Empirical evidence of the impact of early urban charging schemes comes
from Singapore (Holland and Watson, 1982) and Norway (Larsen, 1988). Furthermore,
a good deal of our understanding of the impacts of road pricing come from
modelling studies, such as Richards et al (1996).
A Singapore Area Licence Scheme
Context
Singapore introduced an Area Licensing Scheme in 1975 to reduce congestion
in the city centre. Drivers had to purchase licences for a day or a month
to allow them to enter the defined area between 0730 and 1015. The initial
charge was S$3; this was raised to S$4 in 1976. Vehicles with four or
more occupants were exempt. Police at the 22 entry points observed vehicles
and recorded those without licences; they were then fined. Subsequent
modifications involved extensions to the evening peak, the working day
and Saturdays, to a set of charging points on expressways, and to all
cars however many occupants they had. Different charges were levied for
different types of vehicle. A major study was conducted in 1975 (Holland
and Watson, 1978); the evidence below comes from this.
Impacts on demand
Pattern
Most affected drivers continued to travel to the city
centre; there were no recorded reductions in numbers or length of
journey or destination.
Mode
19% of drivers travelling to the city
centre switched to bus; 17% switched to car sharing to take advantage
of the exemption for cars with four or more people.
Timing
22% of drivers travelling to the city centre switched
to travelling before or after the charged period, resulting in some
increases in congestion then.
Route
Many drivers travelling through the city centre diverted
to the ring road, resulting in some increases in congestion on that
route; a few changed mode or time of travel.
Surprisingly there were very few changes in evening peak travel; it appeared
that people continued to use their cars to leave the city centre in the
evening peak, even though they had made changes in the morning.
Impacts on supply
Only minor adjustments were made to the road network, and no delays were
caused at the entry points. However, drivers did need to spend time purchasing
licences.
Contribution to objectives
Objective
Scale of contribution
Comment
The reduction of 44% of traffic
entering the centre resulted in an increase in speeds of 22% in the
centre and 10% on the approaches. Speeds fell by up to 20% on the
inner ring road. No comprehensive cost-benefit analysis was conducted,
but it is clear that there were substantial reductions in congestion
costs and increases in benefits. It is possible that charges were,
in practice, too high and that greater benefits could have been obtained
by a smaller reduction in car-use and a
The scheme’s impact was primarily
on the commercial and business centre of the city. Residential streets
were therefore little affected, but there was an improvement in conditions
in shopping streets.
This was not a key objective
and no attempt was made to assess impacts. However, it can be expected
that it was improved in the city centre in the morning peak, with
some minor deterioration outside the controlled periods and on the
inner ring road.
The study attempted to identify
gainers and losers, but found little evidence of differential impacts,
and suggested that the range of alternatives offered reduced the scale
of any inequities. A subsequent study, however, suggested that poorer
car-drivers had been adversely affected (Wilson, 1988).
This was not a key objective
and no attempt was made to assess impacts. However, it can be expected
that it was improved in the city centre in the morning peak, with
some minor deterioration outside the controlled periods and on the
inner ring road.
An attempt was made ten years
later to identify impacts on the urban economy and business relocation.
None were found; they had been dwarfed by the expansion of Singapore’s
economic base. Businesses were very supportive of the scheme.
In 1975 prices costs were approximately
S£60M, the operating costs S£1m PA, and revenues approximately S£7M
PA. Although revenue-raising was never an objective, the scheme raised
substantial net revenues; operating costs being only 12% of revenues.
= Weakest
possible positive contribution,
= strongest
possible positive contribution
= Weakest
possible negative contribution
= strongest
possible negative contribution
=
No contribution
Singapore Electronic Road Pricing
Context
In 1998 the Area Licensing Scheme was replaced by an Electronic Road Pricing
Scheme. 97% of the 700,000 vehicles in Singapore were fitted with on board
units, in which smart cards were inserted. Gantries at the Area Licensing
Scheme entry points and expressway charging points were equipped to identify,
interrogate, charge and, if necessary for enforcement, photograph, all
vehicles passing. Charges are now levied per crossing rather than per
day, and vary by time of day and vehicle type. Charges are revised quarterly
to maintain speeds at between 20 km/h and 30 km/h in the city centre,
and 45 km/h and 60 km/h on expressways. As a result charges are lower
than with the Area Licensing Scheme for much of the day and have been
waived on Saturdays. Early results are now available (Menon, 2000).
Impacts on demand
Pattern
There is no evidence as yet of any impact on these.
Mode
It may be that there have been some changes in mode,
given the further reduction in traffic levels, which are 15% below
those under the Area Licensing Scheme.
Timing
Drivers appear to be very sensitive to differences in
change by time of day. In addition the number of
multiple entries was substantially reduced
Route
There is no evidence that further changes in route have
occurred, although there are still some congestion problems on the
boundary route.
Impacts on supply
Capacity has been maintained, and the delays involved in purchasing licences
removed.
Contribution to objectives
Objective
Scale of contribution
Comment
No detailed analysis has been
conducted but it seems probable that, by targeting charge levels to
achieve optimal speeds, efficiency has increased.
As with area licensing, there
was little impact on residential streets
This was not a key objective.
There will have been some limited further reduction in environmental
impact through the further reduction in traffic
No assessment of equity impacts
has been made, but those making occasional journeys off-peak and on
Saturdays will have benefited, while costs will have increased for
those making multiple journeys.
This was not a key objective.
There will have been a limited further reduction in accidents through
the further reduction in traffic.
It seems very unlikely that
there will have been significant impacts on the urban economy.
The cost of introducing electronic
road pricing was substantial, at S£200M. Revenues are, in practice,
lower than with area licensing, at S£8M PA, but revenue generation
is not an objective.
= Weakest
possible positive contribution,
= strongest
possible positive contribution
= Weakest
possible negative contribution
= strongest
possible negative contribution
=
No contribution
Toll rings in Bergen, Oslo and Trondheim
Context
In 1986 Bergen, Norway's second largest city, was the first city in Europe
to introduce a toll ring (or cordon) charging system. It was introduced
with the objective of raising the finances required to accelerate the
implementation of a wide-ranging programme of transport investment. The
system charges all vehicles (other than buses in regular service) a flat
fee for entering the city's central business district and operates between
6AM and 10PM Monday-Friday. Toll rings were subsequently also introduced
in Oslo and Trondheim. As in Bergen, the main objective is to raise revenue
so charges are set according to revenue goals, though both Oslo and Trondheim
use electronic toll collection and in Trondheim tolls are differenciated
by time of day.
Impact on demand
Pattern
In Bergen, whilst it was expected that the ring would
decrease traffic volumes by around 3%, other than a slight decrease
in the beginning there has been an average annual traffic growth of
2-3%. In Trondheim there have been significant impacts on peak hour
traffic levels, with reductions of 10% immediately following the introduction
of the differenciated charges, increasing over time to 17%.
Mode
It is likely that there has been some change in modal
share over the period, though this will have been in part due to the
investment in public transport, using the revenues from the toll rings,
which has taken place over the period.
Timing
Reductions in peak traffic in Trondheim resulting from
the differenciated charge were outweighed by increases in traffic
in off peak periods.
Route
In Bergen, there are no natural detours so there has
been little impact on route choice.
Impact on supply
The toll rings themselves have not affected supply, though the finance
they have generated has enabled a series of major transport projects to
be implemented.
Contribution to objectives
Objective
Scale of contribution
Comment
No detailed analysis has been
conducted but it seems probable that, in Trondheim at least, efficiency
will have been increased via the targeting of the charge on peak period
traffic.
The schemes are focused on the
central business districts. Residential streets are therefore likely
to have been little affected, though there may have been an improvement
in conditions in the central shopping streets.
This was not a key objective.
There is likely to have been some reduction in environmental impact
through the reduction of traffic.
No assessment of equity impacts
has been made, but those making occasional journeys outside the charging
periods, eg on Saturdays, will have benefited while costs will have
been imposed on those travelling during the charging periods, eg during
the peak periods in Trondheim. Opinion polls originally indicated
that approximately two thirds of Bergen’s population were against
the toll ring, though it has now been widely accepted by the majority.
The change in opinion is thought to be connected with the visible
improvements in the local transport network benefiting everyone and
is despite relatively high levels of tax on motoring.
This was not a key objective.
There may have been a limited reduction in accidents through the deterance
of car travel, though this is likely to have been offset by increase
in road capacity.
In Bergen, Saturday was deliberately
kept free from tolls in order to support the city’s shops. However,
the effect on city centre shopping is not known.
In Bergen the initial investment
to establish the ring was approximately NOK 15M. Annual income has
been higher than expected and is approximately NOK 70M. Of this, NOK
50M is spent on roads, NOK 14M is taken up in operating costs and
NOK 7M is stored in a fund (the use of which attracts great political
debate).
Financial and technical support from the US Federal Highways Administration
(FHWA) has been used to facilitate the implementation of three pricing
projects which have come to be known as 'value pricing' schemes. In addition,
the FHWA has supported a comprehensive study of an additional, privately
operated pricing project.
The first of the three value pricing projects was implemented in 1996
along the 13km high-occupancy vehicle (HOV) section of Interstate 15 (I15)
in San Diego. Access to the HOV lane was extended to include a limited
number of solo drivers who were able to pay for a monthly pass for use
of the HOV lane during peak periods. The number of available passes rose
gradually from 500 to 900 over the first year and the cost of the pass
rose from $50 to $70. Then, in March 1998, the pricing scheme was upgraded
to become an automated, dynamic system. Congestion in the HOV lane is
monitered and forms the basis of the toll levels. Tolls are set with the
aim of maintaining 'free-flow' conditions in the HOV lane and range between
$0.50 and $4. They vary as often as every 6 minutes and the current toll
level is displayed on a real-time sign post in advance of the entry to
the lane. Tolls are deducted using transponders and over-head readers
A similar scheme has since been introduced in Houston, Texas. In addition,
higher peak fees on existing toll roads and bridges have been introduced
in Lee County.
Impacts on demand
Pattern
Traffic volumes along the section of I15 increased 'moderately
(by approximately 6%)', comprising a significant (48%) increase in
volumes in the HOV lane (as paying users took up the spare capacity
which existed in the HOV lane prior to the introduction of value pricing)
and a slight decrease in volumes in the other I15 lanes. The overall
increase was smaller than was observed in the 'control' corridor.
Mode
The main impact would appear to have been an increase
in car-pooling (ride-sharing), particularly under the monthly pass
system. Some diversion to express bus services is indicated, though
this would appear to have been a relatively minor impact.
Timing
The main impact has been to divert trips from the peak;
both from the middle of the peak to the 'shoulder' of the peak and
from peak to off-peak.
Route
No evidence reported
Impacts on supply:
By freeing up the spare capacity in the HOV lane for use by non-HOV users,
the I15 value pricing scheme has, in effect, increased the overall capacity
of the road. In addition, the revenues it has generated have provided
funding for a new express bus service along the corridor.
Contribution to objectives
Objective
Scale of contribution
Comment
No detailed analysis has been
conducted but it seems probable that, in Trondheim at least, efficiency
will have been increased via the targeting of the charge on peak period
traffic.
The schemes are focused on the
central business districts. Residential streets are therefore likely
to have been little affected, though there may have been an improvement
in conditions in the central shopping streets.
This was not a key objective.
There is likely to have been some reduction in environmental impact
through the reduction of traffic.
No assessment of equity impacts
has been made, but those making occasional journeys outside the charging
periods, eg on Saturdays, will have benefited while costs will have
been imposed on those travelling during the charging periods, eg during
the peak periods in Trondheim. Opinion polls originally indicated
that approximately two thirds of Bergen’s population were against
the toll ring, though it has now been widely accepted by the majority.
The change in opinion is thought to be connected with the visible
improvements in the local transport network benefiting everyone and
is despite relatively high levels of tax on motoring.
This was not a key objective.
There may have been a limited reduction in accidents through the deterance
of car travel, though this is likely to have been offset by increase
in road capacity.
In Bergen, Saturday was deliberately
kept free from tolls in order to support the city’s shops. However,
the effect on city centre shopping is not known.
In Bergen the initial investment
to establish the ring was approximately NOK 15M. Annual income has
been higher than expected and is approximately NOK 70M. Of this, NOK
50M is spent on roads, NOK 14M is taken up in operating costs and
NOK 7M is stored in a fund (the use of which attracts great political
debate).
= Weakest
possible positive contribution,
= strongest
possible positive contribution
= Weakest
possible negative contribution
= strongest
possible negative contribution
=
No contribution
London congestion charging study
Context
A major desk study in London in 1992-5 provided one of the most comprehensive
studies of cordon charging options and technologies. The simplest was
a single cordon around Central London; the most complex a system of three
concentric cordons and four radial screenlines, with the outer cordon
surrounding Inner London. Most aspects were studied in detail (Richards,
et al, 1996); the transport impacts are described below (May, et al, 1996).
Impacts on demand
Pattern
Destination -a small proportion of the reduction in
traffic in Central and Inner London was explained by changes to destinations
in Inner and Outer London.
Mode
The majority of journeys no longer made by car to Central
London were predicted to switch to bus (which benefits from reduced
congestion) with a further substantial switch to rail.
Timing
Where changes differed by time of day, significant numbers
were predicted to switch away from peak period travel.
Route
Where possible, changes in route are the most common
alternative. With the single cordon much through traffic was predicted
to divert to the Inner Ring Road, reducing the benefits outside Central
London. With the more complex scheme, the screenlines reduced diversions,
but traffic levels still increased on the orbital outside the outer
cordon.
Freight impacts
Surveys suggested that freight vehicles were largely
insensitive to charges, except where they could reroute to avoid them.
Overall traffic
With the single cordon at the highest charge tested
(£10 throughout the working day) car traffic in Central London
was predicted to fall by 40% and all traffic by almost 25%. Two thirds
of these reductions were achieved with a £4 charge.
Impacts on supply
No effects on capacity were predicted.
Contribution to objectives
Objective
Scale of contribution
Comment
For the Central London cordon,
economic benefits were at their highest at a charge of £6 per day.
They fell slightly at higher charges as charging began to discourage
drivers whose benefits exceeded their marginal costs. At a charge
of £8 per day, speeds were predicted to be 32% higher in Central London,
but only 4% higher in Inner London. The same charge with £2 charges
on each of the other cordons in the 3 cordon system, increased speeds
by 26% in Central London and 10% in Inner London.
Traffic levels were predicted
to fall in residential streets within the charged area, but to increase
in those just outside the boundary. The overall reduction in traffic
would enable some re-routing of traffic away fropm residential streets,
at the cost of a small reduction in efficiency benefits.
The Central London cordon with
an £8 charge reduced carbon monoxide in Central and Inner London by
12%, and carbon dioxide by 6%. A £4 charge achieved reductions of
approximately two thirds of these levels. The 3 cordon scheme had
around double the impact.
Impacts were assessed for 3
income groups. The highest income group suffered the greatest loss
of benefits with both schemes. Conversely, the lowest income group
experienced small benefits from both, largely because they are more
likely to use buses and benefit from reduced congestion.
The Central London cordon, at
an £8 charge, reduced London-wide accidents by around 3%; the 3 cordon
scheme again had around double the impact.
Neither scheme was predicted
to have significant impacts on the distribution of land-use. A £4
charge on the Central London cordon would increase employment in Central
London by 2% and reduce households there by under 1%, but with increases
in the number of high income households. A 3 cordon scheme at double
the charge would increase employment in Central London by 2% and reduce
it in Inner London by 1%, again with very small changes in residential
locations.
At £8 per day, the Central London
cordon was estimated to generate gross revenues of £420M PA; the 3
cordon system £740M. Capital costs were estimated as £85M to £140M
and £240M to £335M; operating costs £55M PA and £155M PA. Both options
thus more than covered their costs within the first year.
= Weakest
possible positive contribution,
= strongest
possible positive contribution
= Weakest
possible negative contribution
= strongest
possible negative contribution
=
No contribution
Demonstration projects
A number of demonstration projects throughout Europe have been undertaken
in recent years, the results of which are summarised below.
The Leicester Environmental Road Tolling Scheme (LERTS)
involved monitoring the use of a small scale electronic tolling scheme
by 100 volunteers who were given money which they could either use to
pay the toll or which they could save by using the alternative services
provided (park and ride and bus priority). Interim results indicate:
Patronage on the park and ride service grew from 1300 to 3150 passengers
per week, illustrating a latent demand for the service even in the absence
of the toll scheme;
Journey time savings for bus of 25% (24 to 18 minutes);
Journey time savings for cars of 3% (14.5 to 14 minutes);
Transfer from car to park and ride amongst commuters during the tolling
period
ranged between 16% and 32%, with transfer to bus and to ride-sharing
each in the range 1-4%;
31.9% of people made no change;
25.2% changed route;
12.8% changed departure time;
14.9% used park and ride;
2.1% used other Bus;
The Stuttgart 'MobilPass'
field trial involved 350 users in a test of different pricing strategies
using an electronic fee collection system. Simulations using the MobilPass
data have shown that, in principle, corridor tolls on peri-urban roads
accessing Stuttgart will result in greater transport and emission reductions
than a cordon toll does. This is because, in the case of a cordon, the
toll paid includes any subsequent trips made within the cordon, where
as in a corridor the effect of pricing schemes will be noticed directly.
The Athens cordon pricing experiment, undertaken as part of the TRANSPRICE
project, involved two groups of 50 selected users, each made up of current
car users who commute to central Athens from northern Athens suburbs.
The users were allocated a nominal budget and presented with a choice
of either driving to central Athens in their car, as before, but having
to use their budget to pay to cross a cordon or switching to a metro-based
park and ride service with the opportunity to save their budget. The results
indicate:
65% of users made no change;
24.0% of users switched to the park and ride service;
5.5% of users switched to other public transport;
5.5% of users switched to other modes (e.g. ride
sharing, taxi).
The Bristol ELGAR
demonstration project, undertaken as part of the CONCERT
project, involved an electronic road pricing scheme whereby 116 volunteers
could be rewarded for switching to modes other than private car for their
journey into the city centre, a package of public transport improvements
and a series of variable message signs giving information on pollution
levels. Findings indicate that
The majority of users continued to use their car as before;
15.1% of journeys switched from private cars to other modes (though
it was found that most of the reduction resulted from a few of the participants
switching nearly all of their trips from private car);
over half of the respondents to the follow-up questionnaire thought
that the revenue from any future road pricing scheme should be spent
on improving public transport; and
three-quarters of those switching to public transport expressed a
positive opinion about this experience.
Gaps and weaknesses
The limited range of real world evidence on urban road charging represents,
in itself, a weakness of the body of knowledge on this policy instrument.
However, a number of cities are now quite close to implementing urban
road charging, e.g. London and Rome, which should provide further evidence
within the next few years.
Furthermore, there are gaps in our knowledge of the full range of impacts
which have been experienced in those few situations where urban road charging
has been introduced, particularly with regard to the Norwegian examples.
We will continue our efforts to gather evidence on these examples and
invite users of this knowledge base to send us any evidence that they
are aware of.
Text edited at the Institute for
Transport Studies, University of Leeds, Leeds LS2 9JT
