The Stockholm experience

A congestion tax system was put into trial operation in the Stockholm inner city area between 3 January and 31 July 2006. The Swedish Road Administration (SRA) was commissioned by the Government for the implementation. An extensive, sophisticated technical system was developed in cooperation with IBM Svenska AB, the head contractor, to be able to manage the daily automatic registration of 500,000 passages and 200,000 tax decisions generated. Birger Höök was appointed project manager.

Sensitive traffic system

Water divides Stockholm. This means that all through-traffic must drive past the city over a few bridges. The present traffic volume on the main thoroughfares often greatly exceeds the capacity for which they were originally built. In other words, the Stockholm traffic system has almost reached its maximum capacity. This means that even a minor traffic incident during rush hour can have major repercussions in the form of heavy congestion and overstrained public transport services. During peak hours it can take up to three times longer than normal to cover a stretch on the city access roads, thoroughfares or inner city bridges.

Full-scale trial in Stockholm

In the summer of 2003, the Stockholm City Council passed a decision to conduct a trial implementation of environmental charges in the Stockholm inner city area. It was subsequently found that a municipal charge was not legally permissible, and the charge would have to be collected as a state tax. The Government then assumed responsibility for the matter and issued the Congestion Tax Act on 17 June 2004.

The trial implementation of a congestion tax was included in what has been called the full-scale trial in Stockholm, which aimed at reducing congestion and improving the environment. The objectives were formulated as follows:

• reduce traffic to and from the city by 10–15% during rush hour
• better level of service in Stockholm city traffic
• reduction in the emissions of carbon dioxide, nitric oxide and particulate matter
• city residents will experience a better city environment.

The full-scale trial began on 22 August 2005 with an expansion of public transport services and more park and ride facilities at city access roads. The trial implementation of a congestion tax, for which the Swedish Road Administration was responsible, was launched on 3 January 2006 and ended on 31 July 2006.

Government commission

The Swedish Road Administration was assigned responsibility for the trial implementation of a congestion tax by the Government. This entailed designing and developing a system to register vehicle passages in and out of the Stockholm inner city zone along with a system to administer payments. It also meant informing all Swedish drivers and vehicle owners about the upcoming congestion tax and how the system would function. IBM Svenska AB was contracted by the SRA to design, build and operate the technical system.

Within the SRA a special project organisation was set up, reporting directly to SRA top management. The technical side of the project was divided into an implementation and an operations component. Apart from this, a specific information project was organised. The technical challenge was far from insignificant – new technology and a sophisticated system with high demands on safety and precision were to be developed. This did not, however, overshadow the great challenge of reaching the general public with information about the system and its structure. The information project was crucial for vehicle owners to be able to understand the rules that applied and how they were to pay the congestion tax.

High demands

The work on building up the congestion tax system was conducted within an extremely tight time schedule. In just over a year an entirely new system had been designed and built to handle the registration of half a million vehicle passages, some 2 million photo images and 200,000 tax decisions every day in addition to the entire administration of payments, reminders and other reports and information that were sent out.

Stringent demands were imposed on the system. All automatic functions were to be operational at least 99.9% of the time during which the congestion tax was collected. It was imperative that everything worked on a daily basis. Strict demands were also placed on correct vehicle identification. A few factors were of utmost importance for the success of the project, specifically that:

• the system was ready in time
• the technology worked
• the system was functional and efficient
• the organisation was customer-oriented
• the system was accepted by drivers.

To be able to meet the tight time schedule, milestones were planned precisely. The time schedule became an important project management tool, both at the general and detailed level. A specific contract time schedule in the terms of reference with suppliers was the basis for time management.

Technically sophisticated system

The system that was put into operation on 3 January 2006 was complex, but also highly flexible. It included over 40 automated interfaces and was in operation around the clock, 7 days a week.

Eighteen automatic control points had been set up in a ring around the Stockholm inner city area. The technical equipment at the control points was installed on gantries above the carriageway as well as in roadside control cabinets. A total of 164 cameras and 159 laser detectors monitored 78 traffic lanes at Stockholm city entrance and exit points.

Laser detectors sensed when a vehicle drove past a control, at which point the number plates were photographed, both front and rear. A sequence was simultaneously initiated, which activated an onboard unit, if the vehicle had one, and registered its signal. (See facts box). The number plates were identified directly in the cameras using Optical Character Recognition (OCR) software and then once again in the data system. The number plate was then matched with the right vehicle owner using information extracted from the national vehicle registry. Those number plates that could not be identified automatically were sent to Customer Services for manual examination. The congestion tax applied only to vehicles registered in Sweden and it was the vehicle owner who was responsible for paying the tax.

The accumulated passages by any particular vehicle per day generated a tax decision. The amount was decided at 7 p.m. (after the time the tax was charged) and published by 7 a.m. the following morning. The congestion tax was to be paid retroactively and within 5 days. (From 1 June 2006 this was extended to 14 days). All this imposed major demands on the payment system. There were more direct debit transactions per day for the congestion tax alone than for any other single item in the Swedish direct debit system.

Development of OCR technology

The system had been procured on the premise that onboard units would have to be widely used if the system were to function optimally, as the method of identifying number plates through photo images was originally not sufficiently sophisticated for this purpose. However, a legal investigation revealed that, according to provisions in Swedish law, the number plates on all vehicles would have to be photographed. A decision was then made to photograph both the front and rear plates. In consequence, OCR technology for identification purposes was further refined and the data system enhanced. Through this, it can be said that the congestion tax helped promote technological development, and with the conditions that prevail in Stockholm, it is now both technically and economically possible to base all identification on photo images alone. Even on the days of inclement weather during the trial period, 99% of the vehicles that crossed a control point could be identified, either automatically or manually, from the photo images of the number plates.

Widespread media interest – information project

Well-functioning communication with both individuals and the media was a factor that was vital to the success of the project. Information was therefore organised into a project on its own. A great amount of time and energy was devoted to media contact and other public relations measures.

The overall objective was that everyone entering the Stockholm inner city zone from 3 January 2006 in a Swedish registered vehicle would be aware of the congestion tax and know that it had to be paid without having been prompted to do so. The focus was on providing objective, factual and neutral information devoid of any value judgements. It was to be perfectly clear that the SRA did not take a stand for or against the congestion tax itself, and that its sole mission was to implement a well-functioning system while assisting vehicle owners through providing as much information as possible.

Reaching out with information to all vehicle owners in the country was a challenge. The communication effort was intensive, including information meetings, advertising, information material, media contact and an extensive customer services organisation. Letters were sent to all owners of Swedish vehicles with information about the approaching trial implementation of a congestion tax. Even the media was active. During the autumn of 2005 and spring of 2006 the congestion tax was one of the most widely covered issues in the country. Major effort was therefore expended on responding to the demands of the press and radio/television.

Stable start

The start button was pushed on the cold winter morning of 3 January 2006. Public interest on the part of politicians, stakeholder organisations and innumerable media representatives was huge. When the system was launched it was not without nervous apprehension. In record time a project without precedent in Sweden had crossed the finishing line. But we had no idea how the system would work in real-life. Had the information about not having to stop at the control points been assimilated? How would vehicle owners ever be able to manage the short payment deadline – five days? Many questions hung in the air, and there were deep misgivings about how traffic would work, particularly from the media where total chaos had been predicted.

It was found that the information had in fact made an impact. There was no chaos. Quite the contrary. Many experienced a tangible improvement in the traffic situation. Neither did the short payment deadline cause any problem – 96% of all tax decisions were paid on time. Nonetheless, the system was not entirely free from teething problems. Many had ordered an electronic onboard unit for direct debit payment, but the distribution had not initially worked properly. Further, some vehicle owners had encountered difficulties in connection with the direct debit contract; either they had not filled in their personal identity number correctly or they had tried to connect the direct debit to the wrong kind of bank account. At first, this resulted in numerous calls, which were handled by both Customer Services staff as well as individual project members. However, these problems were gradually solved, and a couple of months later an Internet-based payment service was also developed to simplify payments.

High capacity

The highest number of control point passages occurred on 2 June 2006, when 394,000 were registered. This still left a comfortable margin to the peak capacity of the system, 500,000 passages a day. On three occasions during the trial period there was a serious risk of losing all the registered passages, but the built-in redundancy in the system made it possible to bypass problems, both then and at other times.

The technology and processes proved to be operationally reliable and dependable. This, combined with successful information activities, meant that the number of people employed for the trial implementation could gradually be reduced. In July 2006 there were about ten people in project management, about 100 engaged by the contractor and less than 200 working at Customer Services, compared to the late autumn of 2005 when the peak figure of 700 was reached, 450 of whom were working at Customer Services.

Effects observed

The Stockholm City was responsible for the evaluation of the full-scale trial. Extensive analysis was conducted and Stockholm now has a unique store of traffic data and knowledge about cause-effect relationships in traffic.

It was found that the objectives were essentially achieved. The following observations could be made in comparison to corresponding periods the previous year (2005):

• traffic to and from the Stockholm inner city decreased by 20–25%
• the time spent in traffic queues in and around the inner city was reduced by 30–50%
• 100,000 fewer passages to/from the inner city
• vehicle emissions decreased by 14% in the inner city.

What lies ahead?

As the trial period came to an end on 31 July 2006, the SRA closed down the technical system pending a decision by the Government as to whether or not there would be a congestion tax in the future.

In conjunction with the general election on 17 September 2006, the City of Stockholm and a number of other surrounding municipalities held a referendum on whether or not to reinstate environmental charges/a congestion tax. In Stockholm City, 51% voted in favour of the congestion tax while 46% were against it. In the other municipalities the overall result was a 60% no vote and 40% yes (blank votes not counted).

On 10 November 2006, the Government appointed a mediator to investigate the conditions for an overall traffic solution for the Stockholm region. In conjunction with this, the Government stated its intention to reinstate the congestion tax during the first half of 2007 to help finance future investments in the road network in the Stockholm region.

Much time has been spent this autumn on analysing and evaluating the full-scale trial. During the entire trial period, improvements were consistently being made in the system. This involved over 200 modifications, including developing a method to simplify administration procedures for companies with large vehicle fleets. But there is still room for improvement. Procedures for paying the tax could be simplified. It could also be made more convenient for both individuals and companies to access the information in the system. Wider use of e-identification could, for example, facilitate access to passage and payment information for vehicle owners.

In many respects, the SRA broke new ground during the development and operation of the congestion tax system. This has provided invaluable experience. In the face of a new start, there is a firm foundation to stand on, both as regards technology and knowledge. It is now possible to build a future system that, in addition to high quality, can also be designed so that it is easy to handle for the individual driver. Apart from the system contributing to a more efficient use of the road network, it will also be able to contribute to the future development of the transport system in Stockholm.