|
on Experimental Economics |
| By: | Dwenger, Nadja; Sjursen, Ingrid Hoem; Vietz, Jasmin |
| Abstract: | Many societies aim to design policies based on meritocratic fairness, which involves two principles: (i) paying individuals with equal performance equally (fair equality) and (ii) paying individuals with higher performance more (fair inequality). Yet, often it is impossible to respect both simultaneously. This paper provides novel evidence on the importance individuals attach to each principle from a large-scale experiment in the United States. We document large heterogeneity in preferences. Individuals incur substantial personal costs to implement their preferred principle. Republican supporters are more likely to prefer fair inequality. The findings offer insights into the political economy of redistribution and public policy design. |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19472 |
| By: | Cilliers, Jacobus; Elashmawy, Nour; McKenzie, David |
| Abstract: | The post-double selection Lasso estimator has become a popular way of selecting control variables when analyzing randomized experiments. This is done to try to improve precision, and reduce bias from attrition or chance imbalances. We re-estimate 780 treatment effects from published papers to examine how much difference this approach makes in practice. We find it reduces standard errors by less than one percent compared to standard Ancova on average and does not select variables to model treatment in over half the cases. We discuss and provide evidence on the key practical decisions researchers face in using this method. |
| Keywords: | Attrition |
| JEL: | O12 C93 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19505 |
| By: | Bigoni, Maria; Bortolotti, Stefania; Parisi, Francesco; Zhang, Xin |
| Abstract: | Rational parties enter into a contract if the agreement is mutually beneficial. However, after the contract is formed, changes to the costs and/or benefits of performance may render the original contract undesirable. In this paper, we carry out an incentivized experiment to study the effect of alternative remedies on the parties’ ability to renegotiate their contractual obligations. After entering into a contract, experimental subjects observe symmetrical changes to the original costs and/or benefits, which create a misalignment of their performance vs. breach incentives. Renegotiation of the original contract would allow parties to realign their interests and to capture some additional surplus. Our experimental design compares the effects of damage and specific performance remedies on the parties’ ability to renegotiate. Our results confirm Coase’ (1960) irrelevance of remedies proposition, providing novel insights for the choice of contract remedies in the face of possible market shocks. |
| Keywords: | Damages; Breach remedies; Specific performance |
| JEL: | C90 D86 D91 K12 K41 |
| Date: | 2024–08 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19365 |
| By: | Bazley, William; Cuculiza, Carina; Korniotis, George |
| Abstract: | We examine the impact of perceived social discrimination on U.S. households' preferences, focusing on key aspects of prospect theory. Utilizing both field and experimental data, we find that perceiving discrimination increases risk tolerance, decreases loss aversion, and excessively distorts objective probabilities. These effects are primarily observed in racial/ethnic minorities, with no significant impact on White individuals. Emotional mechanisms, particularly anger, play a role in transmitting the effects of discrimination on preferences. Overall, our findings underscore how social factors, such as discrimination, can systematically shape fundamental preferences, and ultimately influencing economic decision-making. |
| Keywords: | Cumulative prospect theory; Experiment |
| JEL: | D01 D90 D91 G40 G41 G50 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19514 |
| By: | Henrik Guhling; Fabian Herweg |
| Abstract: | We propose a controlled and incentive-compatible design to study racial bias in team-selection decisions. In an online experiment, participants select a soccer lineup from fictitious players whose objective quality is described by numerical attributes. For each player profile, we randomly assign an AI-generated portrait depicting either a Black or a White player, allowing us to isolate the effect of racial appearance. Overall, we find no systematic evidence that White player versions are selected more frequently than otherwise identical Black versions. We also compare generic-team and national-team framing, but find no systematic amplification of racial bias. |
| Keywords: | incentivized experiment, hiring and selection, racial discrimination, soccer |
| JEL: | C91 D91 J15 J71 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_12914 |
| By: | Yuhao Fu; Nobuyuki Hanaki |
| Abstract: | This experimental study investigates how people rely on different sources of advice when detecting AI-generated fake news (deepfake news). In a laboratory deepfake detection task, student participants identified the proportion of human-written (non-AI-generated) content in synthetic deepfake news articles and received advice from ChatGPT (GPT-4), human peers, or linguistic experts. The results show that participants rely more on ChatGPT than on human peers when detecting GPT-2-generated deepfake news. Participants also rely more on linguistic experts than on peers, while the relative reliance on experts versus ChatGPT is mixed across experimental waves, potentially reflecting time trends in beliefs about AI-based detection. Importantly, in the additional experiment conducted in 2025 under the same experimental procedure, participants relied more on linguistic experts than on ChatGPT. Moreover, performance improvements reflect the joint role of reliance and advice quality, arising primarily when participants rely on high-quality advice. Overall, relying on AI to detect AI-generated deepfakes can improve detection outcomes, but only when AI-based detection tools are of sufficiently high quality. These findings highlight the dual role of GAI as both a source of deepfakes and a tool for mitigating related risks. |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2608.01540 |
| By: | Almås, Ingvild; Hufe, Paul; Weishaar, Daniel |
| Abstract: | Many of our everyday life experiences suggest that humans are social individuals who care about themselves and others. Yet, for a long time, the economic discipline has focused on modeling the purely self-interested ``homo economicus''. In this review, we summarize the extensive experimental literature on deviations from such purely self-interested behavior, with a particular focus on attitudes toward inequality and fairness. This literature shows that many individuals are willing to forgo their own economic gain for a fairer distribution of resources. We further illustrate empirical evidence showing that people differ in what they consider to be fair, and how these fairness views vary within and across countries. |
| Keywords: | Social preferences; Experimental evidence |
| Date: | 2024–10 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19620 |
| By: | Duflo, Esther; Dupas, Pascaline; Spelke, Elizabeth; Walsh, Mark |
| Abstract: | We provide experimental evidence on the intergenerational impacts of secondary education subsidies in a low-income context, leveraging a randomized controlled trial and 15-year longitudinal follow-up. For young women, receiving a scholarship for secondary school delays childbearing and marriage, and reduces unwanted pregnancies. Female scholarship recipients are more likely to marry a partner with tertiary education and their children have better early childhood development outcomes. In particular, we document a 45% reduction in under-three mortality as well as cognitive development gains of 0.25 standard deviations of test scores once children are of school age. The primary mechanism seems to be that more-educated caregivers have the knowledge and skills to safeguard their children’s health and stimulate their cognitive development. In contrast, we find no evidence of a positive impact for the children of male scholarship recipients, who tend to marry less educated partners. Together, these results suggest a key role for maternal education in child outcomes. We also estimate the cost-benefit ratio for secondary school scholarships and find that the impact on child survival alone is sufficient to make them a highly cost-effective investment. |
| JEL: | J13 O15 I26 |
| Date: | 2024–08 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19345 |
| By: | Biais, Bruno; Mariotti, Thomas; Moinas, Sophie; Pouget, Sebastien |
| Abstract: | We study asset pricing and risk sharing in experimental financial markets. We design our experiment to test the key equilibrium implications of rational choice and competitive behavior in complete markets without making parametric assumptions on preferences. We find that participants behave competitively but deviate from rationality, as around 25% of their actions are first-order stochastically dominated. We propose a random-choice model predicting that, as the number of participants grows large, prices and average per-participant trades converge to those in the rational-choice competitive equilibrium. This prediction is supported by our experimental data. We structurally estimate a special case of the random-choice model with CRRA utilities and logit weighting functions and find that only around 80% of participants benefit from participating in the market. |
| JEL: | C92 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19420 |
| By: | Urbina Florez, Maria Jose; Moya, Andrés; Rozo, Sandra |
| Abstract: | This paper studies a randomized controlled trial in which undocumented Venezuelan migrants in Colombia were assigned to receive informational videos via WhatsApp encouraging registration in a regularization program. The intervention backfired. Receiving a video reduced take-up by 8 percentage points, a 15 percent decline relative to the control mean. The negative effect operated through two channels. Among individuals who watched, exposure to procedural content discouraged registration. Among those who did not engage, unsolicited contact itself reduced take-up. Both effects were concentrated among individuals at the margin of the decision to participate. These findings challenge the assumption that well-designed informational nudges are, at worst, neutral. |
| Date: | 2026–06–30 |
| URL: | https://d.repec.org/n?u=RePEc:wbk:wbrwps:11423 |
| By: | Yuhao Fu; Nobuyuki Hanaki; Haitao Wang |
| Abstract: | Artificial intelligence increasingly participates in economic interactions not only as a tool, but also as an autonomous bargaining counterpart negotiating on behalf of firms, platforms, and consumers. Yet little is known about how humans respond psychologically and strategically when bargaining with such agents in dynamic settings. We study this question in a laboratory experiment using a three-stage alternating-offer bargaining game in which participants negotiate in real time with either another human or a GPT-based AI agent. We also introduce a human-beneficiary condition in which the AI agent's earnings may affect another participant's payment. Agreements are not reached earlier in human-human bargaining than in human-AI bargaining, but they are reached significantly earlier when the AI's payoff affects another participant's payoff. Human proposers offer more to human opponents than to AI agents, whereas responders become significantly more willing to accept unfair AI offers when AI earnings may benefit another human. These findings suggest that fairness and reciprocity toward AI are weaker and more conditional than toward humans, but partially remerge when AI outcomes affect real people. The results have implications for the design of AI negotiation systems and broader human-AI economic interactions. |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2608.01212 |
| By: | Sebastian M. Peters; Jürgen Huber; Michael Kirchler |
| Abstract: | The climate crisis is one of the major challenges mankind currently faces and, through green investments, the finance industry can play a crucial role in tackling it. However, this topic still appears to be controversially discussed in finance, and only very few studies have investigated the behavior of finance professionals with respect to green investments. In this experiment, we investigate the behavior of 174 finance professionals and 192 participants of the general population, measuring drivers of decision making in green investments. We find in both participant pools that individuals who say it feels better to invest in green assets also do so. Between the two subject pools, we find no difference in investment propensity in green assets. Furthermore, we observe that financially literate participants invest significantly more green and achieve higher portfolio returns. Interesting, while 89 percent of participants say they would not be ready to forego returns for greener assets, only 10 percent select the portfolio with the highest possible return. |
| Keywords: | Green investments, finance professionals, experimental finance |
| JEL: | C90 G40 G41 |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:inn:wpaper:2026-07 |
| By: | Randolph Sloof (University of Amsterdam); Julian Tait (University of Amsterdam); Jeroen van de Ven (University of Amsterdam) |
| Abstract: | Organisations benefit from the timely disclosure of mistakes, yet workers often conceal their errors. We examine the role of blame avoidance, hypothesising that mistakes associated with a larger loss in potential earnings increase blame and discourage disclosure. This creates a paradox: mistakes are concealed precisely when concealment is most costly to the organisation. We formalise this argument in a model and test it in a vignette study and two incentivised laboratory experiments. The vignette study supports the proposed mechanism, showing that blame increases with the loss in potential earnings. In the laboratory experiments, larger losses do not reduce disclosure, but we do document an important honesty penalty in team settings: the desire to avoid blame induces free-riding by leaving the disclosure of mistakes up to their team members. This pattern emerges both when new teams are formed after each interaction and when teams stay together. |
| Keywords: | organisational failures, blame, experiment, mistakes, lying |
| JEL: | D91 M54 |
| Date: | 2026–08–02 |
| URL: | https://d.repec.org/n?u=RePEc:tin:wpaper:20260051 |
| By: | Lindner, Thijs (Erasmus University Rotterdam); de Koster, Willem; van der Waal, Jeroen |
| Abstract: | The relationship between ethnic diversity and welfare attitudes has been the subject of much research. However, little is known about how exposure to information on the overrepresentation of ethnic minority individuals among welfare recipients affects attitudes to welfare recipients in general. We scrutinize a potential information exposure effect, and assess two contrasting scenarios, i.e., exposure to aforementioned information might strengthen either negative stereotypes of welfare recipients or the notion that welfare dependency is the result of structural conditions. Additionally, we hypothesize that the former (latter) response is more likely among those with more (less) ethnocentric attitudes. We tested this by conducting an original pre-registered survey experiment among members of a high-quality probability-based panel designed to represent the Dutch population (n = 1, 527). The treatment group was exposed to information depicting the overrepresentation of Dutch citizens with a Moroccan migration background among welfare recipients in the Netherlands, while the control group was not. It was only among the more ethnocentric respondents that exposure to the treatment increased support for the notions that welfare recipients: 1) are lacking work ethic; and 2) are themselves to blame for their condition. There was 3) no treatment effect on support for welfare schemes in general. We discuss the implications of these findings for a variety of academic debates and make suggestions for further research. |
| Date: | 2026–07–29 |
| URL: | https://d.repec.org/n?u=RePEc:osf:socarx:6r32b_v1 |
| By: | Acemoglu, Daron; Aksoy, Cevat Giray; Baysan, Ceren; Molina, Carlos; Zeki, Gamze |
| Abstract: | This paper investigates whether enduring authoritarian regimes are in part rooted in the population’s misperceptions about their social and economic costs–as opposed to a general preference for authoritarianism. We explore this question using online and field experiments in the context of Türkiye’s May 2023 presidential and parliamentary elections. We confirm that voters, especially those supporting the incumbent authoritarian government systematically underestimate both the extent to which democracy and media freedom have been eroded in Türkiye and their usefulness in dealing with natural disasters and corruption (two salient issues in Türkiye). We find that providing (accurate) information about the state and implications of democracy and media freedom have significant effects on beliefs and increase the likelihood of voting for the opposition by about 3.7 percentage points (6.2 percent) in the online experiment. In the field experiment, the ballot-box level opposition vote share in neighborhoods receiving the information treatment increases by 0.8 percentage point (1.5 percent). Interestingly, both in the field and online, the results are driven not by further mobilizing opposition supporters, but by influencing those likely to vote for the governing coalition and those holding more misperceived beliefs about democracy and media freedom in Türkiye. The evidence suggests that at least part of the support for authoritarian regimes may be coming from misperceptions about their institutions and policies, and may be more malleable than typically presumed. |
| Keywords: | Democracy; Misperceptions; Elections; Institutions |
| JEL: | P16 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19536 |
| By: | Fan, Tony Q.; Liang, Yucheng; Peng, Cameron |
| Abstract: | Evidence from the laboratory and the field has uncovered both underreaction and overreaction to new information. We provide new experimental evidence on the underlying mechanisms of under- and overreaction by comparing how people make inferences and revise forecasts in the same information environment. Participants underreact to signals when inferring about underlying states, but overreact to the same signals when revising forecasts about future outcomes— a phenomenon we term “the inference-forecast gap.” We show that this gap is largely driven by different simplifying heuristics used in the two tasks. Additional treatments suggest that the choice of heuristics is affected by the similarity between statistics in the information environment and the statistic elicited by the belief-updating problem. |
| Keywords: | belief updating;inference;forecast revision;attribute substitution |
| JEL: | F3 G3 J1 |
| Date: | 2026–07–01 |
| URL: | https://d.repec.org/n?u=RePEc:ehl:lserod:131071 |
| By: | Gönül Doğan (University of Cologne); Louis Strang (University of Cologne) |
| Abstract: | We propose a novel behavioural measure of authoritarian preferences and deploy it in a large German sample comparing East and West. We identify preference for three types of decision rules and discrimination when assigning decision power. Support for authoritarian decision rules is high with 73% of participants assigning positive weight to them. When assigning decision power within authoritarian rules, groups with a large number of minorities as well as minority members themselves face substantial discrimination, ranging between 12 to 33 percent lower decision weight allocated to them. East and West German participants do not significantly differ in their discrimination in decision power. Our main treatment variation exposes individuals to minority success stories; neither authoritarian rule support nor discrimination in decision power is significantly impacted by such positive information. Drawing on an extensive questionnaire, we identify attitudes toward decision power allocation within the society, and beliefs about immigrants’ origin and religion asthe strongest predictors of authoritarian behaviour. |
| Keywords: | Authoritarianism, Discrimination, Resource allocation, Political prefer- ences, Social preferences, Online experiment |
| JEL: | C91 H89 P00 |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:ajk:ajkdps:425 |
| By: | John Conlon; Peter Schwardmann |
| Abstract: | We examine whether sycophantic AI advice distorts decisions. Our experiment involves 1, 500 participants in 30 decision environments spanning core domains in economics and the social sciences. Contrary to the vast majority of predictions in an expert survey we conduct, we find that AI advice depolarizes choices on average, moving participants away from their initial leanings. This depolarization arises despite the LLM being measurably sycophantic: it disproportionately offers considerations that support users' initial leanings and uses agreeable and flattering language. Depolarization occurs across moral and non-moral, objective and subjective, strategic and non-strategic, and complex and simple tasks. Increasing sycophancy weakens depolarization, showing that sycophancy is behaviorally relevant, even if it is generally outweighed by the informativeness of AI advice. Finally, several results mitigate the concern that market forces will generate greater polarizing effects outside the experiment or in the future. On the supply side, our baseline AI's level of sycophancy is typical of leading models, and these models are not becoming more sycophantic over time. On the demand side, participants do not prefer greater sycophancy, do not select into AI advice in tasks where it is more polarizing, and exhibit greater depolarizing effects when they are more frequent AI users outside the experiment. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.28133 |
| By: | Bruno Cr\'epon; Aur\'elien Frot; Christophe Gaillac |
| Abstract: | Most digital job-search assistance encourages unemployed workers to broaden their search toward related occupations, targeting one important source of search inefficiency: insufficient occupational diversification. Our analysis suggests that the relevant margin of adjustment depends on the underlying search problem. Building on a detailed analysis of job seekers' beliefs and search behavior, we identify a large group of pessimistic workers for whom the main constraints are low search effort and low aspirations, rather than insufficient occupational diversification. This diagnosis points to an unexpected intervention: rather than encouraging these workers to search in new occupations, we encourage them to search more intensively and apply for better-paying jobs within the occupations they already consider. We evaluate this diagnosis-based intervention, alongside a standard occupational recommendation, in a large-scale randomized experiment conducted with the French Public Employment Service. The motivational intervention increases search effort, raises reservation wages, and improves reemployment outcomes along the predicted margins. Occupational recommendations, by contrast, primarily benefit workers whose search problem lies in the allocation of attention across occupations and operate by activating existing perceptions rather than correcting beliefs. More broadly, our findings show how digital platforms can combine subjective expectations, behavioral data, targeted interventions, and randomized experimentation to diagnose job seekers' needs and iteratively improve intervention design. |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2608.16849 |
| By: | Van Campenhout, Bjorn; Ariong, Richard M.; Kariuki, Sara Wairimu; Chamberlin, Jordan |
| Abstract: | Quality in agri-food supply chains is often unobservable at first sale and early aggregation limits traceability, weakening incentives for quality provision. We study whether making milk quality visible and traceable creates a market for quality in Uganda’s dairy sector. Increasing observability reduces adulteration and improves quality, but no premium emerges. In a follow-up experiment, we introduce trader quality premiums. This increases quality when binding, yet informed intermediaries capture the gains and farm-gate prices do not rise. Observability is necessary but insufficient: without downstream demand for quality and pass-through by intermediaries, incentives for quality upgrading remain weak. |
| Keywords: | dairying; value chains; enforcement; quality management; quality assurance; quality control; Uganda; Africa; Sub-Saharan Africa |
| Date: | 2026–06–11 |
| URL: | https://d.repec.org/n?u=RePEc:fpr:ifprid:183322 |
| By: | Bohren, Noah; Hakimov, Rustamdjan; Lalive, Rafael |
| Abstract: | Generative artificial intelligence (AI) has made substantial progress, but some capabilities of AI are not well understood. This study compares the ability of AI to a representative population of US adults in creative and strategic tasks. The creative ideas produced by AI chatbots are rated more creative than those created by humans. Moreover, ChatGPT is substantially more creative than humans, while Bard lags behind. Augmenting humans with AI improves human creativity, albeit not as much as ideas created by ChatGPT alone. Competition from AI does not significantly reduce the creativity of men, but it decreases the creativity of women. Humans who rate the text cannot discriminate well between ideas created by AI or other humans but assign lower scores to the responses they believe to be AI-generated. As for strategic capabilities, while ChatGPT shows a clear ability to adjust its moves in a strategic game to the play of the opponent, humans are, on average, more successful in this adaptation. |
| Keywords: | ChatGPT; Creativity; Experiment |
| JEL: | I24 J24 D91 C90 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19507 |
| By: | Jason Sockin; Pawel Adrjan; Mária Balgová; Simon Jäger; Jonas Jessen |
| Abstract: | Discrete choice experiments are widely used to estimate workers' willingness to pay (WTP) for job amenities under the assumption that varying an attribute does not change workers' beliefs about other job attributes. We test this assumption by embedding an amenity with a known market price-a popular monthly public transport pass-in a large-scale discrete choice experiment with German workers. Many workers, including public transport users, overvalue the ticket by more than 100%, despite WTP for other attributes aligning with the literature. A complementary belief-elicitation experiment shows that advertising an amenity, such as the pass but also common amenities like work from home, causally shifts beliefs about unlisted attributes of the job. Posted wages similarly signal unlisted attributes so that wage variation, the money metric for WTP calculation, is itself contaminated by belief spillovers-such as higher pay signaling heightened stress. These spillovers imply that discrete choice estimates capture perceived bundles rather than isolated attributes, and distort current estimates of non-wage compensation and monopsony power. |
| Keywords: | Amenities, discrete choice, job postings |
| JEL: | J31 J32 J42 C83 C90 D83 |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:crm:wpaper:26184 |
| By: | Andersen, Erik T. J.; Graffy, Simon; Kerwin, Jason T.; Lambon-Quayefio, Monica |
| Abstract: | Addressing the massive test score gaps between rich and poor countries will require programs that are both high-impact and scalable. This study uses the results of a randomized controlled trial in low-fee private schools in Ghana to study a program that meets both needs. The Tools for Foundational Learning Improvement program increased test scores by 0.5 standard deviation after just nine months of intervention. A machine learning method decomposes the effects by predicted test scores if the students did not receive the treatment, and the findings show that the gains were larger for weaker students. Moreover, the program’s impacts scale roughly linearly with time compared to a shorter-term, smaller-scale pilot randomized controlled trial. The program’s developers used generative artificial intelligence to accelerate lesson plan development and adaptation to new settings. An observational pilot test of this adaptation in Uganda yielded comparable results to those of this study’s randomized controlled trial. The study developed a model in which basic skills constrain the development of advanced skills, which predicts the pattern of effects observed across early reading capabilities, and it makes forecasts about the future impacts of the program as it continues into second grade. |
| Date: | 2026–07–30 |
| URL: | https://d.repec.org/n?u=RePEc:wbk:wbrwps:11433 |
| By: | Dickinson, David (Appalachian State University) |
| Abstract: | We examine time-of-day (TOD) effects on moral choice by comparing afternoon (DAY: 3pm-5pm) and nighttime (NIGHT: 3am-5am) decisions across three incentivized tasks. While the NIGHT treatment successfully induced sleepiness, we report no evidence that circadian suboptimality shifts decisions towards one’s default ethical tendencies. Rather, the NIGHT treatment increased dishonesty in a victimless Die-Roll task but not when dishonesty harms another participant (modified Coin-Flip task) or in consequential moral dilemmas affecting only others’ outcomes (Payoff-Trolley task). Exploratory evidence from response times suggests immorality results from deliberative processes. Overall, these results indicate that moral choice exhibits task-dependent stability during circadian fatigue. |
| Keywords: | behavioral economics, honesty, trolley problem, moral choice, sleep, time-of-day |
| JEL: | C90 D01 D91 |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp18839 |
| By: | Hattori, Keisuke |
| Abstract: | Organizations invest in cohesion so that team members care about one another's outcomes, yet such concern may affect not only how much members work but also what they work on. We study this distinction in a team-production model in which members allocate effort between safe routine work and risky experimentation. Each member values a teammate's payoff and is risk averse. Greater mutual concern raises the return to contributing to team output but magnifies the payoff risk that experimentation imposes on teammates even more. As a result, members work harder while experimenting less. Expected team output can fall even as total effort rises and can be U-shaped in the degree of mutual concern. Under a material-payoff criterion, the welfare-maximizing degree of mutual concern is generally interior. The same mechanism implies that larger teams can be more diligent yet less experimental when concern toward each teammate remains fixed. With heterogeneous exposure to team performance, more exposed members work harder but experiment less; when experimental authority must be concentrated in one member, assigning it to the less exposed member yields higher expected output. Finally, the core diligence--conservatism mechanism survives when experimental output is only partially shared across teammates, while differentiating experimental approaches encourages experimentation. Mutual concern therefore affects not only how hard team members work but also how they allocate initiative inside teams. |
| Keywords: | team production, other-regarding preferences, experimentation, risk taking, organizational design |
| JEL: | D23 D81 J24 M54 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:zbw:esprep:342603 |
| By: | Hongseok Choi; Jeongbin Kim; Matthew Kovach; Kyu-Min Lee; Euncheol Shin; Hector Tzavellas |
| Abstract: | We study how differences in AI-generated financial recommendations are transmitted into individual portfolio choices. In an experiment with 400 employed adults enrolled in workplace defined contribution pension plans in South Korea, participants allocate a hypothetical pension balance across eleven products and may revise it after receiving one of two fixed AI-generated recommendations. A $2 \times 2$ design randomizes recommendation content and whether the recommendation includes a short rationale. Approximately 37$\%$ of the experimentally induced difference between the aggressive and conservative recommendations passes through to final portfolios. This causal contrast changes expected portfolio return, volatility, allocations across risk grades, and the number of products held, but produces no detectable difference in computed Sharpe ratios. 81$\%$ of participants revise. Among revisers, 95$\%$ move toward the assigned recommendation and implement about half of the suggested adjustment. Rationales do not detectably alter pass-through. These results show that users partially and selectively transmit recommendation content into economically meaningful differences in risk exposure while retaining substantial weight on their initial choices. |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2608.11371 |
| By: | Costa-Ramon, Ana; Schaede, Ursina; Slotwinski, Michaela; Brenøe, Anne Ardila |
| Abstract: | The "child penalty" significantly reduces women’s lifetime earnings and pension savings, but it remains unclear whether these gaps are the deliberate result of forward-looking decisions. This paper provides novel evidence on the role of information constraints in mothers’ labor supply decisions. We first document descriptively that mothers are largely inattentive to the long-term financial consequences of reduced hours. In a large-scale field experiment that combines rich survey and administrative data, we then provide mothers with objective, individualized information about the long-run costs of reduced labor supply. The treatment increases demand for financial information and future labor supply plans, in particular among women who underestimate the long-term costs. Leveraging linked employer administrative data one year post-intervention, we observe that mothers who underestimate the long-term costs increase their labor supply by 6 percent over the mean. |
| Keywords: | Field experiment; Labor supply; Inattention; Child penalty |
| JEL: | J16 J22 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19543 |
| By: | Manshu Khanna; Bozhang Xia |
| Abstract: | We study how the similarity of individuals' information shapes collective action. When people draw on a common source of information, such as social media, each becomes more confident about what others have seen and will do. This can help them coordinate, but it can also tempt them to free-ride. We show that which force prevails depends on how demanding the collective goal is. In a content-moderation experiment, subjects decide whether to pay a cost to report harmful content, which is removed only if enough reports are received. We vary the similarity of group members' information, holding fixed what each learns on her own, and independently vary the removal threshold. More similar information impedes reporting when few reports suffice and facilitates it when many are required, lowering reporting by 17 percentage points under an easy threshold and raising it by 34 points under a demanding one. This confirms the central comparative static of the theory of information similarity (Basak, Deb and Kuvalekar, 2026). Elicited beliefs trace the reversal to perceived pivotality and document systematic miscalibration of it. Subjects overestimate pivotality across all regimes, and their beliefs respond to similarity in line with actual pivotality only at intermediate thresholds: easy thresholds produce unrecognized pivotality, and near-unanimous thresholds produce illusory pivotality. The two response-miscalibrated patterns coincide with welfare losses; only under aligned pivotality does greater participation translate into greater collective success and higher welfare. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.25472 |
| By: | Dickinson, David (Appalachian State University); Norwood, Jacob (Appalachian State University) |
| Abstract: | Food insecurity affects roughly 18 million people in the United States and 2.3 billion globally. Given its prevalence, this study aims to examine how food insecurity and resulting anxiety affect cognitive control and decision-making. Samples of n=102 food-insecure and n=102 food-secure individuals were recruited from the Prolific online subject pool to complete an incentivized Stroop Task and a Monetary Risk Taking task. Randomized treatment assignments administered either a High or Low Cognitive Load manipulation prior to decision making by prompting participants to consider a more or less expensive unexpected financial expense. While food insecurity did not impact cognitive control and risky choices as hypothesized, it may indirectly affect outcomes via anxiety. And, exploratory analysis of response times data suggest food insecurity may promote some increased risk taking due to less deliberative decision processing. The role of anxiety and its effect on choice process among the food-insecure likely merits further investigation. |
| Keywords: | food insecurity, cognitive load, anxiety, cognitive control, risky-choice |
| JEL: | D00 D91 I30 |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp18841 |
| By: | Silvia Dominguez-Martinez (University of Amsterdam); Randolph Sloof (University of Amsterdam) |
| Abstract: | This paper investigates how managers choose between communication and delegation when informed workers may have different objectives. We study a laboratory principal-agent environment in which only the worker observes the payoff-relevant state of the world. In three equally likely states, preferences are aligned; in a fourth state, preferences are opposed. The ex ante degree of interest alignment is exogenously varied via the likelihood that the conflict state occurs. We allow managers either to retain authority, delegate authority, or restrict the worker’s discretion under delegation. We find that the manager’s choice between communication and delegation per se exhibits no significant monotone relationship with interest alignment, but conditional on delegation managers impose tighter restrictions when interests become less aligned. Under communication, workers almost always communicate truthfully when interests are aligned, but distort information when interests conflict; overall, messages become more informative and influential as ex ante interest alignment increases. The comparison with equilibrium benchmarks is nuanced: communication is more informative than the pure-strategy benchmarks predict, while the mixed-strategy equilibrium benchmark accounts for part of the observed aggregate pattern. The remaining communication patterns are qualitatively consistent with limited behavioral heterogeneity, especially a modest role for credulous managers, suggesting that small departures from the fully strategic benchmark may help sustain informative communication even when workers remain largely strategic. Overall, managers often retain authority and rely on communication; when they delegate, restrictions increase as interests become less aligned. Communication remains viable because workers distort information mainly when interests actually conflict. |
| Keywords: | Delegation, Communication, Laboratory Experiment, Organizational Economics |
| JEL: | C90 D80 M20 |
| Date: | 2026–08–02 |
| URL: | https://d.repec.org/n?u=RePEc:tin:wpaper:20260047 |
| By: | Quy Lam (University of California, Merced); Ketki Sheth (Department of Economics, University of Tennessee) |
| Abstract: | Deterrence theory assumes that harsher penalties reduce misconduct by raising its cost. Yet enforcement often depends on individuals who witness or respond to wrongdoing, and their behavior may also be shaped by the severity of punishment. Using a high-powered online experiment framed in a workplace setting, we test whether increasing penalties reduces willingness to report. We find that harsher penalties suppress reporting: employees are 1 percentage point less likely to report for every 10 percentage point increase in the penalty. This effect is mediated by perceived fairness: reporting declines only when penalties are seen as excessive. Despite the decrease in reporting, higher penalties still deter misconduct overall, as their direct deterrent effect outweighs the indirect increase in law-breaking caused by lower reporting rates. We also find that individuals support harsher penalties before a violation occurs, when deterrence is relevant, but prefer more lenient consequences after the offense has happened. This helps explain how policies can arise and persist in equilibrium, even when people are unwilling to enforce them in practice. These findings advance understanding of how social norms and perceptions of fairness shape the decisions of those who observe misconduct, and highlight the need to design sanctions that align with public willingness to act on violations. More severe penalties can discourage reporting, weaken enforcement, and ultimately undermine the effectiveness of deterrence mechanisms. |
| Keywords: | Reporting; Whistle-blowing; Deterrence; Penalty Severity; Perceived Fairness; Misconduct; Enforcement; Online Experiment |
| JEL: | C91 D63 D91 K42 M54 |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:ten:wpaper:2026-02 |
| By: | Alfonsi, Livia; Bassi, Vittorio; Rasul, Imran; Spadini, Elena |
| Abstract: | The Covid-19 pandemic represents one of the most significant labor market shocks to the world economy in recent times. We present evidence from a field experiment to understand whether and why skilled and unskilled workers were differentially impacted by the shock, in the context of a low-income economy, Uganda. We leverage a panel of workers and firms, tracked from 2012 to 2022, including high frequency surveys over the pandemic. In 2013, workers were randomly assigned to receive six months of sector-specific vocational training, in one of eight high productivity sectors. We document that over the pandemic, employment and earnings margins follow V-shaped dynamics, whereby the outcomes of treated (skilled) workers are more severely impacted by lockdowns, they recover more quickly between lockdowns, and remain resilient to the shock as the economy recovers. Cumulatively over the pandemic, skilled workers spend 61% more time than controls employed in one of our study sectors, and their total earnings are 17% higher. We explore supply- and demand-side mechanisms through which the returns to skills are maintained through the crisis. We document that skilled workers are more exposed to the shock because they are more likely to be laid off during the first lockdown as firms respond to the rapid, severe and uncertain shock by immediately laying off higher earning workers. However, skilled workers recover quickly because of their greater accumulation of sector-specific experience pre-pandemic, and the certifiability of their skills that allows them to switch employers in the same sector during the crisis. Our findings have implications for understanding the returns to skills acquired through vocational training in good economic times and times of crisis. |
| Keywords: | Vocational training; Skills; Employment; Resilience; Pandemic; Experiment; Uganda |
| JEL: | J24 O12 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19441 |
| By: | Elias Fern\'andez Domingos; The Anh Han |
| Abstract: | Technological races create tension between speed and safety: actors may gain by moving faster than competitors, even when risky development is harmful. This is prominent in debates about artificial intelligence (AI), where competitive pressure is often argued to incentivise riskier, less safety-conscious development. We study this using a framed behavioural experiment based on an idealised AI race, in which paired participants repeatedly chose between Safe and Unsafe development under an uncertain time horizon. Unsafe development gave faster progress and higher immediate payoffs but accumulated private risk up to a treatment-specific maximum of 10\%, 60\%, or 90\%; the race's competitive structure was held constant, and only this maximum risk varied. Neither the pre-registered comparison between risk levels nor the role of elicited risk preferences was supported by the data. Instead, exploratory analyses motivated by the task's repeated structure show that Unsafe behaviour is shaped less by risk preferences than by the evolving strategic state of the race: participants are more likely to choose Unsafe after their opponent does so, being ahead reduces Unsafe play while falling behind increases it, and first-round choices predict later behaviour. To interpret these effects we introduce a reduced evolutionary model with four strategies -- Always Safe, Always Unsafe, Conditionally Safe, and Conditionally Antisocial Safe -- which reproduces the treatment effect and shows how conditional Unsafe behaviour can be favoured by competitive race dynamics. Together, the experiment and model show that unsafe development can emerge from early behavioural momentum, opponent behaviour, and fear of falling behind, rather than from risk preferences alone, suggesting policy should focus on reducing competitive pressure and promoting cooperation in AI development rather than only individual risk. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.26034 |
| By: | Selva Demiralp; Orkun Saka; Michael Weber |
| Abstract: | We study how politically polarized television news shapes household inflation expectations in Turkey. We combine new household panel data with transcript-based measures of channel-level inflation coverage during prime-time news and a large-scale randomized information experiment. Inflation coverage differs sharply across politically aligned channels, and viewers sort into outlets consistent with their political affiliations. Within-person estimates show that greater coverage on a respondent’s initially preferred channel leads to higher inflation expectations, with effects concentrated among viewers of politically neutral channels. Experimental evidence indicates that belief updating depends on prior anchoring as well as on source credibility and reveals the political costs of rising inflation expectations. |
| Keywords: | inflation, expectations, media, polarization |
| JEL: | D84 E31 E52 E58 L82 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_12854 |
| By: | Juan C. Yamin |
| Abstract: | Experimenters often run pilots, but how much a small pilot should shape the main-wave design has no settled answer. This paper shows how noisy pilot evidence should guide treatment assignment probabilities in two-wave experiments. Two canonical rules mark the extremes. Balanced assignment guards against worst cases but ignores evidence that one arm is noisier. Feasible Neyman allocation adapts, but with a finite pilot it can overreact to noise, producing arbitrarily large precision losses. I propose a Conditional Minimax Regret (CMR) rule that minimizes worst-case regret over a finite-sample confidence set for the treatment and control variances. CMR retains balance's worst-case protection with high probability, converges to the Neyman allocation as the pilot grows, and attains the minimax-regret rate up to constants. It extends to multi-arm and stratified designs, and simulations calibrated to four field experiments show it avoids feasible Neyman's severe small-pilot losses while matching its large-pilot gains. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.16982 |
| By: | Bazley, William; Jannati, Sima; Korniotis, George |
| Abstract: | We examine whether grit affects the preferences and trading decisions of U.S investors. Grit is an important non-cognitive personality trait that is malleable and captures the sustained effort toward a goal despite setbacks. Using experiments formalized within prospect theory, we find that grit reduces loss aversion. Gritty investors are also more willing to exit losing investments and accumulate about 7% more wealth relative to control participants. Overall, our results suggest that grit affects the quality of investment decisions. Therefore, interventions cultivating grit could improve households' financial outcomes. |
| Keywords: | Prospect theory; Loss aversion; Disposition effect; Personality traits |
| JEL: | G11 G40 G41 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19513 |
| By: | Philip Oreopoulos; Ruochong Dong; Nina Low |
| Abstract: | We explore the efficacy of an online after-school tutoring program that complements teacher-provided computer-assisted learning practice. The aim is to scaffold students struggling with online Khan Academy practice at school with low-cost, one-on-one virtual supervision for about an hour a week at home. In a two-year randomized trial with the Toronto District School Board, teachers identified struggling Grade 4–8 students, whose parents were then randomly offered additional online tutoring to support their practice at home. The main obstacle was participation. Only 45% of assigned students reached a first session in Year 1. After we reframed the invitation and simplified enrollment, initial take-up in Year 2 rose to 83%, though weekly attendance remained below half—not because students dropped out, but because they attended intermittently. On average, the offer raised Khan Academy practice by about 10 minutes per week. This translated to an average gain in math assessment scores of 0.06 SD, with a confidence interval too wide to rule out null or medium-sized gains (ITT 0.055 SD; 95% CI [−0.09, 0.20]). We find a marginally significant gain in math report-card marks (about 0.08 SD; 95% CI [−0.01, 0.17]) and improvements in math satisfaction and enjoyment, with effects larger in Year 2 than in Year 1, consistent with the higher take-up that year. We therefore conclude that the program holds promise, but more is needed to ensure students actually engage with tutors on a regular basis. |
| JEL: | C93 D91 I21 I24 I28 |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35622 |
| By: | Sili Zhang; Andreas G. B. Ziegler |
| Abstract: | Do consumers naively reward corporate value endorsement? In an online market experiment, U.S. consumers reason about and buy from sellers who endorse values with varying strategic incentives and costs. Consumers reward value-endorsing sellers but discount strategically motivated endorsements; surprisingly, costly endorsements do not outperform cheap ones. Consumers' beliefs appear sophisticated: they anticipate pandering, understand selection, but also expect a larger strategic response when endorsement is costly. A rational-inference benchmark approximates observed willingness-to-pay for strategic sellers, indicating that consumers reward a calibrated probability of genuine alignment. Open-ended narratives reveal consumers' sophistication outside the experiment, where profit and genuine motives are both recognized. These findings help explain why cheap value appeals persist and suggest that the main friction is not consumer credulity, but limited responsiveness to the intensity of costly commitment. |
| Keywords: | markets, value endorsement, strategic inference, sophistication |
| JEL: | D12 D83 C90 M14 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_12909 |
| By: | Shan Huang; Renke Schmacker; Hannes Ullrich |
| Abstract: | AI can raise productivity by extracting information from rich data, yet little is known about how experts weigh AI-generated signals against established decision-support tools. We conduct a nationwide survey experiment with 372 Danish primary care physicians (21.5% of all clinics), who make diagnostic and treatment decisions on urinary tract infection vignettes before and after receiving a diagnostic signal. Holding accuracy constant, we randomize between-subjects whether the signal appears as an AI prediction or as a commonly used dipstick test result. Physicians update beliefs 41% less in response to AI than to dipstick signals, consistent with AI skepticism. Roughly one-third of physicians largely ignore the AI tool; linked administrative data show that these non-adopters exhibit lower technology use at their clinics, but that they are similar to adopters in clinical practice and prescribing measures. When physicians use the AI tool, they ignore asymmetry in informativeness between positive and negative signals and, when shown both the AI and a redundant signal, exhibit correlation neglect. These frictions in information processing lead to increased antibiotic prescribing with the AI signal in our setting. Our findings highlight the importance of training and information design for AI implementation. |
| Keywords: | expert decision-making, artificial intelligence, healthcare, mental models |
| JEL: | D81 I11 D83 J24 O33 |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:crm:wpaper:26199 |
| By: | Fergusson, Leopoldo; Guerra, José-Alberto; Robinson, James A. |
| Abstract: | Since formal rules can only partially reduce opportunistic behavior, third-party sanctioning to promote fairness is critical to achieving desirable social outcomes. Social norms may underpin such behavior, but they can also undermine it. We study one such norm —the “don’t be a toad†norm, as it is referred to in Colombia— that tells people to mind their own business and not snitch on others. In a set of fairness games where a third party can punish unfair behavior, but players can invoke the “don’t be a toad†norm, we find that the mere possibility of invoking this norm completely reverses the benefits of third-party sanctioning to achieve fair social outcomes. We establish this is an anti-social norm in a well-defined sense: most players consider it inappropriate, yet they expect the majority will invoke it. To understand this phenomenon we develop an evolutionary model of endogenous social norm transmission and demonstrate that a payoff advantage from adherence to the norm in social dilemmas, combined with sufficient heterogeneity in the disutility of those who view the norm as inappropriate, can generate the apparent paradox of an anti-social norm in the steady-state equilibrium. We provide further evidence that historical exposure to political violence, which increased the ostracization of snitches, raised sensitivity to this norm. |
| Keywords: | Social norms; Fairness; long-run effects of conflict; Laboratory experiment |
| JEL: | C91 D30 D91 D74 N46 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19519 |
| By: | Lundgren, Magnus; Tallberg, Jonas |
| Abstract: | Artificial intelligence (AI) is rapidly transforming economies, societies, and polities, raising fundamental questions about how it should be regulated. Policymakers face choices over whether to prioritize innovation or safety, rely on public oversight or private self-regulation, and govern nationally or internationally. Yet little is known about how citizens evaluate these competing priorities. Here we report a conjoint survey experiment conducted in seven countries with diverse political and economic profiles. We find that citizens strongly support regulating AI and generally prioritize safety over innovation, public governance over private self-regulation, and international over national approaches. The preference for safety is strongest among those who perceive AI as risky, unpredictable, and personally consequential. These findings reveal a systematic misalignment between dominant regulatory approaches and citizen preferences. |
| Date: | 2026–07–16 |
| URL: | https://d.repec.org/n?u=RePEc:osf:socarx:b5re6_v1 |
| By: | Shan Huang; Renke Schmacker; Hannes Ullrich |
| Abstract: | AI can raise productivity by extracting information from rich data, yet little is known about how experts weigh AI-generated signals against established decision-support tools. We conduct a nationwide survey experiment with 372 Danish primary care physicians (21.5% of all clinics), who make diagnostic and treatment decisions on urinary tract infection vignettes before and after receiving a diagnostic signal. Holding accuracy constant, we randomize between-subjects whether the signal appears as an AI prediction or as a commonly used dipstick test result. Physicians update beliefs 41% less in response to AI than to dipstick signals, consistent with AI skepticism. Roughly one-third of physicians largely ignore the AI tool; linked administrative data show that these nonadopters exhibit lower technology use at their clinics, but that they are similar to adopters in clinical practice and prescribing measures. When physicians use the AI tool, they ignore asymmetry in informativeness between positive and negative signals and, when shown both the AI and a redundant signal, exhibit correlation neglect. These frictions in information processing lead to increased antibiotic prescribing with the AI signal in our setting. Our findings highlight the importance of training and information design for AI implementation. |
| Keywords: | Directed expert decision-making, artificial intelligence, healthcare, mental models |
| JEL: | D81 I11 D83 J24 O33 |
| Date: | 2026–08–07 |
| URL: | https://d.repec.org/n?u=RePEc:bdp:dpaper:0106 |
| By: | Malani, Anup; Kinnan, Cynthia; Conti, Gabriella; Imai, Kosuke; Miller, Morgen; Swaminathan, Shailender; Voena, Alessandra; Woda, Bartek |
| Abstract: | Universal health coverage is a widely shared goal across lower-income countries. We conducted a large-scale, 4-year trial that randomized premiums and subsidies for India’s first national, public hospital insurance program, called RSBY. We find substantial demand (∼ 60% uptake) even when consumers were charged a price equal to the premium the government paid for insurance. We also find substantial adverse selection into insurance at positive prices. Insurance enrollment increases insurance utilization, partly due to spillovers from use of insurance by neighbors. However, healthcare utilization does not rise substantially, suggesting the primary benefit of insurance is financial. Many enrollees attempted to use insurance but failed, suggesting that learning is critical to the success of public insurance. We find very few statistically significant impacts of insurance access or enrollment on health. Because there is substantial willingness-to-pay for insurance, and given how distortionary it is to raise revenue in the Indian context, we calculate that our sample population should be charged a premium for RSBY between 67-95% of average costs (INR 528-1052, $30-60) rather than a zero premium to maximize the marginal value of public funds. |
| JEL: | D1 I13 |
| Date: | 2024–08 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19326 |
| By: | Claudia Curi; Andreas Dibiasi; Francesco Nicolini; Matteo Ploner; Mirco Tonin |
| Abstract: | Using original survey data from working-age individuals in Northern Italy, we study longevity beliefs and other key knowledge components relevant for retirement decisions. We find substantial dispersion in population longevity beliefs, with 40% of individuals misestimating life expectancy by more than five years, and highly fragmented knowledge across domains. Providing actuarial life expectancy information – tailored by age, gender, and county of residence – does not affect beliefs about individuals’ own longevity or retirement plans, reflecting a disconnect between beliefs about one’s own longevity and that of others. We compare our results with forecasts provided by 262 academic experts. On average, experts accurately anticipate the levels and correlations of longevity literacy with other knowledge components, but they mispredict a strong updating response of own longevity expectations to the provision of population information that deviates markedly from what we observe in the data. |
| Keywords: | subjective longevity, expert forecasts, financial literacy, retirement |
| JEL: | D83 J26 D15 G50 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_12907 |
| By: | Cevat Giray Aksoy; Nicholas Bloom; Steven Davis; Victoria Marino; Cem Özgüzel |
| Abstract: | Remote work has expanded rapidly, but the value of regular in-person contact remains unclear. We report a randomized controlled trial in which a large multinational randomly assigned 248 customer-service employees to remain fully remote or to work at the office together one day a month. Monthly office days gradually increased productivity, with treated employees handling 7.8% more calls per hour in the post-intervention period. Office days also strengthened workplace communication: treated employees spent 36 additional minutes communicating with colleagues in the week after an office visit, were more likely to report receiving manager feedback, and employee pairs randomly assigned as desk neighbors were 11 percentage points more likely to communicate afterward. In addition, monthly office days reduced attrition by a third. The resulting gains in productivity and retention generated a benefit-cost ratio of about 5:1. These findings show that limited, coordinated in-person contact can improve communication, performance, and retention in remote teams. |
| Keywords: | remote work, productivity, retention |
| JEL: | M54 J24 J63 |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:crm:wpaper:26177 |
| By: | Bharti, Nitin Kumar (University of Western Australia); Malik, Samreen (New York University Abu Dhabi); Mukhopadhyay, Abhiroop (Indian Statistical Institute, Delhi); Prakash, Nishith (Northeastern University) |
| Abstract: | Despite near-universal school enrollment across many developing economies, the provision of quality education that cultivates lifelong learning and the capacity to apply knowledge in novel circumstances remains elusive. We conduct a cluster-randomized controlled trial in 132 public schools in Uttar Pradesh, India, to evaluate a guided, discovery-based science pedagogy at two intensity levels: a high-intensity Mobile Science Lab (MSL) and a lower-intensity Lab on Bike (LoB). MSL improves motivational beliefs and self-confidence by 0.15-0.18 standard deviations, reduces perceived barriers to education by 0.23 standard deviations, raises engagement by 0.17-0.22 standard deviations, and increases standardized test scores by 0.22-0.34 standard deviations across all subjects. LoB produces limited average effects, with gains concentrated among students completing all sessions. These findings demonstrate that pedagogical design and delivery intensity are critical determinants of multidimensional human capital formation, and that discovery-based pedagogy can shift motivational beliefs, engagement, and achievement in low-capacity public school systems. |
| Keywords: | experiential pedagogy, curiosity, student engagement, randomized controlled trial, human capital, India |
| JEL: | C93 D83 I21 I24 O15 |
| Date: | 2026–05–20 |
| URL: | https://d.repec.org/n?u=RePEc:idd:wpaper:1 |
| By: | Nguyen, Mai Anh |
| Abstract: | Nowadays, robots are much more than simple machine tools that assist people in their daily lives. With technological advances, they have taken on a central role in many areas, assisting users in their decisions and actions as social partners. While this can be advantageous when robots function flawlessly, enhancing efficiency and performance for users, it also poses risks when robots make mistakes and these are accepted without hesitance. In this context, human error occurs as a result of overreliance on the robot. This dissertation addresses the problem of overreliance on faulty robots from a psychological perspective. It aims to uncover psychological mechanisms, based on cognitive biases, and their influencing factors that contribute to this problem and are specific to human-robot interaction (HRI). An overarching conceptual framework was first developed through a literature review in robotics psychology (conceptual research) to identify and structure general psychological processes within HRI. This showed that cognitive biases in particular are a bottleneck for erroneous robot use. Consequently, two empirical research projects, each consisting of two empirical studies, were conducted to investigate the phenomenon of automation bias (AB) (empirical research project 1) and the belief and spread of fake news (empirical research project 2) in the context of HRI. This contributes to gaining insights into the underlying psychological mechanisms of erroneous reliance on faulty robots with individual and collective implications. The first empirical research project uses a laboratory experiment in a company (study 1.1) and an online experiment (study 1.2) and is based on the theoretical foundations of the “Automation Authority and Superiority Theory” and the “Diffusion of Responsibility Theory”. Study 1.1 shows that users are subject to AB in their perception and behavior when dealing with a faulty robot. Users engage in an implicit ability comparison with the robot, in which they perform worse than the robot and, as a result, defer to the robot’s supposed superiority. On this basis, they comply with the faulty robot. AB is also influenced by an interaction between perceived robot competence and task complexity. Study 1.2 shows that when dealing with a faulty robot, users are so strongly influenced by AB in their perception and behavior that their decision-making quality is impaired. Emotional tasks strengthen this effect, while high reliability of the robot weakens it. It is also shown across both studies that AB indirectly reduces satisfaction with the robot, indicating a negative consequence of AB. The second empirical research project consists of an online experiment (Study 2.1) and a laboratory experiment (Study 2.2) and is based on the “Signal Detection Theory” to measure the belief in fake news using the two indicators “sensitivity” and “response bias”. It shows that people are more susceptible to fake news in both online and face-to-face situations when it is delivered by robots. Compared to fake news transmitted by humans, they have a high tendency to carelessly accept news from robots and a reduced ability to identify fake news. Both studies show that this discriminatory ability is reinforced by a triad of cognitive biases due to robot-, user-, and interaction-specific factors. Finally, the study emphasizes that a reduced ability to discriminate increases the spread of fake news, indicating the amplyfying role of the robot as a medium for fake news dissemination. In sum, this dissertation contributes to the research field of robotic psychology by expanding the understanding of the mutual influence between humans and robots from a psychological perspective. Specifically, it derives a coherent theoretical framework of psychological processes that takes into account the specific characteristics of HRI. In addition, this dissertation advances the understanding of the psychological mechanisms of biased HRI – characterized by the erroneous reliance on a faulty robot due to cognitive biases – by extending the phenomena of AB as well as belief and spread of fake news to the domain of HRI and identifying robot-, user-, and context-related influencing factors. This allows for a more holistic view of biased HRI. Finally, practical implications are derived to guide the handling of biased HRI more towards trustful and responsible HRI. |
| Date: | 2025–11–10 |
| URL: | https://d.repec.org/n?u=RePEc:dar:wpaper:160844 |
| By: | Aleksei Chernulich; Nobuyuki Hanaki; Yuki Tamura |
| Abstract: | Obvious strategy-proofness makes truth-telling, in theory, an obvious choice. We experimentally study its role in the object-reallocation problem with single-peaked preferences. We compare two efficient and strategy-proof mechanisms, Top Trading Cycles and the Crawler, under static and gradual implementations. In our experiment, all four mechanisms generate the same outcome under truth-telling, but only the Gradual Crawler is obviously strategy-proof. The Gradual Crawler achieves significantly higher empirical efficiency than the other mechanisms. By contrast, the Gradual TTC does not outperform the Static TTC, suggesting that gradual implementation alone is not sufficient and that the Gradual Crawler benefits from its stop-or-continue procedure. |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:dpr:wpaper:1318 |
| By: | Alejandro Rodriguez Dominguez; Miquel Noguer i Alonso |
| Abstract: | How much capital a trading strategy can absorb before its edge disappears is a causal question about how much is deployed, but it is answered with observational proxies that rest on incompatible assumptions. We ask what experiment would answer it instead, and show that two features of the problem interact to constrain any answer. Deployed capital erodes the edge gradually, so a trial of fixed length measures less than the eventual effect; and parallel implementations of one strategy trade the same securities, so they are not independent units. Comparing implementations on the same date removes market-wide shocks, which is what makes the comparison credible. But the crowding created by the strategy's own accumulated position is common to those implementations too, and an arbitrary date effect absorbs it exactly: the comparison that makes the experiment robust is the one that prevents it from measuring the crowding capacity is about. A same-date design recovers one implementation's private response at the prevailing level of aggregate positioning, and reaching the aggregate effect requires either implementations with deliberately different exposure to that position or variation in it over time. We characterise what each route identifies and what it costs, establish how far a fixed holding period understates the eventual effect and how to correct for it, and show what a finite set of deployment levels can and cannot reveal. A calibration on a purpose-built panel illustrates the resulting design rules and prices a study that would follow them. |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2608.08405 |
| By: | Wang, Jingbin; Hu, Wuyang |
| Abstract: | Non-pecuniary effects of sugar-sweetened beverage (SSB) taxes can alter consumer behavior beyond the price mechanism, thereby affecting the overall effectiveness of the tax. Recognizing and strategically leveraging these effects is essential for enhancing policy performance. The strong linkage between tax salience and non-pecuniary effects provides a promising lever for addressing this challenge. In a controlled experimental setting, we varied tax salience across five levels and applied a difference-in-differences framework to identify non-pecuniary effects and compare their magnitudes across treatment arms. Results show that non-pecuniary effects are present under most salience conditions and significantly reduce SSB demand. However, the strength of this effect follows an inverse U-shaped pattern as salience increases—initially intensifying, then diminishing. Further analysis suggests that the turning point may be attributable to boomerang mechanisms triggered by overly salient tax framing. These findings underscore the critical role of tax salience in shaping behavioral responses to corrective food taxes and offer practical insights for the design of more effective public health policies. |
| Keywords: | Food Consumption/Nutrition/Food Safety |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ags:aaea26:404572 |
| By: | Lee Crawfurd (Center for Global Development); Theo Mitchell (Center for Global Development) |
| Abstract: | Approximately half of children in low- and middle-income countries have elevated blood lead levels. The WHO recommends calcium and iron supplementation for lead-exposed individuals with nutritional deficiencies, but acknowledges this is based on “very low-certainty evidence.†We conducted a systematic review and meta-analysis of 11 randomized controlled trials (n=3, 666 enrolled). Across all seven calcium studies, the pooled effect on blood lead was −1.33 μg/dL (95% CI: −2.87 to +0.21 ; p=.078 ), attenuating to −0.36 μg/dL when restricted to 4 low-risk-of-bias studies. Iron supplementation reduced blood lead by −0.31 μg/dL (95% CI: −0.61 to −0.02 ; p=.045 ). The three studies from settings with widespread calcium deficiency—which are also the three high-risk-of-bias studies—showed substantially larger effects. We rate certainty of evidence as very low for calcium and moderate for iron. These interventions offer modest potential benefits at low cost but cannot substitute for primary prevention. |
| Date: | 2026–05–14 |
| URL: | https://d.repec.org/n?u=RePEc:cgd:wpaper:746 |
| By: | Gabriela Smarrelli (Center for Global Development); Esme Kadzamira (University of Malawi); Thi Le (Center for Global Development); Tionge Saka (University of Malawi) |
| Abstract: | Approximately one billion children worldwide report having experienced physical, emotional, or sexual violence, yet existing evidence suggests this figure substantially underestimates true prevalence, as individuals often avoid disclosing their experiences of violence. We study which survey methods yield more reliable measures of violence against children. We randomly assigned 6, 000 children aged 8 to 12 years across 251 schools in Malawi to respond to a survey using face-to-face (F2F) interviews or audio computer-assisted self-interviews (ACASI), administered either at the child’s home or school. We find that ACASI, relative to F2F, substantially increases the disclosure of physical and emotional violence from peers and school staff, regardless of survey location. Disclosure of sexual violence more than doubles when using ACASI and when surveys are conducted at school rather than the child’s home. These effects are consistent with the greater privacy provided by ACASI and the additional reassurance offered by school settings, which provide physical spaces that may more strongly signal to respondents that their answers are unlikely to be overheard, reducing social and emotional barriers to disclosure. Our results are robust to multiple checks examining social desirability bias, lapses in attention, and misunderstanding of survey questions, and suggest that ACASI offers a cost-effective approach to reducing systematic measurement error in surveys on sensitive topics. |
| Keywords: | measurement, survey methods, survey experiment, school-based violence, sexual violence, economics of education |
| JEL: | C81 C82 C83 I29 J13 |
| Date: | 2026–06–30 |
| URL: | https://d.repec.org/n?u=RePEc:cgd:wpaper:751 |
| By: | Wright, Liam; Tomova, Georgia D; Brown, Matt; Tsigaridis, Konstantinos G. (University of Cambridge); Henderson, Morag; Silverwood, Richard J. |
| Abstract: | Background: Mixed-mode designs can generate bias in all types of analysis of survey data due to systematic differences in how survey participants may respond according to mode – a phenomenon referred to as a mode effect. There is widespread evidence of mode effects upon item measurement (e.g., item means and proportions), but less on associations between survey items, despite estimating associations being a primary use of survey data. Methods: We estimated average differences (means and proportions) in item responses according to mode using data from an experiment in which participants were randomized to respond via web questionnaire, video interview, or in-person interview. We then used these estimates to simulate (counterfactual) data from two web-then-in-person sequential mixed-mode surveys of young adults in the United Kingdom (Next Steps and the Millennium Cohort Study, MCS) as if they were collected using a single mode, web. We estimated associations between variables using this simulated data and compared these against associations obtained using the observed (i.e., factual) data; the difference between the two provides a measure of variation in variable correlations reflecting the mixed-mode survey design. Results: Mode effects were observed for an array of items comparing web with in-person and video modes; differences between video and in-person were more muted. Participants in web were more likely to report negatively valent values for sensitive items, in particular (e.g., poorer physical and mental health, lower life satisfaction, and greater financial difficulty), though effect sizes were generally small with few exceeding 0.2 SD and all below 0.3 SD. The impact of mode effects on variable associations in the mixed-mode surveys was also generally small: a 0.004 SD change in absolute effect sizes on average in both Next Steps and the MCS. Conclusion: Items related to sensitive characteristics showed small mode effects when comparing web and in-person or video interview modes. Mode effects did not appear to translate into substantial changes to associations between variables as assessed in a sequential web-then-in-person mixed-mode survey compared with a counterfactual web-only survey. |
| Date: | 2026–07–31 |
| URL: | https://d.repec.org/n?u=RePEc:osf:socarx:r97ht_v1 |
| By: | Fleck, Lara (ROA, Maastricht University); Becker, Dominik (Federal Institute for Vocational Education and Training (BIBB)); Fregin, Marie-Christine (ROA, Maastricht University); de Grip, Andries (ROA, Maastricht University); Pfeifer, Harald (Federal Institute for Vocational Education and Training (BIBB)); Weis, Kathrin (Federal Institute for Vocational Education and Training (BIBB)) |
| Abstract: | With the advent of generative artificial intelligence, prompting skills are becoming increasingly relevant in the workplace. Using a discrete choice experiment (DCE), we asked decision-makers on hiring in German firms in all sectors of the economy are asked to choose between job applicants with different skills bundles. Applicant profiles vary in five attributes: prompting skills, occupation-specific skills gaps, social skills, gender and salary expectations. We find that prompting skills increase applicants’ hiring probability by 4 percent and employers are willing to pay 2 percent above the average salary of a skilled worker in their firm. This WTP is modest compared to the WTP for having matching occupation-specific skills or high social skills. However, in large firms as well as firms that have adopted or are planning to adopt AI, high prompting skills increase applicant’s hiring probability by 9 percent. Moreover, prompting skills have some leverage to balance out low to intermediate social skills. Yet, higher prompting skills do not compensate for occupation-specific skills gaps. Instead of such a compensatory effect, the demand for prompting skills seems to upgrade skills requirements in jobs. |
| Keywords: | prompting skills, generative AI, discrete choice experiment, willingness to pay, skills complementarities |
| JEL: | J23 J24 M51 O33 |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp18871 |
| By: | Ambler, Kate; Bakhtiar, M. Mehrab; DeBrauw, Alan; Uddin, Mohammad Riad |
| Abstract: | Credit market failures may reflect voluntary withdrawal by risk-averse borrowers in addition to supply-side constraints. We conduct a randomized trial with 1, 517 Bangladeshi households, offering cattle financing through conventional loans or profit-sharing contracts that spread risk between the farmer and the financial partner. Overall, interest in and take-up of the profit-sharing contracts were modestly higher than the conventional loans. However, conventional loan take-up was much lower among risk-averse farmers, and profit-sharing eliminated the take-up gap between risk-averse and non-risk-averse farmers. We find that it is male risk preferences that are associated with these decisions even when contracts explicitly target women. Livestock investment increases under both contracts with no evidence of moral hazard under profit-sharing. |
| Keywords: | Agricultural Finance, Farm Management |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ags:aaea26:404341 |
| By: | Ekaterina Siniakova (University of Amsterdam); Oda Sund (Nordic Institute for Studies in Innovation); Joël van der Weele (University of Amsterdam) |
| Abstract: | Inequality statistics highlight specific aspects of economic reality. Economists traditionally rely on technical indices such as the Gini coefficient, valued for their formal properties. More recently, scholars have emphasized measures like the income share of the Top 10%, which more directly evoke political conflict. We investigate whether inequality metrics shape public perceptions and support for redistribution, and how they are used in the media. In an online experiment (N = 1, 246) among the general public, we show that "political'' measures reduce acceptance of inequality relative to the Gini coefficient, which itself has a desensitizing effect on fairness and emotional ratings. The behavioral effect of switching from the Gini to the Top 10\% share is quantitatively similar to switching 0.12-0.17 Gini points, e.g. from an equal society like the Netherlands to much more unequal Argentina. Academic experts and policy makers are less susceptible, but not immune to the metric effect, and predict a strong behavioral effect on the general public. Finally, we show that left-wing media are more likely to report top-share metrics and less likely to report the Gini coefficient than right-wing media, likely amplifying political divisions. Our findings show that inequality statistics do not merely describe economic reality, but shape normative assessments of the income distribution. |
| JEL: | C91 D63 P16 |
| Date: | 2026–08–02 |
| URL: | https://d.repec.org/n?u=RePEc:tin:wpaper:20260050 |
| By: | Basiglio, Stefania; Del Boca, Daniela; Pronzato, Chiara |
| Abstract: | This paper evaluates the impact of the “Coding Girls†educational enrichment program designed to address the underrepresentation of women and girls in the fields of science, technology, engineering, and math (STEM). Implemented in Italy, the program seeks to stimulate the interest of young female students in programming and science, encouraging them to consider STEM†related fields of study. Through a randomized controlled trial involving 28 classes across 10 secondary schools in Turin, Italy, we assess the program’s effectiveness. Our results show that Coding Girls significantly enhances the programming skills of both male and female students. Moreover, it raises awareness among students about gender disparities in STEM†related professions. While the program positively influences boys’ aspirations to pursue higher education in STEM disciplines, it does not have a similar effect on girls. |
| Keywords: | Gender; Stem; Higher educational choice |
| JEL: | J16 I23 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19477 |
| By: | Adsera, Alicia (Princeton University); Arenas, Andreu (University of Barcelona); Boix, Carles (Princeton University) |
| Abstract: | We examine how individuals value political and economic goods and institutions through a conjoint experiment in eight countries: Argentina, urban China, India, Mexico, the Philippines, Singapore, South Africa, and the United States. Respondents choose between societies that randomly vary in personal and country income, inequality, social mobility, safety, public healthcare, and free elections. We estimate marginal effects and willingness to pay for outcomes at the country and respondent levels. Valuations of personal and country incomes and safety are similar across countries. By contrast, the value of democracy, defined as free elections, varies substantially. The average Chinese respondent is willing to give up democracy for an increase in individual income equal to one quarter of average country income, whereas the average American would do so only for twice the U.S. average country income. Unlike income and safety, for which most individual-level variation is explained by personal characteristics, variation in preferences for free elections is substantially explained by respondents’ country. Our findings contribute to research on human preferences, demand for democratic institutions, and policy and redistributive preferences. |
| Keywords: | democracy, safety, willingness to pay, conjoint experiment, income |
| JEL: | H11 H51 I13 |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp18844 |
| By: | Goulas, Sofoklis; Gunawardena, Bhagya; Megalokonomou, Rigissa; Zenou, Yves |
| Abstract: | Using Greek administrative data, we examine the impact of being randomly assigned to a classroom with a same-gender top-performing student on both short- and long-term educational outcomes. These top performers are tasked with keeping classroom attendance records, which positions them as role models. Both male and female students are influenced by the performance of a same-gender top performer and experience both spillover and conformist effects. However, only female students show significant positive effects from the presence of a same-gender role model. Specifically, female students improved their science test scores by 4 percent of a standard deviation, were 2.5 percentage points more likely to choose a STEM track, and were more likely to apply for and enrol in a STEM university degree 3 years later. These effects were most pronounced in lower-income neighbourhoods. Our findings suggest that same-gender peer role models could reduce the underrepresentation of qualified females in STEM fields by approximately 3 percent. We further validate our findings through a lab-in-the-field experiment, in which students rated the perceived influence of randomized hypothetical top-performer profiles. The results suggest that the influence of same-gender top performers is primarily driven by exposure-related factors (increased perception of distinction feasibility and self-confidence) rather than direct interactions. |
| JEL: | J24 J21 J16 I24 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19432 |
| By: | Harsh Parikh; Gabriel Levin-Konigsberg; Nilesh Tripuraneni; Dhruv Madeka; Michael I. Jordan; Dean Foster; Dominique Perrault-Joncas; Alexander Volfovsky |
| Abstract: | Randomized controlled trials (RCTs) are fundamental tools for causal inference across technology companies, pharmaceutical research, and federal agencies. While the standard difference-in-means estimator provides unbiased treatment effect estimates, it often lacks precision, particularly when treatment effects are heterogeneous or outcomes exhibit heavy-tailed distributions. Although numerous precision-enhancing methods exist---from covariate adjustment techniques to variance reduction strategies---recent research demonstrates that no single estimator performs optimally across all datasets. Rather than seeking the best estimator for individual RCTs, which risks compromising scientific validity through convenient selection, we propose a principled framework for identifying optimal estimators within families of RCTs based on specific analytical goals. Our approach uses sample splitting to estimate the distribution of evaluation metrics (e.g., mean squared error, regret) across RCT families, enabling systematic comparisons between estimators while maintaining asymptotic guarantees. We demonstrate this framework using a sample of Amazon's Supply Chain Optimization Technology trials and the Strengthening Democracy Challenge dataset (25 interventions). Results reveal that optimal estimators vary significantly by analytical objective: weighted least squares performs best for inference goals, while difference-in-means minimizes regret for decision-making contexts. This work provides actionable guidance for estimator selection while preserving methodological rigor across diverse research applications. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.23254 |
| By: | Yuehao Bai; Xun Huang; Joseph P. Romano; Azeem M. Shaikh; Max Tabord-Meehan |
| Abstract: | This paper considers design-based inference on the average treatment effect in finely stratified experiments, where uncertainty arises only from the randomized treatment assignment. We focus on settings in which units are first stratified into groups of fixed size according to baseline covariates and, then within each group, exactly one unit is assigned to treatment. In this setting, we introduce a class of graph-Laplacian variance estimators in which strata form the vertices of a weighted graph and edge weights determine how between-stratum comparisons are aggregated. The canonical estimator of Imai (2008) corresponds to a complete graph with edge weights normalized so that each stratum has weighted degree one, while a paired-stratum estimator arises from a perfect matching graph. For the subclass of degree-calibrated graphs, in which each vertex has weighted degree one, we derive an exact bias identity showing that the corresponding estimators are upward-biased, with bias governed by squared differences in the true stratum-level treatment effects across adjacent strata. As a result, any such estimator may be used for valid inference. The identity further suggests that paired-stratum estimators constructed from a covariate-based perfect matching can induce small biases when treatment effects vary smoothly with the covariates. Without such smoothness, however, we show that paired-stratum estimators can exhibit large worst-case bias, and that, within the class of degree-calibrated estimators, the complete-graph estimator is minimax optimal for normalized bias under a weak bound on treatment-effect heterogeneity. Motivated by this contrast, we propose a regularized graph estimator that controls worst-case normalized bias while preserving much of the locality of the paired-stratum estimator. Simulations illustrate the resulting tradeoff between locality and worst-case protection. |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2608.10177 |
| By: | Berriochoa, Kattalina M.; Morris, Katy; Mcneil, Andrew |
| Abstract: | People often misjudge their place in the national income distribution, but national distributions are composites of heterogenous local contexts. Drawing on evidence that people tend to anchor to more immediate reference groups, we examine whether perceptions of local income position are more accurate and whether correcting local misperceptions has a stronger effect on inequality beliefs and policy preferences than correcting national misperceptions. Using a survey experiment in the United States, we find that perceptions of household position are equally inaccurate at both local and national levels, with respondents defaulting to the middle rung of the income ladder. Correcting these misperceptions shifts attitudes towards meritocracy and redistribution, with effects more consistently significant under local framing. However, local effects are not always statistically different from comparable national treatments, suggesting both corrections matter. Analyses of mechanisms indicate that corrections change perceptions of own financial wellbeing but do not change beliefs about societal inequality. |
| JEL: | J1 N0 |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:ehl:lserod:140425 |
| By: | Bet Caeyers; Lina Cardona Sosa; Sarah Cattan; Sonya Krutikova; Abu Siddique |
| Abstract: | Preschool enrollment has expanded rapidly in low-income countries, but quality has lagged, and evidence on how governments can improve it remains limited. We evaluate Lively Minds, a scalable early childhood program adopted and implemented by Ghana’s Ministry of Education that trains community mothers to deliver structured, play-based activities in public preschools and support children’s learning at home. Using a cluster randomized controlled trial in 80 rural preschools, we find that the program increases cognitive development by 0.12 standard deviations and reduces problem behaviors by 0.10 SD on average. Cognitive gains are similar for children of participating and non-participating mothers, whereas behavioral improvements are concentrated among children of participating mothers. Consistent with this pattern, only participating mothers improve their knowledge of child development and engage in more developmentally supportive interactions with their children at home. These findings suggest that engaging parents — even those with limited formal education — can generate additional benefits beyond those of approaches that rely solely on teachers and paraprofessionals. A cost-benefit analysis indicates that the program’s private benefits substantially exceed its costs, with no adverse effects on mothers’ well-being. |
| Keywords: | early childhood development, preschool quality, parenting, scalability |
| JEL: | J13 I10 I20 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_12910 |
| By: | Francis J. DiTraglia; Craig McIntosh; Isaac Meza; Joyce Sadka; Enrique Seira |
| Abstract: | Pawn loans offer borrowers a substantial degree of repayment flexibility in exchange for a harsh penalty in case of default: forfeit of collateral worth more than the loan amount along with any payments made toward recovery. Using a large RCT conducted in Mexico City, we document key stylized facts about pawn lending and explore the merits of replacing flexibility with structured repayment contracts in this important but understudied form of credit. Our experimental design includes a mandatory frequent-payments arm, a (status quo) flexible payments arm, and a choice between the two. This design point-identifies not only the average treatment effect, but also the effects of treatment on the treated and the untreated along with the average selection on gains, allowing a rigorous study of mandates versus choice. Although the average treatment effect of assigning borrowers to structured payments is a 19% decrease in their financial cost and a 17.5% decrease in the probability of default, only 11% of borrowers choose structured repayment contracts voluntarily. We show that structured repayment reduces financial costs for nearly all borrowers, including those who would not freely choose it, and find no evidence of selection on gains in cost savings. |
| Date: | 2026–08 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2608.13775 |
| By: | Jones, Rachel |
| Abstract: | Although new agricultural technologies are widely expected to enhance the productivity of rural households, evidence on how households reallocate resources to capitalize on these innovations is limited. To explore how households respond to a profitable, but risky new technology, I conduct a framed field experiment played with maize growing couples that simulates the exchange of labor and income in Mozambican households. When presented with a familiar local maize variety that is relatively insensitive to the wife’s labor input, households allocate labor to husbands’ maize production fairly efficiently. But when faced with an improved maize variety whose higher returns would induce cooperative households to increase maize labor, couples forego 17% of expected cooperative income. Analysis of husbands’ income sharing rules reveals that wives would have received just 23% of the productivity gains that would have accrued to households, had they allocated their labor so as to fully exploit the New technology, and would have borne 2.4 times the downside risk faced by husbands. These results suggest that stickiness in intra-household income sharing may dampen the reallocation of productive resources in households that adopt profitable new technologies. Novel empirical evidence on real life intra-household income sharing from sample communities corroborates key findings from the game. |
| Keywords: | Consumer/Household Economics |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ags:aaea26:404783 |
| By: | Brockmeyer, Anne; Garfias, Francisco; Suárez Serrato, Juan Carlos |
| Abstract: | Can the provision of public goods strengthen fiscal capacity and foster tax compliance in developing countries? We study this question using a large-scale randomized infrastructure investment in Mexico City and administrative property tax data. Despite substantial improvements in local amenities, property values, and economic conditions, we find no evidence that infrastructure investments increased tax compliance---even when the tax-benefit link was made salient. These null results hold across different measures, subgroups, and empirical strategies, and we can rule out even small causal effects. By precisely estimating the limits of reciprocity-based compliance, our findings refine the fiscal contract theory and challenge its applicability beyond narrow elites. Equipped with this evidence, policymakers can redirect efforts toward more effective approaches for strengthening state capacity, such as enforcement and administrative reform. |
| JEL: | H41 H71 O23 |
| Date: | 2024–08 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19354 |
| By: | Edoardo Gallo; Rebecca Heath; Jonathan Lusthaus; Federico Varese |
| Abstract: | Market design research in economics naturally focusses on how to improve market efficiency. Our objective here is exactly the opposite - how to design interventions that make a market less efficient. Our research is inspired by the growth of illicit markets online where reducing their efficiency may reduce societal harm. Using a web-based experiment, we find that a partial disruption to delivery is an effective method to decrease market efficiency. The decrease is borne by sellers who sell fewer goods and have lower earnings. A consequence of a disruption to delivery, however, is an increase in market concentration because it facilitates the emergence of a dominant seller. In contrast, we find that attacks on seller ratings are ineffective at reducing market efficiency. This study paves the way for evidence-based, causally driven investigations to aid policies to disrupt cybercrime and other illicit markets. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.24389 |
| By: | Hernan Bejarano (Center of Economics Research and Teaching, Economics Division (CIDE) and Economic Science Institute, Chapman University); Joris Gillet (Middlesex University); Ismael Rodriguez-Lara (Universidad de M´alaga and Economic Science Institute, Chapman University) |
| Abstract: | We report the results of three preregistered experiments examining the efects of negative random endowment shocks on generosity. Using a dictator game, we show that generosity is greater towards those who experience a negative random shock, and this efect is larger than the efect inequality has on giving. Our data provide evidence that people do not expect this additional level of generosity after a shock and that injunctive norms cannot explain it either. Taken together, our fndings suggest that the (high) level of generosity observed after negative random shocks is unexpected and exceeds what people consider to be socially appropriate. |
| Keywords: | generosity, dictator game, negative random shocks, beliefs, injunctive norms |
| JEL: | C91 D03 D60 D81 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:chu:wpaper:26-09 |
| By: | Piyush Akimitsu |
| Abstract: | A red herring, an irrelevant passage added to a problem, corrupts a language model's reasoning and, through it, its final answer, while the form of the response survives untouched. The benchmark, called the Graduate Economic Reasoning Benchmark (GERB), is sixty graduate-level microeconomics problems, each a detailed setup with a verified final answer and a step-by-step reference solution. Each problem has two versions, one with the red herring and one without, and each of those is asked in two ways, one requesting an explanation and one not. This is a within-subject $2\times2$ factorial experimental design. Thirty-eight language models answer all four versions of every problem. The clean problems (the control group) are already hard, with the models answering under sixty percent correctly on average. The red herring lowers the probability of a correct final answer by 12.3 percentage points, about a quarter of the models' mean accuracy of 0.525. The damage is largest on the problems the model rates as easy. Reasoning ability confers no protection, as the red herring's effect does not differ detectably across models with and without reasoning ability. It does change how the failure looks, since a model with no reasoning mode repeats one wrong answer across waves while a reasoning model wavers. The red herring also leads a model to rate a problem as easier than its clean version, while answering it wrong more often. Although open- and closed-weight models reach the same accuracy, the open-weight models reach it at a substantially lower cost per correct final answer. The form of the response is preserved even as its substance fails. The model still produces an explanation (explanation given), the final answer still follows from the reasoning shown (coherence), and, in the aggregate, it remains the same across waves (consistency). |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.23424 |
| By: | Pattnaik, Subhransu; Kramer, Berber; Ward, Patrick |
| Abstract: | Digital credit has emerged as a promising tool for addressing liquidity constraints among smallholder farmers, yet evidence on its longer-term effects on agricultural investment and production remains limited. This paper examines how access to digital agricultural credit shapes farmers’ investment decision and production outcomes over time using a two-phase cluster randomized intervention implemented in Odisha, India. We combine difference-in-differences, and staggered treatment estimators to assess how impacts evolve over years. We found limited evidence of immediate increases in aggregate agricultural input expenditure and durable asset accumulation. In contrast, we find robust evidence of positive effects on agricultural production among existing producers. Taken together, the findings indicate that digital credit primarily operates by relaxing liquidity constraints and improving the productivity of existing farming operations rather than inducing new entry into agriculture or triggering immediate capital deepening. The results highlight the importance of distinguishing between short-run and long-run responses to financial interventions and suggest that agricultural investment is dynamic nature and can take time to materialize. |
| Keywords: | Agricultural Finance, Farm Management |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ags:aaea26:404346 |
| By: | Barba, Francesca R.; Drichoutis, Andreas; Palma, Marco A.; Cerroni, Simone |
| Abstract: | Risk preferences play a central role in farmers’ decisions, such as technology adoption, diversification, and uptake of insurance and other risk-management tools; yet agricultural economics has largely focused on risk aversion, while higher-order risk preferences remain less explored. This paper studies the external validity of experimentally elicited risk aversion, prudence, and temperance in a sample of 162 Italian winegrowers. Preferences are measured through incentivized binary lottery choices and linked to stated intentions, self-reported risk-management behavior, and an observed preventive-technology adoption. Preliminary results show that all three attitudes are present in the sample and capture distinct behavioral dimensions. Risk aversion has no systematic association with external outcomes. Prudence is positively related to stated intentions, but not to observed behavior. Temperance is associated with current insurance status and with the absence of active defense practice adoption. Overall, higher-order risk preferences show limited, outcome-specific external validity for agricultural risk-management behavior. |
| Keywords: | Risk and Uncertainty |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ags:aaea26:404409 |
| By: | Lars Meyer-Waarden (TSM - Toulouse School of Management Research - UT Capitole - Université Toulouse Capitole - Comue de Toulouse - Communauté d'universités et établissements de Toulouse - CNRS - Centre National de la Recherche Scientifique - TSM - Toulouse School of Management - UT Capitole - Université Toulouse Capitole - Comue de Toulouse - Communauté d'universités et établissements de Toulouse); Julien Cloarec (UJML - Université Jean Moulin - Lyon 3 - Université de Lyon); Manon Ferreira (TSM - Toulouse School of Management Research - UT Capitole - Université Toulouse Capitole - Comue de Toulouse - Communauté d'universités et établissements de Toulouse - CNRS - Centre National de la Recherche Scientifique - TSM - Toulouse School of Management - UT Capitole - Université Toulouse Capitole - Comue de Toulouse - Communauté d'universités et établissements de Toulouse) |
| Abstract: | The rise of AI-driven technologies, particularly generative AI (GenAI) like ChatGPT, is transforming customer service in hospitality and tourism sectors. These systems enhance user experiences by offering real-time, personalized support and 24/7 assistance. Yet, as sustainability becomes a key factor in travel decisions, consumers increasingly face information overload due to the abundance of eco-friendly options. This study examines how GenAI can alleviate this cognitive strain in sustainability-focused tourism contexts. Across two experiments (i.e., one on trip planning and another on on-site shopping), the findings show that more advanced GenAI tools significantly reduce information overload, improving decision-making and thus enhancing user perceptions of trust which in turn trust increases both well-being and behavioral intentions to use GenAI tools. On the other hand, the reduction of perceived decision risk increases behavioral intentions to use GenAI tools. In particular, by simplifying access to and comparison of eco-certified travel and shopping options, GenAI encourages more environmentally responsible consumer behavior. |
| Keywords: | Generative AI, Sustainable tourism, Information overload, Well-being, Trust Risk |
| Date: | 2026–12 |
| URL: | https://d.repec.org/n?u=RePEc:hal:journl:hal-05703102 |
| By: | Campbell, Arthur; Thornton, D.J.; Zenou, Yves |
| Abstract: | We study the role of influence in a model of the diffusion of social behaviors in a network. Individual behavior creates either positive spillovers (public goods) or negative spillovers (public bads). Our notion of influence captures the causal effect of an agent's adoption decision on the adoption decision of others in the network. We study a phase transition in equilibrium behavior around which viral equilibria—where diffusion occurs among a nontrivial fraction of the population—emerge. Public goods exhibit a continuous phase transition in equilibrium adoption, while public bads exhibit a discontinuous transition-- they emerge suddenly. Our findings reconcile disparate evidence that attending a public protest is a strategic complement in some settings and a substitute in others. |
| Keywords: | Diffusion; Public goods; Public bads; Narratives; Social Networks |
| JEL: | D43 D85 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19443 |
| By: | Hamed, Allaudeen; Massa, Massimo; Ni, Zhenghui; Zhou, Zhou |
| Abstract: | We study the impact of subsidiaries on the information environment of the firm using a natural experiment: the pandemic-triggered, localized lockdowns of subsidiaries. We find that the randomized lockdowns of subsidiaries enhanced price informativeness, improved analyst forecast precision and increased trading and profitability of short-sellers. At the same time, these lockdowns reduce the informational advantage of insiders, leading to a decline in their trades. The combined effect of these factors ultimately results in an overall increase in stock liquidity. Our findings challenge the conventional wisdom, revealing an unexpected impact of the firm's organizational structure on informational efficiency. Specifically, firms with more localized information (via subsidiaries) are less informational efficient. |
| Keywords: | Information; Liquidity; Analysts; Insiders; COVID-19 |
| JEL: | G12 G22 G23 G32 |
| Date: | 2024–08 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:19391 |
| By: | Elizabeth Burland; Nora Delaney; Susan Dynarski; Katherine B. Leu; CJ Libassi; Katherine Michelmore; Stephanie Owen; Mary A. Quiroga; Elizabeth Salinas; Susan M. Dynarski |
| Abstract: | Earnings inequality among college graduates has grown, and returns to a bachelor's degree (BA) vary substantially across colleges. We experimentally evaluate a program that shifted high-achieving, low-income students into higher-quality colleges. Students randomly offered a guarantee of four years of tuition and fees at the University of Michigan were 21 percentage points more likely to enroll at a UM campus. The offer shifted students across colleges rather than into college: most compliers would otherwise have attended less-selective four-year colleges. HAIL increased four-year BA completion by 4 percentage points, but had no effect on attainment after six years. It substantially increased degree quality: students offered the scholarship were 12 percentage points more likely to earn a bachelor's degree from a highly or most selective college. IV estimates indicate that students induced to attend a University of Michigan campus were 56 percentage points more likely to earn such a degree. The increase in college quality predicts earnings gains of 18 to 25 percent. Faster degree completion adds about another 1 percent to the present discounted value of lifetime earnings. |
| Keywords: | college quality, financial aid, inequality, college completion, experiment |
| JEL: | D31 I22 I24 I26 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_12884 |
| By: | Jack Rossiter (Center for Global Development); David K. Evans (Center for Global Development); Susannah Hares (Independent); Catherine Henny (Independent consultant) |
| Abstract: | Policymakers, practitioners, and donors rely on standardised effect sizes to compare education programmes and allocate resources. Yet in early-grade reading interventions in low- and middle-income countries, these metrics reflect differences in measurement tools, sample composition, and scaling choices as much as genuine learning gains. Drawing on data from 197 studies, we find that a single additional word per minute of reading fluency corresponds to anywhere from 0.03 to 0.55 standard deviations across studies. For tests designed by researchers—the most common type of test in this literature and the hardest to compare—a single correct response on a reading assessment can result in a standardised gain anywhere from 0.08 to 0.80 standard deviations. Converting raw to standardised effects also reorders which programmes appear most effective. Fewer than one in four papers (23 percent) present both standardised and raw effects, leaving readers unable to judge specifically what children can do differently as a result of the programme and whether that change matters educationally. We propose that researchers report raw effects and reference distributions alongside standardised estimates, that work on reading benchmarks be extended across more languages and countries, and that shared measurement frameworks be developed faster. Together, these reporting changes can improve interpretation for policymakers and allow evidence syntheses to explicitly model differences in test design and sample composition across studies. Each of these changes would help the field show what children actually gained from a programme, not just how large the effect appears. |
| Date: | 2026–06–03 |
| URL: | https://d.repec.org/n?u=RePEc:cgd:wpaper:748 |