Microgeographic speed, reliability, and traffic externalities

Abstract

This paper estimates congestion externalities on a broad sample of roads in England’s capital region. A fixed effects approach that compares across times of day finds that 10% more traffic increases travel time by 4%, erodes reliability, and finds smallest marginal effects on high-capacity roads. I incorporate both expected travel time and reliability into a model of congestion externalities. Combining the model with travel data, I find that Central London’s Congestion Charge exceeds tolled drivers’ external costs if they only reflect travel time and reliability. Despite the large toll in London’s city centre, I find that deadweight loss in untolled parts of the city remains substantial, and that accounting for reliability increases deadweight loss by as much as 55%.