A Behavioral Theory of the Firm

The firm stops looking like a single mind

Standard economic theory often made the firm analytically convenient: it had a goal, usually profit; it knew the alternatives available to it; and it selected the best one. Richard Cyert and James March begin from the opposite experience. A real firm contains owners, executives, divisions, workers, suppliers, and customers whose demands do not collapse into one stable preference. Its information is partial, its attention is scarce, and many of its consequential choices emerge from ordinary routines rather than a comprehensive calculation.¹

A Behavioral Theory of the Firm asks what becomes visible once that convenient fiction is removed. It extends the behavioral program of Organizations and Herbert Simon's bounded rationality into a theory close enough to operating life to model prices, output, budgets, and search. The book's original bibliographic record shows its empirical and simulation chapters alongside its general concepts. It is not merely a collection of memorable terms; it tries to explain how those terms generate observable decisions.²

Goals are negotiated, then temporarily stabilized

Cyert and March's firm is a coalition. Participants contribute resources and continue participating while the bargain remains acceptable. Because their claims conflict, organizational goals are produced through bargaining rather than discovered by asking what “the company” wants. Sales may seek volume, production may seek schedule stability, finance may seek cash, and managers may seek discretion. Decisions become possible through a quasi-resolution of conflict: different goals are attended to at different times, constraints are handled locally, and apparent consistency is often postponed rather than achieved.³

The resulting goals are usually expressed as aspiration levels. Performance is not judged against an abstract global optimum but against a threshold shaped by past results, comparable organizations, and recent expectations. Meeting the threshold reduces pressure to reconsider the system. Missing it creates a problem and directs attention. This makes aspirations both psychologically plausible and organizationally political: the threshold chosen determines what counts as a problem.⁴

Organizations also avoid uncertainty. They use contracts, standard procedures, budgets, inventories, and short-run feedback to make an unknowable future more manageable. These devices do not eliminate uncertainty; they convert some of it into rules and negotiated environments. The firm learns by updating routines, goals, and expectations from experience, often without anyone designing a single master lesson.⁵

Trouble starts a local search

When results fall below aspiration, the organization engages in problemistic search. Search is motivated by a specific shortfall, begins near the symptom and current practice, and expands when nearby remedies fail. A sales miss might first prompt more selling effort, then a price change, and only later a product or strategy reconsideration. That sequence explains why adaptation can be intelligent yet myopic. The search process is shaped by departmental vocabulary, available data, and who has authority to redefine the problem.⁶

Organizational slack—resources beyond the minimum needed to satisfy current claims—helps keep the coalition together. In good times, slack can absorb conflict and provide room for experimentation; in hard times, it becomes a buffer that can be drawn down. The term is easily misread as simple waste. In the behavioral theory it is also stored adaptability. But the book does not settle who gets the buffer: spare engineering time and executive perks can both be called slack while serving very different lives.⁷

The legacy is a research program, not a recipe

The book helped establish what later became the Carnegie tradition in organization studies. Its vocabulary travels into research on routines, attention, learning, innovation, and strategy. James March's later Exploration and Exploitation in Organizational Learning asks why adaptive systems can become too good at using present knowledge to discover the next one. A sixtieth-anniversary review shows both the reach of the original mechanisms and the many empirical programs that modified them.⁸

That success creates a reading hazard. Later ideas about routines, attention, and institutional politics are often projected backward into the 1963 book. The book's concrete decision models anchor what the original argument actually claimed. A scholarly assessment of the program is useful for separating the initial propositions from their descendants.⁹

Structured reading paths distinguish lineage from comparison

The reading dependency on Organizations is an editorial behavioral extension: Cyert and March move from the earlier work's general account of bounded rationality and organization toward coalition goals, aspiration feedback, and firm-level models. Herbert A. Simon and Administrative Behavior identify the intellectual context, but the links do not assign every mechanism to Simon or establish a single linear inheritance.²

Paths to cooperation, incentives, and organizational equilibrium, decision-making, judgment, and bounded rationality, and executive attention, information, and organizational sensing are editorial comparisons to coalition bargaining, satisficing, and selective attention. Organizational intelligence frames their combination as an adaptive system, while benefit for all life tests whether the coalition's negotiated goals are legitimate beyond its members. No structured impact observations or typed relations are encoded, so these paths support neither measured effects nor claims of historical influence.

The deepest limit is the coalition's boundary. People with bargaining leverage enter the model more readily than communities, ecosystems, future generations, or workers whose exit is costly. Describing how goals arise does not justify them.¹⁰ The cited record consists of the work, edition records, and scholarly reviews; it contains no representative testimony or outcome data from workers, suppliers, customers, communities, or ecologically affected life. Use the book as a set of questions: Which claims are represented? What performance lies below aspiration? Where will search begin? Which routines filter the evidence? And what protected slack would let the organization learn before a harmed party has to force the problem into view?