Colin Camerer – 1959 Present

Biography

In the history of late twentieth- and early twenty-first-century economic thought, few scholars have disrupted the discipline’s axiomatic foundations as profoundly as Colin F. Camerer. Born in 1959, Camerer has functioned as an intellectual iconoclast and architect, systematically dismantling the hyper-rational abstraction of Homo economicus to construct a scientifically empirical alternative. Where orthodox neoclassical economics historically treated the human mind as an unobservable, hyper-calculating black box governed by invariant mathematical axioms, Camerer has forced economics to confront the messy, computationally bounded, and neurobiologically constrained realities of actual human cognition. His career represents a persistent drive to bridge microeconomic theory, cognitive psychology, experimental methodology, and neuroscience into a unified behavioural science of strategic choice.

Operating primarily from the California Institute of Technology, where he serves as the Robert Kirby Professor of Behavioral Finance and Economics, Camerer did not simply contribute to the behavioral revolution; he directed its expansion into strategic interactions and biological substrate analysis. His seminal works—ranging from the codification of behavioral game theory and the formulation of the Cognitive Hierarchy and Experience-Weighted Attraction models to the pioneering of neuroeconomics and large-scale metascience—have rewritten how economists understand bargaining, market pricing, probability weighting, and strategic contemplation. By insisting that economic hypotheses must withstand laboratory testing, field-level natural experiments, eye-tracking verification, functional neuroimaging scrutiny, and rigorous cross-laboratory replication, he transformed behavioral economics from a collection of observational anomalies into an empirically disciplined, predictive science.

This comprehensive monograph traces the intellectual trajectory, theoretical breakthroughs, methodological innovations, and enduring epistemological legacy of Colin Camerer from 1959 to the present. By examining his accelerated intellectual formation within the quantitative bastions of Johns Hopkins and the University of Chicago, his systematic deconstruction of the classical Nash equilibrium, his exploration of neural valuation pathways, and his leadership in the open science replication movement, this analysis provides an exhaustive assessment of a scholar whose work has fundamentally redefined the study of strategic decision-making.

1. Biographical Background, Early Education, and Intellectual Foundations

1.1 Formative Years and Accelerated Undergraduate Studies

Colin F. Camerer displayed extraordinary intellectual precocity from an early age, demonstrating mathematical and analytical aptitudes that quickly exhausted standard secondary school curricula. His early academic trajectory was marked by rapid acceleration, leading to his admission to Johns Hopkins University at the age of fourteen through programs designed for mathematically gifted youth. Immersed in an elite quantitative environment, Camerer bypassed traditional introductory curricula to engage directly with advanced applied mathematics, probability theory, and formal decision analysis, completing his Bachelor of Arts in quantitative studies in 1977 at just seventeen years of age.

During these undergraduate years at Johns Hopkins, Camerer was exposed to the mathematical rigor of formal decision theory, operations research, and nascent paradigms within cognitive science. The late 1970s was an era when neoclassical economics was solidifying its axiomatic grip over the social sciences, anchored by expected utility theory, rational expectations, and complete markets. Yet, Camerer developed an early skepticism regarding the descriptive accuracy of these models. By cross-referencing mathematical decision models with contemporary developments in cognitive psychology, he noted a recurring divergence between how human beings mathematically ought to optimize and how computational minds actually navigated stochastic environments. This foundational skepticism toward unconditional rational choice theory laid the conceptual bedrock for his career.

1.2 Graduate Training at the University of Chicago Booth School of Business

Following his undergraduate studies, Camerer enrolled in the doctoral program at the University of Chicago Booth School of Business (then the Graduate School of Business). Immersed in the intellectual epicenter of classical free-market economics and rational-actor orthodoxy, Camerer engaged directly with the intellectual traditions established by Milton Friedman, George Stigler, and Eugene Fama. Rather than retreating into purely abstract economic theory, Camerer completed an MBA in finance in 1979 before earning his Ph.D. in behavioral decision theory in 1981 at the age of twenty-two.

At Chicago, Camerer’s primary intellectual anchors were the pioneering psychologists Hillel Einhorn and Robin Hogarth, who led the Center for Decision Research. Einhorn and Hogarth were revolutionizing descriptive decision research by applying psychological principles to business and economic choices. They demonstrated how heuristics, contextual framing, and cognitive biases systematically distort judgment in professional domains. Operating within the physical perimeter of the Chicago School, Camerer experienced the acute intellectual tension between neoclassical dogma—which maintained that competitive markets would rapidly discipline and eradicate any departures from rationality—and the rigorous empirical findings of cognitive psychology demonstrating that human systematic errors were robust, pervasive, and market-resilient. This tension forged his distinctive research program: taking economic theory seriously enough to formalize it, while taking human psychological data seriously enough to rewrite the underlying axioms.

1.3 Early Academic Appointments and Transdisciplinary Synthesis

Upon completing his doctorate, Camerer embarked on a rapid series of academic appointments at top institutions. He served as an assistant professor of decision sciences at Northwestern University’s Kellogg School of Management (1981–1983), where he interacted with leading game theorists and microeconomists, before joining the faculty at the Wharton School of the University of Pennsylvania (1983–1989). He subsequently returned to the University of Chicago as a full professor (1989–1994) before accepting the Axline Professorship at the California Institute of Technology in 1994, where he established his long-term intellectual home.

Throughout these formative appointments, Camerer advanced a transdisciplinary synthesis that had previously been resisted by mainstream economics. Rather than treating psychology as an external critique of economics, he imported cognitive psychological concepts into formal microeconomic models, subjecting them to experimental tests. Influenced by early experimentalists such as Vernon Smith and Charles Plott, Camerer insisted that microeconomic propositions could be empirically validated in controlled laboratory environments. His research from this era systematically tested subjective expected utility, probability weighting anomalies, and dynamic portfolio selections, converting qualitative behavioral observations into mathematically disciplined alternatives to standard expected utility theory.

2. The Foundational Architecture of Behavioral Game Theory

2.1 Critique of Standard Nash Equilibrium and Rational Actor Theory

In standard non-cooperative game theory, the Nash equilibrium assumes that all players possess complete rationality, hold mutually consistent beliefs about one another’s strategies, and execute optimal responses based on those beliefs. Subgame perfection and backward induction further postulate that players can iterate reasoning processes infinitely through complex extensive-form trees without computational or emotional interference. Camerer challenged this orthodoxy by demonstrating that in experimental settings, real human players systematically violate backward induction, fail to coordinate on theoretical equilibria, and routinely deviate from subgame perfection.

Camerer highlighted the critical epistemological distinction between the normative utility of classical game theory (how idealized agents should behave to guarantee mutual best-responses) and its descriptive failure (how biological humans actually calculate and choose under strategic tension). Rather than dismissing laboratory deviations as random error or white noise, Camerer argued that these departures were systematically structured by bounded rationality, computational limits, visceral emotional responses, and social motivations. He pioneered a methodological pivot within game theory from pure mathematical deduction toward laboratory-based inductive testing, requiring that any viable model of strategic interaction account for the cognitive architectures and physiological realities of human decision-makers.

2.2 Publication and Impact of ‘Behavioral Game Theory: Experiments in Strategic Interaction’

In 2003, Camerer published his magnum opus, Behavioral Game Theory: Experiments in Strategic Interaction, under the auspices of Princeton University Press and the Russell Sage Foundation. This 550-page treatise synthesized three decades of disparate experimental economics literature, transforming a collection of counter-examples into a mathematically unified, empirically grounded subdiscipline. The work established behavioral game theory as an autonomous field, systematically categorizing strategic encounters across bargaining games, coordination matrices, signaling dynamics, and social dilemmas.

The impact of Behavioral Game Theory was immediate and profound. Camerer demonstrated that deviations from Nash predictions were not chaotic anomalies, but could be mathematically predicted by incorporating three psychological mechanisms: social preferences (valuing others’ payoffs), bounded iterated thinking (cognitive hierarchies), and dynamic learning (adjusting strategies via experience). The volume provided empirical microeconomists with the mathematical blueprints necessary to structurally estimate psychological parameters within real-world data, directly influencing empirical industrial organization, market design, and macro-level strategic management.

2.3 Social Preferences, Fairness, and Reciprocity in Strategic Settings

A central pillar of Camerer’s game-theoretic critique addressed the assumption of pure, self-interested opportunism. Through extensive experimental implementations of the Ultimatum Game, Dictator Game, and Public Goods Game, Camerer provided robust empirical proof that human beings are fundamentally motivated by other-regarding preferences, equity concerns, and reciprocity. In Ultimatum experiments, responders frequently reject offers below twenty to thirty percent of the total stake—choosing to destroy absolute economic wealth to punish perceived unfairness—while proposers systematically offer substantial fractions of the endowment to preempt retaliation.

Working alongside European collaborators such as Ernst Fehr and Klaus Schmidt, Camerer contributed to the mathematical formalization of inequity aversion models. These models parameterized individuals’ intrinsic disutility regarding both unfavorable and favorable payoff disparities:

$$U_i(x) = x_i – \alpha_i \max{x_j – x_i, 0} – \beta_i \max{x_i – x_j, 0}$$

Crucially, Camerer pushed this paradigm beyond Western university undergraduate pools. In a landmark multi-year collaborative project across fifteen small-scale societies worldwide (ranging from Amazonian foragers to pastoralist nomads), Camerer and his co-authors demonstrated that while self-interested Nash predictions failed universally, the specific manifestations of fairness and reciprocity varied across societies according to their market integration and societal organization. This work established that altruistic punishment, strong reciprocity, and fairness have deep evolutionary roots tailored to sustain cooperation within human cultural ecologies.

3. Cognitive Hierarchy and Level-k Models of Strategic Thought

3.1 Theoretical Framework of the Cognitive Hierarchy Model

Recognizing that players rarely execute infinite iterations of strategic deduction upon first encountering a game, Camerer, along with Teck-Hua Ho and Juin-Kuan Chong, formulated the Cognitive Hierarchy (CH) model in their landmark 2004 paper. The CH model replaced the equilibrium assumption—that all players possess mutually consistent, perfectly aligned beliefs—with a non-equilibrium parametric model of bounded depth-of-thought, structurally organizing human decision-makers into stratified reasoning tiers.

In this framework, Level-0 ($L_0$) players are non-strategic, selecting actions randomly from a uniform probability distribution over the available strategy space. Level-1 ($L_1$) players assume all other participants are $L_0$ and compute an optimal response based on that assumption. Level-2 ($L_2$) players recognize the existence of both $L_0$ and $L_1$ agents, forming subjective internal distributions over their frequencies and best-responding accordingly. More generally, a Level-$k$ player assumes that their opponents are distributed across lower-order levels ($0, 1, dots, k-1$). Camerer and his colleagues modeled the population frequency of these reasoning steps using a one-parameter Poisson distribution:

$$f(k) = \frac{e^{-\tau} \tau^k}{k!}$$

Here, the single parameter $tau$ (tau) represents both the mean and the variance of reasoning depth within the population. By calibrating this model against experimental datasets, Camerer demonstrated that a $tau$ value clustered between 1.5 and 2.0 accurately predicts human behavior across diverse game matrices, resolving many out-of-equilibrium anomalies and empirical coordination failures that standard Nash frameworks could not explain.

3.2 Empirical Applications: The p-Beauty Contest Game

The definitive empirical battleground for testing the Cognitive Hierarchy and Level-$k$ frameworks was the Keynesian p-Beauty Contest Game. In this classic experimental paradigm, $N$ participants choose an integer between 0 and 100. The winner is the player whose selected number is closest to a given fraction $p$ (frequently $p = 2/3$) of the mean of all chosen numbers. Under classical game-theoretic analysis, the unique subgame perfect Nash equilibrium is 0, reached through an infinite cascade of iterated dominance: if the average cannot exceed 100, the maximum target is 67; recognizing this, no rational agent will choose above 45; this logic iterates downward until it converges to zero.

Camerer executed and analyzed this game across dozens of distinct subject demographics, including college students, corporate CEOs, financial portfolio managers, high-school students, and professional game theorists. The results systematically falsified the classical equilibrium assumption while confirming Cognitive Hierarchy predictions. Rather than choosing 0, raw choices spiked predictably at 33 ($2/3$ of the $L_0$ average of 50, reflecting $L_1$ reasoning), 22 ($2/3$ of 33, reflecting $L_2$ reasoning), and occasionally 15 ($L_3$ reasoning). While trained game theorists and finance professionals exhibited slightly higher $tau$ parameters, the population mean remained centered around 1.5 to 2 steps. Camerer linked these findings directly to financial markets, demonstrating that successful speculative asset trading does not reward infinite rationality (choosing 0), but rather the ability to forecast the immediate next cognitive step of the market median (choosing 22 or 33).

3.3 Policy and Mechanism Design Implications of Iterated Thinking

The empirical validation of the Cognitive Hierarchy model had immediate implications for economic mechanism design, auction theory, and public policy. Standard mechanism design proceeds from the “revelation principle” and assumes that participants will play Bayesian Nash equilibria. However, Camerer demonstrated that when economic mechanisms—such as spectrum auctions, procurement bidding, or emissions cap-and-trade exchanges—are implemented, non-equilibrium bidding driven by low-level ($L_1$ and $L_2$) thinking can produce severe revenue shortfalls, inefficient allocations, or unexpected winner’s curses.

Applying bounded depth-of-thought models allowed economists to engineer “foolproof” mechanisms that are robust to non-equilibrium play. In transportation network economics and traffic flow coordination, Camerer’s models explained why dynamic highway signaling often triggered over-concentrated route-switching, as drivers clustered at low cognitive levels when responding to congestion warnings. In industrial organization, Cognitive Hierarchy clarified why pricing wars persist in newly deregulated markets: competitor firms often assume their rivals are less strategically sophisticated than themselves, underestimating the speed with which competitors will match aggressive undercutting. By integrating the $tau$-parameter into structural models, economists can anticipate strategic friction and design more resilient mechanisms.

4. Dynamic Learning Models: The Experience-Weighted Attraction (EWA) Framework

4.1 Bridging Reinforcement and Belief Learning Paradigms

While the Cognitive Hierarchy model illuminates initial, one-shot strategic choices, human agents placed in recurring strategic interactions adapt their behavior over time. Throughout the 1990s, the experimental literature on strategic adaptation was divided into two competing paradigms: reinforcement learning (derived from behavioral psychology, wherein players simply repeat strategies that yielded positive payoffs in the past) and belief learning (derived from classical economics, such as Cournot dynamics or fictitious play, wherein players form explicit probability distributions over opponents’ potential actions and choose best-responses).

In 1999, Colin Camerer and Teck-Hua Ho unified these competing approaches by creating the Experience-Weighted Attraction (EWA) learning framework. Camerer and Ho demonstrated that reinforcement and belief learning were not mutually exclusive theories, but rather extreme parametric special cases of a broader, mathematically unified cognitive architecture. The core innovation of EWA was the introduction of the “imagination parameter,” $\delta$ (delta). While standard reinforcement learning only reinforces strategies that were physically chosen and rewarded, the EWA model tracks the foregone payoffs of unchosen strategies—what the player *would* have received had they chosen a different action:

$$A_i^j(t) = \frac{\phi \cdot N(t-1) \cdot A_i^j(t-1) + \left[\delta + (1 – \delta) \cdot I(s_i^j, s_i(t))\right] \pi_i(s_i^j, s_{-i}(t))}{N(t)}$$

Here, $A_i^j(t)$ represents the internal cognitive “attraction” player $i$ has toward strategy $j$ at time $t$. The parameter $phi$ measures memory decay, $N(t)$ represents the cumulative weight of experience, and $I(s_i^j, s_i(t))$ is an indicator function tracking whether strategy $j$ was actually chosen. When $\delta = 0$, the player ignores unchosen counterfactuals, collapsing EWA into a pure law-of-effect reinforcement model. When $\delta = 1$, the player weights foregone payoffs identically to real payoffs, transforming the model into generalized fictitious play belief learning. By parameterizing the human capacity for counterfactual reasoning, EWA provided a comprehensive mathematical framework for strategic learning.

4.2 Econometric Estimation and Model Comparison

To substantiate the validity of the EWA framework, Camerer and Ho subjected the model to rigorous econometric estimation, avoiding pure simulations in favor of empirical maximum likelihood estimations applied to granular, individual-level dynamic experimental data. Using logit choice probability functions linked to attraction vectors:

$$P_i^j(t+1) = \frac{e^{\lambda \cdot A_i^j(t)}}{\sum_{k} e^{\lambda \cdot A_i^k(t)}}$$

Camerer tested the model against extensive experimental datasets spanning normal-form matrix games, extensive-form bargaining problems, coordination dilemmas, and zero-sum competitive matches (such as patent races and matching pennies).

Crucially, Camerer evaluated EWA via rigorous out-of-sample prediction metrics. Classical models of learning frequently suffered from in-sample overfitting, looking convincing retroactively but failing to forecast dynamic play when tested against new experimental cohorts or variations in payoff structures. EWA consistently outperformed pure reinforcement models and standard Cournot fictitious play in out-of-sample forecasting accuracy. By estimating the structural parameters $(\phi, \delta, \rho, \lambda)$ across games, Camerer proved that human learning is hybrid: economic actors actively simulate counterfactual realities ($\delta > 0$) while progressively decaying outdated experiences through a memory-discounting mechanism ($phi < 1$).

4.3 Self-Tuning EWA and Modern Extensions

Despite its empirical success, the original EWA formulation faced an economic critique: it contained multiple free parameters that had to be econometrically estimated, limiting its predictive utility in new, uncalibrated economic environments. In response to this challenge, Camerer, Ho, and Chong developed the “Self-Tuning EWA” model in 2007. This iteration eliminated the free parameters by replacing them with dynamically adjusting functional representations linked to environmental information and cognitive attention.

In Self-Tuning EWA, the parameter tracking counterfactual attention ($\delta$) and the memory decay parameter ($phi$) are endogenized as state-dependent functions of the environment. The decay parameter $\phi_i(t)$, for instance, is driven by a change-detector index that monitors whether opponents are playing stationary or shifting strategies. If an opponent changes their behavior, the change-detector spikes, driving $phi$ down to instantly discount historical data and allow rapid adaptation. In contrast, if the strategic environment remains stable, $phi$ approaches unity, reinforcing stable beliefs. Modern extensions of this work have integrated Self-Tuning EWA into multi-agent machine learning architectures, automated algorithmic trading engines, and artificial intelligence frameworks designed to interact with boundedly rational human counterparts in real-time pricing and security markets.

5. Pioneering Neuroeconomics: Conceptual Framework and Methodological Synthesis

5.1 The Epistemological Rationale for Neuroeconomics

By the turn of the millennium, Colin Camerer realized that behavioral economics had reached a methodological boundary. While field studies and laboratory experiments had successfully mapped the boundaries of classical rationality, economists remained restricted to behavioral inputs and outputs, inferring mental structures without directly observing them. Economists continued to rely on the concept of “revealed preference”—a conceptualization popularized by Paul Samuelson that treated internal neurocognitive mechanisms as an unknowable black box. Neoclassical purists argued that as long as a mathematical model could predict choice outputs, the physical validity of the internal computational mechanisms was irrelevant to economic science.

Camerer decisively rejected this methodological instrumentalism. Partnering with cognitive neuroscientists, computational biologists, and forward-looking economists like George Loewenstein and Drazen Prelec, Camerer argued that the human brain’s neural architecture constitutes the foundational economic primitive. In an era where functional neuroimaging, electrophysiology, and computational neuroscience were decrypting the biophysical mechanics of information processing, economists could no longer afford to rely on biologically implausible utility constructs. Camerer argued that directly measuring the brain’s internal computational variables—such as reward prediction errors, somatic stress signals, and visual attention—would resolve long-standing theoretical stalemates, allowing economists to distinguish between models that make identical behavioral predictions but postulate fundamentally different underlying mechanisms.

5.2 Neuroimaging Technologies in Economic Science

To implement this research program, Camerer helped construct a neuroeconomic methodology that mapped economic variables onto physical neural substrates. The primary instrument in this empirical pivot was functional Magnetic Resonance Imaging (fMRI). By measuring blood-oxygen-level-dependent (BOLD) signal changes in specific brain areas, Camerer and his collaborators observed real-time hemodynamic responses as human subjects calculated mixed strategies, evaluated financial gambles, and navigated bargaining interactions inside scanners.

Recognizing the temporal limits of fMRI (which tracks hemodynamic responses across seconds rather than milliseconds), Camerer incorporated complementary methodologies to build a multi-layered measurement apparatus:

• Electroencephalography (EEG) and Event-Related Potentials (ERPs): Deployed to track high-temporal-resolution electrical fluctuations during strategic decision-making and loss realization.
• Galvanic Skin Response (GSR) and Pupillometry: Used to record autonomic sympathetic nervous system arousal and pupil dilation during deception, high-stakes risk evaluation, and asset bubble market corrections.
• Transcranial Magnetic Stimulation (TMS) and Transcranial Direct-Current Stimulation (tDCS): Utilized to move beyond pure correlational imaging to establish direct causal links. By temporarily disrupting targeted cortical regions (such as the dorsolateral prefrontal cortex), Camerer’s teams tested whether specific neural networks were mechanistically necessary to enforce fairness, override selfish impulses, or compute iterated dominance.

5.3 Foundational Neuroeconomic Publications and Debates

The neuroeconomic movement reached mainstream economic discourse with the 2005 publication of Camerer, Loewenstein, and Prelec’s landmark paper in the Journal of Economic Literature, “Neuroeconomics: How Neuroscience Can Inform Economics”. This 56-page paper outlined the conceptual architecture for the field. The authors challenged the standard view that human decision-making is controlled by a single, unified cognitive utility-maximizing engine, replacing it with a bi-modular neurobiological matrix divided along two primary axes: controlled versus automatic processes, and cognitive versus affective processes.

This formulation provoked significant debate among traditional economists, culminating in the prominent critique by Princeton economic theorists Faruk Gul and Wolfgang Pesendorfer in their paper “The Case for Mindless Economics.” Gul and Pesendorfer argued that neuroscience was irrelevant to economics because economic hypotheses only evaluate choices made in the marketplace, not the physiological processes within the skull. Camerer robustly countered this view, demonstrating that understanding neural mechanisms enables economists to predict when standard choice models will systematically break down under specific physiological triggers—such as stress, cognitive fatigue, emotional arousal, or chemical addiction. Furthermore, Camerer helped solidify the global infrastructure of this emerging subdiscipline by serving as a founding figure in the Society for Neuroeconomics, advancing an empirical paradigm where economic models must align with biological realities.

6. Neural Substrates of Value, Risk, and Ambiguity

6.1 Neural Valuation and the Dopaminergic Reward Pathway

A core breakthrough in neuroeconomics, heavily driven by Camerer’s work in collaboration with neuroscientists such as Wolfram Schultz and Antonio Rangel, was the discovery of a common neural currency for subjective economic value. Classical microeconomics assumes that consumers evaluate heterogeneous goods—such as comparing an apple to a financial bond—through a universal utility metric, but this construct had historically been viewed as an abstract mathematical convenience. Through neuroimaging paradigms, Camerer helped demonstrate that the human brain physically instantiates a common utility metric within specific, interconnected dopaminergic networks.

This common valuation network is centered primarily in the ventromedial prefrontal cortex (vmPFC) and the ventral striatum (incorporating the nucleus accumbens). These regions compute and encode real-time subjective values, scaling their activation directly with an agent’s reported willingness-to-pay (WTP) across diverse domains, including food choices, charitable contributions, and financial assets. Working with computational implementations of the reward prediction error (RPE) derived from reinforcement learning algorithms, Camerer mapped how midbrain dopamine neurons systematically update valuation vectors:

$$\text{RPE} = R(t) – V(t)$$

Where $R(t)$ is the realized biological or economic reward and $V(t)$ is the expected reward. When choices involve tradeoffs between immediate consumption and delayed rewards (intertemporal choice), Camerer showed that hyperbolic discounting is biological: the ventral striatum and limbic structures drive impulsive, present-biased choices, while the lateral prefrontal cortex executes cognitive control to discount future payoffs more consistently.

6.2 The Neural Divergence of Risk and Knightian Ambiguity

One of Camerer’s most celebrated experimental and neuroimaging investigations tackled the long-standing microeconomic distinction between known risk (where probability distributions are known, such as rolling a fair die) and Knightian ambiguity (where probabilities are unknown, as articulated in the Ellsberg Paradox). Standard expected utility theory and Bayesian models assume that decision-makers assign subjective probabilities to ambiguous events, treating risk and ambiguity through identical computational pathways.

In a seminal 2005 paper published in Science, Ming Hsu, Meghana Bhatt, Ralph Adolphs, Daniel Tranel, and Colin Camerer exposed human subjects to both risky and ambiguous financial gambles while scanning them with fMRI, alongside complementary studies of patients with targeted brain lesions. The neural data conclusively proved that the human brain does not treat risk and ambiguity identically. When evaluating ambiguous options, the brain exhibits heightened activation in the amygdala and the anterior insula—regions heavily associated with vigilance, fear, and somatic discomfort. In contrast, choices governed by known probabilistic risk exhibited elevated engagement in the dorsal striatum, tracking expected mathematical values.

Camerer corroborated these functional findings by examining neurological patients with focal damage to the amygdala and orbitofrontal cortex: these patients were ambiguity-neutral, evaluating gambles based purely on mathematical payoffs. This provided definitive causal proof that ambiguity aversion is not a general computational preference, but a distinct biological reaction triggered by missing information. This finding offered a biological foundation for non-expected utility formulations, such as Maxmin Expected Utility and smooth ambiguity models.

6.3 Theory of Mind and Strategic Brain Networks

Beyond individual decision-making, Camerer advanced the neurobiology of strategic interactions, or what he termed “Theory of Mind” (ToM) within game theory: the human ability to mentally model the intentions, beliefs, and emotional states of an adversary. In classical game theory, players are treated as abstract calculators who simulate opponents’ strategies automatically. Camerer used neuroimaging to demonstrate that strategic calculations activate a distinct social-cognitive brain network that is physically distinct from general mathematical calculation pathways.

Through strategic matrix games, Camerer’s scans revealed that calculating a rival’s move consistently engages the temporoparietal junction (TPJ), the medial prefrontal cortex (mPFC), and the precuneus. In controlled experiments contrasting human-versus-human gameplay with human-versus-computer gameplay (where the computer executed identical mathematical algorithms), Camerer demonstrated that these mentalizing networks activate exclusively when subjects believe they are facing an intentional, biological human opponent. The mPFC tracks subtle expectations of strategic deception, while the anterior insula and anterior cingulate cortex register elevated activity when a player encounters social betrayal or unfair offers in bargaining. This research established that human strategic interaction is deeply anchored in evolutionary adaptations for navigating social hierarchies and preserving collective reputation, rather than general-purpose arithmetic deduction alone.

7. Empirical Field Experiments and Market Anomalies

7.1 Labor Supply Decisions: The NYC Taxi Driver Studies

While the laboratory provided a controlled setting for testing theoretical axioms, Colin Camerer insisted that behavioral models must also account for high-stakes decisions in real-world markets. In 1997, alongside Linda Babcock, George Loewenstein, and Richard Thaler, Camerer published a groundbreaking empirical investigation into the labor supply decisions of New York City taxi drivers, which became an iconic study in applied behavioral economics.

Standard neoclassical labor economic models assume that workers optimize intertemporally: they should work longer hours on days when their marginal wage rate is high (such as rainy days or busy holidays) and cut back on hours when the wage rate is low (a positive wage elasticity of labor supply). Camerer and his co-authors examined daily trip sheets containing granular data on hours worked and fares collected. Strikingly, they discovered an empirical *negative* wage elasticity of labor hours. Cab drivers frequently worked long hours on slow days when wages were low, and cut their shifts short on high-demand, high-wage days once they accumulated their daily earnings target.

To explain this behavior, Camerer and his colleagues applied Daniel Kahneman and Amos Tversky’s Prospect Theory to labor supply. They demonstrated that drivers operate using narrow framing and daily mental accounting: they set a daily monetary reference point (target income). When drivers are below that target, they perceive themselves as operating in the psychological domain of losses; driven by steep loss aversion, they endure long, tiring hours to break even. Once they cross their target threshold, they enter the domain of diminishing marginal gains; experiencing lower subjective utility for subsequent dollars, they cease working early. While this paper sparked significant methodological debates—such as those by Henry Farber, who questioned the econometric specifications and driver experience levels—subsequent high-resolution analyses using modern rideshare and dispatch data have repeatedly confirmed that reference-dependent, loss-averse stopping rules are prevalent among non-salaried, flexible-schedule workers.

7.2 Laboratory Asset Bubbles and Speculative Behavior

In market dynamics, Camerer made extensive contributions to understanding speculative asset price bubbles. Extending the classic experimental asset market framework developed by Vernon Smith, Gerry Suchanek, and Arlington Williams, Camerer engineered complex, multi-period laboratory markets where the underlying fundamental value of an asset was fully observable, deterministic, and systematically declined toward zero as the trading periods elapsed. Standard rational expectations models predict that prices should track this fundamental trajectory via backward induction.

Camerer demonstrated that even when the underlying fundamental valuation trajectory is explicitly explained to all participants, laboratory markets routinely generate massive speculative bubbles, with transaction prices spiking far above fundamentals before crashing precipitously near the end of the asset’s lifecycle. Camerer investigated the psychological and physiological drivers of this phenomenon, integrating physiological monitoring and eye-tracking into trading environments. His findings demonstrated that asset bubbles are not merely driven by cognitive confusion or ignorance, but by a strategic dynamic he termed the “riding the bubble” phenomenon: sophisticated traders often recognize that an asset is grossly overvalued, but intentionally buy into the bubble because they believe they can time the market and sell to a less sophisticated trader (a Level-1 agent) before the eventual crash.

Furthermore, Camerer demonstrated that introducing short-selling mechanisms and structural experience mitigates, but does not fully eradicate, these speculative excesses. Scanner studies tracking subjects during asset market trading revealed that during the formation of a bubble, traders exhibit elevated insular activation—a biological warning signal of impending risk and volatility—yet many continue trading aggressively, displaying physiological herd dynamics that override rational risk budgeting until the bubble collapses.

7.3 Betting Markets, Pari-Mutuel Wagering, and Probability Weighting

Field data from racetrack betting and sports wagering provided Camerer with another empirical environment to evaluate subjective probability transformations. In extensive empirical evaluations of pari-mutuel wagering systems, Camerer documented the persistent existence of the favorite-longshot bias: bettors systematically undervalue horses that have a high mathematical probability of winning (the favorites) while overvaluing horses with remote chances of victory (the longshots).

Neoclassical finance had struggled for decades to explain this phenomenon, often relying on implausible assumptions regarding the extreme risk-loving preferences of everyday gamblers. Camerer demonstrated that the favorite-longshot bias can be mathematically explained using the non-linear probability weighting function $w(p)$ formalized in Cumulative Prospect Theory, particularly as formulated by Drazen Prelec:

$$w(p) = \exp\left( -(-\ln p)^\alpha \right)$$

Human beings display an inverted S-shaped probability weighting curve: they systematically overweight low probabilities (converting an objective 1% probability into a subjective 5% weight) and underweight moderate-to-high probabilities. At the racetrack, this cognitive bias leads to an over-allocation of capital toward longshots. Camerer expanded these insights to real-world prediction markets, options pricing anomalies, and state lotteries, demonstrating that non-linear probability weighting is a robust feature of financial decision-making that influences asset prices across diverse settings.

8. Behavioral Finance, Heuristics, and Decision Biases

8.1 Prospect Theory in Strategic and Financial Decisions

Throughout his career, Camerer has been a primary architect in transitioning Prospect Theory from a descriptive model of laboratory gambles into a tool for understanding financial markets. Neoclassical finance, built on the Capital Asset Pricing Model (CAPM) and the Efficient Market Hypothesis (EMH), assumes that investors evaluate portfolio risks globally, with asset allocations driven by symmetric risk aversion over terminal wealth. Camerer dismantled this approach by showing that financial market participants process investments through narrow framing, mental accounting, and acute loss aversion.

Loss aversion is parameterized by the coefficient $lambda$ (lambda), typically estimated near $\lambda \approx 1.5 – 2.5$, meaning that the subjective psychological pain of a monetary loss is more than twice as intense as the pleasure of an equivalent gain:

$$v(x) = \begin{\cases} x^\alpha & \text{for } x ge 0 \ -\lambda (-x)^\beta & \text{for } x < 0 \end{\cases}$$

Camerer deployed this framework to clarify the disposition effect—the empirical tendency of retail and institutional investors to prematurely sell winning assets to secure a certain gain (operating in the risk-averse domain of gains) while holding losing assets indefinitely to avoid realizing a loss (operating in the risk-seeking domain of losses). He demonstrated that reference points are dynamic rather than static: as asset prices fluctuate, investors constantly update their mental purchase baselines, causing their risk tolerance to shift depending on their recent paper gains or losses.

8.2 Overconfidence, Optimism, and Hubris in Corporate Environments

In his 1999 paper with Dan Lovallo, “Overconfidence and Excess Entry: An Experimental Approach”, published in the American Economic Review, Camerer addressed a classic economic puzzle: why do thousands of new businesses launch every year when empirical data shows that the vast majority fail within five years? Standard economic theory dictates that rational entrepreneurs should forecast competitive entry and avoid investing in overcrowded industries where profits fall below the cost of capital.

Camerer and Lovallo designed an experiment that isolated the causal drivers of business entry. They allowed subjects to enter a market where payouts depended either on random chance or on the subjects’ actual skill (such as general knowledge or logical reasoning). The authors discovered that when market success was framed as dependent on individual competence, entry rates escalated dramatically, leading to negative aggregate payoffs for entrants. This phenomenon was not caused by simple risk-seeking preferences, but by two cognitive biases:

• Reference Group Neglect: Entrants focused on their own perceived skill while ignoring the reality that their competitors were equally skilled.
• Comparative Optimism: A pervasive belief among most participants that their abilities ranked in the top percentiles of the competitive distribution.

Camerer expanded these findings to corporate finance, demonstrating that CEO hubris and the “curse of knowledge” (the inability of an expert to model the perspective of an uninformed counterparty) systematically distort mergers and acquisitions. Many value-destroying corporate takeovers are driven by executive overconfidence, where leadership teams assume their operational abilities will easily eliminate complex corporate inefficiencies, underestimating the strategic frictions of integration.

8.3 Heuristic Valuation and Fast-and-Frugal Economic Choice

Working at the intersection of cognitive psychology and economics, Camerer explored how agents deploy heuristics to navigate information overload. Challenging the idea that heuristics are merely defective departures from optimal rationality, Camerer drew on the work of Herbert Simon and Gerd Gigerenzer to show that heuristics often represent “ecologically rational” adaptations—computationally efficient rules designed to make robust decisions under severe time and processing constraints.

In consumer environments, Camerer demonstrated that choice processes do not involve exhaustive, compensatory evaluations of every possible product attribute. Instead, consumers use non-compensatory heuristics, such as the *lexicographic rule* (filtering options based on the single most salient dimension) or the *satisficing threshold*. Using eye-tracking and computational search methodologies, he showed that visual salience, default settings, and cognitive load heavily shape product evaluations. This research laid the groundwork for policy interventions often described as “nudges” or asymmetric paternalism—a framework Camerer co-developed in 2003 with Samuel Issacharoff, George Loewenstein, Ted O’Donoghue, and Matthew Rabin. Asymmetric paternalism argues for structural regulatory interventions that offer significant benefits to boundedly rational individuals while imposing little to no cost on fully rational actors.

9. Process Tracing and Methodological Innovations

9.1 Information Board and Mouse-Tracking Architectures

To establish how economic decisions are formed, Camerer insisted that economists must observe the intermediate cognitive steps that precede an eventual choice. In the 1990s, along with Eric Johnson and other collaborators, Camerer transformed process-tracing methodology by developing and deploying computerized information platforms, most notably Mouselab.

In traditional experimental setups, an extensive-form game or payoff matrix is presented fully to the subject, obscuring which elements of the payoff matrix are actually evaluated. The Mouselab architecture hid payoffs behind digital boxes that subjects could reveal only by moving their cursor over a specific box. The system recorded high-resolution data on:

• Information acquisition sequences: The exact sequential path a player used to explore strategic options.
• Dwell times: The duration of cognitive processing spent on each payoff cell.
• Payoff neglect: The complete failure to inspect specific payoff cells within extensive-form trees.

These studies delivered critical insights. For instance, when players were evaluated in multi-stage alternating-offer bargaining games, backward induction required them to first inspect the final-period payoffs to determine terminal values before calculating backwards to round one. The Mouselab data demonstrated that subjects rarely opened the final-round payoff boxes at the start of the interaction. Instead, they began by evaluating first-round payoffs and calculating fair splits, mathematically validating that backward induction fails as a cognitive reality, and revealing that strategic deviations stem from how information is sequentially searched.

9.2 Eye-Tracking and Visual Salience in Strategic Interactions

As micro-sensor technologies evolved, Camerer advanced from computer-mouse tracking to high-speed, non-invasive eye-tracking systems. By tracking gaze fixations, micro-saccades, and pupil dilations at sampling frequencies exceeding 500 Hz, Camerer observed visual attention allocation in matrix games, auctions, and consumer shopping tasks.

These eye-tracking experiments yielded valuable data regarding how individuals implement depth-of-thought models. In Level-$k$ games, eye tracking revealed that Level-1 thinkers dedicate almost no visual fixation time to the payoff cells of their opponents; their gaze fixates almost entirely on their own potential returns across candidate actions. Conversely, players who execute higher-order strategic reasoning (Level-2 and above) exhibit structured saccades alternating between their own payoffs and those of their rivals, visually confirming that strategic depth correlates with specific attentional mechanisms.

Furthermore, Camerer demonstrated that visual salience can resolve strategic coordination challenges. In coordination games with multiple equilibria, subtly manipulating the visual brightness, spatial placement, or color saturation of an option established immediate Schelling points (focal coordination points). This proved that economic coordination can often be steered by sensory cues rather than mathematical deduction alone.

9.3 Biological and Physiological Biomarkers

Camerer expanded his process-tracing methodologies to include biochemical and physiological markers, measuring how autonomic and endocrine variations modulate economic choices. He tracked heart-rate variability (HRV) and skin conductance during high-stakes bargaining, revealing that impending structural deadlocks are preceded by acute spikes in sympathetic autonomic arousal, as the threat of an unfair settlement triggers measurable physiological stress.

In parallel, Camerer explored the endocrine drivers of risk-taking, aggression, and cooperation through targeted hormonal assays and pharmacological interventions. He evaluated how baseline variations and exogenous spikes in testosterone, cortisol, and oxytocin influence strategic profiles:

• Cortisol: Chronic biological stress and elevated cortisol amplify risk aversion in gain frames while deepening cognitive entrenchment during financial crises.
• Testosterone: Exogenous testosterone administration impairs cognitive reflection, increasing reliance on uncalibrated intuitive heuristics and amplifying overconfidence in competitive games.
• Oxytocin: Intranasal administration of oxytocin was shown in social dilemma paradigms to increase financial trust and capital transfers in collaborative interactions, though Camerer remained careful to highlight the contextual boundaries and ecological limits of these hormonal effects.

10. The Metascience Revolution: Replicability and Open Science

10.1 Co-Leadership of the Reproducibility Project: Psychology

During the early 2010s, the behavioral sciences encountered what became widely known as the “replication crisis.” Multiple high-profile experimental claims across social psychology failed to replicate, raising fundamental concerns about standard scientific methodologies. Rather than avoiding these criticisms, Colin Camerer emerged as an active reformer within the open science movement, helping to lead the metascience revolution across the social and behavioral disciplines.

Camerer joined the steering committee of the historic Reproducibility Project: Psychology, coordinated through the Center for Open Science and published in Science in 2015. This massive collaborative effort rigorously replicated 100 empirical studies published in top-tier psychology journals. The results revealed systemic vulnerabilities: only 36% of the replications yielded statistically significant findings, compared to 97% of the original studies, with average effect sizes shrinking by half. Camerer and his colleagues identified publication bias (favoring novel, statistically significant outcomes), underpowered sample sizes, and pervasive “p-hacking” (adjusting analyses post-hoc to achieve $p < 0.05$) as structural drivers of these false positives. The study sparked structural institutional reforms across social science journals, promoting preregistration protocols, open data sharing, and methodological transparency.

10.2 The Social Sciences Replication Project (SSRP) and Economics Replications

Recognizing the need to evaluate economics with equal rigor, Camerer led the Experimental Economics Replication Project (2016) and subsequently the Social Sciences Replication Project (SSRP, 2018), testing studies published in Science and Nature between 2010 and 2015. These projects were conducted under high-powered protocols, with pre-registered analytical designs and sample sizes often several times larger than the original studies.

The findings demonstrated that while experimental economics replicated at a substantially higher rate (~61% to 67%) than social psychology, it remained vulnerable to significant effect-size shrinkage. On average, the effect sizes of successful replications were roughly 50% to 75% of the originally reported magnitudes. Camerer documented that while the fundamental theoretical insights of behavioral economics generally held up, published estimates had frequently been inflated by selective reporting and small sample sizes. Through these systematic metascience projects, Camerer established gold-standard replication architectures, including blinded analytical pipelines, registered reports, and publicly archived data repositories.

10.3 Prediction Markets for Scientific Replicability

Leveraging his deep background in information aggregation and market mechanisms, Camerer integrated his metascience work with behavioral economics by pioneering the use of prediction markets to forecast scientific replication. Prior to conducting high-powered replications, Camerer, Anna Dreber, Magnus Johannesson, and their team established real-money prediction markets, allowing participating researchers and domain experts to trade contracts linked directly to whether a target study would successfully replicate.

The outcomes proved the power of market mechanisms for information aggregation: the prediction markets forecasted replication success with accuracy rates consistently between 70% and 85%, significantly outperforming subjective peer reviews and traditional expert panel evaluations. Contracts tied to unreplicable studies quickly fell toward zero, while solid, high-powered experiments traded at high probabilities. Camerer demonstrated that the collective wisdom of researchers, synthesized through real-money betting mechanisms, can effectively detect weak scientific findings. This work pointed toward alternative institutional approaches for allocating research grants and improving peer review across the scientific ecosystem.

11. Academic Leadership, Caltech Tenure, and Honors

11.1 The Caltech Environment and Interdisciplinary Culture

Colin Camerer’s arrival at the Division of the Humanities and Social Sciences (HSS) at the California Institute of Technology in 1994 provided an optimal institutional ecosystem for his transdisciplinary research program. Caltech’s culture—characterized by a low tolerance for unverified mathematical assumptions, a tradition of experimental physics and biology, and minimal administrative divisions—allowed Camerer to easily bridge economics, computation, and neural engineering.

At Caltech, as the Robert Kirby Professor of Behavioral Finance and Economics, Camerer established world-class experimental laboratories and served as a foundational leader within the Caltech Brain Imaging Center (CBIC). He transformed Caltech into an intellectual hub for neuroeconomics and quantitative behavioral social science, mentoring a generation of doctoral students, postdoctoral scholars, and research fellows who now direct behavioral economics and neuroimaging laboratories worldwide. The Caltech environment enabled Camerer to engage in continuous cross-pollination with computer scientists, theoretical neurobiologists, and applied mathematicians, ensuring that his models incorporated modern developments in computational neuroscience and machine learning.

11.2 The MacArthur Fellowship (‘Genius Grant’) and Professional Accolades

In recognition of his pioneering work, Camerer was awarded the prestigious MacArthur Fellowship (the “Genius Grant”) in 2013. The MacArthur Foundation highlighted his foundational role in establishing behavioral game theory and pioneering neuroeconomics, emphasizing his ability to leverage empirical and biological techniques to expand the explanatory power of economic science.

Beyond the MacArthur Fellowship, Camerer has received numerous academic honors:

• Election as a Fellow of the Econometric Society (1999).
• Induction into the American Academy of Arts and Sciences (2003).
• Election to the National Academy of Sciences (2018).
• Service as President of the Economic Science Association (ESA) (2001–2003).
• Service as President of the Society for Neuroeconomics (2005–2006).

His keynotes across the American Economic Association, the Society for Neuroscience, and international finance congresses have helped shape contemporary empirical agendas in social science.

11.3 Policy Advisory and Real-World Institutional Engagements

Throughout his career, Camerer has applied behavioral insights to corporate strategy, public policy, and institutional design. He has advised behavioral insights teams (often termed “nudge units”), central banking systems, and international regulatory bodies, emphasizing the practical implications of bounded rationality and choice architecture.

His contributions to policy include:

• Asymmetric Paternalism in Consumer Finance: Designing default contribution rules and automatic escalation programs for pension plans, preserving choice autonomy while protecting boundedly rational savers from under-saving.
• Strategic Procurement and Auction Engineering: Advising governmental entities on spectrum and resource auctions to mitigate winner’s curse anomalies driven by non-equilibrium bidding.
• Regulatory Standards for Financial Disclosures: Developing financial risk metrics that account for non-linear probability weighting and reference-dependent framing, helping ensure that retail investors understand asymmetric portfolio risks.
• Evidence-Based Open Science Policy: Consulting with scientific agencies to mandate preregistration, data accessibility, and high-powered statistical protocols across state-funded empirical behavioral programs.

12. Contemporary Frontiers and Future Directions of Camerer’s Work

12.1 Artificial Intelligence, Deep Learning, and Cognitive Modeling

In his recent work, Colin Camerer has engaged directly with the frontiers of artificial intelligence and machine learning. As algorithms play a larger role in coordinating economic systems, Camerer has evaluated the predictive boundaries between classical behavioral formulations and deep neural networks. In landmark papers exploring machine learning applications in behavioral economic datasets, Camerer demonstrated that while deep neural networks can often predict human economic choice from raw inputs with high accuracy, they function as non-interpretable black boxes that lack structural economic primitives.

To resolve this tension, Camerer has explored the synthesis of machine learning architectures with behavioral game theory. By embedding Cognitive Hierarchy and Self-Tuning EWA models directly within the structural nodes of deep neural networks, his research generates models that match the out-of-sample predictive accuracy of machine learning while retaining interpretability. Furthermore, Camerer has integrated these behavioral models into multi-agent reinforcement learning environments, training AI systems to coordinate with boundedly rational human beings in shared mobility systems, algorithmic security exchanges, and human-in-the-loop autonomous platforms.

12.2 Computational Neuroimaging and High-Dimensional Behavioral Data

Technologically, Camerer continues to expand the methodological tools of neuroeconomics by moving beyond standard univariate fMRI analyses. His laboratory utilizes multivoxel pattern analysis (MVPA) and real-time decoded neurofeedback, allowing researchers to track the spatial distributions of neural activity that represent complex strategic variables, such as social deceit, risk tolerances, and subjective certainty, before an overt behavioral choice is executed.

Simultaneously, Camerer is taking neuroeconomics “into the wild.” By leveraging wearable biometric sensors, eye-tracking smart glasses, and mobile physiological monitors, his research captures continuous streams of high-dimensional data from economic actors operating in real-world, high-stress environments—such as foreign exchange traders and emergency logistics coordinators. These initiatives aim to construct dynamic models that trace how physiological exhaustion, biological stress, and cognitive load degrade strategic decision-making in real time, moving computational economics beyond static laboratory sessions.

12.3 The Enduring Epistemological Legacy of Colin Camerer

The overarching legacy of Colin Camerer is marked by a fundamental transformation of economics from a purely axiomatic-deductive discipline into an empirical, mechanistic, and interdisciplinary science. When Camerer began his career in the late 1970s, economics treated psychological data and biological mechanisms as irrelevant to rigorous economic theory. Through persistent theoretical formalization, laboratory experimentation, field studies, neuroimaging innovation, and metascience leadership, Camerer demonstrated that an economic theory cannot be considered complete if it relies on demonstrably false models of human computation.

By establishing behavioral game theory, he provided the tools needed to analyze strategic interactions as they actually unfold in human societies. By founding neuroeconomics, he opened the black box of human valuation, showing that economic choices are anchored in evolutionary biology and neural signaling. And through his leadership in metascience, he helped ensure that behavioral discoveries are supported by rigorous replication architectures. In reshaping the foundations of decision theory, Colin Camerer established an enduring truth: to understand how economies function, we must understand the biological mind that makes decisions within them.

Conclusion

Across more than four decades of research, Colin Camerer has fundamentally challenged and expanded economic science. His intellectual journey—from an accelerated student at Johns Hopkins and Chicago Booth to the Robert Kirby Professorship at Caltech—has bridged disparate academic domains into a cohesive, empirically grounded science of strategic choice. By refusing to accept theoretical convenience at the expense of empirical reality, he transformed behavioral game theory from a collection of experimental paradoxes into an axiomatic subdiscipline, captured in the formulation of the Cognitive Hierarchy and Experience-Weighted Attraction frameworks.

His contributions go far beyond the boundaries of standard economics. By co-founding neuroeconomics, Camerer dismantled the long-standing “revealed preference” black box, revealing the neural circuits that govern value representation, risk processing, Knightian ambiguity, and Theory of Mind. Moreover, his recent focus on open science and prediction markets has helped establish a more replicable, methodologically transparent standard for the social and behavioral sciences. As economics continues to evolve alongside artificial intelligence, computational neuroscience, and large-scale data environments, the empirical foundations laid by Colin Camerer will remain essential to deciphering the complexities of human strategic choice.

References

• Camerer, C. F. (1995). Individual decision making. In J. H. Kagel & A. E. Roth (Eds.), The Handbook of Experimental Economics (pp. 587–703). Princeton University Press.
• Camerer, C. F. (2003). Behavioral Game Theory: Experiments in Strategic Interaction. Princeton University Press.
• Camerer, C. F., Babcock, L., Loewenstein, G., & Thaler, R. (1997). Labor supply of New York City cabdrivers: One day at a time. The Quarterly Journal of Economics, 112(2), 407–441.
• Camerer, C. F., Dreber, A., Forsell, E., Ho, T.-H., Huber, J., Johannesson, M., Kirchler, M., Almenberg, J., Altmejd, A., Chan, T., Heikensten, E., Holzmeister, F., Imai, T., Isaksson, S., Nave, G., Pfeiffer, T., Razen, M., & Wu, H. (2016). Evaluating replicability of laboratory experiments in economics. Science, 351(6280), 1433–1436.
• Camerer, C. F., Dreber, A., Holzmeister, F., Ho, T.-H., Huber, J., Johannesson, M., Kirchler, M., Nave, G., Nosek, B. A., Pfeiffer, T., Altmejd, A., Buttrick, N., Chan, T., Chen, Y., Forsell, E., Gampa, A., Heikensten, E., Hummer, L., Imai, T., … Wu, H. (2018). Evaluating the replicability of social science experiments in Nature and Science between 2010 and 2015. Nature Human Behaviour, 2(9), 637–644.
• Camerer, C. F., & Ho, T.-H. (1999). Experience-weighted attraction learning in normal form games. Econometrica, 67(4), 827–874.
• Camerer, C. F., Ho, T.-H., & Chong, J.-K. (2004). A cognitive hierarchy model of games. The Quarterly Journal of Economics, 119(3), 861–898.
• Camerer, C. F., Issacharoff, S., Loewenstein, G., O’Donoghue, T., & Rabin, M. (2003). Regulation for conservatives: Behavioral economics and the case for ‘asymmetric paternalism’. University of Pennsylvania Law Review, 151(3), 1211–1254.
• Camerer, C. F., & Lovallo, D. (1999). Overconfidence and excess entry: An experimental approach. American Economic Review, 89(1), 306–318.
• Camerer, C. F., Loewenstein, G., & Prelec, D. (2005). Neuroeconomics: How neuroscience can inform economics. Journal of Economic Literature, 43(1), 9–64.
• Dreber, A., Pfeiffer, T., Almenberg, J., Isaksson, S., Xiao, B., Tran, B., & Camerer, C. F. (2015). Using prediction markets to estimate the reproducibility of scientific research. Proceedings of the National Academy of Sciences, 112(50), 15343–15347.
• Fehr, E., & Schmidt, K. M. (1999). A theory of fairness, competition, and cooperation. The Quarterly Journal of Economics, 114(3), 817–868.
• Henrich, J., Boyd, R., Bowles, S., Camerer, C., Fehr, E., Gintis, H., & McElreath, R. (2001). In search of Homo economicus: Behavioral experiments in 15 small-scale societies. American Economic Review, 91(2), 73–78.
• Ho, T.-H., Camerer, C. F., & Chong, J.-K. (2007). Self-tuning experience weighted attraction learning in games. Journal of Economic Theory, 133(1), 177–198.
• Hsu, M., Bhatt, M., Adolphs, R., Tranel, D., & Camerer, C. F. (2005). Neural systems responding to degrees of uncertainty in human decision-making. Science, 310(5754), 1680–1683.
• Open Science Collaboration. (2015). Estimating the reproducibility of psychological science. Science, 349(6251), aac4716.