Bounded Rationality

Core components

• Three constraints on rationality: cognitive limits (limited memory, computation, attention), information costs (acquiring and processing information is costly), time pressure (decisions must be made before all alternatives can be evaluated)
• Satisficing: searching sequentially and accepting first alternative meeting aspiration level, contrasted with optimizing across full alternative set
• Aspiration-level adjustment: aspiration levels rise after success and fall after failure, producing dynamic adjustment to environmental conditions
• Procedural rationality: judging decision quality by the procedure used rather than the outcome — appropriate when decisions occur under fundamental uncertainty
• Substantive rationality contrast: classical economics' assumption of optimization given preferences and constraints
• Heuristics: rules of thumb that bounded rational agents use to navigate complex decisions, foundational concept for subsequent heuristics-and-biases research
• Search costs: explicit modeling of information acquisition as costly activity that ends at aspiration-level satisfaction
• Computational limits as binding constraint: foundational link to artificial intelligence and bounded-optimality formalizations (Russell-Subramanian 1995)
• Organizations as bounded-rational structures: organizational design as response to individual cognitive limits, with division of labor, standard operating procedures, and routines as bounded-rationality solutions
• Connection to administrative behavior: Simon's earlier work on hierarchical decision-making in organizations as bounded-rationality application

Primary use case

Foundational framework for behavioral and organizational decision theory; applied principally in: behavioral economics (Kahneman, Tversky, Thaler subsequent traditions), organizational theory (March-Simon Organizations as canonical text), artificial intelligence (bounded-optimality formalizations, satisficing search algorithms), public administration (Simon's original applied domain), management science (decision-aid design under bounded rationality), economics of information; standard reference in graduate-level economics, organizational theory, and decision-science curricula; intellectual foundation for the heuristics-and-biases research programme that produced two subsequent Nobel Prizes (Kahneman 2002, Thaler 2017).

Common criticisms

• Bounded Rationality has substantial acceptance but specific critiques exist — classical-economics defenders including Milton Friedman argued (1953 essay) that the realism of decision-maker assumptions does not matter if predictions of the model hold, contesting the relevance of bounded-rationality observations to economic theory
• the satisficing concept has been argued to be definitionally slippery — without specifying how aspiration levels are set and adjusted, the framework can rationalize any observed choice as satisficing at the implicit aspiration level
• the relationship between bounded rationality and the heuristics-and-biases programme (Kahneman, Tversky) is sometimes argued to be a divergence rather than continuation, with Gerd Gigerenzer arguing that bounded rationality should emphasize the ecological appropriateness of heuristics rather than treating them as systematic deviations from a normative optimum
• formal models of bounded rationality (rational inattention, sparse maximization, level-k thinking) have proliferated but consensus on a unified formalization has not emerged, with critics arguing the framework remains a research program rather than a settled theory
• some critics including Ariel Rubinstein have argued bounded rationality is too often used as a wildcard to explain whatever classical models cannot, without providing positive predictive content
• the link between cognitive psychology and economic behavior remains methodologically contested, with experimental economics, neuroeconomics, and behavioral economics taking different approaches
• applied policy implementation through nudge interventions (Thaler-Sunstein, building on bounded-rationality foundations) has produced mixed empirical results, with some flagship interventions failing replication
• the boundary between bounded rationality and other choice-theoretic concepts (preference incompleteness, ambiguity aversion, robust control) is sometimes unclear.