Optimal congestion pricing with diverging long-run and short-run scheduling preferences

Traffic congestion is among the main market failures in modern cities. Dynamic marginal external cost pricing is the textbook economic response to this externality. Recent empirical work has shown that there is an important distinction between short-run departure time choice versus long-run routine formation of commuters, also characterized by differences in values of time and schedule delays for the short-run versus the long-run problem. This paper investigates whether this affects optimal pricing of congested roads. Using a dynamic model of congestion, and integrating it with a dynamic model of routine formation, it is found that contrary to expectation, short-run optimal congestion pricing alone cannot optimally decentralize the optimal formation of long-run routines. A separate instrument is therefore needed to optimize routine formation.