In this paper, the arguments used in the literature pro and contra congestion pricing are analysed. Although it is a first-best instrument in theory, it is argued that the assumptions needed to arrive at this conclusion oversimplify reality. In practice, congestion pricing is a second-best instrument with some advantages over other second-best instruments, but it will also give rise to numerous problems, as discussed in the paper. These problems will be illustrated with the Dutch attempts to implement an electronic road-pricing system. Some research issues relating to congestion pricing have been overlooked in the past. In the first instance, the behavioural side (motorists' responses) of congestion pricing has not been paid much attention. In general, it is argued that individuals are aiming to maximise utility (or minimise travel time). However, there seems to be an increasing recognition that this assumption does not properly describe reality. Second, the impact of the compensation scheme - used to compensate drivers who are worse off under congestion pricing - on the behavioural responses should be analysed more carefully in future work. This scheme might partly reverse the behavioural responses induced. Third, the welfare-generating properties of simple schemes should be looked at in future work. Fourth, given the potential opposition, we conclude that a cordon system, in which the price is dependent on the time of the day, is currently the most attractive option for pursuing a kind of congestion pricing. A cordon system might increase the individual's awareness of the costs of mobility during congested periods and be the first step towards more sophisticated pricing systems. (A)
Abstract