Abstract: This paper studies decision-making in standardized tests. I model test-taking as a multitasking problem, where students select questions guided by subjective confidence. Standard belief-formation models predict efficiency: while students may attempt too many or too few questions, they should attempt those they are objectively more likely to hit. Using data from a large field experiment, I show that multiple versions of this prediction fail. To account for the puzzle, I introduce imperfect metacognition, allowing errors in how students rank their own productivity across tasks. Structural estimation reveals substantial heterogeneity in metacognitive skill, accounting for over a third of test score dispersion. Moreover, whereas both genders are equally metacognitive, females are better at identifying harder questions, while males easier ones. This gender asymmetry implies that test score gaps should widen with productivity - a distinctive prediction confirmed in the data. My findings inform the design of scoring rules that restore fairness while rewarding both productivity and metacognitive skill.

Presentations: Northwestern Kellog School in Economic Theory*, Bocconi-CCA-Cornell Winter Workshop*, Princeton Junior Micro Seminar, Boston College PhD Conference, HEC PhD Conference, Milan PhD Workshop, 1st Berlin Micro Theory and Behavioral Econ PhD Conference, Transatlantic Theory Workshop, Monash-Warwick AYEW, CESifo / ifo Junior Workshop on the Economics of Education 2026*, CEPR Workshop on Frontiers in Measurement and Survey Methods* (u), BRIC 2026* (u), CEBI PhD Workshop in Behavioral Labor Economics (u). 

Abstract: We develop a dynamic model of narrative-based political competition. Voters recall policy and outcome frequencies, but not their correlation, allowing for multiple causal accounts. Politicians exploit this limitation by crafting plausible narratives that inflate their policy's effectiveness. Plausibility disciplines rhetoric in a history-dependent way: the less a policy is implemented, the more optimistic the narrative. Equilibrium narratives take a merit-stealing/blame-shifting form and are maximally polarized. Incumbency erodes narrative advantage, producing endogenous policy cycles; in the long-run tenure is independent of true policy effectiveness. Results are robust to narratives induced by analogy-based expectations, preferences for truth-telling, and heterogeneous voter feedback.

Presentations: 16th UniTo-CCA PhD Workshop, UC Davis Microeconomic Theory Seminar, EEA-ESEM 2024 Rotterdam, XVIII GRASS Workshop Turin, 2025 Lisbon Meetings in Game Theory and Applications, BRIC 2026* (u), ESSET 2026 (u)

Abstract: We study the supply and effects of causal rhetoric in U.S. politics. We conceptualize causal rhetoric through blame and merit -- claims that assign negative or positive outcomes to political action -- and train a supervised classifier to detect these forms in over a decade of congressional tweets. Causal rhetoric has risen rapidly and pervasively, displacing affective messaging. Its production trades off revenues and costs. Quasi-random variation in Twitter adoption shows that blame increases small-donor revenues by expanding donor count, while merit raises average donation size. Fine-grained legislative data suggest that policy ownership determines relative costs: blame is cheaper for opponents, merit for proposers. Finally, causal rhetoric has downstream effects on societal outcomes, influencing protest activity, polarization, and institutional trust.

Presentations: Riederau Workshop on Beliefs Narratives and Memory*, Bocconi-CCA-Cornell Winter Workshop, Princeton Junior PE Seminar, Workshop on Text-as-Data at Lancaster, 2nd Verona Early Career Workshop, Potsdam Text-as-Data in Behavioral Economics, XIV Alghero IBEO Workshop, Milan PhD Workshop, 1st Berlin Micro Theory and Behavioral Econ PhD Conference, 10th Monash-Paris-Warwick-Zurich-CEPR Text-As-Data Workshop, 1st CEPR Annual Interdisciplinary Conference at WZB, Ludwig Erhard ifo Research Workshop, Monash-Warwick AYEW, 9th Workshop on the Economics of Media Bias (u), RIDGE Economics Workshops 2026 (u)

Abstract: A designer relies on an experimenter to provide information to a decision maker, but the experimenter has incentives to persuade rather than merely transmit information. Anticipating this motive, the designer can restrict the set of admissible experiments, but cannot prevent the experimenter from garbling any admissible experiment. We model this situation as delegation over experiments. The optimal delegation set can be obtained by comparing maximally informative experiments among those the experimenter has no incentive to garble. When the experimenter's preferences are S-shaped, we fully characterize such experiments as double censorship. Relative to the full delegation outcome, upper censorship, double censorship features an intermediate pooling region, inducing a smaller pooling region for the highest states. We show that the designer strictly benefits from imposing a nontrivial delegation set to constrain the experimenter's ability to persuade while retaining valuable information provision. 

Presentations: University of Exeter, University of Bielefeld, CSEF-IGIER Symposium on Economics and Institutions (u) 

Abstract: We study joint decision-making when agents agree on all primitives other than signal likelihoods (i.e. they have different subjective Blackwell experiments). We characterize a measure of disagreement between decision-makers by considering signal-contingent plans which satisfy ex ante individual rationality for both agent relative to the prior-optimal outside option, uniformly across decision problems. We show our measure is prior-independent, and provide a linear test: inclusion holds if and only if each joint distribution is a convex combination of the other pair's. We then show that the unique complete, rotation-invariant extension is the cosine similarity of the two joint distributions. Applications show that greater agreement expands a normalized ex ante Pareto frontier and enlarges the set of single-model rationalizations. Our order is independent of Blackwell's and selects quadratic over KL-type Bregman divergences.

Presentations: University of Pennsylvania, Barcelona Summer Forum (u)