Comparison of Second-Best and Third-Best Tolling Schemes on a Road Network

Abstract

Much of the road pricing literature has focused on deriving second-best optimal tolls when only parts of a network can be tolled or tolling is constrained in other ways. A drawback of second-best tolling is that it requires extensive information on speed–flow curves and demand elasticities throughout the network. Such information is often not readily available, and errors in estimating key parameters could result in tolls that are nonoptimal or even welfare reducing. The purpose of this paper is to explore a simpler alternative policy, dubbed “no-queue” tolling, whereby time-varying tolls are imposed selectively on a road network with the objective of eliminating queuing on the tolled links. No-queue tolling is an example of third-best pricing because the effects of the tolls on other links are disregarded. To explore the merits of no-queue tolling, a dynamic traffic simulator (METROPOLIS) was used to compute no-queue tolls for individual links and cordon rings on a laboratory network. For comparison, second-best flat and time-varying tolls were also computed on the same sets of links. Two results stood out. First, even without accounting for the likely computational and acceptability advantages of no-queue tolling, it appeared to dominate flat tolling and performed relatively well compared to step tolling. Second, the benefits from no-queue tolling exhibited approximately constant returns with respect to the number of links that are tolled. This suggests that no-queue tolling could fruitfully be selectively implemented now, rather than waiting a decade or more for comprehensive road pricing to become feasible.