With respect to the prevailing congestion problems in the more urbanised regions of the European Union, transportation planners and policymakers are facing a dilemma. Supply-side measures, i.e. increasing the capacities, might shorten the congestion duration, especially if bottlenecks can be removed. However, the addition of extra lanes in motorway bottlenecks is generally very costly, not only in terms of investment costs, but also in terms of space consumption and environmental effects. The strategy to build to provide, that is, adjusting the capacities to the expected peak demand, is sub-optimal from an economic point of view. Economists have shown that users tend to respond to congestion in a way that is not optimal from a system point of view (see for instance Waiters, 1968). In many cases, there is a mismatch between the congestion costs experienced by an individual user and the congestion costs caused by the same user, and demand-side measures might help to stimulate 'social' behaviour. The problem is, however, that it is hard to convince the public that measures like congestion pricing are really effective in preventing congestion. The allegation that congestion pricing would only result in paid queuing instead of free queuing has successfully been used in a campaign against the introduction of road pricing in the Netherlands. The objections against congestion pricing, especially the doubts about the effectiveness of this solution, are less irrational than some advocates of the pricing mechanism try to let us believe. An example is the mismatch between the willingness- to-pay of a potential road user and the utility of the planned trip, which results primarily from income differences. A related problem is the redistribution of welfare. Many (potential) road users are 'direct losers'; that is, their direct transportation costs are higher, though this might be (partially) compensated by a form of tax refunds (Thomson, 1998). Even more convincing, however, is the argument that a knowledge gap prevents even the most flexible congestion-pricing scheme to be optimal. In the ease of simultaneous pricing, the congestion costs can be calculated exactly, but the users cannot base their choices on the exact costs (Thomson, 1998). Hence, the tolls will have to be based on demand estimations, and should not vary too much. This might result in under-utilisation of the road if the tolls are too high, or, more likely, congestion if the tolls are not high enough. Moreover, since users are not capable of perfectly performing the utility-maximisation task assumed in classical economic theory, road pricing will never be a perfect way to avoid congestion.