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

Network & Environment

The overall reduction in traffic crossing the congestion-charge cordon during the congestion-charge period (06.30–18.29 weekdays), was 22 % during the trial.1] This reduction was lower during the morning peak period (16%) and higher during the afternoon/evening peak (24%). The reduction stabilized quickly after the introduction of charges, and resettled at almost original levels as soon as the trial ended but not entirely (see figure below). During the permanent scheme from August 2007, traffic reductions have been almost as during the trial, but slightly lower. There are many ways for motorists to adapt in a situation with charges. During the trial approximately half of the disappearing motorists changed to public transport which increased by 6 %, and the other half changed in less traceable ways like fewer trips, trip chaining and other destinations. However, motorists did not increase car-pooling, work at home or change departure times.

A common assumption before the trial was that the time-differentiated charges would have significant impact on time-of-day choices, so that traffic volumes would increase during periods when passage was free of charge. Surprisingly enough, though, no such compensatory increases in traffic volume were observed for any time of day during the Stockholm trial. Rather, charging seemed to have a (small) reducing effect on traffic volumes over the cordon also after charging hours. This can be explained by the inherent linkage between trips over the day: The motorist, who does not take his /her car into the city in the morning, will not drive it out from the city in the evening, either. This effect must have outweighed the effect of substituting car trips between times of day.

The effect of charging is, naturally, at its largest just over the cordon. For all other types of streets, (inside the zone as well as outside) the effects were “diluted”. Only part of the car trips using those streets were subject to charges, and for other car trips no reduction should be expected1]. Therefore, the total reduction of vehicle kilometres traveled (VKT) within the charging zone, will be less than the reduction in number of passages over the cordon. Based on samples of link volumes, it was estimated during the trial that the effect on daily total VKT within the charging zone was approximately 14%, and that the corresponding figure for the region as a whole was 2%.

As a consequence of reduced demand, travel times are significantly reduced. These reductions are particularly large on the access (approach) roads to and from the inner City. Queuing times on these roads have fallen by one third for inbound traffic during the morning peak period (see next figure), and by half for outbound traffic during the afternoon/evening peak.

Also, travel time variability was reduced significantly in both AM and PM peaks. Travel time variability is known to constitute a major part of road users’ negative assessment of congested traffic conditions.

In the discussions before the trial, problems with redistribution of traffic was one of the main issues since it was decided to exclude a major motorway bypass through the city, from charging. This bypass – the Essinge link – was under severe congestion already before the trial. It was feared that through traffic that previously had used the inner city network would shift route to the Essinge link to avoid charges, and thereby further increase congestion problems there. Contrary to those expectations, however, volume increases on the Essinge link during the trial were very limited.

Air quality measurements are very sensitive to weather conditions, and do vary considerably from day-to-day and year-to-year. Therefore, the larger part of the environmental evaluation of the trial was model-generated. Emissions, concentrations and immissions (health effects) of a number of pollutants were computed for the situations before and during the trial, respectively. Real measured traffic volumes were used as input to the models.

The reduction in vehicle km during the trial, contributed basically proportionally to a reduction of the pollutant emissions from road traffic. Thus, such emissions were reduced for the region with 1%-3%, and for the inner city (charging area) with 8%-14%, see the following table. The reduction of Nitrogene oxides (NOx) were in the lower end of that range, as a consequence of additional bus traffic during the trial, which contributed to extra emissions.

Emission reductions were concentrated to the inner city, which is densely populated both day and night. Therefore, the relative effects on traffic related immissions and health problems in the Stockholm region were much larger than the effect on emissions. Based on recent estimates of dose-response relationships from several consistent international studies, it was estimated that in total for the entire Greater Stockholm area (1.44 million inhabitants, 35 x 35 km), between 25 and 30 fewer premature deaths would occur per year as a result of a reduction in long-term exposure to particles.

The congestion charging system managed to press average yearly concentrations of pollutants below what is required according to environmental quality standards (legally binding by European agreement). However, also when congestion charging is implemented, there are a number of locations in Stockholm at which maximum daily levels, as defined by European environmental quality standards, will remain to be exceeded.

1] In fact, there is reason to expect that non-charged trips may even have increased to some extent, as a consequence of reduced travel times. However, the evaluation program was not able to identify and/or quantify such rebound effects.

1]Depending on how you define the demand calculations from Stockholm show that the cost elasticity vary from - 0,27to - 0,41. Elasticities for different groups have not been estimated for Stockholm.