The Paradigm Shift: From Homo Economicus to Bounded Rationality

The edifice of neoclassical economics has long rested upon a singular, elegant foundation: the axiom of *Homo economicus*—the hypothetical agent who, endowed with stable preferences, unlimited cognitive capacity, and complete information, consistently maximizes utility through a process of unemotional calculation. This construct, refined over a century into the formal machinery of expected utility theory and rational choice models, yielded unprecedented theoretical progress, providing a coherent language for analyzing markets, welfare, and resource allocation. Yet, beneath this surface of mathematical rigor, a profound disconnect emerged between the predicted behaviors of these idealized agents and the observed choices of actual human beings. The rise of experimental economics in the mid-twentieth century, coupled with the importation of psychological principles, began to expose systematic, predictable deviations from rationality that could not be dismissed as mere noise. These deviations were not random errors around a rational mean but rather structured patterns—biases—that suggested a different underlying decision-making architecture. This chapter narrates the intellectual pivot from the world of perfect rationality to the realm of bounded rationality, a shift that birthed behavioral economics as a distinct field. We will critically deconstruct the philosophical origins and formal axioms of expected utility theory, highlight their empirical vulnerabilities through seminal paradoxes, and introduce Herbert Simon’s foundational concept of bounded rationality as a necessary corrective. The narrative will then accelerate through the programmatic research of Daniel Kahneman and Amos Tversky, whose collaborative work crystallized into prospect theory and identified anchoring, loss aversion, and framing as the empirical pillars of the new paradigm. Finally, we will confront the major critiques leveled from both within and outside the behavioral camp, questioning the scope, stability, and normative implications of these cognitive biases. The journey from *Homo economicus* to bounded rationality is not merely an academic correction; it represents a fundamental reorientation in how we understand human agency, with profound implications for policy, finance, and the very definition of rational action.

The Philosophical and Formal Architecture of Rational Choice

The assumption of rational self-interest did not originate in economics but was inherited from the Enlightenment’s philosophical project. René Descartes’s *cogito* established the thinking subject as the foundation of certainty, while Immanuel Kant’s notion of practical reason posited that moral and practical decisions should be derived from universalizable maxims through dispassionate deliberation. This Cartesian-Kantian tradition valorized consistency, coherence, and the subordination of passion to reason. In the utilitarian lineage of Jeremy Bentham and John Stuart Mill, this philosophical stance found its economic expression: individuals were seen as utility-maximizers, with the “greatest happiness principle” serving as a normative guide for both personal choice and social welfare. The “felicific calculus” implied that preferences could, at least in principle, be placed on a cardinal scale and that rational action consisted in choosing the alternative with the highest expected utility.

The formal translation of these philosophical ideas into economic models occurred during the marginal revolution of the late nineteenth century, through the independent works of William Stanley Jevons, Carl Menger, and Léon Walras. They introduced the concept of marginal utility and provided a mathematical framework for consumer choice. However, the definitive axiomatization of rationality under uncertainty arrived with John von Neumann and Oskar Morgenstern’s *Theory of Games and Economic Behavior* (1944) and Leonard Savage’s *The Foundations of Statistics* (1954). The von Neumann-Morgenstern (vNM) utility theorem derived expected utility maximization from a set of seemingly plausible axioms governing preferences over lotteries (risky prospects). These axioms are:

1. Completeness: For any two lotteries $A$ and $B$, the decision-maker can express a preference: $A \succeq B$, $B \succeq A$, or $A \sim B$ (indifference). 2. Transitivity: If $A \succeq B$ and $B \succeq C$, then $A \succeq C$. 3. Continuity: If $A \succeq B \succeq C$, then there exists a probability $p \in [0,1]$ such that the decision-maker is indifferent between $B$ and a lottery that yields $A$ with probability $p$ and $C$ with probability $1-p$. 4. Independence (or Savage’s Sure-Thing Principle): If $A \succeq B$, then for any third lottery $C$ and any $0 < p < 1$, the mixed lottery $pA + (1-p)C \succeq pB + (1-p)C$. This axiom is crucial; it implies that preferences between $A$ and $B$ are independent of what happens in the common consequence $C$.

From these axioms, it follows that there exists a cardinal utility function $u(\cdot)$ such that the decision-maker chooses the lottery with the highest *expected utility*: $$ EU(L) = \sum_{i=1}^{n} p_i u(x_i) $$ where $p_i$ is the probability of outcome $x_i$. Savage extended this to subjective probabilities, assuming that individuals update beliefs via Bayes’s rule and that probabilities obey the standard axioms of probability calculus.

This formal structure provided an unparalleled tool for prediction: given a specification of the utility function and beliefs, the model yields a unique, consistent choice. It became the bedrock of microeconomic theory, finance (CAPM, option pricing), and game theory. Its power lies in its parsimony and internal consistency. However, its very elegance rests on assumptions that, upon closer inspection, describe a being of almost superhuman capacities: infinite computational power to calculate expected utilities, perfect knowledge of all relevant probabilities and outcomes, the ability to hold complex lotteries in mind, and a stability of preferences unaffected by the context of decision or the manner of presentation. The *Homo economicus* is a creature of pure logic, untethered from the flesh-and-blood constraints of human cognition.

Empirical Cracks in the Axiomatic Foundation: Paradoxes and Anomalies

The first serious empirical challenges came from within economics itself, through “paradoxes of choice” that demonstrated systematic violations of the independence axiom. The most famous is the Allais Paradox (1953), which Maurice Allais presented to a group of leading economists, including Savage. Consider two pairs of choices:

Choice 1:

• Option A: 1,000,000 Francs for sure.
• Option B: 10% chance to get 5,000,000 Francs, 89% chance to get 1,000,000 Francs, 1% chance to get nothing.

Choice 2:

• Option C: 11% chance to get 1,000,000 Francs, 89% chance to get nothing.
• Option D: 10% chance to get 5,000,000 Francs, 90% chance to get nothing.

Most people strongly prefer A over B (risk aversion for a sure gain) and also prefer D over C (risk seeking to avoid a sure loss). However, if we denote the utility of 1,000,000 as $u(1M)$ and 5,000,000 as $u(5M)$, the expected utilities are:

• $EU(A) = u(1M)$
• $EU(B) = 0.1 \cdot u(5M) + 0.89 \cdot u(1M) + 0.01 \cdot u(0)$
• $EU(C) = 0.11 \cdot u(1M) + 0.89 \cdot u(0)$
• $EU(D) = 0.1 \cdot u(5M) + 0.9 \cdot u(0)$

Preferring A over B implies $u(1M) > 0.1 \cdot u(5M) + 0.89 \cdot u(1M) + 0.01 \cdot u(0)$, which simplifies to $0.11 \cdot u(1M) > 0.1 \cdot u(5M) + 0.01 \cdot u(0)$. Preferring D over C implies $0.1 \cdot u(5M) + 0.9 \cdot u(0) > 0.11 \cdot u(1M) + 0.89 \cdot u(0)$, which simplifies to $0.1 \cdot u(5M) + 0.01 \cdot u(0) > 0.11 \cdot u(1M)$. These two inequalities are mutually exclusive. Thus, the pattern A > B and D > C violates the independence axiom because the common consequence (1M in the first pair, nothing in the second) changes the relative attractiveness of t