Loading field map
Behavioral decision theory is the study of how people actually make judgments and choices, and of how those actual decisions compare with the standards of rationality. It sits at the intersection of psychology and economics, but its core subject matter is narrower than either parent discipline: it concerns the mental processes that produce a decision, the systematic ways those processes depart from normative ideals, and the attempts to model both the departures and the processes themselves.
The field’s central question is deceptively simple: when a person chooses among options, estimates a probability, or evaluates a gamble, what determines the outcome? The stakes are substantial. If human decisions reliably deviate from what rational calculation would prescribe, then models of economic behavior, public policy design, legal liability, and medical consent all rest on assumptions that may be false. Behavioral decision theory supplies the evidence and the theoretical vocabulary for assessing those assumptions.
To understand behavioral decision theory, one must first understand the standard against which it measures behavior. That standard is expected utility theory, developed in its modern form in the mid-twentieth century. Expected utility theory specifies how a rational agent should choose under uncertainty: assign a numerical utility to each possible outcome, multiply each utility by the probability of that outcome, sum these products across all possible outcomes, and choose the option with the highest total. The theory also specifies how rational beliefs should be updated in light of new evidence, a requirement formalized in Bayes’ theorem.
Expected utility theory was never primarily a description of how people decide. It was a normative theory—a set of axioms that any coherent preference ordering should satisfy. But for several decades, many economists treated it as a reasonable approximation of actual behavior as well. Behavioral decision theory emerged from the accumulating evidence that this approximation fails in systematic, predictable ways. People do not merely make random errors; they make the same kinds of errors in the same kinds of situations, and those errors often follow patterns that can be modeled.
The first sustained research program in behavioral decision theory was built around the concept of heuristics—mental shortcuts that reduce complex judgment problems to simpler operations. This program, associated most strongly with Daniel Kahneman and Amos Tversky from the early 1970s onward, argued that people routinely substitute an easy question for a hard one. Asked how likely an event is, people may instead ask how easily examples of that event come to mind (the availability heuristic). Asked to judge whether an object belongs to a category, people may ask how similar it is to a stereotypical member of that category (the representativeness heuristic). Asked for a numerical estimate, people may start from an initial value and adjust insufficiently (the anchoring-and-adjustment heuristic).
These heuristics are not occasional lapses. They are the normal machinery of judgment, and they produce characteristic biases: systematic errors that persist even when the correct answer is available in principle. For example, people overestimate the frequency of dramatic causes of death (because those causes are vivid and easily recalled) and underestimate the frequency of mundane causes. They ignore base rates when judging whether someone has a disease, focusing instead on how typical the person’s symptoms are of that disease. They are overconfident in their predictions, especially when they have a plausible story that ties the evidence together.
The heuristics-and-biases program was not merely a catalogue of errors. It offered a unified explanation: the human mind, adapted to a world of limited time and information, uses fast and frugal procedures that work well in many contexts but fail in predictable ways when the context is unusual. The program’s influence came from its demonstration that these failures are not noise but structure. The biases are reproducible in laboratory experiments, they appear across cultures and age groups, and they can be described mathematically.
The program’s limits became clearer over time. Critics noted that the heuristics were often defined after the fact: if people made one kind of error, it was attributed to one heuristic; if they made the opposite error, it was attributed to a different heuristic. The theory did not always specify in advance which heuristic would apply in a given situation. Moreover, the framing of the program as a catalogue of human irrationality drew objections from researchers who argued that the apparent biases often disappeared when the problem was presented in a more natural format or when the participants had relevant experience.
The most influential single achievement of behavioral decision theory is prospect theory, developed by Kahneman and Tversky in 1979. Prospect theory was an explicit attempt to replace expected utility theory as a descriptive model of choice under risk. It retains the structure of expected utility—options are evaluated by combining probabilities and values—but it modifies both components in ways that reflect observed behavior.
The theory has three central features. First, people evaluate outcomes relative to a reference point, typically their current state or their expectations, rather than in terms of final wealth. Gains and losses are coded relative to that reference point, and the reference point can shift depending on how a problem is framed. Second, the value function is asymmetric: losses hurt more than equivalent gains please, a property called loss aversion. Third, probabilities are transformed nonlinearly. People overweight small probabilities and underweight moderate and large probabilities, which helps explain why people buy lottery tickets (overweighting the tiny chance of winning) and also buy insurance (overweighting the small chance of a disaster).
Prospect theory was a turning point because it showed that a formal, mathematically precise model could capture the systematic deviations from expected utility. It was not a vague claim that people are irrational; it was a specific claim about how they are irrational. The theory has been revised and extended since its original formulation, and it remains the standard descriptive model of risky choice in behavioral economics.
The relationship between prospect theory and the heuristics-and-biases program is worth making explicit. The heuristics program was primarily about judgment—how people estimate probabilities and infer causes. Prospect theory is primarily about choice—how people select among options. The two are complementary: judgment supplies the beliefs that feed into choice, and choice reveals the values that judgment cannot directly measure. But they are distinct research agendas with distinct methods and distinct theoretical commitments.
A parallel tradition, older than the heuristics-and-biases program, approaches behavioral decision theory from a different angle. Herbert Simon, writing in the 1950s, argued that human rationality is bounded by the limitations of the human mind: limited memory, limited computational capacity, and limited time. A perfectly rational agent would consider all possible options, compute all consequences, and choose optimally. A real agent cannot do this. Instead, people satisfice: they search for an option that meets some minimum standard of acceptability, and they stop searching once they find one.
The bounded rationality tradition differs from the heuristics-and-biases program in its emphasis. Simon and his successors did not primarily study errors. They studied the design of decision procedures under resource constraints. The question was not "Why do people deviate from rationality?" but "How do people manage to decide at all, given their limitations?" The answer often involves simple rules that exploit the structure of the environment. A chess player does not evaluate all possible moves; she uses pattern recognition and selective search. A consumer does not compare all available products; she uses a few salient criteria and stops when a product is good enough.
This tradition has a different normative stance. Where the heuristics-and-biases program tends to treat deviations from normative models as errors, the bounded rationality tradition treats them as adaptations. A heuristic that works well in most real-world situations is not a bias; it is a rational response to the cost of information and computation. The two traditions have often been in tension, and the tension is genuine. The same behavior—using a simple rule instead of a full calculation—can be described as a bias (if the full calculation is the standard) or as an efficient strategy (if the cost of calculation is counted).
The bounded rationality tradition also connects behavioral decision theory to a broader research agenda. Simon’s work on organizational decision-making, on the design of artificial systems, and on the philosophy of the social sciences all grew out of this core insight about human limitations. In recent decades, the tradition has been extended by researchers who study "fast and frugal heuristics"—simple decision rules that are deliberately designed to ignore most available information and that nonetheless perform well in specific environments. This research program, associated with Gerd Gigerenzer and colleagues, has been a persistent critic of the heuristics-and-biases program, arguing that many apparent biases are artifacts of unrealistic experimental tasks and that the same heuristics that produce errors in the laboratory produce good decisions in natural settings.
Behavioral decision theory is defined as much by its methods as by its theories. The field is built on controlled experiments, typically conducted in laboratories, in which participants make judgments or choices under carefully manipulated conditions. The experimental method allows researchers to isolate specific cognitive processes, to test precise predictions, and to rule out alternative explanations.
The experimental tradition has several distinctive features. One is the use of hypothetical scenarios: participants are asked what they would do in a described situation, rather than actually facing the situation. This method has been criticized for lacking realism, but it has a crucial advantage: it allows the researcher to control every aspect of the situation, including aspects that could not be controlled in the real world. A second feature is the use of monetary incentives. Many experiments pay participants based on their decisions, so that the choices have real consequences. The finding that incentives reduce but do not eliminate the standard biases is itself an important result. A third feature is the systematic manipulation of framing: the same objective situation is described in different ways, and the researcher observes whether the description changes behavior. The classic demonstration is the Asian disease problem, in which people choose differently depending on whether outcomes are described in terms of lives saved or lives lost, even though the objective outcomes are identical.
The experimental method has also been extended beyond the laboratory. Field experiments, conducted in real-world settings with real decision-makers, test whether the laboratory findings generalize. Natural experiments, in which the researcher exploits a naturally occurring change in circumstances, provide evidence from situations that could not be experimentally manipulated. Neuroeconomics, a newer subfield, uses brain imaging to observe the neural correlates of decision processes. These methods are complementary: laboratory experiments provide control, field experiments provide realism, and neuroimaging provides a window into the underlying mechanisms.
Contemporary behavioral decision theory is not a single unified theory but a family of approaches that share a common subject matter and a common commitment to empirical evidence. The field has moved beyond the early debates about whether people are rational or irrational. The more productive question is: under what conditions do which decision processes operate, and what are the consequences?
Several developments characterize the current landscape. One is the increasing sophistication of formal models. Prospect theory has been extended and refined, and new models of choice have been developed that incorporate reference dependence, loss aversion, and probability weighting in more flexible ways. Another development is the integration of behavioral decision theory with other fields. Behavioral economics applies the findings to market behavior, public policy, and finance. Behavioral finance explains anomalies in asset prices, such as excessive volatility and predictable patterns in returns, by reference to investor psychology. Behavioral law and economics examines how legal rules and institutions can be designed to account for systematic decision errors.
A third development is the rise of "nudging" as a policy application. If people systematically make decisions that are not in their own interest—failing to save for retirement, choosing unhealthy foods, forgetting to enroll in insurance—then policymakers can design the choice environment to steer people toward better outcomes without restricting their freedom. This application, associated with Richard Thaler and Cass Sunstein, has been influential in government policy in several countries. It has also been controversial, raising questions about paternalism, manipulation, and the legitimacy of using behavioral insights to influence citizens.
A fourth development is the growing attention to individual differences and cultural variation. Early behavioral decision theory tended to treat biases as universal features of human cognition. More recent research has shown that the size and even the direction of some biases vary across individuals and across cultures. This variation does not undermine the core findings, but it complicates the picture. The same person may be highly susceptible to framing effects in one domain and resistant in another; the same bias may be strong in one culture and weak in another.
The field also faces unresolved questions. One is the relationship between the different research traditions. The heuristics-and-biases program, the bounded rationality tradition, and the formal modeling approach often study the same phenomena from different angles, but they have not been fully integrated. Another unresolved question is the boundary between descriptive and normative claims. Behavioral decision theory describes how people do decide, but it does not settle the question of how they should decide. A person who is subject to loss aversion may be making a mistake, or may be expressing a legitimate preference for stability; the descriptive finding alone cannot tell us which.
The most durable contribution of behavioral decision theory is not any single finding or model. It is the demonstration that the study of decision-making must be an empirical science. The question of how people choose is not settled by reflection, by introspection, or by mathematical derivation from axioms of rationality. It is settled by observation. The field has built a substantial body of evidence about the regularities in human judgment and choice, and it has built the theoretical tools to organize that evidence. The result is a picture of human decision-making that is more complex than the rational agent of classical economics, but also more structured than the image of a purely erratic or impulsive creature. People are neither fully rational nor randomly irrational; they are systematic, and the task of behavioral decision theory is to characterize that system.