations execute programs ; programs cannotbe substantially changed in a particular situa-tion. The more complex the action and t,hegreater the number of individuals involved, themore important are programs and repertoires asdeterminants of organizational behavior.

5. Uncertainty Avoidance. Organizations donot attempt to estimate the probability distribu-tion of future occurrences. Rather, organizations

**The stability of these constraints is dependenton such factors as rules for promotion and reward,budgeting and accounting procedures, and mun-dane operating procedures.

1969 CONCEPTUAL MODELS AND THE CUBAN MISSILE CHISIS 701

avoid uncertainty. By arranging a negotiatedenvironment, organizations regularize the reac-tions of other actors with whom they have todeal. The primary environment, relations withother organizations that comprise the govern-ment, is stabilized by such arrangements asagreed budgetary splits, accepted areas of re-sponsibility, and established conventional prac-tices. The secondary environment, relationswith the international world, is stabilized be-tween allies by the establishment of contracts(alliances) and “club relations” (U.S. State andU.K. Foreign Office or U.S. Treasury andU.K. Treasury). Between enemies, contracts andaccepted conventional practices perform a simi-la’r function, for example, the rules of the “pre-carious status quo” which President Kennedyreferred to in the missile crisis. Where the inter-national environment cannot be negotiated, or-ganizations deal with remaining uncertainties byestablishing a set of standard scenarios that con-stitute the contingencies for which they prepare.For example, the standard scenario for TacticalAir Command of the U.S. Air Force involvescombat with enemy aircraft. Planes are designedand pilots trained to meet this problem. Thatthese preparat.ions are less relevant to moreprobable contingencies, e.g., provision of close-inground support in limited wars like Vietnam,has had little impact on the scenario.

6. Problem-directed Search. Where situationscannot be construed as standard, organizationsengage in search. The style of search and the so-lution are largely determined by existing rou-tines. Organizational search for alternativecourses of action is problem-oriented: it focuseson the at*ypical discomfort that must beavoided. It is simple-minded : the neighborhoodof the symptom is searched first; then, theneighborhood of the current alternative. Pat-terns of search reveal biases which in turn re-flect such factors as specialized training or ex-perience and patterns of communication.

7. Organizational Learning and Change.The parameters of organizational behaviormostly persist. In response to non-standardproblems, organizations search and routinesevolve, assimilating new situations. Thus learningand change follow in large part from existingprocedures. But marked changes in organiza-tions do sometimes occur. Conditions in whichdramatic changes are more likely include: (1)Periods of budgetary feast. Typically, organiza-tions devour budgetary feasts by purchasing ad-ditional items on the existing shopping list. Nev-ertheless, if committed to change, leaders whocontrol the bud-get can use extra, funds to effectchanges. (2) Periods of prolonged budgetaryfamine. Though a single year’s famine typically

results in few changes in organizational struc-ture but a loss of effectiveness in performingsome programs, prolonged famine forces majorretrenchment. (3) Dramatic performance fail-ures. Dramatic change occurs (mostly) in re-sponse to major disasters. Confronted with anundeniable failure of procedures and repertoires,authorities outside the organization demandchange, existing personnel are less resistant tochange, and critical members of the organizationare replaced by individuals committed tochange.

E. Central Coordination and Control. Actionrequires decentralization of responsibility andpower. But problems lap over the jurisdictions ofseveral organizations. Thus the necessity for de-centralization runs headlong into the require-ment for coordination. (Advocates of one hornor t.he other of this dilemma-responsive actionentails decentralized power vs. coordinated a’c-tion requires central control-account for a con-siderable part of the persistent demand for gov-ernment reorganizat.ion.) Both the necessity forcoordination and the centrality of foreign policyto national welfare guarantee the involvement ofgovernment leaders in the procedures of t,he or-ganizations among which problems are dividedand power shared. Etich organization’s propensi-ties and routines can be disturbed by govern-ment leaders’ intervention. Central direction andpersistent control of organizational activity,however, is not possible. The relation among or-ganizations, and between orgalnizations and thegovernment leaders depends critically on a num-ber of structural variables including: (1) thenature of the job, (2) the measures and infor-mation available to government leaders, (3) thesystem of rewards and punishments for organi-zational members, and (4) the procedures bywhich human and material resources get com-mitted. For example, to the extent that rewardsand punishments for the members of an organi-zation are distributed by higher authorities,these authorities can exercise some control byspecifying criteria in terms of which organiza-tional output is to be evaluated. These criteriabecome constraints within which organization alactivity proceeds. But constraint is a, crude in-strument of control.

Intervention by government leaders doessometimes change the act,ivity of an organiza-tion in an intended direction. But instances arefewer than might be expected. As FranklinRoosevelt, the master manipulator of govern-ment organizations, remarked :

The Treasury is so large and far-ffung and in-grained in its practices that I find it is almost im-possible to get the action and results I want. . . .


But the Treasury is not to be compared with theState Department. You should go through the ex-perience of trying to get any changes in the think-ing, policy, and action of the career diplomats andthen you’d know what a real problem was. But theTreasury and the State Department put togetherare nothing compared with the Na-a-vy . . . Tochange anything in the Na-a-vy is like punching afeather bed. You punch it with your right and youpunch it with your left until you are finally ex-hausted, and then you find the damn bed just as itwas before you started punching.”

John Kennedy’s experience seems to have beensimilar : “The State Department,,” he asserted,“is a bowl full of jelly.“45 And lest the McNa-mara revolution in the Defense Departmentseem too striking a counter-example, the Navy’srecent rejection of McNamara’s major inter-vention in Naval weapons procurement, the F-111B, should be studied as an antidote.

F. Decisions of Government Leaders. Orga-nizational persistence does not exclude shifts ingovernmental behavior. For government leaderssit atop the conglomerate of organizations.Many important issues of governmental actionrequire that these leaders decide what organiza-t.ions will play out which programs where. Thusstability in the parochialisms and SOPS of in-dividual organizations is consistent with someimportant shifts in the behavior of governments.The range of these shifts is defined by existingorganizational programs.

III. Dominant Inference PatternIf a nation performs an action of this type to-

day, its organizational components must yester-day have been performing (or have had estab-lished routines for performing) an action onlymarginally different from this action. At any spe-cific point in time, a government consists of anestablished conglomerate of organizations, eachwith existing goals, programs, and repertoires.The characteristics of a government’s action inany instance follows from those established rou-tines, and from the choice of government leaders-on the basis of information and estimates pro-vided by existing routines-among existing pro-grams. The best explanation of an organization’sbehavior at t is t - I; the prediction of t + Iis t. Model II’s explanatory power is achieved byuncovering the organizational routines and reper-toires that produced the outputs that comprisethe puzzling occurrence.

ti Marriner Eccles, Beckoning Frontiers (NewYork, 1951)) p. 336.

M Arthur Schlesinger, A Thousand Days (Boston,1965), p. 406.

IV. General PropositionsA number of general propositions have been

stated above. In order to illustrate clearly thetype of proposition employed by Model II ana-lysts, this section formulates several more pre-cisely.

A. Organizational Action. Activity accordingto SOPS and programs does not constitutefar-sighted, flexible adaptation to “the issue” (asit is conceived by the analyst). Detail and nu-ance of actions by organizations are determinedpredominantly by organizational routines, notgovernment leaders’ directions.

1. SOPS constitute routines for dealing withstandard situations. Routines allow large num-bers of ordinary individuals to deal with numer-ous instances, day after day, without consider-able thought, by responding to basic stimuli.But this regularized capability for adequate per-formance is purchased at the price of standardi-zation. If the SOPS are appropriate, averageperformance, i.e., performance averaged over therange of cases, is better than it would be if eachinstance were approached individually (givenfixed talent, timing, and resource constraints).But specific instances, particularly critical in-stances that typically do not have %tandard”characteristics, are often handled sluggishly orinappropriately.

2. A program, i.e., a complex action chosenfrom a short list of programs in a repertoire, israrely tailored to the specific situation in whichit is executed. Rather, the program is (at best)the most appropriate of the programs in a pre-viously developed repertoire.

3. Since repertoires are developed by paro-chial organizations for standard scenarios de-fined by that organization, programs availablefor dealing with a particular situation are oftenill-suited.

B. Limited Flexibi l i ty and IncrementalChange. Major lines of organizational action arestraight, i.e., behavior at one time is marginallydifferent from that behavior at t - I. Simple-minded predictions work best: Behavior at t + 1will be marginally different from behavior at thepresent time.

1. Organizational budgets change incremen-tally-both with respect to totals and with re-spect to intra-organizational splits. Though or-ganizations could divide the money availableeach year by carving up the pie anew (in thelight of changes in objectives or environment), .in practice, organizations take last year’s budgetas a base and adjust incrementally. Predictionsthat require large budgetary shifts in a singleyear between organizations or between unitswithin an organization should be hedged.

2. Once undertaken, an organizational in-


vestment is not dropped at the point where “ob-jective” costs outweigh benefits. Organizationalstakes in adopted projects carry them quitebeyond the loss point.

C. Administrative Feasibility. Adequate ex-planation, analysis, and prediction must includeadministrative feasibility as a major dimension.A considerable gap separates what leaderschoose (or might rationally have chosen) andwhat organizations implement.

1. Organizations are blunt instruments. Proj-ects that require several organizations to actwith high degrees of precision and coordinationare not, likely to succeed.

2. Projects that demand that existing organi-zational units depart from their accustomedfunct,ions and perform previously unpro-grammed tasks are rarely accomplished in theirdesigned form.

3. Government leaders can expect that eachorganization will do its “part” in terms of whatthe organization knows how to do.4. Government leaders can expect incomplete

and distorted information from each organiza-tion concerning its part of the problem.

5. Where an assigned piece of a problem iscontrary to the existing goals of an organization,resistance to implementation of that piece willbe encountered.

V. Specific Propositions.1. Deterrence. The probability of nuclear at-

tack is less sensitive to balance and imbalance,or stability and instability (as these conceptsare employed by Model I strategists) than it isto a number of organizational factors. Exceptfor the special case in which the Soviet Unionacquires a credible capability to destroy the U.S.with a disarming blow, U.S. superiorit,y or infe-riority affects the probability of a nuclear attackless than do a number of organizational factors.

First, if a nuclear attack occurs, it will resultfrom organizational activity: the firing of rock-ets by members of a missile group. The enemy’scontrol system, i.e., physical mechanisms andstandard procedures which determine who canlaunch rockets when, is critlical. Second, the ene-my’s programs for bringing his strategic forcesto alert status determine probabilities of acci-dental firing and momentum. At the outbreak ofWorld War I, if the Russian Tsar had under-stood the organizational processes which hisorder of full mobilization triggered, he wouldhave realized t.hat he had chosen war. Third, or-ganizational repertoires fix the range of effectivechoice open to enemy leaders. The menu avail-able to Tsar Nicholas in 1914 has two entrees:full mobilization and no mobilization. Partial mo-bilization was not an organizational option.

Fourth, since organizational routines set thechessboard, the training and deployment oftroops and nuclear weapons is crucial. Giventhat the outbreak of hostilities in Berlin is moreprobable than most scenarios for nuclear war,facts about deployment, training, and tacticalnuclear equipment of Soviet troops stationed inEast Germany-which will influence the face ofthe issue seen by Soviet leaders at the outbreakof hostilities and t,he manner in which choice isimplemented-are as critical as the question of“balance.”

2. Soviet Force Posture. Soviet force posture,i.e., the fact that certain weapons rather thanothers are procured and deployed, is determinedby organizational factors such as the goals andprocedures of existing military services and thegoals and processes of research and design labs,within budgetary constraints that emerge fromthe government leader’s choices. The frailty ofthe Soviet Air Force wit,hin the Soviet militaryestablishment seems to have been a crucial ele-ment in the Soviet failure to acquire a largebomber force in the 1950s (thereby faultingAmerican intelligence predictions of a “bombergap”). The fact that missiles were controlleduntil 1960 in the Soviet Union by the SovietGround Forces, whose goals and procedures re-flected no interest in an intercontinental mission,was not irrelevant to the slow Soviet buildup ofICBMs (thereby faulting U.S. intelligence pre-dictions of a “missile gap”). These organiza-tional factors (Soviet Ground Forces’ control ofmissiles and that service’s fixation with Euro-pean scenarios) make the Soviet deployment ofso many MRBMs that European targets couldbe destroyed three times over, more understand-able. Recent weapon developments, e.g., thetesting of a Fractional Orbital BombardmentSystem (FOBS) and multiple warheads forthe SS-9, very likely reflect the activity and in-terests of a cluster of Soviet research and devel-opment organizations, rather than a decision bySoviet leaders to acquire a first strike weaponsystem. Careful attention to the organizationalcomponents of the Soviet military establishment(Strat.egic Rocket Forces, Navy, Air Force,Ground Forces, and National Air Defense), themissions and weapons systems to which eachcomponent is wedded (an independent weaponsystem assists survival as an independent ser-vice), and existing budgetary splits (whichprobably are relatively stable in the SovietUnion as they tend to be everywhere) offer po-tential improvements in medium and longerterm predictions.

THE tr.S. l3LOCICAI)E OF CUBA: A SECOND CUT

Orgnnixationd hte%gence. At 7 : 00 P.M. on


October 22, 1962, President Kennedy disclosedthe American discovery of the presence of So-viet strategic missiles in Cuba, declared a “strictquarantine on all offensive military equipmentunder shipment to Cuba,” and demanded that,“Chairman Khrushchev halt and eliminate thisclandestine, reckless, and provocative threat toworld peace.““6 This decision was reached at thepinnacle of the U.S. Government after a criticalweek of deliberation. What initiated that pre-cious week were photogra.phs of Soviet missilesites in Cuba taken on October 14. These pic-tures might not have been taken until a weeklater. In that case, the President speculated, “Idon’t think probably we would have chosen asprudently as we finally didY4? U.S. leadersmight have received this information threeweeks earlier-if a U-2 had flown over San Cris-tobal in the last week of September? What de-termined the context in which 4merican leaderscame to choose the blockade was the discoveryof missiles on October 14.

There has been considerable debate over al-leged American “intelligence failures” in theCuban missile crisis. 4g But what both critics anddefenders ha.ve neglected is the fact that the dis-covery took place on October 14, rather thanthree weeks earlier or a week later, as a conse-quence of the established routines and proce-dures of the organizations which constitute theU.S. intelligence community. These organiza-tions were neither more nor less successful thanthey had been the previous month or were to bein the months to follow.50

The notorious “September estimate,” approvedby the United States Intelligence Board (USIB)on September 19, concluded that the SovietUnion would not introduce offensive missilesinto Cuba.51 No U-2 flight was directed over thewestern end of Cuba. (after September 5) before

46 U.S. Department of State, Bdetin, XLVII, pp.715-720.

‘7 Schlesinger, op. cit., p. 803.4s Theodore Sorensen, Kennedy, p. 675.” See U.S. Congress, Senate, Committee on

Armed Services, Preparedness Investigation Sub-committee, Interim Report on Cuban MilitaryBuild-up, 88th Congress, 1st Session, 1963, p. 2;Hanson Baldwin, “Growing Risks of BureaucraticIntelligence,” The Reporter (August 15, 1963), 4%50; Roberta Wohlstetter, ‘Cuba and Pearl Har-bor,” Foreign Afkzirs (July, 1965)) 706.

w U.S. Congress, House of Representatives, Com-mittee on Appropriations, Subcommittee on De-partment of Defense Appropriations, Hearings,88th Congress, 1st Session, 1963,25 ff.

m R. Hilsmq To Move a Nation (New York,1967). nx>. 172-173.

October 4.52 No U-2 flew over the western endof Cuba until the flight that discovered the So-viet missiles on October 14/s Ca.n these “fail-ures” be accounted for in organizational terms?

On September 19 when USIB met to con-sider the question of Cuba, the “system” con-tained the following information: (1) shippingintelligence had noted the arrival in Cuba oftwo large-hatch Soviet lumber ships, which wereriding high in the water; (2) refugee reports ofcountless sightings of missiles, but also a reportthat Castro’s private pilot, after a night of drink-ing in Havana, had boasted: “We will fight tothe death and perhaps we can win because wehave everything, including atomic weapons” ;(3) a sighting by a CIA agent of the rear profileof a strategic missile; (4) U-2 photos producedby flights of August 29, September 5 and 17showing the construction of a number of SAMsites and other defensive missiles.54 Not all ofthis information was on the desk of the estima-tors, however. Shipping intelligence expertsnoted the fact that large-hatch ships were ridinghigh in the water and spelled out the inference:the ships must be carrying (‘space consuming”cargo.55 These facts were carefully included inthe catalogue of intelligence concerning shipping.For experts sensitive to the Soviets’ shortage ofships, however, these facts carried no special sig-nal. The refugee report of Castro’s private pi-lot’s remark had been received at Opa Locka,Florida, along with vast reams of inaccurate re-ports generated by the refugee community. Thisreport and a thousand others had to be checkedand compared before being sent to Washington.The two weeks required for initial processingcould have been shortened by a large increase inresources, but the yield of this source was al-ready quite marginal. The CIA agent’s sightingof the rear profile of a strategic missile had oc-

62 Department of Defense Appropriations, Hear-ings, p. 67.

83 Ibid., pp. 6647.64For (1) Hilsman, op. cit., p. 186; (2) Abel, op.

cit., p. 24; (3) Department of Defense Appropria-tions, Hearings, p. 64; Abel, op. cit., p. 24; (4) De-partment of Defense Appropriations, Hearings, pp.l-30.

“The facts here are not entirely clear. This as-sertion is based on information from (1) “Depart-ment of Defense Briefing by the Honorable R. S.McNamara, Secretary of Defense, State Depart- .menf Auditorium, ZOO p.m., February 6, 1963.” Averbatim transcript of a presentation actuallymade by General Carroll’s assistant, John Hughes;and (2) Hilsman’s statement, op. cit., p. 186. Butsee R. Wohlstetter’s interpretation, “Cuba andPearl Harbor,” 766.


curred on September 12; transmission timefrom agent sighting to arrival in Washingtontypically took 9 to 12 days. Shortening thistransmission time would impose severe cost interms of danger to sub-agents, agents, and com-munication networks.

On the information available, the intelli-gence chiefs who predicted that the SovietUnion would not introduce offensive missilesinto Cuba made a reasonable and defensiblejudgment.56 Moreover, in the light of the factthat these organizations were gathering intelli-gence not only about Cuba but about potentialoccurrences in all parts of the world, the infor-mational base available to the estimators in-volved nothing out of the ordinary. Nor, froman organizational perspective, is there anythingstartling about the gradual accumulation of evi-dence that led to the formulation of the hypoth-esis that the Soviets were installing missiles inCuba and the decision on October 4 to directa special flight over western Cuba.

The ten-day delay between that decision andthe flight is another organizational story57 Atthe October 4 meeting, the Defense Depart-ment took the opportunity to raise an issue im-portant to its concerns. Given the increaseddanger that a U-2 would be downed, it would bebetter if the pilot were an officer in uniformrather than a CIA agent. Thus the Air Forceshould assume responsibility for U-2 flights overCuba. To the contrary, the CIA argued thatthis was an intelligence operation and thuswithin the CIA’s jurisdiction. Moreover, CIAU-2’s had been modified in certain ways whichgave them adva,ntages over Air Force U-2’s inaverting Soviet SAM%. Five days passed whilethe State Department pressed for less risky al-ternatives such as drones and the Air Force (inDepartment of Defense guise) and CIA engagedin territorial disputes. On October 9 a flightplan over San Cristobal was approved byCOMOR, but to the CIA’s dismay, Air Force pi-lots rather than CIA agents would take chargeof the mission. At this point details becomesketchy, but several members of the intelligencecommunity have speculated that an Air Forcepilot in an Air Force U-2 attempted a high alti-tude overflight on October 9 that “flamedout”, i.e., lost power, and thus had to descend inorder to restart its engine. A second round be-tween Air Force and CIA followed, as a resuIt of

WSee H&man, op. cit., pp. 172-174.“Abel, op. cit., pp. 26 ff; Weintal and Bartlett,

Facing the Brink (New York, 1967), pp. 62 ff ;Cuban Military l3uiZd-up; J. Daniel and J. Hub-bell, &r&e in the West (New York, 1963), pp.15 ff.

which Air Force pilots were trained to fly CIAU-2%. A successful overflight took place onOctober 14.

This ten-day delay constitutes some form of“failure.” In the face of well-founded suspicionsconcerning offensive Soviet. missiles in Cuba thatposed a critical threat to the United States’most vital interest, squabbling between organi-zations whose job it is to produce this informa-tion seems entirely inappropriate. But for eachof these organizations, the question involved theissue : “‘Clrhose job was it to be?” Moreover, theissue was not simply, which organization wouldcontrol U-2 flights over Cuba, but rather thebroader issue of ownership of U-2 intelligenceactivities-a very long standing territorial dis-pute. Thus though this delay was in one sense a“failure,” it was also a nearly inevitable conse-quence of two facts: many jobs do not fallneatly into precisely defined organizational juris-dictions ; and vigorous organizations are imperi-alistic.

Organizational Options. Deliberations of lead-ers in ExCom meetings produced broad out-lines of alternatives. Details of these alternativesand blueprints for their implementat.ion had tobe specified by the organizations that would per-form these tasks. These organizational outputsanswered the question : What, specifically, couldbe done?

Discussion in the ExCom quickly narrowedthe live options to two: an air strike and ablockade. The choice of the blockade instead ofthe air strike turned on two points: (1) the ar-gument from morality and tradition that theUnited States could not perpetrate a “PearlHarbor in reverse”; (2) the belief that a “surgi-cal” air strike was impossible.5* Whether theUnited States might strike first was a questionnot of capability but of morality. Whether theUnited States could perform the surgical strikewas a factual question concerning capabilities.The majority of the members of the ExCom, in-cluding the President, initially preferred the airstrike.5g What effectively foreclosed this option,however, was the fact that the air strike theywanted could not be chosen with high confidenceof success ?O After having tentatively chosen thecourse of prudence-given that the surgical airstrike was not an option-Kennedy reconsid-ered. On Sunday morning, October 21, he calledthe Air Force experts to a special meeting in hisliving quarters where he probed once more for .the option of a %rgicaZ” air strike.61 General

” Schlesinger, op. cit., p. 804.“Sorensen, Kennedy, p. 684.@‘Ibid., pp. 684 ff.Vbid., pp. 6a-6w


Walter C. Sweeny, Commander of Tactical AirForces, asserted again that the Air Force couldguarantee no higher than ninety percent effec-tiveness in a surgical air strike? Tha.t “fact”was false.

The air strike alternative provides a classiccase of military estimates. One of the alterna-tives outlined by the ExCom was named “airstrike.” Specification of the details of this alter-native was delegated to the Air Force. Startingfrom an existing plan for massive U.S. militaryaction against Cuba (prepared for contingencieslike a response to a Soviet Berlin grab), AirForce estimators produced an attack to guaran-tee success.63 This plan called for extensivebombardment of all missile sites, storage depots,airports, and, in deference to the Navy, the ar-tillery batteries opposite the naval base atGuantanamo .64 Members of the ExCom repeat-edly expressed bewilderment at military esti-mates of the number of sorties required, likelycasualties, and collateral damage. But the “sur-gical” air strike that the political leaders had inmind was never carefully examined during thefirst week of the crisis. Rather, this option wassimply excluded on the grounds that since the So-viet MRBM’s in Cuba were classified “mobile” inU.S. manuals, extensive bombing was required.During the second week of the crisis, careful ex-amination revealed that the missiles were mobile,in the sense that small houses are mobile: that is,they could be moved and reassembled in 6 days.After the missiles were reclassified “movable”and detailed plans for surgical air strikes speci-fied, this action was added to the list of live op-tions for the end of the second week.

Organizational Implementation. Ex-Cornmembers separated several types of blockade:offensive weapons only, all armaments, and allstrategic goods including POL (petroleum, oil,and lubricants). But the “details” of the opera-tion were left to the Navy. Before the Presidentannounced t*he blockade on Monday evening,the first stage of the Navy’s blueprint was inmotion, and a problem loomed on the horizon?The Navy had a detailed plan for the blockade.The President had several less precise butequally determined notions concerning whatshould be done, when, and how. For the Navythe issue was one of effective implementation ofthe Navy’s blockade-without the meddling andinterference of political leaders. For the Presi-dent, the problem was to pace and manage

6’ Ibid., p. 697; Abel. op. cit., pp. 100-101.63 Sorensen, Kennedy, p. 669.gq Hilsman, oy,. tit ., p. 204.63 See Abel, op. cit., pp. 97 ff.

events in such a way that the Soviet leaderswould have time to see, think, and blink.

A careful reading of available sources uncov-ers an instructive incident. On Tuesday theBrit.ish Ambassador, Ormsby-Gore, after hav-ing attended a briefing on the details of theblockade, suggested to the President that theplan for intercepting Soviet ships far out of reachof Cuban jets did not facilitate Khrushchev’shard decision9 Why not, make the interceptionmuch closer to Cuba and thus give the Russianleader more time? According to the public ac-count and the recollection of a number of indi-viduals involved, Kennedy “agreed immediately,called McNamara, and over emotional Navyprotest, issued the appropriate instructions?As Sorensen records, “in a sharp clash with theNavy, he made certain his will prevailed.“68 TheNavy’s plan for the blockade was thus changedby drawing the blockade much closer to Cuba.

A serious organizational orientation makesone suspicious of this account. More careful ex-amination of the available evidence confirmsthese suspicions, though alternative accountsmust be somewhat speculative. According to thepublic chronology, a quarantine drawn close toCuba became effective on Wednesday morning,the first Soviet ship was contacted on Thursdaymorning, and the first boarding of a ship oc-curred on Friday. According to the statementby the Department of Defense, boarding of theMar&a by a party from the John R. Pierce“took place at 7:50 A.M ., E.D.T., 180 milesnortheast of Nassau.“6s The Marcula had beentrailed since about 10: 30 the previous evening.70Simple calculations suggest that the Pierce musthave been stationed along the Navy’s originalarc which extended 500 miles out to sea fromCape Magsi, Cuba’s eastern most tip.71 Theblockade line was not moved as the Presidentordered, and the accounts report.

What happened is not entirely clear. Onecan be certain, however, that Soviet ships passedthrough the line along which American destroy-ers had posted themselves before the official“first contact” with the Soviet ship. On Oc-tober 26 a Soviet tanker arrived in Havana andwas honored by a dockside rally for “runningthe blockade.” Photographs of this vessel showthe name Vinnitsa on the side of the vessel in

c6 Schlesinger, op. cit., p. 818.67 iTbid.” Sorensen, Kennedy. p. 710.“New York Times,, October 27, 1962.“Abel, op. cit., p. 171.‘* For the location of the original arc see Abel,

op. cit., p. 141.


Cyrillic letters? But according to the officialU.S. position, the first tanker to pass throughthe blockade was the Bucharest, which washailed by the Navy on the morning of October25. Again simple mathematical calculation ex-cludes the possibility that the Bucharest and theVimzitsa were the same ship. It seems probablethat the Navy’s resistance to the President’sorder that the blockade be drawn in closer toCuba forced him to allow one or several Sovietships to pass through the blockade after it wasofficially operative .73

This attempt to leash the Navy’s blockadehad a price. On Wednesday morning, Oct,o-ber 24, what the President had been awaitingoccurred. The 18 dry cargo ships heading to-wards the quarantine stopped dead in the water.This was the occasion of Dean Rusk’s remark,“We are eyeball to eyeball and I think the otherfellow just blinked .“74 But the Navy had an-other interpretation. The ships had simplv.stopped to pick up Soviet submarine escorts.The President became quite concerned lest theNavv-already riled because of Presidentialmeddling in its affairs-blunder into an incident.Sensing the President’s fears, McNamara be-came suspicious of the Navy’s procedures androutines for making the first interception. Call-ing on the Chief of Naval Operations in theNavy’s inner sanctum, the Navy Flag Plot,McNamara put his questions harshbY Who.would make the first interception? VVere Rus-sian-speaking officers on board? How wouldsubmarines be dealt with? At one point McNa-mara asked Anderson what he would do if a So-viet ship’s captain refused to answer questionsabout his cargo. Picking up the Manual of NavyRegulations the Navy man waved it in McNa-mara’s face and shouted, “It’s all in there.” Towhich McNamara replied, “I don’t give a damnwhat John Paul Jones would have done; I wantto know what you are going to do, now.“TG Theencounter ended on Anderson’s remark : “Now,Mr. Secretary, if you a,nd your Deputy will goback to your office the Navy will run the block-.ade?

MODEL III : BUREAUCRATIC POLITICS

The leaders who sit on top of organizations‘2 Fncts on File, Vol. XXII, 1962, p. 376, pub-

lished by Facts on File, Inc., New York, yearly.73 This hypothesis would account for the mystery

surrounding Kennedy’s explosion at the leak of thestopping of the Bucharest. See Hilsman, op. cit.,p. 45.

“Abel, op. cit., p. 153.” See ibid., pp. 154 ff.761bid., p. 156.77 Ibid.

are not a monolithic group. Rather, each is, inhis own right, a player in a central, compet,itivegame. The name of the game is bureaucraticpolitics : bargaining along regularized channelsamong players positioned hierarchically withinthe government. Government behavior can thusbe understood according to a third conceptualmodel not as organizational outputs, but as out-comes of bargaining games. In contrast withModel I, the bureaucratic politics model sees nounitary actor but rather many actors as players,who focus not on a single strategic issue but onmany diverse intra-national problems as well, interms of no consistent set of strategic objectivesbut rather according to various conceptions ofnational, organizational, and personal goals,making government decisions not by rationalchoice but by the pulling and hauling that) ispolitics.

The apparatus of each national governmentconstitutes a complex arena for the intra-na-tional game. Political leaders at the top of thisapparatus plus the men who occupy positions ontop of the critical organizations form the circleof central players. Ascendancy to this circle as-sures some independent standing. The necessarydecentralization of decisions required for actionon the broad range of foreign policy problemsguarantees that each player has considerablediscretion. Thus power is shared.

The nature of problems of foreign policy per-mits fundamental disagreement among reason-able men concerning what ought to be done.Analyses yield conflicting recommendations. Sep-arate responsibilities laid on the shoulders of in-dividual personalities encourage differences inperceptions and priorities. But the issues are offirst order importance. What the nation doesreally matters. A wrong choice could mean ir-reparable damage. Thus responsible men areobliged to fight for what they are convinced isright.

Men share power. Men differ concerningwhat must be done. The differences matter. Thismilieu necessitates that policy be resoIved bypolitics. What the nation does is sometimes theresult of the triumph of one group over others.More often, however, different groups pulling indifferent directions yield a resultant distinctfrom what anyone intended. What moves thechess pieces is not simply the reasons which sup-port a course of action, nor t>he routines of orga-nizations which enact an alternative, but thepower and skill of proponents and opponents ofthe action in question.

This characterization captures the thrust ofthe bureaucratic politics orientation. If problemsof foreign policy arose as discreet issues, and de-cisions were determined one game at a time, this


account would suffice. But most “issues,” e.g.,Vietnam or the proliferation of nuclear weapons,emerge piecemeal, over time, one lump in onecontext, a second in another. Hundreds of issuescompete for players’ attention every day. Eachplayer is forced to fix upon his issues for thatday, fight them on their own terms, and rush onto the next. Thus the character of emerging is-sues and the paoe at which the game is playedconverge to yield government “decisions” and“actions” as collages. Choices by one player,outcomes of minor games, outcomes of centralgames, and “foul-ups”-these pieces, when stuckto the same canvas, constitute government be-havior relevant to an issue.

The concept of national security policy as po-litical outcome contradicts both public imageryand academic orthodoxy. Issues vital to nationalsecurity, it is said, are too important to be set-tled by political games. They must be “above”politics. To accuse someone of “playing politicswith nationa,l security” is a most serious charge.What public conviction demands, the academicpenchant for intellectual elegance reinforces. In-ternal politics is messy; moreover, according toprevailing doctrine, politicking lacks intellectu.alcontent. As such, it constitutes gossip for jour-nalists rather than a subject for serious investi-gation. Occasional memoirs, anecdotes in his-torical accounts, and several det’ailed case stud-ies to the contrary, most of the literature of for-eign policy avoids bureaucratic politics. The gapbetween academic literature and the experienceof participants in government is nowhere widerthan at this point.

BUREAUCRATIC POLITICS PARADIGM 7s

I. Basic Unit of Analysisa: Policy asPolitical Outcome

The decisions and actions of governments areessentially intra-national political outcomes :

“This paradigm relies upon the small group ofanalysts who have begun to fill the gap. 1My pri-mary source is the model implicit in the work ofRichard E. Neustadt, though his concentration onpresidential action has been generalized to a con-cern with policy as the outcome of political bar-gaining among a number of independent players,the President amounting to no more than a “su-perpower” among many lesser but considerablepowers. As Warner &hilling argues, the substantiveproblems are of such inordinate difficulty that un-certainties and differences with regard to goals, al-ternatives, and consequences are inevitable. Thisnecessitates what Roger Hilsman describes as theprocess of conflict and consensus building. Thetechniques employed in this process often resem-ble those used in legislative assemblies, though

outcomes in the sense that what. happens is notchosen as a solution to a problem but rather re-sults from compromise, coalition, competition,and confusion among government officials whosee different faces of an issue; political in thesense that the activity from which the outcomesemerge is best characterized as bargaining. Fol-lowing Wittgenstein’s use of the concept of a“game ,” national behavior in international af-fairs can be conceived as outcomes of intricateand subtle, simultaneous, overlapping gamesamong players located in positions, the hierar-chical arrangement of which constitutes thegovernment. 79 These games proceed neither atrandom nor at leisure. Regular channels struc-ture the game. Deadlines force issues to the at-tention of busy players. The moves in the chessgame are thus to be explained in terms of thebargaining among players with separate and un-equal power over particular pieces and with sep-arable objectives in distinguishable subgames.

II. Organizing ConceptsA. Players in Positions. The act,or is neither

a unitary nation, nor a conglomerate of organi-zations, but rather a number of individual play-

Samuel Huntington’s characterization of the pro-cess as “legi&tive” overemphasizes the equalityof participants as opposed to the hierarchy whichstructures the game. Moreover, whereas for Hunt-ington, foreign policy (in contrast to military pol-icy) is set by the executive, this paradigm main-tains that the activities which he describes as leg-islative are characteristic of the process by whichforeign policy is made.

“The theatrical metaphor of stage, roles, andactors is more common tha.n this metaphor ofgames, positions, and players. Nevertheless, therigidity connotated by the concept of “role” bothin the theatrical sense of actors reciting fixed linesand in the sociological sense of fixed responses tospecified social situations makes the concept ofgames, positions, and players more useful for thisanalysis of active participants in the determinationof national policy. Objections to the terminologyon the grounds that “game” connotes non-seriousplay overlook the concept’s application to mostserious problems both in Wittgenstein’s philosophyand in contemporary game theory. Game theorytypically treats more precisely structured games,but Wittgenstein’s examination of the “languagegame” wherein men use words to communicate is .quite analogous to this analysis of the less speci-fied game of bureaucratic politics, See LudwigWittgenstein, Phih~phknl Investigations, andThomas &helling, “What is Game Theory?” inJames Charlesworth, Contemporary PoliticalIAnal&.


ers. Groups of these players constitute the agentfor particular government decisions and actions.Players are men in jobs.

Individuals become players in the national se-curity policy game by occupying a critical posi-tion in an administration. For example, in theU.S. government the players include “Chiefs”:the President, Secretaries of State, Defense, andTreasury, Director of the CL4, Joint Chiefs ofStaff, and, since 1961, the Special Assistant forNational Security AfYairs ; 8O “Staffers” : theimmediate staff of each Chief; “Indians”: thepolitical appointees and permanent governmentofficials within each of the departments andagencies ; and “Ad Hoc Players”: actors in thewider government game (especially “Congrcs-sional Influentials”), members of the press,spokesmen for important interest groups (espe-cially the “bipartisan foreign policy esta.blish-men? in and out of Congress), and surrogatesfor each of these groups. Other members ofthe Congress, press, interest groups, and publicform concentric circles around the central arena-circles which demarcate the permissive limitswithin which the game is played.

Positions define what players both may andmust do. The advantages and handicaps withwhich each player can enter and play in variousgames stems from his position. So does a clusterof obligations for the performance of certaintasks. The two sides of this coin are illustratedby the position of the modern Secretary ofState. First, in form and usually in fact, he isthe primary repository of political judgment onthe political-military issues that are the stuff ofcontemporary foreign policy; consequently, he isa senior personal advisor to the President. Scc-ond, he is the colleague of the President’s othersenior advisers on the problems of foreign pol-icy, the Secretaries of Defense and Treasury,and the Special Assistant for National SecurityAffairs. Third, he is the ranking U.S. diplomatfor serious negotiation. Fourt,h, he serves as anAdministration voice to Congress, the country,and the world. Finally, he is “Mr. State Depart-ment” or “Mr. Foreign Office,” “leader of

8o Inclusion of the President’s Special Assistantfor National Security Affairs in the tier of ‘Chiefs”rather than among the ‘Staffers” involves a de-batable choice. In fact he is both super-stafferand near-chief. His position has no statutory au-thority. He is especially dependent upon goodrelations with the President and the Secretaries ofDefense and State. Nevertheless, he stands astridea genuine action-channel. The decision to includethis position among the Chiefs reflects my judg-ment that the Bundy function is becoming institu-tionalized.

officials, spokesman for their causes, guardian oftheir interests, judge of their disputes, superin-tendent of their work, master of their careers?But he is not first one, and then the other. Allof these obligations are his simultaneously. Hisperformance in one affects his credit and powerin the others. The perspective stemming fromthe daily work which he must oversee-thecable traflic by which his department maintainsrelations with other foreign offices-conflictswith the President’s requirement that he serveas a generalist and coordinator of contrastingperspectives. The necessity that he be close tothe President restricts the extent to which, andthe force with which! he can front for his de-pbrtment. When he defers to the Secretary ofDefense rather than fighting for his depart-ment’s position- as he often must-he strainst’he loyalty of his officialdom. The Secretary’sresolution of these conflicts depends not onlyupon the position, but also upon the player whooccupies the position.

For players are also people. Men’s metabo-lisms differ. The core of the bureaucratic politicsmix is personality. How each man manages tostand the heat in his kitchen, each player’s basicoperating style, and the complementarity orcontra.diction among personalities and styles inthe inner circles are irreducible pieces of the pol-icy blend. Moreover, each person comes to hisposition with baggage in tow, including sensitivi-ties to certain issues, commitments to variousprograms, and personal standing and debts withgroups in the society.

B. Parochial Priorities, Perceptions and Is-sues. Answers to the questions: “What is theissue ?” and “What must be done?” are coloredby the position from which the questions arecconsidered. For the factors which encourage or-ganizational parochialism also influence theplayers who occupy positions on top of (orwithin) these organizations. To motivate mem-bers of his organization, a player must be sensi-tive to the organization’s orientation. The gamesinto which the player can enter and the advan-tages with which he plays enhance these pres-sures. Thus propensities of perception stemmingfrom position permit reliable prediction about aplayer’s stances in many cases. But these pro-pensities are filtered through the baggage whichplayers bring to positions. Sensitivity to boththe pressures and the baggage is thus requiredfor many predictions.

*’ Richard EL. Neustadt, Testimony, United StatesSenate, Committee on Government Operations,Subcommittee on National Security Staffing, Ad-ministration of National Security, March 26, 1963,pp. 82-83.

710 THE AMERICAN POLITICAL SCIENCE REVIEW VOL . 63

C. Interests, Stakes, and Power. Games areplayed to determine outcomes. But outcomesadvance and impede each player’s conception ofthe national interest%, specific programs to whichhe is committed, the welfare of his friends, andhis personal interests. These overlapping inter-ests constitute the stakes for which games areplayed. Each player’s ability to play successfullydepends upon his power. Power, i.e., effective in-fluence on policy outcomes, is an elusive blendof at least three elements: bargaining advan-tages (drawn from formal authority and obliga-tions, institutional backing, constituents, exper-tise, and status), skill and will in using bargain-ing advantages, and other players’ perceptionsof the first two ingredients. Power wisely in-vested yields an enhanced reputation for effec-tiveness. Unsuccessful investment depletes boththe stock of capital and the reputation. Thuseach player must pick the issues on which hecan play with a reasonable probability of suc-cess. But no player’s power is sufficient to guar-antee satisfactory outcomes. Each player’s needsand fears run to many other players. What en-sues is the most intricate and subtle of gamesknown to man.

D. The Problem and the Problems. “Solu-tions” to strategic problems are not derived bydetached analysts focusing coolly on the prob-lem. Instead, deadlines and events raise issues ingames, and demand decisions of busy players incontexts that influence the face the issue wears.The problems for the players are both narrowerand broader than the strategic problem. Foreach player focuses not on the total strategicproblem but rather on the decision that must bemade now. But each decision has critical conse-quences not only for the strategic problem butfor each player’s organizational, reputational,and personal stakes. Thus the gap between theproblems the player was solving and the prob-lem upon which the analyst focuses is often verywide.

E. Action-Channels. Bargaining games do notproceed randomly. Action-channels, i.e., regular-ized ways of producing act,ion concerning typesof issues, structure the game by pre-selectingthe major players, determining their points ofentrance into the game, and distributing partic-ular advantages and disadvantages for eachgame. Most critically, channels determine“who’s got the action,” that is, which depart-ment’s Indians actually do whatever is chosen.Weapon procurement decisions are made withinthe annual budgeting process; embassies’ de-mands for action cables are answered accordingto routines of consultation and clearance fromState to Defense and White House; requests forinstructions from militarv groups (concerning

assistance all the time, concerning operationsduring war) are composed by the military inconsultation with the Office of the Secretary ofDefense, State, and White House; crisis re-sponses are debated among White House, State,Defense, CIA, and Ad Hoc players; major polit-ical speeches, especially by the President butalso by other Chiefs, are cleared through estab-lished channels.

F. Action as Politics. Government decisionsare made and government actions emerge nei-ther as the calculated choice of a unified group,nor as a formal summary of leaders’ preferences.Rather the context of shared power but separatejudgments concerning important choices, deter-mines that politics is the mechanism of choice.Note the environment in which the game isplayed : inordinate uncertainty about what mustbe done, the necessity that something be done,and crucial consequences of whatever is done.These features force responsible men to becomeactive players. The pace of the game-hun-dreds of issues, numerous games, and multiplechannels-compels players to fight to “get oth-er’s attention,” to make them “see the facts,” toassure that they “take the time to think seri-ously about the broader issue.” The structure ofthe game-power shared by individuals withseparate responsibilities-validates each player’sfeeling that “others don’t see my problem,” and“others must be persuaded to look at the issuefrom a less parochial perspective.” The rules ofthe game-he who hesitates loses his chance toplay at that point, and he who is uncertainabout his recommendation is overpowered byothers who are sure-pressures players to comedown on one side of a 51-49 issue and play. Therewards of the game-effectiveness, i.e., impacton outcomes, as the immediate measure of per-formance-encourages hard play. Thus, mostplayers come to fight to “make the governmentdo what is right.” The strategies and tactics em-ployed are quite similar to those formalized bytheorists of international relations.

G. Streams of Outcomes. Important gov-ernment decisions or actions emerge as collagescomposed of individual acts, outcomes of minorand major games, and foul-ups. Outcomeswhich could never have been chosen by an actorand would never have emerged from bargainingin a single game over the issue are fabricatedpiece by piece. Understanding of the outcome.requires that it be disaggregated.

III. Dominant Inference PatternIf a nation performed an action, that action

was the outcome of bargaining among individu-als and groups within the government. Thatoutcome included results achieved by groups


commit&ted to a decision or action, rem&antswhich emerged from bargaining among groupswith quite different positions and foul-ups.Model III’s explanatory power is achieved byrevealing the pulling and hauling of variousplayers, with different perceptions and priorities,focusing on separate problems, which yielded theoutcomes that constitute the action in question.

IV. General Propositions1. Action and Intention. Action does not pre-

suppose intention. The sum of behavior ofrepresentatives of a government relevant to anissue was rarely intended by any individual orgroup. Rather separate individuals with differentintentions contributed pieces which compose anoutcome distinct from what anyone would havechosen.

2. Where you stand depends on where y o usit? Horizontally, the diverse demands uponeach player shape his priorities, perceptions, andissues. For large classes of issues, e.g., budgetsand procurement decisions, the stance of a par-ticular player can be predicted with high reli-ability from information concerning his seat. Inthe notorious B-36 controversy, no one was sur-prised by Admiral Radford’s testimony that‘(the B-36 under any theory of war, is a badgamble with national security,” as opposed toAir Force Secretary Symington’s claim that “aB-36 with an A-bomb can destroy distant objec-tives which might require ground armies yearsto take.“83

3. Chiefs and Indians. The aphorism “whereyou stand depends on where you sit” has verti-cal as well as horizontal application. Vertically,the demands upon the President, Chiefs, Staf-fers, and Indians are quite distinct.

The foreign policy issues with which thePresident can deal are limited primarily by hiscrowded schedule: the necessity of dealing firstwith what comes next. His problem is to probethe special face worn by issues that come to hisattention, t.0 preserve his leeway until time hascla,rified the uncertainties, and to assess the rele-vant risks.

Foreign policy Chiefs deal most often withthe hottest issue de jour, though they can getthe attention of the President and other mem-bers of the government for other issues whichthey judge important. What they cannot guar-antee is that “the President will pay the price”or t.hat “the others will get on board.” They

82 This aphorism was stated first, I think, by DonK. Price.

83 Paul Y. Hammond, “Super Carriers and B-36Bombers,” in Harold Stein (ed.), American Civil-Mililary Decisions (Birmingham, 1963).

must build a coalition of the relevant powersthat be. They must “give the President confi-dence” in the right course of action.

Most problems are framed, alternatives speci-fied, and proposals pushed, however, by Indians.Indians fight with Indians of other depart-ments ; for example, struggles between Interna-tional Security Affairs of the Department of De-fense and Political-Military of the State Depart-ment are a microcosm of the action at higherlevels. But the Indian’s major problem is how toget the attention of Chiefs, how to get an issuedecided, how to get the government “to do whatis right .”

In policy making then, the issue looking downis options : how to preserve my leeway untiltime clarifies uncertainties. The issue lookingsideways is commitment: how to get otherscommitted to my coalition. The issue lookingupwards is confidence: how to give the bossconfidence in doing what must be done. To par-aphrase one of Neustadt’s assertions which canbe applied down the length of the ladder, the es-sence of a responsible official’s task is to induceothers to see that what, needs to be done is whattheir own appraisal of their own responsibilitiesrequires them to do in their own interests.

V. Specific Propositions1. Deterrence. The probability of nuclear at-

tack depends primarily on the probability ofattack emerging as an outcome of the bureau-cratic politics of the attacking government. First),which players can decide to launch an attack?Whether the effective power over action is con-trolled by an individual, a minor game, or thecentral game is critical. Second, though ModelI’s confidence in nuclear deterrence stemsfrom an assertion that, in the end, govern-ments will not commit suicide, Model III re-calls historical precedents. Admiral Yamamoto,who designed the Japanese attack on Pearl Har-bor, estimated accurately: “In the first sixmonths to a year of war against the U.S. andEngland I will run wild, and I will show youan uninterrupted succession of victories; I rnustalso tell you that, should the war be prolongedfor two or three years, I have no confidence inour ultimate victory.“84 But Japan attacked.Thus, three questions might be considered. One:could any member of the government solve hisproblem by attack? What patterns of bargainingcould yield attack as an outcome? The major -difference between a stable balance of terror anda questionable balance may simply be that int,he first case most members of the government

” Roberta Wohlstctter, Pearl Harbor (Stanford,1962), p. 350.

T.


appreciate fully the consequences of attack andare thus on guard against the emergence of thisoutcome. Two: what strea,m of outcomes mightlead to an attack? At what point in that streamis the potential attacker’s politics? If members ofthe U.S. government had been sensitive to thestream of decisions from which the Japanese at-tack on Pearl Harbor emerged, they would havebeen aware of a considerable probability of thatattack. Three: how might. miscalculation andconfusion generate foul-ups that yield attack asan outcome? For example, in a crisis or afterthe beginning of conventional war, what hap-pens to the information available to, and the ef-fective power of, members of the central game.

THE U.S. BLOCKADE OF CUBA: A THIRD CUT

The Politics of Discovery. A series of over-lapping bargaining games determined both thedate of the discovery of the Soviet missiles andthe impact of this discovery on the Administra-tion. An explanation of the politics of the dis-covery is consequently a considerable piece ofthe explanation of the U.S. blockade.

Cuba was the Kennedy Administration’s “po-litical Achilles’ heel.“85 The months precedingthe crisis were also months before the Congres-sional elections, and the Republican Senatorialand Congressional Campaign Committee hadannounced that Cuba would be “the dominantissue of the 1962 campaign.“86 What the admin-istration billed as a “more positive and indirectapproach of isolating Castro from developing,democratic Latin America,” Senators Keating,Goldwater, Capehart, Thurmond, and others at-tacked as a “do-nothing” policy.87 In statementson the floor of the House and Senate, campaignspeeches across the country, and interviews andarticles carried by national news media, Cuba-particularly the Soviet program of increasedarms aid-served as a stick for stirring thedomestic political scene.88

These attacks drew blood. Prudence de-manded a vigorous reaction. The President de-cided to meet the issue head-on. The Adminis-tration mounted a forceful campaign of denialdesigned to discredit critics’ claims. The Presi-dent himself manned the front hne of this offen-sive, though almost all Administration officialsparticipated. In his news conference on August19, President Kennedy attacked as “irresponsi-ble” calls for an invasion of Cuba, stressingrather “the totality of our obligations” andpromising t,o “watch what happens in Cuba with

85 Sorensen, Kennedy, p. 670.” Tbid.s7 Ibid., pp. 67Off.” Nczo York Times, August, Sept,embcr, 1962.

the closest attention.“8g On September 4, h eissued a strong statement denying any provoca-tive Soviet action in Cuba.go On September 13he lashed out at “loose talk” calling for an inva-sion of Cuba .gl The day before the flight of theU-2 which discovered the missiles, he cam-paigned in Capehart’s Indiana against those“self-appointed generals and admirals who wantto send someone else’s sons to war.“gz

On Sunday, October 14, just as a U-2 wastaking the first pictures of Soviet missiles,McGeorge Bundy was asserting :

I know that there is no present evidence, and Ithink that there is no present likelihood that theCuban government and the Soviet governmentwould, in combination, attempt to install a majoroffensive capabilityY3

In this campaign to puncture the critics’charges, the Administration discovered that thepublic needed positive slogans. Thus, Kennedyfell into a tenuous semantic distinction between“offensive” and “defensive” weapons. This dis-tinction originated in his September 4 staternentthat there was no evidence of “offensive groundto ground missiles” and warned “were it to beotherwise, the gravest issues would arise.“g4 HisSeptember 13 statement turned on this distinc-tion between “defensive” and “offensive” weap-ons and announced a firm commitment to actionif the Soviet Union attempted to introduce thelatter into Cuba .g5 Congressional. committees elic-ited from administration officials testimonywhich read this distinction and the President’scommitment into the Congressional Record.96

What the President least wanted to hear, theCIA was most hesitant to say plainly. On Au-gust 22 John McCone met privately with thePresident and voiced suspicions that the Sovietswere preparing to introduce offensive missilesinto Cuba.g7 Kennedy heard this as what itwas: t,he suspicion of a hawk. McCone left

” New YorFc Times, August 20,1962.w New York Times, September 5,1962.O1 New Yodb Times, September 14, 1962.O2 New York Times, October 14,1962.O3 Cited by Abel, 0~. cit., p. 13.O4 New York Times, September 5,1962.OS New York Times, September 14, 1962.~6 Senate Foreign Relations Committee ; Senate

Armed Services Committee; House Committee onAppropriation; House Select Committee on ExportControl.

“I Abel, op. cit., pp. 17-18. According to McCone,he told Kennedy, “The only construction I can puton the material going into Cuba is that the Rus-sians are preparing to introduce offensive missiles.”See also Weintal and Bartlett, op. cit., pp. 60-61.


Washington for a month’s honeymoon on the Riv-iera. Fretting at Cap Ferrat, he bombarded hisdeputy, General Marshall Carter, with tele-grams, but Carter, knowing that McCone hadinformed the President of his suspicions and re-ceived a cold reception, was reluctant to distrib-ute these telegrams outside the CIA?8 On Sep-tember 9 a U-2 “on loan” to the Chinese Na-tionalists was downed over mainland China.ggThe Committee on Overhead Reconnaissance(COMOR) convened on September 10 witha sense of urgency.loo Loss of another U-2 mightincite world opinion to demand cancellation ofU-2 flights. The President’s campaign againstthose who asserted that the Soviets were actingprovocatively in Cuba had begun. To risk down-ing a U-2 over Cuba was to risk chopping offthe limb on which the President was sitting.That meeting decided to shy away from thewestern end of Cuba (where SAMs were becom-ing operational) and modify the flight pattern ofthe U-2s in order to reduce the probability thata U-2 would be lost Jo1 USIB’s unanimous ap-proval of the September estimate reflects similarsensitivities. On September 13 the Presidenthad asserted that there were no Soviet offensivemissiles in Cuba and committed his Administra-tion to act if offensive missiles were discovered.Before Congressional committees, Administra-tion officials were denying that there was anyevidence whatever of offensive missiles in Cuba.The imp1ication.s of a National Intelligence esti-mate which concluded that the Soviets were in-troducing offensive missiles into Cuba were notlost on the men who constituted America’s high-est intelligence assembly.

The October 4 COMOR decision to directa flight over the western end of Cuba in ef-fect “overturned” the September estimate, butwithout officially raising that issue. The decisionrepresented McCone’s victory for which he hadlobbied with the President before the September10 decision, in telegrams before the September19 estimate, and in person after his return toWashington. Though the politics of the intelli-gence community is closely guarded, severalpieces of the story can be to1d.1°2 By September27, Colonel Wright and others in DIA believedthat the Soviet Union was placing missiles in the

* Abel, op. cit., p. 23.~9 New Yorlc Times, September 10,1962.loo See Abel, op. cit., pp. 25-26; and Hilsman, op.

cit., p. 174.lo* Department of Defense Appropriation, Hear-

ings, 69.lo3 A basic, but somewhat contradictory, account

of parts of this story emerges in the Department ofDefense Appropriations, IleviTtgs, l-70.

San Cristobal area .lo3 This area was markedsuspicious by the CIA on September 29 and cer-tified top priority on October 3. By October4 McCone had the evidence required to raise theissue officially. The members of COMORheard McCone’s argument, but. were reluctantto make the hard decision he demanded. Thesignificant probability that a U-2 would bedowned made overflight of western Cuba a mat-ter of real concern.lo4

The Politics of Issues. The U-2 photographspresented incontrovertible evidence of Soviet of-fensive missiles in Cuba. This revelation fellupon politicized players in a complex context. Asone high official recalled, Khrushchev hadcaught us “with our pants down.” What each ofthe central participants saw, and what eachdid to cover both his own and the Administra-tion’s nakedness, created the spectrum of issuesand answers.

At approximatelv 9 :00 A.M., Tuesday morn-ing, October 16, l&George Bundy went to thePresident’s living quarters with the message:“Mr. President, there is now hard photographicevidence that the Russians have offensive mis-siles in Cuba.“lo5 Much has been made of Ken-nedy’s “expression of surprise,“lO6 but “surprise”fails to capture the character of his initial reac-tion. Rather, it was one of startled anger, mostadequately conveyed by the exclamation: “Hecan’t do that to me !“lo7 In terms of the Presi-dent’s attention and priorities at that moment,Khrushchev had chosen the most unhelpful actof all. Kennedy had staked his full Presidentialauthority on the assertion that the Sovietswould not, place offensive weapons in Cuba.Moreover, Khrushchev had assured the Presi-dent through the most direct and personal chan-nels that he was aware of the President’s domes-tic political problem and that nothing would bedone to exacerbate this problem. The Chairmanhad Lied to the President. Kennedy’s initial reac-tion entailed action. The missiles must beremoved.lOS The alternatives of “doing nothing”or “taking a diplomatic approalch” could nothave been less relevant to his problem.

These two tracks-doing nothing and taking

lo3 Department of Defense Appropriations, Hear-ings, 71.

lo4 The details of the 10 days between the October4 decision and the October 14 flight must be heldin abeyance.

lo5 Abel, op. cit., p. 4.lo6 Ibid., pp. 44.8.lo7 See Richard Neustadt, “Afterword,” Presiden-

tial Power (New York, 1964).lofi Sorensen, Kennedy, p. 676; Schlesinger, 013. cit.,

p. 801.


a diplomatic approach-were the solutions advo-cated by two of his principal advisors. For Sec-retary of Defense McNamara, the missilesraised the spectre of nuclear war. He firstframed the issue as a straightforward strategicproblem. To understand the issue, one had tograsp two obvious but difficult points. First, themissiles represented an inevitable occurrence:narrowing of the missile gap. It simply hap-pened sooner rather than later. Second, theUnit-ed States could accept this occurrence sinceits consequences were minor: “seven-to-one mis-sile ‘superiority,’ one-to-one missile ‘equality,’one-to-seven missile ‘inferiority’-the three pos-tures are identical.” McNamara’s st.atement ofthis argument at the first meeting of the ExComwas summed up in the phrase, “a missile is a&sile."l09 cJIt makes no great difference,” hemaintained, “whether you aIre killed by a missilefrom the Soviet Union or Cuba.“l10 The impli-cation was clear. The United States should notinitiate a crisis with the Soviet Union, risking asignificant probability of nuclear war over anoccurrence which had such small strategic impli-cations.

The perceptions of McGeorge Bundy, thePresident’s Assistant for National Security Af-fairs, are the most difficult of all to reconstruct.There is no question that he initially argued fora diplomatic track?I1 But was Bundy laboringunder his ,acknowledged burden of responsibilityin Cuba I? Or was he playing the role of dev-il’s advocate in order to make the Presidentprobe his own initial reaction and consider otheroptions ?

The President’s brother, Robert Kennedy, sawmost clearly the political wall against whichKhrushchev had backed the President. But he,like McNamara, saw the prospect of nucleardoom. Was Khrushchev going to force the Presi-dent to an insane act? At the first meeting ofthe ExCom, he scribbled a note, “Now I knowhow Tojo felt when he was planning Pearl Har-bor .“l12 From the outset he searched for an al-ternative that would prevent the air strike.

The initial reaction of Theodore Sorensen, thePresident’s Special Counsel and “alter ego,” fellsomewhere between that of the President andhis brother. Like the President, Sorensen felt thepoignancy of betrayal. If the President had beenthe architect of the policy which the missilespunctured, Sorensen was the draftsman.Khrushchev’s deceitful move demanded a strong

lo9 Hilsman, op. cit., p. 195.‘lo Ibid.‘11 JVeintal and Bartlett, op. cit., p. 67 ; Abel,

op. cit., p. 53.‘12 Schlesinger, op. cit., p. 803.

counter-move. But like Robert Kennedy, Soren-sen feared lest the shock and disgrace lead to dis-aster.

To the Joint Chiefs of Staff the issue wasclear. Now was the time to do the job for whichthey had prepared contingency plans. Cuba Ihad been badly done; Cuba II would not be.The missiles provided the occasion to deal withthe issue: cleansing the Western Hemisphere ofCastro’s Communism. As the President recalledon the day the crisis ended, “An invasion wouldhave been a mistake-a wrong use of our power.But the military are mad. They wanted to dothis. It’s lucky for us that we have McNamaraover there.“113

McCone’s perceptions flowed from his con-firmed prediction. As the Cassandra of the inci-dent, he argued forcefully that the Soviets hadinstalled the missiles in a daring political probewhich the United States must meet with force.The time for an air strike was now.114

The Politics of Choice. The process by whichthe blockade emerged is a story of the most sub-tle and intricate probing, pulling, and hauling;leading, guiding, and spurring. Reconstruction ofthis process can only be tentative. Initially thePresident and most of his advisers wanted theclean, surgical air strike. On the first day ofthe crisis, when informing Stevenson of the mis-siles, the President mentioned only two alterna-tives : “I suppose the alternatives are to go inby air and wipe them out, or to take other stepsto render them inoperable.“ll5 At the end of theweek a sizeable minority still favored an airstrike. As Robert Kennedy recalled: “The four-teen people involved were very significant. . . .If six of them had been President of the U.S., Ithink that the world might have been blownUP* "116 What prevented the air strike was a for-tuitous coincidence of a number of factors-theabsence of any one of which might have permit-ted that option to prevail.

First, McNamara’s vision of holoca.ust sethim firmly against the air strike. His initial at-tempt to frame the issue in strategic termsstruck Kennedy as particularly inappropriate.Once McNamara realized that the name of thegame was a strong response, however, he and hisdeputy Gilpatric chose the blockade as a fall-back. When the Secretary of Defense-whosedepartment had the action, whose reputation inthe Cabinet was unequaled, in whom t,he Presi-dent demonstrated full confidence-marshalled ’

‘I3 Ibid., p. 831.11’ Abel, op. tit ., p. 186.‘15 Ibid., p. 49.11’ Interview, quoted by Ronald Steel, hTew York

Review of Books, March 13,1969, p. 22.


the arguments for the blockade and refused tobe moved, the blockade became a formidable al-ternative.

Second, Robert Kennedy--the President’sclosest confidant-was unwilling to see hisbrother become a rrTojo.” His arguments againstthe air strike on moral grounds struck a chordin the President. Moreover, once his brother hadstated these arguments so forcefully, the Presi-dent could not have chosen his initially pre-ferred course without, in effect, agreeing to be-come what RFK had condemned.

The President learned of the missiles onTuesday morning. On Wednesday morning, inorder to mask our discovery from the Russians,the President flew to Connecticut to keep acampaign commitment, leaving RFK as the un-official chairman of the group. By the time thePresident. returned on Wednesday evening, acritical third piece had been added to the pic-ture. McNamara had presented his argumentfor the blockade. Robert Kennedy and Sorensenhad joined McNamara. A powerful coalition ofthe advisers in whom the President had thegreatest confidence, and with whom his st,ylewas most compatible, had emerged.

Fourth, the coalition that had formed behindthe President’s initial preference gave him rea-son to pause. who supported the air striketheChiefs, McCone, Rusk, Nitze, and Acheson-asmuch as how they supported it, counted. Fifth,a piece of inaccurate information, which no oneprobed, permitted the blockade advocates tofuel (potential) uncertainties in the President’smind. When the President returned to Washing-ton Wednesday evening, RFK and Sorensen methim at the airport. Sorensen gave the Presidenta four-page memorandum outlining the areas ofagreement and disagreement. The strongest ar-gument was that the air st,rike simply could notbe surgical .I11 After a day of prodding andquestioning, the Air Force had asserted that itcould not guarantee the success of a surgical airstrike limited to the missiles alone.

Thursday evening, the President convenedthe ExCom at the White House. He declared histentative choice of the blockade and directedthat preparations be made to put it into effectby Monday morning. 118 Though he raised aquestion about the possibility of a surgical airstrike subsequently, he seems to have acceptedthe experts’ opinion that this was no liveoption.llg (Acceptance of this estimate suggeststhat he mav have learned the lesson of the Bayof Pigs-“Never rely on experts”-less well than

11’ Sorensen, Kennedy, p. 686.‘= Ibid., p. 691.‘lg Ibid., pp. 6914392.

he supposed.) 12* But this information was incor-rect. That no one probed this estimate duringthe first week of the crisis poses an interestingquestion for furt,her investigation.

A coalition, including the President, thusemerged from the President’s initial decisionthat something had to be done; McNamara, Rob-ert Kennedy, and Sorensen’s resistance to the airstrike; incompatibility between the Presidentand the air strike advocates; and an inaccuratepiece of information.121

CONCLUSIONThis essay has obviously bitten off more than

it has chewed. For further developments andsynthesis of these arguments the reader is re-ferred to the larger study.122 In spite of the lim-its of space, however, it would be inappropriateto stop without spelling out several implicationsof the argument and addressing the question ofrelations among the models and extensions ofthem to activity beyond explanation.

At a minimum, the intended implications ofthe argument presented here are four. First, for-mulation of alternative frames of reference anddemonstration that different analysts, relyingpredominantly on different models, producequite different explanations should encouragethe analyst’s self-consciousness about the nets heemploys. The effect of these “spectacles” in sen-sitizing him to particular aspects of what isgoing on- framing the puzzle in one way ratherthan another, encouraging him to examine theproblem in terms of certain categories ratherthan others, directing him to particul.ar kinds ofevidence, and relieving puzzlement by oneprocedure rather than another-must be recog-nized and explored.

Second, the argument implies a position onthe problem of “the state of the art.” While ac-cepting the commonplace characterization of thepresent condition of foreign policy analysis-personalistic, non-cumulative, and sometimes in-sightful-this essay rejects both the counsel ofdespair’s justification of this condition as aconsequence of the character of the enterprise,and the “new frontiersmen’s” demand for CIptio~i theorizing on the frontiers and ad hoc ap-propriation of “new techniques.“123 What is re-

120 Schlesinger, op. cit., p. 296.12’ Space will not permit an account of the path

from this coalition to the formal government de- ’cision on Saturday and action on Monday.

‘2~ Bureaucracy and PO&Z/ (forthcoming, 1969).123 Thus my position is quite distinct from both

poles in the recent “great debate” about interna-tional relations. While many “traditionalists” ofthe sort Kaplan attacks adopt the first posture and

716 TIIE AMERICAN POLITICAL SCIEKCE HEVIEW V O L. 6 3

quired as a first step is non-casual examinationof the present product: inspection of existing ex-planations, articulation of the conceptual modelsemployed in producing them, formulation of thepropositions relied upon, specification of thelogic of the various intellectual enterprises, andreflection on the questions being asked. Thoughit is difficult to overemphasize the need for moresystematic processing of more data, these pre-liminarv matters of formulating questions withclarity and sensitivity to categories and assump-tions so that fruitful acquisition of large quan-tities of data is possible are still a major hurdle inconsidering most important problems.

Third, the preliminary, partial paradigms pre-sented here provide a basis for serious reexami-nation of many problems of foreign and militarypolicy. Model II and Model III cuts at problemstypically treated in Model I terms can permitsignificant improvements in explanation andprediction .124 Full Model II and III analyses re-quire large amounts of information. But even incases where the information base is severely lim-ited, improvements are possible. Consider theproblem of predicting Soviet strategic forces. Inthe mid-1950s, Model I style calculations led topredictions that the Soviets would rapidly de-ploy large numbers of long-range bombers.From a Model II perspective, both the frailty ofthe Air Force within the Soviet military estab-lishment and the budgetary implications of sucha buildup, would have led analysts to hedge thisprediction. Moreover, lMode1 II would havepointed to a sure, visible indicator of such abuildup : noisy struggles among the Servicesover major budgetary shifts. In the late 1950sand early 196Os, Model I calculations led to theprediction of immediate, massive Soviet deploy-ment of ICBMs. Again, a Model II cut wouldhave reduced this number because, in the earlierperiod, strategic rockets were controlled by theSoviet Ground Forces rather than an indepen-dent Service, and in the later period, this wouldhave necessitated massive shifts in budgetary

many “scientists” of the sort attacked by Bulladopt the second, this third posture is relativelyneutral with respect to whatever is in substantivedispute. See Redly Bull, “International Theory :The Case for a Classical Approach,” World Politics(April, 1966) ; and Morton Kaplan, “The NewGreat Debate: Traditionalism vs. Science in In-ternational Relations,” World Politics (October,1966).

I’* A number of problems are now being examinedin these terms both in the Bureaucracy StudyGroup on Bureaucracy and Policy of the Instituteof Politics at Harvard University and at the RandCorporation.

splits. Today, Model I considerations lead manyanalysts both to recommend that an agreementnot to deploy ABMs be a major American ob-jective in upcoming strategic negotiations withthe USSR, and to predict success. From aModel II vantage point, the existence of an on-going Soviet ABM program, the strength of theorganization (National Air Defense) that con-trols ABMs, and the fact that an agreement tostop ABM deployment would force the virtualdismantling of this organization, make a viableagreement of this sort much less likely. A ModelIII cut suggests that (a) there must be signifi-cant differences among perceptions and prioritiesof Soviet leaders over strategic negotiations, (b)any agreement will affect some players’ powerbases, and (c) agreements that do not. requireextensive cuts in the sources of some majorplayers’ power will prove easier to negotiate andmore viable.

Fourth, the present formulation of paradigmsis simply an initial step. As such it leaves a longlist of critical questions unanswered. Given anyaction, an imaginative analyst should always beable to construct some rationale for the govern-ment’s choice. By imposing, and relaxing, con-straints on the parameters of rational choice (asin variants of Model I) analysts can construct alarge number of accounts of any act as a ra-tional choice. But does a statement of reasonswhy a rational actor would choose an actionconstitute an explanation of the occurrence ofthat action? How can Model I analysis beforced to make more systematic contributions tothe question of the determinants of occurrences?Model II’s explanation of t in terms of t - 1 isexplanation. The world is contiguous. But gov-ernments sometimes make sharp departures.Can an organizational process model be modifiedto suggest where change is likely? Attention toorganizational change should afford greater un-derstanding of why particular programs andSOPS are maintained by identifiable types oforganizations and also how a manager can im-prove organizational performance. Model IIItells a fascinating “story.” But its complexity isenormous, the information requirements areoften overwhelming, and many of the details ofthe bargaining may be superfluous. How cansuch a model be made parsimonious? Thethree models are obviously not exclusive alter-natives. Indeed, the paradigms highlight thepartial emphasis of the framework-what eachemphasizes and what it leaves out. Each concen-trates on one class of variables, in effect, rele-gating other important factors to a ceteris parn-bus clause. Model I concentrat,es on “market’factors : ” pressures and incentives created byt,he “international strategic marketplace.” Mod-


els II and III focus on the internal mechanismof the government that chooses in this environ-ment. But ca,n these relations be more fullyspecified? Adequate synthesis would require atypology of decisions and actions, some of whichare more amenable to treatment in terms of onemodel and some to another. Government behav-ior is but one cluster of factors relevant to oc-currences in foreign affairs. Most students offoreign policy adopt this focus (at least whenexplaining and predicting). Nevertheless, the di-mensions of the chess board, the character ofthe pieces, and the rules of the game-factorsconsidered by international systems theorists-constitute the context in which the pieces aremoved. Can the major variables in the full func-tion of determinants of foreign policy outcomesbe identified?

Both the outline of a partial, ad hoc workingsynthesis of the models, and a sketch of theiruses in activities other tha,n explanation can besuggested by generating predictions in terms ofeach. Strategic surrender is an important prob-lem of international relations and diplomatichistory. ‘War termination is a new, developingarea of the strategic literature. Both of these in-terests lead scholars to address a central ques-tion: FVhy do nations surrender when? Whetherimplicit in explanations or more explicit in anal-ysis, diplomatic historians and strategists relyupon propositions which can be turned forwardto produce predictions. Thus at the risk of beingtimely-and in error-the present situation(August, 1968) offers an interesting test case:Why will North Vietnam surrender when?125

In a nutshell, analysis according to Model Iasserts: nations quit when costs outweigh thebenefits. North Vietnam will surrender when sherealizes “that continued fighting can only gener-ate additional costs without hope of compensat-ing gains, this expectation being largely theconsequence of the previous application of forceby the dominant side.“l26 U.S. actions can in-crease or decrease Hanoi’s strategic costs.Bombing North Vietnam increases the pain andthus increases the probability of surrender. Thisproposition and prediction are not withoutmeaning. That-“other things being equal”-na-tions are more likely to surrender when the

*23 In response to several readers’ recommenda-tions, what follows is reproduced verb&n from thepaper delivered at the September, 1968 Associa-tion meetings (Rand P-3919). The discussion isheavily indebted to Ernest R. May.

“’ Richard Snyder, Deterrence and Defense(Princeton, 1961)) p. 11. For a more general presen-tation of this position see Paul Kecskemeti, Stra-tegic Surrender (New York, 1964).

strategic cost-benefit balance is negative, is true.Nations rarely surrender when they are winning.The proposition specifies a range within whichnations surrender. But over this broad range,the relevant question is: why do nations surren-der?

Models II and III focus upon the governmentmachine through which this fact about the in-ternational strategic marketplace must be filteredto produce a surrender. These analysts are con-siderably less stt.nguine about the possibility ofsurrender at the point that the cost-benefit cal-culus turns negative. Never in history (i.e., innone of the five cases I have examined) havenations surrendered at that point. Surrender oc-curs sometime thereafter. When depends on pro-cess of organizations and politics of playerswithin these governments-as they are affectedby the opposing government. Moreover, the ef-fects of the victorious power’s action upon thesurrendering nation cannot be adequately sum-marized as increasing or decreasing strategiccosts. Imposing additional costs by bombing anation may increase the probability of surren-der. But it also may reduce it. An appreciationof the impact of the acts of one nation upon an-other thus requires some understanding of themachine which is being influenced. For moreprecise prediction, Models II and III requireconsiderably more information about the organi-zations and politics of North Vietnam than ispublicly available. On the basis of the limitedpublic informattion, however, the.se models canbe suggestive.

Model II examines two sub-problems. First,to have lost is not sufficient. The governmentmust know that the strat,egic cost-benefit calcu-lus is negative. But neither the categories, northe indicators, of strategic costs and benefits areclear. And the sources of information aboutboth are organizations whose parochial prioritiesand perceptions do not facilitate accurate infor-mation or estimation. Military evaluation ofmilitary performance, military estimates of fac-tors like “enemy morale,” and military predic-tions concerning when “the tide will turn” or“the corner will have been turned” are typicallydistorted. In cases of highly decentralized guer-rilla operations, like Vietnam, these problemsare exacerbated. Thus strategic costs will be un-derestimated. Only highly visibZe costs canhave direct impact on leaders without being fil-tered through organizational channels. Second,since organizations define the details of optlionsand execute actions, surrender (and negotiation)is likely to entail considerable bungling in theearly stages. No organization can define optionsor prepare programs for this treasonous act.Thus, early overtures will be uncoordinated with


the acts of other organizations, e.g., the fightingforces, creating contradictory “signals” to thevictor.

Model III suggests that surrender will notcome at the point that strat’egic costs outweighbenefits, but that it will not wait until the lead-ership group concludes that the war is lost.Rather the problem is better understood interms of four additional propositions. First,strong advocates of the war effort, whose careersare closely identified wit)11 the war, rarely cometo the conclusion that costs outweigh benefits.Second, quite often from the outset of a war, anumber of members of the government (partic-ularly those whose responsibilities sensitize themto problems other than war, e.g., economic plan-ners or intelligence experts) are convinced thatthe war effort is futile. Third, surrender is likely

to come as the result of a political shift thatenhances the effective power of the latter group(and adds swing members to it). Fourth, thecourse of the war, particularly actions of thevictor, can influence the advantages and disad-vantages of players in the loser’s government.Thus, North Vietnam will surrender not whenits leaders have a change of heart, but whenHanoi has a change of leaders (or a change ofeffective power within the central circle). HowU.S. bombing (or pause), threats, promises, oraction in the South affect the game in Hanoi issubtle but nonetheless crucial.

That these three models could be applied tothe surrender of governments other than NorthViet’nam should be obvious. But that exercise isleft for the reader.

Conceptual Models and the Cuban Missile Crisis...Conceptual Models and the Cuban Missile Crisis...

Documents

Transcript of Conceptual Models and the Cuban Missile Crisis...Conceptual Models and the Cuban Missile Crisis...