The Welfare Impact of Using Second-Best Uniform Taxes to Address Traffic Congestion

This paper uses a novel GPS-coded dataset for more than three million car trips in the four largest German cities to measure the efficiency losses from using uniform fuel taxes as a second-best optimal tax to reduce traffic congestion. We estimate the short-run price elasticity of vehicle-kilometres travelled (VKT) across different hours of the day from a panel gravity equation. Identification of the price elasticity relies on comparing VKT changes between granular trip origin and destination pairs within each city. We find that the VKT price elasticity differs strongly across trips taken at different hours of the day, and that this elasticity correlates with contributions to the congestion externality. Using a policy simulation, we show that 52% of the deadweight loss of the congestion externality remains after levying a uniform fuel tax