Research

Job Market Paper

Disagreement or Misunderstanding? A Field Experiment on Parent-Child Differences over School Choice with H. Vidinova

  • Whenever parents make decisions that do not reflect their children's preferences, they have previously been presumed to do so because of preference misalignment. We posit a new explanation: Parents hold inaccurate beliefs about their children’s preferences despite valuing such information. We demonstrate the economic relevance of such misinformation through a field experiment in the high-stakes context of the centralized high school application process in Bulgaria. We find parents underestimate how much their school preferences differ from their children’s, even for schools that families have previously discussed. We design an informational intervention that conveys children’s preferences in a transparent way: we share with parents their child’s personal ranking of schools. We find the intervention significantly increases the alignment between children’s school rankings and the application ranking parents ultimately send to the Ministry of Education. Not only do parents prioritize their children’s top choices, but also accommodate for differences in schools with heterogeneous projected career tracks (e.g., vocational vs. college-prep). School assignment counterfactuals based on estimating a structural model of parent-child preference aggregation show that misinformation can lead to changes in school assignment for up to 30% of children (depending on child's academic performance) and effectively reduce child's weight in the aggregation process almost 3-fold. Overall, our findings suggest that disregarding misinformation about preferences can have important economic consequences on joint household decision-making.

Freedom of Choice or Pay-off Maximization? An Experiment on Third-Party Restrictions in Public Good Games with M. Cersosimo

Working Papers

  • We experimentally investigate the objective function of third parties who can impose binding minimum contribution requirements on a group of ten players playing a linear Public Goods Game (PPG). We find that third parties exhibit a preference for honoring players’ choice autonomy, even though allowing for autonomy carries the risk of lower group pay-off. Average requirements decrease by 28.8% relative to a baseline condition where third parties have unambiguous incentives to maximize material pay-offs (without concerns about autonomy violation). Mandates do not vary substantially by the marginal net group return generated by an extra contributor. Instead, variation in intermediate mandates is better explained by a novel social preference: third parties require the minimum contribution level that ensures players' payoff is no less than their endowment. We design an additional treatment condition to understand how much of the effect is driven by (1) a distaste for violating other's exercise of choice, vs (2) a distaste for interfering with other's outcome and find the effect is largely driven by the former. We also show evidence that typical survey measures of autonomy are more correlated with the non-interference component than the freedom of choice component. Overall, our study suggests that respect for other's freedom of choice can be a non-trivial component of policy maker's objective function.

Do People Honor Other’s Autonomy Out of Principle? An Experimental Investigation

  • The decision to intervene (or not) on other’s behalf has traditionally been examined via outcome-based preferences: will the intervention improve other’s material outcome? In this paper, I examine the role of procedural preferences based on moral principles: do people have the right to intervene on other’s behalf? I design a novel experimental “intervention” task which allows the clean separation of the procedural from the outcome-based motivation: a disinterested third-party decides whether to (1) choose envelope A or B, one of which contains a bonus payment for another participant, or (2) let the other participant choose the envelope on their own. As no one knows which envelope contains the bonus, the decision to intervene cannot be driven by outcome-based motivations. I find evidence of a strict preference for avoiding to choose on other’s behalf: 30.5\% of third-parties in the main treatment condition have a positive willingness to pay to avoid choosing for others. Additional treatments suggest that more than half of the refraining third-parties do so because of a desire to honor other’s autonomy rather than because of responsibility aversion or a preference for satisfying other’s desire for agency. The existence of such procedural preference may help explain why in many real life situations individuals may not intervene even if interventions would improve other’s material outcomes.

Improving Posterior Estimation of COVID-19 Policy Effects with Synthetic Control Guided Compartmental Models with M. Charpignon, M. Cersosimo, S. Newman, C. Rastogi, J. Gimenez and J. Zou

  • It is critical to accurately estimate the impact of non-pharmaceutical mitigation policies on the COVID-19 pandemic. Direct estimation of these effects using standard epidemiological models can lead to biased estimates. This is because epidemiological model estimation is not tailored to disentangle effectively naturally occurring changes in behavior unrelated to the shelter-in-place interventions from the effect of the interventions themselves. In this paper, we propose a novel framework that combines synthetic control with compartmental models to jointly estimate policy effects. Our approach detects a 14..3 % decrease in R0, the expected number of secondary infections an infected individual inflicts. This is an epidemiologically significant effect given the multiplicative nature of R0.

Selected Work in Progress

Moral Suasion in the Field: Leveraging the Language of Promising for Job Search in Egypt

  • Descriptive data from Job Search NGOs in Egypt reveals that only 25-30% of job applicants who have been offered a job actually complete the hiring process and anecdotal evidence suggests this is due to procrastination rather than rational cost-benefit calculus. Through a field experiment, I compare the effect of two verbiages for expressing commitment to finish the process, "I will be able to” vs "I promise I will be able to”, on follow-through rates. First pilot results suggest that the morally-loaded one “I promise” increases follow-through rates due to the moral salience of the phrase. I plan to conduct a full welfare analysis of the implications of the moral nudge, with carefully studying the benefits (e.g., job seekers successfully landing the job, fiscal externalities) as well as the costs (e.g., foregone job opportunities, emotional stress induced by the promise).

Multinational Firms and The Gender Glass Ceiling Effect: Evidence from Bulgaria

with H. Vidinova

  • We study the case of Bulgaria where according to the most recent Eurostat data almost half of managers are women, way above the EU average of 33%. We hypothesise this may in part be due to the increased number of multinational firm branches on the Bulgarian labor market between 2010 and 2020, a period which has coincided with the increase in the share of female managers. Using a dataset on the full universe of operating firms in Bulgaria between 2012 and 2018, we document that the share of female managers in multinational firms is 10% higher compared to the share in domestic firms, conditional on firm-level characteristics. We plan to use a complementary employer-employee matched dataset to study the mechanisms of this heterogeneity, including employee selection as well as firm-level hiring and promotion practices.