|
on Experimental Economics |
| By: | Weidmann, Ben; Vecci, Joseph; Said, Farah; Bhalotra, Sonia; Adhvaryu, Achyuta; Nyshadham, Anant; Tamayo, Jorge; Deming, David |
| Abstract: | We introduce and validate a novel approach to identifying good managers. In a preregistered lab experiment, we causally identify managerial contributions by randomly assigning managers to teams and controlling for individual skill. We find that manager contributions are crucial for team success, and that people who self-select into management roles perform worse than randomly assigned managers. Managerial performance is strongly predicted by economic decision-making skill but not by demographic characteristics. Two validation studies support our experimental results. Participants who succeed in the lab receive more real-world promotions and, in a separate study of retail store managers, skill measures strongly predict store sales. A one standard deviation increase in manager quality increases annual per store sales by US$4.1 million (25% increase). Selecting managers on skills rather than demographic characteristics or the desire to lead could substantially improve organizational performance. |
| Keywords: | management; treamwork; skills; measurement; experiment |
| JEL: | M54 J24 C90 C92 |
| Date: | 2026–05–31 |
| URL: | https://d.repec.org/n?u=RePEc:ehl:lserod:137018 |
| By: | Dominik Atella-Suri (University of Bonn); Sebastian Kube (University of Bonn) |
| Abstract: | Artificial intelligence (AI) is increasingly becoming part of economic decision-making. Yet, in many strategic interactions, individuals may not know whether others rely on AI when forming their decisions. We examine whether decision-makers who are themselves not allowed to use AI behave differently when other group members may consult AI. In an incentivized experiment with a public goods game and a weakest-link game, we exogenously vary whether group members are allowed to use AI to inform their decisions. We find that AI can affect strategic interaction even when it is not directly used by the decision-maker: merely knowing that others may use AI reduces cooperation in the public goods game and effort provision in the weakest-link game. Participants also perceive group members who may use AI as socially more distant and report lower beliefs about appropriate and expected contributions and effort levels. At the same time, the shares of conditional cooperators and conditional coordinators remain largely stable across treatments. These findings suggest that AI is not only a private decision aid but can also shape the social and strategic environment in which economic decisions are made. |
| Keywords: | Cooperation, Coordination, Human-AI Interaction, Artificial Intelligence, Experiment |
| JEL: | C71 D83 D91 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:ajk:ajkdps:407 |
| By: | Catia Batista; Lara Bohnet; Jules Gazeaud; Julia Seither |
| Abstract: | International migration can promote development in both origin and destination countries. We hypothesize that migrant integration in destination countries is an important constraint on these gains. Using a randomized controlled trial, we study the effects of a low-cost, scalable digital intervention designed to reduce information frictions among Cape Verdean immigrants in Portugal. Access to the intervention improves migrants’ labor market outcomes, legal status, social integration with native-born individuals, and aspirations. These integration gains generate international spillovers, increasing political participation and leading to more egalitarian gender norms in the migrants’ origin-country. Leveraging variation in official destination country electoral data, we show that political participation transmits through increased exposure of better-integrated migrants to prevalent local norms at destination. These international turnout spillovers are weaker in localities with higher far-right support, consistent with a less migrant welcoming political climate attenuating norm diffusion. |
| Keywords: | International migration, Migrant integration, Randomized field experiment, Employment, Immigrant regularization, Remittances, Voting, Gender norms |
| JEL: | F22 J61 O15 |
| Date: | 2025 |
| URL: | https://d.repec.org/n?u=RePEc:unl:novafr:wp2503 |
| By: | Cockx, Bart (Ghent University); Egebark, Johan (Swedish Public Employment Service); Van Hoye, Greet (Ghent University); Videnord, Emilie (Swedish Public Employment Service); Vikström, Johan (IFAU and Uppsala University) |
| Abstract: | educed motivation among jobseekers over the unemployment spell may lead to declining job-finding rates. We report findings from a low-cost digital intervention with motivational emails aimed at enhancing and sustaining motivation and search effort among job seekers in Sweden. Using a randomized controlled trial that included 200, 720 job seekers, we evaluate both carrot messages aimed at encouraging the pursuit of personal goals and intrinsic motivation and stick messages focusing on external pressure and constraints. A large share of job seekers opened the emails, and they triggered behavioral responses. Both types of messages backfired, reducing search effort and job-finding rates. The carrot messages reduced both the number of job applications and job finding, particularly among men. One likely explanation is that these messages signal to job seekers that the Public Employment Service was less controlling than initially perceived, prompting a reduction in effort. The stick messages backfired for job seekers who, at the onset of unemployment, reported that they were motivated by an inner drive rather than by constraints. These findings underscore the challenges of motivating job seekers to actively search for jobs and suggest that low-cost digital interventions, in isolation, are inadequate and may even be counterproductive. |
| Keywords: | Job search motivation; Job-finding rates; Digital interventions Behavioral interventions Randomized controlled trial |
| JEL: | A12 D01 D91 J64 J68 |
| Date: | 2026–04–10 |
| URL: | https://d.repec.org/n?u=RePEc:hhs:ifauwp:2026_008 |
| By: | Sudeep Ghosh (The Hong Kong Polytechnic University); Piet Sercu (KU Leuven); Tom Vinaimont (Nazarbayev University, Graduate School of Business) |
| Abstract: | In our experimental order-driven stock market, company news can be either high-quality and fact-based ('expert' news) or low-validity and survey-based ('social' news).Further, such messages can be provided in either a compact/matter-of-fact versus a florid/vivid form. We expect the latter to elicit a stronger interest and to boost volumes and prices. In the experiment, we find that the impact of vividness on order-submission activity is statistically and economically important. Across the experimental sessions, we also observe that the behavioral reluctance to sell exerts a powerful influence, with sell-side orders being fewer than buy-side ones, and less affected by vividness. Unexpectedly, the impact of expert news is not more marked than social news, and a confirmation bias in survey-based news does not provide the explanation. However, our result provide little cause for concern regarding market efficiency. The overlap between the buy and sell sides grows roughly proportionally with the total book size, implying there is no tangible net effect on executed order volume as a fraction of total book volume. Similarly, there is no statistically distinct effect on pricing. The main effect is on order volumes and traded volumes - the brokers' main objective as newsmongers, plausibly. |
| Keywords: | Vividness, Financial Information, Trading Behavior, Laboratory Experiment |
| JEL: | C91 C92 G41 |
| Date: | 2025–11 |
| URL: | https://d.repec.org/n?u=RePEc:asx:nugsbw:2025-09 |
| By: | Kassaballi, Zouhier |
| Abstract: | This paper evaluates the effectiveness of a low-cost intervention combining Mobile-Assisted Language Learning (MALL) with behavioral economics insights, specifically motivational framing and near-peer role models. Using a randomized controlled trial among 317 migrants and refugees in German integration courses across six cities, I examine whether such an intervention can shift learning behavior and improve German language proficiency. The intervention raised adoption of the official state-recommended app "Ankommen" from 5% to 46% and doubled daily mobile learning time from 15 to 30 minutes. Intention-to-treat estimates show treatment increased self-reported German proficiency by 0.3 standard deviations, while Complier Average Causal Effect estimates indicate actual MALL adoption improved language skills by 0.76 to 0.83 standard deviations. These findings demonstrate that a brief, replicable encouragement intervention can meaningfully complement formal language instruction at negligible cost, offering a scalable policy tool for integration programs across Europe. |
| Keywords: | Mobile-Assisted Language Learning, Migration, Integration, Language Acquisition, Randomized Controlled Trial |
| JEL: | C93 I29 J15 J24 J61 |
| Date: | 2026–05–11 |
| URL: | https://d.repec.org/n?u=RePEc:pra:mprapa:129066 |
| By: | James Alm (Tulane University); Jubo Yan (Lingnan College, Sun Yat-sen University); William D. Schulze (Cornell University); Melissa Vigil (Internal Revenue Service); Carrie von Bose (Public Company Auditing Oversight Board) |
| Abstract: | What tax preparer characteristics are most important to taxpayers in their decision to use a tax preparer, and how does this choice of a tax preparer affect subsequent taxpayer compliance? We use laboratory experiments to examine these questions. We find that individuals in this environment simultaneously choose a preparer and their compliance based in part on factors predicted by the standard expected utility theory of individual behavior under uncertainty. However, we find that factors based on psychological considerations –- which we refer to as “behavioral factors” -– also play an important role in this setting: participants prefer tax preparers who are “credentialed, ” even when the cost is high or the credential has no impact on outcomes; participants fear an audit, regardless of its likelihood; participants often choose high-cost preparers even when they are fully compliant; and many participants forego substantial expected earnings rather than underreport income. |
| Keywords: | tax compliance; tax preparer; experimental economics; expected utility theory; behavioral economics |
| JEL: | H2 H26 C91 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:tul:wpaper:2606 |
| By: | Sandra Bohmann; Lars Felder; Peter Haan; Merve Kucuk; ; Laura Schmitz; Jürgen Schupp |
| Abstract: | Carbon pricing can deliver large emissions reductions, but public opposition remains a key barrier. We study how support for carbon tax-and-transfer schemes depends on policy design and information provision in a large-scale survey experiment with German respondents. Explaining the policy mechanism robustly increases support across price levels. Information on distributional consequences raises support only when revenue recycling is sufficiently generous, and can secure majority approval even at high carbon prices. Individualized cost information increases support among those who overestimated costs, with no backlash for under-estimators when redistribution is high. These effects operate through distinct fairness channels: information shapes both self- and other-regarding justice perceptions, and while self-interest predicts support, other-regarding concerns — particularly for the poor — are an independent driver of policy acceptance. Our findings suggest that political feasibility hinges not only on policy design, but on making the mechanism understood and its distributional implications visible. |
| Keywords: | Climate policy, distributional effects, public support, justice perceptions |
| JEL: | Q52 Q58 H23 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:diw:diwwpp:dp2164 |
| By: | Joshua S. Gans |
| Abstract: | Entrepreneurs learn by experimenting, but experiment choices are often public. A closed beta, private pilot, or public launch not only generates evidence; it also reveals what kind of entrepreneur would choose that action. We develop a dynamic model in which a founder chooses between stealthy and public experiments while potential entrants infer from both actions and outcomes. Public outcomes are modelled as garblings of the founder's private experimental evidence, so public leakage informs outsiders without giving the founder information beyond the private signal already observed. The key state variable is competitive exposure: the public runway before entry becomes attractive. Exposure is depleted by two forces, leakage burn from public outcomes, and action burn from public inference about experiment choice. This distinction implies that competition can distort experiment design without forcing earlier scale: lower-confidence founders choose stealthier tests, while higher-confidence founders spend exposure to obtain faster private learning through more public tests. Scale is accelerated only when exposure reduces the value of waiting more than it reduces the value of scaling. Finally, exogenous funding gates make observable scale more selective and, therefore, more informative to entrants. The analysis shows that entrepreneurial experimentation is not merely private learning under uncertainty; it is public action under competitive inference. |
| JEL: | O30 O33 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35172 |
| By: | Jasper Z. Siol (University of Cologne); Angela R. Dorrough (University of Cologne & FernUniversität in Hagen); Louis Strang (University of Cologne); Jennifer Brunne (FernUniversität in Hagen); Andreas Glöckner (University of Cologne); Bernd Irlenbusch (University of Cologne & London School of Economics and Political Science); Shaul Shalvi (University of Amsterdam); Joscha Beckmann (FernUniversität in Hagen); Nils Köbis (University Duisburg-Essen & Max Planck Institute for Human Development Berlin) |
| Abstract: | Punishment is commonly believed to deter bribery. Yet, in transnational contexts with fragmented enforcement responsibility across countries, punishment effectiveness depends on public perceptions about the enforcing countries. We bridge behavioral experimentation and computational social science by combining an incentivized behavioral experiment across 20 countries (N = 4, 081; 81, 620 decisions) alongside a large-scale media sentiment analysis spanning 16 years. Participants’ expectations about corruption and punishment in different countries predicted bribery behavior. These expectations aligned with media narratives portraying countries as more or less corrupt, revealing a close relationship between media discourse and bribery decisions. These findings suggest that anti-corruption efforts must address not only legal frameworks but also information environments influencing public perceptions, highlighting the complex interplay between enforcement credibility and media discourse in transnational bribery. |
| Keywords: | Transnational Bribery, Corruption, Experiment, Punishment, Media Narratives |
| JEL: | C71 C90 F0 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:ajk:ajkdps:408 |
| By: | Teufel, Julia (RS: GSBE other - not theme-related research, Microeconomics & Public Economics); Beeder, Monica |
| Abstract: | Temporary migrants power key industries in many countries, yet often face poor working conditions. Vulnerability to such conditions is increased by limited knowledge of workplace rights, language barriers, and, in some cases, lack of prior work experience. One common policy response is the use of information interventions aimed at increasing rights awareness and empowering migrants to seek better working conditions where choice is possible. Despite their widespread use, evidence on their effectiveness remains limited. We present results from a randomized controlled trial evaluating a workplace rights information intervention in a longitudinal study of temporary migrants in Australia. Using multilingual online surveys, we recruited a diverse sample of workers across sectors that were understudied to date. Consistent with previous research, workers with limited rights awareness, and little prior work experience experienced worse employment conditions. While the intervention increased participants’ perceived sense of control over workplace decisions, workplace rights knowledge increased in both treatment and control groups over time. Increased rights knowledge was associated with improved working conditions, particularly among workers with limited English proficiency and less prior work experience. However, structural constraints, including dependence on employers for visa extension requirements, may limit the extent to which information alone can improve outcomes. These findings suggest that information interventions alone may be insufficient without complementary measures addressing institutional vulnerability. |
| Keywords: | temporary migration, working conditions, exploitation, working rights knowledge |
| JEL: | C93 D83 J28 J47 J83 |
| Date: | 2026–05–11 |
| URL: | https://d.repec.org/n?u=RePEc:unm:umagsb:2026003 |
| By: | Ron Berman; Walter W. Zhang; Hangcheng Zhao |
| Abstract: | Decision-makers often deploy the best-performing treatment from a randomized experiment, creating a winner's curse: selection favors treatments whose observed outcomes are high partly because of statistical noise, so the na\"ive estimate of the winner is upward biased. We distinguish two forms of winner's curse, bias relative to the true best treatment (global) and bias relative to the selected treatment's true mean (selective), and link them to regret from deploying a suboptimal treatment. This framework defines seven decision-relevant evaluation targets: mean bias, mean squared error, and confidence interval coverage for the global and selective winner's curse, and mean regret. We then show that methods that perform well on one target can perform poorly on others, so corrections should be matched to the manager's objective. Across simulations with varying effect sizes, multiple-arm settings, and data calibrated to an online A/B testing platform, no method dominates uniformly: the plug-in estimator performs best when treatment differences are large, cross-fitting performs best when treatments are similar, and resampling methods often achieve low mean squared error for moderate differences. We also introduce an adaptive empirical likelihood procedure that delivers asymptotically valid confidence intervals across settings without the tuning sensitivity of resampling-based methods. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.18887 |
| By: | Hanna Brosch; Elisabeth Grewenig; Philipp Lergetporer; Katharina Werner; Helen Zeidler |
| Abstract: | Gender norms about parental labor supply are central to explaining persistent gender inequalities in the labor market, yet their causal determinants remain poorly understood. We examine whether people’s gender attitudes are driven by mothers’ and fathers’ earnings, which may shape views about the efficient allocation of paid work and care. In a large-scale representative vignette experiment in Germany (N > 10, 000), we randomly vary pre-childbirth earnings and measure whether respondents recommend that the mother (father) stay home with the child while the father (mother) works full-time. Without specifying earnings, 90% recommend that the mother stay home. This share remains high when we specify that the mother earns less (93%). When she earns more, the share drops sharply to 47%, yet nearly half of respondents still recommend that the mother stay home. This asymmetric response rejects a purely income-based explanation of gender norms. Thus, economic circumstances shape gender attitudes, but deeply rooted norms persist even when they conflict with financial incentives. |
| Keywords: | gender norms, labor supply, gender, survey experiment |
| JEL: | C90 D13 J16 J22 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_12674 |
| By: | Javier Olivera (Departamento de Economía de la Pontificia Universidad Católica del Perú); Paola Villa-Paro (Universidad de Michigan) |
| Abstract: | We conducted an online survey experiment in Lima to study how perceptions of social class shape support for economic redistribution. Participants were randomly informed about either their actual socio-economic status (SES) or the true share of affluent households in their district. Respondents substantially overestimated their own SES and, to a lesser extent, the prevalence of affluent households. Correctingthese misperceptions generally increased support for redistribution, with no effect on a wealth-tax proposal. Effects were especially strong when respondents had misjudged their SES by two or more levels: even those predisposed against redistribution (e.g., right-leaning, individualistic, or sceptical of government) increased their support. Similar patterns also emerged when correcting beliefs about the local distribution of SES. Palabras claves: Preferencias for redistribución, percepciones de desigualdad, creencias, impuesto a la riqueza, Peru. JEL Classification-JE: H24, D31, D63, E62, H53 |
| Keywords: | Preferences for redistribution, inequality perceptions, beliefs, wealth taxes, Peru |
| Date: | 2025 |
| URL: | https://d.repec.org/n?u=RePEc:pcp:pucwps:wp00551 |
| By: | Hanna Brosch (Technical University of Munich (TUM)); Elisabeth Grewenig (Kreditanstalt für Wiederaufbau (KfW), Frankfurt); Philipp Lergetporer (ifo Institute, CESifo, Technical University of Munich (TUM)); Katharina Werner (ifo Institute, CESifo, Pforzheim University, IZA Bonn); Helen Zeidler (Technical University of Munich (TUM)) |
| Abstract: | Gender norms about parental labor supply are central to explaining persistent gender inequalities in the labor market, yet their causal determinants remain poorly understood. We examine whether people’s gender attitudes are driven by mothers’ and fathers’ earnings, which may shape views about the efficient allocation of paid work and care. In a large-scale representative vignette experiment in Germany (N > 10, 000), we randomly vary pre-childbirth earnings and measure whether respondents recommend that the mother (father) stay home with the child while the father (mother) works full-time. Without specifying earnings, 90% recommend that the mother stay home. This share remains high when we specify that the mother earns less (93%). When she earns more, the share drops sharply to 47%, yet nearly half of respondents still recommend that the mother stay home. This asymmetric response rejects a purely income-based explanation of gender norms. Thus, economic circumstances shape gender attitudes, but deeply rooted norms persist even when they conflict with financial incentives. |
| Keywords: | gender norms, labor supply, gender, survey experiment |
| JEL: | C90 D13 J16 J22 |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:aiw:wpaper:48 |
| By: | Rachel Glennerster (Center for Global Development); Joanna Murray (Development Media International); Victor Pouliquen (Paris Dauphine University) |
| Abstract: | This paper tests whether the arrival of mass media triggers a decline in fertility, a central prediction of modernization theory. Using a field experiment, we vary exposure to mass media and its content in a quarter of Burkina Faso. We provide radios to 1, 600 women without previous access to mass media. Half live in status quo areas and half in areas where the local radio station was randomly selected to air a science-based family planning campaign. Contrary to modernization theory and previous literature, gaining access to status quo mass media decreases contraception use by 14 percent and reinforces traditional gender norms. In contrast, receiving a radio in campaign areas boosts contraception use by 16 percent. The campaign also led to a 9 percent reduction in births and a 0.3 standard deviation increase in reported welfare. Reduced belief in misinformation rather than shifts in attitudes and preferences drives the result. |
| Keywords: | Mass media campaign, radio, modern contraception, family planning |
| JEL: | L82 J13 J16 |
| Date: | 2026–04–07 |
| URL: | https://d.repec.org/n?u=RePEc:cgd:wpaper:743 |
| By: | Assenza, Tiziana; Huber, Stefanie J.; Mogilevskaja, Anna; Schmidt, Tobias |
| Abstract: | We use a randomized experiment in the Bundesbank Online Panel-Households (n ≈ 3, 900) to show that the estimated link between inflation expectations and household consumption flips sign depending on survey wording. This finding reconciles prior contradictory results and has direct implications for central bank survey design. Our experiment systematically varies elicitation framing of consumption question along three dimensions: the reference unit (individual vs. household), the time horizon (past one, 3, or 12 months), and the question type (attitudinal, planned, qualitative and quantitative recall-based). We find that the time horizon and question type significantly influence the estimated relationship between inflation expectations and durable consumption. While the average effect is weak, its sign and magnitude vary strongly with question design. Planned spending and attitudinal questions, such as whether it is a good time to buy, produce very similar negative associations, suggesting that respondents interpret the former as a proxy for future intentions. In contrast, quantitative recall-based questions on past spending yield a modestly positive link, especially for shorter horizons. These results highlight the critical role of survey design in shaping behavioral measurements, offering a novel explanation for mixed findings in the literature and guidance for both research and policy. |
| Keywords: | expectations, inflation, consumption, household decision making, survey methodology, framing effects, measurement |
| JEL: | C83 D12 D84 E31 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:zbw:bubdps:341099 |
| By: | Arendt, Jacob (Rockwool Foundation Research Unit); Bolvig, Iben (The Danish Center for Social Science Research) |
| Abstract: | This study estimates the effects of an employment programme for disadvantaged unemployed individuals. The programme emphasized on-the-job training and contracting the unemployed for a few paid work hours as a stepping stone into the labour market. Evaluated through a randomised controlled trial, the programme was found to accelerate transitions into part-time work. Contrary to its intention, it permanently increased the share of participants receiving disability pensions among the most disadvantaged groups. To explain this finding, we suggest that training, while enhancing productivity for some, simultaneously provided information of employability used in the assessment of disability pension eligibility. |
| Keywords: | unemployed, active labour market policy, disability pension, immigration |
| JEL: | J14 J15 J64 D61 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp18618 |
| By: | Elliott Ash; Francesco Capozza; Sergio Galletta |
| Abstract: | We estimate demand for five conversational AI use cases — administrative assistance, studying, wellness, friendship, and romance — using a preregistered, within-respondent conjoint experiment in a U.S. online sample (N=1, 989). Purpose dominates interest: instrumental uses rank highest, romance lowest. Free pricing raises interest sharply but is attenuated for relational purposes, revealing non-monetary signaling costs that price reductions cannot eliminate. Privacy and personalization shift demand; interaction modality does not. Second-order beliefs exceed own interest in proportion to the purpose penalty, consistent with a single stigma parameter suppressing stated below latent demand. Corrected consumer surplus for a free romantic companion turns positive once suppression costs are recovered — the welfare mirror image of social media’s collective trap. |
| Keywords: | AI, conjoint experiment, WTP, second-order beliefs |
| JEL: | C81 C93 D82 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_12673 |
| By: | Marina Agranov; Gabriel Lopez-Moctezuma; Philipp Strack; Omer Tamuz |
| Abstract: | We compare how well agents aggregate information in two repeated social learning environments. In the first setting agents have access to a public data set. In the second they have access to the same data, and also to the past actions of others. Despite the fact that actions contain no additional payoff-relevant information, and despite potential herd behavior, free riding and information overload issues, observing and imitating the actions of others leads agents to take the optimal action more often in the second setting. We also investigate the effect of group size, as well as a setting in which agents observe private data and others' actions. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.17662 |
| By: | Kausar, Shafiya (INSEAD) |
| Abstract: | We evaluate LLM-augmented reinforcement learning for stock trading on Nasdaq- 100 (2019–2023) and report a previously unmeasured experimental phenomenon: the relationship between LLM signal coverage density and trading performance is non-monotonic, with a clearly identifiable harmful regime. In a controlled coverage sweep over {0%, 5%, 20%, 50%, 80%, 100%}, signal injection at 5% and 20% coverage degrades performance below the no-signal baseline, becoming net-positive only at ≥ 50% coverage. The FNSPID dataset’s 9.7% non-neutral coverage sits inside this harmful regime—meaning that for typical research configurations available today, adding LLM signals to the RL pipeline reduces returns. Beyond this density finding, we report three further negative results that the LLMRL trading literature has not adequately addressed. First, our LLM-augmented RL agent (158.11% cumulative return as a 3-seed ensemble) is outperformed by three standard non-RL baselines that prior work in this thread does not report: momentum top-10 (250.45%), equal-weight buy-and-hold (235.00%), and equal-weight monthly rebalanced (214.06%), all of which also exceed the Nasdaq- 100 buy-and-hold benchmark (164.52%). Second, we control for the daily-vs.- monthly rebalancing-frequency confound by deploying the same trained agents under matched-frequency monthly execution; the monthly variant underperforms its daily counterpart by 47pp (111.01% vs. 158.11%), confirming that the baseline gap is not driven by transaction-cost differences. Third, a v3-matched ablation finds that removing the CVaR tail-risk constraint produces a difference within the seedto- seed variability of the experiment. Across two independent runs, the sign of this difference flipped, providing direct empirical evidence that the algorithmic risk-tail machinery contributes no detectable return benefit in this setting. A regime decomposition reveals one clear win for the agent: in the 2023 recovery period, the 3-seed ensemble (52.6%) outperforms all non-RL baselines, suggesting the learned policy may have regime-specific advantages that single-window evaluation obscures. We argue that LLM-RL trading research should adopt non-RL baselines as standard practice, report signal coverage density as a first-class experimental variable, and decompose results by regime. Code and trained models are available at https: //anonymous.4open.science/r/signal-densi ty-llm-trading-9966/. |
| Date: | 2026–05–14 |
| URL: | https://d.repec.org/n?u=RePEc:osf:socarx:nxvdp_v1 |
| By: | Jeff Tjiong; Thijs Dekker (University of Leeds, Institute for Transport Studies); Stephane Hess (University of Leeds, Institute for Transport Studies); Marek Giergiczny (University of Warsaw, Faculty of Economic Sciences); Manuel Ojeda-Cabral (University of Leeds, Institute for Transport Studies); Mikołaj Czajkowski (University of Warsaw, Faculty of Economic Sciences) |
| Abstract: | Stated choice surveys commonly used in public policy appraisal – such as in transport or environmental economics – often contrast a ‘free’ status quo alternative against a range of (policy) interventions which can be implemented at a cost. Limited attention has, however, been paid to the fact that the ‘free’ nature of the status quo (SQ) alternative may make the SQ alternative overly attractive due to the zero-price (ZP) effect. The ZP effect is a well-established notion in behavioural economics explaining the phenomenon that individuals tend to over-react to free alternatives. We present an experimental design setup allowing the separation of the ZP effect from the SQ effect together with the identification of non-linear sensitivities to costs. Choices made by students between different mobile broadband packages are used for illustrational purposes. Our analysis shows that the ZP effect is significant and the observed preference to remain in the SQ is largely due to the ZP effect. In practice, this may lead to biased welfare estimates for public policy packages if the ZP effect is not explicitly accounted for. Moreover, we also show that misspecification of the functional form for cost can lead to significant bias in WTP estimates and the ZP and SQ effects. |
| Keywords: | zero-price effect, stated choice experiments, status quo bias, willingness-to-pay, welfare analysis, discrete choice modelling, non-linear cost sensitivity, Box-Cox transformation, mobile broadband |
| JEL: | C25 D91 D61 Q51 L96 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:war:wpaper:2026-15 |
| By: | Croke, Kevin; Daniels, Benjamin; Lipinski, Robert; Rogger, Daniel |
| Abstract: | This paper examines the impacts of family doctors writing an explicit “care contract” with at-risk patients for increased holistic primary care. The intervention was designed to shift the care relationship between the two parties away from episodic curative care and towards a holistic plan for patient welfare. The experiment tracked healthcare utilization, diagnosis, prescription, hospitalization, and mortality outcomes through the universe of patient records. The program caused increased screening, diagnosis and treatment of chronic health issues among enrolled patients by about 10% across these categories, with suggestive evidence that hospitalization declined by 8%. Among ‘mild-risk’ patients, the treatment group experienced annual reductions in all-cause mortality of 20%, in an analysis pre-specified as exploratory. |
| Date: | 2026–05–19 |
| URL: | https://d.repec.org/n?u=RePEc:wbk:wbrwps:11388 |
| By: | Phoebe Koundouri; Theodoros Daglis; Conrad Landis; Akrivi Katifori; George Gkanias |
| Abstract: | The present work examines urban residents' preferences for biodiversity conservation, heatwave-related living conditions, and traffic-related noise and pollution, using a discrete choice experiment (DCE), also considering willingness-to-pay (WTP) and the influence of immersive virtual reality (VR) exposure, focusing on the metropolitan area of Athens. More importantly, VR exposure examined whether individuals' preferences were affected, focusing on its methodological contribution to the better planning of urban sustainability interventions. According to the results, biodiversity emerges as the most valued attribute, followed by heatwave measures, and then, traffic reduction, and VR exposure affects individuals' preferences for traffic measures, emphasizing its methodological value for improving urban sustainability decisions. |
| Keywords: | Discrete choice experiment, Urban sustainability, Biodiversity protection, Heatwave mitigation, Traffic pollution, Virtual reality, Environmental preferences |
| Date: | 2026–05–14 |
| URL: | https://d.repec.org/n?u=RePEc:aue:wpaper:2614 |
| By: | van den Berg, Anthony; Doroc, Karlo; Fu, Changfa; Grossmann, Max R.P.; Miller, Joshua B.; Pavlovic, Lana |
| Abstract: | Xiang et al. (2025) investigate whether the gambler's fallacy-the false belief that a random event is less likely to occur if it has occurred recently-is robust to using probabilistic (versus point) predictions and independently and identically distributed (versus non-IID) sequences. In five experiments with 150 participants each observing 18 sequences of colored balls, the authors find strong evidence of the gambler's fallacy when eliciting point predictions or using non-IID sequences, but fail to observe a robust gambler's fallacy when eliciting probabilistic predictions over IID sequences. We successfully reproduce all main results from the processed data and analysis code. The absence of raw data precludes robustness checks incorporating response times, attention checks, or demographic controls. We conduct two main robustness analyses that both validate the main finding that point predictions are fundamentally different from probabilistic predictions, but also findstrong evidence of the gambler's fallacy for probabilistic predictions. First, we investigate the simplest form of the gambler's fallacy, finding that participants do expect negative autocorrelation based solely on the outcome of the final draw, which applies to both probabilistic and point predictions. Second, we find that this expectation of negative autocorrelation strengthens as the streak gets longer, and that this also applies to both probabilistic and point predictions. |
| Keywords: | Gambler's fallacy, replication, probabilistic prediction, decision making, cognitive bias |
| JEL: | D91 C90 D83 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:zbw:i4rdps:295 |
| By: | Howai, Niko; Bian, Alice; De Guzman-Mortillero, Arnica; Robinson, Elizabeth |
| Abstract: | Mangroves, especially in coastal areas, provide collective benefits to households, not just individuals. In this study, we undertake a comparison of individuals’ and couples’ intra-household decision-making on preferences for mangrove preservation expenditure and benefits using a discrete choice experiment (DCE) in Palawan province in the Philippines. We find that men’s and women’s individual preferences differ when responding separately to the survey, and that their joint preferences align more with the men’s preferences. We also conducted in-depth interviews with a subset of the population considered to be marginalised and exempt from contributing to mangrove preservation payments under the DCE. The findings from the exemption interviews suggest strong support for community co-management of mangrove marine protected areas (MPAs), provided that income-generating alternative livelihood projects are created. This, in turn, is combined with the couples’ preferences in the DCE. The resulting preferences for mangrove benefits and their valuation can be used to inform the design and financing of MPAs that include co-managed mangrove protection and restoration projects with locals, as well as policies for the use of mangrove resources on the island. |
| Keywords: | discrete choice experiment; intra-household preferences; in-depth interviews; hierarchical Bayesian logit; mangroves |
| JEL: | C11 C52 D12 Q57 |
| Date: | 2026–05–20 |
| URL: | https://d.repec.org/n?u=RePEc:ehl:lserod:138521 |
| By: | Dmitry Dagaev; Egor Ivanov; Petr Parshakov; Alexey Savvateev; Gleb Vasiliev |
| Abstract: | The emergence of large language models (LLMs) has spurred economists to study how humans and LLMs behave in strategic settings. We organized a series of round-robin tournaments in the Colonel Blotto game. This game attracts game theorists' attention due to high-dimensional action space and the absence of pure strategy Nash equilibria. In the first tournament, more than 200 human participants competed against one another. In the second tournament, several popular LLMs were invited to submit strategies. In the third tournament, we matched the number of LLM strategies to the number submitted by humans. We find that humans more often employ better-calibrated intermediate-level allocation heuristics and outperform the simpler, more stereotyped strategies submitted by LLMs. Strategic sophistication is key to success if and only if the necessary level of reasoning depth is reached, while lower and higher levels of reasoning offer no clear advantage over the primitive strategies. Among humans, field of study weakly predicts success: participants with STEM backgrounds perform better in the first tournament. Surprisingly, humans almost do not adjust their strategies across tournaments with different sets of opponents. This result suggests that humans base their choices primarily on the game's rules rather than on the identity of their opponents, treating LLMs much like human competitors. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.22095 |
| By: | Dupuy, Arnaud (University of Luxembourg) |
| Abstract: | We study how individuals trade off outcome (what) and process (how) utility in high-stakes strategic decisions. We exploit optimality conditions and high-frequency choices in professional tennis to derive nonparametric bounds on process utility and implement a structural approach to estimate player-specific preferences. Under mild shape restrictions, these bounds imply that a large majority of players place positive weight on process utility. Our structural estimates further show that most players systematically sacrifice success probabilities to increase process utility, generating economically meaningful effects on match outcomes and expected earnings. |
| Keywords: | process utility, intrinsic motivation, outcome utility, salience weight, strategic behavior, nonparametric, structural estimation |
| JEL: | D91 D81 D01 C57 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp18625 |
| By: | Mehrzad Khosravi; Max Kleiman-Weiner; Hema Yoganarasimhan |
| Abstract: | Many consumer decisions are repeated choices under uncertainty. Standard models capture these decisions using Bayesian learning and dynamic programming: consumers update beliefs from feedback and use those beliefs to guide future choices. In many markets, however, learning does not restart when consumers enter a new context: prior experience with a brand, product, or provider can shape beliefs in later, related decisions. We study this cross-context knowledge transfer, or meta-learning, in sequential choice. We design a hierarchical laboratory task in which participants repeatedly choose among airlines across routes and observe noisy binary outcomes. Reduced-form evidence shows that participants improve not only within routes, but also across routes: they choose better airlines earlier in later routes and reduce pseudo-regret. To identify the mechanism behind this transfer, we compare human choices to a no-transfer benchmark and a fully integrated Bayesian meta-learning benchmark. In particular, we introduce a class of boundedly rational meta dynamic programming policies, BRMDP(D), that approximate full integration using a limited number of hyper-posterior draws, denoted by D. Trial-by-trial likelihood comparisons show that low-D boundedly rational meta-learning, especially BRMDP(1), fits participant behavior better than both no transfer and fully integrated Bayesian transfer. Consumers, therefore, transfer brand-level regularities across contexts, but through coarse representations of prior uncertainty. The findings imply that models of consumer learning should allow for approximate cross-context transfer, and that managerial counterfactuals based on either no-transfer or fully integrated learning can be misleading. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.16532 |
| By: | Joshua S. Gans; Luca Gius |
| Abstract: | How should an organisation choose the breadth of its experimentation portfolio? Breadth has two distinct margins—the number of paths kept alive, and the degree of contrast among them—and prior research has largely studied them in isolation. We bring them into a single framework and show that they need not move together. Under a fixed experimentation budget, adding paths creates more chances to find a strong direction, but it also dilutes learning across paths and weakens the strongest feasible contrast. When the task is primarily ranking among already-viable alternatives, broader portfolios become more attractive as the budget rises. When paths share common viability uncertainty, and experimental signals track payoff relatedness, however, additional paths partly repeat the same viability test rather than provide independent information. We identify conditions under which testing exactly two sharply contrasting paths is optimal, dominating both a single deep test and broader portfolios. The framework reconciles competing prescriptions—many parallel shots versus a few sharp comparisons—by clarifying when each applies, and shows why empirical measures of breadth should not treat the number of options and their relatedness as separable margins. |
| JEL: | D81 D83 L26 O31 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35207 |