www.curacaoproject.eu                      CURACAO - coordination of urban road-user charging organisational issues                   Funded by the EU

Road Pricing Context













Decision Making



Implementation and Evaluation



Case Studies






Dutch National Case









The Hague


Urban Road User Charging Online Knowledge Base

Impacts Predictability

Based on the available information, it seems that no predictions of effects were made. However, in the following paragraphs the elaborated predictions of adjusted and/or larger Spitsmijden follow-up are described.

One model is based on economic welfare theory and was used to determine the optimal reward level. The second model is a traffic model that allowed the simulation of different reward levels and an assessment of the global impact of the corresponding reward schemes.

The INDY model

The macroscopic dynamic traffic model, named INDY, was used to forecast the traffic conditions that would result from the introduction of a Spitsmijden reward scheme. The main aim was not to forecast what would happen during the pilot phase, but what would have happened if (a) a different reward scheme would have been used and/or (b) a larger number of people would have participated.

This model calculates an equilibrium state via an iterative process, taking into account that the congestion decrease due to behavioural adjustments of rewarded drivers is (partly) compensated by other drivers: the return-to-the-peak effect.

The INDY model shows that a participation level of 10% and 100% both lead to travel time savings, while a participation of 50% generates travel time losses. The latter effect is caused by the fact that a small number of participants changing their departure time can alleviate the congestion for many, but if too many people change, they cause congestion for themselves and for others. Changing the level of rewards causes a similar effect as changing participation level.

A high level of participation with high rewards will probably lead net travel time losses for the whole network. In practice, this combination of high reward and high participation would also be very expensive. The greatest travel time savings can be achieved by shifting a number of travellers low enough to not cause congestion for themselves or others, while decreasing demand during rush-hours and solving bottlenecks. An interesting question would then be what the best combination of participation and reward levels would be, given a certain budget.

The Bottleneck Model

The main function of the bottleneck model, an economic model, is its use as a complementary tool for traffic analysis. The complexity of the INDY model – which provides a much more detailed picture of traffic congestion in the relevant area than the bottleneck model does – complicates the determination of optimal tolls.

The model calculates an equilibrium in which all drivers can reach the same utility: the disutility of having to spend some time in the queue for those who arrive at work exactly on time is equal to that of those who arrive earlier or later and have to spend less time in the queue.

The four road segments between the inflow from highway A4 (junction Prins Clausplein) and the Voorburg ramp act as a bottleneck because of the relatively large amounts of weaving traffic here.

The optimal fine toll is estimated to be € 2.25 for passing the bottleneck at the beginning or at the end of rush-hours. Drivers who pass through the bottleneck between the beginning and the end of rush-hours receive a lower reward, and drivers who pass through exactly at the time for arriving at work at the preferred time, do not receive anything. It is clear that in this case the toll is in fact a reward: nobody has to pay, and all except the drivers who arrive at the preferred time receive some money.

The optimal toll, whose value does not change over time, equals to €1.28 between 8:05a.m. and 9:24 a.m. This toll can be easily transformed into an equivalent reward. This reward equals to €0.84 and is given to drivers who pass through the bottleneck between 7.24a.m. and 8:05a.m., or between 9:24a.m. and 9.36a.m.. No reward is granted outside this period.

Target group analysis

Moreover, a target group analysis was carried out using a logistical regression model. This showed that the amount of reward has no influence on the probability of displaying a certain reaction, but probably more effect on the frequency of behavioural adjustment. Well-educated participants were more likely to choose to work from home (monetary variant) and to travel before rush-hours (Yeti variant). Men were more likely than women to avoid traffic.