“Exposure to poverty and productivity” with  Patricio Dalton and Charles Noussair , PLoS ONE.

We study whether poverty can induce affective states that decrease productivity. In a controlled laboratory setting, we find that subjects randomly assigned to a treatment, in which they view a video featuring individuals that live in extreme poverty, exhibit lower subsequent productivity compared to subjects assigned to a control treatment. Questionnaire responses, as well as facial recognition software, provide quantitative measures of the affective state evoked by the two treatments. Subjects exposed to images of poverty experience a more negative affective state than those in the control treatment. Further analyses show that individuals in a more positive emotional state exhibit less of a treatment effect. Also, those who exhibit greater attentiveness upon viewing the poverty video are less productive. The results are consistent with the notion that exposure to poverty can induce a psychological state in individuals that adversely affects productivity.

Link to the paper

Working Papers


The Dark Side of Monetary Bonuses: Theory and Experimental Evidencewith Patricio Dalton and Charles Noussair.

To incentivize workers and boost performance, firms often offer monetary bonuses for the achievement of production goals. Such bonuses appeal to two types of motivations of the worker. On the one hand, the existence of a goal, on its own, triggers an intrinsic motivation associated with the desire to not fall short of the target. On the other hand, the money paid to achieve the goal is an extrinsic motivation for the worker to work harder. This paper studies the possibility that these two effects are substitutes: whether and how monetary bonuses incentivizing the achievement of goals can crowd out the intrinsic motivation that goals create. We develop a theoretical model which predicts that if the worker is sufficiently loss-averse, offering a monetary payment to reach a goal set by the worker would be counterproductive. This is because the loss-averse worker sets lower goals, which in turn negatively affect subsequent performance. The results from a controlled laboratory experiment corroborate this prediction. This paper highlights the limits of monetary bonuses as an effective incentive when goals are chosen by the worker

Download the current version here


“Contracting Probability Distortions” (Submitted)

  I introduce an incentive contract designed to take advantage of the regularity that individuals distort probabilities. The contract includes risk in the agent’s compensation and the principal can specify the degree of such risk. Specifically, the principal can choose the probability that determines the extent to which the compensation of the agent depends on his performance on the delegated task. The principal can exploit this contract feature by targeting probabilities that are distorted by the agent such that the agent’s motivation to perform the task is enhanced. A theoretical framework and an experiment demonstrate that the proposed contract generates higher performance than a traditional contract. However, the specified probability is critical to achieve this result. Small probabilities generate higher levels of performance, whereas medium-ranged or high probabilities do not generate performance differences between the contracts. I show that probability distortions caused by likelihood insensitivity-cognitive limitations restricting the accurate evaluation of probabilities-explain these findings.  

Download the current version here


Social Status, Performance, and Motivated Beliefs” 

 This paper shows that individual performance in relevant and profitable tasks is affected by the social status to which individuals belong. I propose that such influence takes place through a psychological mechanism and, as a result, occurs above and beyond the material constraints that are inherent to social status. Specifically, belonging to higher social status grants individuals the access to better beliefs about their abilities, and these beliefs are crucial to performance. A theoretical framework formalizes the proposed mechanism. Two laboratory experiments show that individuals exhibit higher performance as well as higher beliefs about their own performance when assigned to a high rather than to a low status treatment. These results are robust to the provision of feedback about relative and absolute performance. Such dependence of performance with respect to social status corroborates the validity of the proposed theory against alternative explanations.

(This paper supersedes “Believe me you are (not) that bad”. CentER Discussion Paper Series No. 2016-032.)

Download the current version here



Work in progress


    “Evaluating Treatment Effects and Replicability” with Karl Schlag.

We shed new light on replications of laboratory experiments in economics and recall largely ignored methodological problems in the current praxis of analyzing experimental data. The basis of our investigation is a recent study by Camerer et al.(2016). We propose that most of the papers therein refer to a treatment effect that cannot be identified with the used test. We reinvestigate the original data sets and the replicated ones using tests that uncover treatment effects. We find that 6 of the papers in Camerer et al. (2016) do not have a treatment effect in the original data sets with the proposed test. Moreover, among those with an effect in the original data set, three present a treatment effect in the replication data whereas 8 do not exhibit a treatment effect. We comment on the importance of adequate testing before replication.

Status: Manuscript in preparation


We investigate the emergence of echo chambers in a controlled laboratory environment. A theoretical model based on Williams (2019) shows that in settings in which types are unknown but the actions of others can be observed, rational (Bayesian) agents choose to observe the actions of those with whom they expect to share similar beliefs. This is because those with similar beliefs are more likely to be of a similar type as one’s own; hence, their actions are more informative. We test the validity of this finding with two experiments. In the first experiment, types are exogenously assigned to subjects. This is advantageous in that it guarantees heterogeneity and allows us to manipulate the degree of heterogeneity. The subjects’ task in this experiment is to i) choose another subject to observe her action and ii) accurately infer their own type after observing an informative signal and observing the action of the chosen subject. The second experiment considers preferences that are naturally formed outside of the laboratory. Specifically, we implement pairwise lotteries that investigate violations of expected utility and stochastic dominance. We analyze whether subjects who are not informed about the structure of the lotteries choose to observe the action of subjects who share similar choice patterns in the past and thus those who exhibit similar deviations from expected utility.

Status: Experiments coming soon!


“Group decision-making and GARP”  with Wieland Muller

Status: Experimental implementation. 


“Alternative theories of Risk and GARP” with Wieland Muller

Status: Experimental design.


Resting Papers

“Self-Chosen Goals: Incentives and Gender Differences”with Patricio Dalton and Charles Noussair, CentER Discussion Paper; No. 2016-036.

To boost employees’ performance, firms often offer monetary bonuses when production goals are reached. However, the available evidence indicates that the particular level at which a goal is set is critical to the effectiveness of this practice. Goals must be challenging yet achievable. Computing optimal goals when employees have private information about their own abilities may be impossible for an employer. To solve this problem, we propose a compensation scheme, in which workers set their own production goals and bonuses. We provide a simple model of self-chosen goals and test its predictions in the laboratory. The model predicts that (a) the self-chosen goal contract is more cost effective than a piece rate contract for an employer interested in attaining a desired level of output, and that (b) workers set goals that they systematically outperform. Our experimental data support both predictions. We also observe sharp gender differences in the experiment. The self-chosen goal contract increases the performance of men but not of women relative to a piece rate contract. Women set lower goals, but outperform them to a greater extent than men.

Download the current version here