Working Papers

1. Self-Presentation (with Leah Cao) 

We study how individuals form self-presentation strategies when publicly reporting their self-confidence about their personal attributes. In a laboratory experiment, subjects first privately state their self-confidence and then report it publicly to the other participants in the session. Subjects bias their public reports toward the middle, with low-confidence individuals tending to over-report and high-confidence individuals tending to under-report. We provide evidence that under-reporting is motivated by social image concerns, whereas we find no such evidence for over-reporting. There is moderate heterogeneity in self-presentation by gender and socioeconomic status, with under-reporting more common among female subjects and over-reporting less common among those with higher socioeconomic status. Overall, self-presentation strategies make subjects appear more homogeneous in self-confidence than they actually are, making it difficult for others to infer individuals’ true attributes and reducing the accuracy of reported confidence.

2. Motivated Beliefs: The Role of Information Accuracy  

I design an experiment to study how the uncertainty about information accuracy impacts the formation of motivated beliefs. Subjects receive noisy signals about an ego-relevant state (IQ test performance) and an ego-irrelevant state (suitcase prices) and are asked to update their beliefs. The experiment includes three treatments: (i) certain accuracy, where information is correct 70% of the time; (ii) compound uncertainty, where it is correct 50% or 90% of the time with equal probability; and (iii) ambiguity, where the accuracy is also 50% or 90%, but the probability of each is unknown. I find evidence of motivated reasoning in the certain accuracy and ambiguous information treatment, but not in the compound uncertainty treatment. In the ego-relevant task, under the certain accuracy and ambiguity treatments, subjects respond rationally to positive feedback but update too little after negative feedback. In contrast, in the compound uncertainty treatment, subjects update too little after both types of signals. In addition, I find evidence of confirmation bias, which appears to be consistent with motivated reasoning in the certain accuracy and ambiguity treatments but not in the compound uncertainty treatment. The results suggest that the relationship between uncertainty about information accuracy and motivated reasoning does not appear to be monotonic. This may reflect two opposing forces, as individuals generally dislike uncertainty, while greater ambiguity may provide more flexibility to interpret information in ways that support preferred beliefs.

3. Price ceiling regulation: Evidence from the retail gasoline market (with Georgios Gatsios and Christos Genakos) (Under Review)

We examine the short-run impact of price ceilings on retail gasoline prices in isolated oligopolistic markets, uniquely observing for which stations the regulation was binding and for which it was not. Leveraging daily pricing data and a difference-in-differences methodology, we find that, while binding stations naturally lowered prices, non-binding stations increased theirs, though there is substantial heterogeneity. Among non-binding stations, those with more favorable characteristics for collusion adjust prices faster, move closer to the ceiling, and exhibit lower price dispersion, consistent with more effective coordination. Our results provide evidence that the price ceiling acted as a focal point for collusion among non-binding stations, consistent with channels identified in tacit collusion theory.