This article was downloaded by: [University of Cambridge]

On: 19 November 2008

Access details: Access Details: [subscription number 902174449]
Publisher Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Transport Reviews
Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t713766937

Transport and urban form in thirty-two of the world's principal cities

Peter W. G. Newman a; Jeffrey R. Kenworthy b
a
Associate Professor in Environmental Science, Murdoch University, Murdoch, Western Australia b Research
Fellow, Institute for Science and Technology Policy, Murdoch University, Murdoch, Western Australia
Online Publication Date: 01 July 1991

To cite this Article Newman, Peter W. G. and Kenworthy, Jeffrey R.(1991)'Transport and urban form in thirty-two of the world's principal

cities',Transport Reviews,11:3,249 272

To link to this Article: DOI: 10.1080/01441649108716787
URL: http://dx.doi.org/10.1080/01441649108716787

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf
This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.

TRANSPORT REVIEWS,

1991, VOL. 11, No. 3, 249-272

Transport and urban form in thirty-two of the

world's principal cities
By PETER W. G. NEWMAN

Associate Professor in Environmental Science, Murdoch University,

Murdoch, Western Australia
and JEFFREY R. KENWORTHY

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

Research Fellow, Institute for Science and Technology Policy,

Murdoch University, Murdoch, Western Australia
A study of 32 major world cities shows that there are very clear relationships
between transport and urban form. Economic factors such as income and petrol
price are less important than the direct policy instruments of the transport planner
and urban planner, such as the relative provision of infrastructure for automobiles
and rapid transit, or the density of population and jobs. Transport and urban
planning policies are developed with quantitative guidelines that can help cities ease
their dependence on the automobile, for example, by increasing population
densities where these are under 30 per hectare.

1. Introduction

This paper presents some of the findings from an Australian Government funded
study of transport and land use in 32 of the world's principal cities. The major purpose
of the study was to establish policies at the urban level for reducing transport energy
use. However, there are so many social, economic and environmental implications in
the relationships between transport and urban form, that the project has significance
for many aspects of urban and transport policy. Thus for this paper it is possible at the
outset to suggest that the following aims are considered to be desirable outcomes from
improving transport and urban form:
lessening the vulnerability of a city to oil supply disruptions thus improving its
sustainability in energy terms;
minimizing the effect of transport-related inflation and the national balance of
payments due to imported oil;
reducing the level of dependence on the private car;
improving the balance between public and private transport, thus reducing the
public transport deficit;
increasing the amount of non-motorized transport i.e. walking and bicycling;
improving the level of accessibility to the transport disadvantaged i.e. the elderly,
children, poor people and handicapped, who cannot use a car;
reducing the quantity of emissions including those that contribute to the
'greenhouse effect' and to smog;
0144-1647/91 $3.00 1991 Taylor & Francis Ltd.

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

250

P. W. G. Newman and J. R. Kenworthy

lessening the possibilities of road accidents; and

enhancing several less quantifiable variables concerning the 'human' aspects of a
city, especially the attraction of the central city.
The full list of parameters chosen for this study are summarized in table 1. The data
covering 1960, 1970 and 1980 were collected over a five-year period, principally by
personal visits to each city in 1983 and 1986, and a long series of letters and telephone
calls. The utmost care was taken to ensure the data were supported by relevant
transport and planning authorities in each city, and numerous internal consistency
tests were carried out to check the reliability of the data.
Considerable work was done to ensure the parameters were defined in a
comparable way in each city and in particular, urban land-use data were defined to
exclude all rural land uses such as farms, forests, undeveloped land and large bodies of
water. The raw data and standardized data (reduced to transport planning parameters)
amounting to over 100 data items per city is now available complete with detailed
comments on source material, assumptions and computations in a book (Newman and
Kenworthy 1989).
It should be stressed that the great majority of parameters collected are not readily
available from published sources but must be sought out and compiled from detailed
documents within government agencies. A major part of this work is often the
definition of the relevant area of the city (e.g. inner area) and the compilation of data
such as population and employment to match the definition. The book discusses these
issues in detail and provides clear descriptions of how the data on each city were drawn
together from many diffuse and unrelated sources.
This paper will review the main transport and land-use patterns in the 32 cities for
1980. For many of the parameters Moscow is not available and hence the total is
generally made up of 12 European, 10 American, 1 Canadian, 5 Australian and 3 Asian
cities. An analysis of the patterns in the ten US cities has been published (Newman and
Kenworthy 1988 a). The major questions being assessed by the study which will be
examined here are:
How much variation is there in the transport and land use of the world's major
cities?
How closely does transport relate to land use in these cities?
Table 1. Parameters collected in world cities study for 1960, 1970 and 1980.
Population, urbanized area and employment for central business district (CBD), inner area and
total city
Parking (on-street and off-street) in CBD
Length of road network in whole city
Passenger cars and total vehicles on register
Total annual VKT (vehicle kilometres of travel) by passenger cars and other vehicles
Average vehicle occupancy
Average speed of travel in the road traffic network
Total annual gasoline consumption and diesel consumption for whole city
Journey to work modal split (%) and other modal split data where available
Average trip lengths (km) for the journey to work and other trips
Annual vehicle kilometres, passengers carried, average travel distance of passengers, average
speed of travel and annual energy consumption for all bus, train, tram and ferry operations
(including publicly and privately operated transit services)

Transport of the world's principal cities

251

How does automobile usage relate to the provision of infrastructure for

automobiles?
Are economic factors such as income and petrol price so dominant in determining
transport patterns that transport infrastructure and urban form cannot be
addressed directly for urban policy?
What are the direct policy implications from the study?
These questions have been the subject of many academic debates as well as soulsearching within transport and planning authorities. This study is able to throw some
light on the issues as it is probably the first time there has been a reliable set of urban
data for this number of cities with such an international cross-section.

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

2. Results and discussion

Table 2 presents the main transport variables emphasizing the modal split.
Passenger km of private car use and public transport use enable a real comparison of
the relative importance of these modes and the percentage of workers bicycling and
walking to work gives some idea of the relative priority of these non-motorized modes.
Petrol use gives an overall feel for the transport system in each city by acting as a kind of
barometer which rises with increasing automobile emphasis.
Table 3 sets out the parameters which relate to the provision of infrastructure for
automobile usage i.e. how many roads and car parking spaces are provided, the level of
congestion, and the relative speed of the public transport options that are available.
Table 4 provides the main urban form parameters: population and job densities in
the city as a whole and then by central city, inner city and outer area.
3. Transport patterns

3.1. Petrol use

US cities use on average twice as much petrol per capita as in Australian cities, four
times as much as in European cities and ten times as much as in Asian cities. Moscow
which has almost no private car use, manages on a mere 380 MJ per capita which is
nearly two-hundred times less than some US cities.
Toronto uses more than Australian cities but this raises the question about the role
of car size as Toronto private car use is some 8% less than in Australian cities and it has
more than double the public transport usage. Hence if the cities are adjusted for vehicle
efficiency as in table 5 the comparisons are a little more meaningful in purely transport
terms.
The table shows that if all cities had vehicles like US cities then Toronto's petrol use
is less than Australian cities and the variation from US cities to Asian cities would be
reduced from a factor often to a factor of seven, and US to European cities reduces from
four times to three times. Thus this technological factor is a relevant parameter but
clearly is not the dominant factor often described in energy conservation literature (see,
for example, La Belle and Moses 1982, Chandler 1985). Other economic factors will be
discussed shortly to see their contribution to this variation in petrol usage.
3.2. Modal split
The highly automobile-oriented US cities at the top of the list in table 2 have
virtually no public transport as a percentage of their total passenger km of travel e.g.
Houston 0-8%, Phoenix 0-5%, Detroit 0-8%; even in a city like Denver where there is a

P. W. G. Newman and J. R. Kenworthy

252

Table 2. Transport patterns in the world's major cities (1980).

Petrol use
(MJ per capita)

Total vehicles
(per 1000 people)

Car ownership
(per 1000 people)

74510
69908
65978
63466
58474
55365
54185
51241
48246
44033

797
689
691
853
667
681
557
645
518
459

603
499
594
666
542
543
465
561
445
412

15968
13170
14017
11630
13865
13200
12570
11670
11122
7856

128
66
112
218
384
926
518
616
971
1285

0-8
0-5
0-8
1-8
2-7
6-6
40
50
80
141

58541

656

533

12507

522

4-4

32610
30653
29104
28791
27986

614
595
528
568
489

475
458
446
475
412

11477
11721
10128
10625
9450

592
745
779
655
1511

4-9
60
71
5-8
13-8

Average

29829

559

453

10680

856

7-5

Canadian city
Toronto

34813

554

463

9850

1976

16-7

London
Munich
West Berlin
Copenhagen
Vienna
Amsterdam

16671
16093
15709
15574
14744
14091
12426
12372
11331
11106
10074
9171

382
427
432
390
408
383
356
398
306
296
374
342

344
387
375
347
361
338
288
360
269
246
311
308

7470
6810
7254
6570
5706
4199
4452
5235
4 572
6231
4262
4441

1516
1713
2157
2124
1396
1827
1717
1592
2159
1657
1828
1801

170
201
22-9
24-4
19-7
30-3
27-8
23-3
32-1
210
30-0
28-9

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

City
US cities
Houston
Phoenix
Detroit
Denver
Los Angeles
San Francisco
Boston
Washington
Chicago
New York
Average
Australian cities
Perth
Brisbane
Melbourne
Adelaide
Sydney

European cities
Hamburg
Frankfurt
Zurich
Stockholm
Brussels

Paris

Public transport
(passenger km
per capita)

Private car/public
transport balance
(% of total
passenger km on
public transport)

Private car
(passenger km
per capita)

Average

13280

375

328

5 595

1791

24-8

Asian cities
Tokyo
Singapore
Hong Kong

8488
6003
1987

267
155
66

156
65
42

2993
1789
615

5191
1942
2043

63-4
52-1
76-9

Average

5493

163

88

1799

3059

641

USSR city
Moscow

380

40

20

230

t Moscow commuter rail data are missing.

>4262t

>95

253

Transport of the world's principal cities

Table 2 (concluded)

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

City
US cities
Houston
Phoenix
Detroit
Denver
Los Angeles
San Francisco
Boston
Washington
Chicago
New York

Public transport
vehicle km of
service per person

Public transport
passenger trips
per person

Proportion of
public transport
passenger km on
trains

Proportion of
workers using
public transport

(7.)

\/a}

Proportion of
workers using
private transport
(%)

Proportion of
workers using
foot or bicycle
(%)

9
7
17
25
27
50
26
40
42
58

15
9
26
27
59
115
80
91
115
122

00
00
2-7
00
00
33-9
52-3
37-2
66-7
78-0

3-3
2-2
41
6-5
7-7
170
161
141
18-3
28-3

93-9
94-6
931
881
880
77-5
741
80-7
75-5
63-6

2-8
3-2
2-8
5-3
4-2
5-5
9-8
5-2
6-2
81

30

66

27-1

11-8

82-9

5-3

53
48
53
51
77

71
79
95
83
142

14-4
55-8
631
33-5
69-9

120
16-6
20-6
16-5
29-5

840
781
73-7
77-7
651

40
5-3
5-7
5-8
5-4

Average

56

94

47-3

190

75-7

5-2

Canadian city
Toronto

81

178

40-2

31-2

630

5-8

80
55
62
119
54
47
120

69
74

248
306
363
302
266
259
284
307
395
201
313
345

64-4
51-3
55-6
60-4
43-2
83-8
63-7
611
58-3
480
26-6
45-5

410
190
340
460
26-7
39-8
390
42-0
370
310
44-9
140

43-9
540
450
34-0
57-7
36-4
380
380
480
36-8
40-4
580

15-3
270
210
200
15-6
23-8
230
200
150
32-2
14-7
280

Average

79

299

554

34-5

44-2

21-3

Asian cities
Tokyo
Singapore
Hong Kong

94
98
116

472
353
466

94-9
00
17-1

590
59-6
62-2

161
24-6
3-3

24-9
15-8
34-5

Average

103

430

37-3

60-3

14-7

25-1

USSR city
Moscow

>131

>678

740

20

240

Average
Australian cities
Perth
Brisbane
Melbourne
Adelaide
Sydney

European cities
Hamburg
Frankfurt
Ziirich
Stockholm
Brussels
Paris
London
Munich
West Berlin
Copenhagen
Vienna
Amsterdam

75
83

no

>75

Downloaded By: [University of Cambridge] At: 14:26 19 November

Table 3. Provision for the automobile in the world's major cities (1980).
Parking spaces

Average speed of public transport (km/h)

Road supply
(m/person)

(per 1000 CBD

workers)

Average speed of
traffic (km/h)

Total vehicles
per km of road

10-6
10-4
5-8
9-4
4-5
4-9
5-2
51
50
4-7

370
1033
473
498
524
145
322
264
91
75

51
42
44
45
45
46
39
39
41
35

76
66
119
107
158
140
112
127
103
99

6-6

380

43

13-3
6-9
7-9
91
6-2

562
268
270
380
156

8-7

Toronto

2-7

European cities
Hamburg
Frankfurt
Zurich
Stockholm
Brussels
Paris
London
Munich
West Berlin
Copenhagen
Vienna
Amsterdam

2-2
20
2-6
2-3
1-7
0-9
1-9
1-7
1-5
4-3
1-7
21

City
US cities
Houston
Phoenix
Detroit
Denver
Los Angeles
San Francisco
Boston
Washington
Chicago
New York
Average
Australian cities
Perth
Brisbane
Melbourne
Adelaide
Sydney
Average

Car kilometres per ~

km of road

Bus

Train

Tram

Ferry

Total system

939428
818455
1714024
1002509
1989979
1923096
1586674
1562228
1505631
1267248

22
23
21
21
21
22
18
18
18
15

42

45
45
40
47
35

15
20

25

20

22
23
22
21
21
29
30
26
37
31

HI

1430927

20

42

18

23

26

43
48
48
43
39

46
94
67
64
82

497392
930096
740564
658970
870836

22
23
21
21
20

35
37
33
45
45

18
28

14

23

24
31
28
29
37

p
z
s

327

44

71

739572

21

39

23

19

30

198

204

2262597

20

34

16

25

149
242
140
153
186
201
130
285
438
212
190
208

30
30
36
30
?
28
31
35
28
45
30
39

171
214
165
171
246
410
186
238
208
69
216
161

1974143
2136111
1647922
2128 378
2376794
2997666
1 321401
1961237
1850153
810087
1539623
1378489

22
22
20
25
20
13
18
20
20
24
19
18

36
44
46
36
38
45
38
55
32
54
38
57

17
15
26
17

17

17
15

Canadian city

I
<I

12

11

31
37
33
32
27
40
31
44
27
38
23
36

Average

21

211

30

205

1843 500

20

43

18

12

33

Asian cities
Tokyo
Singapore
Hong Kong

1-9
10
0-2

66
97
37

21
30
21

140
158
290

1122092
727886
1 518142

12
19
15

40

31

13

10

14

38
19
17

Average

10

67

24

196

1122707

15

36

12

14

25

0-4

45

93

281895

21

41

18

>37

USSR city
Moscow

T)

CI

o-t

Downloaded By: [University of Cambridge] At: 14:26 19 November

Table 4. Urban form in the world's major cities (1980).

Whole city density
City
US cities
Houston
Phoenix
Detroit
Denver
Los Angeles
San Francisco
Boston
Washington
Chicago
New York

Population

Jobs

Central city density

Inner area density

Population

Jobs

Population

Jobs

Outer area density

Population

Jobs

Proportion of
jobs in CBD (%)

Proportion of
population in inner
area (%)

Proportion of

Proportion of
population in CBD (%)

jobs in inner
area(%)

9
9
14
12
20
16
12
13
18
20

6
4
6
8
11
8
6
8
8
9

6
17
11
19
29
90
126
8
16
217

443
67
306
263
472
713
383
584
938
828

21
19
48
19
30
59
45
44
54
107

26
24
20
17
14
48
33
38
26
53

8
8
11
10
18
13
10
11
11
13

4
4
5
5
9
5
4
6
5
6

01
0-5
01
0-4
01
11
2-7
01
0-1
2-8

11-6
3-9
6-6
11-6
4-8
170
15-9
161
12-3
22-9

16-6
3-7
31-6
30-9
31-3
21-3
24-3
21-4
42-3
39-5

411
10-6
29-6
49-7
43-3
344
34-6
32-5
44-9
41-9

14

54

500

45

30

11

0-8

12-3

26-3

36-3

11
10
16
13
18

5
4
6
5
8

8
15
25
8
11

121
346
647
251
434

16
19
29
19
39

15
16
40
25
39

10
9
16
12
16

3
3
4
4
5

0-7
0-3
0-2
0-2
0-1

24-1
13-9
15-2
14-4
13-2

22-9
21-7
90
11-6
16-7

510
45-7
33-2
37-3
39-3

14

13

360

24

27

13

0-3

16-2

164

41-3

Toronto

40

20

25

757

57

38

34

14

0-2

13-4

35-7

47-9

European cities
Hamburg
Frankfurt
Zurich
Stockholm
Brussels
Paris
London
Munich
West Berlin
Copenhagen
Vienna
Amsterdam

42
54
54
51
67
48
56
57
64
30
72
51

24
43
33
34
42
22
30
34
27
16
38
23

26
65
44
97
74
235
66
HI
133
85
65
108

407
389
422
280
592
400
397
231
333
325
403
153

88
63
79
58
101
106
78
159
84
59
133
83

106
74
66
62
85
60
62
192
46
38
113
46

35
49
42
46
50
26
48
48
57
24
59
32

12
25
17
16
16
8
19
21
20
11
23
10

0-7
2-5
0-9
6-4
1-9
5-4
2-7
5-9
0-8
2-2
1-3
9-7

200
184
13-6
26-3
24-6
20-2
29-7
20-5
4-8
160
14-9
29-9

26-8
43-3
474
49-3
51-8
60-9
37-2
21-4
31-8
37-3
31-9
59-2

560
64-2
65-2
74-7
75-9
751
55-1
42-9
41-8
44-8
50-8
71-7

Average
Australian cities
Perth
Brisbane
Melbourne
Adelaide
Sydney
Average

Canadian city

~
to"

Average

54

31

92

361

91

79

43

17

3-4

19-9

41-5

59-9

Asian cities
Tokyo
Singapore
Hong Kong

105
83
293

66
37
110

82
204
160

477
339
1259

153
202
1037

114
?
478

58
63
224

20
?
66

1-3
6-6
0-4

26-6
24-3
7-3

32-3
35-2
300

84-8
?
45-3

Average

160

71

149

692

464

296

115

43

2-8

194

32-5

651

139

155

3-7

USSR city
Moscow

pri
a
2.

<-*

256

P. W. G. Newman and J. R. Kenworthy

Table 5. Adjusted average 1980 petrol use per capita in cities by region to account for vehicle efficiency
(relative to US vehicle efficiencies, using national values and adjusted for average speed in cities).
Petrol use per capita
with US vehicle
efficiency (MJ)

Average vehicle Adjusted for

Unadjusted petrol
efficiency
average speed use per capita (national values)
in cities
National
(MJ)
(1/100 km)
(1/100 km)
values

City

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

US cities
Australian cities
Toronto
European cities
Asian cities

58 541
29829
34813
13280
5493

15-35
12-50
16-30
10-66
7-63

19-33
15-33
21-72
16-38
15-05

58 541
33446
32784
19123
11051

Adjusted for
average speed
58541
37612
30982
15727
7248

Note 1: Adjustments for average speed are made by using y=l-0174x + 37-4291, where j = fuel
consumption in ml/km and x is the inverse of average speed in s/km (Kenworthy and Newman 1982) and
national fuel efficiencies are assumed to be at an average speed of 60 km/h.
Note 2: Detailed data on vehicle efficiencies are contained in table 4.1 of Newman and Kenworthy (1989).

strong policy to encourage bus usage due to the smog, only 1-8% of total passenger
travel is by public transport. It is only in the US cities with rail systems that any
significant proportion of transport is by non-automobile modes e.g. San Francisco 7%,
Chicago 8%, New York 14%, (the proportion of total transit passenger km by trains in
these cities is San Francisco 34%, Chicago 67% and New York 78%). At the same time
the bicycling/walking proportion for journey to work trips rises (up to 10% in Boston, 6
to 8% in others).
Australian cities overall are a little less automobile oriented, though Perth (5%
transit, 4% bicycle/walking) is virtually an average US city. Sydney with 14% public
transport use is the most non-car oriented Australian city with once again a high
proportion on rail (70% of transit). Toronto is significantly different to its North
American neighbours with 17% transit use (in particular the comparison with its
nearest neighbour Detroit at 0-8% is quite stunning).
European cities on average have 25% public transport use for the total passenger
transport task (passenger km) and for the work journey 21% of trips are by
bicycling/walking. This ranges from 17% public transport in Hamburg to 32% in West
Berlin and 30% in Paris and Vienna; for bicycling/walking to work Copenhagen at 32%
and 28% in Amsterdam are the best. In the European cities 55% of public transport
passenger km are on trains. On average, people in US cities travel nearly 7000 km
further by car and nearly 1200 km less by public transport than in European cities.
Among other things this suggests urban travel distances are shorter in Europe and in
fact work journey average distances are 30 to 40% shorter in European cities (8 km)
compared with US (13 km) and Australian cities (12 km).
All these comparisons are even more striking when the Asian cities are examined
where 64% of the transport task is by public transport and 25% of people go to work by
walking or biking (35% in Hong Kong). In the modern metropolis of Tokyo only 16%
of the people use a car to go to work and in the public transport system 95% of
passenger km are by train.

Transport of the world's principal cities

257

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

3.3. Congestion and public transport speeds

The very clear pattern distinguishing automobile dominated cities from those with
significant public transport use (particularly rail) can be related in purely transport
terms to how easy it is to travel by car and how the transit option competes in time. The
data in table 3 show the automobile based cities to have average traffic speeds of
43km/h (US) and 44km/h (Australia) compared with European cities 30km/h and
Asian cities 24 km/h. On the other hand, the bus-only cities of the US provide little
competition for cars with 21 to 23 km/h transit speeds. Only the rail option can
compete with cars as the average speed of urban trains is 42 km/h in the US, 45 km/h in
Sydney, 43 km/h in Europe and 40 km/h in Tokyo (compared with 21 km/h for cars).
Tram speeds are much lower but they act usually as distributors in central areas linking
in to the major train stations (Vuchic 1981), and typically operate with very high
passenger loadings especially compared with buses. It is also interesting that the
average speed of buses in US, Australian and European cities as well as Toronto and
Moscow is 20 to 21 km/h, a remarkably constant figure considering the enormous
diversity in urban conditions in these cities. In the very much denser and congested
Asian cities it drops to 15 km/h. It would thus appear that, in general, bus-based public
transport systems seem to have an in-built limit on operating speed of no more than
25 km/h, and thus cannot be considered genuine competitors in speed to the car in any
city. It could be concluded that any city seriously wishing to change the private
car/public transport equilibrium in favour of public transport, must move in the
direction of rail-based systems.

3.4. Road supply and parking

Table 3 also looks at how cities provide for their transport modes in terms of road
supply and central city parking. Here again the automobile cities of the US and
Australia provide around three to four times as much road per capita as in European
cities and nearly seven to nine times as much as in Asian cities. Central city parking
does not have quite such a large variation with the US cities having some 80% more
spaces per 1000 workers than European cities and six times that provided in the three
Asian cities. Perth is the outstanding city in the sample as far as automobile provision is
concerned with by far the highest road supply per capita and a central city parking
provision second only to Phoenix which does not in fact have a true central city area.

4. Urban form patterns

4.1. Total density

The main parameter describing the form of a city is its density which has significant
effects on travel distances and modal split (e.g. Pushkarev and Zupan 1977). The overall
shape of the US and Australian automobile city is of low density in population and jobs
with European cities generally being three to four times more dense. Newer cities like
Houston, Phoenix, Perth and Brisbane have densities around half that of the older
cities like Chicago, New York and Sydney. Toronto tends to be more like a European
city in its overall urban form. The Asian cities are again even more extreme with
densities some ten times those of the US and Australian cities. Hong Kong is by far the
highest density city in the sample and probably in the world.

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

258

P. W. G. Newman and J. R. Kenworthy

4.2. Central city density

One of the significant differences between the US/Australian automobile cities and
the more transport balanced European and Asian cities is that the former have central
cities which have become areas of very high job concentration with generally few
residents and the latter have a much better balance between central city jobs and
residences. The central city high rise office block is a characteristic mainly of the
automobile city and it gives to US cities much higher average central city job
concentrations than in Europe. In fact, the average job density profile of the US city is
extremely sharp going from 500 per ha in the central city to 30 per ha in the inner city
and 5 per ha in the outer areas, compared with European cities which have 361 per ha,
79 per ha and 17 per ha. On the other hand the residential density of US and Australian
central cities is generally less than 20 per ha (except for New York, Boston and San
Francisco), whilst in Europe they average around 90 per ha and in Asia about 150 per
ha. One of the questions to examine in the correlations between transport and urban
form is whether this extreme pattern ofjob densities in US and Australian cities has any
significant effect on balancing automobile use patterns as public transport is favoured
by strong central cities (Thomson 1977) or whether the residential density pattern is
more dominant.
4.3. Inner city and outer area density
As well as being different in overall density there are clear differences in urban form
between US/Australian cities and European/Asian cities in terms of their inner cities
and their outer areas.
The US/Australian inner city is generally two to three times less dense than in
European, and ten times less than that in Asian cities. However, the old inner cities of
San Francisco, Washington, Boston, Chicago, Sydney and particularly New York are
similar to many European cities (as is Toronto), while the inner cities of newer US and
Australian cities are generally little more than their overall density. This confirms the
generally accepted picture of older cities as having steeper population density gradients
(e.g. Clark 1982).
The outer area densities of US and Australian cities are amazingly uniform in all
cases with very low land-use intensity. European cities are marked by much more
intensively utilized outer areassome four times more on average than in US and
Australian cities. Toronto is again more like a European city in its outer area; from
observation it appears to develop its density partly through a number of intensively
utilized sub-centres linked by rapid transit to the city centre. Tokyo's outer areas are
similar to European outer areas although Hong Kong is once again at the extreme in
land-use intensity with an overall outer area density greater than the central cities of
some US and Australian cities.
The variations in total density and in inner/outer area patterns suggest that these
may be highly significant in determining the overall transport patterns. Thus the next
section examines the correlations between transport and urban form as revealed by our
data.
5. Transport and urban form correlations

Table 6 provides the linear correlations between the transport and urban form
variables and table 7 provides the linear correlations between the transport variables so
that the question of petrol use and private car dependence can be linked to the degree of
provision for the automobile. More sophisticated statistical analysis of the data is

Downloaded By: [University of Cambridge] At: 14:26 19 November 2

Table 6. Transport and urban form correlations.

Urban
density

Job
density

CBD
population
density

CBD job
density

Inner area
population
density

Inner area
job density

Outer area
population
density

Outer area
job density

Proportion of
population
in CBD {%)

Proportion of
jobs in
CBD <%)

Proportion of
population in
inner area (%)

Proportion of
jobs in inner
area <%)

Petrol use
(MJ per capita)

-0-6099
s=0000

-0-6627
s=0-000

-0-4827
5=0-003

-00301
5=0-436

-0-3914
s=0-015

-0-4849
5=0-003

-0-5752
s=0-000

-0-5913
5=0-000

-0-4810
s=0-003

-0-5067
5 = 0-002

-0-4561
5=0-005

-0-6412
s=O000

Total vehicles
(per 1000 people)

-0-7619
5=0-000

-0-7649
5=0-000

-0-6523
5=0-000

-0-2325
s=0-104

-0-5930
s=0000

-0-6524
s = 0-000

-0-7177
s=0-000

-0-7322
s=0-000

-0-4726
s=0-003

-0-3864
s=0-016

-0-4349
s=0-007

-0-5031
s=O002

Car ownership
(per 1000 people)

-0-7801
s=0-000

-0-7792
s=0-000

-0-6350
s=0-000

-0-20OO
s=0-140

-0-6129
5=0-000

-0-6758
5=0-000

-0-7305
s=0000

-0-7501
s=0-000

-0-4395
5=0-006

-0-3637
s=0-022

-0-3503
s=0-027

-0-4879
5=0-003

Private car
(passenger km per capita)

-0-7438
5=0-000

-0-7793
s=0-000

-0-6698
5=0-000

-01360
5=0-233

-0-5409
5=0-001

-0-6204
s=00O0

-0-6931
5=0-000

-0-7164
s=0-000

-0-5379
s=0-001

-0-4694
s=0-004

-0-5186
s=0-001

-0-6214
3=0000

Public transport
(passenger km per capita)

+0-5234
s=0-001

+0-6773
5=0-000

+0-3959
s=0-014

+0-1452
5=0-218

+0-2927
5=0-055

+0-3675
5=0-023

+ 0-4602
s=0-005

+ 0-4897
5=0-003

+0-2981
5=0-052

+0-4500
s=O006

+0-4185
5=0-010

+0-6935
s=0-000

Private car/public transport

balance (% of total passenger
km on public transport)

+0-8537
5=0-000

+0-9077
5=0-000

+ 0-5948
s=0-000

+0-3155
s=0-042

+0-7405
5=0-000

+0-7810
s=0-000

+ 0-8402
s=0-000

+0-8458
s=0-000

+ 0-3549
5=0-023

+ 0-2902
5=0-057

+ 0-3639
s=0-022

+ 0-5352
3=0-001

Public transport vehicle km of

service per person

+0-5785
5=0-000

+0-6205
5=0-000

+0-3879
s=0015

+01762
s=0-172

+0-4365
5=0-007

+ 0-4850
s = 0-003

+0-5893
s=0-000

+0-5975
s=0-000

+0-3772
s=0-018

+0-4813
5=0-003

+0-3333
s=0-033

+ 0-4660
5=0-005

Public transport passenger

trips per person

+ 0-7390
5=0-000

+ 0-8414
s=O000

+0-5314
5=0-001

+ 01574
5=0-199

+ 0-5289
s=0001

+ 0-6151
5 = 0-000

+0-7185
s=0-000

+ 0-7564
s=0-000

+ 0-4515
5=0-005

+ 0-3893
s=0-015

+0-5179
s=0-001

+0-6653
5 = 0-000

Proportion of public
transport passenger
km on trains (%)

+ 00863
s=0-319

+0-1287
s=0-245

+0-3275
s=0-034

+ 0-1683
s=0-183

-00921
s=0-311

+ 0-0257
5=0-446

-00277
5=0-441

+0-0295
s=0-439

+0-2317
s=0-101

+0-4543
5 = 0-005

+0-3186
s = 004

+ 0-4788
s=0-004

Proportion of workers using

public transport (%)

+ 0-7143
s=0000

+0-7835
s=0-000

+0-5725
5=0-000

+ 0-2989
s=0-051

+ 0-5582
s=0-001

+ 0-6661
5=0-000

+ 0-6941
s=0-000

+ 0-7044
s=0-000

+0-3244
s=0-037

+0-4000
s=O013

+0-3554
s = 0-025

+0-5667
3=0001

Proportion of workers using

private transport (%)

-0-7615
s=00O0

-0-8395
s=0-000

-0-6114
s=0-000

-0-2378
s=0099

-0-5963
+0000

-0-6883
5=0-000

-0-7387
s=0-000

-0-7606
5=0-000

-0-4359
5=0-007

-0-4334
5=0-007

-0-4685
5=0-004

-0-6239
s=0-000

Proportion of workers using

foot or bicycle (%)

+0-6636
s=0-000

+ 0-7387
s=0-000

+ 0-5349
s=0-001

+0-0725
3=0-349

+0-5218
s=0-001

+ 0-5767
s=0-000

+0-6417
s=0-000

+ 0-6900
s=00O0

+ 0-5306
s=0-001

+ 0-3898
5=0-015

+ 0-5577
s=0-001

+0-5843
s=0000

&O

3-

1
1

Petrol
use

Total
vehicle
ownership

Car
ownership

Private car
(pass, km
per capita)

Public trans,
(pass, km
per capita)

Private car/
public trans,
balance

Public trans,
vehicle km
per person

Public trans,
pass, trips
per person

Prop, of public
trans, pass.
km on trains

Prop, of
workers using
public trans.

Prop, of
workers using
private trans.

260

Downloaded By: [University of Cambridge] At: 14:26 19 November

Table 7. Intercorrelations between the transport variables.

Petrol use
Total vehicle
ownership

+ 0-8950
5=0-000

Car
ownership

+ 0-8555
s=0-000

+0-9813
5=0-000

Private car (pass,

km per capita)

+0-9185
5=0-000

+0-9437
s=0000

+ 0-9258
5=0-000

Public trans, (pass,

km per capita)

-0-7328
5=0-000

-0-7328
s=0-000

-0-7450
5=0-000

-0-7633
s=0-000

Private car/public
trans, balance

-0-7340
5=0-000

-0-8889
s=0-000

-0-9205
5=0-000

-0-8746
s=0-000

+0-8046
s=0-000

Public trans, vehicle

km per person

-0-8305
s=0-000

-0-8256
5=0-000

-0-8082
5=0-000

-0-7981
s=0-000

+ 0-7026
5=0-000

+0-7347
s=0-000

Public trans, pass,

trips per person

-0-8750
s=0-000

-0-8908
s=0-000

-0-8712
s=0000

-0-9234
s=0-000

+0-8551
5=0-000

+ 0-8995
s = 0-000

+0-7635
s=0-000

Prop, of public trans,

pass, km on trains

-0-5169
s=0001

-0-4633
s=0004

-0-3828
s=0015

-0-4452
5=0-005

+0-5858
5=0-000

+ 0-3312
5=0-032

+0-4101
s=0-011

+0-4043
5=0-012

Prop, of workers
using public trans.

-0-8216
s=0-000

-0-8775
5=0-000

-0-8775
s=0-000

-0-8835
s=0-000

+0-7906
s=0-000

+ 0-8721
5=0000

+ 0-8098
s=0-000

+ 0-8547
s=0-000

+ 0-4153
s=0010

Prop, of workers
using private trans.

+ 0-8831
s=0000

+0-9339
s=0-000

+0-9218
s=0-000

+ 0-9430
5=0-000

-0-8148
s=00O0

-0-9120
s=0-000

-0-8459
s=0-000

-0-9244
s=0-000

-0-4336
s=0-007

-0-9536
s=0000

Prop, of workers
using foot or bicycle

-0-7822
s=0000

-0-8117
s=0-000

-0-7809
s=0000

-0-8241
5=0-000

+ 0-6635
s=0-000

+ 0-7654
5=0-000

+0-7083
s=0-000

+0-8279
5=0-000

+ 0-3634
s = 0-022

+0-6434
s=0-000

-0-8441
s=0-000

Road
supply

+0-7081
5=0-000

+ 0-7737
s=0000

+0-7026
s=0-000

+0-7854
5=0-000

-0-6621
5=0-000

-0-6744
s=0-000

-0-6257
s=0000

-0-8205
s=0-000

-0-4563
s=0-004

-0-7559
5=0-000

+0-8033
5=0-000

Central city
parking

+0-5775
s=0-000

+0-5935
0-0000

+0-5271
s=0-001

+0-5574
s=0-001

-0-5795
s=0-000

-0-5394
s=0-001

-0-6238
s = 0-000

-0-5694
s=0-000

-0-5699
s=0-000

-0-6546
s=0-000

+0-6472
s=0-000

Average speed
of traffic

+ 0-6340
5=0-000

+0-6502
5=0-000

+0-6232
5=0-000

+ 0-7212
5=0-000

-0-7737
5=0-000

-0-5660
s=0-001

-0-6472
s=0-000

-0-8601
s=0-000

-0-3465
s=0-030

-0-8438
s=0-000

+ 0-8510
s=0-000

Total vehicles
per km of road

-0-4661
s=0-004

-0-4093
s=0-010

-0-3437
s=0-027

-0-5317
s=0-001

+0-4038
s=0-012

+0-3547
5=0-023

+0-3143
s=0-042

+0-6219
s=0-000

+ 0-2006
5=0-135

+0-5736
s=0-000

-0-6173
s=O-0O0

Car km per km
of road

-0-0733
s=0-345

+0-0059
s=0-487

+ 01145
5=0-266

-0-0675
5=0-357

+ 0-1951
s=0-146

-01138
s=0-268

+00508
5 = 0-393

+0-3098
5=0-045

+ 01739
s=0-170

+0-2436
s = 0-093

-0-2667
5=0-073

Average speed of
buses

+ 0-2362
s=0-096

+0-3437
s=0-027

+0-3402
5=0-028

+0-4156
s=0-009

-0-5043
s=0-002

-0-4005
s=0-011

-01553
s=0-202

-0-4244
s=0-009

-0-3508
s=0-024

-0-4298
s=0-008

+0-4304
s=0-008

Average speed of
trains

-00636
s=0-376

+0-0174
5=0-466

+0-0557
5=0-391

-00163
5=0-468

-00108
5=0-479

-00997
s=0-310

-00767
s=0-355

+ 00153
s=0-470

+0-0849
s=0-337

-0-2367
s=0122

+ 00317
5=0-439

Average speed of
trams

+ 0-2533
5=0-181

+0-3386
s=0-076

+0-4255
s=0-056

+ 0-4145
s=0-062

-0-4124
5=0-071

-0-4417
s=0-049

-01964
5=0-250

-0-5527
s=0-020

+ 00290
s=0-459

-0-3562
s = 0-015

+ 0-4550
s=0-051

Average speed of
ferries

+0-7962
s=0-014

+0-6123
5=0-069

+ 0-6095
s=0070

+ 0-6297
5=0-062

-0-4504
s=0-153

-0-4483
5=0-154

-0-5159
s=0-115

-0-6257
s=0-064

+01908
s = 0-340

-0-4915
s=O129

+ 0-5196
s=0113

Average speed
overall

-0-3708
5=0-020

-0-2481
5=0-091

-01594
s=0-196

-0-2503
s=0087

+ 0-4083
5 = 0011

+0-0521
s=0-0390

+0-2428
s=0094

+0-2633
s=0-076

+ 0-8079
s=0-000

+01623
s=0191

-0-2698
s=0-071

?
to

3
P

Downloaded By: [University of Cambridge] At: 14:26 19 November

Table 7 (concluded)
Prop, of
workers using
foot or bicycle

Road
supply

city
parking

Average
speed of
traffic

Total
vehicles per
km of road

Car km
per km
of road

Average
speed of
buses

Average
speed of
trains

Road
supply

-0-6960
5=0-000

Central city
parking

-0-4800
s=0-003

+ 0-6489
s=0-000

Average speed
of traffic

-0-6797
s=0-000

+0-6983
s=0-000

+0-4555
s=0-006

Total vehicles
per km of road

+ 0-5474
s=0-001

-0-7127
s=0-000

-0-3537
5=0025

-0-6975
s=0000

Car km per km
of road

+0-2438
5=0093

-0-5166
s=0-001

-0-2498
5=0-088

-0-4479
5=0-006

+ 0-7941
s=0-000

Average speed of
buses

-0-3287
5=0-035

+0-4005
5=0-012

+0-4331
5=0-007

+ 0-5686
5=0-001

-0-4814
5=0-003

-0-2210
s=0-112

Average speed of
trains

+0-2913
s=0-074

-00773
5=0-351

-0-0417
5=0-120

+0-2740
5=0-092

-0-0245
5=0-452

+00031
s=0-494

+ 00632
s=O-377

Average speed of
trams

-0-4638
5=0-047

+ 0-5586
s=0-015

+0-6753
s=0004

+0-3702
5=0-106

-0-4890
s=0-032

-01230
5=0-331

+ 0-6368
s = 0005

+0-0436
s=0-439

Average speed of
ferries

-0-5458
s=0100

+ 0-2380
s=0-303

-0-3846
5=0195

+0-6853
s=0-042

-0-4678
5=0-143

-01532
5=0-371

+ 00236
s = 0-480

+0-8810
s=0004

Average speed
overall

+0-3974
5=0-013

-0-2795
s=0-064

-0-3509
s=0026

-01560
s=0-209

+0-1453
s=0-218

+ 0-2206
5=0-116

-01467
5 = 0-215

+0-6980
5=0-000

Average
speed of
trams

Average
speed of
ferries

Average
speed
overall

STO

I
O

+ 0-1302
s=00328

+0-5392
s=0-103

K)
Os
hi

P. W. G. Newman and J. R. Kenworthy

262

currently underway but these correlations provide sufficient basis for clarifying and
confirming the patterns already discussed.

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

5.1. Urban form

The correlations suggest that strong negative relationships exist between petrol use
or private vehicle use and all the density variables, except central city job densities. This
highlights the question of the US and Australian cities with their central city high rise
office blocks. There is no correlation between central city job density and any of the
transport variables (including public transport use) suggesting that this factor has little
overall effect on transport patterns despite the apparent importance of peak hour CBD
oriented public transport activity in these cities. It would appear to be more important
to have higher residential densities mixed in with the employment activity if there is to
be much less dependence on the automobile. Residential density in the central city does

Dilrati

Denvor

9 San Francisco
Boston

t Washington DC

Chicago

"3.

I New York

s
g_ 40,000

9!

100

125

150

175

200

Urban density (persons per hectare)

Figure 1. Petrol use per capita versus urban density (1980).

275

300

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

Transport of the world's principal cities

263

correlate strongly with all the transport patterns, including the amount of
walking/bicycling. The case of Boston highlights this as it is the highest US city for
bicycling/walking to work (10%) and it has for its 72000 central city residents the
highest population to jobs ratio (0-33) for US CBDs, thus it is more like a European city
in this regard. It is not hard to see that city centres with plenty of employment activity
that also have high residential densities (e.g. Paris 235 per ha) would have a significantly
higher proportion of people walking and having little need for a car.
The relationship between density and petrol use may be more complex than a
purely linear linkage. Figure 1 suggests that it may in fact be closer to an exponential
relationship particularly under 30 or so people per hectare. This is conceptually quite
possible as a city with density in the less than 30 per ha range does not just have longer
distances for all types of journey, it is ensuring that modes other than the automobile
are not feasible because of the sheer lack of people living near a transit line and the time
required for walking and biking. Thus the effects of lowering density are multiplicative.
This cut off around 30 per ha we have also found to be significant for transport within
different parts of urban areas (Newman and Hogan 1987). It means that in terms of
transport energy saved or private car use curtailed the effects of increasing density can
be considerable if they move urban areas into at least the 30 per ha density range i.e.
more like the old inner area densities.
The significance of the inner area (with its mixture of jobs and residences at medium
densities) as a model for directing policy in transport and urban form is highlighted by
transport data collected in a few US cities on an inner/outer area basis. Table 8 shows
that the New York Tri State region petrol usage per person is 44 030 MJ, however for
the inner area residents (City of New York) this reduces to 20120 MJ and for the 1-4
million residents of Manhattan, their average petrol consumption drops to an
extraordinary 11 860 MJ. The average outer area New Yorker consumes 59 590 MJ
and in Denver it is possible to distinguish the 240 000 ex urban residents who live on the
fringe of the city and consume some 137 000 MJ on average. The linkage to density
would appear to be very strong.

5.2. Provision for the automobile

There is a strong correlation between petrol consumption and provision for the
automobile in terms of road supply and parking. Also the significant positive
correlation between average speed and petrol use highlights one of the traffic
management controversies.

Table 8. Petrol use and urban density by city region in the New York Tri State Metropolitan
Area, 1980.
Area
Outer area New York
Whole city (New York Tri
State Metro Area)
Inner area (City of New York)
Central city (New York County
including Manhattan)

Petrol use
Urban density
(MJ per capita) (persons per hectare)
59590
44033

13
20

20120
11860

107
251

Downloaded By: [University of Cambridge] At: 14:26 19 November 2

Table 9. Average value for per capita petrol use in cities by region 1980, compared with adjusted values (for US petrol prices,
incomes and vehicle efficiency).
Adjusted petrol use for
US petrol prices, incomes
and vehicle efficiency
(MJ]per capita)
Cities

Actual petrol use

(MJ per capita)

Short-term
elasticities

% Difference between
US petrol use and
adjusted petrol use by
other cities

Long-term
elasticities

Short-term
elasticities

Long-term
elasticities

ho
hfl

JS

. P
p
Z
z
CD
et

US cities
Australian cities
Toronto
European cities
Asian cities

58541
29829
34813
13280
5493

58541
38488
29995
17082
7676

58 541
43 680
26090
31080
12340

51%
49%
71%
87%

25%
55%
47%
79%

Average for non-US cities

17133

21450

31160

63%

47%

JS

Note 1: Petrol consumption elasticities used were: petrol price 0-20 short-term, 10 long-term; incomes + 0 1 1 short-term,
+ 0-6 long-term.
Note 2: As petrol consumption elasticities include a component due to vehicle efficiency, it is necessary to subtract this when
adjusting other cities for US vehicle efficiencies otherwise it would be accounted for twice.
Vehicle efficiency elasticities used were: petrol price +0-11 short-term, +1-0 long-term; incomes 0-11 short-term, 10 longterm.
Vehicle efficiencies used were national values adjusted for average speed in each city (table 5). In all cases vehicle efficiencies in the
long-term became more than equivalent to US levels and hence the vehicle efficiency factor in the long-term is cancelled out.
Source: Pindyck (1979), Dahl (1982), Archibald and Gillingham (1981) and Wheaton (1982).

a
p
13
Q-

o.
<i
<i

S
o

Transport of the world's principal cities

265

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

There is clearly less petrol use in cities with low average speeds which contradicts
those traffic planners who suggest freeing up congestion to increase average speeds will
save petrol (see Newman and Kenworthy 1984,1988 b, c). Although free flowing traffic
may improve individual vehicle efficiencies, the evidence suggests that it also causes
overall fuel consumption increases presumably due to greater private vehicle use.
Before examining the policy implications of these relationships it is necessary to
show to what degree the patterns are explained by economic parameters.

6. Economic factors versus physical planning factors

The ten US cities in which petrol use per capita varies some 40% were examined first
to see if there were any direct economic relationships. No significant correlation was
found with per capita income (01219) or petrol price (0-1765) which varies a little
across the nation. Thus transport patterns appear to be clearly related to other noneconomic factors in this US sample. In the global sample there are significant
correlations between petrol use and price ( 0-8500), income per capita (0-7477) and
average vehicle fuel efficiency (1/100 km) (0-5906). The question is thus how much do
these economic factors influence transport.
To compare across the international sample requires the use of accepted price and
income elasticities for petrol consumption together with vehicle efficiencies which also
vary with petrol price. Table 9 sets out the expected petrol consumption if US incomes,
US petrol price and US vehicle efficiency were found in all the 31 cities (excluding
Moscow), i.e. if these economic factors were the sole or primary factors then all the cities
should have the same petrol use. They clearly do not. On average, the economic factors
explain at the most around half the petrol use. It can be argued that these long-term
elasticities are overestimates as they incorporate some degree of anticipated urban
form change but even they do not adequately explain the variations in petrol use in the
sample. What is suggested by these results is that a purely economic approach to
transport matters will be inadequate, that matters of urban form and provision for the
automobile have direct and independent influence on transport patterns. Thus
planners who provide the transport infrastructure or who set out the physical plan of a
city are directly and actively influencing transport patterns, they are not just
responding to economic factors.

7. Policy implications
As outlined in the introduction, if a city is to move towards less transport energy
use, this will almost certainly imply less use of the private car and more use of public
transport, bicycling and walking; it will more than likely also mean greater accessibility
by the transport disadvantaged, less emissions and road accidents and most probably a
more 'human' city especially in the central city (Schaeffer and Sclar 1975, Newman and
Hogan 1981, Kenworthy 1986). The analysis followed so far will be further developed in
terms of what transport planners and urban planners can do to assist cities to move in
this direction. In particular, the assumption will be how to shift the automobile
dependent cities of the US and Australia into something more like the transport
balance found in European cities (and Asian cities), though the notional numbers
suggested in the policies tend to be less extreme than those in European cities and are
more like those found in Toronto (and in other Canadian cities).

266

P. W. G. Newman and J. R. Kenworthy

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

7.1. Transport infrastructure policies

As the data suggest, the provision of transport infrastructure has a direct effect on
the balance of transport modes and travel distances, thus it would seem appropriate to
directly address the provision of transport infrastructure.
It would appear from the data presented that the transport patterns (and urban
form) of a city would be shifted more in the direction outlined above if the following
were adopted.
A policy to restrict the amount of road supply within a city to something around 2 m
to 3 m per capita. This would essentially mean curtailing new road projects that
pass through the city. It would depend also on urban form policies outlined
below. In particular, since roughly three quarters of the road kilometrage in a city
is comprised of minor and local streets this would imply a move away from the
single family house with a typical wide street frontage, towards innovative family
housing which shares road access more intensively (and which also has less need
of roads because of its public transport and walking/bicycling orientation). For
example, in Toronto single family housing is around 60% and road provision is
2-7 m per person. According to the Real Estate Research Corporation (1974),
single family housing requires around seven times as much road length as that for
high rise apartments. The money generated from road project savings and from
the selling of major road reserves through cities could be put into traffic
management and some of the other transport projects listed below.
A policy to restrict central city parking to a level around 200 spaces or less per 1000
CBD workers. This would require a concurrent policy that provides good public
transport access and a series of central city policies on housing, cultural
attractions, urban design, pedestrianization and commercial activity that allows
a central city to compete strongly with suburban centres where easy parking is
available (McNulty, Penne and Jacobson 1986).
A policy accepting that average speeds in a city of around 30 km/h are adequate.
This means rejecting the notion that fuel is saved by increasing average speeds.
A policy that provides a rapid transit option (most likely to be rail) which is
substantially faster than the average traffic speed in the city and together with other
improvements slowly builds up public transport in stages so that it provides
something more like 20 to 30% of total passenger km. This would be a considerable
change for many cities where the present situation has public transport at less
than 10% and in some cities less than 1%. However, in almost every city examined
by us public transport utilization went up between 1970 and 1980 after decades of
decline, and the indications are that it is on a continued trend upward.
A policy that encourages pedestrianization, traffic-calming treatment of streets and
bicycle facilities so that the proportion of work journeys by bicycling and walking
rises to something more like 20% rather than its present 5% in automobile cities.
This policy like the others is not really very contentious, however such policies are
unlikely to be successful unless they are wedded to the following land use policies.
7.2. Urban form policies
Each of the transport infrastructure policies outlined will have an influence on
urban form and should not be seen as separate. In the same way, the following major

Transport of the world's principal cities

267

urban form policy directions are a necessary adjunct to the transport infrastructure
policies as outlined.
The primary urban form policy theme is Reurbanization, a policy emphasized
mainly in Europe but with even greater application to US and Australian cities (e.g. van
den Berg et al. 1982,1987). This policy is pictured in the two diagrams (figure 2 and 3)
and in the following detailed policies.

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

A policy to increase by stages the intensity of urban activity overall so that

population densities of around 30 per ha to 40 per ha and job densities of around 20
per ha are obtained This will mean an immediate policy of restricting or at least
slowing urban development at the urban fringe and concentrating on
redevelopment; this consolidation generally has the added benefit of considerable
capital savings due to the better use of present urban infrastructure rather than
requiring new infrastructure at the fringes (Wilmoth 1982, Bunker 1983).

Figure 2. Conceptual plan for re-urbanization.

Post
ww 11

1960s

1970s

Time

Figure 3. Stages of urban development (after van den Berg et al. 1982).

268

P. W. G. Newman and J. R. Kenworthy

A policy to build up the central city activity intensity so that job densities are
maintained at more than 300 per ha and population densities are built up to over 50
to 60 per ha. The provision of housing in central city areas of automobile cities
seems to be harder than providing for jobs though some outstanding successes
have been achieved in recent years in places like Boston, Toronto and San
Francisco.

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

A policy to build up or maintain the inner area at population densities of 40 to 50 per

ha and job densities of similar levels. Most of the older automobile cities have inner
cities with these densities, however in many cases, particularly in the US, these
have been declining in recent decades. Policies to contain inner city decline
appear to have been relatively successful in Sydney, Melbourne and Toronto. In
newer automobile cities like Houston, Denver, Brisbane, Adelaide and Perth
there is enormous development potential in their inner areas.
A policy to build up outer area urban activity to population densities of around 20 to
30 per ha and job densities of around 15 per ha. As outlined in figures 2 and 3 the
way that this is most likely to be effective in transport and land use terms is (a) to
slowly expand the present inner area type of development (i.e. mixed and more
intensive) into the outer area, and (b) building up densities around rapid transit
routes. For low density cities like Brisbane and Adelaide with present rail
systems, and in cities like Los Angeles and Perth which are building new lines, this
policy would appear to be of primary importance. Washington and Toronto are
good examples where this policy has made major changes in transport and urban
form in less than a decade.

8. Conclusion
A range of transport infrastructure and urban planning policies have been
developed for those cities with an excessive dependence on the automobile. This paper
does not attempt to outline the implementation of such policies, it merely attempts to
suggest what are feasible goals for automobile-oriented cities. The time scale for such
changes could possibly be gauged from cities like Detroit and Los Angeles which over a
30 to 40 year period were transformed from being compact rail-oriented cities to
dispersed automobile-oriented cities. The re-urbanization process could arguably be
even faster with modern technology and more rapid turnover of buildings. The forces of
dispersal and greater automobile reliance are obviously very powerful and to overcome
them cities must first begin by recognizing what is possible (Newman 1988). This study
has attempted to set out the patterns of transport and urban form which currently exist
in more balanced cities and which could thus be emulated in substance, but in style and
manner fitting to the social, cultural and environmental conditions peculiar to each
city.

Acknowledgments
This work was partially funded by the Australian Department of Primary Industry
and Energy through the National Energy Research Development and Demonstration
Council. Their support is gratefully acknowledged.

Transport of the world's principal cities

269

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

Foreign summaries
L'examen de trente deux villes parmi les plus importantes de la plante rvle la relation
troite entre l'organisation du transport et la morphologie de la ville. Des facteurs conomiques
tels que les niveaux de revenus et le prix des carburants marquent moins l'espace que les
instruments de l'urbaniste et de l'organisateur de transport, tels que la cration d'infrastructures
routires ou ferroviaires, ou la densit d'emploi et de rsidence. Les politiques de planification
urbaine et de transport ont leur propres rgles quantifies qui peuvent soulager les villes de leur
dpendance l'gard de l'automobile, par example, en accroissant la densit l o l'habitat est
infrieur trente par hectare.
Eine Studie ber 32 Hauptstdte der Welt zeigt, da es sehr deutliche Zusammenhnge
zwischen dem Verkehr und der Stadtform gibt. Wirtschaftliche Faktoren wie Einkommen und
Kraftstoffpreise sind weniger bedeutend als die direkten Manahmen der Verkehrs- und
Stadtplaner wie z.B. das relative Infrastrukturangebot im individuellen Personen- und
Schienenschnellverkehr oder die Dichte von Einwohnern und Arbeitspltzen. Verkehrs- und
Stadtplanung handen nach quantitativen Mastben, die dazu beitragen knnen, da die Stdte
sich von der Abhngigkeit vom Auto lsen, indem sie beispielsweise die Einwohnerdichten dort
anhaben, wo sie unter 30 Einwohner je Hektar liegen.
Un estudio de las treinta y dos ciudades ms importantes del mundo muestra que existen
relaciones muy claras entre el transporte y la forma urbana. Factores econmicos, tales como el
ingreso y el precio del petrleo, tienen menor importancia que los instrumentos directos de
poltica de los planificadores de transporte y urbanistas, tales como la provisin relativa de
infraestructura vial y de transporte pblico, o la densidad de poblacin y empleo. Las polticas de
planifiacin de transporte y urbanismo se desarrollan en base a direccionamientos
cuantitativos que pueden ayudar a aliviar la dependencia en el automvil de las ciudades, por
ejemplo, mediante el incremento de las densidades poblacionales en sitios en que stas sean
menores a 30 habitantes por hectrea.

References
and GILLINGHAM. R., 1981, Decomposition of the price and income elasticities of
the consumer demand for gasoline. Southern Economic Journal, 47 (4), 1021-1031.
BUNKER, R., 1983, Urban Consolidation: the experience of Sydney, Melbourne and Adelaide,
Australian Institute of Urban Studies, Publication No. 111, Canberra.
CHANDLER, W. U., 1985, Energy Productivity: Key to Environmental Protection and Economic
Progress, Worldwatch Paper 63, Worldwatch Institute, Washington DC.
CLARK, C, 1982, Regional and Urban Location, University of Queensland Press, St Lucia,
Australia.
DAHL, C. A., 1982, Does gasoline demand elasticities vary? Land Economics, 58 (3), 373-382.
KENWORTHY, J. R., 1986, Transport Energy Conservation through Urban Planning and Lifestyle
Changes: some fundamental choices for Perth, edited by C. McDavitt. Towards a State
Conservation Strategy: Invited Review Papers, Department of Conservation and
Environment, Western Australia, Bulletin 251, pp. 69-99.
KENWORTHY, J. R., and NEWMAN, P. W. G., 1982, A Driving Cycle for Perth: Methodology and
Preliminary Results. Presented to Joint SAEA/ARRB Second Conference on Traffic,
Energy and Emissions, Melbourne, May 19-21, Paper 82149.
LA BELLE, S. J., and MOSES, D. O., 1982, Technology Assessment of Productive Conservation in
Urban Transportation (ANL/ES-130), Energy and Environmental Systems Division,
Argonne National Laboratory.
MCNULTY, R. H., PENNE, R. L., JACOBSON, D. R., and Partners for Livable Places, 1986, The
Return of the Livable City (Washington DC: Acropolis Books).
NEWMAN, P. W. G., 1988, Australian Cities at the Crossroads. Current Affairs Bulletin,
December, 4-15.
NEWMAN, P. W. G., and HOGAN, T. L. F., 1981, A Review of Urban Density Models. Human
Ecology, 9 (3), 269-302.
ARCHIBALD, R.,

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

270

P. W. G. Newman and J. R. Kenworthy

NEWMAN, P. W. G., and HOGAN, T. L. F., 1987, Urban Density and Transport: A simple model
based on 3 city types. Transport Research Paper 1/87 pp.36, Environmental Science
Murdoch University.
NEWMAN, P. W. G., and KENWORTHY, J. R., 1984, The Use and Abuse of Driving Cycle Research:
Clarifying the Relationship between Traffic Congestion, Energy and Emissions. Transportation Quarterly, 38, (4), 615-635.
NEWMAN, P. W. G., and KENWORTHY, J. R., 1988 a, Gasoline Consumption and Cities: A
comparison of US Cities with a Global Survey. Journal of the American Planning
Association, 55 (1), 24-37.
NEWMAN, P. W. G., and KENWORTHY, J. R., 1988 b, The Transport Energy Trade-Off: FuelEfficient Traffic versus Fuel-Efficient cities. Transportation Research, A, 22 (3), 163-174.
NEWMAN, P. W. G., and KENWORTHY, J. R., 1988 c, Does Free Flowing Traffic Save Energy and
Reduce Emissions in Cities? Search, 19 (5/6), 267-272.
NEWMAN, P. W. G., and KENWORTHY, J. R., 1989, Cities and Automobile Dependence: An
International Sourcebook (Aldershot, UK: Gower).
PINDYCK, R. S., 1979, The Structure of World Energy Demand (Massachussetts, Cambridge: MIT
Press).
PUSHKAREV, B. S., and ZUPAN, J. M., 1977, Public Transportation and Land Use Policy
(Bloomington and London: Indiana University Press).
REAL ESTATE RESEARCH CORPORATION, 1974, The Costs of Sprawl, Detailed Cost Analysis, US

Goverment Printing Office, Washington, DC.

SCHAEFFER, K. H., and SCLAR, E., 1975, Access for All (Harmondsworth: Penguin).
THOMSON, J. M., 1977, Great Cities and their Traffic (England: Penguin).
VAN DEN BERG, L., DREWETT, R., KLAASSEN, L. H., ROSSI, A., and VIJVERBERG, C. H. T., 1982,

Urban Europe: a Study of Growth and Decline (Oxford: Pergamon Press).

VAN DEN BERG, L., BURNS, L. S., and KLAASSEN, L. H., 1987, Spatial Cycles (England, Aldershot:

Gower).
VUCHIC, V. R., 1981, Urban Public Transportation Systems and Technology (New Jersey,
Englewood Cliffs: Prentice Hall).
WHEATON, W. C , 1982, The long-run structure of transportation and gasoline demand. Bell
Journal of Economics, 13 (2), 439-454.
WILMOTH, D., 1982, Urban Consolidation Policy and Social Equity, in Conference on Urban
Consolidation and Equity, Centre for Environmental Studies, Macquarie University.

Editorial suggestions for further reading

CER VERO, R., 1984, Light rail transit

and urban development, Journal of the American

Planning Association, 50 (2), pp. 133-147.

Recent construction of light rail transit systems in a number of North American

cities raises crucial questions about their possible effects on land use and urban
development. Although serving passengers and keeping construction costs down have
been the primary aims of new rail investments, the possibilities for joint development
and land use are numerous. This paper explores light rail transit's potential influence
on urban growth and revitalizing central city areas. Some cities are integrating light rail
transit with pedestrian malls as part of downtown redevelopment. A significant number
of others, however, are downplaying the development potential of light rail transit by
aligning their systems principally along abandoned railroad rights-of-way and
industrial belts in order to cut costs. For most cities in the preconstruction stages of
their projects, policy-makers need to recognize the trade-offs involved when the lowestcost corridor and alignment are chosen. On the whole, the land-use potential of light
rail is moderately high, where there are pro-development policy environments and
other complementary forces.
(Author)

Transport of the world's principal cities

271

R., 1988, Use of models by French consultants for Urban Transport

planning in developing countries, Transportation Research Record, 1167,
pp. 41-45.

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

DARBERA,

This paper is based on a survey of 39 urban transport planning studies performed

by French consultants in developing countries and on a more detailed analysis of three
major cases. Presented are the local conditions that have influenced the choice of the
models, the difficulties encountered, the final outcomes of the studies, and the extent to
which the plans have actually been implemented. Unless they were constrained by terms
of reference set by some international agency, the consultants often took shortcuts
through the traditional methodology. They tended to favour straightforward
unsophisticated models in which complex equations with many variables are replaced
by normative inputs and in which land use forecasts are exogenous or simply taken for
granted. In this approach, the role of transport demand modelling is only to ensure that
the planned land use would not lead to an unbearable burden on the transport
investment budget and, conversely, that planned transport infrastructures were the
most suitable for the planned land use. In most cases, simple normative models proved
to be quite sufficient and guaranteed good transparence and readability of the
mechanisms at work. A more pressing issue is the problem of accuracy in projections of
exogenous variables such as population, income, and automobile ownership. French
consultants who are confronted with this problem now tend to include modelling
packages and training programs with their studies. In this way, they locally install the
capabilities to follow up the urban transport situation and to monitor the implementation of the plans.
(Author)

L. H., and VAN DER MEER, J., 1987, Urban change and public transport,
International Journal of Transport Economics, 14 (2), pp. 123-132.

KLAASEN,

In recent years city centres have seen a decline in population and, to a lesser extent,
jobs. The increasing spread of population has led to severe peak hour congestion
problems as people travel to and from work, with accompanying social, economic and
environmental costs. However, recently there has been an increase in the
deconcentration of employment, and rising demand for part-time jobs, which may lead
to more efficient use of the transport system. The paper argues that this tendency could
be reinforced by policy moves to emphasize the role of the city centre as a location for
high grade services, which could lead to more balanced transport demands and a
gradual change in peak hour demand.
(Authors)

P. W. G., and KENWORTHY, J. R., 1988, The transport energy trade off: fuel
efficient traffic versus fuel efficient cities, Transportation Research, A, 22 (3),
pp. 163-174.

NEWMAN,

Improving fuel efficiency in vehicular traffic by increasing average speeds is shown

to have a major trade-off through land-use changes and modal shifts that result in an
overall loss in fuel efficiency for the total urban area. In Perth, even though vehicles in
central areas have a 19 per cent lower fuel efficiency than average due to congestion, the
central area residents still use 22 per cent less actual fuel on average due to their
locational advantages. On the other hand, outer suburban traffic is 12 per cent more

272

P. W. G. Newman and J. R Kenworthy

efficient than average but residents use 29 per cent more actual fuel. A comparison of 32
world cities confirms that there is a trade-off between fuel efficient traffic and fuelefficient cities. The implications for traffic engineering programmes and road funding
are discussed.
(Authors)
M. R., 1988, Changes in the relationships between transport, communications
and urban form, Transportation, 14 (4), pp. 395-417.

Downloaded By: [University of Cambridge] At: 14:26 19 November 2008

WIGAN,

Transport, communications and urban form cover an overlapping area of rising

academic and practical concern. This paper traces several of the many themes brought
together under different professional banners, and shows how a confluence of interest is
emerging.
The themes are the developments in urban planning analysis, transport and time
use studies, telecommunications and industrial location, all of which contribute to the
area of locational effects of improved telecommunications technology, and are affected
by alterations in the nature of work and the uses made of time by individuals. The
convergence of geographical, planning, transportation and communication
developments now requires explicit investigation, as the timings of technological and
theoretical developments appear to coincide with the emergence of a significant need to
do so. Some of the gaps between present knowledge and expertise that need to be filled
are specified.
These include the investigation of the testing and use of current land use integrated
analyses for road and development assessment, investigation of the nature and
characteristics of work and education which are most affected by telecommunications
and computer support, development of longitudinal monitoring methods for overall
urban development leading indicators, exploitation of newly-available cross-sectional
household and city data sets in conjunction with historical data for longitudinal
investigations and forward projections, accounting for altered family structure and
activity patterns and the anticipation and assessment of probable further technological
change, which can and will undermine many current long term commitments.
(Author)
Readers are also referred to the ISGLUTI series published in Transport Reviews:
ECHENIQUE, M. H., FLOWERDUE, A. D. J., HUNT, J. D., MAYO, T. R., SKIDMORE, I. J., and
SIMMONDS, D. C , 1990, The MEPLAN models of Bilbao, Leeds and Dortmund,

Transport Reviews, 10 (4), pp. 309-322.

R. L., 1990, The systematic application of the LILT model to Dortmund,
Leeds and Tokyo, Transport Reviews, 10 (4), pp. 323-338.
MACKETT, R. L., 1991, A model-based analysis of transport and land-use policies for
Tokyo, Transport Reviews, 11 (1), pp. 1-18.
MACKETT, R. L., 1991, LILT and MEPLAN: a comparative analysis of land-use and
transport policies for Leeds, Transport Reviews, 11 (2), pp. 131-154.
PMJLLEY, N. J., and WEBSTER, F. V., 1991, Overview of an integrated study to compare
models and evaluate land-use and transport policies, Transport Reviews, 11 (3),
pp. 197-222.
WEBSTER, F. V., and PAULLEY, N. J., 1990, An international study on land-use and
transport interaction, Transport Reviews, 10 (4), pp. 287-308.
WEGENER, M., MACKETT, R. L., and SIMMONDS, D. C., 1991, One city, three models: a
comparison of land-use/transport policy simulation models for Dortmund,
Transport Reviews, 11 (2), pp. 107-129.
MACKETT,