ELSEVIER

AUTOMOBILE D E P E N D E N C E IN
CITIES: A N INTERNATIONAL
COMPARISON OF U R B A N
TRANSPORT A N D L A N D USE
PATTERNS WITH IMPLICATIONS
FOR SUSTAINABILITY

Jeffrey R. Kenworthy and Felix B. Laube

Institute ]'or Science and Technology Policy
Murdoch University, Perth, Australia

A comparison o f global cities over the period 1980 to 1990 reveals large
differences in automobile dependence with implications for the future sus-
tainability o f cities in different countries. Cities in the United States have acceler-
ated dramatically in their dependence on the automobile between 1980 and
1990, whereas urban transit has remained at very low levels, though with some
modest success stories. US cities in 1990 have levels o f per capita auto use that
are some two times higher than Australian cities, the next most auto-dependent
cities in the study, and have transit use rates that are the lowest in the world
by a significant margin. Toronto, European cities, and modern Asian cities on
the other hand have experienced much smaller increases in automobile use
with some very large increases in transit use; their car use remains much
lower again than in US and Australian cities. Developing Asian cities such as
Bangkok, while still at very low levels o f auto use and high transit use by
international standards, are experiencing rapid increases in automobile depen-
dence, which is not explainable by rising incomes alone. This study explores
some o f the underlying land use, transport, and economic reasons for these
different transport patterns. It briefly reviews what the sustainability agenda
means for transport and land use patterns in cities and suggests a suite o f targets
or goals for sustainability by which cities might measure their current directions
and plans. © 1996 Elsevier Science lnc.

Address requestsfor reprints to: Jeffrey R. Kenworthy, Institute for Science and Technology
Policy, Murdoch University, Perth, Western Australia 6150, Australia.

ENVIRON IMPACT ASSESS REV 1996;16:279-308

© 1996 Elsevier Science Inc. 0195-9255/96/$15.00
655 Avenue of the Americas, New York, NY 10010 PII S0195-9255(96)00023-6
280 J E F F R E Y R. K E N W O R T H Y A N D FELIX B. L A U B E

Introduction
Cities around the world are subject to increasing levels of environmental
impact from dependence on the automobile. In the highly auto-dependent
cities of the US and Australia, this is manifested in problems such as urban
sprawl and its destruction of prime farming land and natural landscapes,
photochemical smog that can be primarily attributed to auto emissions,
and many other traffic-related impacts such as noise, neighbourhood sever-
ance, and visual intrusion (Newman, Kenworthy, and Vintila 1992). Many
social problems, such as isolation in suburbs, lack of community, and diffi-
culties for those without an automobile (the young, old, poor, and people
with disabilities), can also be largely attributed to auto-based develop-
ment patterns.
In the rapidly developing cities of Asia such Bangkok, Jakarta, Surabaya,
Manila, Kuala Lumpur, and now the Chinese cities, the automobile poses
a serious environmental threat. Bangkok is already suffering extremely
high levels of impact from its auto use, as well as intense human impacts
from the gridlock that grips the city for many hours every day (Kenwor-
thy 1995).
On top of the more local impacts of the automobile, the global dimension
should not be forgotten. Perhaps the two most pressing issues in this regard
are the oil problem and the greenhouse problem (Newman 1991). Global
oil production is set to peak early next century and then enter into a period
of declining annual production at a time when there is likely to be increasing
demands for this resource, especially from China and other developing
regions and their rapidly expanding fleets of motor vehicles (Fleay 1994;
Campbell 1991). Greenhouse gases from automobiles are an increasing
contributor to global CO2 and are very difficult to reduce in the face of
accelerating automobile dependence.
In light of these issues, this study examines the patterns of automobile
dependence in cities around the world and explores some of the underlying
reasons for the differences and relationships between the various land use
and transport factors. It provides 1980 and 1990 data on the land use and
transport patterns of 39 cities from the United States, Canada, Australia,
Europe, and the developed and developing parts of Asia, summarized
according to these regional groupings. The study discusses briefly the sig-
nificance of each factor and the salient trends between 1980 and 1990.
Finally, it looks briefly at the need to work toward sustainability in urban
transport and suggests a series of specific annual goals for land use, private
transport, public transport, and nonmotorized modes that emerge out of
the results.

Land Use and Transport Characteristics in International

Cities, 1980-1990
Table 1 provides land use, transport infrastructure, private transport, and
public transport data for 1980 and 1990 in a global sample of cities, excluding
C A R S IN CITIES: IMPLICATIONS F O R SUSTA1NABILITY 281

the developing Asian cities. These data represent a preliminary insight into
the complete update of Cities and Automobile Dependence: An International
Sourcebook by Newman and Kenworthy (1989). The original book con-
tained data on 32 cities for 1960, 1970 and 1980. The present study updates
the 1980 data with 1990 data, where available, and adds several more cities
to the perspective. Although not all cities in the sample are included for
every variable in each regional grouping (see footnote to Table 1), there
are enough data in each case to gain a strong indication of the trends in
some of the key variables describing land use and transport patterns.
Table 2 shows the same 1990 data (where available) for a set of five
developing Asian cities which form part of the update of Cities and Automo-
bile Dependence. These data are referred to briefly in the discussion that
follows and are included in the statistical correlations.
The following sections outline the differences between the cities, first in
land use, then in transport infrastructure, followed by a discussion of the
private and public transport trends that flow from these patterns.

L a n d Use Characteristics
URBAN DENSITY. One of the key indicators of urban land use as it affects
transport is urban population density, particularly the density of the whole
city, but also density in different parts of the city such as the CBD and
inner area. Density is a key determinant of transport through its effect on
a number of factors. High densities tend to be associated with lower average
trip distances for all modes, improved public transport through higher
potential patronage around each stop and in particular, enhanced viability
of walking and cycling. Dense, compact urban environments also tend to
be more mixed in their land use, rather than heavily zoned and segregated
as in low density suburban areas. This further enhances reduced trip lengths
and the viability of transit and nonmotorized modes. Thus, density is a key
factor in the future of cities as regards sustainability. Most of the literature
on new trends in urban development to reduce auto dependence and im-
prove physical and social amenity stresses the density issue (e.g., Katz 1994;
Calthorpe 1993; ACT City Services 1994; Holtzclaw 1994).
The data reveal that overall metropolitan densities in the United States
and Australia have remained very low and basically static between 1980
and 1990 (13 to 14 per ha). Metropolitan Toronto's density is some three
times higher than US and Australian city densities (41 per ha), having risen
marginally over the 10 years. New development in European cities has
spread out to some extent, though it is still occurring at densities significantly
higher than in the United States and Australia (Newman and Kenworthy
1991). Combined with falling household occupancies, densities in the Euro-
pean sample of cities dropped some 7% from 54 to 50 persons per ha,
though they are still some 3.5 times higher than in the United States and
Australia. In the wealthy Asian cities (Singapore, Tokyo, and Hong Kong),
282 JEFFREY R. K E N W O R T H Y A N D FELIX B. L A U B E

. oO0
'U
~<.oo
o
<
~ ~ o

Z ~. ~ ~
e,i ,.e
"U

c-I
.o Q~----.>
o~

o~
~ .=
~ ~: 0

= o
o~
r4
o o <.~
0
OO 0

oo
~ re5
Ng

"2.

O ~
o g
~ t--, u.-, ~

"d

~ ~ o

°~
o N ;

b~ "~
~ o ~
".~-~ ~

~'~ ~.~ ~ ~ ~
o o
N ~.~ ~- ~ ~ ' ~ ~ ~" ~-.~.
o.~ E
.~s o2a
,~.- o
g _~- ~~
.~ ~ ,., o . _ ~. ~ ~
M
._~ .~ ~
,1
~ ..- ~ < < < - ~ ~oc o~
 CARS IN CITIES: IMPLICATIONS FOR SUSTA1NABILITY 283

oa~

2_~ =~ -=

oF. ~
,,t3 , 0

•~--00 ~

~.~ ~ ~ E ~

u ~ ~ ~ .~ ~ ~ ~.~

-~ = o ' -_~ ~ _ ~~ ~
~=.~
~ ~ ~ ~ ~-~ -~.~

. .=_~_ ~ ;==

~0 . ~~ . : -= o E~ - ~~ ~ ~ -S K

o,.

~ ~.-'=' o
z.~: ~
_
>~ ~ 2
.-~ ~ ~ ~ - . ~ . ~ ~u = = ~ ~
.~_'~-

~~ E
~ _ E~ ~ ~ o . . . 0~ . .o 0 ~~
284 JEFFREY R. KENWORTHY AND FELIX B. LAUBE

TABLE 2. Land Use and Transport System Characteristics of Developing

Asian Cities, 1990
Kuala
Bangkok Jakarta Lumpur Manila Surabaya

Land use characteristics

U r b a n density (people/ha) 144 171 59 198 177
C B D density (people/ha) 325 235 123 445 314
Inner area density (people/ha) 237 267 69 372 265
Transport Infrastructure
Length of road/capita (metres) 0.6 0.5 1.5 0.6 0.3
C B D parking spaces/1000 jobs 338 ? 305 ? 9
Private transport
Passenger cars/1000 people 153 75 184 66 40
Total vehicles/1000 people 348 205 433 86 211
A n n u a l car kms/capita 1514 523 2920 860 250
A n n u a l car pass. kms/capita 4002 1380 6007 1978 1167
% workers using foot and bike 10 22 17 19 31
Public transport
A n n u a l service kms/capita 110 62 104 248 55
A n n u a l transit trips/capita 423 235 269 448 129
A n n u a l transit pass. km/capita 2313 1285 2235 1906 513
Public/Private transport balance
% of total pass. k m on transit 36.6 48.2 27.1 49.1 30.5

Source: Barter et al. (1994), Barter and Kenworthy (1995), Poboon et al. (1994), plus unpublished data
from Barter (1995) and Guia (1995).
The developing Asian cities, except for Manila, have large numbers of motor cycles which add significantly
to the picture of vehicle ownership. This is reflected in the much higher total vehicle ownership figures in
developing Asian cities, compared to cars alone.
The annual passenger kms per capita in cars in these cities also includes the passenger kms on motor cycles
as these are a significant part of the private transport picture, unlike in other cities where they tend to be
very minor. Exclusion of motor cycle passenger kms would overstate the proportion of total passenger
kms performed by transit.
Some data for the developing Asian cities may be subject to minor revisions.

densities rose marginally (2.5%) to 164 per ha as development continued

to occur at very high densities compared with the western world (12 times
US and Australian urban densities and over three times European densi-
ties). In the five developing Asian cities in Table 2, urban densities are also
high (average 150 per ha), with only Kuala Lumpur atypically low for this
group (59 per ha).
Figures 1 and 2 demonstrate the importance of urban density in ex-
plaining annual per capita auto and transit use (data points are those cities
in our study where 1990 data are available for these variables, including
the developing Asian cities in Table 2). Auto use is strongly correlated
with density (r 2 = 0.799) as is transit use (r 2 = 0.669). The different levels of
auto and transit use per capita shown in Table i are discussed in detail later.

CBD DENSITY.The population living in the central area of a city can be a

significant factor in shaping transport patterns by bringing more people
CARS IN CITIES:IMPLICATIONSFOR SUSTAINABILITY 285

15000

10000-

Annual car kms

per capita

5000-

%
0 I I I

0 100 200 300

Urban density (persons per ha)

F I G U R E 1. Urban density versus car use per person in world cities, 1990.

closer to what is still the m a j o r center o f e m p l o y m e n t , business, social,

cultural, and e n t e r t a i n m e n t facilities in m o s t cities. This increases p e o p l e s '
ability to walk, cycle, or use transit, with a high potential to live a lifestyle
free o f the car. F o r example, data f r o m N e w Y o r k show that residents of
M a n h a t t a n living at 251 persons per ha with h o u s e h o l d incomes greater
than $US75,000 o w n only 0.81 vehicles p e r household, whereas those on
the same income in o u t e r areas of the region at 13 persons p e r ha o w n

600
[]
[]
500-
On [] O ~
400 S
Annual transit
trips per capita 300 -

200 -
y °
m
100- 1
q
0 I I f
0 100 200 300 400

Urban density (persons per ha)

F I G U R E 2. Urban density versus annual transit trips per person in world cities,
1990.
286 J E F F R E Y R. K E N W O R T H Y A N D F E L I X B. L A U B E

2.61 vehicles per household. This general pattern occurs at every income
level (New York Metropolitan Transportation Commission 1995). Nowlan
and Stewart (1992) show that the dramatic growth in downtown housing
in Toronto in the last 20 years has resulted in a measurable reduction in
inbound commuting trips (120 trips for every extra 100 dwellings).
Many large US cities, especially New York, San Francisco, and Boston
have, on the whole, sizable concentrations of people still residing in the
downtown area. The data from this sample of eleven cities suggest that on
average, US downtown areas have been holding their own or perhaps
increasing marginally in downtown population. For example, Houston has
tripled its CBD population density in the 10 years from a very low six
persons per ha to 18 per ha, Los Angeles has increased by 11% to 28 per
ha, San Francisco by 24% to 111 per ha, Chicago almost 100% to 30 per
ha, New York 4% to 227 per ha, and even San Diego has been developing
a downtown residential character (up by 46% to 27 per ha).
Australian cities show positive trends in this direction too (six cities up
by an average 13%), though in absolute terms they do not have the high
density concentrations of downtown population characteristic of the historic
cores of US cities (Australian cities in 1990 averaged only 14 persons per
ha in their core, compared with US cities with 54 per ha). Most Australian
cities such as Melbourne and Perth are addressing this issue with concerted
efforts to build up central area population (Murphy 1994; Perth Inner City
Housing Taskforce 1992).
In European and especially wealthy Asian cities, CBD population density
is declining (11% and 42%, respectively), though in absolute terms they
are still high by international standards (78 and 87 persons per ha in 1990).
Most European cities do have policies and strategies to reduce or reverse
this decline (e.g., Hamburg and Stockholm rose in CBD density), though
the small and still decreasing household sizes of central city apartments
means that new dwelling construction has to be quite aggressive to counter
the downward trend. CBD population in wealthy Asian cities appears to
be being forced out primarily by a trend toward greater commercial uses
(all three cities dropped considerably in downtown residential character).
The developing Asian cities in 1990 had the highest CBD population densi-
ties of all (average of 288 per ha or 5.3 times more than US cities, ranging
up to greater than 400 per ha). Such densities, especially if combined with
efforts to improve housing livability and public environments, can be very
positive features in terms of their potential to reduce dependence on the au-
tomobile.
Of course, another reason why a reasonable build up, rather than a
decline in CBD residential character is desirable from a sustainability per-
spective, is that it can enhance the social aspects of city life and improve
personal security in city streets (Jacobs 1961). Most cities that are attempting
to pedestrianize or traffic calm their central areas and to give transit priority
are doing so at least partly to help make central city living more attractive.
CARS IN CITIES: I M P L I C A T I O N S F O R SUSTAINABILITY 287

INNER AREA DENSITY. In common with the tightly confined CBD area, the
wider inner area or pre-automobile area is also important in affecting
overall metropolitan transport patterns. Most cities in the developed world
have for many years been depopulating their pre-automobile areas, which
were built before the automobile became the dominant mode of transport
after the Second World War. These inner areas were constructed largely
around rail and tram (streetcar) systems and are even today generally much
denser and more mixed in their land use patterns (Newman and Kenworthy
1996). In most cities they are the focus of the metropolitan transit system
and are also where a high proportion of the walking and cycling is done
due to the fine-grained mixing of activities.
Inner area density can thus be a key factor in helping to shape the overall
transport patterns of a city. Those cities having a sizable, high density inner
area that is clearly distinct from the auto-based sprawl around them, will
tend to have lower overall auto use (e.g., New York is America's lowest
per capita user of cars and Sydney with its extensive, dense inner area is
"Australia's New York").
Unfortunately, most inner areas are still in decline or stable in density
terms, despite significant dwelling construction activity throughout the
eighties in many cities. US cities in this sample declined 11% in inner area
density from 1980 to 1990 (to 38 per ha), whereas the six Australian cities
in this sample remained static (again, Australian inner city densities are
on average only around half those of US cities--22 per ha). Australian
cities have had a policy of urban consolidation since the 1980s, which
has arrested, and in some case reversed, inner city population decline in
certain areas.
E u r o p e a n cities also declined a little (4%), but remained moderately
high at more than twice the US level (87 per ha) and 4 times the Australian
level. Inner density in the three wealthy Asian cities decreased by 24%,
but remain at very high levels (353 per ha). The five developing Asian
cities also show very high inner area densities relative to other cities in the
study (average 242 per ha).
Some Canadian cities such as T o r o n t o (and also Vancouver, not shown
in Table 1) have had significant reversals in inner city density decline due
to very strong reurbanization policies, especially high densities around
transit stations. Toronto's inner density rose 7%, whereas Vancouver's
inner city added a net 43,275 people from 1981 to 1991 to increase in density
by 11% (City of Vancouver 1995). These are positive trends in terms of
urban sustainability, because they mean that sprawl is at least being slowed
and population is being added in areas that tend to minimize the need to
travel by car. They demonstrate that a reversal of the auto-based sprawl
pattern can begin to occur, but it requires a strong commitment to effective
public planning.
288 JEFFREY R. KENWORTHY AND FELIX B. LAUBE

15000

[] [] []
[] [] [] [] []

10000- D

Annual car kms

per person

5000- ° °
(o°°
0 1"31" I !
0 5 10 15

Length of road per person (metres)

FIGURE 3. Road availability versus car use per person in world cities, 1990.

Private Transport Infrastructure

Infrastructure for the automobile is represented here by the length of road
per person and the amount of C B D parking relative to e m p l o y m e n t levels.

LENGTH OF ROAD PER CAPITA. It can be seen that US and Australian cities
clearly provide most for the automobile with some 3 to 4 times the level
of roads found in E u r o p e a n cities, 6 to 8 times that found in the three
wealthy Asian cities, and 9 to 12 times m o r e than the five developing Asian
cities. T h e y are even 2.5 to 3 times higher than the less auto-dependent
T o r o n t o (see later discussion). The data suggest that road supply in Toronto,
and wealthy Asian cities between 1980 and 1990 has remained m o r e or
less constant, whereas in US cities and Australian cities it has declined a
little (6 to 7 % ) and in E u r o p e a n cities it has increased slightly (9%).
Per capita auto use significantly increases with increasing road availabil-
ity, as shown in Figure 3 (r 2 = 0.668).

PARKING SPACES PER IOOOC B D JoBs. Parking in the C B D can have a signifi-
cant effect on the level of auto use for central city b o u n d trips, as well as
the overall use of transit in a city, given that radial transit is still the
dominant part of m a n y urban transit systems.
In c o m m o n with the other data discussed so far, Table I shows that there
is a systematic variation in C B D parking levels across cities around the
world. Australian cities lead the world in this factor (489 spaces per 1000
jobs), followed closely by US cities with 462, E u r o p e a n cities with 216,
T o r o n t o with 176, and the wealthy Asian cities with a miniscule 80 spaces
per 1000 C B D jobs. Cities in all groupings have been, on average, increasing
C A R S IN CITIES: I M P L I C A T I O N S F O R SUSTAINABILITY 289

600 []
[]
500- []
[] []

Annualtransit 400- ~ D []
tripsper capita 300 -
[]

[] 0°

10001 d~ ° o
I I I
0 250 500 750 1000

CBDparkingspacesper 1000jobs

FIGURE 4. Parking in the CBD versus annual transit use in world cities, 1990.

in this factor (US cities up by 15%, Australian by 22%, European by 13%,

and Asian cities by 19%). Toronto has however reduced its central city
parking supply by 11%. There is not a lot of data yet on the developing
Asian cities, but what is available suggests that cities such as Bangkok
and Kuala Lumpur (338 and 305 spaces per 1000 jobs) are attempting to
accommodate the automobile in a way that is approaching western levels
in this factor, unlike Singapore, Tokyo, and Hong Kong where the auto
remains firmly under control and CBD parking is held at extremely low
levels (see Kenworthy et al. 1994, 1995; Poboon et al. 1994; Barter and
Kenworthy 1995).
We have shown elsewhere how there tends to be a tradeoff between the
use of land in central cities for parking versus residential development
(Newman and Kenworthy 1988). Using 1990 data on CBD parking and
density, those cities with more CBD parking tend to have lower CBD
population densities (r 2 = 0.493), and Figure 4 shows how overall transit
use per capita is lower in cities with higher levels of CBD parking (r2 =
0.656). These data suggest that increased provision of parking in central
areas not only increases car use and diminishes transit use, but is also
associated with reduced potential for housing in the central city.
Overall, increased CBD parking tends to work against efforts to improve
urban sustainability.

Private Transport Characteristics

We have attempted to characterize the private transport system through a
number of key variables related to vehicle ownership and use, as well as
modal split.
290 JEFFREY R. KENWORTHY AND FELIX B. LAUBE

CARS AND TOTAL VEHICLES PER I O 0 0 PEOPLE. Table 2 shows how there is a
universal increase in passenger and total vehicle ownership. The US cities
generally have the highest overall ownership (602 per 1000 for cars and
755 overall, both having increased about 10% between 1980 and 1990).
T o r o n t o is also high with 606 cars per 1000 people and 706 total vehicles
(31% and 27% increase, respectively). Australian cities are next with 491
cars per 1000 and 595 total vehicles (increases of 10% and 9%, respectively).
E u r o p e a n cities are still significantly behind US and Australian cities in
vehicle ownership with only 392 cars per 1000 (35% less than US cities)
and 452 total vehicles (40% less than US cities). The increase in car and
total vehicle ownership in E u r o p e a n cities was 18% between 1980 and 1990.
Wealthy Asian cities that have the financial capacity to purchase large
numbers of vehicles are still far behind all the other cities with a mere 123
cars per 1000 and 217 total vehicles per 1000 (one-fifth to one-third the
level of ownership in US cities in 1990).
The developing Asian cities tend to be lower in car ownership than the
wealthy Asian cities (104 per 1000 cf 123), though Bangkok and Kuala
L u m p u r are comparatively high with 153 and 184 per 1000 respectively.
Also, total vehicle ownership is higher than their wealthier Asian counter-
parts at an average of 257 per 1000, due especially to the large number of
motor cycles in developing Asian cities, popular for their affordability and
ability to weave through the chronically congested streets.
Overall, in the global sample, higher car owership is associated with
higher car use, as becomes clear in the next section on annual travel in
different cities. However, the question of why wealthy Asian cities such as
Singapore, Tokyo, and H o n g Kong remain very low in car ownership and
use and much less wealthy cities such as Bangkok burgeon in these factors is
an important question to consider. Is increasing wealth inevitably associated
with rampant auto dependence? This question is dealt with briefly later.

A N N U A L CAR KMS AND PASSENGER KMS PER CAPITA. T h e s e factors measure

in the most direct way, the degree to which a city is dependent on the
automobile. The first factor is a primary measure, whereas the second is
used with transit passenger kms to calculate an important overview factor,
the relative proportion of travel in a city undertaken by transit. This is the
last variable and is discussed at the end of this section.
Car use per person has increased in all cities between 1980 and 1990,
though the extent of the variations is very large. Figure 5 summarizes the
data in Table 1 and reveals that US cities with 11,133 kms per person in
1990 are by far the biggest car users in the world (some 1.7 times Australian
cities, 2.5 times E u r o p e a n cities, 7.5 times wealthy Asian cities, and 9 times
the developing Asian cities).
However, the trend is also a real concern. US cities on average were
2091 kms higher in per capita car use in 1990 than they were in 1980. By
CARS IN CITIES: IMPLICATIONS FOR SUSTAINABILITY 291

Car Use Per Capita in World

Cities, 1980-1990

12000 11133

10O0O
Car
kms 8000
per 6000
capita
4000
2000
0
US Cities Aust. Metro Euro. Wealthy Dev'ping
Cities Toronto Cities Asian Asian
Cities Cities

FIGURE 5. Car use per capita in world cities, 1980-1990.

contrast Australian cities grew by 720 kms per capita, T o r o n t o by 781 kms,
E u r o p e a n cities by 993 kms and Asian cities by 564 kms. T h e US increase
can be put into perspective by considering that in 1990, T o k y o ' s total car
use per person was 2,103 kms. Los Angeles, which is reasonably typical of
other US cities in this factor, has been doubling its increase in car use per
person every decade between 1960 and 1990 (Kenworthy and N e w m a n
1993). This is exponential growth at a time when a whole host of pressures
are demanding that cities b e c o m e less dependent on the automobile and
less damaging of their physical and social environments.

PERCENTAGE OF WORKERS USING FOOT A N D BICYCLE. This factor provides

some measure of the importance of nonmotorized modes in private trans-
port. Table 1 shows that US and Australian cities, as well as Toronto, are
a b o u t equal at only 5% of workers getting to work on foot and bicycle,
and there has been a marginal reduction in all these cities in this factor
over 10 years. On the other hand, the ten E u r o p e a n cities represented in
Table 1 have some 17% of workers using these modes (and this excludes
one of the most bicycle-oriented cities, A m s t e r d a m , due to present lack
of data).
However, there has generally been a reduction in walking and cycling
to work in E u r o p e a n cities from 1980 to 1990. Zurich stands out as having
achieved an increase from 21% to 24%, which appears to be associated
with a concurrent dramatic i m p r o v e m e n t in transit (see next section).
292 JEFFREY R. KENWORTHY AND FELIX B. LAUBE

15000

10000
Annual car kms
per person

5000
- °[]~ rl
0 10 20 30 40

Percentage of workers walking and cycling

FIGURE 6. Percentage of workers walking and cycling versus car use per person
in world cities, 1990.

There are problems with obtaining journey-to-work mode data on the

Asian cities, though Bangkok is very low for a dense Asian city with only
10% of workers using foot or bicycles, whereas Surabaya and Kuala Lumpur
are more healthy at 31% and 17% respectively. The low use of walking
and cycling generally in Bangkok is partly a reflection of the horrific physical
environment conditions of air pollution, noise, danger, and lack of adequate
facilities along most streets in Bangkok, even within the residential sois,
most of which are very narrow and overwhelmed by the automobile (Ken-
worthy 1995).
Correlating this factor with annual car use per person (Figure 6), there
is a reasonably strong negative relationship (r 2 = 0.660).
For cities to be truly sustainable in the future, there needs to be a policy
focus on the nonmotorized modes (e.g., see H o o k 1993; Hillman 1990).
Cities such as Freiburg, Germany show what can be achieved with a strong
policy focus on walking and cycling (see later case example).

Public Transport Characteristics

These data are measured by three key variables: the annual service kilome-
ters provided per person, the annual usage of transit in terms of trips per
capita (unlinked trips), and the annual per capita passenger kilometers.
The data again show a highly systematic trend in all factors.

ANNUAL SERVICE KMS/CAPITA. The US cities are by far the worst in terms
of transit provision, ranging progressively upward to the wealthy Asian
CARS IN CITIES: IMPLICATIONS FOR SUSTAINABILITY 293

Transit Trips Per Capita in

World Cities, 1980-1990
496 ~5~
500
Annual 400
transit
trips 300
per
capita 200

100

US Cities Aust, Metro Euro. Wealthy Dev'ping

Cities Toronto Cities Asian Asian
Cities Cities

FIGURE 7. Annual transit trips per capita in world cities, 1980-1990.

cities, which have very high levels. US cities provide a meager 28 kilometers
of service per person each year, whereas the Australian cities provide more
than double this level (60 kms), the European cities 3.3 times more (93
kms), Metro Toronto a large 3.6 times more (106 kms), and the wealthy
Asian cities some 4 times more service (114 kms). Data for developing
Asian cities are difficult to get but suggest high service kms (average of
116 kms per person for the cities in Table 2).
The trends for 1980 to 1990 show all city groupings increasing in service
provision, though mostly only by small to moderate margins (United States
up by 4%, Australian up by 5%, European by 16%, and Asian cities
by 11%).

ANNUAL TRANSITTRIPS/CAPITA. Transit use per capita in 1980 and 1990 is

graphed in Figure 7. It shows a similar trend to service kms with US cities
at very low levels internationally (63 trips per person annually), Australian
cities a little better at 92 trips, European cities a major leap ahead with
322 trips annually, Metro Toronto with 324 trips, and the wealthy Asian
cities with a very large 496 trips per person (or nearly 8 times more than
US cities). It should be noted that the Greater Toronto Area had 210
transit trips per person in 1991, which is still very high by North American
standards--New York, the next highest had 155 trips per capita (see Ken-
worthy and Newman 1993, 1994).
Developing Asian cities are also high in transit use with an average of
301 trips annually in 1990, though there are large variations from only 129
294 J E F F R E Y R. K E N W O R T H Y A N D FELIX B. L A U B E

trips in Surabaya with its poorly developed transit system, up to 448 in

Manila with its highly developed jeepney system.
Within this sample there are some exceptional success stories in transit.
For example, under very strong transport and urban planning policy support
throughout the 1980s, Swiss cities prioritized transit and put large sums of
m o n e y into network development, rolling stock, operational improvements
such as full signal priority in traffic, as well as some extremely successful
marketing campaigns based on appeals to peoples' environmental con-
sciousness (Taplin 1992; Hass-Klau 1990; H o p p e 1993; Newman and Ken-
worthy 1996). They also had a strong emphasis on integrating development
around transit stations. Zurich as a result increased in annual transit use
from 344 to 515 trips per capita from 1980 to 1990, an increase of some
170 trips or 50%. This 10-year increase alone can be seen in perspective
by considering that it is similar to New York's total annual per capita transit
use in 1990 and over 2.5 times the US cities' average. Transit as a percentage
of total passenger travel in Zurich rose from 20% to 23%, a difficult reversal
to achieve for a whole metro area (data from present study).
Singapore and Hong Kong achieved similar large increases in per capita
transit use over 10 years, each increasing by 104 trips per capita, or 29% and
22%, respectively. Transit ridership of 104 trips per capita is approximately
equal to Melbourne's and Washington's total annual transit use in 1990.
The development of extensive mass rapid transit systems in Hong Kong
and Singapore, the exceptional integration of development around stations
in very high density residential and mixed use nodes, plus physical and
economic restraints on automobiles all contributed strongly to these cities'
success with transit (Kenworthy et al. 1994; Urban Redevelopment Author-
ity 1991; Pun 1980; Bygate 1980).

ANNUALTRANSIT PASS. KM/CAP1TA. These data show the same trend as transit
trips and are used primarily to calculate the next factor, which is a measure
of the balance between transit and the private car in cities.

Public~Private Transport Balance

This factor to some extent
P E R C E N T A G E OF T O T A L P A S S E N G E R KMS ON TRANSIT.
summarises many of the other factors discussed in the paper. It shows a
very dramatic difference in the patterns of transport within cities.
US cities in 1990 had on average a mere 3.2% of total motorized passen-
ger travel on transit (down from 3.9% in 1980). The lowest city was Sacra-
mento with 0.6%, though a number were only around 1 % - - H o u s t o n , Phoe-
nix, Detroit, Denver, and San Diego. New York was the highest US city
in this factor with 11%. Australian cities are not very high either (7.7% in
1990), though they are more than double the average US figure and did
improve marginally between 1980 and 1990.
E u r o p e a n cities rank very much higher in this factor at 22.9% or over
7 times US cities, and transit appears to have basically held its own over
CARS IN CITIES: IMPLICATIONS FOR SUSTAINABILITY 295

100.0-

[]

75.0-

Percentage of
total pass. kms 50.0-
on transit

25.0 -

0.0 i I I

0 100 200 300 400

Urban density (persons per ha)

FIGURE 8. Urban density versus percentage of total passenger travel on transit

in world cities, 1990.

the 10-year period. The cities that stand out most in this factor are the
wealthy Asian cities with an average of 64.1% of all motorized passenger
travel on transit (20 times higher than US cities), though they have declined
marginally from 67.0% in 1980.
Developing Asian cities, with an average for the five cities of 38.3%, are
also high internationally in this factor, though they are much less than their
wealthier cousins, and are the subject of very rapid increases in automobile
dependence. Without the well-developed transit systems, the traffic re-
straint schemes (both physical and economic), and the generally strong
public planning traditions of Singapore, Tokyo, and Hong Kong, the devel-
oping Asian cities are in danger of experiencing extreme impacts from
auto-based environmental problems (as already evident in Bangkok). This
is discussed in detail (Poboon et al. 1994; Barter and Kenworthy 1995;
Kenworthy 1995; Kenworthy et al. 1995) and summarized below in the
discussion of the wealth factor in automobile dependence.
This key overview factor, like previous factors, is strongly related to
urban density across an international sample of cities, as shown in Figure
8 (r 2 = 0.800) and tends to suggest the conclusion that cities must address
the density issue if they are to seriously tackle automobile dependence.

Car Ownership, Car Use, and Wealth

Many urban commentators suggest that there is an inevitability between
rising incomes and increasing car use. They also argue that transit will
automatically decline as the financial capacity to use private modes grows
(Gomez-Ibafiez 1991; Lave 1992). Transport models build these assump-
296 JEFFREY R. K E N W O R T H Y AND FELIX B. L A U B E

tions into their predictions of future traffic and road needs. The tendency
to use wealth to purchase cars is clearly a very strong force. However, the
extent of their use in relation to wealth is rarely studied on a city-wide basis.
We have dealt with the issue of rising wealth and automobile dependence
in detail in a recent study, with particular reference to Asian cities (Kenwor-
thy et al. 1995). It is important to briefly summarize the results of this work
as they tend to demonstrate a more widely applicable conclusion, i.e., that
rising wealth is not necessarily associated with uncontrollable increases in
automobile dependence. However, for this to hold true, strong physical
planning policies that favor non-auto modes and economic constraints are
required. These will ensure that the price of auto use and ownership more
adequately reflects its impacts, and that good quality transit systems as well
as safe and attractive environments for walking and cycling are provided
so that auto use is not essential.
Kenworthy et al. show that although there is a not unexpected general
linear rise in per capita car use with increasing wealth (r z = 0.44 with real
gross domestic product ( R G D P ) and 0.62 with city median income), the
relationship is not straightforward. Within the Asian cities sample there is
no significant correlation between car use and wealth (r 2 = 0.12 with R G D P ,
and 0.18 with city median income). As well, transit use is not correlated
at all with wealth across the global sample (r 2 = 0.02 with R G D P and 0.09
with city median income), and within the Asian cities sample there is a
moderate correlation (r 2 = 0.24 with R G D P and 0.19 with city median
income), but it shows transit use increasing with wealth which tends to go
against conventional wisdom. Many wealthy cities such as Zurich have
excellent transit systems as already discussed.
A few examples serve to demonstrate some of these points.

• US cities are leaders in both automobile dependence and wealth and

have the lowest use of public transport. However, many cities with
wealth levels around 85% that of the US cities, have very much lower
car use and dramatically higher transit use. For example, Greater
T o r o n t o has 50% the car use and 3 times the transit use of US cities,
the E u r o p e a n cities have 40% of the car use and 5 times the transit
use. And, significantly, H o n g Kong, Singapore, and Tokyo average
only 13% of the car use and 8 times the transit use of US cities,
despite having wealth levels between 67% and 88% of US levels.
• Within the Asian cities sample, car ownership levels vary greatly. For
example, Hong Kong with 84% of the R G D P of US cities has a tiny
43 cars/1000 people versus 602/1000 in US cities. Conversely, Kuala
Lumpur has 184 cars/1000 people and Bangkok 153, yet Malaysia's
and Thailand's R G D P averages only 28% that of the US. Singapore
with 67% the R G D P of US cities has only 101 cars/1000 people.
• Wealthy Tokyo has a very large 45% of all trips by foot and bicycle,
CARS IN CITIES: IMPLICATIONS FOR SUSTAINABILITY 297

whereas much poorer Bangkok and Kuala Lumpur have only 14%
and 20%, respectively. Private transport accounts for over 50% of
total trips in these two cities, whereas in Tokyo it is only 27%.
Wealth alone is therefore an inadequate explanation for the broad pat-
terns just outlined. We argue in particular that a number of physical plan-
ning factors, as discussed in this article, can help us to understand the
patterns of automobile dependence in international cities. These include:
• land use patterns, in particular urban densities and mixed land use
with a special emphasis on the development of strategic nodes of high
density mixed uses around transit stations, i.e., urban villages (e.g.,
see Newman, Kenworthy, and Robinson 1992; Newman and Kenwor-
thy 1992);
• transport infrastructure priorities, in particular road and parking pro-
vision levels;
• transit provision, in particular emphasis on rail with its higher speeds
and passenger attraction;
• the opportunities for nonmotorized modes in terms of basic land use
conducive to short trip making, as well as good physical environments
to encourage safe, pleasant walking and cycling.
Within the Asian cities in particular, we argue that a number of factors
related to the above points (as well as economic factors to do with the
price of car ownership and use), help provide an understanding of why
some cities such as Bangkok are burgeoning in car ownership and use (see
Poboon et al 1994; Kenworthy 1995), whereas others such as Singapore,
Hong Kong, and Tokyo seem to have the automobile much more under
control.

Working Toward a Sustainable Transport System

We can now take this international perspective on automobile dependence
and its causes and explore what it means for the development of sustainable
urban transport systems.
There are still many arguments over the concept of sustainable develop-
ment or sustainability, its applicability in different situations and in some
cases its very definition (e.g., Lowe 1991; Newman 1990). Despite this, the
idea of sustainability appears to be gaining ground and influence in policy
making, particularly at a global level. Increasingly at a local level, mecha-
nisms and processes are being developed for implementing sustainable
development principles (e.g., Parenteau 1994). International organizations
such as the OECD, the ECE, the World Bank, and the United Nations
Center for Human Settlements all have sustainable cities programs. In 1994
the Global Forum on Cities and Sustainable Development in Manchester
received reports from 50 cities on their progress toward sustainability and
298 JEFFREY R. K E N W O R T H Y A N D FELIX B. L A U B E

in June 1996 "The City Summit" or Second United Nations Conference

on Human Settlements (Habitat II) in Istanbul heard from all nations on
progress toward sustainability in cities.
Newman (1995) has suggested that sustainability has become a "global
guiding principle." In terms of cities, one implication is that many of the
individual and perhaps disjointed policies and strategies suggested over the
years for helping to alleviate auto-dependence in cities, can be brought
together under this unifying concept and perhaps be given greater currency
and urgency.
Under the rubric of sustainability, there is thus now a need for cities to
try to fulfill the global agenda of "reducing the city's use of natural resources
and its production of wastes while simultaneously improving its human
amenity and human environment" (Newman 1995). This means bringing
together the economic, social, and environmental goals of cities.
As part of the preparations for The City Summit, the World Bank and
the United Nations Center for Human Settlements undertook a research
exercise entitled "Indicators Program: Monitoring the City," which at-
tempted to collect 74 key indicators and 76 extensive indicators covering
poverty, employment, productivity, social development, infrastructure,
transport, environmental management, and local government (World Bank
and UN Center for Human Settlements 1994). There is thus a strong trend
to try to quantify various characteristics of cities that relate to sustainability
and thereby ultimately to be able to measure trends in each city on key
parameters to assess progress toward sustainability.
The comparisons between cities presented here suggest certain basic
directions in the way cities should evolve if they are to become more
sustainable. However, in line with the above trend toward quantification
of basic city characteristics, it is useful to have a set of guidelines whereby
to annually assess a city's progress toward a sustainable transport system.
The development of Local Action Plans for sustainability as required by
Agenda 21 stresses the need to have a clear set of measurable goals (Paren-
teau 1994; Newman 1995).
Based mostly on the comparisons of cities set out here, Figures 9 to 12
provide a series of specific measurable annual goals or objectives whereby
cities might judge their performance in moving toward sustainability
(adapted from Newman 1995). It should be stressed that they are by no
means exhaustive, but they do provide a picture of the general direction
required in the physical planning of cities if they are to move towards sus-
tainability.
• Land use objectives (more transit-oriented, higher density, mixed land
uses that help to halt the growth in auto-based development).
These objectives relate to the data on urban density, CBD density, and
inner area density by suggesting the need to reduce the consumption of
CARS IN CITIES: IMPLICATIONS FOR SUSTAINABILITY 299

Sustainable Transport Planning:

Land Use Objectives Each Year
Loss of agdoulmral land and bushland at urban I
trfnge~ r ~ u ~

Number of specially zoned, translt-<)~ented

lotmtlons, Increased (urban villages, maln streets
1;1
Denslty of population and emp4oyment In tmnslt-
oriented locations, Increased
I Proportion of city mployment In translt,<~tlented
sub-centres of more than 5,000 jobs, Incrolmed
Portton of urban land under single use zoning
fused
"Proportion of ud:mn land under fine-grained, 1
t...m~x~ ,u__.~u,~,
9 zon , I ~ . . ~ ....................................
j

FIGURE 9. Land use objectives for sustainable transport planning.

land by sprawl on the urban fringe and instead, to concentrate on redevelop-

ment and infilling within the existing urban boundary. In particular, the
objectives stress the need to reintegrate development around transit systems
in the form of high density, mixed use urban villages. They imply the
need to progressively make more opportunities available for this kind of
development, while reducing the areas of single use, low density housing.
• Private transport objectives (stabilized or lower car use and less em-
phasis on infrastructure for cars).
The need here is to gradually reduce the amount of auto travel by
reducing single occupant vehicle trips, shortening trip distances, decreasing
the modal share of cars and discouraging auto trip-making through lower
provision of roads and parking.
• Public transport objectives (higher quality transit systems, especially
rail, which is more competitive with cars).
Of course, the private transport objectives are difficult to achieve unless
there is simultaneously more provision of good quality transport and land
use alternatives. Better transit systems are essential to this. These objectives
stress the need to make transit more competitive with cars in speed terms
and to provide higher levels of transit service, in particular to increase
transit's visibility and presence in the road system through new light rail
300 JEFFREY R. KENWORTHY AND FELIX B. LAUBE

Sustainable Transport Planning:

Private Transport Objectives Each Year

Car use per capita (vehicle kilometres travelled), decreased 1

[ Modal share of"car driver only" work trips, decreased l
[ Modal share of cars for all trips, decreased I

[ Average distance tot other trips ishoppin8, business etcL decreased! |I Infrastructure
Less I
[ Length of road relative to population, decreased ] L . f°r CaT__~

I Parking spaces per 1000 CBD jobs, decreased ]

I Parkin spaces per 1000 jobs in sub~centre~, decreased }

FIGURE 10. Private transport objectives for sustainable transport planning.

systems where appropriate (Newman and Kenworthy 1991). Although not

dealt with here, there is also a need to put transit on a more financially
firm basis through better cost recovery, which perhaps involves increased
income or capital contributions from joint development and value capture

Sustainable Transport Planning:

Public T r a n s p o r t O b j e c t i v e s Each Year
........................ : ::!"":::!!!~!~::
:

Relative average speed of transit compared to cars, Increased]

Length of road system having rail transit (light rail), tncraaead1
i Servlce kllometres of transit par kllometra of road, Incraased I
~.o.a, shSrao, trans, woe,rips, ,nora.,~I ( . . . . . . 7-..;,~i;:--~

~ . : : : ..... ;;: ...........;m;:;;;: "-......................I Higher t

[Propo.lon of transit pas.nger kllomatrall ori rail, tn~a;e-d'J ~ . ~ . n s L ~
[ Cost recovery on transit fares, Increased]

FIGURE 11. Public transport objectives for sustainable transport planning.

CARS IN CITIES: IMPLICATIONS FOR SUSTAINABILITY 301

FIGURE 12. Nonmotorized mode objectives for sustainable transport planning.

mechanisms based on the integration of high density development around

stations (Keefer 1985; Cervero, Hall, and Landis 1992).
• Nonmotorized mode objectives (greater safety and amenity for walk-
ing and cycling and increased use of these modes).
In auto-dependent cities it is easy to forget the potential role of walking
and cycling, which are the most sustainable of all modes. As well, the
transport problems of developing Asian cities could be alleviated by greater
attention to these modes, which are ideally suited to the dense, mixed use
urban fabric of these cities (Kenworthy et al. 1995).
In order to enhance the role of nonmotorized modes for work and other
trips, there needs to be greater attention to providing environments that
are safe and attractive to pedestrians and cyclists, through pedestrianization
and traffic calming schemes (preferably area-wide), and by increasing the
provision of cycling facilities (cycleways, bike parking, carriage of bikes on
transit, etc.) (Pharaoh and Russell 1989; TEST 1989; Monheim 1988).
Many of the above four sets of objectives relate directly to the data and
analysis provided in this article. The relative importance of each objective
in different cities will vary and will change over time, but the basic thrust
of the objectives is generally applicable to most cities in the world, if
sustainable transport is to become a reality. These are of course city planning
objectives only, and as stated before, are thus not exhaustive (for example,
they do not deal with transport technology objectives such as greater fuel
302 J E F F R E Y R. K E N W O R T H Y A N D FELIX B. L A U B E

Freiburg:
Reducing Automobile Dependence
I III IIII IIIII IIII II II

loo

Federal
Republic
of
Germany
% of n0n-walk % of non*walk % change in use
trips, 1976 trips, 't99t t976-199~
Freiburg
Source; D~ta in Pucher and C~orer(1992~
Taming the Atl~omobite in Germany

FIGURE 13. Reducing automobile dependence in Freiburg, Germany.

efficiency and lower emissions from vehicles), but they do give some basic
physical planning directions whereby cities might begin to judge in which
direction they are heading: toward sustainability or toward increased auto-
mobile dependence and auto-based resource consumption and environmen-
tal impact.
One brief case example serves to highlight the potential effectiveness of
a coordinated suite of policies based around the above objectives.

Toward Sustainable Urban Transport: A Case Example

from Freiburg, West Germany
Freiburg, West Germany has shown that it is indeed possible to virtually
stop the growth of car use, even when car ownership is growing. Pucher
and Clorer (1992) provide data that show how Freiburg's car ownership
has risen from 113 per 1000 people in 1960 to 422 per 1000 in 1990, only
a little under the average for the Zurich agglomeration, and only 12% less
than the national average for West Germany (481 per 1000).
Figure 13 shows how, despite this growth in availability of cars, car
use has virtually remained constant since 1976. Transit passengers have
increased 53%, and bicycle trips have risen 96% between 1976 and 1991.
Freiburg's growth in car trips in 15 years was only 1.3%, yet total trips
increased 30%. Freiburg's growth in mobility was supplied principally by
increased public transport and bicycling. In fact the share of trips by car
reduced over the 15 years from 60% to 47%. Pucher and Clorer also show
C A R S IN CITIES: IMPLICATIONS F O R SUSTAINABILITY 303

how the growth in car ownership has also begun to slow down (Freiburg
had previously been higher in car ownership than West G e r m a n y as a
whole, whereas now it is less).
Pucher and Clorer (1992) attribute Freiburg's success at "taming the
automobile" to a combination of transportation and physical planning strat-
egies; these relate strongly to the four sets of sustainability objectives
just discussed:
First, it has sharply restricted auto use in the city. Second, it has provided
affordable, convenient, and safe alternatives to auto use. Finally, it has
strictly regulated development to ensure a compact land use pattern that
is conducive to public transport, bicycling, and walking (p. 386).
Restricted auto use has been achieved through mechanisms such as pedes-
trianization of the city center, area-wide traffic calming schemes (citywide
speed limit of 30 km/h in residential areas) and more difficult, expensive
parking. Freiburg's improvements to transit have focused on extending and
upgrading its light rail system as opposed to buses. Buses are used as feeders
to the light rail system. Land use regulations are similar to those in many
other parts of Europe and have involved limiting the overall amount of
land available to development and strictly zoning land for agriculture,
forests, wildlife reserves, or undeveloped open space.
Pucher and Clorer stress the important automobile use savings of the
more compact urban patterns that have resulted from these latter policies.
It is also worth noting that after the Second World War it was decided to
rebuild Freiburg, totally destroyed by the war, on the old model, not on
an auto-orientated model. They note that even in the post 1960s period,
as Freiburg expanded on flatter land to the west, the resulting development
" . . . is at a much higher density than outlying portions of American metro-
politan areas", as well as being within easy reach of public transport and
well-served by bikeways.

Conclusion
The automobile represents a major cause of environmental impact in devel-
oped and developing cities around the world. This article has examined
some key land use and transport differences for 1980 and 1990 across a
large international sample of cities and shown how automobile dependence
and its underlying factors vary in a systematic way. US cities are by far the
most auto-dependent with the highest levels of auto use and lowest levels
of transit provision and use. They have also increased by far the most in
their use of cars over the 10-year period. Following US cities in auto
dependence are the Australian cities, then Toronto (the one Canadian city
so far analyzed in full in the study), European cities, wealthy Asian cities
(Singapore, Tokyo, and Hong Kong), and developing Asian cities, which
have all added much more modest extra car kilometers per person.
304 JEFFREY R. K E N W O R T H Y AND FELIX B. L A U B E

The key land use, infrastructure, and public transport factors that lie
behind this picture also tend to vary in the same systematic way. Densities
are lowest in the US and Australian cities (by a factor of three even
compared to the next lowest, Metro Toronto). Per capita metropolitan
road provision and CBD parking are also highest in these cities and drop
dramatically in Toronto and the European cities to reach very low levels
in the Asian cities. Vehicle ownership generally diminishes in a similar
way, though some developing Asian cities have atypically high levels for
their wealth (especially compared with their much more wealthy Asian
neighbours in Singapore, Tokyo, and Hong Kong). Walking and cycling
for the journey to work is low in US and Australian cities and in Toronto
but increases in European and Asian cities.
The very systematic increase in transit provision and use is quite marked
with US cities clearly at the bottom and wealthy Asian cities an order of
magnitude higher. Transit has basically been struggling to hold its small
though important role in the overall transport systems of US and Australian
cities over the period 1980 to 1990. On the other hand, European cities,
especially Zurich, have generally improved their service provision and use,
as have the wealthy Asian cities. Transit use, though high in developing
Asian cities, is lower than in their wealthier cousins and tends to suffer
from a lack of quality systems and severe competition from the automobile.
The overall relative role of transit in cities varies from a low of 3% of all
passenger travel in US cities to a high of 64% in wealthy Asian cities, a
massive difference in automobile dependence.
The article demonstrates significant statistical relationships between the
key transport and land use variables: urban density is a key explanatory
variable in auto and transit use as well as the relative role of transit (auto use
increasing and transit decreasing with decreasing density). Road provision,
parking, and nonmotorized mode use are all also strongly associated with
the pattern of auto-dependence across cities.
From the analysis, the study offers a series of directions for land use and
transport in cities if the goal of increased sustainability is to be met. It
suggests that the concept of sustainability is becoming a key global guiding
or unifying principle under which to take action to reduce auto dependence
in cities. The specific sustainability goals are summarized under:
• Land use objectives (more transit-oriented, higher density, mixed land
uses which help to halt the growth in auto-based development).
• Private transport objectives (stabilized or lower car use and less em-
phasis on infrastructure for cars).
• Public transport objectives (higher quality transit systems, especially
rail, which are more competitive with cars).
• Nonmotorized mode objectives (greater safety and amenity for walk-
ing and cycling and increased use of these modes).
CARS IN CITIES: IMPLICATIONS FOR SUSTAINABILITY 305

It is o n l y t h r o u g h a c o n c e r t e d set o f a c t i o n s on all t h e s e f a c t o r s (as in

the case e x a m p l e o f F r e i b u r g , W e s t G e r m a n y ) p l u s o t h e r s n o t d i s c u s s e d
in this p a p e r , c a n cities h o p e to b e g i n to r e d u c e their a u t o m o b i l e d e p e n -
dence and become more sustainable.

The authors gratefully acknowledge the contribution of three PhD students in this work: Paul
Barter for his collecting and assembling of the data on Kuala Lumpur, Surabaya, Jakarta,
and Manila (with Bendicto Guia Jr.), and Cbamlong Poboon for similar work on Bangkok.
As well, we would also like to thank Toni Kasim for her additional assistance in the data
work on the developing Asian cities. We also would like to thank the innumerable people
and agencies in cities all around the world who have given generously and freely of their
time and expertise in providing the basic data behind this study. Without their help, this work
would not have been possible. Finally, the work was partly funded by a grant from the Asia
Research Centre at Murdoch University and the World Bank which is gratefully acknowledged.
The authors remain responsible for the presentation and interpretation of results.

References
A C T City Services. 1994. Making Canberra a Better Place: Ideas for a More Sustain-
able Future. Australian Capital Territory, Canberra.
Barter, P. 1995. Urban transport, land use and energy in four Southeast Asian
cities: Jakarta, Kuala Lumpur, Manila, and Surabaya. PhD thesis in preparation.
Barter, P., and Kenworthy, J. 1995. Transport and urban form in A S E A N cities:
An international comparative perspective. Paper presented to City Trans Asia
'95 Conference, World Trade Center, Singapore, September 21-23.
Barter, P., Kenworthy, J., Poboon, C., and Newman, P. 1994. The challenge of
Southeast Asia's rapid motorisation: Kuala Lumpur, Jakarta, Surabaya, and Ma-
nila in an international perspective. Presented to Asian Studies Association of
Australia Biennial Conference, July 13-16, 1994, Asia Research Centre, Murdoch
University, Perth, Western Australia.
Bygate, J. 1980. Transport and urban development integration. Mass Transit Rail-
way Corporation, MTR Planning report, Hong Kong.
Calthorpe, P. 1993. The Next American Metropolis: Ecology, Community and the
American Dream, Princeton Architectural Press.
Campbell C.J. 1991. The Golden Century of Oik 1950-2050, Dordreeht: Kluwer
Academic.
Cervero, R., Hall, P., and Landis, J.D. 1992. Transit-linked joint development: A
national study. National Transit Access Center, monograph 42.
City of Vancouver. 1995. Vancouver local areas, 1981-1991. City of Vancouver
Planning Department, July.
Fleay, B.J. 1994. Liquid petroleum is peaking: Decline of the age of oil. ISTP
Occasional Paper 3/94, Murdoch University, Perth, Western Australia.
Gomez-Ibafiez, J.A. 1991. A global view of automobile dependence. Journal of the
American Planning Association 57(3):376-379.
306 JEFFREY R. KENWORTHYAND FELIXB. LAUBE

Guia, B. 1995. Towards a sustainable transportation system for Metro Manila. PhD
thesis in preparation, Murdoch University, Perth, Western Australia.
Hass-Klau, C. 1990. Greening the streets with public transport: How the Swiss and
Germans have learnt to promote the benefits. Urban Transport International
May/June:12-17.
Hillman, M. 1990. Planning for the green modes: A critique of public policy and
practice. In The Greening of Urban Transport: Planning for Walking and Cycling
in Western Cities, R. Tolley (ed). London: Belhaven Press, Chap. 4, pp. 64-74.
Holtzclaw, J. 1994. Using residential patterns and transit to decrease auto depen-
dence and costs. Natural Resources Defense Council, June.
Hook, W. 1993. Non-motorized transport and traffic congestion management. Exec-
utive Conference on Traffic Congestion Management, October 14.
Hoppe, K. 1993. The importance of public transport in a strongly ecological-orien-
tated traffic policy: The case of Berne. ITE 1993 Compendium of Technical
Papers, 383-387.
Jacobs, J. 1961. The Death and Life of the Great American Cities, New York:
Random House.
Katz, P. 1994. The New Urbanism: Toward an Architecture of Community, New
York: McGraw Hill.
Keefer, L.E. 1985. Joint development at transit stations in the United States. Trans-
portation 12:333-342.
Kenworthy, J.R. 1995. Automobile dependence in Bangkok: An international com-
parison with implications for planning policies. World Transport Policy and Prac-
tice 1(3):31--41.
Kenworthy, J.R., Barter, P., Newman, P.W.G., and Poboon, C. 1994. Resisting
automobile dependence in booming economies: A case study of Singapore, Tokyo
and Hong Kong within a global sample of cities. Paper presented to Asian Studies
Association of Australia Biennial Conference, "Environment, State and Society
in Asia: The Legacy of the Twentieth Century," July 13-16, Asia Research Centre,
Murdoch University, Perth, Western Australia.
Kenworthy, J., and Newman, P. 1993. Automobile dependence: "The irresistible
force"? Commissioned report for the University of Technology, Sydney, Faculty
of Design, Architecture and Building, Institute for Science and Technology Policy,
Murdoch University, Perth, Western Australia, 49 pp.
Kenworthy, J., and Newman, P. 1994. Toronto---Paradigm regained. Australian
Planner 31 (3):137-147.
Kenworthy, J.R., Newman, P.W.G., Barter, P., and Poboon, C. 1995. Is increasing
automobile dependence inevitable in booming economies?: Asian cities in an
international context. IATSS Research 19(2):58-67.
Lave, C. 1992. Cars and demographics. Access 1:4-11. University of California
Transportation Center, 108 Naval Architecture Building, Berkeley.
Lowe, I. 1991. What is sustainable development? Consuming Interest June/July:4--6.
C A R S IN CITIES: I M P L I C A T I O N S F O R SUSTAINABILITY 307

Monheim, R. 1988. Pedestrian zones in West Germany--The dynamic development

of an effective instrument to enliven the city centre. In New Life for City Centres:
Planning Transport and Conservation in British and German Cities. C. Hass-Klau
(ed). London: Anglo-German Foundation, pp. 107-130,
Murphy, D. 1994. Sale of two cities. The Bulletin January 18:32-33.
Newman, P. 1990. Sustainable development. Environmental Education and Informa-
tion 8(4):250-261.
Newman, P. 1991. Greenhouse, oil and cities. Futures May:335-348.
Newman, P.W.G. 1995. Sustainable cities for a sustainable world: From visions to
reality. Presented to "Eco-Fire: Sustainable Cities, Sustainable World Confer-
ence," City of Waitakere, New Zealand, March.
Newman, P.W.G., and Kenworthy, J.R. 1988. Parking and city centre vitality. Paper
presented to International Parking Conference, Perth, October.
Newman, P., and Kenworthy, J. 1989. Cities and Automobile Dependence: An Inter-
national Sourcebook. Aldershot, England: Gower Publishing.
Newman, P., and Kenworthy, J. 1991. Towards a More Sustainable Canberra: An
Assessment of Canberra's Transport, Energy, and Land Use. Institute for Science
and Technology Policy, Murdoch University, Perth, Western Australia.
Newman, P., and Kenworthy, J. 1992. Transit-oriented urban villages: Design model
for the 90s. Urban Futures 2(1):50-58
Newman, P.W.G., and Kenworthy, J.R. 1996. The land use-transport connection:
An overview. Land Use Policy 13(1):1-22.
Newman, P., Kenworthy, J., and Robinson, L. 1992. Winning Back the Cities, Sydney:
Australian Consumers Association and Pluto Press.
Newman, P., Kenworthy J., and Vintila, P. 1992. Housing, transport and urban
form. Background Paper 15, National Housing Strategy, Commonwealth of Aus-
tralia, Canberra.
New York Metropolitan Transportation Commission. 1995. Report by Yuri Teles-
hevsky on 1990 NPTS data that NYMTC has processed (memorandum dated
February 9, 1995.
Nowlan, D.M., and Stewart, G. 1992. The effect of downtown population growth
on commuting trips: Some recent Toronto experience. Journal of the American
Planning Association 57(2):165-182.
Parenteau, R. 1994. Local action plans for sustainable communities. Environment
and Urbanization 6(2):183-199.
Perth Inner City Housing Taskforce. 1992. City living. Report of the Perth Inner
City Housing Taskforce, City of Perth, March.
Pharaoh, T., and Russell, J. 1989. Traffic calming: Policy evaluation in three Euro-
pean countries. Occasional Paper 2/89, Department of Planning, Housing and
Development, South Bank Polytechnic, London.
Poboon, C., Kenworthy, J.R., Newman, P.W.G., and Barter, P. 1994. Bangkok:
Anatomy of a traffic disaster. Paper presented to Asian Studies Association of
Australia Biennial Conference, "Environment, State and Society in Asia: The
308 JEFFREY R. KENWORTHYAND FELIXB. LAUBE

Legacy of the Twentieth Century," July 13-16, Asia Research Centre, Murdoch
University, Perth, Western Australia.
Pucher, J., and Clorer, S. 1992. Taming the automobile in Germany. Transportation
Quarterly 46(3):383-395.
Pun, K.S. 1980. Urban planning principles in Hong Kong--Concepts and ap-
proaches. Asian Architect and Builder December:48-56.
Taplin, M. 1992. Model system: The Ziirich way. Light Rail Review 3:5-11, Platform
5 Publishing Ltd and the Light Rail Transit Association.
TEST. 1989. Quaffty Streets: How Traditional Urban Centres Benefit from Traffic
Calming, London: Transport and Environment Studies, May, Abstract and pp.
5-21.
Urban Redevelopment Authority. 1991. Living the Next Lap: Towards a Tropical
City of Excellence, Singapore: Urban Redevelopment Authority.
World Bank and UN Center for Human Settlements. 1994. Indicators programme:
Monitoring the city. Vol. 2, Urban Indicators Review, The Survey Instrument,
Preliminary Worksheets, United Nations Center for Human Settlements, Nai-
robi, September.