What Is Bounded Rationality?
Bounded Rationality: Reasoning Within Human and Organisational Limits
Bounded rationality describes decision making by people whose information, time, attention, and computational capacity are limited. The concept, associated especially with Herbert A. Simon, replaces the assumption of an all-knowing optimiser with a more realistic account of purposeful but constrained reasoning.
A perfectly rational decision-maker in a formal model may know every alternative, consequence, probability, and preference and then select the option that maximises utility. Real decision-makers cannot usually enumerate the full possibility space or calculate all downstream effects, particularly in novel, complex, or rapidly changing situations.
People therefore simplify. They attend to selected cues, use rules and heuristics, search alternatives sequentially, rely on representations of the problem, and stop when an option meets an aspiration level. Simon called this satisficing: finding a satisfactory and sufficient option rather than proving that no better option exists.
The bounds are not purely individual. Organisations shape rationality by controlling information, dividing labour, setting goals, establishing procedures, and directing attention. These structures can make coordinated action possible, yet they can also suppress inconvenient evidence, fragment responsibility, or keep obsolete assumptions in place.
Bounded rationality does not imply that people are irrational or incapable of improvement. Decision aids, training, diverse perspectives, better information design, simulation, and well-chosen routines can extend effective reasoning. However, every aid also embodies assumptions and must fit the user's task and environment.
The concept matters because it changes how decisions are evaluated and supported. Instead of asking why a person failed to perform an impossible optimisation, analysts can ask what information was available, what representation guided search, what constraints shaped attention, and how the decision environment could be redesigned.
Back to reading