The Garbage Can Model of Organizational Choice: Looking Forward at Forty: Volume 36

The papers collected in this volume celebrate the 40th anniversary of “A Garbage Can Model of Organizational Choice” – one of the most influential and sustained attempts to represent organizational decision-making processes in a way that accounts for generally recognized but hard to accept features of organizational life. In our overview of the volume we emphasize ways in which the garbage can model (GCM) differs from more generally accepted models of organizational decision making. We suggest that future progress in linking the GCM to specific empirical settings might be facilitated by attempts to model explicitly the interdependencies connecting participants, problems, solutions, and decision opportunities in organizations. We discuss examples of current work in which this strategy is followed in a way that is consistent with the original spirit of the model. We present the overall organization of the volume and discuss how the various chapters contribute to the further development of organizational research inspired by ideas contained in the original GCM and in some of its more recent variants and critiques.

The main task of scholars is to help good ideas forged by their predecessors find a new life in the imaginations of their successors. In this essay, we consider some aspects of this process from our experience with garbage can ideas of organizational decision making. We record our memories of initial encounters with them, impressions of their current condition, and some thoughts on convolutions they may experience in the years ahead.

President John F. Kennedy navigated through the Cuban missile crisis with the help of his advisers in the so-called ExComm. While ExComm attendance was very stable and its goal, the removal of the missiles, clear, true to the garbage can model the options available were socially constructed and were ambiguously related to the objective they purportedly served. An analysis of the recorded discussions reveals that Kennedy's choice of a blockade required the ExComm to suppress talk about the perils it entailed; his decision not to intercept a Soviet tanker was based less on caution than unsustainable indecision; and when Kennedy squared off against his advisers regarding the best way to respond to Khrushchev's conflicting offers on October 26 and 27, the latter worked to exclude him from the very decision he was about to make. The analysis points to a natural affinity between the garbage can model and ethnomethodological attention to the fine-grained details of deliberative talk.

The garbage can model showed that what appears to be irrational and unpredictable choices can be explained by processes that regulate attention allocation and the availability of choice alternatives. Because attention to alternatives fluctuates, the model generates context-dependent choices: evaluations of alternatives depend on the mix of other alternatives considered. I re-examine the mechanisms by which fluctuating attention can cause context-dependent choices. Using insights from behavioral decision theory I demonstrate how adding fluctuating attention to a well-known model of organizational decision making generates context-dependent choices of a kind that could not be explained by a maximizing process.

We build three stochastic models of garbage can processes in an organization populated by boundedly rational agents. Although short-run behavior in our models can be quite chaotic, they generate systematic, testable predictions about patterns of organizational choice. These predictions are determined, in fairly intuitive ways, by the degree of preference conflict among agents in the organization, by their patterns of attention, and by their tendencies to make errors. We also show that nontrivial temporal orders can arise endogenously in one of our models, but only when some form of intentional order, based on agents’ preferences, is also present.

This chapter reconstructs the garbage can model (GCM) of organizational choice as an agent-based model. Subsequently, it modifies the original model by establishing behavioral rules that regulate processes of organizational founding, growth, and disbanding in an artificial garbage can ecology. This population-level GCM reproduces some of the core features of the original GCM. Furthermore, it produces aggregate regularities that are broadly consistent with the historical trajectories followed by actual organizational populations.

We examine the garbage can model as a case of organizational decision making under temporal order constraints such as in the case of a dispatcher on whose desk problems, solutions, and choice opportunities arrive in a stochastic order and get resolved at particular points in time. We follow the original (Cohen, March, & Olsen, 1972) simulation except their particular setup where the problems and decision makers have access to information regarding the decisions made up to the end of the previous period a rule that we consider to be myopic. Making the assignment (or rule) less myopic would be to provide information about problems that were attached in the current period and about decision makers that were assigned a choice this period to problems and to decision makers. We also bring in an expediter who can select one or two choices to expedite each period. The expediter is endowed with a total energy of 55 that can be expended on resolving any choice to decision. Each period after the problems are attached and decision makers assigned to choices, the expediter scans the choices. The expediter selects the two choices that are the closest to decision and expends energy to move them to decision. The simulations we ran show improvement in certain situations as measured by the number of unsolved problems. We discuss our results in the sense of providing simple design features to a complex decision situation. We also discuss the paradigm of the garbage can model in the larger context of organizational decision making.

We develop an experimental setting where the assumptions and predictions of the garbage can model can be tested. A careful reconstruction of the original simulation model let us select parameters that leave room for potential variations in individual behavior. Our experimental design replicates these parameters and thereby facilitates comparison of human behavior with the original model. We find that the majority strategy of human subjects is consistent with the original model, but exhibits some behavioral diversity. Human subjects exhibit fluid diverse behaviors that improve coordination in the face of uncertainty, but hinder collective learning that can improve group performance.

This chapter replaces the fixed access and decision structures of the original garbage can model with the possibility of forming teams based on considerations of the characteristics of problems and the skills of organizational members. In real organizations, members often solve problems in collaboration, either in parallel or sequentially. In an extension of earlier simulation models, I assume that a task can be accepted by an individual member of an organization who can form a team by attempting to match the skill sets of members with the combination of skills necessary for solving the problem. Analysis of the model examines task performance and the efficient use of skills for individual and team efforts under conditions of varying workloads.

We present a simulation designed to capture the impact of both formal authority ties and informal socialization ties on the performance of an organization adapting to a turbulent world. We present a summary of three key models that informed our approach and then outline and describe the operation of our resulting simulation. Using an experiment that manipulated both the authority network structure and the stress the organization placed on socialization, we show inefficient authority structures harm performance, and also that socialization has a strong and nonlinear impact on peak organizational performance and on the performance of top management. We also present a case study, instantiating the general model with the specific context of a real-world organization. Finally, our integrated multimodel suggests that companies should pursue different strategies in hiring key strategic actors than they do for other actors.

This chapter first examines the role of attention in the garbage can model of decision making and compares it both to prior approaches in the Carnegie School tradition and the attention-based view of the firm. Both the garbage can model and the attention-based view rely on the same assumption, one that is rarely recognized nor understood – that organizational decision making is characterized by situated attention, where organizational participants vary across time and place in what they attend to. In the garbage can model, decision opportunities are the temporal contexts for situated attention; in the attention-based view, attention is situated in both time and place within the organization's communication channels. In the garbage can, situated attention is also shaped by the ecology of problems and opportunities competing for attention. The final part examines the determinants and consequences of tight versus loose coupling of channels in organizations and its effects on participants’ situated attention. Attention structures external to channels and the architecture of channel structures shape the degree of coupling found in organizations. In viewing coupling as a variable, the chapter suggests that a modified garbage can model, combined with an increased focus on situated attention, provides the foundations for a more general theory of nonroutine decision making.

We extend the classical garbage can model to examine how individual differences in ability and motivation will influence organizational performance. We find that spontaneous coordination provided by an organized anarchy is superior when agents are equally competent. The Weberian bureaucracy of planned coordination is effective when problems require specialist knowledge. However, errors in matching problems to specialized agents are a central challenge for bureaucracies. Actual organizations, therefore, combine elements of organized anarchies and bureaucracies. Heterogeneous motivation compounds coordination problems, but is usually less important than competence. Our findings point to matching and interactive learning as fruitful areas for further study.

The garbage can model of organizational decision making was initially framed as being particularly applicable to universities and other “illegitimate” organizations that were likely subject to unclear technology and goals. It is argued here that corporate strategy making, a most traditional organizational process, may be subject to garbage can processes as well. While business firms may have a goal of profit maximization, the “technology” of profit maximization, particularly over moderately distant horizons, is typically sufficiently poorly understood so as to permit garbage can-like mechanisms to come into play. Consideration of garbage can processes in a corporate context also suggests some enrichments of the original conceptualization, particularly the role of organizational politics. Problem resolution is not merely an issue of the mobilization of sufficient collective energy, but is also a problem of legitimation. In the corporate context, legitimation is importantly driven by the links that can be made between a possible solution and the, typically ambiguous, corporate strategy.

When considered as organized solutions to problems of provision of public goods, Free/Open Source Software (F/OSS) productions share a number of their defining features with the organized anarchies described by Cohen, March and Olsen in their “Garbage Can Model” (GCM). The open and voluntary contribution of software developers creates constant fluctuations in levels of attention and an extremely fluid participation. The lack of predefined hierarchical access to organizational problems determines a fundamental uncertainty about how collective goals may be linked to individual activities, and in how responsibilities and tasks may be allocated efficiently within the project. Finally, the complexity involved in the collective production of tens of thousands of lines of computer code without explicit coordination creates a situation of technological ambiguity supported by a radically decentralized activity of organizational problem finding and problem solving. In this paper we take these broad similarities as point of departure to specify an empirical model that captures some of the garbage can properties of organizational problem-solving activities in the context of a specific F/OSS project followed throughout a complete release cycle. We examine the interconnected system of individual decisions emerging from problem-solving activities performed by the 135 contributors involved in the F/OSS project on the 719 software bugs reported during the period of observation. We treat the evolving two-mode network produced by encounters between carriers of organizational solutions (contributors) and organizational problems (software bugs) as a dynamic opportunity structure that constrains and enables organizational decision making. We document how stable local configurations linking problems and solutions are induced by – and at the same time sustain – decentralized problem-solving activities with meaningful self-organizing properties.

Competitions celebrate meritocratic values. Letting the best man or woman win leaves little room for human choice, since presumably the result follows from ascertaining the fact that someone did better than the rest. But in architectural competitions, appointing a winner involves human choice. An in-depth empirical investigation demonstrates that such human choice has the character of intuition and judgment. The choice of the winner preceded the process by which the winning design proposal was established as being better than the other proposals. We discuss the role of intuitive choices in architectural competitions and claim that they reflect necessity more than vice. They are ways around the fundamental incommensurability of the alternative design proposals. The garbage can model is used as a framework for making sense of the observed counterintuitive ways of decision making. Its attempt to theorize alternative forms of orderliness proves helpful, but on certain points our observations also suggest modifications to the model.

This chapter explores geopolitics, garbage cans, the need for interdisciplinary insight, and the lures and limitations of one-sided mono-disciplinary conceptual models in understanding strategic decision making. We argue that a combination of the garbage can model and Nathan Leites’ psycho-cultural approach to decision making might be useful in giving insights for events and for organizational behavior. As a decision making case, we consider the 1941 decision of the Empire of Japan to declare war on the Allied Powers. We find that there could be useful lines of integration between the garbage can framework and other perspectives in geopolitical decision making. In using a historical example to illustrate the possible integration, we argue that there are inherent limits to single-model decision making approaches. Developing interdisciplinary frameworks for understanding foreign policy decision making may lead to better insights in real-world processes and seems like a step in a fruitful direction.