Over the past decades increasing congestion and pollution levels in urban areas have widened the gap between the private and social costs of urban traffic. Economists typically suggest that an appropriate pricing policy, reflecting the relevant external costs of the various modes, should be an important part of any policy proposal. This raises several questions. The first problem is the computation of all relevant marginal social costs. This includes the valuation of time lost because of congestion, the valuation of reduced pollution and noise, and the valuation of changes in accident risks. Not surprisingly, these issues have been dealt with by several authors (see, for example, Newbery, 1988, 1990; and Jones-Lee, 1990). A second question is how to determine the optimal prices for car use and public transport. Under ideal conditions the economic prescription is to make transport users pay the marginal social costs of their consumption. However, in practice the problem may be more complicated if distributional considerations have to be taken into account, or if pricing restrictions impose additional constraints on the optimal pricing problem (see, for example, B? s, 1986; Glaister and Lewis, 1978). Price differentiation between peak and off-peak traffic may not be technically or politically feasible, or budget constraints may apply to the public transport authority. Finally, a third question relates to the numerical calculation of optimal prices. As the level of the marginal social cost of congestion and other externalities is itself a function of the intensity of car and bus use, an equilibrium optimum tax or price has to be computed which takes into account all demand and supply responses.