Research Projects
[PAPER]
Households in hurricane-prone areas face rare but potentially devastating losses, yet flood insurance coverage remains low and unstable. This paper links hurricane forecasts to flood insurance uptake, introducing forecast errors as novel, exogenous reference points for household expectations. Using administrative records of NFIP policies in Florida merged with storm forecasts, I show that deviations between predicted and realized storm outcomes generate strong behavioral asymmetries: unanticipated impacts (“false misses”) drive large increases in demand, often exceeding the response to correctly predicted strikes (“true hits”), while predicted threats that fail to materialize (“false hits”) reduce uptake, suggesting that false alarms erode perceived risk. These patterns cannot be explained by actuarial risk alone and are consistent with reference-dependent preferences and loss aversion. Salience further amplifies these dynamics, as recent storms and hurricane-classified systems elicit the strongest responses, while near-miss “close calls” often depress demand. Together, the findings demonstrate that household insurance decisions are shaped as much by the psychological impact of forecast errors as by objective risk, with implications for forecast communication, policy timing, and disaster preparedness.
This paper provides experimental evidence on how salience and firsthand experience interact to shape insurance demand. Using a controlled real-effort task with certain risk, I independently manipulate the salience of potential losses and observe naturally occurring variation in prior experience. Contrary to standard theoretical predictions, both the salience enhancements and prior exposure to loss reduce willingness to pay (WTP) for insurance on average. However, when salience and negative experience coincide, WTP increases sharply, revealing strong interactive effects. This nonlinear response suggests that salience amplifies the influence of adverse outcomes but suppresses demand when recent experience is reassuring. The results point to a behavioral framework in which attention, reference points, and recent outcomes jointly determine the perceived need for protection. These findings challenge the predictions of expected-utility theory and highlight the importance of attentional signals in shaping insurance behavior.
(data collection is currently underway)
This paper examines how reference points influence effort provision through a real-effort laboratory
experiment. In the first round, participants are asked to assemble 15 LEGO figures within 25 minutes and are compensated based on the number of minutes remaining upon completion.
In a second round, participants are randomly assigned to one of two treatments. In Treatment 1, the required output is reduced to 10 figures, with the same per-minute reward.
In Treatment 2, the original 15-figure target is retained, but participants receive a $0.75 reward per minute saved. By independently manipulating performance expectations
and pay, the design identifies whether a downward shift in the performance goal, leads to reduced effort, even when economic incentives
remain unchanged. The experiment offers a novel and transparent test of reference-dependent effort provision, isolating the behavioral effects of goal-setting from standard
monetary responses.
(data collection is currently underway)
Teaching
Winter 2025 | Class Size: 56 | Mode: Online (Synchronous) Overall Rating: 4.2/5 | Full Course Evaluation Show Selected Comments ▼
Fall 2024 | Class Size: 140 | Mode: In-Person Overall Rating: 4.39/5 | Full Course Evaluation Show Selected Comments ▼
Spring 2024 | Class Size: 87 | Mode: In-Person Overall Rating: 4.55/5 | Full Course Evaluation Show Selected Comments ▼
Fall 2022 | Class Size: 75 | Mode: In-Person Overall Rating: 1.8/1 | Full Course Evaluation Show Selected Comments ▼
About Me
I’m a behavioral and experimental economist by day, and a hurricane-data-sleuth and Lego-specalist by… well, also by day.
My job market paper explores how reference dependence shapes flood insurance demand through hurricane forecast errors. I love teaching economics, especially when I
can bring data to life, spark a good debate, or sneak a behavioral experiment into the classroom. I live with two cats: Nash (yes, named after the economist)
and Lulu. I also have more than two houseplants. When I’m not in my office, you can find me basking in the sun, baking something sweet, reading something fun, or nursing a plant back to life.