|
on Cognitive and Behavioural Economics |
|
Issue of 2026–03–02
nine papers chosen by Marco Novarese, Università degli Studi del Piemonte Orientale |
| By: | Pietro Bini; Lin William Cong; Xing Huang; Lawrence J. Jin |
| Abstract: | Do generative AI models, particularly large language models (LLMs), exhibit systematic behavioral biases in economic and financial decisions? If so, how can these biases be mitigated? Drawing on the cognitive psychology and experimental economics literatures, we conduct the most comprehensive set of experiments to date$-$originally designed to document human biases$-$on prominent LLM families across model versions and scales. We document systematic patterns in LLM behavior. In preference-based tasks, responses become more human-like as models become more advanced or larger, while in belief-based tasks, advanced large-scale models frequently generate rational responses. Prompting LLMs to make rational decisions reduces biases. |
| Date: | 2026–02 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2602.09362 |
| By: | Thomas R. Cook; Sophia Kazinnik; Zach Modig; Nathan M. Palmer |
| Abstract: | Large language models (LLMs) are now used for economic reasoning, but their implicit "preferences" are poorly understood. We study these preferences by analyzing revealed choices in canonical allocation games and a sequential job-search environment. In dictator-style allocation games, most models favor equal splits, consistent with inequality aversion. Structural estimation of Fehr-Schmidt parameters suggests this aversion exceeds levels typically observed in human experiments. However, LLM preferences prove malleable. Interventions such as prompt framing (e.g., masking social context) and control vectors reliably shift models toward more payoff-maximizing behavior, while persona-based prompting has more limited impact. We then extend our analysis to a sequential decision-making environment based on the McCall job search model. Here, we recover implied discount factors from accept/reject behavior, but find that responses are less consistently rationalizable and preferences more fragile. Our findings highlight two core insights: (i) LLMs exhibit structured, latent preferences that often align with human behavioral norms, and (ii) these preferences can be steered, albeit more effectively in simple settings than in complex, dynamic ones. |
| Keywords: | Behavioral economics; Game theory; Search and matching models |
| JEL: | C63 C68 C61 D14 D83 D91 E20 E21 |
| Date: | 2026–01–30 |
| URL: | https://d.repec.org/n?u=RePEc:fip:fedgfe:102439 |
| By: | Dominik Suri (University of Bonn); Simon Gächter (University of Nottingham); Sebastian Kube (University of Bonn) |
| Abstract: | AI-driven systems are rapidly moving from decision support to directing human behavior through rules, recommendations, and compliance requests. This shift expands everyday human–AI interaction and raises the possibility that AI may function as an authority figure. However, the behavioral consequences of AI as an authority figure remain poorly understood. We investigate whether individuals differ in their willingness to comply with arbitrary rules depending on whether these rules are attributed to an AI agent (ChatGPT) or to a fellow human. In a between-subject design, 977 US Prolific users completed the coins task: they could earn a monetary payoff by stopping the disappearance of coins at any time, but a rule instructed them to wait for a signal before doing so. There are no conventional reasons to follow this rule: complying is costly and nobody is harmed by non-compliance. Despite this, we find high rule-following rates: 64.3% followed the rule set by ChatGPT and 63.9% complied with the human-set rule.Descriptive and normative beliefs about rule following, aswell as compliance conditional on these beliefs, are also largely unaffected by the rule’s origin. However, subjective social closeness to the rule setter significantly predicts how participants condition their behavior on social expectations: when participants perceive the rule setter as subjectively closer, conditional compliance is higher and associated beliefs are stronger, irrespective of whether the rule setter is human or AI. |
| Keywords: | Artificial intelligence, AI-human interaction, ChatGPT, rule-following, coins task, CRISP framework, social expectations, conditional rule conformity, social closeness, IOS11, online experiments. |
| JEL: | C91 D91 Z13 |
| Date: | 2026–02 |
| URL: | https://d.repec.org/n?u=RePEc:ajk:ajkdps:391 |
| By: | Grace E. Steward; Mario Macis; Nicola Lacetera; Jeffrey P. Kahn; Vikram S. Chib |
| Abstract: | We conducted a within-subject laboratory experiment in which participants decided whether to experience physical discomfort for charity, with or without additional personal compensation. Acceptance decreased with greater discomfort and increased with both larger charitable donations and personal payments. We show that private monetary incentives and prosocial benefits interact in a less-than-additive way: personal compensation raises participation but attenuates the marginal impact of charitable donations, making the combined impact of private and social rewards smaller than the sum of their separate effects. We also find suggestive evidence that the sequencing of compensated and uncompensated choices may change the responsiveness to charitable benefits. Overall, our results indicate that context, especially the presence (and timing) of private rewards, can affect the relationship between incentives and prosocial behavior. |
| Keywords: | prosocial behavior, incentives, altruism, motivation, decision-making |
| JEL: | C91 D64 D91 M52 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_12433 |
| By: | Daniele Caliari; Valentino Dardanoni; Carla Guerriero; Paola Manzini; Marco Mariotti |
| Abstract: | We study experimentally how children’s ability to avoid choice errors develops over time, focusing on both riskless and risky decisions among primary school children. We identify four types of rationality violations: cycles and menu effects in the riskless domain; and dominance and framing effects compatible with correlation neglect in the risky domain. We find that types of violations are correlated within domains but broadly independent across domains. To interpret our results we build and estimate a structural model of limited consideration. We identify an index of error avoidance and study how it develops with age and socioeconomic background, providing a new tool to understand the development of choice errors. |
| Date: | 2026–01–30 |
| URL: | https://d.repec.org/n?u=RePEc:bri:uobdis:26/823 |
| By: | Noémi Berlin (EconomiX - EconomiX - UPN - Université Paris Nanterre - CNRS - Centre National de la Recherche Scientifique); Mamadou Gueye (LEDa - Laboratoire d'Economie de Dauphine - IRD - Institut de Recherche pour le Développement - Université Paris Dauphine-PSL - PSL - Université Paris Sciences et Lettres - CNRS - Centre National de la Recherche Scientifique); Stéphanie Monjon (LEDA-CGEMP - Centre de Géopolitique de l’Energie et des Matières Premières - LEDa - Laboratoire d'Economie de Dauphine - IRD - Institut de Recherche pour le Développement - Université Paris Dauphine-PSL - PSL - Université Paris Sciences et Lettres - CNRS - Centre National de la Recherche Scientifique) |
| Abstract: | In this paper, we use a laboratory experiment to analyze the effect of information provision (feedback) on individual sorting behavior. Effective sorting requires both quantity and quality, yet increasing quantity may reduce quality due to the higher risk of contamination. We conduct a collective sorting behavior experiment consisting of a two-stage coordination game in which two subjects are paired and then individually decide whether or not to participate in a collective sorting task. The performance achieved depends on the quantity and quality of sorting, and the payoff depends on the decision and performance of both subjects in the task. Information about the subject's own past performance, and information about the partner's past performance, are included as feedback treatments. Using a between-subjects experimental design, we find that the feedback type has very different effects on participation, performance and coordination (defined as both subjects succeeding in the sorting task). Only feedback about one's own performance leads to better performance and more coordination. Although this experiment is not contextualized, the results provide useful pointers for waste sorting policies. |
| Keywords: | experiment, sorting task, cooperation, informational feedback |
| Date: | 2024–12–30 |
| URL: | https://d.repec.org/n?u=RePEc:hal:journl:hal-05460234 |
| By: | Thomas Dohmen; Frauke Meyer; Gari Walkowitz |
| Abstract: | This paper provides clear evidence that concerns for basic needs satisfaction (BNS) represent a distinct distributional motive. Using a unified theoretical and experimental framework across five dictator-game experiments in Germany and Georgia (N=446), we disentangle BNS from motives such as maximin, selfishness, efficiency, generosity, and envy. A substantial share of participants displayed BNS-driven choices and were willing to forgo income and efficiency to satisfy others’ basic needs. BNS remained robust across contexts, incentive schemes, and countries, and increased when needs satisfaction had strategic relevance. The results highlight the importance of BNS for understanding distributional preferences and policy design. |
| Keywords: | Basic Needs, Redistribution, Distributional Motives, Maximin, Public Policy, Field Experiment, Laboratory Experiment |
| JEL: | D31 D63 H23 C93 C91 D01 D91 |
| Date: | 2026–02 |
| URL: | https://d.repec.org/n?u=RePEc:bon:boncrc:crctr224_2025_729 |
| By: | Theresa Hübsch (University of Bonn); Robert Mahlstedt (University of Copenhagen); Pia Pinger (University of Cologne, ECONtribute & MPI for Behavioral Economics Bonn); Sonja Settele (University of Cologne, ECONtribute & MPI for Behavioral Economics Bonn); Helene Willadsen (National Research Centre for the Working Environment) |
| Abstract: | Using surveys with Danish students transitioning to secondary education, we study mental models of how gender and parental education shape academic performance. Students hold heterogeneous beliefs about performance gaps by gender and parental background, which appear to be shaped by within-family transmission and broader social environments. Open-text responses reveal that respondents link strong performance by girls and less socioeconomically privileged students to effort, while attributing privileged students' success to external advantages. Mental models matter: beliefs about performance gaps predict enrollment in upper secondary education by gender and parental education and causally affect students’ self-assessments, intended effort, and educational aspirations, as shown in an information experiment with girls. We highlight two mechanisms: updating about the returns to effort and about gender-specific effort costs in response to observed gender performance gaps. Our findings advance the understanding of education choices and shed light on the determinants and effects of mental models in a high-stakes setting. |
| Keywords: | Beliefs, Education, Inequality |
| JEL: | D83 D84 I24 |
| Date: | 2026–02 |
| URL: | https://d.repec.org/n?u=RePEc:ajk:ajkdps:392 |
| By: | Giuseppe Attanasi (Sapienza University of Rome); Annamaria Nese (University of Salerno); Patrizia Sbriglia (Luigi Vanvitelli University of Campania); Luigi Senatore (University of Salerno) |
| Abstract: | In this paper, we report the results of a field experiment conducted in Southern Italy in 2023, analysing the behavioural effects of earthquakes as far as trust, trustworthiness, and risk aversion are concerned. The experiments were conducted in an area where a disastrous earthquake occurred in 1980 within the Campania region. Our working hypotheses aim at testing whether there are long-term effects of an earthquake. The experimental design comprised two treatments. For the first treatment, we recruited subjects living in towns close to the earthquake epicentre that had experienced significant damage from the disaster. For the second treatment, we recruited subjects living in towns with similar socio-economic characteristics but located far from the epicentre. Our results indicate that subjects who experienced the earthquake and its aftermath are more willing to trust, reciprocate kindness, and are more risk-averse than subjects in the alternative treatment. Overall, our results shed new light on the long-term effects of catastrophes and bear relevant implications for public and health policies. |
| Keywords: | Field Experiments, Environmental Disasters, Trust, Risk Aversion |
| JEL: | C90 C91 C93 D15 |
| Date: | 2026–02 |
| URL: | https://d.repec.org/n?u=RePEc:ahy:wpaper:wp67 |