Foreign Policy Engineering: From Theory to Practice and Back Again Author(s): Philip Zelikow Reviewed work(s): Source:

International Security, Vol. 18, No. 4 (Spring, 1994), pp. 143-171 Published by: The MIT Press Stable URL: http://www.jstor.org/stable/2539180 . Accessed: 30/04/2012 15:15
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to International Security.

http://www.jstor.org

Foreign Policy

Engineering

Philip Zelikow

From Theory to Practiceand Back Again

Decades of modern

scholarshipabout foreignpolicies have actuallyproduced relatively few suggestions about how such policies ought to be constructed.Many volumes discuss the ways in which the process works, frombureaucraticturf battles to the influenceof television. More volumes suggest what should be done about some particularissue, frompreventingmilitary use of outer space to dividingproperty rightsat the bottomof the sea. But few books offer explicit general advice about just how policies should be crafted,divorced from opinions about a particularissue. This is one reason why internationalpolicies are usually not very well understood, even by academic experts.Indeed many diplomats and officials do not themselves fullygrasp the conceptual framework or pivotal choices implicitin their own country's policies. Lacking any common vocabulary about the building blocks of policies, practitioners have difficulty communicating to others what is going on, so they consciously and unconsciously simplifythe policy into broad themes or vivid anecdotes, and then feel frustrated at the gulf in understandingbetween themselves and outsiders, whetheracademics or newspaper editors. For theirpart, outsiders may sense that policies are going well or faring badly,but findit difficult to diagnose the cause of the success or the failures, especially if the mattercannot be reduced to a clash in personal values. The outsiders (and many insiders) usually have one of three responses. They may turnto caricaturesof the people involved: Leaders have "vision" or they are "preoccupied"; an official is "canny" or he is "bookish"; the diplomats are "misguided" or they are "shrewd." Alternately, theymay resortto caricatures of the policies involved, condensing several kinds of challenges to-

Philip Zelikowis AssistantProfessor of Public Policyat Harvard'sKennedySchoolof Government. Formerly a foreign service officer withtheDepartment of State,from 1989 to 1991 he was director for European security affairs on thestaff oftheNationalSecurity Council. The author acknowledges gratefully the insightshe has derived fromdiscussions with Robert Art,RobertBlackwill,Ashton Carter,Ernest May, Mark H. Moore, Richard Neustadt, Condoleezza Rice, Alfred P. Rubin, Richard Zeckhauser, and the students in his fall 1993 class on Politicaland Organizational Analysis. None of these people are, of course, responsible forthe misuse of theirideas.
Internationial Security, Vol. 18, No. 4 (Spring 1994), pp. 143-171 C 1994 by the Presidentand Fellows of Harvard College and the Massachusetts Instituteof Technology.

143

18:4 | 144 Security International

getherinto "underestimationof those revolutionary forces" or "undue reliance on that dictator"or "back-roombargaining." Thirdly,they may write offthe problem as unintellectualor "judgmental," and raise the study of where the vagaries of particular international politicsto a level of abstraction the are policies episodes eclipsed by granderforcesshaping the global "system." This is not good enough. If citizens are to understand the choices being and ifmore ofthe vast storeofknowledge outside made by theirgovernment, and improvepublic policy performance, officials of governmentis to inform that helps them and observers need a relativelysimple, usable framework appraise the analyticalcomponents of a public policy. This essay offerssuch a framework.It can be outlined as follows: First with each other:problem recthinkof three streams, constantlyinteracting ognition, politics (bureaucratic and otherwise), and policy "engineering." The essay concentrateson just one of these streams, that of policy "engineering," defined as the application of knowledge, principles,and methods (including both policy analysis and institutional analysis) to the solution of The "engineering" specificpublic problems in a given politicalenvironment. task has seven parts: national interest,objectives, strategy, design, implementation,maintenance,and review. Particular descriptionsof each of these below. parts, or policy components, are offered This framework is prescriptive.Policies ought to include all these components but sometimes they do not. Yet though there are a few suggestions here and there, the essay does not tryto say how good policies should be which analyticalpropomade. Its framework is just a tool for determining sitions are trulymaterialto the success of a policy. Once the materialpropositionsare forcedinto the open, it becomes easier forinsidersand outsiders alike to see just what kinds of social science, what "if-then" generalizations, are relevant.The framework draw up a checkmightalso help policymakers list of questions they may ask themselves,or others. An example of how social science can be related to concretepolicy propositionsis Alexander George's new book, Bridging theGap: Theory and Practice in Foreign the applicationof general knowlPolicy.'George's work illustrates
1. Alexander L. George, Bridging theGap: Theory and Practice in Foreign Policy(Washington,D.C.: U.S. Institute of Peace Press, 1993), p. 108. From the timehe worked forthe RAND Corporation lifeon the frontiers between political duringthe 1950s, George has spent most of his intellectual analysis and the use of historyin focused case studies, psychology,and practical advice for policymakers,usually about the efficacy of using or threatening to use force.He long directed

Foreign PolicyEngineering 1145

edge to one of the seven policy engineeringcomponents, that of strategy. George and a few otherscholars are indeed buildingbridgesbetween theory and practice.This essay triesto help otherswho wish to followtheirexample.

Bridging theGap
Unless explicit technical or scientific questions are at issue, policymakers tend to draw on academic knowledge principallyin the formof historical study or familiarity with the culture and language of particularregions, leavened occasionally by a bit of economics or comparativepoliticaltheory. As in domestic policymaking,"analysts not familiarwith the government decisionmaking process are surprised and often shocked by how small a directcontribution research makes."2 "Only rarelyhave I witnessed serious governmentalattentionbeing given to serious social science research," said JamesQ. Wilson; "I will make an even stronger I have only rarely statement: observed serious social science being presented to government agencies."3
a researchprogramat Stanforddevoted to "Theoryand Practicein International Relations"; see Alexander L. George, "Bridgingthe Gap between Theoryand Practice,"in JamesRosenau, ed., In Searchof Global Patterns(New York: Free Press, 1976), pp. 114-119. Some of the leading examples of George's scholarship are Alexander L. George and Juliette L. George, Woodrow Wilsonand ColonelHouse: A Personality Study(New York: Dover, 1964); Alexander L. George, David K. Hall and William E. Simons, The Limitsof Coercive Diplomacy (Boston: Little,Brown, 1971; 2d ed. forthcoming); Alexander L. George and Richard Smoke, Deterrence in American Foreign Policy:Theory and Practice(New York: Columbia UniversityPress, 1974); Alexander L. George, Ole Holsti, and R.M. Siverson, eds., Changein theInternational System (Boulder: Westview Press, 1980); Alexander L. George, Presidential Decisionmaking in Foreign Policy:TheEffective UseofInformation and Advice(Boulder: WestviewPress, 1980);AlexanderL. George, ed., Managing U.S.-SovietRivalry:Problems of Crisis Prevention (Boulder: Westview Press, 1983); Alexander L. George, PhilipJ.Farley,and AlexanderDallin, eds., U.S.-Soviet Cooperation: Achievements, Security Failures, Lessons(New York: OxfordUniversity Press, 1988); Gordon A. Craig and Alexander L. George, Forceand Statecraft: Diplomatic Problems ofOur Time(New York:OxfordUniversity Press, 1990); Alexander L. George, ed., Avoiding War:Problems ofCrisisManagement (Boulder: Westview Press, 1991); Alexander L. George, Forceful Persuasion: Coercive as an Alternative to War Diplomacy (Washington,D.C.: U.S. Instituteof Peace Press, 1991). 2. Henry J. Aaron, Politicsand theProfessors: The GreatSociety in Perspective (Washington,D.C.: Brookings,1978), p. 165. This intuitiveimpressionis reinforced by the writingsof many perceptive practitioners who, afterhaving significant foreignpolicymaking experiencesin govemment,joined or rejoined the world of scholarship,includingGeorge Kennan, RaymondGarthoff, Henry Kissinger,William Hyland, Leslie Gelb, Zbigniew Brzezinski,RobertBlackwill,William Quandt, RobertPastor, and GregoryTreverton.Careful narrativehistoryis the dominant analyticaldiscipline. 3. JamesQ. Wilson, "Social Science and Public Policy: A Personal Note," in Laurence E. Lynn, Jr.,ed., Knowledge and Policy:The Uncertain Connection (Washington,D.C.: National Academy of Sciences, 1978), pp. 82-83.

International Security 18:4 1146

in turntend to have littleregardforthe researchthat Governmentofficials is presented to them. Paul Nitze recentlyreflectedthat "most of what has been written and taughtunder the heading of 'politicalscience' by Americans to experience and to common sense. since World War II has been contrary as a guide to the It has also been of limitedvalue, if not counterproductive, actual conductof policy."4 Nitze helped build one of the majorgraduate schools devoted to internationalaffairsand knows a good deal about the example of academic world, and thus his commentis an especially arresting veteransof public service. a view quite widely held among long-time ifthe vast storesof outside would benefit Yet it does seem thatgovernment knowledge could somehow be broughtto bear on daily policy challenges. It has happened. A young student described lunch with the great economist John Maynard Keynes one summer day in 1922. Keynes and his French counterparts could turnfrom"the latestgossip about Continentalstatesmen, theirmistresses,theirneuroses, as well as theirpoliticalmanoeuvres" to "the internationalmovement of money." Most astonishing, they "seemed able and ready to relate theiritems of financialinterestto theoreticaldoctrine" with "subtle points of criticism."Realizing that he "was in the presence of in the latesttheorieswith somethingquite unusual-this mixtureof expertise inside knowledge of day-to-dayevents," the student recalled that "the excitementwas unbearable."5 the "expertise"mustbe connectedto particular To reignite such excitement, problems of public policy. Bridging theGap, George's latest book, identifies findthe rightstrategy threetypes of knowledge that can help policymakers for influencingadversaries. The firstis to conceptualize strategies,expose theirinternallogic, and deduce what elements-in theory-such strategies appear to require. The second is genericknowledge, derived fromsystematic conditionalgeneralizationsabout the comparisonof past cases, which offers strengthsand weaknesses of various chains of strategiclogic. Third is the knowledge thatcan help policymakers develop more nuanced and individual ("actor-specific") models of how otherswill behave. Modest about how much theorycan contribute to finalpolicy judgments, George nonetheless shows how valuable theorycan be, using the case of
ofPolitics(New 4. Paul H. Nitze, Tension Between Opposites: Reflections on thePractice and Theory York: Charles Scribner'sSons, 1993), p. 3. Maynard 5. Roy Harrod describinga lunch on July27, 1922, quoted in RobertSkidelsky,John Keynes: TheEconomist as Saviour1920-1937 (London: Macmillan, 1992), pp. 114-115 (emphasis in original).

Foreign PolicyEngineering 1147

U.S. policytowardIraq before,during,and after the GulfWar. Less interested in the details of either the means used or the ends being sought, George triesto conceptualize the relationshipbetween them. In Bridging theGap, George describes how Washingtonfirst triedto moderate the behavior of an outlaw state with moderate incentives. George defines the essential strategicquestion: how does one go about attempting to moderate the behavior of an outlaw regime?6He criticizesthe Bush administration foreither"poor implementation ofthe strategies" or "an inability to cope with severe situationalconstraintson more effective application of the strategies"(p. 40). He also criticizesthe intelligencecommunity fornot
6. It is important,however, that Iraq was not actually perceived as an "outlaw state" by the moderateArab states or WesternEurope untilthe springor summerof 1990, at the veryearliest. Iraq, viewed as the paladin of the Arab world in early 1989 and the host forfoundationof the Arab Cooperation Council, was actually much closer to the Arab "establishment"than Syria. Iraq retained this status at least until the mask began to drop at the May 1990 ACC summit. Second, George misunderstandsthe nature of the incentivesbeing offered, though his description of them is importantto his narrative.George questions why the Bush administration did not "at least substantially reduce the flow of economic and indirectmilitary assistance to Iraq" assistance and the flow beforethe invasion of Kuwait (p. 34). But therewas no indirect military in May of agricultural creditsactuallywas reduced in 1989 beforebeing suspended indefinitely 1990. The agriculturalcredits program did not add to Iraq's net cash inflow in 1989-90. The cash flow possible under the programwas one-way, fromIraq to the United only international States. Agricultural exporterswere paid directly by Americanbanks who then sought payments on Iraqi lettersof credit.The lettersof creditwere secured by U.S. agricultural creditguarantees, a formof creditinsurance. Iraq received no money. Nor could the guaranteesbe used to secure loans of cash, such as those made to Iraq by the Atlantabranchof the Banca Nazional de Lavoro (BNL). Also, in FY 1990 Iraq received $392 million worth of agricultural goods like wheat and rice under the agricultural creditguarantee program,but made hard currency paymentsof $855 million to American banks on its outstanding lettersof credit. No evidence has emerged to substantiateother assertions that Iraq's military was aided in 1989-90 by kickbacksfromgrain arms market.See "USDA or by bartering Americanwheat and riceon the international exporters Administrative Review of Iraq GSM-102 Program," May 21, 1990 (released by congressional in U.S. Agricultural Trade: Iraq's Participation staff);General AccountingOffice,"International ExportPrograms,"GAO/NSIAD-91-76(November 1990); KennethJuster, "Iraqgate: Anatomyof a Myth," Foreign Policy(forthcoming). The small amount of publiclylicensed Iraqi commercial purchases of dual-use technology in 1989-90 (about $75 million worth) were also restricted further in the spring and summer of 1990, afterdiscovery of several illicitIraqi arms buys, leading to the U.S. government's Enhanced Proliferation Control Initiativeof July1990. See LetterfromJames Baker to Robert Mosbacher, July25, 1990 (released by congressional staff); Departmentof Commerce, "BXA Facts: Fact Sheet on ExportLicensing forIraq," December 16, 1991. George believed, relyingon a misinformed source, that the U.S. actuallycontinued most formsof assistance afterits spring 1990 policy review. In fact,most assistance was cut offin May 1990 as a result of this review. See, e.g., State Departmentnote fromRobertKimmittto James Baker, "DC Meeting on Iraq," April 17, 1990; Commerce Departmentmemo to Dennis Kloske, "InteragencyMeeting on Iraq," June 11, 1990 (both released by congressional staff). Assertions that the United States had a policy of aiding Iraq by willfully ignoringmisconduct in giving loans at BNL-Atlanta are also unsupported. See Senate Select Committeeon Intellion Banca Nazionalde Lavoroand theIntelligence gence, Report Community (1992).

International Security 18:4 | 148

fully pursuing the implications of Saddam's extraordinary dedication of scarce resources to ambitiousmilitary programs,includingweapons of mass
destruction.7

But George's main point is that the strategiesadopted to moderate Iraqi behavior were not well thought out. Thereforeseveral analyticallydistinct approaches, such as graduated reciprocationin tension reduction,bribes, and behavior modification, conditionalreciprocity, were blurred,impeding evaluation. He concentrates coherentexecutionand systematic on the potential fora strategy of conditionalreciprocity, and offers hypothesesabout the risksof such a strategy and how to minimizethem. of confrontation Given the lack of preparationfora strategy and containment of Iraq, it is no surprise to George that the United States was unable to mount a credible strategyof deterrenceor reassurance in the last weeks before the August 1990 invasion of Kuwait. Policymakersunderstood the well enough, George concludes, but conceptual requirementsof the strategy neitherCongress nor local Arab states were willing to support the military deploymentsthat could have displayed a credibledeterrent. The "sobering" lesson is that "what the United States was willing and able to do after the aggression against Kuwait, it was not able to threatento do on behalf of deterrencebefore the invasion occurred" (pp. 73-74). Efforts to reassure Saddam Hussein about America's desire forfriendshiponly mixed the already weakened signals. George then turns to the new range of strategiesthat became available afterthe invasion of Kuwait, which fallunder the heading of coercivediplomacy. Here his conceptualizationhelps again, distillingthe need to decide: (1) what to demand of the opponent; (2) whetherand how to create in the
7. A possible excuse George offersfor the intelligencefailureis that "the commitment at the highest levels of the administrationto pursue friendshipwith Saddam Hussein may have discouraged intelligenceand policy specialists froma more vigorous challenge to the policy" (p. 42). He later refersagain to this "strong presidentialcommitment" (p. 50). George is too generous to the analysts. There are numerous examples of analysts not being so deterred.But a key featureof America's Iraq policy beforeJuly1990 is the relativeinattention of top officials. PresidentBush participatedin only one discussion of policy toward Iraq beforethe invasion of Kuwait (in the summer of 1989) and his contribution was to wonder aloud whetherSaddam's behavior could be changed. The consensus response, even then,was that"the leopard will not change his spots." Interviewswith RobertGates, RichardKerr,RichardHaass, and U.S. intelligence officials. The relevantpolicydocument,NSD 26, was neveritself discussed at the Cabinet level, was so ambiguously draftedas to be almost completelynon-controversial, and sat in White House in-boxes formonths beforeBush botheredto sign it.

Foreign PolicyEngineering | 149

adversary'smind a sense of urgencyabout complyingwith the demand; (3) how to createa sufficiently crediblethreatof punishmentfornoncompliance; and (4) whetherto couple the threatenedpunishmentwith positive inducements. Neitherthe United States and its key allies on the one hand, nor Saddam Hussein on the other, were ever willing to negotiate down fromabsolute positions. The U.S. choice was deliberate:aggression had to be seen clearly to fail. Saddam could be allowed to save himselfand his army,but he could not be allowed to save face or be rewarded in any way. George does not argue thatWashingtonshould have softenedits approach. He does spell out the difficult challenge coercive diplomacy faced in trying to impose such demands, as the coalition shiftedfromthe strategyof economic sanctions to an ultimatumbacked by a threatof military action. Still, the main miscalculationwas Saddam's and, "ironically, the failureof coercive diplomacy was necessary [for the United States] to gain support for war when war became the last resort"(p. 88).8 George demonstrates that conceptualizing strategiescan illuminate the logic and key variables associated with theirsuccess. He is therefore justified in claimingto have created a "basic framework forunderstandingthe nature and general requirementsfordesigning an effective strategy"(p. 137). Both genericknowledge and analysis of the actors involved can then be brought to bear to improve this component of policymaking.Unfortunately, "adequate scholarlyknowledge of the conceptual and generictypes does not yet existformany of the standard strategies and instruments of policy" (p. 138). Why is thisso? Much of the problemin the dialogue between policymakers and academic expertsstem, as Adam Yarmolinsky noted long ago, fromthe inabilityof officialsto formulatethe kind of questions that academics can usefullyanswer.9

The Questions Academics Ask

Many observers have noted, of course, that social scientistsand historians kinds of questions pursue fieldsof intellectual inquirythatask quite different
8. See also JaniceGross Stein, "Deterrenceand Compellence in the Gulf, 1990-91: A Failed or Impossible Task?" International Security, Vol. 17, No. 2 (Fall 1992), pp. 147-179. 9. See Adam Yarmolinsky'sexcellentessay, "How Good Was the Answer? How Good Was the and Decisions:The Social Sciencesand Public Question?," in Charles Frankel, ed., Controversies Policy(New York: Sage Foundation, 1976), pp. 259-272.

International 18:4 | 150 Security

than those thatinterestpolicymakers.There is a striking resemblanceto the dichotomybetween the sciences, such as physics and mathematics,and the professionof engineering.Formeraerospace engineer and currentStanford professorJames Adams has pointed out how "the motivationsof pure scientists are markedlydifferent fromthose of practicingengineers" because "pure scientists desire to understandphenomena. Theirproductis published knowledge, and theiraudience and judges are theircolleagues. They are not necessarilyconcernedwith the applicationof theirknowledge." On the other hand, "engineers are motivatedto solve theirproblem successfully withina to understandthe microscopic given schedule and budget. They would prefer phenomena that cause macroscopic behavior,but they must complete their work whetherthey do or not."10 Two distinctproblems are presented in applying "outside" knowledge to "inside" problems. The firstis to develop knowledge that can readily be in the development of public policy.11 applied by practitioners But a second problemis to understandthe ingredients of policydevelopmentwell enough to see how to use "outsiders"' stores of knowledge to help answer material
questions. 12

10. JamesL. Adams, Flying Buttresses, Entropy, and O-Rings:TheWorld ofan Engineer (Cambridge: Harvard UniversityPress, 1991), pp. 40-41. For comparable descriptionsin the public policy area see Seymour J.Deitchman, The Best-Laid A Tale of Social Research Schemes: and Bureaucracy theGap, pp. 3-18; and Yarmolinsky, (Cambridge:MIT Press, 1976), pp. 440-41; George, Bridging "How Good Was the Answer? How Good Was the Question?" pp. 262-263. AfterPhilip Heymann returnedfroma sojourn in government(as head of the JusticeDepartment's Criminal Division), he wrote about just how to maintain legislativeand public support in the political process foran agency and its leader's objectives. He observed that "the literature in the area of my concern is oftenbrilliantbut always addressed to a somewhat different set of questions," so thatbooks like Allison's EssenceofDecision,forexample, "has much to teach, but its purpose is explainingnot prescribing."Philip B. Heymann, ThePolitics ofPublicManagement (New Haven: Yale University Press, 1987), pp. xiii-xiv. 11. For some examples of how this can be done, see Richard E. Neustadt and Ernest R. May, in Time:The Uses ofHistory Thinking forDecision-Makers (New York: Free Press, 1986); RichardE. FDR toReagan(New York:JohnWiley Neustadt, Presidential Power:ThePolitics ofLeadership from and Sons, 1988); Robert L. Rothstein,Planning,Prediction, and Policymaking in ForeignAffairs: and Practice Theory (Boston: Little,Brown, 1972); and Alexander George's own work. 12. The goal should be to draw out those propositionsthat are "material" to the development offoreign policies. Americanlegal rules ofevidence draw a nice distinction between the concepts of materiality and relevance. "In the courtroomthe termsrelevancyand materiality are often in its more precise meaning looks to the relationbetween used interchangeably, but materiality the propositionforwhich the evidence is offeredand the issues in the case. If the evidence is offeredto prove a propositionwhich is not a matterof issue or probativeof a matterat issue, the evidence is properlysaid to be immaterial.... Relevancyin logic is the tendencyof evidence to establish a proposition which it is offeredto prove. Relevancy,as employed by judges and lawyers,however, is the tendencyof the evidence to establisha materialproposition." Edward

Foreign PolicyEngineering | 15

Much of the study of public policy is concerned with developing theories about why nations, governments,or officialsact the way they do, rather than theories about how they should do better.Many of the theoriesabout national behavior have such a deterministcast that the effort to improve policymakingperformancecan even seem to be intellectually vain or uninteresting. Afterall, if one has painstakingly developed propositionsabout patterns of behavior of nation-statesover centuries, the attemptto apply such theories to a particularbilateral policy in a given month must seem 13 One result,as George observes in Bridging appallinglyreductionist. theGap, is that "not a few policy specialists exposed to the scholarlyliterature have concluded that most universityprofessors seem to write largely for one anotherand have littleinclinationor abilityto communicatetheirknowledge in termscomprehensibleto policymakers"(p. 7). The kinds of questions academics ask have influencedthe methodology theyemploy in answering them. The dominantmethods of scientific inquiry into foreignpolicymakingcompare the inputs into policymakingand the outputs fromit, or just observe outputs, so as to postulate the patternsof behavior thatvalidate or disprove theory.Anotherversionis to use historical case studies to develop axioms about the way nations, particulargovernments, or officials tend to behave.14 The literatureon inputs into policymakingis certainlyextensive. Such inputs can be traced back to the culture of a nation or society,15extend to

W. Cleary,ed., McCormick's Handbook oftheLaw ofEvidence (St. Paul: West PublishingCo., 1972), pp. 434-435. 13. On the divide between the tasks of providinginformation useful in making policy choices and the task of studying social processes, see Mark H. Moore, "Social Science and Policy in Daniel Callahan and BruceJennings, Analysis: Some Fundamental Differences," eds., Ethics, theSocial Sciences, and PolicyAnalysis(New York: Plenum Publishing,1983), pp. 271-291. 14. On the underlyinglimitations of this method of studyingpolitics,and the oftenfalse view that techniques forthe scientific study of natural phenomena can readilybe applied to human subject matter,see Gabriel A. Almond and Stephen J. Genco, "Clouds, Clocks, and the Study of Politics," WorldPolitics,Vol. 30 (1977), pp. 489-522. George himself helped develop the methodologyforusing historicalcases to distillvaluable axioms or conditionalgeneralizations. See Alexander L. George, "Case Studies and Theory Development: The Method of Structured, Focused Comparison," in Paul Gordon Lauren, ed., Diplomacy: in History, NewApproaches Theory, and Policy(New York: Free Press, 1979). 15. Contrast,forexample, the magisterial distillations of RaymondAron,Peaceand War:A Theory ofInternational Relations, tr.RichardHoward and AnnetteBaker (Garden City:Doubleday, 1966), PartTwo; or Stanley Hoffmann,Gulliver's Troubles, Or theSetting ofAmerican Foreign Policy(New York: McGraw Hill, 1968), Parts Two and Three; with the more scientific approach exemplified in Ronald Inglehart, "The Ren'aissance of Political Culture," American PoliticalScienceReview, Vol. 82, No. 4 (December 1988), pp. 1203-1230; or Harry Eckstein, "A CulturalistTheory of

International Security 18:4 j 152

the role of public opinion,16 and encompass prevailingintellectualcurrents and ideas, including the way sets of ideas are selected and promoted by communities of experts.17 The individual backgrounds and psychological motivationsof officials have not been neglected either.18All of these factors can sometimes be brought togetherto create a compelling picture of the settingin which a nation's foreignpolicy is made.19 Yet many scholars do tryto probe more deeply into the way in which policies are made. They tend to do so in two ways. First,they look at the organization of the policymakingprocess. Second, they turn their microscopes on "decision making."

Political Change," American PoliticalScienceReview,Vol. 82, No. 3 (September 1988), pp. 789804. 16. Public opinion is, of course, mediated through the various institutionsfound in quite different democracies. See, e.g., Richard Eichenberg, Public Opinionand National Security in Western Europe(Ithaca: Cornell University Press, 1989); Bruce Russett,Controlling theSword:The Democratic Governance of National Security (Cambridge: Harvard UniversityPress, 1990); and Thomas Risse-Kappen, "Public Opinion, Domestic Structure,and Foreign Policy in Liberal Democracies," World Politics, Vol. 43, No. 4 (July1991), pp. 479-512. 17. On the general role of ideas, using the example of the persistentAmericancommitment to freetrade and how it has been filtered throughstate structures createdwith particular purposes in mind, see JudithGoldstein, "Ideas, Institutions, and American Trade Policy," International Vol. 42, No. 1 (Winter1988), pp. 179-218. An effort Organization, to calibratesystematically the relationship of goals to policy strategies and execution is John P. Lovell, ForeignPolicy in Perspective: Strategy, Adaptation, DecisionMaking(Hinsdale, Ill.: Dryden Press, 1970). On the role of transnationalnetworksof experts, or "epistemic communities,"see Peter M. Haas, "Introduction: Epistemic Communities and InternationalPolicy Coordination," International Organization,Vol. 46, No. 1 (Winter1992), pp. 1-35; Haas, "Do Regimes Matter?EpistemicCommunitiesand MediterraneanPollution Control,"International Organization, Vol. 43, No. 3 (Summer 1989), pp. 377-404; and Emmanuel Adler, "The Emergenceof Cooperation: National Epistemic Communitiesand the International Evolutionof the Idea of Nuclear Arms Control,"International Organization, Vol. 46, No. 1 (Winter 1992), pp. 101-145. For an interesting synthesis of how ideas, epistemiccommunities,and politicalstructures interact to reshape a national agenda, see Sarah E. Mendelson, "Internal Battles and External Wars: Politics, Learning, and the Soviet WithdrawalfromAfghanistan,"World Vol. 45, No. 3 (April 1993), pp. 327-360. Politics, 18. A finesummaryof the literature on general psychologicaltheoriesof motivation is in James Dougherty and Robert Pfaltzgraff, Theories A Comprehensive Contending ofInternational Relations: 3rd ed. (New York: HarperCollins, 1990), pp. 272-310; on generalpsychologicaltheories Survey, of perception(which apply to background motivation,cognitionof information, and analytical processes), see Robert Jervis,Perception and Misperception in International Politics (Princeton: PrincetonUniversityPress, 1976); Richard Ned Lebow, Between Peace and War: The Natureof International Crisis (New York: Free Press, 1981); and Thomas Wiegele, Gordon Hilton, Kent Layne Oots, and Susan Kiesell, LeadersUnderStress: A Psychophysiological Analysis ofInternational Crisis(Durham: Duke UniversityPress, 1985). 19. E.g., Hoffmann,Gulliver'sTroubles; and Michael Brecher,The Foreign PolicySystem ofIsrael: Setting, Images,Process(New Haven: Yale University Press, 1972).

Foreign Policy j 153 Engineering

The policymaking process is undoubtedly important.20 Organizations within it vie with one another for dominance and autonomy. Meanwhile theyestablishimportant capacities and routinesforpolicy implementation.21 It is striking,though, how often the craftof policymakingis routinely equated simplywith the organizationof this process.2 This is true forpractitionersand academics alike. "How can United States foreignpolicy be improved?"Roger Hilsman asked, and then answered: "Wheneverthe question comes up, attentionturnsfirst to the question of organization."23 The other usual method fordissectingpolicymaking is to dissect the process of "decision making." The implicitimage in most writing is of a host of factorsthat converge upon some climacticmoment or period when a great
20. The best album of images, whatever their relative explanatorypower, is still Graham T. Allison, Essence of Decision: Explainingthe Cuban Missile Crisis (Boston: Little, Brown, 1971) (explanation of Models II and III). See also Graham Allison and Morton H. Halperin, "BureaucraticPolitics:A Paradigm and Some Policy Implications,"World Politics, Vol. 24, supp. (Spring 1972), pp. 40-79; Graham Allison and Peter Szanton, Remaking Foreign Policy:The Organizational and Foreign Connection (New York: Basic Books, 1976); I.M. Destler,Presidents, Bureaucrats, Policy: ThePolitics ofOrganizational Reform (Princeton:Princeton University Press, rev. ed., 1974); Morton H. Halperin, National Security Policy-Making: Analyses,Cases, and Proposals(Lexington, Mass.: D.C. Heath, 1975); Morton Halperin with PriscillaClapp and Arnold Kanter,Bureaucratic Politics and ForeignPolicy (Washington, D.C.: Brookings, 1974); and Francis Rourke, Bureaucracy and Foreign Policy(Baltimore:JohnsHopkins University Press, 1972). 21. See James Q. Wilson, Bureaucracy (New York: Basic Books, 1989), p. 221; and Samuel R. Williamson,Jr.,"Theories of OrganizationalProcess and ForeignPolicy Outcomes," in Lauren, Diplomacy, pp. 137-161. 22. See, e.g., Charles F. Hermann, "Decision Structureand Process Influences on American Foreign Policy," in Maurice East, Stephen Salmore, and Charles Hermann, eds., WhyNations Act:Theoretical Perspectives forComparative Foreign PolicyStudies (BeverlyHills: Sage, 1978), pp. 69102; MargaretG. Hermann and Charles F. Hermann, "Who Makes ForeignPolicyDecisions and How: An EmpiricalInquiry,"International Vol. 33, No. 4 (December 1989), pp. StudiesQuarterly, Subcommittee 361-388; Henry M. Jackson, ed., The National Security Council:Jackson Papers on at thePresidential Level(New York:Praeger,1965); Amos Jordan, WilliamL. Taylor, Policy-Making National Security: Jr.,and Lawrence Korb, eds., American Policyand Process(Baltimore:Johns Hopkins UniversityPress, 3rd ed., 1989), Part Two; Roy Macridis, ed., Foreign Policyin World Politics,3rd ed. (Englewood Cliffs:Prentice Hall, 1967); and John F. Reichartand Steven R. 5th ed. (Baltimore: Sturm,American Defense Policy, JohnsHopkins University Press, 1982), chapter 6. 23. Roger Hilsman, ThePolitics ofPolicyMakingin Defense and Foreign Affairs (New York: Harper and Row, 1971), p. 151. Zbigniew Brzezinski,recountinghis years as national securityadviser to President Carter,entitleda section of his memoirs,"Making Policy." This promisingtitleis then followed by the words: "Coordination is predominance." The section speaks only of how the process was organized and who tended to talkto whom and how often,affirming-naturally enough-that Brzezinski's coordinatingrole conferred upon him dominance. No mentionwas made of the substance of how policies were crafted.Zbigniew Brzezinski,Powerand Principle: Memoirs Adviser1977-1981 (New York: Farrar,Straus and Giroux, 1985), of theNationalSecurity pp. 63, 63-74.

Security 18:4 | 154 International

decision, or a set of related decisions, is made.24In a discipline dominated by concerns about conflict,the tendency naturallyis to focus upon "crisis decisionmaking,"episodes that are laden with the potentialforconflict and compressed in space and time. These episodes are then taken apart to see how the policy machineryactuallyworks.25 Experts have concluded that the machinerydoes not work mechanically. Theorists long ago showed that decision making processes do not match stylized or economic models of synoptic,efficient, and utilitarian choice.26 The next stage was to show how, given the inherentflawsin human perception and cognitive assimilation of information as well as the constrained decision environment,officialscould grope for months, even years, down paths that mightwell seem quite irrational, even "bizarre," to an outsider.27
24. Literatureon negotiations also focuses on a key event-the agreement-and sometimes follows patternsrecognizable fromthe writingson decisionmaking.See, e.g., the descriptions of the "diagnostic phase," "formula phase," and "detail phase," in I. William Zartman and Maureen R. Berman, ThePractical Negotiator (New Haven: Yale University Press, 1982); but also see Linda P. Brady,ThePolitics America's ofNegotiation: and Friends Dealingswith Allies,Adversaries, (Chapel Hill: University of North Carolina Press, 1991). 25. They have been taken apart in almost every conceivable way, fromperceptions of information,to patternsofbargaining,to measurementsofinteraction between decisionmaking units, the effectsof stress and surprise, and styles of "crisis management." Leading examples are Allison, Essenceof Decision;Alexander George, ed., Avoiding War:Problems ofCrisisManagement (Boulder: Westview Press, 1991); Charles Hermann, ed., International Crises:Insights from Behavioral Research (New York: Free Press, 1972); Ole Holsti, Crisis, Escalation, and War (Montreal: McGill-Queens UniversityPress, 1972); Richard Ned Lebow, Between Peace and War:The Nature of International Crisis (Baltimore:Johns Hopkins UniversityPress, 1981); Glenn D. Paige, The KoreanDecision, June 24-30, 1950 (New York: Free Press, 1958); J.David Singer,ed., Quantitative International Politics (New York: Free Press, 1968) (especially cases on 1914 and Berlin);and Oran Young, ThePolitics ofForce:Bargaining DuringInternational Crises(Princeton:PrincetonUniversity Press, 1968). 26. The groundbreakingstudies were Herbert A. Simon, Administrative Behavior (New York: Macmillan, 1958); RichardC. Snyder,H.W. Bruck,and BurtonSapin, eds., Foreign PolicyDecision Making:An Approach to theStudyofInternational Politics (New York: Free Press, 1962); and David Braybrookeand Charles E. Lindblom, A Strategy of Decision:PolicyEvaluation as a Social Process (New York: Free Press, 1963). For a more nuanced definition of decision makers' "rationality," see Paul Diesing, Reason in Society: Five Typesof Decisionsand TheirSocial Conditions (Urbana: Universityof Illinois Press, 1962). This work formedmuch of the basis forAllison, Essenceof Decision;and Glenn H. Snyder and Paul Diesing, Conflict AmongNations:Bargaining, DecisionMakingand System Structure in International Crises(Princeton:PrincetonUniversity Press, 1977). 27. JohnSteinbruner, TheCybernetic Theory ofDecision: New Dimensions ofPolitical Analysis (Princeton: PrincetonUniversityPress, 1974), p. 332. See also Yaacov Y.I. Vertzberger, The Worldin Their Minds:Infornation and Perception Processing, Cognition, in Foreign PolicyDecisionmaking (Stanford, Calif.: Stanford UniversityPress, 1990); and note the introductionof the concept of "disjointed incrementalism"in Braybrookeand Lindblom, A Strategy of Decision,pp. 81-146. Much ofthisworkin the fieldofpublic policyhas reliedheavilyon organizationtheoryoriginally postulated to explain and describe the behavior of private firms.An excellent distillationis JamesG. March, Decisionsand Organizations (Oxford:Basil Blackwell, 1988).

Foreign PolicyEngineering | 155

The net result of such examinations is a better understanding of why littleto the policies come out the way they do, though adding surprisingly insights that can be gleaned from direct perusal of the better works of diplomatichistory.But describingor even explainingoutcomes is stillquite different fromprescribinghow policies ought to be constructed.One can understand how a political scientistcould confess that "the policy-making process has been like a black box to many of us since we see what comes out but not much of what happens inside. . . . Our habits of inquiryhave imprisonedus."28

The SevenComponents ofPolicymaking

If policies are the settled course of action to be followed by a government, then policymakingis more than the politicalprocesses that determinehow choices are made. Writingabout domestic governance, John Kingdon has pointed out thatthe "threemajor process streamsin the federalgovernment are (1) problem recognition,(2) the formation and refining of policy proposals, and (3) politics." Politics includes bureaucraticpolitics.29 Though much of the time these streamsmay flow independentlyfromeach other,theycan interact,or be coupled together,at any point in ways that determinethe finaloutcome. Policymakingis also different frompolicy analysis. The policy analyst's a problem, specifyalternaclassic approach to policy analysis is to identify select the best one, tives, evaluate them according to some explicitcriteria, and implement the decision. The components of policymakingdescribed here, however, may in the real world be more disorderlyand more fine28. Charles W. Kegley, Jr.,"Decision Regimes and the Comparative Study of Foreign Policy," in Charles F. Hermann, Charles W. Kegley, Jr.,and JamesN. Rosenau, eds., New Directions in theStudyofForeign Policy(Boston: Allen and Unwin, 1987), p. 248. Philip Heymann noted one model forusing systematic analysis to suggest how to get results:RichardNeustadt, Presidential Power:ThePolitics ofLeadership from FDR to Carter (New York:JohnWiley and Sons, 1980). 29. John Kingdon, Agendas,Alternatives, and Public Policies(New York: HarperCollins, 1984), p. 92. "Politics," as I use it here, includes the factors in Allison's Model III, such as the channel forproducing action on the problem, the procedures forreconcilingagency views, the players who are centrallyinvolved, the histories and personalities of the players, and the external deadlines forcingthe issue to resolution. See, e.g., Allison, Essenceof Decision,p. 257. The concerns of Allison's Model II about organizationalresponsibilities and routinesapply here as well, but some Model II elements also featurein the policymaking process itself,in constraints on design choices, and in the separate dimension of policyanalysis, such as how organizational capacities constrainthe availabilityof information or evaluation criteria.

International Security 18:4 | 156

grained. While using a simplified diagram of policymaking to help framehis analysis and formulate the rightquestions, the analyst should not substitute that frameworkfor the more formalmethodologies he has developed for answeringthem,whetherhis goal is "quick analysis" forimmediate,practical application or longer-term"researched analysis" to probe the answers to deeper questions.30 Finally,policymakingis more than just a decision. Choices may be presented to a president,but the resultingdecision will only be one aspect of policymaking, a process that began before the decision memorandum reached the president's desk and continues afterit has gone into the outbox. Policymakingthus bears littleresemblance to the commonlyencountered model whereby options percolate up throughcontendingbureaucraciesto a culminating decision. During the year ofAmericandiplomacyassociated with German unification, forexample, PresidentBush never received an options paper. But this did not mean thatoptions were not considered, or thatBush did not know about them. Nor do formalevaluations of options find much reflection in the storyof the Carter administration's handling of the Camp David peace process. Even where a choice among clearcutoptions does take place, that choice is oftenover just one of the ingredientsin the policy.31 There have been a few effortsto itemize the components of national or foreignpolicymaking.32 security Buildingon elementsof priorscholarship

30. There is a large literatureon the purposes, nature, and utilityof policy analysis. For variationson the "classic" elements of policy analysis, see Percy H. Hill, ed., MakingDecisions: A Multidisciplinary Introduction (Reading, Mass.: Addison-Wesley,1978), p. 22; Duncan MacRae and JamesA. Wilde, PolicyAnalysis forPublicDecisions(North Scituate,Mass.: Duxbury Press, 1979), pp. 7-12; Carl V. Pattonand David S. Sawicki, BasicMethods ofPolicy Analysis and Planning, 2d ed. (Englewood Cliffs,N.J.: PrenticeHall, 1993), p. 3, Fig. 1-1; Edward S. Quade, Analysis forPublicDecisions,2d ed. (New York: Elsevier,1982), pp. 47-62; and Edith Stokeyand Richard Zeckhauser, A Primer forPolicyAnalysis(New York: Norton, 1978), pp. 5-6. On the distinction between "quick analysis" and "researched analysis," see RobertD. Behn and JamesW. Vaupel, QuickAnalysis forBusyDecisionMakers(New York:Basic Books, 1982), pp. 3-7. On the legitimacy of accepting"political"parametersas part of the framework forvalid policyanalysis, see Arnold J. Meltsner,PolicyAnalystsin the Bureaucracy (Berkeley:Universityof CaliforniaPress, 1976), pp. 307-309. See generallyGiandomenico Majone, Evidence, and Persuasion in the Argument, Policy Process(New Haven: Yale University Press, 1989). 31. See William Quandt, Camp David: Peacemnaking and Politics(Washington,D.C.: Brookings, 1986); forothergood narrativesof policymaking, see, e.g., Leslie H. Gelb with RichardK. Betts, TheIrony ofVietnam: TheSystem Worked (Washington,D.C.: Brookings,1979); I.M. Destler,Making Foreign Economic Policy(Washington,D.C.: Brookings,1980). 32. Charles Hermann has divided the "decision process," for example, into "(1) initialpolicy

Foreign PolicyEngineering | 157

on both domesticand foreignpolicy,I suggest that"policy engineering"can be seen as one of three streams,each of which interactconstantly with the other. The other two are "problem recognition,"and "politics" (defined as the way choices are made), as portrayedin Figure 1. In addition to its interaction with the evolvingperceptionsor assessments of the problem and the constraintsimposed by the realitiesof politics, the streamof "policy engineering"is informedby two major analyticmethods: policy analysis and institutional intervention in Soanalysis. A multilateral malia, for example, would require analysis both of the efficacy of certain competingapproaches and of the culture,capacities, personalities,and routines of relevantinstitutions such as the United Nations. The policy would then, ideally,be developed to take advantage of both kinds of analysis.33 The "problem recognition"streamis fairly well understood, although the literatureis thinner on the criticalchallenge of assessing foreigngovernments. The roles of both experts and the media have been repeatedly examined and the bulk of "researched analysis" in the academic community
expectations;(2) externalactor/environmental stimuli;(3) recognition of discrepantinformation; (4) postulationof a connectionbetween problemand policy; (5) developmentof alternatives; (6) building authoritative consensus forchoice; and (7) implementation of new policy." His taxonomy focuses on a particular decision situationand is brokeninto categoriesof theoretical inquiry ratherthan stages in policymaking.Charles Hermann, "Changing Course: When Governments Choose to RedirectForeign Policy," International StudiesQuarterly, Vol. 34, No. 1 (March 1990), pp. 3, 14. For a betterbut too elaborate list of thirteen steps in the policymakingprocess, see Elmer Plischke, ForeignRelations: Analysisof its Anatomy (Westport:Greenwood Press, 1988), p. 257. For a list of five stages of the "decision-makingprocess" culminating in a decision, see Paul A. Anderson, "What Do Decision Makers Do When They Make a ForeignPolicyDecision?" in Hermann, Kegley, and Rosenau, New Directions in theStudyofForeign Policy,p. 304. Among practitioners, Dean Rusk listed seven items in a 1965 address to the AmericanPoliticalScience Association, "Anatomy of Foreign Policy Decisions," Department Vol. 53 (Sepof StateBulletin, tember27, 1965), pp. 502-509; and Theodore Sorensen offereda list of eight steps in DecisionMakingin theWhite House (New York: Columbia University Press, 1963), pp. 18-19. To resolve disputes about the natureof arms races, Matthew Evangelistahad to categorizethe actual stages of policymaking in cases of weapons innovationin both the United States and the Soviet Union. He outlines fivestages in both countries,fromtechnocratic initiative to high-levelendorsement in the United States, and fromstifledinitiativeto mass productionin the USSR. His outline is specificto the process of weapons innovation. But the diagnostic power of the technique was revealing. Innovation and theArmsRace: How the UnitedStatesand theSovietUnionDevelopNew Military Technologies (Ithaca: CornellUniversity Press, 1988), p. 52, Table 3; or Evangelista,"Issuearea and Foreign Policy Revisited," International Vol. 43, No. 1 (Winter1989), pp. Organization, 147, 156. 33. Policy analysis is what George calls "substantivetheory,"which provides knowledge about standard foreignpolicy undertakings,instruments of policy, and strategies.At the same time the politics stream should ideally be informedby what George calls "process theory,"which suggests how to structurethe'management of information and to make the choices in the manner dictatedby the politics stream. George, Bridging theGap, pp. 20-21.

18:4 | 158 Security International

of Policymaking. Figure1. Components

| Policy | Analysis

|
|

| Institutional
|

Analysis

Problem
Recognition

E>

Policy Engineering

E-o

olitics! Choice

Policy Engineering
h

|m

naoInterests

NatioIi

Objetives

Statgy E7 ew
[

Design

| Implementation|

Maintenance ||

Review

| 159 Foreign PolicyEngineering

can already be brought to bear on the recognitionof problems.34 As mentioned earlier,the processes in the "politics" streamhave also been studied intensively. So this essay concentrates on the "policy engineering" stream, whose components are as follows: * * * * * * * National interest Objectives Strategy Design, preliminary and detailed Implementation Maintenance Review

This is a schematic,idealized outline. The components will not be found in all policies. These components need not occur in neat chronologicalsequence. They influenceeach otherand revisioncan occur. Politicalfailureto achieve agreementon criticaldesign featuresmight,forexample, cause the policymakerto amend the policy's objectives. "Strategy"and "design" also turn out to be practicallyinseparable, since each shapes the other. The components are treated separately here in order to distinguish reflection about the conceptual relationshipbetween ends and means (as in George's book) from the assorted questions that always arise about the particular designs to carryforwardany chosen strategy.35
34. On participants in problemrecognition and how itoccurs,see Kingdon,Agendas, Alternatives, and PublicPolicies,pp. 23-121. For insightsinto a centraldimension of foreignpolicy problem recognition-foreignassessment-good startingplaces are Aaron Friedberg,The WearyTitan: Britainand the Experience of RelativeDecline 1895-1905 (Princeton:PrincetonUniversityPress, 1988), an extended case studyin the assessment of global power and adaptation to the changing recognition of Britain's"problem"; forthe synthesisof a numberof cases of military assessment, in May, ed., Knowing ErnestR. May, "Conclusions: Capabilities and Proclivities," One's Enemies: Intelligence Assessment Before the Two WorldWars (Princeton:PrincetonUniversityPress, 1984), in May, TheMakingoftheMonroeDoctrine pp. 503-542; and the templateforassessment offered (Cambridge: Harvard University Press, 1975). 35. This set of components can also be contrastedwith HerbertSimon's three-phasedivision of intelligence,design, and choice; HerbertA. Simon, The ShapeofAutomation (New York: Harper and Row, 1965), p. 54; or with Mintzberg,Raisinghani,and Theoret's parallel considerationof identification, development, and selection; Henry Mintzberg,Duru Raisinghani, and Andre Theoret, "The Structureof 'Unstructured'Decision Processes," Administrative ScienceQuarterly, Vol. 21, No. 2 (June 1976), pp. 246, 252. Both conceptions are oriented to a decision setting, not to the process of policymakingas a whole. They also tryto be descriptiveand explanatory, not normative.Their schemes apply best to understandingonly the first threecomponents in the outline presented here.

International Security 18:4 1160

The "level of analysis" is also important. This essay assumes thatthe level of analysis is a national government, not particularagencies or officials.

NATIONAL

INTEREST

The national interestis a "non-operationalgoal" which is often used as a forwhateverpreferences generalrationalization actuallyundergirda policy.36 A nation "never admits it is doing violence to its moral instincts."37 at least in this centuryhas voiced Hence every American administration its solemn regard forworld peace, the nation's defenses, America's growing and the progress of democracyaround the world. In September prosperity, 1993, in a much-heraldedaddress, PresidentClinton's national securityadviser said that the United States would support the "enlargementof the world's free communityof marketdemocracies." One can assume the administration did not lingerlong over the alternative of "shrinkage."38 Effective statementsof national interestseparate more important interests fromless important ones. It is therefore somewhat surprising thatacademics have tended to neglect the importanttask of formulating hierarchies of national interestthat can support such judgments.39
36. Alexander George and Robert Keohane, "The Concept of National Interests: Uses and Limitations,"in George, Presidential DecisionMakingin Foreign Policy,p. 219. See also Arnold Wolfers,"National Security as an Ambiguous Symbol," in Wolfers,Discordand Collaboration (Baltimore:Johns Hopkins UniversityPress, 1962), pp. 147-165; Thomas Robinson, "National Interests,"in JamesN. Rosenau, ed., International Politics and Foreign Policy:A Readerin Research and Theory (New York: The Free Press, 1969), pp. 182-190. A StudyofNationalist 37. AlbertK. Weinberg,Manifest Destiny: Expansionism in American History (Baltimore:Johns Hopkins UniversityPress, 1935), p. 73. "According to certain diplomatic the abstractformulation officials, of national goals is the easiest and least demanding step in the conduct of foreignaffairs."Plischke, Foreign Relations: Analysis ofitsAnatomy, p. 98. 38. AnthonyLake, "From Containmentto Enlargement,"address at JohnsHopkins University, School of Advanced InternationalStudies (SAIS), Washington,D.C., September 21, 1993, in Department ofStateDispatch,Vol. 4, No. 39 (September27, 1993), pp. 658-664. The explanation of "enlargement"was equally amorphous. The first of fourstated strategiesto achieve the goal is to "strengthenthe communityof major marketdemocracies." But one finds the only policy contentto be a call for"the major marketdemocracies [to] act together-updating international economic institutions, coordinatingmacro-economic policies, and striking hard but fairbargains on the ground rules of open trade." No guiding principlesforthese efforts can be discerned. A possible inference is in the use of the code phrase "open trade" instead of "freetrade," yet even this fragileinferenceappears belied by the free-trade tone of this section of Lake's speech. 39. For two exceptions,see RobertBlackwill,"A TaxonomyforDefiningU.S. National Security Interestsin the 1990s and Beyond," in WernerWeidenfeldand JosefJannings,eds., Europein GlobalChange(Gutersloh: BertelsmannFoundation, 1993), pp. 100-119; and Donald Nuechterlein, AmericaRecommitted: United States National Interests in a Restructured World(Lexington: of KentuckyPress, 1991), pp. 13-33. University

PolicyEngineering Foreign 1161

Expressionsof nationalinterest thus onlybecome a meaningful component of policymaking when the pronouncementtellspeople somethingsignificant thattheydid not already know about the futuredirectionand commitments of the government.Washington's Farewell Address, the Monroe Doctrine, Wilson's Fourteen Points, Truman's address to Congress seeking aid for Greece and Turkey, the 1950NSC 68 directive, and PresidentCarter'sJanuary 1980 statementthatthe security of the Persian Gulfranked among America's vitalinterests are all examples of declarationsthatmarkedout new reference points forthe formation of policies. Statementsneed not be historicallandmarks to communicate important guidance. Bush's unqualifiedand unequivocal endorsementsof German unificationin May, September, and October 1989 before the fall of the Berlin Wall were statementsabout America's national interest, which disappointed some allies while they emboldened another. in 1989,itquestioned the Reagan When the Bush administration took office administration'sfirstefforts to move fromdetente to rapprochementin its relations with the Soviet Union.40 Then the Bush administrationrapidly shiftedgears, beginning with a May 1989 statementthat the United States would move "beyond containment";with subsequent arms control initiatives, this confirmedthe shiftto rapprochementwhile hintingat the possibility for even greater change.41 Secretary of State Baker's October 1989 speech on "points of mutual advantage" indicated for the firsttime that toward a genuine Washingtonwas prepared to look beyond rapprochement entente with the Soviet Union.42This entente was a realityby the end of 1990, as treatiessettledthe fateof Germanyand of Soviet conventionalarms in Europe, and active cooperationwas undertakenin the confrontation with Iraq. PresidentClintonhas also chosen to emphasize certaingoals and priorities, such as the world economy, nonproliferation, and the global environment. It is stilltoo early to tell how these declarationsof interestwill fareas they are turned into policies, because none of these statementsabout national interestwere operational. They did not, standingalone, establish new policies. To do that more was needed, startingwith the specificationof the objectives forsuch policies.
40. These and the followingtermsforthe stages in improvementof relationsare drawn from theGap, and in turnfromCraig and George, Forceand Statecraft, George, Bridging p. 250, Fig. 8. 41. PresidentBush, commencementaddress at Texas A&M University, May 12, 1989. 42. SecretaryBaker, address to the Foreign Policy Association, New York, October 16, 1989.

18:4 1162 International Security

OBJECTIVES

of policy objectives should converta general sense of the The formulation national interest,a "non-operational goal," into a prioritizedagenda for action. It is the phase which often gives such expressions of interesttheir The objectives say what the policyis supposed to accomplish. real content.43 They are concreteand give operational guidance.44 of a problem. It Specifyingobjectives goes beyond mere identification forcespolicymakersto define what the problem really is by imaginingthe and analysts to define "success," policymakers problem's solution. In trying redefinethe problem and at the same time open up the specifictradeoffs between certain objectives and the obstacles likely to be encountered in achieving them. frommere desires: I may want a new Objectives are thereforedifferent Ferrariautomobile and encourage others to buy one forme, but that does not mean I have actually set myselfthe objectiveof purchasingone. Desires with Iraq in 1990and objectivesare frequently confused. During the conflict 91, the United States deliberatelyconsidered whether to extend American rid ofSaddam Hussein by moving and UN policyobjectivesto include getting on to Baghdad and installinga new government.Many observers feltthat the failureto remove Saddam meant that "an importantpoliticalobjective" was "not accomplished."45But when desires are confused with objectives, the resultis a brew called "problematicpreferences."46
43. "From the Monroe Doctrine to the CarterDoctrine, students of Americanforeignrelations doctrinesand perhaps the cumulativeimpact have been puzzled about the meaning of particular A common featureof these doctrineshas of them collectively forthe conduct of foreignaffairs. been theirhighlyambivalent and flexiblecharacter."Cecil V. Crabb, The Doctrines ofAmerican Foreign Policy:TheirMeaning,Role, and Future(Baton Rouge: Louisiana State UniversityPress, 1982), p. 394. 44. See Plischke, Foreign Relations: AnalysisofitsAnatomy, p. 111: "Foreign policy objectives are in the pursuance the concreteand actionable aims of the nation, decided upon by governments, of national interests. . . for the attainmentof which foreignpolicies . . . are formulatedand implemented." 45. George, Bridging theGap, p. 91. George laternotes, however,thatthe Bush administration's "refusal to escalate its political and military objectives" was the result of a deliberate decision to "strictly limitits objectives." Ibid., p. 95. 46. On "problematic preferences" see Michael Cohen, James March, and Johan Olsen, "A ScienceQuarterly, Garbage Can Model of Organizational Choice," Administrative Vol. 17 (March 1972), pp. 1-25; Charles Lindblom, "The Science of Muddling Through," PublicAdministration Review, Vol. 14 (Spring 1959), pp. 79-88; Kingdon, Agendas, Alternatives, and PublicPolicies, p. 89. For more on the distinction between aspirationsand policy objectives,see Wolfers,Discordand Collaboration, p. 71.

Foreign PolicyEngineering 1163

The Bush administration chose not to set the removal of Saddam as one of the concreterequirementsof war or peace. Officialsundoubtedly hoped Saddam would fall from power. When asked, they admitted this. They encouraged others to bringhim down. That did not make it an "objective." If any American president of the last thirty years was asked whether he hoped Castro's dictatorshipcould be brought down, the answer would be yes. They have regularlyencouraged the Cuban people to give the same answer. But no presidentsince Kennedy has set in motion a policy with the operational objective of securingthe overthrowof Castro's dictatorship. Such opportunities for confusion underscore the need for precision in formulating objectives. Precision forces real preferencesinto the open for timelydebate. When the Kennedy administration persuaded Khrushchevto remove nuclear missiles from Cuba in 1962, it was vague about whether nuclear-capable IL-28 bombers were included in the deal. The subsequent clash over removal of the IL-28s almost caused another major crisisin November 1962.47 Precision also clarifiesthe true scope of a policy. Somalia is an obvious example. The original policy objectives set for American interventionin December 1992 were "to establish a secure environmentfor humanitarian "48 reliefoperations. Yet when the UN peacekeeping force was created in March 1993, its policy objectiveswere breathtakingly broad. The forcewould "assume responsibility forthe consolidation,expansion and maintenanceof a secure environment throughoutSomalia." The forceswould implementan arms embargo on the parties "fromwithin Somalia" and help enforceUN "demands" that all Somali parties, "including movements and factions," disarmand complywiththe politicalpromisestheyhad made to each other.49
47. Raymond L. Garthoff, Reflections on the Cuban Missile Crisis,rev. ed. (Washington,D.C.: BrookingsInstitution, 1989), pp. 104-115. 48. U.S. forces would "address a major humanitariancalamity,avert related threatsto international peace and security,and protectthe safetyof Americans and others engaged in relief operations." Letter fromPresident Bush to congressional leaders, December 10, 1992, in Department of StateDispatch,Vol. 3, No. 50 (December 14, 1992), p. 877; see also United Nations SecurityCouncil Resolution No. 794 (December 3, 1992), para 10. 49. United Nations SecurityCouncil ResolutionNo. 814 (March 26, 1993). The "hunt" forSomali National Alliance leader Aideed, ordered in June1993 by the UN SecurityCouncil, was simply the logical corollaryof the objectives established in March, as the Council noted in Resolution No. 837 (June 6, 1993), which affirmed the force'sduty to disarm all parties and "establish the effective authorityof UNOSOM II throughoutSomalia." These objectives were held up as a before the Senate Foreign "model worth promoting"by Undersecretary of State Peter Tarnoff

International Security 18:4 1164

Close attentionto the precise formulation of policy objectives could have opened the way to expert analysis, asking questions such as: What is the past experiencewith multinational undertakingswith similarobjectivesin a country tornby civilstrife? What variablesare likelyto determinethe success or failurein attainingsuch objectives? Do these warrantrenewed attention to other components of the policy, such as coercive strategyand military forcedesign? The questionerwould quicklyhave noticedthatAmericanpolicyobjectives in, forexample, Lebanon in 1983, were strikingly similarto those adopted forthe multinationalforcein Somalia ten years later. Other analyses of UN peace enforcementexperience could have helped in identifying some key variables.50But even top subcabinet officialsreportedlydid not bother to deliberateabout the March reformulation of objectives, nor did outside experts appear to appreciate the vital significance of this new policy component.51

Some mightargue withthiscall forprecision,and praise instead the virtues ofvagueness in avoiding unwanted controversy. Vagueness can indeed serve to deflectboth controversy and attention froma passive policy. Leaders may also not care enough about the problem to defineits solution,or mighthave not made up their minds just what problem it is they want to solve. But active policies are endangered whenever governments in theirprivatecouncils substituteproblematicpreferences forconcrete,explicitobjectives.

STRATEGY

Strategies are those mechanisms, those theories of the relation between government action and the behavior of others,by which it is hoped thatthe
Relations Committee on July29, 1993. Department of StateDispatch,Vol. 4, No. 32 (August 9, 1993), p. 567. 50. U.S. objectives for the multinationalforcein Lebanon were to facilitate the withdrawal of foreignforces "and to assist the Lebanese Central Governmentin reestablishingits control throughout Lebanon." Departmentof State press statement, March 15, 1983, in American Foreign Policy: CurrentDocuments(Washington, D.C.: U.S. GovernmentPrintingOffice [U.S. GPO], 1983), p. 747. For fine case studies of UN peacekeeping, including the peace enforcement operationin the Congo, see WilliamJ.Durch, ed., TheEvolution ofUN Peacekeeping: Case Studies and Comparative Analysis(New York: St. Martin's Press, 1993). 51. "Asked iftherewere any Cabinet or deputy-levelmeetingsdevoted to discussing the March 'nation-building'resolution,or the implicationsit would have on the ground in Somalia, three senior administration officialscould recall none." Thomas W. Lippman and Barton Gellman, "How Somalia Started Biting the Hand That Fed It," Washington Post NationalWeekly Edition, October 18-24, 1993, p. 14.

Foreign PolicyEngineering 1165

policywill act upon its object to produce the desired result.They are theories of persuasion. Each strategy,then, is an analyticallydistinctivepathway toward the policy objectivesbeing sought. The strategiescan include appeasement, deterrence,payoffsor reassurances, tit-for-tat trades of rewards and punishments, marshaling "world opinion" to impose diplomatic isolation, and the use of economic or other formsof coercive diplomacy,and can extend to military compellence or even punitiveaction. One scholarhas catalogued twenty-five different "programs" or "procedures" for internationalconflictresolution.52 The menu of diplomatic strategiesalso includes the choice of which internationalnorms or decision rules will apply. Althoughthe operationof these channels will then fallinto the "politicsstream,"officials can use institutional analysis to decide which kinds of politicalprocesses are most likelyto attaintheirobjectives.53 Designing a policyto carryforward a strategy-e.g., determining the scope and enforcement of economic sanctions-is a different task fromthe analytical conceptualization of the strategy.Although design and strategiceffectiveness are interdependent,separating them analytically allows the expert
52. No one has done more to conceptualize these alternatives than George himself.In addition to Bridging the Gap, see George, Forceful Persuasion;George, The Limitsof CoerciveDiplomacy; and Practice; and Gordon A. George and Smoke, Deterrence in American ForeignPolicy:Theory Craig and Alexander George, Forceand Statecraft, Part Two. For the "twenty-five" approaches, see Lynn Wagner,"Processes forImpasse Resolution,"citedin George, Bridging theGap,pp. 161162, n. 12. For illustrations of other types of strategies,see Ariel E. Levite, Bruce W. Jentleson, and Larry Berman, eds., Foreign MilitaryIntervention: The Dynamicsof Protracted Conflict (New York:Columbia University Press, 1992); Durch, TheDynamics ofUN Peacekeeping; J.L.Richardson, "New Perspectives on Appeasement: Some Implications for InternationalRelations," World Politics,Vol. 40, No. 3 (April 1988), p. 312; Stephen Rock, WhyPeace BreaksOut: GreatPower Rapprochement in Historical Perspective (Chapel Hill: Universityof North Carolina Press, 1989); and Janice Gross Stein, "Reassurance in InternationalConflictManagement," PoliticalScience Quarterly, Vol. 106, No. 3 (Fall 1991), pp. 431-451. 53. On the influenceof internationalnorms and institutions in shaping strategicchoices, see Friedrich Kratochwil,Rules,Norms, and Decisions:On theConditions and LegalReasoning ofPractical in International Relationsand DomesticAffairs (Cambridge: Cambridge UniversityPress, 1989), pp. 10-11. On the manipulation of decision settings,see Ze'ev Maoz, "Framing the National Interest:The Manipulation of Foreign Policy Decisions in Group Settings,"WorldPolitics,Vol. 43, No. 1 (October 1990), pp. 77-110. Maoz looks at how manipulation of domestic political processes explains nationaldecisions,but his argumentcan be applied to international diplomacy as well. Examples are the January-February 1990 choice to negotiateGermanunification through the Two-Plus-Fourprocess ratherthan Four Power talks (the preference of the USSR and some West Europeans) or a CSCE conference (Genscher's proposal); or the American decision to bypass the Geneva multilateral forumfornegotiatingArab-Israelipeace in favorof reliance on U.S.-Egyptian cooperation and trilateral diplomacy, which culminated in the September 1978 to theTable:The Camp David summit. See Quandt, PeaceProcess; JaniceGross Stein, ed., Getting Processof International Prenegotiation (Baltimore: Johns Hopkins UniversityPress, 1989); and WilliamH. Riker,TheArtofPolitical Manipulation (New Haven: Yale University Press, 1986).

International Security 18:4 1166

to ask more fundamentalquestions about the strategic concept,distinct from critiquesabout the particularway a strategy's potentialis being realized.
DESIGN

Withobjectives set and a notion of the path the policy should followtoward these goals, the policymakermust decide just what the government is going to do. Design occurs when a strategy is convertedinto operationalplans that specifyjust what should happen in the real world. Returning to the case of U.S. policy toward Iraq in 1989-90, the diplomatic strategyfor moderating Iraqi behavior through constructiveengagement rested almost entirelyon a particulardesign, the extension of creditguarantees for agriculturalpurchases (CCC). But when an AgricultureDepartment investigationturned up possible minor irregularities in Iraqi transactions backed by CCC guarantees, the resultwas that in May 1990 the CCC design foundered in the wake thrown up by the investigationsand the turbulencecaused by otherquestionable Iraqi behavior.No alternative design was available thatcould have carriedforward the old strategy of constructive therefore found itself, engagement.The Bush administration duringthe next two months, standing inertly by an emptystrategy. Thus, the details of design matter.It is no wonder that, in the field of engineering,a large portionof available talentworks on the preliminary and detailed design of products. It is surprising,though, that no equivalent recognitionexists forthe nature of policy design, a challenge which encompasses but extends well beyond the empiricaland economic tools highlighted in the policy analysis literature. The significanceof design work is illustratedin the story of the U.S. deploymentof intermediate-range nuclear forcesto Europe. Responding to the threatperceived fromSoviet nuclear forcesin Europe, in 1978 U.S. and NATO defense officialschose a strategyof deterrence of the USSR and reassurance of America's allies throughdeploymentof new American militaryforcesto Europe. The preliminary design forthis deploymentwas U.S. long-rangetheaternuclear forces.5 Then detailed design began. By April 1979, this phase of work settledthe general nature of the modernization,the kinds of missiles that would be
54. The best accounts are Raymond L. Garthoff, Detente and Confrontation: American-Soviet RelationsfromNixon to Reagan (Washington, D.C.: Brookings, 1985), pp. 849-859; and David N. Schwartz,NATO's NuclearDilemmas(Washington,D.C.: Brookings,1983), pp. 194-225.

Foreign PolicyEngineering 1167

deployed (ground-launchednot sea-launched systems,and using both cruise and ballisticmissiles), the numberof missiles to be deployed (200-600), and that a number of allied countries had to accept deployment, not just the Germans. All NATO allies, and especially the deploying countries,had to approve the final design because they were indispensable to implementing it, given the normsand rules governingNATO as an institution and U.S. deployments in allied countries. The Europeans, led by Helmut Schmidt,persuaded the Americans early in 1979 to modify the original strategyand include an additional concept: parallel arms controlefforts. The new arms controlstrategy then also had to be filledout witha specific design. The Americansused ad hocNATO institutions theychairedas settings forrefining and then consolidatingallied agreementupon both the deployment and the arms controlelements of the policy design.55They were then able to use the forcingevent of a special December 1979 meetingof NATO foreignand defense ministersin order to cement agreementto a complete integrateddesign: the "dual-track"policy. The arms controlelement of the design had anotherdesign layerbeneath it: fiveprincipleswere laid out to guide the U.S. negotiating positionin talks with the Soviets, specifyingthe procedures and appropriate outcome for these talks. Each of these principles,in turn,required detailed designs of its own. Thus policy design, like otherpolicymaking components,can have several layers, just as a general blueprintfor a product can be followed by more detailed blueprints of its individual parts. The components of a primary policy can also spawn secondary or tertiary policies, each of which includes all the components of policymakingbut on a smaller scale. For example, as it sought to win allied agreementto the overallpolicy design fordeployment ofnew U.S. nuclear forces,the United States had to craft subordinatepolicies to deal with the concernsof some of the countriesacceptingthe deployment. One of these was the promise to remove 1000 existingnuclear systemsfrom the NATO stockpile in Europe in order to hold fragileBelgian and Dutch
support.56 55. These were the High Level Group, chaired by a U.S. assistant secretary of defense (David McGiffert), and the Special Group, chaired by a U.S. assistant secretaryof state (Leslie Gelb, then Reginald Bartholomew). See Garthoff, Detente and Confrontation, pp. 860-863; Schwartz, NATO's NuclearDilemmas, pp, 225-232. 56. Garthoff, Detente and Confrontation, p. 866 and note 34; Schwartz,NATO's NuclearDilemmas,

International Security 18:4 1168

Not all policy designs are so elaborate. The presentation of designs offered here, with layers laid out so neatly one beneath another,also imposes an appearance of order on a much messier reality.Others might define the layers differently or not even see them at all. Thinkingabout design, and layers,is no more than an analyticaltool to help one see the structure of the policy and the relationshipbetween its parts. This also makes it easier to compare a policy against its rivals or its predecessors.
IMPLEMENTATION, MAINTENANCE, AND REVIEW

Policy, like any product, must be produced, maintained, and eventually eitherreplaced or discarded. Policies are usually made to be implemented. Design ends and implementation the content begins when those specifying of the policy actually begin doing, or delegate the task of doing, what is to be done. Implementationcan only begin when all parties who must carry out a policy have agreed to the design-the workplan. "One of the oldest topics in the study of organizations"is that "policy as implemented often seems different frompolicy as adopted."57 In 1980, for example, the White House decided to use a military of directaction strategy to end the Iranian hostage crisis. The president and his advisers carefully designed the elements of the "rescue" policy, specifying numerous requirements. Even so, "the decision to attemptthe rescue is not the same thingas the plan to employ eight helicopters, train in a certain way, decentralize command, and so on. Decisions on such key aspects of the plan thatwas to be 'carriedout' were themselvesdecided in the process of implementation."58 Similarly,organizational routines may have led Soviet forces deploying to Cuba in 1962 to bringalong tacticalnuclear weapons may not have been part of Khrushchev's policy design, but these operatingprocedures could have had fateful consequences forthe entireworld.59
p. 238. A generic diagram of just such a multilayered"foreignpolicy complex" is offeredin Plischke,Foreign Relations: Analysisofits Anatomy, p. 136. 57. Vicki Eaton Baier,JamesG. March, and Harald Saetren, "Implementation and Ambiguity," in March, Decisionsand Organizations, p. 150. The general point is covered well in Jeffrey L. Pressman and Aaron Wildavsky, 3rd ed. (Berkeley: Implementation, of California University Press, 1984); Wilson, Bureaucracy; and an extensive literature in the domestic policy arena. The comparable literatureon foreignpolicy implementationis modest, although it has been decades since Allison and others highlightedthe issue. See e.g., Steve Smithand Michael Clarke, eds., Foreign PolicyImplementation (London: George Allen and Unwin, 1985). 58. Michael Clarke and Steve Smith, "Conclusion," in Smith and Clarke, Foreign PolicyImplementation, p. 168. 59. BruceJ.Allynand JamesG. Blight,"Closer Than We Knew" (letter), New York October Times,

Foreign PolicyEngineering 1169

Once it is recognized as a distinctcomponentof policymaking, implementationtoo can be subjected to systematic analysis. Studentsofdomesticpolicy have already observed that obstacles in implementation tend to arise from: (1) the operational demands implied by a particulardesign; (2) the nature and availabilityof the resources used in implementingthe design; and (3) the need to share authoritywith or retain the support of the other actors involved in implementation.60 Where implementation fails, the failuremay oblige the parties to develop a new design. For example, in implementingthe Egypt-Israelpeace treaty, the planned United Nations peacekeeping forceforSinai could not be created. So the United States established its own peacekeeping forceto take the UN's place. Once a policy has been put in place, it needs to be maintained. For example, as the war against Iraq ended in 1991, coalitionpolicies continued to evolve. A new set of objectives was formulatedforpostwar supervision of Iraq and a strategy was selected forwinningIraqi agreement,namely,the coercive use of ultimatacrediblythreatening the renewal of hostilities.The design forthe supervision of Iraq went forward,includingthe creationof a new UN organization, and then the design was implemented. As of early 1994, it was stillbeing maintainedwith money,manpower,and the political will and military readiness to keep the coercivestrategy credibleand effective. Most activityin bilateral relations or in internationalorganizations falls into the component of maintenance.It is the stuff of routinediplomacy,with the hope of keeping a policy comfortably settledin the "maintenance"mode. Policies thatare not effectively maintainedmightturninto problemsthatstir national policymakersinto developing a new policy. Naturally,top officials hope to avoid such exertions.This is the concept former Secretaryof State George Shultz refersto as "gardening." It is, he explains, "one of the most underrated aspects of diplomacy" because "the way to keep weeds from and in theirearlystages."61 overwhelming you is to deal withthemconstantly

29, 1992. There is still some dispute about whether the weapons were there, and about what authority local commanders had to use them. 60. This list is adapted fromGordon Chase, "Implementinga Human Services Program:How Hard Will It Be?" Public Policy,Vol. 27, No. 4 (Fall 1979), pp. 385-435. Chase also identifies fifteen areas to receive special attentionand forty-four "factorsforconsideration"in a model to make predictionsabout the prospects forimnplementation of a given program. 61. George P. Shultz, Turmoil 'and Triumph: My Years as Secretary of State (New York: Charles Scribner'sSons, 1993), p. 128.

Security 18:4 1170 International

The finalstep in a sound policy is periodicreview. Problemsin implementation, for example, can lead to the development of entirelynew policies, not just a new design. The failureto implementthe comprehensivepeace settlement dimension of the Camp David Accords led the Reagan administration to develop a new policy to deal withthe issue of Palestinianautonomy in the late summer and fall of 1982, with a different formulation of national and a new policy design. objectives,a different diplomaticstrategy, Though the need to review policies from time to time is intellectually irreproachable, governmentpolicy reviews have a sad reputation.They are usually perceived eitheras pickingup the pieces aftera policy disasteror as a plodding way for bureaucracies to assure new political leaders that they are indeed doing the right thing.HenryKissingereven used "policyreviews" as a way of keeping bureaucracies busy while the "real" policies were being made elsewhere. There is obviously some need to be able to review policies withoutwaiting fora severe crisis. Policy toward Iraq was reviewed in 1989 and again in the spring of 1990, but the fundamentalobjectives and strategicconcept were not strongly challenged. Yet thereare cases of success. Again, disaggregating this componentallows formore systematic analysis. An obvious question is: What are the criteria forevaluation? The characteristic policy review simultaneouslyreviews the extentof any problems, reflectson the national interest,and considers policy objectives, and design, in what amounts to a large, almost indigestible,lump. strategy, Breakingout all the possible questions only makes mattersworse. Review documents containinglists of scores of questions under numerous different headings, usually circulated at the beginning of new administrations, are viewed by bureaucracies (and indeed by the harried new appointees) with varyingmixturesof disgust and horror.This is as much because of concern about the inefficient expenditure of scarce energy as fear that cherished policies will be lost. Analysis can therefore help, in this instance,by narrowing the issues to be reviewed, selectingkey criteria forevaluation,and giving the policy review a clear set of focal points.

Conclusion
With his new book, Bridging theGap, Alexander George poses a challenge both to policymakersand to academics. Can theyfinda common and informative medium of communication?It is hard, in the best of circumstances,

| 171 Foreign PolicyEngineering

forpolicymakersto step back and thinkanalyticallyabout theirwork. The numberof officials involved in any given foreign policy,fromtop to bottom, can be surprisingly small. For example, fewer than a dozen people below the cabinet level were deeply involved in managing the U.S. diplomacy associated with German unification.Even if they had known how to ask good analyticalquestions, none of them would have had the time to write detailed conceptual papers trying to answer them.Analystsofthe intelligence communityshun such work too, because they see theirmain job as lying mainlyin the streamof "problem recognition,"not "policy engineering." The comparative advantage of academics is that they have the general knowledge, disciplinarytraining,and opportunityto provide the needed analysis. They may have to look forways to offer help. "If policy makersare unclear about what they expect fromresearch, researchershave to ferret, guess, and improvise."62 But this potential comparativeadvantage has not been adequately realized in foreign policymaking. Career officials have found will offer too often that, "good social scientists,unless watched carefully, guesses, personal opinions, and politicalideology under the guise of 'expert advice'. "63 A comprehensivepictureof the componentsof foreign policymaking helps one to spot the opportunitiesforanalyticalimprovement. The task now is to work these ideas into a broader literature, synthesizedinto texts,and carried over into academic and professionaltraining.Once both officials and academics have internalizedthese patternsof thinking,they mightask better questions and receive more useful answers. The array of questions is not impossibly long. Dean Rusk once spoke wistfully of the idea of a checklistof questions, like the checklista pilot runs throughbefore launching a plane on its flight.Even combat pilots under pressure forimmediate takeoff go throughat least a dozen questions; commercialfliershave lists with hundreds of items, and formalprocedures for being sure theyare asked. Surelyit is not too much to ask thatpolicymakers display equal caution beforetheircountry"takes offon a policy."64
62. Carol H. Weiss, "Improving the Linkage Between Social Research and Public Policy," in Lynn, Knowledge and Policy,p. 44. 63. Wilson, "Social Science and Public Policy: A Personal Note," p. 91. 64. Dean Rusk, "The Formulationof Foreign Policy," remarksto officers of the Departmentof State, February20, 1961, in American Foreign Policy:Current Documents (Washington,D.C.: U.S. GPO, 1961), p. 29.