Graham Allison’s Essence of Decision offered alternative conceptual models on foreign policy decision making and a specific discussion on the Cuban missile crisis; and has been one of the influential book in history of foreign policy analysis. It gives a significant contribution to political science study, as it has been heavily cited in most international relations textbook and also discussed by foreign policy analysts. However, despite the model’s strong influence in foreign policy study, it has been heavily criticised by foreign policy analysts about its utility and value in decision making analysis.
Number of criticism has risen regarding Allison’s conceptual framework, ranging from its originality until the problem of evidences that have been used by Allison in explaining the Cuban Missile Crises. Cornford and Horelick, for example, argue that Allison’s model is not wholly original work, rather than it is developed from previous study. Moreover, another group of criticism have questioned the account of the Cuban Missile Crises that explained by Allison. Despite these two criticisms, there are number of criticism that will be discussed in the following section. By looking at number of criticism about Allison’s model, there is a big question about the usefulness of the model in foreign policy decision making process analysis.
This essay aims to evaluate the utility of Allison’s conceptual policy in foreign policy decision making. This essay will also critically discuss each of the three models by looking from some perspectives. Moreover, taking into account that this essay relies on Allison’s “Essence of Decision”, this essay will also look at the decision making process regarding the Cuban Missile Crisis.
Allison’s Conceptual Framework
Model I: The Rational Actor (RAM)
Model I is the basic yet critical conceptual framework that mostly utilized in foreign policy decision making analysis. RAM is the best model in explaining and predicting of an individual behaviour, as well as purpose generalization in state’s action. The model “reduces the organizational and governmental political complications” by looking at government as unified actor.  Thus, a complete-informed government -regarded as ‘black box’- will process information to optimize rational action. The internal structure within decision making process will calculate theaˆˆpros and cons, and afterward, rank all the options by their chance to succeed. 
Its feature of being simple and easy to utilize, RAM could be useful when a state has limited or even no available information about the enemy. Moreover, RAM which stresses on interaction among states, will immediately produce prudent decision after considering the pro and con. Therefore, since it does not require much information to analyse a case, RAM would be very suitable in a crisis situation. Its simplicity in analysing a case makes RAM one of the popular methods in foreign policy decision making process.
On the other side, some foreign policy analysts argue that in the real foreign policy formation, number of external and individual interest factor will eventually impact the policy making process. Moreover, RAM tends to ignore a large state with complex bureaucratic nature that has various kinds of departments with their own different political and ideological perspectives.  Therefore, along with the argument that intra-national factors are “very importantaˆ¦yet critical when one is concerned with planning policy”, Allison has proposed so-called, Bureaucratic Politics Model. 
Model II: Organizational Process
Difficulties will arise when the cases that are going to be examined is not the behaviour of an individual or a state with simple bureaucracy model, but the behaviour of one organization or government with a complex structure inside. Therefore, Allison provides two alternative conceptual frameworks that will open up the black box to evaluate internal structure inside the government, which is later known as Model II and Model III.
Model II or Organizational Process Model focuses on the existing organization and their standard operating procedures (SOP) for gaining information, defining possible option and implementing programme.  Each organization has its own mission and function and series of program are developed to carry out those missions. In defining feasible option, Model II is restricted based on SOP that they believe, will enhance performance and efficiency. Moreover, Model II is not optimizing rational actor, as model I does, but rather, it is satisfying decision making actors.
Its account that foreign policy outcomes are derive from bureaucratic programs, routines and SOP, produces some advantages. It emphasizes the importance of domestic political influences in foreign policy decision making process that sometimes missed out in RAM. Therefore, model II reminds the analysts that the policy was formed not only by a high level decision-maker, but sometimes it is formed by organization. 
Nevertheless, a set of criticism has arisen in the utility of model II. Its emphasis on organisational culture may ill-serve higher level officials and finally can lead to impair the analyst’s understanding of organizations and their behaviour.  Even though this kind of problem does not occur for most of the time, but we can take it as a consideration of the effectiveness of the model II.
Model III: Governmental Politics
Governmental politics or government bargaining model focuses on key individual decision makers with their great influence in deciding on organizational action. Moreover, the model assumes that decision makers have different perceptions, priorities, commitments and also organizational positions (“where you stand depends on where you sit”).  Therefore, model III assumes that governmental actions are the result of a political bargaining process among key players. Furthermore, bargaining and negotiation processes will result in satisfying rather than optimizing decision making result. It obviously explained because personal ambition of key actor may diverge from public policy position and may lead to personal power considerations when making decision. 
Between Allison’s three conceptual frameworks, model III adds important detail about domestic politics that obviously, cannot be found in model I. In addition, model III does not only explain the roles of key individuals, but it also explains why sometimes individuals are working at contrary purposes to the interest of the government as a whole. Lastly, model III gives us explanation why policy sometimes appears to be irrational if we look it from a unitary government perspective.
Nevertheless, model III also received many critics, especially on the complexity of the model. It is focus on individual key actor that makes it difficult to study and analyse. Moreover, it requires too many variables, some variable are unknown and it is hard to apply for other countries with unclear bureaucratic politics inside.
Criticism toward Allison’s conceptual framework
Allison’s conceptual framework has been attacked by number of criticism, varying from the originality of the model, different interpretation of the Cuban Missile Crisis and the US political system, and also methodological criticism. As Stephen D. Krasner has argued that Allison’s model “aˆ¦is misleading, dangerous, and compelling”.  Therefore, in this section, number of criticism of Allison’s model will be critically discussed and assessed.
Some foreign policy analysts, such as, Cornford, Horelick, Ball and Art have claimed that Allison did not present a completely brand new approach to analysing foreign policy process; but rather it just development from previous theories. Cornford has claimed that organisational process mode is previously done by writers such as, Simon, March and Simon, Cyert and March.  Furthermore, Cornford has claimed that “aˆ¦Model IIIaˆ¦is pure Neustadt.”  Horelick et al. supported Cornford’s argument by suggesting that the bureaucratic model is closely related to previous work done by Kremlinologists.  Ball and Art also mentioned names of analyst that originally make the bureaucratic policy model, such as Huntington, Hilsman, Schilling, and Neustadt. 
Nevertheless, Allison has dedicated a section in his book to acknowledge previous scholars that become his foundation in developing his methods. He is fully aware that he utilized and developed earlier scholars’ work as he mentioned in his book, “aˆ¦this encourages much repackaging of existing theoriesaˆ¦”  Therefore, he identifies a group of writers such as March and Simon, Barnard, Cyert and Simon and so on for foundation of model II.  Furthermore, Allison also acknowledged his intellectual debt to previous scholars that related to model III paradigm, “aˆ¦model III variety have attracted increasing attention since 1960aˆ¦ the publication of Presidential Power by Richard E. Neustadt.”  Moreover, Bernstein has argued that the model is a ‘helful summary’ of earlier decision making research to present a practical mode of analysis and guide to understand business and organizational decision.  Hence, we could argue that Allison’s originality does not claim the originality of development of the model. The originality does not lie in his model, but rather in his approach to apply his models consistently to one particular case study, the Cuban Missile Crisis. 
Another criticism can be seen from methodological perspective; numbers of similarities between model II and model III have shaped ambiguity between those two models. In many occasions, some foreign analysts will combine model II and model III to analyse a case study, including Allison himself. In his article with Halperin, Allison combines those two models and become one major model – the bureaucratic politics paradigm- as an alternative model to RAM.  As Cornford argues that the three models is not totally incommensurable model to analyse foreign policy making process.  Therefore, even though Allison distinguishes three kinds of model in foreign policy analysis, those models is not easily separable in their actual application.
In bureaucratic politics model, decision is not arise from one unitary actor, but through some bargaining between organisation structures with their own agenda. Model II and model III have identical characteristics that enable them to be grouped as bureaucratic politics model. The two models are similar in a sense that both models focus on departments and organizations inside the decision maker; however, it is slightly different, in a sense that, if model II will reach a decision through Standard Operation Program, model III will make a decision through bargaining between various players within government. Despite the insignificant difference between those models, they are usually combined as the bureaucratic politics model. 
However, Caldwell has raised bureaucratic politics model’s major problem regarding the use of evidence and data. The model requires detailed data that hardly available in term of quantity and quality. In addition, Caldwell argued that there is huge possibility for analyst to imposing the model on the evidence rather than testing the model against it. Therefore, bureaucratic politics model has significant problem in analysing the data and evidence, since previous empirical problems show that data was made to fit the model. 
Allison’s alternative model has also been argued that it eliminated decision-makers’ responsibility toward the policy. The strong criticism has risen from Steel and Krasner, which argued that no one, even the President, holds responsibility of the policy as the outcome from bargaining process among bureaucratic groups. As Steel argued that, “aˆ¦where everyone is responsible for a decision, no one is responsible.”  The same argument also comes from Krusnet who argued that bureaucratic politic eliminates the importance of election: “aˆ¦Elections are a farce not because the people suffer from false consciousness, but because public officials are impotent, enmeshed in a bureaucracy so large that the actions of government are not responsive to their will.”  In contrast, Smith argues that criticism regarding the elimination of bureaucrats’ responsibility “aˆ¦is only valid to the extent to which the President is unable to get his wishes carried out.”  In some cases, the President still has the power and responsibility in deciding the final decision and for most of the cases, the President will be the one who chose the key group of decision maker. Therefore, even though Allison’s model can be an excuse for bureaucrats, we can argue that the criticism from Steel and Kranser is not applicable for all cases and need to be modified.
Furthermore, following previous criticism, there is criticism about Allison’s model utility to other countries. Even though Allison clearly points out his intention to present two additional frameworks to other countries’ foreign policy analysis (not only limited to the US and Soviet Union’s policy making)  , a group of writers has argued the inability of the model to analyse foreign policy behaviour in other countries, to be precise, un-industrialized countries. As Hill has noted that there is “a growing consensusaˆ¦over the inapplicability of the insights of Allison, et al. to foreign policy-making inside less modernised states.”  Migdal has also argued that the model cannot be applied to the countries that do not have stability of organizational structure, routine, and even bargaining process.  Moreover, Brenner also argues that Allison’s model is not a universal model and “aˆ¦more distinctive in the United States than elsewhere.”  Despite all criticisms regarding its utility to other countries, Weil has proved, in fact, the model could be utilised in the North Vietnamese foreign policy analysis; as he has noted that “examining North Vietnamese foreign policy decision making from a governmental politics perspective complements understanding gained from a rational actor analysis.” 
Nevertheless, some analysts have argued that the model is not even applicable to the Soviet Union, although the Soviet Union foreign policy has been heavily discussed in Essence of Decision. It is not only because the model requires more specific information than is available, but also as Dawisha has noted that the bureaucracy in some countries (e.g. the Soviet Union) is fundamentally different from its position in the United States because the persistent influence of the Communist Party.  Therefore, there is a doubt about the utility of Allison’s model in other countries, as Wagner has pointed out, “aˆ¦the extension of Allison’s model III to other countries may be a less straightforward enterprise than he implies”. 
Case Study: Singapore’s Defence Posture Change
In the early 1980s, Singapore announced a major important change in its defence policy, from a defensively deterrent strategy (“poisonous shrimp”) to a more actively deterrent strategy (known as the “porcupine”). In an article done by Pak Shun Ng, he applies Allison’s model to analyse Singapore’s domestic decision making process. Pak Shun Ng treats Singapore as unitary rational actor to utilizing model I (RAM); the military organisations as the unit of analysis for model II; lastly, the military and political party leadership (including senior military leaders and civilian leaders of Singapore’s ruling party, the People’s Action Party (PAP)) as units of analysis for model III.  The article argues that model II and model III provide the most reasonable explanation of the change in Singapore’s defence position in the 1980s; while model I has failed to fully explain the change in its defence posture from a ‘poisonous shrimp’ to a ‘porcupine’. Model II first reveals the appropriate development of both Singapore’s military capability and military planning ability. Furthermore, model III then proves details how the Singapore Armed Force (SAF) could announce the change convincingly to improve its stature among Singaporeans and foreigners by persuasive them that Singapore has appropriate capability to defend and survive any potential threat.  Even though the article heavily honours the utility of Allison’s model, but it still proposes modification of the models in order to be able to analyse a decision making process in a small and non-western states under absence of crisis condition. Pak Shun Ng has argued Allison’s assumption that policy is the outcome of bargaining process is not applicable policy making in reality. In fact, decision makers do not automatically have different missions.  Therefore he argued that “political bargaining model assumptions are too restrictive to explain real-life government decisions sufficiently, they should be relaxed so as to increase the explanatory power of the model.”  Furthermore, the case study of Singapore’s policy making shows the evidence of model II and III complete each other and make one alternative model against RAM. Therefore, it supports the criticism that previously discussed that Allison’s models, especially model II and III, have strong similarities and hardly separate.
The essay has discussed Allison’s conceptual frameworks, by looking at each model and its pros and cons. Moreover, a number of major criticisms about the utility of Allison’s model have also discussed and critically assessed. The first criticism is regarding the originality of the model, which has been criticised that, in fact, Allison did not bring anything new to the table; instead the models are just developed from earlier scholarship. However, we can argue that the originality of Allison model does not lie in the formulation of the models, but at the consistency in applying the model to one case study, the Cuban Missile Crisis. Hence, the fact that it focus on Cuban Crisis has leaded us to the following criticism, the flexibility of the model to be utilized to other countries.
Some foreign policy analyst have argued that Allison’s conceptual framework is not applicable to other country that is not industrialized enough to have a complex bureaucratic politics, like the United States. A group of analyst also argued that the model actually is not applicable for country with strong influence of communist party, namely the Soviet Union, even though it is heavily discussed in Essence of Decision.
Another criticism also attacked the relationship between Allison’s model, especially model II and model III. There is an argument that model II and model III are hardly separated from each other, and in some cases, they are combined into one major model, the bureaucratic politics model. Moreover, problem arose from the model about the availability of the data. The bureaucratic politics model requires specific data and evidence that hardly available.
The model has also been argued to eliminate responsibility of top level bureaucrats in policy making. However, we can argue that in some cases, the President will be the one who holds the final decision and responsibility toward the outcome (policy).
In the final section of the essay, there is a case study about Singapore’s policy change in early 1980s. By looking at the case study, we can conclude that after some modification, we can apply Allison’s model to small and non-western countries like Singapore. Moreover, after Ng modified the model to be more applicable to non-western countries, he argued that model II and model III give better explanation on the policy change, rather than model I. However, from the case study we can also point out that model II and model III are completed each other and hardly separated.
This essay has discussed some general criticisms, aside from specific criticism about its utility on Cuban Missile Crisis. Even though it’s a complex model and has been heavily criticized by some analyst, Allison has successfully provided fresh yet provocative alternative conceptual frameworks in decision making process. Allison does not intend to supplant any previous model, rather just provide a supplement framework in decision making study. Furthermore, in his book, Allison strongly emphasized that the model itself is unfinished; therefore he encourages foreign policy analysts as well as the reader to join and carry on the discussion about the model. Moreover, even though it cannot be fully utilized in all states, as case study of Singapore has shown, with small modification, the model can be utilized and proved to give better explanation than RAM.