Research
Working papers
This paper uses a lab experiment to study how incidental affects influence choices and beliefs in two classical choice situations - a dictator game and an effort task. We employ music to provoke an exogenous shift in participants’ affective state after they enter the lab. The experimental design is between-subject where we vary the valence - positive or negative - of the affect induced and the moment, in the timeline of the experiment, when participants listen to the music. Decisions and beliefs under the effect of the exogenous affective activation are compared to a baseline where participants’ affective state remains untouched. We collected data for 313 participants. The analysis reveals that incidental affects reduce altruism. Moreover, our treatments show that participants with high trait emotional intelligence (Petrides et al., 2007) strategically manage their beliefs when under the influence of the exogenous affective activation. Both in the dictator game and effort task, they appear to overestimate the likelihood of the negative outcome. We conjecture that they adopt a strategic pessimism approach as a way of protecting themselves from the emotional reaction should the negative outcome materialize. We also find that high trait emotional intelligence is predictive of choices in the effort task. Here, the decrease in confidence induced by incidental affects in participants with high trait emotional intelligence generates a decrease of 9% in effort provision. This result confims the importance of confidence for performance.
An increasing amount of empirical studies suggests that in many strategic interactions, people's preferences depend on guilt aversion. This paper explores the presence of heterogeneity in the way people experience guilt under different incentive schemes and studies the influence of personality traits on elicited guilt aversion. I use an experiment where participants play three rounds of the investment game by Berg, Dickhaut and McCabe (1995) with different rates of return. At the end, participants complete a Big-Five questionnaire. The results reveal significant heterogeneity in the way people trade-off guilt and monetary incentives with a prevalence of players for which guilt aversion decreases as the stakes in the game become higher. The analysis of the personality data suggests that guilt aversion is negatively correlated to openness to experience and positively correlated with neuroticism. All the results are robust upon controlling for standard measures of guilt proneness and compliance used in psychology.
We study how matching affects confidence. Our lab experiment allows us to identify the effect of being matched with others of either similar or dissimilar performance (assortative or disassortative matching) on people's confidence in their own ability. Across a variety of tasks we find that assortative matching does not have a substantial nor statistically significant effect on confidence compared to a control group with random matching. By contrast, disassortative matching has a negative effect on confidence on average that is driven by the bottom half of performers. This group becomes substantially less confident compared to random matching. We discuss potential mechanisms and implications of this result.
Work in progress
"Do Monetary Penalties Discourage Absenteeism Behaviour?" with Xiao Yu
We study the impact of a policy introduced by the Department of Education of England in 2013. The policy is intended to discourage term-time absenteeism by the deterrence effect of monetary penalties. As a consequence, parents receive fines for pupils' unauthorised holiday absences. Since the introduction of the regulation, the number of penalties issued has increased by more than six-fold to the point that in 2018-2019 local councils collected more than 5 million pounds from fines. However, reliable evidence concerning the efficacy of the policy is still lacking. Our identification strategy exploits the highly heterogeneous application of the penalty policy across Local Authorities to define treated and control group. Using school-level data on absences released by the Department of Education, we run a diff-in-diff model combined with a coarsened exact matching procedure and perform several robustness checks. Our results indicate that penalties are not effective in reducing the rate of absence for holiday reasons.