Garbage can Theory

Michael Cohen, James March, and Johan Olsen originally proposed the garbage can theory of organizational choice to extend Herbert Simon's discussion of factors comprising the rational model of organizational decision making. The rational model suggests that decisionmakers define a problem, come up with a variety of alternative solutions, then identify and choose the best alternative. Simon observed that although organizations try to make rational decisions to achieve their goals, their capacity to do so is ultimately limited by two main problems. First, decision-makers may have unclear or ambiguous preferences—within a group of decisionmakers, individuals may have different ideas about what the goal of the decision should be, or an individual may even be unclear about the precise nature of his or her own preferences. Second, decisionmakers have too little information about their options to make perfectly rational decisions. This leads to an element of randomness or trial and error in their decision outcomes.

Extrapolating from Simon's observations, Cohen, March, and Olsen described an ideal-typical organizational form called the organized anarchy, which is characterized by the constant presence of these two problems as well as the additional complication of the fluid participation of decision-making members. In the organized anarchy, decisions are not rational; rather, each decision opportunity resembles an irrational “garbage can” made up of a random draw from each of four elements (or “streams”) of decision making. Instead of being defined through group deliberation, problems arise from both inside and outside of the organization. Solutions, instead of being generated in response to the statement of a problem, exist independently and can be said to go actively in search of problems to which they can be applied. Choice opportunities, rather than emerging after the consideration of various alternatives, occur regularly as a result of structural processes and not on a timetable controlled by decisionmakers. Finally, participants in the decision-making process come and go, depending on the time they have available and their interest in individual problems or solutions. As a result of these assumptions, choice opportunities often do not result in a problem being solved. Rather, the authors claim, decisions either do not result in a solution being offered for a problem, or solutions are offered where there is no existing problem.

The garbage can model has served several purposes in studies of organizations and public administration. Most directly, Cohen, March, and Olsen identified the university as an organization often exemplifying the conditions of organized anarchy and, in several books, analyzed the implications of garbage can model for university decision making. However, their basic theory has been applied to several other organizational contexts, including military decision making, the development of psychological research programs, and the social construction of intergovernmental organizations. Scholars have also investigated the effects of such variables as strong hierarchies, planning, and organizational learning on the basic model.

The model gained new currency with John Kingdon's influential revision of the garbage can model for the decision arena of public policy making. Kingdon identified similarities between Cohen, March, and Olsen's organized anarchy and the United States policy-making process. However, Kingdon reformulated the streams entering the garbage can, naming instead three major processes of governmental agenda setting—problem recognition, policy formulation, and politics—and then specifying two additional elements: policy entrepreneurs and policy windows. The process of problem recognition can occur in a number of locations, including individual political offices, parties, or governmental agencies. Policies, as in the original garbage can model, do not develop in response to problems but, rather, have mysterious, complicated origins. Specific policies are often brought out of this soup by policy entrepreneurs, who are members of the policy community with an attachment to particular policies. The political stream—the environment in which policy is made—is made up of larger, external political conditions including the national mood, organized interest group pressure, and changes resulting from elections. The coupling of problem, policy, and political streams can [Page 336]only occur during the opening of a policy window—a particular moment of opportunity when an issue is salient because of a change in administration (a “political window”) or because of a pressing problem (a “problem window”). The process of coupling policies to problems is performed by policy entrepreneurs, who attach problems to their preferred policies, pay attention to the opening of policy windows, and find appropriate political forums for discussing their new problem-policy formulations.

...