The Decision to Attack Iraq: a Noncompensatory Theory of Decision Making Author(S): Alex Mintz Source: the Journal of Conflict Resolution, Vol

The Decision to Attack Iraq: A Noncompensatory Theory of Decision Making Author(s): Alex Mintz Source: The Journal of Conflict Resolution, Vol. 37, No. 4 (Dec., 1993), pp. 595-618 Published by: Sage Publications, Inc. Stable URL: http://www.jstor.org/stable/174541 . Accessed: 19/03/2013 12:48

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 [email protected].

Sage Publications, Inc. is collaborating with JSTOR to digitize, preserve and extend access to The Journal of Conflict Resolution.

http://www.jstor.org

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions The Decision to Attack Iraq

A NONCOMPENSATORY THEORY OF DECISION MAKING

ALEX MINTZ TexasA&M University

This article attemptsto explain the decision to use force againstIraq using a noncompensatory theory of decision making. Ratherthan choosing an alternativethat maximizes utility on the basis of a holistic comparisonprocess as suggestedby the expectedutility model, or selecting an alternativethat "satisfices"a certain criterion as predictedby the cybernetic model, the noncompensatorytheory suggests that decisions on the use of force are often madebased on the rejectionof undesirablealternatives on the basis of one, or at most a few, criteria.

This article attemptsto explain the decision to use force againstIraq using a noncompensatory,multidimensional theory of decision making. This theory highlights the role of distinctcontextual dimensions associated with the decision to use force and integratesnoncompensatory strategies of decision making into this multidimensionalframework. At the core of this theory is the assumption that policy makers are confronted with a set of feasible alternativesand a set of contextualdimensions that enter into their calculus of decision making. Associated with each policy alternative,there are consequences thatfollow from the pursuitof the alternative.If a certainalternative scores low on a given dimension,then a high score on anotherdimension cannot "compensate"for it, and hence the alternativeis rejected. A decision situation typically consists of a set of "feasible alternatives A = (a,. ..., am) and a set of attributes or evaluators of the alternatives (X1, ..., Xv).Associated with each alternativea in A, there is a corresponding consequence (X,(a), X2(a) ...., Xn(a)) in the n-dimensional consequence space X = X1 x X2 x... X," (Sage 1990, 345). Rather than choosing an

AUTHOR'SNOTE: An earlierversion of this articlewas presentedat the annualmeeting of the InternationalStudies Association,Acapulco, Mexico, March23-27, 1993. I am gratefulto NehemiaGeva for manyinvaluable discussions. JOURNALOF CONFLICT RESOLUTION, Vol. 37 No. 4, December1993 595-618 ? 1993Sage Publications, Inc. 595

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 596 JOURNALOF CONFLICTRESOLUTION alternative that maximizes utility on the basis of a holistic comparison process, or selecting an alternativethat "satisfices"a certain criterion,the theory suggests that foreign policy decisions are often made based on the rejection of undesirablealternatives on the basis of one, or at most a few, criteria. This theory differs significantly from the rationalactor-expected utility model of the use of force (see Bueno de Mesquita 1981, 1985) and the cyberneticsatisficing political use of force model (see Ostromand Job 1986). WhereasBueno de Mesquitaand Lalman (1990) and Ostromand Job (1986) used compensatorystrategies of decisionmaking, the model presentedin this article uses a noncompensatorystrategy.

1. COMPENSATORYSTRATEGIES OF DECISION MAKING

According to Bruce Bueno de Mesquita(1981, 1984, 1985), nations are led by rational,forward-looking, expected-utility-maximizing leaders. The strategies these leaders select are a function of "the values they attach to alternativeoutcomes andthe beliefs they hold regardinghow theiradversary will respond to their strategicdecisions" (Bueno de Mesquita and Lalman 1990, 750). Decision makersevaluate the costs and benefits associatedwith each alternative to obtain "the largest net gain (expected utility) at an acceptablelevel of risk"(Bueno de Mesquita 1984, 228). The rationalist, microeconomic, expected utility approachto decision making is a powerful one, but it has limitations. Cognitive psychologists (e.g., Schwab, Olian-Gotlieb,and Heneman 1979; Klein 1989; Mitchell and Beach 1990) and students of behavioralorganization theory (e.g., Simon 1985) attributeto the expected utility and other analyticaldecision-making strategies the requirementfor extensive processing time, cognitive effort, concentration,and skills that in many cases are not available, especially undertime pressuresand rapidly changing conditions, and in othercases are not worthy to the decision. Accordingly,these scholars argue that decision makersseldom maximize utility. Accordingto the boundedrational model, of which the cyberneticsatisficing model of decision makingis one type (see Simon 1985; Steinbruner1974; Ostromand Job 1986), policymakersoperate under constraints while search- ing for an acceptable outcome (Dougherty and Pfaltzgraff 1990). These limitations or bounds stem from deficits in the informationbase and from limitationof processingcapabilities (e.g., cognitive and motivational).Con- sequently,the decision process is less comprehensivebecause only a subset of alternativesand dimensionsis considered.

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions Mintz/NONCOMPENSATORY THEORY 597

Expected-utility studies (Bueno de Mesquita and Lalman 1990) and cyberetic models of the use of force (Ostromand Job 1986) have relied on what is known in the informationprocessing/cognitive psychology literature as the "compensatory"strategy of decision making (Ford et al. 1990). According to this model, in choice situations,if a certain alternative(e.g., use of force) is low on a given dimension, then a high score on another dimension can compensatefor it, and hence the alternativeis adopted.This is reflectedin the elegant mathematicalequations in Bueno de Mesquitaand Lalman(1990) and Ostromand Job (1986). In these studies, a high score on one dimension (e.g., military) can compensate for a low score on another dimension (e.g., political). These are combined to maximize (Bueno de Mesquitaand Lalman 1990) or "satisfy"(Ostrom and Job 1986) utility. Two models are used in describingthe compensatorystrategy. According to the linear model,each dimensionfor a decision alternativeis given a value and the dimensionsare combinedadditively to producean overall value for each alternative.The decision makercompares the alternativeson the basis of these overall values and chooses the alternativewith the highest value (Einhornand Hogarth1981; Fordet al. 1990). Accordingto this model, then, a high score on the militaryor internationaldimension can "compensate"for a low score on the politicalvariable and vice versabecause the political leader makes a decision to use force based on the overall score. The second model is the additivedifference model, by which alternatives are compared on each dimension and differences across dimensions are summed.This summationresults in a preferencefor one decision alternative (Olshavsky 1979). Compensatorymodels are additive.Compensatory processes are marked by interdimensionalsearch (i.e., a search within an alternative) using a constantamount of informationper alternative(Billings and Scherer 1988). The additive difference model is compensatorybut proceeds from an in- tradimensionalsearch pattern (i.e., across the portfolioof alternatives). Using a compensatorystrategy, the decision makermay tradeoff a high value on one dimension of an alternativewith a low value on another dimension. The essence of this model is that the various components of a multidimensionalobject are combined for a "score"(e.g., "netgain," "critical value,""satisfactory score," etc.) and contributeindependently to its overall worth or utility (Einhornand Hogarth 1981). With a few exceptions, linear equationshave been used to approximatethe cognitive processes underlying decision making. Although the expected utility and other analytic models have produced accuratepredictions of policy decisions and outcomesin numerousinstances (see Bueno de Mesquita 1981, 1984, 1985), such models seldom capturethe

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 598 JOURNALOF CONFLICTRESOLUTION underlyingcognitive processes involved in decision making.An alternative theory of foreign policy decision makingis presentedbelow.

2. THE NONCOMPENSATORY DECISION-MAKING THEORY OF THE USE OF FORCE

A growing number of studies of foreign policy decision making have shown that policymakers adopt cognitive shortcuts to rational decision making (see Jervis 1976, 1985;Jervis, Lebow, and Stein 1985; George 1980; Holsti 1990; Herek, Janis, and Huth 1987; Maoz 1990; Vertzberger1990). George(1980), for example,points out thatin the face of valuecomplexity and uncertainty,policymakers resort to a varietyof simple decision rules.Among these rules are the use of historical analogies (see also Neustadt and May 1986), the use of incrementalpolicies (Wildavsky1974), "consensuspolitics" (George 1980), and differentmodes and mechanismsof defense avoidance (e.g., "bolstering"or "spreadingthe alternatives"and procrastination-see George 1980, chap. 2). According to George (1980), policymakersattempt to avoid value tradeoffsby using "cognitiverestructioning" and by "deval- uating ... one of the values or intereststhat [thepolicymaker] or othersclose to him hold" (1980, 33). Jervis (1985) attemptsto understandthe biases thatinfluence information processingand decision making.According to Jervis,both unmotivated (e.g., cognitive predispositions)and motivated (i.e., affect-driven)biases influence decision making.Implicit in the work of Jervis(1976), Holsti (1990), Jervis, Lebow, and Stein (1985), and George (1980) is the assumptionthat there are cognitive limits on the rationalist,microeconomic theory of decision making. These limits have profoundimplications on foreign policy decision making. No dominant model of the use of force has emerged, however, as an alternativeto the expected utility or cyberetic/satisficing models of the use of force. The noncompensatorydecision-making model presented in this article suggests thatdecisions to use force aretypically based not on a compensatory calculus of decision making, but on a noncompensatorystrategy; that is, in a choice situation,if a certainalternative is unacceptableon a given dimension (e.g., it is unacceptable politically), then a high score on another dimension(e.g., the military)cannot compensate/counteract for it, andhence the alternativeis eliminated. The model capturesthe nonholistic natureof the foreign policy decision-makingprocess. It focuses on a very restrictive set of alternativesand dimensions.Moreover, it recognizes the noncompensatory natureof the decision process. Consequently,rather than relying on

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions Mintz/NONCOMPENSATORY THEORY 599 holistic decision rules (e.g., expected utility theory, subjective expected utility theory,or multiattributeutility theory)that require the evaluationand comparison of all alternatives across different dimensions, the decision makeradopts heuristicdecision rules that do not requiredetailed and com- plicated comparisonsof relevantalternatives, and adoptsor rejects undesir- able alternativeson the basis of one or a few criteria. The noncompensatorydecision-making model employs less cognitively demandingdecision proceduresthan the expected utility or other analytic models. The model is associated with a nonexhaustive decision-making process, implies curvilinearity,and is nonadditive(as a person who suffers from high levels of cholesterol is unlikely to buy himself items rich in cholesterol, even if these are offered at a huge discount). Abelson and Levi (1985) found that decision makers use a variety of choice strategies (see also Payne, Bettman, and Johnson 1988). Whereas compensatorymodels can be appliedto some incidentsof the use of force, a noncompensatorymodel can be appliedto othercases. The use of a particularstrategy of decision making (compensatoryor noncompensatory)is dependenton a numberof factors (such as task complexity or the numberof feasible alternatives).In information-processing parlance, whereas the expected utility and cybernetic/politicaluse of force models are alternative-basedmodels, the model describedin this article is a dimensional or attribute-basedmodel. As Payne, Bettman, and Johnson (1988, 536) correctly pointed out, in alternative-basedprocess models, "multipleattributes of a single alternativeare consideredbefore information about a second alternativeis processed."In a satisficingmodel, a la Ostrom and Job (1986), the alternativethat is above an "acceptable"threshold level is selected. In Bueno de Mesquitaand Lalman (1990), the alternativewith the highest net gain is selected.In contrast,in the noncompensatorydecision- makingmodel presentedin this article,the value of an alternativeon a single dimensionis processedfirst, and alternatives that score below the cutoff value on this dimensionare eliminated. Political leaders review alternativesin light of a political dimension and reject all alternativesthat may damage them politically, then move to a substantivedimension and reject alternativesthat are low on this dimension.1 Tversky's(1972) eliminationby aspects (EBA) theoryimplies a noncompensatorydecision-making strategy.2 EBA will eliminatean alternative"if it

1. Although in most cases politicians do view political supportand survival as the salient factors, there are situationswhere this orderof processingmight be violated. 2. The noncompensatorydecision-making model is not synonymouswith EBA. The con- junctive, disjunctive,and lexicographicdecision rules can also be viewed as noncompensatory (Sage 1990).

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 600 JOURNALOF CONFLICTRESOLUTION is below the cutoff on the first dimensionexamined, even though it may be very good on the remainingdimensions" (Billings and Scherer 1988, 17). EBA involves "searchingall alternativeson the most importantdimension andeliminating those thatdo not have satisfactoryvalues on thatdimension" (Billings andScherer 1988,4). The next mostimportant dimension is selected and the process is continueduntil only one alternativeremains. Because a choice requiresthat one alternativebe selected and the rest rejected(Billings and Scherer 1988, 2), choice involves "lookingfor dissimilarities,meaning those featuresthat distinguish between alternatives (features held in common would be disregarded)"(Billings and Scherer 1988, 4). The EBA model searches intradimensionallyand uses a variableamount of informationper alternative(Payne 1976, 367). A dimensionis selected with a probabilitythat is proportionalto its weight or relativeimportance. Using a noncompensatory strategy,the decision makeris left with fewer alternativesto consider. The noncompensatorydecision-making process does not always con- tinue, however, until only one alternativeis left. There are situationswhere severalremaining alternatives all pass a thresholdvalue (i.e., do not fail) and are then evaluated based on a more analytic processing or by using a lexicographicdecision rule. Even when one alternativeis left (by default), a final refinement of the default choice is typically performedby trying to minimize costs and maximize benefits. The noncompensatorydecision-making model is dimension-based.Di- mensions typically involve groups of similar criteria.The criteriawithin a specific dimensionare processed sequentially to arriveat an overall score on that dimension. A dimension can be conceived as an organizingtheme for related information and variables(T. Ostromet al. 1980). Thus if the decision makeris concernedwith the political implicationsof a decision, then public opinion polls, the leader's popularity,the state of the economy, domestic opposition, and other variables that are related to this general theme may be used to evaluate the consequenceof a chosen alternativeon this organizingtheme. Withinthe EBA context,the decision makerassesses alternativeswith regard to whether they pass or fail a threshold on a given dimension (or OT, organizingtheme). The numberof criteriaand variablesthat are considered for a particularOT may vary across alternatives.One alternativemay be rejected on the basis of only one criterion (e.g., how it scored on public opinion polls), whereas anotheralternative may be rejected or accepted on the basis of polls, editorials,and more. Whatdetermines the sequence of the variablesthat enter the process is usually the availabilityof information,and its structure.If all variablesare available and the sequenceof theirprocessing is determinedby the decision maker,then the items that are consideredmost

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions Mintz/NONCOMPENSATORY THEORY 601 relevant to the theme (highest decrease of uncertainty)will be first to be considered(T. Ostromet al. 1980). Such a process is conductedfor the first OT. The decision maker can then move to the second OT while operating along the same principles.According to the noncompensatorymodel, only relevantand nontrivialOTs (e.g., political, military,etc.) are considered. Foreign policy decision making is often characterizedby decision processes that involve a numberof decisions taken over time (see Maoz 1990), often as steps on an escalatory ladder and typically in an interactive setting. The noncompensatorydecision-making model involves the analysis of alternativespresented at each stage of the process. Decision makers are not necessarily"forward looking," although they are influencedby their own past decisions, which may restricttheir choice set due to increasedsunk costs, cognitive investment, and so on. The greaterthe prospective uncertainty (e.g., regardingthe opponent'sintentions, the outcome of a conflict, etc.), the less the "forwardlooking" tendencies of decisionmakers (Vertzberger 1992). Political leaders almost by definition take into account (explicitly or implicitly)political factors and consequences when makingpolicy decisions. As Morton (1991, 772) pointed out, althoughleaders do care about policy, they areespecially concernedabout the politicalimplications of theiractions. Isard(1992) has similarlyclaimed that in seekingto be or stayin power,or in havinghis/her political party be and remainincumbent, the politicalleader must be primarilyconcerned with constituencysupport. Put otherwise, he is plaguedwith the problem of main- tainingvoter support. This problem is primarilyshort-run and involves constant shiftingwithin a multidimensionalpolicy space.And long-runfactors that shouldbe incorporatedin rationalaction are by andlarge dismissed. (P. 7) We arguethat because organizationalpolitics, bureaucraticpolitics, local politics, and other forms of domestic politics are so pervasive in policy making(Russett 1990a), leaderseliminate options that are below the "cutoff' level on the political dimensionand do not make tradeoffsacross dimensions to compensate for a low score on the political dimension. Thus they use a noncompensatorystrategy. An alternativethat is likely to damagethe political prospects of the leader is rejectedbefore evaluatingthe "score"on the other dimension(s).Thus one set of calculationsthat enters into the political leader's decision-making calculus is political. This dimension has to be satisfied (cannot be eliminated) before other dimensions are considered. Remainingoptions are then evaluatedbased on how they score on a substan- tive dimension (e.g., economic, strategic,diplomatic, legal, social, psychological, and so on).

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 602 JOURNALOF CONFLICTRESOLUTION

The context of the other dimension(s) that enters into the calculus of decision making varies as a function of the substance of the decision. Therefore,it is common to identify at least two distinct (though typically interrelated)dimensions-sets of calculations-that are "nested" in the calculus of decision makingof political leaders. In contrastto othermodels of the use of force, the noncompensatorymodel of decision making suggests thatif the decision to use force is unacceptable politically, it will prevent the decision maker from using force even if the score on the benefit/costratio on the militarydimension is very high (historical examples show that the opposite is not always true, however-leaders had initiatedconflicts when the score on the option to use force was high on the political dimension but low on the militarydimension). Therefore,the relationshipsare not necessarilysymmetrical. Policymakers typically expect the use of force to help them politically (at least in the shortrun) through the "rally 'round the flag" effect, and thus the two dimensions (political and military)are not always orthogonal(independent) of each other.The use of force may increase the leader's utility because it may increase domestic support. Policymakers"weigh" the relativeimportance of each dimension(Xi) of the decision alternative(ai). In the compensatorylinear model, the values across dimensions are summed as utility scores for an alternative.In the additive difference model, the sum of the differences across dimensions should be the greatestfor the alternativethat is being selected. In contrast, the expected gain on a critical dimension should be higher than a certain thresholdlevel in the noncompensatorymodel. In this model, if the expected gain along a critical dimension is below the threshold value, then the alternativeis eliminated. Even if one adoptsthe noncompensatoryapproach, it is still importantto analyze the hierarchyand orderin which dimensionsand/or alternatives are entered into the calculus of decision making. It was shown elsewhere that decision makersonly considerfeasible options, whereasimpractical options are not even raised (Klein 1989). We agree with Bueno de Mesquita(1981) thatthe decision to use force is made by a small group of leaders. We also agree that a crucial contextual dimension underlying many policy decisions is domestic politics, or the desire of politiciansto be popularand get reelected.This is being pursuedby policy makersat an acceptablelevel of risk (Ostromand Job 1986; Russett 1990a). Furthermore,we concurwith John Steinbruner's(1974) assessment thatdecision makersin crisis areconstrained by time pressuresand uncertain outcomes, and are thereforerelying on their inner circle of policy advisors and experts in assessing the threatsand gains associatedwith their decision

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions Mintz/NONCOMPENSATORY THEORY 603

(see also Klein 1989; Mitchell and Beach 1990). One can also incorporate organizationaland bureaucraticaspects of decision making (see Allison 1971) into our conceptual model. These aspects represent instances of political organizing themes during the decision-makingprocess. However, in contrastto all studies of internationalcrises and the use of force,3we use a noncompensatorymodel of decision making that does not allow policymakers to "compensate"for a low score on the political dimension while making decisions. In what follows, we review existing explanationsof the decision to attackIraq and then apply the noncompensatoryutility theory of decision making to the president'sdecision to launch an attackagainst Iraq on January16, 1991 (January17, Baghdadtime).

3. THE DECISION TO ATTACK IRAQ: PREVIOUS EXPLANATIONS

There is a growing literatureon the moves and countermovesthat led to the Gulf Warand the geostrategicconsequences of that war. George (1991) calls the Gulf Waran "inadvertentwar" (because duringthe crisis that led to the Iraqiinvasion of Kuwait,neither side "wantedor expecteda warto result" [p. 567]), Hussein'sinvasion of Kuwaita "miscalculatedfait accompli,"U.S. policy a variantof coercive diplomacythat has been called "agradual turning of the screw" (p. 571), and the processes leading to the war, a diplomatic version of the game of "chicken." Freedmanand Karsh (1991) have also analyzedthe strategiesof bothsides in the war.According to these authors,Hussein's strategiccalculations were shaped by Iraq'seight-year war with Iran (p. 9), whereas U.S. strategyhas been influenced profoundly by the U.S. involvement in the Vietnam War (p. 15). Recent historical analogies (see Neustadt and May 1986) have therefore played an importantrole in shaping U.S. and Iraqi strategies. Freedmanand Karsh(1991) concludedthat the Bush administrationadopted a militarystrategy that would minimize coalition casualties"reinforced by a careful leadershipof the alliance,"whereas Saddam Hussein's strategywas based "on increasing American casualties and sowing division within the coalition" (p. 6). (For a differentinterpretation focusing on Hussein's per-

3. Blechmanand Kaplan(1978) and Mintz andRussett (1992) pointedout that"the political use of force"is not synonymouswith the decisionto employ forcein war. Russett(1990b), Mintz and Russett (1992), Ostromand Job (1986), James and Oneal (1991), and others have all dealt with political use of force cases and decisions. The model presentedin this article may also be applicableto situationswhere a favorablebalance of force gives the decision makera choice of whetheror not to engage in war (e.g., wars between majorand minorpowers).

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 604 JOURNALOF CONFLICTRESOLUTION sonalityand his politicalrise to powerwithin the Ba'athparty, see Miller and Mylroie 1990; othersfocus on the war as an importantelement in Hussein's "state-building"efforts, where external conflict or coercion are used as mechanismsto control conflicting groups in Iraq'sheterogeneous society.) Freedman and Karsh pointed out (1991, 6) that the U.S. was successful because the strength of its military offensive "eased the pressure on the internationalcoalition PresidentBush and Secretaryof State James Baker have founded"(see also Bulloch and Morris 1991). Cooper, Higgott, and Nossal (1991) have analyzed the strategies and motivations of other anti-Iraqcoalition partnersand supporters(e.g., Syria andEgypt, the Gulf states,the WestEuropeans, the Soviet Union, andChina). They pointed out that the behaviorof SaddamHussein himself "increased the costs of defection [from the coalition]" (p. 402) (see also Inman et al. 1992 on this point). Cooper, Higgott, and Nossal (1991) argued,however, that the United States had far fewer genuine followers in the Gulf conflict than "conventionalwisdom" would suggest (p. 408). (For a discussion of political and militarytrends in the post-Gulf WarNear East, see Herrmann 1991.) Kugler's (1991) "realtime" analysis of the Gulf crisis is the most relevant to our study,however. Kugler (1991) has appliedthe rationalactor-expected utility theoryof the use of force to the Gulf crisis. His carefulanalysis of data collected in September-November1990 producedsome accurateforecasts: it showed that Hussein would survivethe war;that the Kurdand Shia would rebel if supportedby outside actors;that the rebellion would collapse; that Hussein "would be defeated in a confrontationwith the U.S.-led coalition; and attemptsby otherforces to negotiatea compromisesolution would fail" (Kugler 1991). His study also disclosed thatHussein had a free handto alter Iraq'spolicy withoutloss of domestic supportmainly because he has had "a secure domesticposition" that permitted wide policy latitude(Kugler 1991). Kugler's analysis of Iraq's political dynamics furtherdisclosed that, given the arrangementof forces, Iraqwas in no positionto initiatemilitary activities against Saudi Arabia or the U.S.-led coalition (1991, 26). According to Kugler,if Husseinwould have followed the patternof unilateralconcessions, the loss for Iraq would have been limited, while maximum benefits were retained.Most important,a disastrouswar could have been avoided (p. 28). However, Kugler's (1991) analysis (based on an assessmentof the distri- bution of forces in the Gulf) predictedonly a low probabilitythat the crisis would lead to a war (p. 28). These 1991 predictions were based on the assumptionsthat (1) the United States could have reached a compromise solution with the help of Saudi Arabia,and (2) that SaddamHussein had a

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions Mintz/NONCOMPENSATORY THEORY 605

"unilateralability to obtain a more favorableoutcome by accommodation than by conflict" (p. 28). While Kugler's expected utility analysis failed to anticipate that the United States would attackIraq, it is unlikely that the cybernetic model of the use of force would have fared better.Whereas in a cyberneticenviron- ment, the decision maker simplifies decisions and focuses on a subset of alternatives,this model allows decision makers to make tradeoffs across dimensions. Furthermore,the cybernetic approach emphasizes not only limited searchof informationbut also hierarchicalprocessing. Thus, using a cybernetic model of the use of force, the order in which alternativesare evaluatedcould affect the alternativechosen. Meetinga satisficingthreshold may lead the decision makerto choose an alternativewith no furtheranalysis of otheravailable alternatives. For example,if the no-use-of-force("containment") strategyhad been evaluatedfirst, based on a satisficing principle,it is likely that because this option was less risky militarily than the "use of force"option, a high score on this dimensioncould have "compensated"for a low score on the political variables.A total "score"may have exceeded, then, a "satisfactory"threshold value and hence, based on a compensatory model, the "containment"alternative could have been adopted. Based on Woodward(1991) andothers, however, there is plentyof evidence to suggest that the containmentstrategy was indeed considered first but eliminated, whereasthe use of force alternativewas approved.Moreover, an analysis of Hussein's options based on a compensatorylogic has predicted a peaceful outcome (see Mintz and Geva forthcoming). In the rest of this article, I offer an alternativeexplanation for President Bush's decision to use force againstIraq. This explanationuses the noncompensatory decision-making theory of the use of force. It is a model of a unitary actor operatingunder imperfect/bounded rationality in an environ- ment characterizedby imperfectinformation (prospective uncertainty).

4. THE DECISION TO ATTACK IRAQ: A NONCOMPENSATORYPERSPECTIVE

Warand peace decisions are rooted not only in internationalpolitics, but also in considerationsof domestic politics (Russett 1990a, chap. 2, 1990b; Mintz andRussett 1992, 179). It obviouslymatters "what the militarybalance of power is, but every political leader must also worry about his base of domestic popularsupport" (Mintz and Russett 1992, 179). Mintz andRussett have thus identified the political dimension and the military dimension as

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 606 JOURNALOF CONFLICTRESOLUTION most influentialin decisions to use force. A decision to use force typically consists of political calculationsbased on conditionsin the domesticpolitical economy and military/strategiccalculations based on an assessment of military capabilities,balance of militaryforces, casualty estimates, and so on. Below, we show thatthe domestic political economy served as a crucial dimensionin the calculus of the decision to use force againstIraq (it has had a powerful impact on U.S. presidents'decisions to use force in the past 200 years; Russett 1990a, chap. 2) and that alternativesthat were below the "satisfactory"level on the political dimension were eliminated. Military/ strategic considerations served as the other contextual dimension in the decision maker'scalculus of decision (they served historicallyas necessary, althoughnot sufficient,conditions to use force). To achieve the policy of forcing SaddamHussein out of Kuwait, three alternativeswere identified:(1) use of force; (2) containment,that is, con- tinue the sanctions and keep the diplomaticpressure on Iraq;and (3) withdrawal.4Mintz and Geva (forthcoming)showed that decisions in the Gulf crisiswere contingent on previousdecisions. For example, the administration's miscalculation of Iraqi intentions in the pre-August 2 invasion led to a tendencyto reject any "passive"or "do nothing"alternatives at a laterstage. Similarly,President Bush's November 8 decision to switch to an offensive posture limited his choice set at a later stage in the crisis, but put him in a "win-win"situation. (Former Chairman of the JointChiefs of Staff General D. Jones claimed in a testimony before the Senate's Committeeon Armed Services [1990], that the very presence of U.S. forces in the Gulf has narrowedU.S. options and the coalition's ability to act with "patientre- solve.") By the end of December,the presidenthad alreadymade a series of such criticalprior decisions. Mintz and Geva (forthcoming)showed thatthe choices made on these (and other) previous occasions conditioned and structuredthe choice madein late December.sIn decidingto use force against Iraq,the presidentfollowed the multidimensional,noncompensatory (EBA) strategy.The political dimension and the military/strategicdimension were

4. Mueller (1992, 8) points out that for many decision makers, the use of force and containmentalternatives were not that differentfrom one anotherbecause of the expectations thatthe use of force would eventuallyfollow containmentif Iraqwould refuseto withdrawfrom Kuwait. One option was thereforepartially nested within another. 5. It is interestingto note that a two-stage, noncompensatorydecision-making process followed most of the previous decisions as well. Such considerationsas U.S. interests in oil supplies and oil markets,U.S. prestige and credibility,strategic concerns regardingU.S. war fighting and pressurefrom U.S. allies (e.g., MargaretThatcher) were reviewed for the decision to deployforces in the PersianGulf in the firstplace, thus narrowingthe numberof options and ensuring that the outcome would not result in a loss of political "capital."The fear that the internationalcoalition would dissolve also influencedBush's decisionto attackIraq immediately afterthe January15 deadline (see Freedmanand Karsh 1991, 41).

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions Mintz/NONCOMPENSATORYTHEORY 607 the crucialOTs influencing the president'sdecision (see Freedmanand Karsh 1991, 39). An interview with Senator William Cohen (R-Maine) on PBS (February7, 1991) has also confirmedthis assessment.

THE POLITICAL DIMENSION

The literatureon the political economy of defense offers several theoret- ical explanationsfor the use of force, each of which can help explain the president'sdecision to eliminate the unilateralwithdrawal and containment strategiesand to use force instead(see below). The politicaleconomy context (dimension)consists of severalfactors. These arederived from the diversion- ary theory of war (see Levy 1989a;Russett 1990a, 1990b); the referendum model (see Williams 1990); the military industrial complex thesis (see Russett and Hanson 1975); the spatial theory of elections (see Hinich and Enelow 1984); and the rally 'round the flag element (see Mueller 1973; Morrow 1991; Brody 1984). (For a comprehensivereview of the literature on the causes of war, see Levy 1989b.) Below, we apply our noncompensatory theory to the president'sdecision to use force againstIraq. By all accounts,the presidenthad made up his mindregarding the decision to attackIraq by December27, 1990. Therewas no comprehensiveevaluation of the alternativesas would be requiredby a compensatoryprocess and there was "little or no process where alternativeswere systematically[and com- paratively]weighed andargued" (Woodward 1991,320), as would have been expected based on an expected utility model. The "withdrawal"strategy would have hurt the presidentbadly because it would have left Mr. Bush vulnerableon both the militaryand political dimensions.It would have been perceived as a failure of U.S. foreign policy (because it would have highlighted the administration's"miscalculation" of Hussein's intentions vis-a- vis Kuwaitin the pre-August2 period)and was thereforeeliminated and not even considered.The politicaleconomy rationale for eliminatingthe containment strategyconsisted of some of the following elements: 1. Presidential popularity. Previous studies of foreign policy decision making showed that the public is more attentive to foreign policy than traditionallydescribed (Aldrich, Sullivan, and Borgida 1989). Leaders'popularity is significantlyincreased in the "job performance"polls in times of internationalcrisis (Russett 1990a). The public gives an emotional backing to the president.This effect is known as the "rally 'round the flag" effect (Mueller 1973; Kernell 1978; Brody 1984). It varies somewhatfor different actions and accordingto how the leader is perceived by the voters. Russett (1990a, 38) pointed out that presidentsmore often use militaryforce when their performancerating is "in the 'critical'40-60 percent range" (when a

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 608 JOURNALOF CONFLICTRESOLUTION boost is most useful) and when recenterosion in thatrating is evident, and is greateramong their supportersthan in the opposition (see also Ostromand Job 1986; Ostromand Simon 1985). Assertive speeches andrelatively short, low-cost uses or threatsof use of militaryforce "almostalways invoke this short-termpopularity increase" (Mintz and Russett 1992, 181).6 Empirical evidence suggests that leaders of democratic countries are likely to use force when their performancerating in public opinion polls is declining (perhapsbecause of the way they are handlingthe economy) and are more likely then to eliminateother policy alternatives.President Bush's performancerating quickly declined in the fall of 1990, largely due to the October fierce budget struggle, "when tax-unfairess and party-of-the-rich debates took a big bite out of [his] job ratings"(Phillips 1991, xii). The president'spopularity ratings dropped to 53% (Jentleson 1992, 66), that is, within the above-mentionedcritical range of public support.There was also a declining willingness of the Americanpeople to supportprolonged deploy- ment of American forces (Jentleson 1992, 66). Problems in defining a domesticagenda have not helpedthe presidenteither. According to Woodward (1991, 42), the president told the chairmanof the Joint Chiefs of Staff, General Colin Powell, that "I don't think there's time politically for [the containment] strategy."Consistent with prospect theory (Kahnemanand Tversky 1979), the presidentwas in the domain of loss and was therefore more likely to take a risk and go to war.7An eliminationprocess therefore guided the president to reject the containmentalternative. According to GeneralPowell (see Woodward1991), therewas no comparativeevaluation of the other alternatives. 2. Thestate of the economy.Governments' popularity is closely relatedto the state of the national economy (Mintz and Russett 1992, 180). In the post-1945 era,U.S. presidentshave been morelikely to use force when faced with economic difficulties (inflation or high unemployment)(Ostrom and Job 1986; Russett 1990a,chap. 2). However,a democraticleader who wishes to raise his level of popular supportby stimulatingthe economy may not always be able to do so (Russett 1990a). Russett (1990b) pointed out that from the perspective of a "rational, self-interestmaximizing" democratic national leader, an alternativeto ma- nipulatingthe economy is the use of force againstforeign adversaries.Under such circumstances,a policy advocatingcompromise will typically not meet a satisfactorycriterion and will therebybe eliminated.

6. Jentleson(1992, 50) pointedout, however,that "theAmerican public is much more likely to supportthe use of force for the restraintrather than the remakingof other governments." 7. Althoughthe presidenthad to make a choice between two risky prospects,it is clear that the attackstrategy involved greatermilitary risks than the alternative.

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions Mintz/NONCOMPENSATORY THEORY 609

Levy (1989a) and Russett (1990b) argue that the use of force can help divertattention from domestic economic adversity.And historically,reduced popularityand a deterioratingeconomy have been two of the most important variables explaining presidents' decisions to use force (Ostrom and Job 1986). The looming recession, the twin deficits, the savings and loan crisis, the collapse of severalmajor banks, the continuedincrease in oil prices (to more than$40 a barrel)due to the situationin the Gulf in the pre-January15 period, the "sunkcosts" associatedwith the deploymentand maintenanceof a very large air, land, and naval force (of 400,000 plus soldiers) on Saudi soil for months, and Saddam Hussein's threatsto blow out Kuwaiti and Saudi oil fields (andBush's expertisein both the oil businessand politics) have led the presidentto concede that he does not have time to wait for the containment strategyto work. This has also guided him, then, in his decision to rejectthat strategy.The situationin the Gulf only temporarilypushed the economy off the frontpage. 3. Partisanpolitics. Mueller (1992, 13) pointedout thatbecause the wars in Korea and Vietnam "were begun under Democratic administrations, Democrats tended to support them more than Republicans."Bendor and Hammond (1992, 314-15) reportedthat during the Cuban missile crisis, decisions were likewise made "in a partisancontext." They recalled the criticismsby Republicanleaders and the administration's"nervousness about the upcoming congressional elections." According to Mueller (1992), the Gulf War was "begun and engineered" by a Republican president, and Republicans,accordingly, were muchmore supportiveof the president'swar policies than Democrats(see Morganand Bickers 1992 for the importance of partisansupport in decisions on the use of force). The Democratic party's leadershipin Congress (Speaker of the House Thomas Foley, MajorityWhip RichardGephart, Majority Leader Senator George Mitchell, and influential senators such as Lloyd Bentsen and Sam Nunn) and most other Democratsin Congress were on record as opposing the use of force againstIraq. They urgedthe presidentto adopt the containment strategy,that is, to give the unprecedentedeconomic sanctions more time to force Hussein out of Kuwait. Consequently,the president had an opportunityto distinguish himself from potential presidentialrivals on a critical issue, to enhance his party'spopularity, to reinforce the image that the Republicanparty protects the Americanpeople with a strongdefense and better manages foreign policy, and to erase the so-called "wimp"image. If the war would indeed be short,decisive, and relatively "low cost," then the presidentwould have an opportunityto humiliatethe Democrats.Although typically,the outcome of a war is not known with certainty,in this case there

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 610 JOURNALOF CONFLICTRESOLUTION was a consensus that the allies would win. In fact, there is substantial evidence that the presidenthimself had thoughtso (see below), and that he was able "to pull off his warbefore the oppositioncould get fully organized" (Mueller 1992, 18). It was sufficientto see how decisive the Israelivictories were againstvarious coalitions of Arab states when Israeltook the initiative (in the Sinai campaignof 1956, in the Six-Day Warof 1967, and on other occasions). 4. Themilitary industrial complex thesis. Despite notabledivisions within the capitalistclass (see Russett and Hanson 1975), capitalistinterests, especially those of militaryindustry, seldom oppose acts of militaryforce against foreign adversaries,especially when the militaryindustrial sector is threat- ened by severe cuts in militarybudgets and contractawards, as was the case in 1990. The reduced threatof the Soviet Union to Westernsecurity, the administration'splans to withdrawforces fromEurope, to slash the defense budget, to sign unprecedentedarms control agreements with the Soviets, to cut major weapons systems, and to close militarybases have obviously threatenedthe militaryindustrial sector. Drew (1991, 190) reportsthat defense contractors have even lobbied for the "president'sposition." The crisis did not initiate, however,from domestic motivations.Nor did the administration"create" a crisis to boost the president'spopularity. The public has largely supportedthe president'shandling of the situation (see Jentleson 1992 and Mueller 1992 for data), and most consideredit as a "just war."In fact, Michael Walzer(1991, 303) arguedthat, from the standpoint of morality,"it is hardto imaginea bettercause or a moreappropriate enemy." But when faced with the options of unilateralwithdrawal, containment, and the use of force, the presidenteliminated the withdrawaland containment strategies (using the noncompensatorystrategy of eliminationby aspects), primarily(although not only) because of the political costs associated with the pursuit of these strategies. The rejection of these alternativesleft the president with the default option (use of force). The costs and benefits associated with this alternativewere then evaluatedto maximize the benefit-to-cost ratio. (It is plausible to conceive of situations in which the decision-makingprocess entails such a two-stepprocess; see Wright 1974; Klein 1989.) Because the benefits had clearly outweighed the costs on the political dimension, the focus had shifted to finding ways to minimize the costs associated with the militarydimension. Domestic political considerations have not stood in the president'sway in making the decision to use force.

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions Mintz/NONCOMPENSATORY THEORY 611

THE MILITARY/STRATEGICDIMENSION

The careers of presidentsare often defined by how they deal with wars. The president'sdecision to use force servedmultiple domestic political goals as well as strategic/military/foreignpolicy aims. Accordingto the Associated Press (February4, 1991), PresidentBush placed his bets on a short,swift war thatcould be largely suppliedfrom existing inventorieswith minimalloss of lives. Drew (1991, 193), Mueller (1992, 16), Freedmanand Karsh (1991), and Jentleson (1992), among others, also pointed out that the presidentwas reasonablyconfident that the overwhelmingmilitary power of the coalition force would secure victory,that the war would be short,and that the number of Americancasualties would be relatively small. Because Americanstypically turn sharply against a war when it becomes "extendedand expensive,"and endangering(rather than benefiting) domestic economic prosperity,the administrationwas determinedto keepAmerican casualties to a minimum.Russett (1990a, 46) pointed out that supportfor leaders dropsfast when it becomes clear thatthe war would be neithershort nor low cost (wars are not popularas they build in duration,casualties and economic cost). The presidentwas told by some of his advisors to expect, even in the worst case, not morethan five times as many aircraftlosses in the initial phase of the war as have been recorded.In fact, realizing the importance of public supportfor the war and the fact thatthe numberof casualties is negatively correlatedwith domestic public support,his top priority in planningfor the war was to keep casualtiesas low as possible. The guiding principlesin the preparationfor war were the maximization of firepower and troops and the minimizationof potential losses-Bush's declarationthat "this will not turninto anotherVietnam" (Woodward 1991, 324). Senior officers on the Joint Staff had estimatedthat the Americanside would sufferabout 1,000 casualties(killed in action)(Woodward 1991). The air campaignwas designed to try to keep losses to a minimum.Avoiding a frontalground forces assaultinto Iraqistrength was anotherway to minimize costs. Such a strategywas supposedto enhancethe chances that the benefit- to-cost ratio would be high. The presidenttried to convince the American people that this would not be anotherVietnam. According to Freedmanand Karsh(1991, 16), the presidentsaid, "thiswill not be a protracteddrawn-out war;the forces arrayedare different;the oppositionis different;the resupply of Saddam'smilitary would be very different... the topographyof Kuwait is different,and the motivationof our all-volunteerforce is superb."Inter- estingly, Jentleson (1992, 68) reports that support for the war was not dependentstrictly on expectationsof low casualties:it was high even when

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 612 JOURNALOF CONFLICTRESOLUTION casualties were estimated at 5,000, and droppedonly when the figure was 10,000. Record (1991) points out that many analystshave grossly overestimated the Iraqimilitary capabilities and creditedthe Iraqimilitary with formidable fightingpower by analyzingpotential military capabilities. They did so while focusing on the quantifiabledimensions of Iraqimilitary power ratherthan on importantunquantifiable indicators such as "moral, intellectual, and organizationalfacets of military power" and existing capabilities (Record 1991). They underestimatedthe U.S. militarystrength by looking at the U.S. experiencein Vietnam,the failed attemptto rescue U.S. hostages in Iran,the involvement of the Marines in Lebanon, and other incidents. They down- played the overwhelmingsuccess of U.S.-made high-tech systems used by the Israelisagainst Arab armies, or the highly sophisticatedweapons andthe technological revolution in warfare developed by the military-industrial sector during the military buildup of the eighties.8 To minimize the risks associatedwith the militaryoption, the Secretaryof Defense was thus going to recommendmilitary action only if he was "sureof success" (Woodward 1991, 319). To whom did the presidentlisten? Therewere basically two groupsin the president'sinner circle and his militaryadvisors: those who advocated the use of force andthose who preferred"containment." Interestingly, those who have servedin the Air Force(former Air Force Chief GeneralMichael Dugan; the new commanderof the Air Force,General Merrill McPeak; Air Force Lt. General [ret.]Brent Scowcroft [the National SecurityAdvisor]; and former Navy pilot George Bush) supportedthe use of force. According to the WashingtonPost (September16, 1990), Dugan and othersfelt thatair power will shatter enemy resistance to the point that will preclude house-to- house fighting in Kuwait. Representatives of the ground forces (Chief of Staff Colin Powell, the Army Chief GeneralKarl Vuono, and even General Norman Schwarzkopf) were more hesitant, however, to recommend an attack.Some have even insistedthat the size of the groundforces be expanded to reduce the risk of an unsuccessfuloperation. The president'sadvisors carefully evaluated the relativecapabilities available to both sides. Real fears thatU.S. forces would sufferheavy or moderate casualtieswere expressedby 78%of respondentsin public opinionpolls just days before the election (Jentleson 1992, 67). But an unusual combination

8. But as RepresentativeLes Aspin put it in a memo to the House Committee on Armed Services, "the U.S. militarywas not as bad as it looked at Desert One and it is probablynot as good as it looks after Desert Storm."

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions Mintz/NONCOMPENSATORY THEORY 613 of "political, diplomatic, strategic, geographic, topographic and climatic conditions"and the Soviet Union's tacit approvalof the U.S.-led warfavored the allies (Record 1991, 4). Top militaryadvisors, including Air Force Chief GeneralMichael Dugan, advised the administrationas earlyas mid-September that the Iraqi Air Force had very limited military capability.According to Dugan, Iraqiair defenses are no matchfor U.S. air power; the Iraqispossess an incompetentarmy that did notdistinguish itself in the Iran-Iraqwar. Dugan arguedthat unlike the situationin jungle-coveredVietnam, air power would have a tremendousadvantage in the desert.According to Dugan (Washington Post, September16, 1990), this view was also sharedby the otherchiefs. Air Force Lt. General and Deputy Chief of Staff Jimmie V. Adams added that attacking Iraq does not involve high risks and that the Air Force could guaranteetremendous devastation in Iraq (WashingtonPost, September16, 1990).9(For an estimateof the numberof casualties,see Cioffi-Revilla 1991.) And although history shows that U.S. presidents used force even when American lives were in danger (duringthe U.S. attack on Tokyo in World War II, tens of thousands of Americans lived there), Saddam Hussein's decision to release the foreign hostages provedto be a strategicmistake that made President Bush's decision to use force easier (as he himself had admitted).There is also evidence that the presidentreceived confirmatory informationfrom other key playersin theregion (Israeli Prime Minister Shamir, EgyptianPresident Mubarak, and Syrian President Assad) thatthe warwould be swift and relatively low cost-a "pushover"(Woodward 1991, 360). The hierarchyof informationand the orderin which it was processed are likewise importantfor understandingthe decision. For example, what if the presidenthad met MotherTeresa or the Pope at the beginning of the crisis and not MargaretThatcher (who, by all accounts, encouragedhim during theirAugust 2 meeting to be "tough"with Saddam)?

CONCLUSION

Studiesof the use of force in internationalrelations have typicallyfocused either on the question of why nations engage in interstateconflicts (i.e., because of political, economic, or internationalfactors) (see Jamesand Oneal

9. In contrast,during the Cubanmissile crisis, PresidentKennedy "was apparentlyuncertain whetheran air strikewould destroyall the Soviet missiles in Cuba.When he sought advice from the TacticalAir Command,its commandertold him that no such guaranteewas possible. This judgmentapparently played a majorrole in Kennedy'sruling out the air strikeoption" (Bendor and Hammond 1992, 320, emphasisadded).

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 614 JOURNALOF CONFLICTRESOLUTION

1991; Russett 1990a), or on the questionof how leadersmake war andpeace decisions (i.e., undertime pressure,group pressure, information uncertainty, and so on; see Janis 1972; Maoz 1990; Schrodt 1991; Steinbruner1974). However, the "why" and "how" of a decision certainly interact with and influence each other and the final outcome (Holsti 1990). It is important, therefore,to incorporateelements of both approachesinto a single model of decision making on the use of force. In this article,I presentedsuch a model and appliedit to the decision to use force againstIraq. The model highlighted the role of noncompensatorystrategies of decision making. The study shows that the decision to use force againstIraq involved:

- A multidimensionalstrategy of decisionmaking (which included, at the minimum,military and political considerations); - A noncompensatory(EBA) strategyof decisionmaking that eliminated optionsthat scored low on thepolitical dimension; - Nonidenticalutility functions. Although most of the variablesidentified as importantin this case are evidentin otherincidents of the use of force throughoutAmerican history, and although recurring patterns are evident, individualdecision makers vary in theirsubjective estimations of costs and benefits,even whenfaced with the sameinformation (Bueno de Mesquita, forthcoming).One cannot conclude therefore, that regardless of whohad been sittingin the WhiteHouse, he or she wouldhave also decidedto use force given identicalmilitary, economic, and politicalconditions (politicians chooseequivalent courses of actiononly whenthey have identicalutility functions-see Morton1991, 772).

Expected utility theories (Bueno de Mesquita 1981, 1985), cybernetic models (Ostromand Job 1986), and political economy of defense theories (such as Russett 1990a) help us identify the critical dimensions and variables that influence decisions to use force.?1 However, decision makers often process information in a different way from that suggested by the expected utility and cybernetic political use of force models. Many analysts therefore expected Hussein to withdraw prior to January 15, 1991. Theory development in the field of the use of force will benefit from incorporating noncompensatory strategies of decision making into this multidimensional framework.The decision to use force against Iraq clearly followed the noncompensatory strategy of elimination by aspects.

10. Nye and Lynn-Jones(1989), Ross (1991), and Mintz (1992), among others, suggested that the gap between the political economy field and the securityfield is one of the most serious problems within the study of internationalrelations and recommendedthat elements of both fields be integrated.

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions Mintz/NONCOMPENSATORY THEORY 615

REFERENCES

Abelson, RobertP., and Ariel Levi. 1985. Decision makingand decisiontheory. In Handbookof social psychology,edited by G. Lindzeyand E. Aronson.Hillsdale, NJ: LawrenceErlbaum. Aldrich,John H., JohnL. Sullivan,and EugeneBorgida. 1989. Foreignaffairs and issue voting: Do presidentialcandidates "waltz before a blind audience"?American Political Science Review 83:123-42. Allison, Graham.1971. Essence of decision: Explainingthe Cubanmissile crisis. Boston:Little, Brown. Bendor, Jonathan,and Thomas H. Hammond. 1992. RethinkingAllison's models. American Political Science Review 86:301-22. Billings, RobertS., and Lisa L. Scherer.1988. The effects of responsemode and importanceon decision-makingstrategies: Judgement versus choice. OrganizationalBehavior and Human Decision Proceses 41:1-19. Blechman, Barry, and S. Kaplan. 1978. Force without war. Washington, DC: Brookings Institution. Brody, Richard. 1984. Internationalcrises: A rallyingpoint for the president?Public Opinion 6:41-43. Bueno de Mesquita,Bruce. 1981. The war trap. New Haven, CT: Yale UniversityPress. .1984. Forecastingpolicy decisions:An expectedutility approach to post-KhomeiniIran. PS: Political Science and Politics 17:226-36. - 1985. The war traprevisited. American Political Science Review 79:157-76. Forthcoming.The game of conflict interactions:A researchprogram. Stanford, CA: Hoover Institution. Bueno de Mesquita, Bruce, and David Lalman. 1990. Domestic opposition and foreign war. AmericanPolitical Science Review 84:747-65. Bulloch, John,and HarveyMorris. 1991. Saddam'swar: The origins of the Kuwaitconflict and the internationalresponse. Boston: Faberand Faber. Cioffi-Revilla, Claudio. 1991. On the likely magnitude,extent, and durationof an Iraq-UNwar. Journal of ConflictResolution 35:387-411. Cooper,Andrew F., RichardA. Higgott,and Kim R. Nossal. 1991. Boundto follow? Leadership and followershipin the gulf conflict. Political Science Quarterly106:391-410. Dougherty,James E., and RobertL. Pfaltzgraff,Jr. 1990. Contendingtheories of international relations.New York:Harper & Row. Drew, Elizabeth. 1991. Washingtonprepares for war. In The Gulf Warreader, edited by Micha L. Siffry and ChristopherCerf. New York:Random House. Einhom,Hillel J., andR. M. Hogarth.1981. Behavioraldecision theory:Processes ofjudgement and choice. AmericanReview of Psychology 32:52-88. Ford,Kevin J., Neal Schmidt,Susan L. Schechtman,Brian M. Hults,and Mary L. Doherty.1990. Processtracing methods: Contributions, problems, and neglected research questions. Organ- izationalBehavior and HumanDecision Processes 43:75-117. Freedman,Lawrence, and EfraimKarsh. 1991. How Kuwaitwas won: Strategyin the Gulf War. InternationalSecurity 16 (2): 5-41. George, AlexanderL. 1980. Presidentialdecisionmaking inforeign policy: The effectiveuse of informationand advice. Boulder,CO: Westview. . 1991. The Persian Gulf crisis, 1990-1991. In Avoiding war: Problems of crisis management,edited by AlexanderL. George.Boulder, CO: Westview.

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 616 JOURNALOF CONFLICTRESOLUTION

Herek, GregoryM., IrvingL. Janis,and Paul Huth. 1987. Decision makingduring international crises. Journal of ConflictResolution 31:203-26. Herrmann,Richard K. 1991. The Middle East and the new world order.International Security 16:42-75. Hinich, Melvin J., and James Enelow. 1984. The spatial theory of voting: An introduction. Cambridge:Cambridge University Press. Holsti, Ole R. 1990. Crisismanagement. In Psychologicaldimensions of war, edited by B. Glad. Newbury Park,CA: Sage. Inman,Bobby R., JosephS. Nye, Jr.,William J. Perry,and Roger K. Smith. 1992. Lessons from the Gulf War.In Worldpolitics 92/93 (13th ed.), edited by Helen E. Purkitt.Guilford, CT: Dushkin. Isard,Walter. 1992. Peace economics: Futuredirections and potentialcontributions to international security.Paper presented at the Economics of InternationalSecurity Conference, The Hague, The Netherlands,May. James, Patrick,and John R. Oneal. 1991. The influence of domestic and internationalpolitics on the president'suse of force. Journalof ConflictResolution 35:307-32. Janis, IrvingL. 1972. Victimsof groupthink.Boston: HoughtonMifflin. Jentleson,Bruce W. 1992. The prettyprudent public: Post post-VietnamAmerican opinion on the use of militaryforce. InternationalStudies Quarterly36:49-73. Jervis, Robert. 1976. Perception and misperceptionin internationalpolitics. Princeton,NJ: PrincetonUniversity Press. . 1985. Perceiving and coping with threat. In Psychology and deterrence, by Robert Jervis,Richard N. Lebow, and JaniceG. Stein. Baltimore:Johns Hopkins University Press. Jervis, Robert, Richard N. Lebow, and Janice G. Stein. 1985. Psychology and deterrence. Baltimore,MD: JohnsHopkins UniversityPress. Kahneman,Daniel, and Amos Tversky. 1979. Prospecttheory: An analysis of decision under risk. Econometrica47:263-91. Kernell, Samuel. 1978. Explainingpresidential popularity. American Political Science Review 72:506-22. Klein, Gary A. 1989. Strategiesof decision making.Military Review May, pp. 56-64. Kugler,Jacek. 1991. FromDesert Shield to Desert Storm:Major victory or political quagmire? Paper presentedat the annual meeting of the Peace Science Society (International),Ann Arbor,Michigan, November 15-17. Levy, Jack S. 1989a. The diversionarytheory of war: A critique.In Handbookof war studies, edited by ManusMidlarsky. London: Unwin Hyman. . 1989b. The causes of war:A review of theoriesand evidence. In Behavior,society, and nuclearwar. Vol. 1, editedby PhilipE. Tetlock,Jo L. Husbands,Robert Jervis, Paul C. Stern, and CharlesTilly. New York:Oxford University Press. Maoz, Zeev. 1990. National choices and internationalprocesses. Cambridge:Cambridge University Press. Miller, Judith,and LaurieMylroie. 1990. SaddamHussein and the crisis in the gulf. New York: RandomHouse. Mintz, Alex, ed. 1992. Thepolitical economyof militaryspending in the UnitedStates. London: Routledge. Mintz, Alex, and Nehemia Geva. Forthcoming.The poliheuristic theory of decision. Mimeo, Texas A&M University. Mintz, Alex, and Bruce Russett. 1992. The dual economy and Arab-Israeliuse of force: A transnationalsystem? In Defense, welfare and growth, edited by S. Chan and A. Mintz. London:Routledge.

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions Mintz/NONCOMPENSATORY THEORY 617

Mitchell, Terence R., and Lee R. Beach. 1990. Do I love thee? Let me count ... toward an understandingof intuitive and automaticdecision making. OrganizationalBehavior and HumanDecision Processes 47:1-20. Morgan,T. Clifton., and KennethN. Bickers. 1992. Domestic discontentand the externaluse of force. Journal of ConflictResolution 36:25-52. Morrow,James D. 1991. Electoraland congressionalincentives and arms control. Journal of ConflictResolution 35:245-65. Morton, Rebecca B. 1991. Groups in rationalturnout models. AmericanJournal of Political Science 35:758-76. Mueller,John E. 1973. War,presidents, and public opinion.New York:Wiley. . 1992. Americanpublic opinion and the Gulf War:Trends and historicalcomparisons. Paperpresented at the Midwest Political Science Association meetings,Chicago, April. Neustadt,Richard E., andErnest R. May. 1986. Thinkingin time:The use of historyfor decision making.New York:Free Press. Nye, Joseph S., and Sean Lynn-Jones. 1989. Internationalsecurity studies: A report of a conference on the state of the field. InternationalSecurity 12:5-27. Olshavsky, R. W. 1979. Task complexity and contingent processing in decision making: A replication and extension. Organizational Behavior and Human Performance 24:300- 316. Ostrom, Charles W., Jr., and Brian Job. 1986. The president and the political use of force. AmericanPolitical Science Review80:541-66. Ostrom,Charles W., and Dennis Simon. 1985. Promiseand performance:A dynamic model of presidentialpopularity. American Political Science Review79:334-58. Ostrom, Thomas M., John H. Lingle, John B. Pryor, and Nehemia Geva. 1980. Cognitive organization of person impressions. In Person memory: The cognitive basis of social perception,edited by R. Hastie et al. Hillsdale, NJ: LawrenceErlbaum. Payne, John W. 1976. Task complexity and contingent procesing in decision-making: an informationsearch and protocol analysis. Organizational Behavior and HumanPerformance 16:366-87. Payne, John W., James R. Bettman,and Eric J. Johnson. 1988. Adaptive strategy selection in decision making.Journal of ExperimentalPsychology 14:534-52. Phillips, Kevin. 1991. Thepolitics of rich and poor. New York:Harper Perennial. Record,Jeffrey. 1991. U.S. lousy at zeroingin on ThirdWorld war prowess.Houston Chronicle, September8. Ross, AndrewL. 1991. The political economy of defense: The natureand scope of the inquiry. In Thepolitical economyof defense:Issues and perspectives,edited by A. Ross. New York: Greenwood. Russett,Bruce. 1990a. Controllingthe sword: Thedemocratic governance of national security. Cambridge,MA: HarvardUniversity Press. . 1990b. Economicdecline, electoralpressure, and the initiationof interstateconflict. In Prisoners of war? Nation-states in the modern era, edited by C. Gochman and A. N. Sabrosky.Lexington, MA: D. C. Heath. Russett, Bruce, and ElizabethHanson. 1975. Interestand ideology: Theforeign policy beliefs of Americanbusinessmen. San Francisco:Freeman. Sage, Andrew P., ed. 1990. Informationprocessing in systems and organizations.New York: Pergamon. Schrodt,Philip A. 1991. Feedback,feedforward and self-organization in internationalbehavior: Sources and constraintson innovationin foreignpolicy. Paperpresented at the International Studies Association meetings, Vancouver,British Columbia, March.

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions 618 JOURNALOF CONFLICTRESOLUTION

Schwab, Donald P., Judy D. Olian-Gotlieb,and HerbertG. Heneman. 1979. Between subjects expectancytheory research: A statisticalreview of studiespredicting effort andperformance. Psychological Bulletin 86:139-47. Simon, HerbertA. 1985. Humannature in politics: The dialogue of psychology with political science. AmericanPolitical Science Review 79:293-304. Steinbruner,John. 1974. The cybernetictheory of decision. Princeton,NJ: PrincetonUniversity Press. Tversky, Amos. 1972. Elimination by aspects: A theory of choice. Psychological Review 79:281-99. U.S. SenateHearings before the Committeeon ArmedServices. 1990. Crisis in the Persian Gulf region: U.S. policy options and implications.Washington, DC. Vertzberger,Yaacov. 1990. The world in their minds: Informationprocessing, cognition and perception in foreign policy decision making.Stanford, CA: StanfordUniversity Press. .1992. Risk in foreignpolicy decisionmaking:The forgottendimension. Paper presented at the annual meeting of the InternationalStudies Association, Atlanta, Georgia, March 31-April 4. Walzer, Michael. 1991. A just war? In The Gulf War reader, edited by Micha L. Sifry and ChristopherCerf. New York:Random House. Wildavsky,Aaron. 1974. Thepolitics of the budgetaryprocess. 2d ed. Boston: Little, Brown. Williams,John. 1990. The politicalmanipulation of macro-economicpolicy. American Political Science Review 84:767-95. Woodward,Bob. 1991. The commanders.New York:Simon & Schuster. Wright,Peter L. 1974. The harasseddecision maker:Time pressures,distractions, and the use of evidence. Journal of AppliedPsychology 59:555-61.

This content downloaded from 152.2.176.242 on Tue, 19 Mar 2013 12:48:53 PM All use subject to JSTOR Terms and Conditions