Allison’s Slow “Waltz” with Structure in Foreign Policy Analysis

Allison’s Slow Waltz with Structure:
A Comparison of the Bureaucratic Politics and Cognitive Models of Foreign Policy Analysis

There are numerous approaches to analyzing foreign policy decision-making. Realism and Liberal Institutionalism are perhaps the two ‘hegemons’ of Foreign Policy Analysis (FPA), but competing theories include Social Constructivism, Cognitive and Psychological, Bureaucratic Politics, New Liberalism, and the Interpretive Actor Perspective. When categorized, however, said theories are typically split into two categories: structure-based approaches (Realism, Neoliberal Institutionalism, and Social Constructivism) and agent-based approaches (Cognitive and Psychological, Bureaucratic Politics, New Liberalism, and the Interpretive Actor Perspective). These distinctions are further mediated through Kenneth Waltz’s three level of analysis (system, unit, and individual). The structure-based approaches address foreign policy at the level of the system and address how the international realm affects the decisions of policy makers; whereas agent-based approaches treat foreign policy at the individual level to address how various actors affect foreign policy-making. Such categorization, however, is problematized when using the levels-of-analysis rubric. The Bureaucratic Politics, often touted as an agent-based approach, demonstrates a distinctive lack of agency when compared to other agent-based approaches. By comparing the Bureaucratic Politics approach with the Cognitive and Psychological approach, one not only demonstrates the systematic nature of the bureaucratic politics model, but also reveals the disparate nature of agency within the two approaches, and the consequences thereof. One first establishes how the Bureaucratic Politics model is systematic by identifying parallels with Waltz’s 1979 Neorealism doctrine, and then compares it to the Cognitive model to illuminate the structure and agency systems.

According to the Bureaucratic Politics model, policy decisions are not made by a single actor within a state, but are made in a push-and-haul process by various actors in a government, particularly senior bureaucrats representing their organization and each vying for position and power. The outcomes are decisions made through bureaucratic competition and bargaining, which undermine Realist and Neoliberal Institutionalist assumptions that the state is a unitary and rational actor. Bureaucratic Politics model scholars David Kozak and James Keagle outline several key characteristics that define the approach. First, since WW2, non-elected officials have increasingly become prominent actors in policy-making. Second, external factors affect an organization’s environment. Third, bureaucrats are motivated by the need to maintain power. Fourth, national security policies are often the result of competing bureaucratic organizations. Fifth, individuals in an organization often will adopt the organization’s goals. Sixth, because of their tenure, expertise, and knowledge, elected officials are often dependent of the skills and abilities of bureaucrats. Seventh, bureaucratic politics are marred by competition and compromise. Eighth, the elected executive official plays the role of coordinating and considering bureaucratic information.[1] Finally, an organization’s primary goal is self-preservation, and it will seek actions that maximize their interests (Kozak and Keagle, 1988, 5-10).

On one hand, Graham Allison suggests that the state is not a rational actor due to the intra-state competition between senior bureaucrats. On the other hand, he is making the same assumptions about bureaucrats sitting around the proverbial table. In essence, Allison has prescribed a Waltzian self-help paradigm to the executive branch of the state. Whereas Kenneth Waltz suggests that main actors in the system are the states (1979, 93); so too does Graham Allison and Morton Halperin claim that the main actors in the executive branch of government are the senior members of bureaucracies (1972, 47). Moreover, there are parallels in the interests of the actor. A realist state’s main priority is survival, and therefore its key drive is the accumulation of power (Waltz 1979, 91). Similarly, the main interest of senior bureaucrats is the health of their organization, which “depends on maintaining influence, fulfilling its missions, and securing the necessary capabilities” (Allison and Halperin, 1972, 48). Furthermore, in Allison and Halperin’s text, the elements that ensure the health of an organization (resources, influence, reputation) are finite and fungible (1972, 50), and the probability of success depends on their acquisition of power (52). Correspondingly, Waltz claims that two key aspects of his international system is the relative nature of power (1979, 90). In other words, power is a zero-sum game. To acquire power, another has to lose power. The other key aspect of his international system is anarchy (1979, 91).[2] Comparably, Jerel Rosati points out that no preponderant individual or organization exists within the bureaucratic realm. While the president may coordinate other players and weighs the information presented, he is only one of several chiefs, and his decisions can be reversed or ignored (1981, 237). The purpose of outlining these parallels is to demonstrate that Allison’s bureaucratic politics model echoes Waltz’s anarchic, self-help system. Waltz succinctly describes the self-help system as “one in which those who do not help themselves, or who do so less effectively than others, will fail to prosper, will lay themselves open to danger, will suffer” (Waltz, 1979, 118). The actors in a self-help system are compelled and conditioned to act according to structural constraints.

While Waltz’s theory originally applied to the international realm, or the level of the system, it also fits within Allison’s bureaucratic realm (the level of the unit). Moreover, the same structures that constrain and condition state behavior at the level of the system also constrain and condition bureaucrats at the level of the unit. There remains, therefore, little agency for the senior bureaucrat. The bureaucrat is compelled to advocate for decisions that will ensure the continued success of their organization. That isn’t to say that the bureaucrat does not have the interests of the nation in mind; in fact, what compels the bureaucrat to act is threats to the nation. But as both Allison and Halperin point out, bureaucrats come to believe that the health of their organization is crucial to the interests of the nation (Allison 1971, 48; Halperin 1971, 73). Senior bureaucrats, then, are rational actors. They are not deciding rationally for the interest of the state (although they might think they are), but rather for the sake of their own department. If they do not act rationally, the structure will diminish their influence, capability, and resources, and threaten them with redundancy. Thus, decisions generated from the bureaucratic politics model are products of the structure of the bureaucratic system.

It is important to acknowledge that according to the Bureaucratic Politics model, most of a players’ interests that affect their decision-making are related to the organization. Allison and Halperin note that the player’s personal interests and his conception of his role do affect, in some measure, where he stands (1972, 48). Indeed, examples of personal interests playing a role in decision-making are evident in Yaacov Verzberger’s 1984 case study of the Sino-Indian conflict, where things such as interpersonal relationships, self-image, and personality played a role in bureaucratic politics (257-285). However, many bureaucratic politic scholars (Allison and Halperin, 1972, 48; Carlsnaes, 2012, 123; Kozak and Keagle, 1988, 7) consider bureaucratic interactions as the key interaction.

The Cognitive model of FPA, on the other hand, offers an individualistic account of decision-making. The Cognitive model suggests that an individual’s perceptions are filtered through lenses of externally acquired beliefs and concepts (Hart, 1977, 116). These beliefs and concepts provide norms, standards, and guidelines that influence the individual when processing information and considering outcomes (George, 1969, 191). According to Rosati, these ‘filters’ alter a decision maker’s cognition in six principle ways: cognitive structures of beliefs, selective memory, selective attention and perception, causal inference, and cognitive stability (2000, 52-53). Although there are a number of different schools of theory within the cognitive model -including mirror imaging, historical analogy, schemas, and operational code – attention will primarily be given to the schema theory and operational codes.

Schemas, as outlined by Rosati, are

mental constructs that represent different clumps of knowledge (or comprehension) about various facets of the environment. They necessarily simplify and structure the external environment, enabling individuals to absorb new information and intelligibly make sense of the world around them (2000, 57).

Moreover, Schema theory posits that different schemas will be used at different times, depending on how stimulus from the environment evoke them. Alternatively, the operational code, outlined by Alexander George, suggests that people’s beliefs contribute to an overarching philosophical system that informs their perceptions and decisions. This philosophical system can be parsed out through specific questions directed at one’s philosophical content and instrumental beliefs (1969, 199-200). It is important to establish that these schemas and operational codes are not consciously or willing created. Rather, the minds of individuals are unable to cope with the vast amount of information presented to them, so they rely of mental ‘shortcuts’ such as schemas and overarching operational codes to simplify the incoming information and rationalize decisions (Rosati 2000, 57).

The Cognitive model, similar to the Bureaucratic Politics model, challenges the assumption that policy makers are rational actors. Unique to the Cognitive model, however, is the level of individualism. While the Cognitive model may successfully identify and explain foreign policy decisions, it is limited to the individual. Each person’s schemas or operational codes are the product of their experiences, their history, and personal philosophies. Every person’s cognitive map,therefore, is idiosyncratic. Thus, there is no structural system in the cognitive model. All constraints come from the actors themselves, and the particularities of those constraints are unique to the actor. This raises an important point regarding agency in FPA theory, namely the connection between agency and accountability.

In structure-based approaches, the nature of the system forces players to act in particular ways, removing agency. If agency is removed, then the ability to hold players accountable for their actions is suspect. Valarie Hudson remarks that FPA theory has been dominated by structure-orientated approaches, and that agent-orientated approaches introduce a measure of accountability into the field.

[I]t is very difficult to grapple with the issue of accountability in international affairs if the theoretical language cannot, in a realistic fashion, link acts of human agency in that realm to the consequences thereof (Hudson, 2005, 4).

Interestingly, Stephen Krasner offers a similar critique of the Bureaucratic Politics approach, suggesting that the approach obscures the powers of the presidency and excuses the key decision makers from the responsibility of their actions. Drawing a further link between the structure and unaccountability, Krasner asserts:

My argument here is that this vision [Bureaucratic Politics approach] is misleading, dangerous, and compelling: misleading because it obscures the power of the President; dangerous because it undermines the assumptions of democratic politics by relieving high officials of responsibility; and compelling because it offers leaders an excuse for their failures and scholars an opportunity for innumerable reinterpretations and publications (1972, 160).

If players in the Bureaucratic Politics approach are forced to act according to the edicts of the self-help system, then they are absolved from taking responsibility for their actions.

The juxtaposition of the two models illustrates the distinctive nature of structure and agency within each model. The Bureaucratic Politics model is typically grouped into an actor-based category of foreign policy analysis approaches, especially when being compared to Realism, Neoliberal Institutionalism, and Social Constructivism. However, upon viewing the approach through a ‘Waltzian lens’ and comparing it with the cognitive approach, we see structure within the framework of the sub-unit system. The presence of structure suggests diminished levels of agency. This would indicate that the agency-structure divide is not as stark as we like to think.


Allison, Graham T, and Morton H Halperin. “Bureaucratic Politics: A Paradigm and Some Policy Implications.” World Politics 24 (1972): 40-79.

Allison, Graham T, and Philip Zelikow. Essence of Decision: Explaining The Cuban Missile Crisis. 2nd Edition. New York: Longman, 1999.

Carlsnaes, Walter. “Actors, Structures, and Foreign Policy Analysis.” In Foreign Policy: Theories, Actors, Cases, by Steve Smith, Amelia Hadfield and Tim Dunne, 113-129. Oxford: Oxford University Press, 2012.

George, Alexander L. “The “Operational Code”: A Neglacted Approach to the Study of Political Leaders and Decision-Making.” International Studies Quarterly 13, no. 2 (1969): 190-222.

Halperin, H Morton. “Why Bureaucrats Play Games.” Foreign Policy 2 (Spring 1971): 70-90.

Hart, Jeffery A. “Cognitive Maps of Three Latin American Policy Makers.” World Politics 30, no. 1 (1977): 115-140.

Hudson, Valerie M. “Foreign Policy Analysis: Actor-Specific Theory and the Ground of International Relations.” Foreign Policy Analysis 1 (2005): 1-30.

Kozak, David C, and James M Keagle. Bureaucratic politics and national security : theory and practice. Boulder: Lynne Rienner, 1988.

Krasner, Stephen D. “Are Bureaucracies Important? (Or Allison Wonderland).” Foreign Policy 7 (Summer 1972): 159-179.

Rosati, Jerel A. “Developing a Systematic Decision-Making Framework: Bureaucratic Politics in Perspectives.” World Politics 33, no. 2 (1981): 234-252.

Rosati, Jerel A. “The Power of Human Cognition on the Study of World Politics.” International Studies Review 2, no. 3 (2000): 45-75.

Vertzberger, Yaacov. Misperceptions in Foreign Policymaking: the Sino-Indian Conflict 1959-1962. Boulder: Westview Press, 1984.

Waltz, Kenneth N. Theory of International Politics. Long Grove, Illinois: Waveland Press, 1979.


[1] Rosati points out that this preeminence of the executive can be contested (1981, 237).

[2] Anarchy, in this sense, is defined as a lack of hierarchy or Hobbesian Leviathan. It does not indicate that it is chaotic. Behaviors are still governed by norms and conventions.

Written by: Mack Clayton
Written at: London School of Economics
Written for: Dr. Chris Emery
Date written: December 2012

Please Consider Donating

Before you download your free e-book, please consider donating to support open access publishing.

E-IR is an independent non-profit publisher run by an all volunteer team. Your donations allow us to invest in new open access titles and pay our bandwidth bills to ensure we keep our existing titles free to view. Any amount, in any currency, is appreciated. Many thanks!

Donations are voluntary and not required to download the e-book - your link to download is below.