with Dusan Paredes, Mark Skidmore, Scott Loveridge
Natural disasters often raise gasoline prices when they disrupt upstream fuel supply, but less is known about disasters that primarily disrupt local mobility and access. We examine the short-run effect of the 2024 Park Fire on retail gasoline prices in Northern California using high-frequency station-level data and a distance-based research design. Benchmark two-way fixed effects difference-in-differences estimates show that stations near the wildfire experienced price declines relative to more distant stations, with the largest effects concentrated within 10 miles of the fire perimeter. Because treated and control stations do not fully satisfy parallel pre-trends in the weekly panel, we use synthetic difference-in-differences as our preferred estimator. The SDID results confirm a negative local price response: gasoline prices fell by 2.5 cents per gallon for stations within 20 miles of the wildfire and by 3.9 cents for stations within 10 miles during the six weeks following fire onset. The effect attenuates as the post-treatment window expands and is concentrated among rural and unbranded stations. Taken together, the spatial, temporal, and heterogeneity patterns are more consistent with a localized contraction in gasoline demand than with a broad supply-driven fuel shock. More broadly, the results show that disaster-induced gasoline price responses depend on the dominant channel of disruption and are not uniformly upward.
Dusan Paredes, Ye-Rim Lee, Mark Skidmore, Scott Loveridge
Despite the integration of national oil markets, gasoline prices in the United States display marked spatial variation. Using weekly data from over 89,000 stations between 2024 and 2025, this paper applies a multilevel framework to decompose rural–urban price disparities across stations, counties, and states. Nearly 70 percent of total price variation is explained by geography, with state-level factors alone accounting for more than half. Contrary to the idea of a universal rural premium, the national rural–urban gap is close to zero once spatial scale is properly modeled. However, results reveal strong heterogeneity across states: in 17 states urban stations charge more, while in 32 states rural stations pay higher prices. These findings highlight the importance of accounting for geographic scale and institutional boundaries when evaluating the welfare implications of energy prices.
with Erika Rosebrook, Fernanda Alfaro, Scott Loveridge, Dusan
Paredes, and Mark Skidmore
This paper analyzes determinants of inflation perceptions among 1,000 Michigan residents surveyed in 2024, in the run-up to the U.S. presidential election. Ordered logistic regressions show that political leaning is the strongest predictor: Republicans and Independents perceive significantly higher inflation than Democrats, and Trump supporters and Biden administration disapprovers report sharply higher inflation than Harris supporters or those with positive approval. Gender differences are also robust, with women more likely than men to report higher inflation across all domains. By contrast, age, income, education, and rural--urban location have weaker or inconsistent effects, while home ownership consistently dampens perceptions of inflation, particularly in housing and gasoline. These findings highlight that perceptions of inflation are politicized and gendered, reflecting partisan and leadership evaluations more than local economic fundamentals.
with Lenin H. Balza, Jose Belmar
The study explores how hydrological drought reshapes electricity generation in Latin America. Using a harmonized plant-level panel for Chile, Brazil, Colombia, Peru, and Ecuador from 2000 to 2024, I link monthly power output to HydroBASINS watershed data and high-resolution runoff anomalies. Droughts are defined as basin-level 12-month cumulative runoff falling one standard deviation below the historical mean. Two-way fixed-effects regressions indicate that hydropower generation declines by 20–25 percent during drought months in Chile and Brazil, while effects in Colombia, Peru, and Ecuador are smaller and statistically insignificant. These findings provide cross-country evidence on drought sensitivity and highlight implications for energy security, substitution patterns, and decarbonization strategies in hydro-dependent systems.
This study investigates how hurricanes disrupt retail gasoline prices across the United States and how infrastructure and policy shape recovery. Using daily station- and county-level price data linked with NOAA hurricane tracks and energy supply networks, we analyze the impacts of major 2024 hurricanes. A stacked event-study framework measures temporal and spatial price responses, while a spatial econometric model captures spillovers through interconnected markets. We further examine how state emergency responses—such as fuel waivers, toll suspensions, and trucking exemptions—affect price stabilization and market recovery. The results shed light on how supply-chain structure and state-level policy actions influence the resilience of U.S. fuel markets to extreme weather events.
Michiganders' Inflation Perception for the Last 5 Years (2019-2024)