Road Pricing under Heterogeneity

Congestion pricing policies are widely debated and critics often argue that toll costs outweigh congestion relief. We show theoretically that welfare gains depend on two forces: the extent of driver heterogeneity—both in preferences and in exposure to highly demanded roads and times of day—and the extent of revenue recycling. We quantify the impact of both forces using data from a large-scale experiment that tracked driving behavior with and without usage-based congestion pricing incentives for nearly 10,000 drivers over two years. Using a model of time-varying traffic equilibrium with heterogeneous drivers and a granular road network, we show that a congestion pricing policy using the experiment's price schedule would raise welfare by as much as adding a new lane to every highway and would raise enough revenue to fully compensate losers. Moreover, the extent of driver heterogeneity entails that optimal prices can improve net welfare with only a small fraction of toll revenue.