Garbage Can Model

The Garbage Can Model is a theory of organizational decision-making under ambiguity, developed by James March, Michael Cohen, and Johan Olsen in 1972. It rejects the assumption that decisions are the rational outcome of a sequence in which problems are identified, alternatives are generated, preferences are evaluated, and choices are made. Instead, the garbage can model treats decisions as the largely accidental product of four independent streams flowing through an organization: problems, solutions, participants, and choice opportunities.

In this model, problems are not solved. They wander through the organization looking for situations in which they might be addressed. Solutions are not responses to problems; they are independent items — technologies, policies, personnel — looking for problems to which they can attach. Participants drift in and out of decision arenas based on their energy and attention, not on their relevance. Choice opportunities — meetings, budget cycles, hiring decisions — are occasions when these streams may collide, producing a decision that may or may not have any logical connection to the problem it appears to address.

The garbage can model is not a theory of irrationality. It is a theory of organized anarchy — situations in which the normal requirements for rational choice (clear preferences, stable technology, defined participation) are not met. Universities, research laboratories, and technology startups are frequently organized anarchies. The model predicts that under these conditions, decision outcomes will be highly sensitive to timing, attendance, and the sheer accident of which solution is currently available.

The model connects to Problemistic search and the behavioral theory of the firm: where Cyert and March treated search as problem-triggered but still purposeful, the garbage can model treats the relationship between problems and solutions as essentially stochastic. It also connects to complex adaptive systems: the organization is a system in which structure is produced not by design but by the temporal coupling of independent processes.

The garbage can model predicts that in organized anarchies — organizations with ambiguous preferences, unclear technology, and fluid participation — decision quality is less important than decision timing. A solution that arrives too early meets no problem; a problem that arrives too late finds no choice opportunity. The temporal coupling of streams, not the rationality of any individual actor, determines what gets decided.