Transportation Economics & Technology

Nobel Prize-winning bottleneck model showing optimal time-varying tolls can eliminate queuing waste. The birth certificate of dynamic congestion pricing.

Revitalized Vickrey's bottleneck model. The most cited paper in congestion economics, establishing the canonical framework for analyzing time-varying congestion and pricing.

Established the α-β-γ framework for empirically estimating schedule delay costs. The standard methodology for valuing time in transportation.

Nobel Prize-winning paper proving multinomial logit derives from random utility maximization. His BART ridership prediction (6.3% vs official 15%; actual 6.2%) demonstrated the approach's power.

Introduced flexible mixed logit overcoming the IIA limitation of standard logit. Enabled random taste variation and correlation patterns essential for modern demand modeling.

Synthesized revealed-preference estimates of value of time from California toll road experiments. Found VOT around 50% of wage rate with substantial heterogeneity.

Established the Mohring Effect: as ridership increases, higher frequency reduces waiting time, creating user-side scale economies that justify transit subsidies.

Derived empirically tractable formulas for optimal transit subsidies accounting for congestion, pollution, accident externalities, and scale economies. Found large fare subsidies typically welfare-improving.

Developed speed-flow relationships for optimal congestion charges. Showed efficient prices far exceed gasoline taxes—foundational for road pricing economics.

Landmark study using 50 million UberX sessions. Exploited surge pricing variation to estimate $6.8 billion annual consumer surplus from UberX in the US.

First comprehensive analysis of Uber driver demographics, earnings, and labor supply using survey and administrative data on 600,000+ drivers.

Quantified flexibility value using discontinuities in surge pricing. Found Uber drivers earn more than twice the surplus they would in less-flexible arrangements.

Found UberX drivers have 30-50% higher capacity utilization than taxis due to more efficient matching through the platform.

Shows ride-hailing is prone to matching failures where high demand creates feedback loops. Surge pricing prevents market collapse—theoretical foundation for dynamic pricing.

Introduced the Expected Marginal Seat Revenue (EMSR) decision model—foundational for airline revenue management practice.

Found the expected absolute difference in fares between passengers averages 36% of ticket price, with dispersion increasing under competition.

Developed foundational methodology for structural estimation of entry models using airline data. Showed how market structure affects competitive behavior.

Documented hub premiums and market power at dominant airports. Showed airlines can charge 10-20% higher fares on routes where they dominate.

Found incumbents cut fares significantly when Southwest merely threatens entry, before actual competition begins. Evidence for contestability theory.

First quasi-experimental study of carbon taxes. Found Sweden's transport CO2 declined ~11% after implementation, with carbon tax elasticity 3x larger than price elasticity.

Comprehensive review of auto externalities—congestion, accidents, local pollution, and climate. Estimated costs per mile and evaluated policy alternatives.

Used electronic toll collection adoption to show reduced tax salience leads to higher toll rates. Pioneering evidence on how payment visibility affects policy outcomes.

Exploited California subsidy variation. Found EV demand is price-elastic (-3.3) with near 100% pass-through. Estimated $12-18 billion needed to reach 1.5 million EVs.

Estimated $936 billion annual U.S. benefits from AVs through reduced crashes, productivity gains, and fuel savings. Foundational policy survey.

Found shared AVs could be cost-competitive with private car ownership at $0.20-0.50/km, potentially transforming urban mobility economics.

Estimated willingness-to-pay for full automation at $4,900 using advanced discrete choice methods, with substantial heterogeneity across consumers.