Content uploaded by Piotr S Olszewski
Author content
All content in this area was uploaded by Piotr S Olszewski on Apr 06, 2014
Content may be subject to copyright.
Abstract The example of Singapore shows that rapid urban and economic growth
does not have to bring traffic congestion and pollution. Singapore has chosen to
restrain car traffic demand due to its limited land supply. Transport policy based on
balanced development of road and transit infrastructure and restraint of traffic has
been consistently implemented for the past 30 years. Combined with land use
planning, it resulted in a modern transport system, which is free from major con-
gestion and provides users with different travel alternatives. As the economic growth
caused a substantial increase in demand for cars, several pricing policies were
introduced with the aim of restraining car ownership and usage. Growth of the
vehicle population is now controlled and potentially congested roads are subject to
road pricing. These measures help to keep the roads free from major congestion,
maintain car share of work trips below 25% and keep the transport energy usage
low. Although Singapore conditions are in many aspects unique, its travel demand
experience can provide useful lessons for other rapidly growing cities in Asia.
Keywords Singapore motorisation Æ Traffic restraint Æ Travel demand management Æ
Modal split Æ Road pricing Æ Vehicle Quota System
Introduction
Rapid pace of urbanisation and economic development in Asia leads to accelerated
motorisation and produces the unwelcome effects of urban traffic congestion and
major environmental problems. It is estimated that in the last decade of the 20th
century the number of city dwellers in East and South Asia increased by 300 million
(41%) and the number of motor vehicles by 20 million (35%) (World Bank 2006).
P. S. Olszewski (&)
School of Civil and Environmental Engineering, Nanyang Technological University, 50
Nanyang Avenue, Singapore 639798, Singapore
e-mail: colsze@ntu.edu.sg
123
Transportation (2007) 34:319–335
DOI 10.1007/s11116-007-9115-y
ORIGINAL PAPER
Singapore motorisation restraint and its implications
on travel behaviour and urban sustainability
Piotr S. Olszewski
Published online: 13 March 2007
Springer Science+Business Media B.V. 2007
Many governments have not been able to handle such a massive growth and as a
result the quality of life in some cities is actually deteriorating. In comparison,
Singapore’s industrialisation and economic growth have not led to serious traffic
congestion and pollution problems because of its successful approach to transpor-
tation system development and travel demand management.
Singapore is a fully urbanised island state with a population of 4.35 million and a
land area of just under 700 km
2
. With a per capita GDP of US$25,800
1
in 2005, it is
classified as a Newly Industrialised Economy (NIE). As evident from Table 1, over
the last three decades, the real Gross Domestic Product (GDP) has been increasing
at an average annual rate of around 7%. Such sustained growth has increased the
demand for both passenger and freight transport as well as car ownership aspirations
of the population. However, with a population density of 6,222 persons per km
2
and
about 12% of its scarce land already being used for road infrastructure, Singapore
has very limited scope for road expansion. This situation resulted in the government
having to adopt tough demand management policies to restrain the growth of car
ownership and usage to reasonable levels. These fiscal measures have generated
substantial revenue, which is partly used for financing new road and public transport
projects.
Singapore’s innovative transportation policies have been closely studied by
transport economists and researchers, as evidenced by a large body of literature on
the subject (e.g. Cochrane et al. 1986; Behbehani et al. 1988; Spencer and Madhavan
1989; Olszewski and Turner 1993; Phang 1993; Willoughby 2001; May 2004). The
objectives of this paper are: to briefly review the development of vehicle restraint
methods over the last 30 years; to investigate the effect of these measures on travel
behaviour, motor vehicle fleet and transport energy consumption and finally to
suggest how the Singapore experience could be useful for other cities. The paper is
organised as follows: Sect. 2 introduces the Singapore transportation system; Sects. 3
and 4 outline the car ownership and usage restraint policies; Sects. 5 and 6 present
some evidence on the effects of these policies and finally Sects. 7 and 8 discuss the
relevance of restraint to other cities and provide conclusions.
Singapore transportation system
Emergence of a planned city
Before achieving self-rule in 1959, Singapore like many Asian cities was character-
ised by overcrowding, poor housing conditions, high levels of unemployment and
poverty. After independence in 1965, the government embarked on an ambitious
programme of industrialisation and construction of public housing. A strict urban
planning and development control system was put in place by means of Planning Act
legislation. Recognizing the need to safeguard land for public housing and infra-
structure development, in 1966 the government enacted the Land Acquisition Act,
by which extensive areas of land have been acquired from private owners (Dale
1999).
In 1971 the first concept plan for long-term urban development was drawn up.
The 25-year development strategy incorporated a comprehensive land use and
1
As of December 2006 US$1 = $1.55 Singapore dollars.
320 Transportation (2007) 34:319–335
123
Table 1 Socioeconomic and transport indicators
1970 1975 1980 1985 1990 1995 2000 2005
Area km
2
586.4 596.8 617.8 620.5 633 647.5 682.7 699.4
Population (‘000) 2074.5 2262.6 2413.9 2736 3047.1 3525.6 4017.7 4351.4
GDP (current prices) S$ million 5805 13373 25091 38923 66175 119470 159840 194360
GDP (2000 prices)
a
S$ million 16230 25565 38844 52439 77158 117745 159840 193453
GDP annual growth
b
% n.a. 9.51% 8.37% 6.19% 8.03% 8.82% 6.30% 3.89%
GDP per capita
a
S$ 7823 11299 16092 19166 25322 33397 39784 44458
US$ 4548 6569 9356 11143 14722 19417 23130 25847
Cars 106900 142045 152574 221279 271174 342245 386780 432827
Motorcycles 78900 83145 118345 127564 122525 129587 131937 139434
Goods vehicles 25600 41363 78020 107146 115536 137913 134756 139098
Other vehicles 6400 13825 22402 30771 33117 32384 39334 43633
Total vehicles 217800 280378 371341 486760 542352 642129 692807 754992
Car ownership Per 1000 51.5 62.8 63.2 80.9 89.0 97.1 96.3 99.5
Road length km 1938 2167 2356 2645 2882 2972 3100 3234
Veh. per km of road Veh/km 112.4 129.4 157.6 184.0 188.2 216.1 223.5 233.5
a
In constant year 2000 prices
b
During the previous 5-year period
Sources: Department of Statistics, Singapore (1974–2006)
Transportation (2007) 34:319–335 321
123
transport plan that envisaged construction of a system of expressways, comple-
mented by a network of arterial roads. In addition to building a string of high-density
housing estates (called New Towns), the plan proposed a Mass Rapid Transit system
to connect the New Towns with the central area and industrial estates. In order to
reduce demand for travel, all the New Towns were planned with industrial zones and
several regional commercial centres were proposed to reduce job concentration in
the CBD. While the concept plan has been revised and updated in 1991 and 2001, the
urban development vision of the early 1970s has been largely realised and in many
aspects exceeded. In today’s modern garden city some 86% of the population lives in
public housing apartments, 90% of which are owned by their occupants (Yuen
2004).
2
The present road network comprising 150 km of expressways, 594 km of major
arterial roads and 2,490 km of other roads serves all forms of transport including an
efficient network of bus services. A 109-km Mass Rapid Transit (MRT) system of
three radial lines was built over the last 20 years and is being extended. Three Light
Rapid Transit (LRT) lines serve as feeders to the MRT. There are continuous efforts
to improve the public transport system by integrating the MRT lines with feeder
modes and providing better accessibility (May 2004; Luk and Olszewski 2003).
As shown in Table 1, the present car ownership is still under 100 cars/1,000
people, which is very low for an industrialised country. In fact, it is lower than in
countries with one third of Singapore’s per capita income (Willoughby 2001).
Figure 1 shows the growth of population, GDP and vehicle fleet, relative to 1975.
While in the last 30 years the real GDP has increased about 7.5 times, the car
population has only tripled and its present growth rate is barely keeping up with the
population growth.
Transport policy
By the early 1970s, the motor vehicle population was near 300,000 and traffic con-
gestion was becoming a serious problem in the city. The first comprehensive
transport study warned that congestion would soon lead to urban strangulation.
Thus, an integrated Land Transportation Strategy was adopted in mid 1970s, with
the following aims:
(1) Integrating land use and transport planning,
(2) Increasing road capacity as far as practicable,
(3) Managing demand for travel, especially that of the private car, and
(4) Providing an efficient public transport system.
The above strategy has been consistently pursued over the last three decades. In
1996 Land Transport Authority (LTA) was established with a mission to create a
world-class transport system to serve the Singapore population (LTA 1996). LTA is in
charge of both roads and public transport planning as well as traffic management and
safety, which gives it a clear institutional advantage in dealing with potentially con-
flicting priorities.
The rapid economic growth in Singapore has naturally increased people’s desire
to own a car. As shown in Table 1, around 755,000 motor vehicles are currently
2
Discussion of Singapore urban planning and development control is beyond the scope of this
paper. Interested readers are referred to: Dale (1999) and Yuen (2004).
322 Transportation (2007) 34:319–335
123
registered resulting in a ‘‘statistical’’ density of 233 vehicles per km of road. With
about 12% of its land area devoted to road transport uses, there is little scope for
major increases in road capacity. Therefore, some control on vehicle ownership and
usage was deemed necessary to prevent traffic gridlock. Over the years several fiscal,
administrative and traffic management measures have been introduced. Ownership
restraints include a variety of taxes and registration fees as well as a direct quota
imposed on the number of new vehicle registrations. Usage restraints include high
gasoline taxes and parking fees, an Off-Peak Car Scheme and road pricing. These
measures will be discussed briefly in the following sections.
Car ownership restraint policies
Fiscal measures
Since Singapore has no domestic car industry, a logical first step was to impose a high
customs duty on cars. This has varied from 45% of the Open Market Value (OMV)
of a vehicle
3
to the current figure of 20%. In addition to a one time registration fee
(now $120), owners pay an Additional Registration Fee (ARF), which at its peak in
1980s was equal to 175% of the OMV but after the introduction of the vehicle quota
system has been reduced to 110% of OMV. Annual road tax depends on engine
capacity and is also substantial: for example, it now amounts to $500 for a 1,000 cc
car and $3050 for a 3,000 cc car.
The above taxes on car ownership drastically reduced the growth in vehicle
numbers. Between 1973 and 1982, the car population grew by 35% as compared to
between 100% and 150% forecast by OECD (Behbehani et al. 1988). However,
except for a brief economic slowdown in 1985–1986, the growth picked up again
despite high car prices. Phang (1993) explained this phenomenon by showing that
demand for cars was more income elastic than price elastic. In the late 1980s the
growth far exceeded 3% per year, which was the rate estimated to be sustainable,
taking into account the expansion of the road network. This necessitated the
0
100
200
300
400
500
600
700
800
1975 1980 1985 1990 1995 2000 2005
Year
Index (1975=100)
Population
GDP
Vehicles
Cars
Fig. 1 Growth in Singapore’s population, real GDP and the vehicle population. Sources:
Department of Statistics, Singapore (1975–2006)
3
The OMV is the value declared by local distributors which includes factory and shipping cost of the
vehicle.
Transportation (2007) 34:319–335 323
123
introduction of more radical restraint measures. Following a public debate on the
issue in 1989, the government decided to put a cap on all new motor vehicle regis-
trations, in the form of a Vehicle Quota System (VQS).
Vehicle quota system
Under the VQS, prospective owners of new vehicles (other than scheduled buses
and emergency vehicles) need to successfully bid for a Certificate of Entitlement
(COE), a fixed number of which are auctioned to the highest bidders each month.
The COE for a new vehicle is valid for 10 years. Its validity can be extended for
another 5 years or 10 years by paying the ‘‘prevailing premium’’ price, calculated as
the 3-month moving average.
Since its introduction in 1990, the VQS has undergone several modifications.
Initially, there were four COE categories for cars, based on engine capacity:
<1,000 cc, 1,001–1,600 cc, 1,601–2,000 cc and >2,000 cc as well as separate categories
for motorcycles and goods vehicles. The combined monthly quota for the four car
categories was 2,070 vehicles. In addition, some 470 COEs were available in the
‘‘open’’ category, which allows one to buy any type of vehicle but practically is used
for car purchases. In 1999 the four car categories were combined into two: up to
1,600 cc and above 1,600 cc. Initially, the certificates were transferable, but following
public concern over speculation, they were made non-transferable.
The Quota System lets market forces determine the premium prices. Figure 2
shows the changes in monthly quota and COE premiums over the last 16 years for
cars under 1,600 cc. The values shown are for December of each year. For the first
8 years the premiums followed an increasing trend (with a dip in 1995) while the
quota available for all the car categories combined (including the ‘‘open’’ category)
remained almost constant at around 2,400 per month. The COE premium for cars
under 1,600 cc reached an all-time-high of $62,200 in July 1997. Together with the
other fiscal measures (ARF, customs duty) this has raised the price of a new med-
ium-size car in Singapore to an exorbitant level of over six times its OMV.
Since their peak in 1997 the COE premiums have eased considerably (Fig. 2) and
now stand at between $9,605 and $12,889, depending on the car category. This
constitutes between 10% and 20% of the on-the-road price of a new car. The main
reasons for the declining trend of COE prices are:
(1) The quota numbers have been increased substantially due to ‘‘recycling’’ of
COEs of de-registered vehicles
4
(Fig. 2).
(2) The introduction of Internet open bidding system in 2002 made the whole
bidding process more transparent. The auction now takes 4 days and bidders
can monitor the current COE price on-line and revise their bids accordingly
(LTA 2006).
It remains to be seen whether the current spell of relatively low car prices can be
sustained. In 2005 a total of 88,076 cars were de-registered while the number of new
registrations was 109,167, giving a net increase in the car population of 21,000. The
huge number of de-registrations of cars purchased during the period of high COE
4
The number of vehicles taken off the road and any unused COEs are added to the next year’s
quota.
324 Transportation (2007) 34:319–335
123
prices (1996–2000) is due to the fact that owners can get a refund for the ‘‘unused’’
part of the COE within the first 10 years of its validity.
Car usage restraint policies
Area licensing scheme
In addition to curbing car ownership, several innovative measures aimed at reducing
the usage of cars were introduced. The most well known and studied of these—the
coupon-based Area Licensing Scheme (ALS)—operated in the central area from
1975 until 1998 when it was transformed into an electronic scheme. It was the only
example of a full-scale urban road pricing system specifically designed to reduce
peak period traffic flows. Over the years the ALS has expanded in terms of types of
vehicles affected, time periods and the area covered. Several studies discussed its
historical development and impacts, e.g.: Holland and Watson (1978), Behbehani
et al. (1988), Hau (1992), McCarthy and Tay (1993), Menon et al. (1993), Polak
et al. (1994), Olszewski et al. (1996).
Under the scheme, motorists entering the Central Business District (CBD) during
peak period (7:30–10:15 AM) had to buy a licence, initially priced at $3 and later at
$5. The pricing levels were set and adjusted experimentally, not based on theoretical
models. Although some doubts were raised about the level of charging (McCarthy
and Tay 1993), the ALS was a simple and effective travel demand management
scheme and showed an extremely attractive rate of return (Hau 1992).
Electronic road pricing
As the coupon system became too cumbersome (many types of coupons for dif-
ferent vehicles), it was upgraded to a fully automated Electronic Road Pricing
(ERP) system in 1998. The planning, implementation and the initial experience
0
1000
2000
3000
4000
5000
6000
19
90
1
99
1
199
2
1993
1
99
4
199
5
19
96
19
97
1
99
8
19
99
20
00
2
00
1
20
02
2
00
3
2
00
4
200
5
2006
Monthly quota
0
10000
20000
30000
40000
50000
60000
COE premium $
Quota
Premium
Fig. 2 Trends in monthly quota and COE premium for cars under 1,600 cc (values for December
each year). Source: LTA (2006)
Transportation (2007) 34:319–335 325
123
with the ERP are described by Phang and Toh (1997), Luk (1999) and Menon
(2000).
Under the ERP, the charge is deducted automatically from a pre-paid smart card
when a vehicle passes under an ERP gantry. At present there are 48 gantries: 30
form a cordon around the CBD, 13 others are located on selected expressway seg-
ments and five on radial arterial roads. The charges at the CBD cordon apply on
working days during the daytime hours: 7.30 AM–7.00 PM, but there are periods
with zero charges (e.g. 10:00 AM–12:00 noon). On other roads charges apply during
the morning peak period (7:30 AM–9:30 AM) and in some locations in the evening
(5.30 PM–8:00 PM).
Variable pricing system
The ERP system makes it possible to vary charges by location, time of day and
vehicle type, so as to relate them to the actual level of congestion. To emphasise the
link between road pricing and congestion, the rates for different types of vehicles are
set to be approximately proportional to their passenger car equivalent (PCE) values
(Menon 2000). A method called ‘‘shoulder pricing’’ is used, which involves
increasing the rate in steps every half an hour before the peak and decreasing it after
the peak. For example, the current basic rates for cars on Central Expressway are:
• 7:30 AM–8:00 AM $1.50
• 8:00 AM–8:30 AM $2.00
• 8:30 AM–9:00 AM $3.00
• 9:00 AM–9:30 AM $1.00
The system is actually even more complicated because to ensure smooth traffic
flow a five-minute transition rate is used whenever the rate increment is $1 or more.
Thus, in the above example, $0.80 is charged from 7:30 to 7:35 and $2.50 from 8:30 to
8:35.
Rate adjustments
Traffic flow is quite sensitive to the ERP even though the charges are relatively
low—the maximum rate for cars is $3.00 on expressways and $2.50 to enter the
CBD, which is comparable to a one-hour parking fee in the city (Olszewski
and Xie 2005). Traffic speeds are monitored by LTA and the rates are revised
every 3 months to maintain speeds within bands corresponding to the desired
level of service. The target speed range was defined as 45–65 km/h for express-
ways and 20–30 km/h for arterials and roads crossing the CBD cordon (Menon
2000). If the observed speed falls below the lower threshold, the rates will
be increased, whereas if the speed exceeds the upper threshold, the rates will be
reduced.
Other measures
In 1991, the ‘‘Off-Peak Car’’ scheme (originally called the ‘‘Weekend Car Scheme’’)
was introduced to enable more people to own cars without contributing to
326 Transportation (2007) 34:319–335
123
congestion (Yee and Menon 1994). Newly purchased cars could be registered as
off-peak cars, with an upfront saving of $17,000 on reduced registration tax fees and
COE.
The off-peak car scheme has initially proved to be popular: by the end of 1991
there were 4,800 such cars but subsequently their numbers fluctuated. At present the
number is 12,900—about 3% of all cars. An ‘‘off-peak car’’ is permitted to use the
road network from 7:00 PM to 7:00 AM on weekdays, after 3:00 PM on Saturdays
and all day on Sundays and public holidays. There are severe penalties for using an
off-peak car outside these hours but daily use permits can be purchased for $20. To
facilitate enforcement off-peak cars have red number plates.
Effects on travel behaviour
Traffic reduction in the central area
The immediate effect of ALS introduction in 1975 was the reduction of car traffic
entering the city centre during the morning peak by over 70%. In 1992, the car
volume was still at 54% of the pre-1975 level, despite the value of ALS fee expressed
as a percentage of daily wage declining in the meantime by 62% (Polak et al. 1994).
With many years of observations it was possible to model the relationship
between car flow entering the CBD, the level of ALS charges and other explanatory
variables. A time series partial-adjust model for assessing the effects of ALS was first
proposed by Polak et al. (1994) and subsequently recalibrated by Olszewski and Xie
(2002) for years 1974–1993. The following best-fit equation was obtained (t-values in
brackets):
Ln Q
t
ðÞ
¼ 5:142þ0:294Ln Q
t1
ðÞ0:214Ln CALS
t
ðÞ0:224Ln CP
t
ðÞþ0:432Ln GDP
t
ðÞþe
t
5:4
ðÞ
2:2
ðÞ
2:9
ðÞ
2:2
ðÞ
5:0
ðÞ
where: Q
t
= volume of cars entering CBD during the ALS period (7:30–10:15) in
year t; CALS
t
= cost of the daily ALS licence for cars in year t, adjusted for infla-
tion; CP
t
= average parking cost in year t, estimated by combining parking fees at
public and private car parks, adjusted for inflation; GDP
t
= Singapore GDP in year t,
adjusted using the national income deflator; and e
t
= a disturbance term.
Given the small sample size (n = 20), the levels of significance of the coefficients
and the overall model fit (R
2
= 0.97) are satisfactory. Several other explanatory
variables were tried, e.g.: parking capacity, MRT operation and overall driving cost
but did not contribute to explaining the variation of traffic volume. Only car own-
ership had a significant effect on Q but because it was strongly correlated with GDP,
it could not be used in the same equation.
The advantage of the above model is that each coefficient b has a direct inter-
pretation as the short-term elasticity of flow with respect to explanatory variable.
The long-term elasticity is obtained as: b
CALS
/(1 b
Q
t1
). The value of the coef-
ficient on lagged flow indicates that there are significant delays in the response of the
peak travel demand to price changes, with around 70% of the full effect of a price
change becoming apparent in the first year.
Transportation (2007) 34:319–335 327
123
The values of elasticity of flow versus ALS fee estimated from the model are:
• Short-term –0.214
• Long-term –0.303
The long-term elasticity is about 42% higher than the short-term value. The
elasticity with respect to parking fees is of a similar magnitude. The sensitivity of
traffic volume entering CBD to price changes is a reflection of the flexibility in trip
scheduling, availability of alternative modes of travel serving the area and alterna-
tive routes for through traffic.
When ERP was started in the CBD area, it was even more effective than ALS
because of its ‘‘per pass’’ charging principle (as compared to a daily licence before)
and time variable rate structure. The rates had to be revised down one month after
the introduction because traffic in the CBD declined too much (Menon 2000). In fact
LTA’s revenue from road pricing decreased after the introduction of ERP, which
proves that flexible systems with time-variable pricing can achieve their objective of
alleviating congestion at a lower cost to users than systems with a flat toll.
Modal split
The restraint policies outlined above had a significant effect on modal split of travel
in Singapore. Before the ALS was implemented in 1975, the share of private car
drivers among commuters entering the CBD was 48%. This share dropped to about
29% immediately after the ALS introduction (Watson and Holland 1978) and has
decreased further to 16% in 1983 (Behbehani et al. 1988).
Figure 3 shows the changes in island-wide modal shares of work trips, based on
Singapore census data. The share of private car trips (including both driver and
passenger) among motorised trips has increased from about 16% in 1980 to 25% in
2000. This increase is approximately proportional to the corresponding increase in
car ownership from 63 to 96 cars per 1,000 population (Table 1). The latest travel
surveys indicate that the overall share of private transport (i.e. including motorcycles
and trucks) among motorised trips for all purposes has increased from 37% in 1997
to 42% in 2004 (LTA 2005). During that period the car ownership level practically
did not change, indicating that besides motorisation the level of income also influ-
ences modal split, especially for non-work trips.
International comparisons made by Kenworthy and Laube (1999) show Singapore
with one of the lowest car shares for work trips among the 46 cities studied. In 1990
lower car shares were only observed in Hong Kong and Seoul, which had even lower
car ownership and much higher urban population density. On the other hand, cities
like Manila and Jakarta, both with car ownership lower than Singapore, had com-
muter trip car shares between 1.5 and 2 times higher. This can be attributed to the
high quality of Singapore public transport system, which benefits from motorisation
restraint in two ways:
(1) There is a large pool of captive transit passengers, in addition to those
‘‘nudged’’ out of their cars, which helps to maintain high frequency of services
and to make transit operations profitable.
(2) Congestion-free roads allow buses to operate at a higher speed, which makes
them more attractive.
328 Transportation (2007) 34:319–335
123
Travel surveys (LTA 2005) reveal interesting trends in modal split. It seems that
high usage of vehicles other than cars is a way of compensating for the high cost of
cars, which puts them beyond reach of many families. These other vehicles are:
• Motorcycles—the number of motorcycles is still slowly growing (Table 1) which
is unusual in a high-income country. Motorcycles provide mobility mainly for the
lower income workers.
• Taxis—there are 870,000 taxi trips per day (11% of all trips). Taxis are efficient,
relatively inexpensive and plentiful and are mostly used for non-commuting trips.
• Chartered buses—these provide an almost-door-to-door commuter service for
people working outside the CBD. They have 8% share in work trips.
• Goods vehicles—pick-ups and even trucks are also used for passenger travel, not
necessarily related to work.
Peak spreading
Another effect of the Area licensing scheme was to spread the morning peak of city-
bound traffic, as many motorists took advantage of free-entry periods before 7:30
and after 10:15 AM. This was evident from traffic counts as mini traffic peaks ap-
peared before and after the ALS hours. When lower-priced licences were introduced
for the mid-day period in 1993, the high peak at 8:00–9:00 AM has reappeared
because commuters found that it was no longer worthwhile to retime their trips
(Olszewski et al. 1996).
When motorists were asked at that time how they responded to newly introduced
road pricing, their first choice was to change the trip timing, followed by the change
of route and then change of mode. Similar results were obtained in stated preference
surveys conducted recently by LTA (2005)—84% of motorists felt that changing
departure time was feasible.
When ERP was started with its time-variable ‘‘shoulder pricing’’ method, the
peak spreading effect was even more visible from traffic observations on express-
ways and arterial roads. Demand elasticity with respect to ERP charges in the
morning peak was estimated as –0.106 for the CBD cordon and –0.195 for
expressways—almost twice as high (Olszewski and Xie 2005). This can be explained
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1980 1990 1995 2000 2005
Other
Public transport
Other private
Car
Fig. 3 Modal split for motorised trips to work based on Singapore census data. Source: Department
of Statistics, Singapore (1990–2000)
Transportation (2007) 34:319–335 329
123
by the fact that at the expressway sites charges applied for 2 h only, so both trip
postponing and bringing it forward resulted in savings whereas at the CBD cordon
the only way to avoid fees was to enter before 7:30 AM.
Travel survey findings confirm that the ERP contributes to peak spreading not
only locally but on an island-wide scale. Household interview surveys conducted in
1997 and 2004 show that the percentage of car trips starting between 6 AM and
7 AM has increased from 4% to 10%, while those starting between 7 AM and 8 AM
decreased from 13% to 10% (LTA 2005).
Other effects
Changes in car fleet composition
The Vehicle Quota System had a profound effect on the composition of private car
fleet in terms of vehicle age and engine capacity. Even before VQS there were strong
incentives for scrapping cars at 10 years as part of ARF is then refunded. The annual
road tax is also getting progressively higher for older cars. However, scrapping even
younger cars can also be attractive because the ‘‘unused’’ part of COE is refunded.
Such operations make economic sense when the COE prices are falling, as happens
now (Fig. 2). Hence, over the last few years thousands of 4- to 9-year-old cars,
bought when the COEs were high, have been scrapped and exchanged for new
models at little or no extra cost to their owners.
Figure 4 shows the cumulative distribution of cars in Singapore by age. In 1990
the distribution between the ages 1 and 9 years was more or less uniform and the
median age of a car was 6.4 years. In 2005 the median age has been reduced to just
2.2 years and only 14% of cars are more than 5 years old! Such a rejuvenation of the
car fleet is of course a positive development. New cars of the same engine size are
more fuel efficient, less polluting (since 2001 they have to meet Euro II emission
standards) and safer to drive.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
123 45678910 11 12 13 14 15 15+
Age x (years)
Percentage of cars less than x years old
1990
1995
2005
Fig. 4 Changes in cumulative age distribution of cars in Singapore. Source: LTA (2006)
330 Transportation (2007) 34:319–335
123
Figure 5 shows the changes in car distribution by engine capacity. It seems that
the small car category (<1,000 cc), which comprised 16.4% in 1980 has now practi-
cally disappeared. On the other hand, the share of cars with over 1,600 cc engines
has grown from 19% to 39%. The fastest growing category now is cars over 2,000 cc.
The main reason for these shifts is that the COE premium is now practically the
same for all cars and constitutes a higher percentage of the final price of small cars
than of big cars. This makes big cars relatively less expensive as compared to their
OMVs. The resulting price distortion is bigger when the COE prices are high. It has
been suggested that to remove this distortion, COE premium could be defined as a
percentage of OMV rather than in dollar terms.
The shift towards higher engine capacity is a cause of some concern. Bigger cars
are bound to consume more fuel, which can reverse the trend towards better energy
efficiency. They are also heavier and while probably safer for their occupants in an
accident, they can potentially do more damage to vulnerable road users (motorcy-
clists, cyclists, pedestrians).
Energy use in transportation
What are the implications of the Singapore motorisation restraint on long-term
transport system sustainability? One way of measuring the efficiency of a transport
system is to calculate the energy input required to keep it moving. Ang and Tan
(2001) showed that per capita energy consumption in road transport in Singapore is
about 42% of that in France and Germany. However, such country-to-country
comparisons can be misleading as in big countries most of the travel mileage occurs
outside cities.
It would be more appropriate to compare Singapore with other cities. This can be
done using data assembled in the invaluable international sourcebook on automobile
dependence (Kenworthy and Laube 1999). The energy use can be expressed
in energy per capita or per unit of the regional income and these two indicators
were calculated for different cities. Figure 6 shows the comparison of energy use
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1980 1985 1990 1995 2000 2005
>3000
2001-3000
1601-2000
1001-1600
<1000
Fig. 5 Changes in the composition of car fleet by engine capacity. Source: LTA (2006)
Transportation (2007) 34:319–335 331
123
indicators in 1990 for Singapore and selected cities. One can clearly distinguish four
groups of cities:
(1) US cities have by far the highest rates of energy use per capita.
(2) Canadian and Australian cities have energy use about half that of US cities,
with European cities doing even better.
(3) A group of developing Asian cities have relatively low energy use per capita
but very high per unit income.
(4) The ‘‘wealthy’’ Asian cities to which Singapore belongs have the lowest
transport energy use rates both on per capita and per unit income basis.
Overall, it seems that Singapore transport system is energy efficient, although
even lower energy use rates are possible as shown by examples of Tokyo and Hong
Kong. Compared to Singapore, both these cities have even higher urban density and
better-developed mass transit systems. However, unlike Singapore, both experience
congestion, which self-regulates the modal split but imposes heavy delay cost on
commuters.
Road transport energy use can also be analysed based on the local petrol and
diesel consumption figures published by the Customs and Excise Department (1980–
1998). The volume of these fuels released each year for the local market are mul-
tiplied by their average energy equivalents and then divided by the population.
Figure 7 shows the trends in road transport energy intensities calculated on per
capita and per vehicle basis (this includes all types of road vehicles, as shown in
Table 1). The effect of the Vehicle Quota System on energy use seems to be clear:
while between 1980 and 1990 per capita energy use has increased by 32%, the
increase was only 2% during the first 7 years under the VQS.
Relevance of the Singapore experience
Several authors (e.g. Willoughby 2001; May 2004) discuss the question how relevant
the Singapore experience is to other countries. Some may argue that Singapore
geographical and political conditions are so unique that its positive experience in
Singapore
Kuala Lumpur
Frankfurt
Tokyo
Bangkok
Jakarta
Manila
Paris
Toronto
Sydney
San Francisco
Hong Kong
Seoul
Washington
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0 10000 20000 30000 40000 50000 60000 70000
Energy per capita (MJ/cap)
Energy per unit GRP (MJ/$)
Fig. 6 Comparison of land transportation energy use for selected cities. Source: calculated based on
Kenworthy and Laube (1999)
332 Transportation (2007) 34:319–335
123
Travel Demand Management is not transferable and of limited value to other
countries. It is certainly true that Singapore is unique in being an island city, which
makes it self-contained and unlike other cities, which are intertwined with wider
metropolitan regions. It also has the advantage of a one-tier government, making
policy setting and implementation more efficient. The government had the benefit of
controlling most of the land resources, which facilitated urban planning and infra-
structure construction. It also had the strong political resolve and courage to
introduce restrictions on private cars for the sake of common good of containing
congestion. However, even if some of these conditions are not present elsewhere,
Singapore’s experience with both low-technology (ALS) and high-tech (ERP)
solutions can provide important lessons to other cities undergoing rapid economic
development. Several elements of the Singapore transport strategy seem to be
universally applicable:
• It is very important to formulate an integrated transport policy at an early stage,
with clear objectives encompassing all transport modes.
• Motorists are more likely to accept restrictions on car usage if offered attractive
travel alternatives, like a high standard rapid public transport and inexpensive
taxis.
• Low technology cordon pricing systems like the ALS should be considered first in
developing countries. They are effective, easy to implement and can generate
substantial revenue. However, they require strict enforcement and may not work
in cities, which do not have a well-defined CBD.
• In electronic road pricing, flexible systems with time-variable charges and ‘‘free
periods’’ are preferable to systems with a flat toll. They can achieve the desired
effect of alleviating congestion with lower overall cost to users.
• Periodic adjustments of ERP rates based on traffic speeds (or level of service)
provide a self-regulating mechanism in the system. Such a goal-oriented
experimental approach to demand management seems to work very well in
practice.
• Restraint policies need to be closely monitored as they may have unexpected
effects such as increase in car engine capacity, use of trucks for passenger
transport, etc.
0
3,000
6,000
9,000
12,000
15,000
19801982198419861988 1990199219941996
Year
Energy intensity per capita
0
10,000
20,000
30,000
40,000
50,000
60,000
70,000
80,000
90,000
100,000
Ener
gy
intensit
y
per vehicle
MJ/cap
MJ/veh
VQS
introduced
Fig. 7 Trends in road energy intensity in Singapore. Source: based on Customs & Excise Dept.
(1980–1997)
Transportation (2007) 34:319–335 333
123
Besides reducing both car ownership and use, the Singapore restraint polices has
been generating a substantial amount of revenue for the government. The figures
published in 1993 show that the revenue collected from the motor vehicle taxes, fees
and tolls amounted to over $3.12 billion (ROV 1994), which made a sizeable con-
tribution to the consolidated revenue fund. Unlike in some other countries, the
Singapore road revenues are not earmarked for transportation related expenditure.
Willoughby (2001) estimated that in the years 1965–1993 the annual road revenues
were at least three to four times higher than road related expenditures.
Conclusions
The example of Singapore shows how motorisation policies in each country are a
product of geographical, economic and political factors. Singapore has chosen to
restrain car usage and ownership mainly due to its very limited land supply. The
transport policy based on balanced development of road and transit infrastructure
and restraint of private car traffic has been consistently implemented for the past
30 years. Combined with land use planning, it resulted in a modern transport system,
which is free from major congestion, energy efficient and provides users with dif-
ferent travel alternatives.
The experience with fiscal measures of restraining car ownership shows that they
are effective only up to a certain point. Under conditions of rapidly rising incomes,
car ownership increases quickly despite high prices. This is because demand for cars
is more income elastic than price elastic. In the end, the growth of car population can
only be arrested by more radical measures such as the car quota system.
However, strong economic growth and rising affluence put a constant upward
pressure on car ownership and usage. While the present use of cars for commuting
(25%) is low, their share of trips for other purposes is increasing. The peak spreading
effect of road pricing means that the potential for congestion to occur spreads more
widely both in time and space. From the point of view of energy efficiency, the
gradual shift of the car fleet towards higher engine capacity is a worrying trend. Given
over 30 years of experience in traffic demand management, it seems that Singapore
has all the necessary system tools in place to manage these challenges efficiently.
Acknowledgments The author would like to thank the Singapore Land Transport Authority for
their assistance and data. Special thanks are due to Mr. A.P.G. Menon who has helped to obtain data
and provided useful comments on the draft manuscript. The author is also grateful to the anonymous
reviewers whose useful comments helped to improve the paper.
References
Ang, B.W., Tan, K.C.: Why Singapore’s land transportation energy consumption is relatively low?
Nat. Resour. Forum 25, 135–146 (2001)
Behbehani, R., Pendakur, V.S., Armstrong-Wright, A.: A case study of Singapore. In: Cities and
Transport. OECD, Paris (1988)
Cochrane, S.R., Ferguson, J.D., Al-Mufti, M.A.: Vehicle growth trends in Singapore. Traffic Eng.
Control 27(3), 122–124 (1986)
Customs and Excise Department, Singapore: Annual Report (various years) (1980–1998)
Dale, O.J.: Urban Planning in Singapore: The Transformation of a City. Oxford University Press,
Oxford (1999)
Department of Statistics, Singapore: Yearbook of Statistics. Singapore (1974–2006)
334 Transportation (2007) 34:319–335
123
Department of Statistics, Singapore: Singapore Census of Population – Statistical Release 4 –
Geographic Distribution and Travel. Singapore (1990–2000)
Hau, T.: Congestion Charging Mechanisms for Roads: An Evaluation of Current Practice. Policy
Research Working Papers. World Bank (1992)
Kenworthy, J.R., Laube, F.B.: An International Sourcebook of Automobile Dependence in Cities
1960–1990. University Press of Colorado, Boulder, Colorado (1999)
LTA: A World Class Land Transport System. Land Transport Authority White Paper presented to
the Parliament of the Republic of Singapore (1996)
LTA: Findings of 2004 Household Interview and Stated Preference Surveys, LTA Planning and
Policy Division Seminar, 28 September 2005
LTA: Facts and Figures – Land Transport Authority website: http://www.lta.gov.sg/corp_info/in-
dex_corp_facts.htm. Cited on 20 Dec 2006
Luk, J.Y.K.: Electronic road pricing in Singapore. Road Trans. Res. 8(4), 28–40 (1999)
Luk, J.Y.K., Olszewski, P.: Integrated public transport in Singapore and Hong Kong. Road Trans.
Res. 12(4), 41–51 (2003)
May, A.D.: Singapore: thedevelopmentof a world class transport system. Trans. Rev.24(1), 79–101 (2004)
McCarthy, P., Tay R.: Pricing road congestion: recent evidence from Singapore. Policy Stud. J. 21(2),
296–308 (1993)
Menon, A.P.G.: ERP in Singapore – a perspective one year on. Traffic Eng. Control 41(2), 40–45 (2000)
Menon, A.P.G., Lam, S.H., Fan, H.S.L.: Singapore’s road pricing system: its past, present and future.
ITE J. 63(12), 44–48 (1993)
Olszewski, P., Turner, D.J.: New methods of controlling vehicle ownership and usage in Singapore.
Transportation 20(4), 355–371 (1993)
Olszewski, P., Xie, L.: Traffic demand elasticity with respect to road pricing – some evidence from
Singapore. In: Proc. International Conference on Seamless and Sustainable Transport. Singa-
pore, Nov 2002 (2002)
Olszewski, P., Xie, L.: Modelling the effects of road pricing on traffic in Singapore. Transport. Res. A
39A, 755–772 (2005)
Olszewski, P., Lam, S.H., Wong, Y.D.: Effects of the recent changes in the Singapore road pricing
scheme. In: Hensher, D.A., King, J. (eds.) World Transport Research, vol 3, pp. 373–384.
Pergamon (1996)
Phang, S.Y.: Singapore’s motor vehicle policy: review of recent changes and a suggested alternative.
Transport. Res. A 27A(4), 329–336 (1993)
Phang, S.Y., Toh, R.: From manual to electronic road congestion pricing: the Singapore experience
and experiment. Transport. Res. E 33(2), 97–106 (1997)
Polak, J., Olszewski, P., Wong, Y.D.: Evidence of the long term effects of the Singapore Area
Licensing Scheme on travel behaviour. In: Proc. 7th International Conference on Travel
Behaviour Research, vol 1, pp. 319–330, Santiago, Chile (1994)
ROV: Singapore Registry of Vehicles Annual Report 1994 (1994)
Singstat: website of the Singapore Department of Statistics: http://www.singstat.gov.sg/keystats/an-
nual/indicators.html (cited on 2 Sept 2006)
Spencer, A.H., Madhavan, S.: The car in Southeast Asia. Transport. Res. A 23(A), 425–437 (1989)
Watson, P.L., Holland, E.P.: Relieving Traffic Congestion: The Singapore Area License Scheme.
World Bank Staff Working Paper 281. World Bank (1978)
Willoughby, C.: Singapore’s motorization policies 1960–2000. Trans. Policy 8, 125–139 (2001)
World Bank: 2005 World Development Indicators online: http://devdata.worldbank.org/wdi2005/
Home.htm (cited on 20 August 2006)
Yee, J., Menon, A.P.G.: Weekend cars. Traffic Eng. Control 35(4), 250–252 (1994)
Yuen B.: Planning Singapore growth for better living. In: Freire, M., Yuen, B. (eds.) Enhancing
Urban Management in East Asia. Ashgate, Aldershot (2004)
Biography
Dr Piotr Olszewski obtained his M.Sc. and Ph.D. degrees in highway engineering from the Warsaw
University of Technology in Poland. After working in an environmental planning institute in
Warsaw, in 1981 he joined the Nanyang Technological University in Singapore, where he is now
Associate Professor in the School of Civil and Environmental Engineering. Dr Olszewski teaches
courses in the areas of transportation engineering, urban planning and logistics. His research
interests include modelling of transportation systems, travel behaviour and sustainable transport
policies.
Transportation (2007) 34:319–335 335
123
