Luke Catalano
Economist Apprentice, Amazon Ads Economics
I work on empirical industrial organization, auction theory, causal inference, and platform and marketplace economics. At Amazon, I calibrate production demand models against randomized experiments and develop new uses for willingness-to-pay measures in advertising.
I hold an M.S. in Quantitative Economics and Finance (2026) and a B.A. in Economics with Honors (2025), both from the University of California, Santa Cruz. I plan to apply to Ph.D. programs in economics in the Fall 2027 cycle.
Curriculum vitae
Research
Experimental calibration of a production demand model in advertising
Validation of the price-elasticity input to a production willingness-to-pay metric against randomized pricing experiments across multiple marketplaces. Category-level calibration coefficients are estimated in each experiment with an IV/Wald estimator, using randomized treatment assignment as an instrument for the mismeasured price shock, and pooled across experiments with DerSimonian–Laird random effects. A pre-registered leave-one-experiment-out test showed that category-specific corrections add no out-of-sample information beyond a single global rescaling, which substantially reduces out-of-sample squared-error loss relative to the uncalibrated model and indicates the metric already ranked product categories correctly.
Willingness-to-pay and brand value in offsite ad creative
Preliminary evidence from third-party offsite ads that naming the brand in an ad's title is associated with lower click-through rates, with the penalty shrinking as the product's estimated willingness-to-pay rises — consistent with the metric capturing brand value.
Spatial heterogeneity in the effects of NYC Congestion Pricing on collisions
Estimated spatially heterogeneous causal effects of NYC's January 2025 Congestion Pricing Act on weekly motor vehicle collisions using Local Linear Causal Forests. Constructed a 62-week post-treatment panel from NYPD collision data across 3,598 hexagonal spatial units; one hex reached the 95% level after honest calibration, providing evidence against a measurable citywide collision effect.
NYC Congestion Pricing and vehicle collisions: a difference-in-differences analysis
Applied causal analysis of NYC's January 2025 Congestion Pricing Act using a difference-in-differences framework. Estimated a ~17% reduction in motor vehicle collisions during the first 10 weeks post-implementation; validated identifying assumptions (parallel trends, no anticipation) and discussed policy implications.
Injunctive norms and information about labor practices in a duopoly market
Designed and implemented a laboratory experiment (oTree/Python) testing how information about unethical labor practices and injunctive social norms affect consumer choice in a duopoly market. On pilot data (n=24), labor-practice information raised the probability of choosing the ethical seller by 57.5 percentage points, and adding an injunctive-norm treatment raised it by 61.5 percentage points, relative to a 12.5% baseline.