SUSTAINABLE TRANSPORTATION AND GLOBAL CITIES

BY PETER NEWMAN
Professor of City Policy Director, Institute for Sustainability and Technology Policy Murdoch University Perth, Australia

Abstract
Introduction
New Data on Global Cities

Car use and wealth

Technology-Vehicle efficiency
Economic Implications
Road expenditure

Transit cost recovery

Transportation deaths
Transport emissions
Proportion of city wealth on transportation
Understanding the patterns
Global information age processes and sustainable transportation
Conclusions
References
Discussion questions

Abstract

The energy, environmental and social benefits of sustainable transportation have always been recognised but are now mainstream as Kyoto and the community politics of transportation indicate. However sustainable transportation has always been seen as secondary in economic terms unless lots of unquantifiable parameters were included. The results of a study for the World Bank now show that cities with significant sustainable transportation systems are least costly in terms of a range of parameters including the amount of road expenditure, transit operating cost recovery, fuel-efficiency, road accidents, air pollution and in overall terms the % of city wealth that goes into transportation. The data show that cities with the most roads hare the most costs for their transportation and the most rail-oriented cities have the least transportation costs. Further, the single most important variable relating to transportation efficiency is the density of the city - the most sprawling cities are the most costly. Thus strategies to contain sprawl, to reurbanise , to traffic calm, to build new light rail systems into car dependent suburbs with focussed sub centres, and to facilitate biking and walking, all appear to add to the economy of a city. Strategies that build freeways and add to sprawl are draining the economy of cities. Global information trends are making the need for these sustainable urban patterns even more necessary. The need to operationalise these strategies in planning and engineering practice and in the politics of infrastructure funding remain the major challenges.

Introduction

Focussing on energy and transport enables us to examine the sustainability of transport systems; energy cuts across the full spectrum of environmental, social and economic aspects of transport.

If the manuals of professional practice in both engineering and town planning are examined it is easy to see a marginalisation of sustainable transportation modes (transit, biking and walking). They are not seen as a key component of the dynamics of the urban economy, they are something you help because you should but the real transportation system is about moving cars and trucks. This paper is about how that professional perception now needs to change from environmental, social and economic perspectives.

Environmentally, sustainable transportation is now mainstream. The sustainable cities movement has brought the environmental movement into the city and focussed their attention on the myriad of local and regional problems associated with automobile dependence (Newman and Kenworthy, 1998). Globally the attention of the world has been focussed on the potential climaxing of oil production and on climate change; the Kyoto agreement now sets all developed nations into a pathway of reducing energy. The biggest single technology causing greenhouse emissions is the automobile and it is the hardest for nations to recognise as the culprit. But as the process of reducing emissions becomes more and more mandated down to local level the need for planners and transportation engineers to find ways of reducing car use will be on the agenda - mainstream.

Socially, the situation with sustainable transportation is similar with the alternative modes very much in the centre of the social agenda. This is for several reasons including the reality that sustainable transportation modes have a stronger element of social justice, that most democratic processes lead to sustainable transportation choices rather than outcomes preferred by transportation bureaucracies, and finally the strength of revival by communities in response to the globalising economy is forcing a more community oriented transportation (see Naisbett 1994).

However, on the economic front the sustainable modes have not been mainstream; the only economists supporting these modes have tended to do so on the basis that there are massive unquantifiables that are left out of standard cost comparisons on the modes. However, ISTP has recently completed a major study for the World Bank which has provided a different perspective - indeed it suggests that sustainable transportation should come in from the cold even on mainstream economic considerations.

New Data on Global Cities

The full data from the World Bank study involved patterns in transportation, infrastructure, landuse, environmental and economic parameters (Kenworthy et al 1997). These have been more fully analysed in our new book along with a more extensive set of transport and land use data on a further 9 cities including 4 from Canada (Newman and Kenworthy, 1998).

A few of the key transportation parameters will be examined before looking in more detail at the economic parameters. The sample of 37 cities are set out in Table 1 below; most data are presented as a summary of the different regions.

The patterns of unsustainable car use are easily seen in the data on private passenger transportation energy use per capita in our global cities sample (Figure 1).

Figure 1 Likewise the patterns of sustainable transportation modes are evident in Figure 2.

We have always found the most persuasive way to explain these patterns is to see the links to the provision of infrastructure for cars (Figure 3) and the associated reductions in urban density (Figure 4) which are found to occur universally. It is not unexpected that most cities end up with about the same commitment of their resources to commuting. It appears to be related to the way commuting times adjust to around 30 minutes on average in all cities independent of how they are provided with transport infrastructure (SACTRA, 1994). Increasing travel speeds means that people on average travel further, they do not save time.

Figure 3

Figure 4

However in professional transport circles the dominant explanation is that such patterns just reflect the nature of the economies involved: those who use more cars and less sustainable modes are wealthier. This is no longer quite so clear.

Car use and wealth

For many years there has been an implicit assumption amongst transport planners, engineers and economists that there is a close link between mobility and wealth (eg Rainbow and Tan, 1993). This leaves very few policy options open to cities for managing growth in car use. However, the data for such assertions tends to be national data and is rather selective.

Below we will examine the link between mobility and wealth by comparing the per capita use of cars in 37 global cities and see how this compares with their per capita city wealth (called Gross Regional Product or GRP ie, the total goods and services in that city/region, which in the US for example is the full SMSA region).

The data on car use and wealth (in 1990 US dollars) are given in Table 1.

Table 1 Car use and Gross Regional Product per capita for 37 global cities, 1990

Correlating the data in Table 1, it is found that there is only a weak positive linear correlation between car use and wealth which only explains 18% of the variance and is therefore not particularly significant in terms of policy implications.

As already outlined, North American and Australian cities have considerably higher car use per capita than European and Asian cities. It is higher than would be expected just considering the level of economic activity or wealth, especially in comparison to the European and developed Asian cities in the sample (ie Tokyo, Singapore and Hong Kong).

The large US cities in this sample have:

• 1.66 times higher car use than the major Australian cities but are only 1.36 times higher in GRP;
• 2.17 times higher car use than Metropolitan Toronto but are only 1.19 times higher in GRP;
• 2.41 times higher car use than the average European city but actually have only 0.85 the level of GRP per capita;
• 7.3 times higher car use than the wealthy Asian cities but have only 1.26 the level of GRP.

Perhaps of even more significance is the comparison between the developing Asian cities of Kuala Lumpur, Surabaya, Jakarta, Bangkok, Seoul, Beijing and Manila and the three wealthy Asian cities of Tokyo, Singapore and Hong Kong: the poorer cities have 108% as much car use but have an average GRP which is only 12% of that in the developed Asian cities. This is even more accentuated in the case of Kuala Lumpur, the most motorised developing Asian city. Kuala Lumpur has 2.7 times the average car use per capita of the wealthy Asian cities, yet only 19% of the per capita GRP.

The car use per capita figures in developing Asian cities in some cases include a reasonable amount of motor cycle use (motor cycle use is also included in other cities but is not as significant). However, this does not fundamentally affect the point being made here, which is that developing Asian cities, despite low levels of wealth compared to their more developed neighbours, are experiencing very much higher levels of private mobility.

Within the US, there is also a significant difference between cities that cannot be explained by simple economic factors alone. For example, New York (the lowest car using US city) has 36% less car use per capita than Houston (the highest car using US city), but is actually 10% higher in GRP.

The fact that transport energy and wealth can be decoupled is good news for sustainability. It is sanguine to be reminded that despite all the massive differences in transportation investment priorities and the large differences in transportation patterns in different types of cities, that cities can be wealthy with considerably less car use than others. Transforming the transportation patterns of a city into one that is sustainable can be achieved without damaging overall economic performance (Serageldin and Barrett, 1993; World Bank, 1996).

Technology-Vehicle efficiency

Technology is often the first factor considered in discussions on fuel use. Obviously the efficiency of vehicles must impact on fuel use, but how important is it in explaining the variation between urban fuel use? The data set out below show it is only a small factor in explaining the variation, if by technology we mean how efficient motor vehicles are; however, if by technology we also consider public transit systems and their associated ridership then we can begin to see some major variations.

It should also be pointed out that the variation between US cities and other cities in terms of fuel efficiency of vehicles has generally diminished between 1980 and 1990, as US vehicle fleets have been downsizing. The 1990 average car fuel efficiencies based on actual 1990 fuel use are provided in Table 3.3 below. What appears to be clear by comparing Tables 3.2 and 3.3 is that all cities have improved their urban car fuel efficiencies. Whereas in 1980 the US cities had cars which operated in city conditions with 1.28 times higher fuel use than the wealthy Asian cities (Singapore, Tokyo and Hong Kong), in 1990 the cars in US cities operated at only 1.04 times higher fuel use than in these same cities.

Despite this flattening out in fuel efficiencies of the urban car fleets between US and wealthy Asian cities, there still remains a very considerable difference in gasoline use per capita (US cities are still 8 times higher than their wealthy Asian counterparts). Obviously, automobile technology cannot be the main factor in this considerable variation.

Likewise, in 1980, US cities had 1.26 times higher fuel consumption in urban cars than in Australia, but in 1990 the situation had changed so that US cities are actually marginally more fuel-efficient in their urban cars than in Australia (1.02 times better, due primarily to Australia having one of the world‬ "!s highest average ages for their car fleet, hence lower penetration rates for new, more fuel-efficient cars). Per capita gasoline use however in US cities is still 1.7 times higher than in Australian cities.

Table 2 Fuel efficiencies of urban cars in global cities, 1990

Through the data that are available on all 46 cities in this study, it is possible to develop an even more detailed picture of actual energy efficiencies by mode of urban transport. These data measure energy efficiency in terms of MJ per passenger km (ie including vehicle occupancy), as opposed to the technological efficiency of the mode as expressed in Tables 2 . These data are summarised here in Table 3.

The data show that energy efficiency by car travel is reasonably similar across all cities and is universally at least less than half the efficiency of transit travel, or even worse when compared to urban rail systems (the only exception is buses in US cities which are comparatively inefficient compared to other cities - see below). North American urban car travel (i.e. US and Canadian cities) is 12 to 15% less fuel-efficient than car travel in Australian and wealthy Asian cities, 33% less efficient than car travel in European cities and 64% less efficient than in newly developing Asian cities.

Table 3 Modal energy efficiencies for regional groupings of cities, 1990

Notes: Rail energy efficiency includes heavy rail, light rail and trams where relevant.

Table 3 shows that fuel efficiency by bus travel varies considerably by region and in comparison to other modes within that region. US bus systems are similar in fuel efficiency to European urban car travel and are 19% less efficient than car travel in newly developing Asian cities. This is mostly because of their low patronage levels, as buses across the world are not very different in technology, though US buses are virtually all equipped with air conditioning. The overwhelming use of standardised large buses for both very low and high density services in US cities also contributes to lower energy efficiency in US bus systems. On the other hand, bus systems in European cities and in places such as Sydney and Toronto, are around 2 to 3 times as efficient as in the US cities. In Asian cities bus travel is around 3 to 4 times as efficient as in US cities with Beijing having an efficiency of 0.15 MJ/pass km which is 17 times more efficient; bus loadings in peak hours in Chinese cities reach 12 persons per square metre (Hu and Kenworthy, 1996). Such data show that the gains in energy to be made from bus transport technology are dwarfed by the possibilities offered through ridership improvements.

Rail modes (trains and trams) are by far the most fuel-efficient motorised transport technology in each regional grouping of cities. The only cities to show exception to this are Perth and Adelaide which in 1990 still had old diesel train systems with low ridership. Perth has since electrified its rail system and is now much more efficient (see Newman, 1992). Apart from these two cities, the others show that rail travel is between 2.5 and 5 times more energy-efficient than buses. Rail energy use reaches a low of 0.06 MJ per passenger km in Manila (0.07 in Beijing), which is some 59 times more energy-efficient than car travel in the US. More typically, rail systems in European cities are 7 times more energy-efficient than car travel in US cities.

Electric rail technology tends to be more energy-efficient because of its speed and capacity which leads to higher ridership. Electric rail also has a demonstrated capacity to induce higher density development around stations due to its ability to bring or take large numbers of people to concentrated nodes of development without harming the pedestrian qualities of an area. Furthermore, electric rail can also be linked to renewable energy (as discussed above) which is a significant advantage as we enter the next century with its reduced oil production.

.

In order to round off this section, Table 4 below summarises modal energy efficiencies in the global sample of cities even further by showing the overall modal averages from all the global cities combined, but separating the rail modes into heavy and light rail, and dividing heavy rail into electric and diesel systems. It also shows the comparative loadings or vehicle occupancies which contribute significantly to the energy efficiency differences between modes.

Table 4 Overall modal energy efficiencies in the global sample of cities, 1990

Note: Rail mode occupancies are given on the basis of the average loading per wagon, not per train. The average occupancy of cars is a 24 hour figure.

These data reveal that urban car travel is on average nearly 2 times as energy consumptive as average urban bus travel, 6.6 times more energy intensive than average urban electric train travel and 3.7 times more than typical light rail or tram system travel. Light rail and tram systems typically operate in environments requiring a lot more stopping and starting than heavy rail, with much closer station spacings than heavy rail. So although their average loading is similar to heavy rail, their energy efficiency is a little poorer.

The data in Table 4 also show that diesel rail is only a little more fuel-efficient on average than an urban bus and that average train wagon occupancies are roughly equal across types; they are on average more than twice that of buses and about 20 times higher than cars. Overall, these data re-emphasise the importance of developing a good backbone of electric rail in cities if energy conservation is to be enhanced. Those cities without such systems are the ones with very high gasoline use.

Economic Implications

The parameters below help to provide some of the detail as to why there is a negative impact on economic performance as a city invests in excessive levels of mobility through automobiles. It will examine direct economic costs such as road expenditure, percentage of GRP spent on the journey-to-work and transit cost recovery. It will then examine the indirect costs due to transport deaths and transport emissions. The detailed data on these items can be found in Kenworthy et al (1997) and Newman and Kenworthy, (1998).

Road expenditure

Road expenditure per capita (Figure 5) follows the pattern of car use and car dependence in the sample of cities, though it does not display such extreme differences (US cities spend $264 per capita each year, Australian $142, Toronto $150, European cities $135, wealthy Asian cities $88 and developing Asian cities spend $39 per capita). There is a higher level of maintenance in North American and Australian cities due to their higher amount of roads per capita, but it is obvious that considerable road building is still occurring in these car-based cities. The sustainability agenda will require a change in these priorities in the future if car dependence is to be eased. It is apparent from the above data and the economic parameters below, that such a change can also constitute a move towards greater transportation efficiency.

Figure 5 Road expenditure in global cities, 1990

Road expenditure in European cities is relatively high as they also have many new areas on their peripheries where a more car-dependent urban form has been created, eg the Copenhagen suburbs and surrounding villages that have been developed into suburbs since the 1940's have densities of 25 and 21 persons per ha and have much greater car use than the old city with a density of 63 persons per ha. Such areas will also require reassessment in the light of the sustainability agenda with a view to redirecting road funds to other modes as part of a strategic plan to reduce car dependence (see Newman et al, 1997).

In developed Asian cities, road expenditure per capita is one-third what it is in US cities, and 50% to 60% of what it is in Australian cities and Toronto. As shown below, it is also the lowest in relation to city wealth.

Road expenditure per capita in newly developing Asian cities appears to be comparatively small in absolute terms, though in Bangkok, Seoul and Beijing, there is evidence of relatively heavy spending on roads compared to other cities in this group ($61 to $72 compared with the average of $39). However, in terms of road expenditure per 1000 dollars of GRP, or in other words, in relation to a city‬ "!s capacity to pay, money spent on roads in these developing cities is high. The figures for all the cities are: $9.84 for the US cities, $7.19 for Australian cities, $6.65 for Toronto, $4.26 for European cities, $4.13 for wealthy Asian cities and $14.76 for developing Asian cities. This latter figure is 1.5 times higher than in US cities, the next highest relative spender on roads. Bangkok is spending $18.56 per $1000 of GRP or 1.9 times US levels and Beijing is spending $46.11 or 4.7 times that in the US.

Transit cost recovery

The indicator of transit cost recovery is one of the most emotionally debated issues of any area of public policy. The data in this survey, which measure operating cost recovery, is one of the first to show a comparative set of numbers from the major cities of the world which has been done on as consistent a basis as is possible. It shows that the percentage transit cost recovery follows very precisely the level of car dependence in the city (Figure 6):

‬ North American and Australian cities average a low 37% and 40% with Toronto standing out at 61% and the most bus-based, low density, car dependent cities of Perth, Phoenix and Houston have a mere 28% and Denver only 19% cost recovery. In such cities, even if fares are set reasonably high, it is difficult to have a high cost recovery because of the inherently higher cost structures of such systems (eg high labour input per passenger kilometre, low occupancy per service unit etc).

Figure 6 Transit operating cost recovery in global cities, 1990

‬ European cities average 54% cost recovery with a variation from 93% in London to 27% in Brussels. It needs to be understood that such variations are not just reflections of inherent economic differences between systems, but are also the result of conscious political choices made by each city as to how much of their public transport expenses they want to recover. London chooses to set high fares and recover almost all their costs (since the Thatcher years), while other cities such as in Germany and Belgium choose to recover a lesser proportion in recognition that roads are also being subsidised. The case of Stockholm with only 33% recovery also reflects a social/political position on the role of transit in the community. Of course, having made a decision to recover a relatively high proportion of transit expenses, it is certainly easier to do so in a city environment which is physically supportive of high transit use and where the quality of transit services enables transit to compete with the car. Thus in London it is extremely expensive to use the underground but it is still the best way to get around.

‬ Asian cities have on average very high transit cost recovery at 105%, with the highest in Hong Kong (136%) and Kuala Lumpur (135%), and the lowest in Beijing at 20% due to its very low fares and high staffing levels. Chinese bus and trolley bus tickets are perhaps the cheapest in the world, the average rate in the early 90s being less than US 0.5 cents per passenger km. This is compared to public transport prices (all modes) in other cities which range from a low of around US 1.7 cents per passenger km in Manila, through averages of about US 6 to 9 cents per passenger km in Australian, US and European cities (Hu and Kenworthy, 1996).

The transit cost recovery debate tends to focus on how to reduce government costs. It often concludes that it would be much cheaper to provide only buses as these have lower capital and sometimes lower maintenance requirements. These data suggest that buses are only effective in transit cost recovery in situations where there are large numbers of captive users, as in newly developing Asian cities such as Manila. The more fundamental way to recover transit costs in developed cities is to influence the form of the city towards a more transit-oriented structure. The role of rail systems in influencing and facilitating this cannot be underestimated.

Transportation deaths

In this section we examine the very real but nevertheless external cost of transportation due to traffic accidents. Many others have done estimates of what these costs actually represent (eg in the US the cost of road accidents was estimated in 1996 as US$150 billion: USA Today, January 3-5, 1997). Here we are just presenting the various patterns of transport deaths in the different cities.

The data show that traffic deaths tend to follow both the degree of automobile dependence and the level of development of the traffic regulatory system (Figure 7). In US cities, despite their highly developed road systems, strictly regulated traffic, and a population generally well-educated in traffic safety issues, traffic deaths are highest of all the regional groupings of cities (14.6 per 100,000 people). This seems to be due to the world's highest level of exposure of the population to auto traffic.

Transport deaths then decline with decreasing car use, though not in a parallel way, due presumably to the level of traffic regulation (Australian cities have 12.0 deaths per 100,000 people, Toronto 6.5, European cities 8.8, wealthy Asian cities 6.6 and developing Asian cities 13.7 deaths per 100,000 people).

Thus in developing cities such as Kuala Lumpur, which are motorising at a very rapid rate with high levels of motor cycle ownership and use and a relatively poorly developed traffic regulatory environment, traffic deaths are also very high at 22.7 per 100,000 people. This is despite the fact that the absolute level of automobile dependence is still very low compared to US and other developed cities. Overall, the newly developing Asian cities have an average transport death rate of 13.7 per 100,000 which is a far worse record than their level of car use would predict

Figure 7 Transport-related deaths in global cities, 1990

Beijing, with 71% of total daily trips by walking and cycling, also has a comparatively low rate of transport deaths compared to other cities, as do most Chinese cities (6.1 deaths per 100,000). A study of seven large Chinese cities suggests a transport death rate of 4.8 per 100,000 (Hu and Kenworthy, 1996). The situation in Chinese cities can however be expected to worsen and perhaps begin to mirror the picture in the other rapidly motorising Asian cities in this sample as more and more traffic begins to mix with the high numbers of pedestrians and cyclists. This is especially true if little or nothing is done to slow down this rate of motorisation or to plan for effective harmonisation of motorised and non-motorised transport.

Overall, the data show how transport deaths decline with car use though not to the same magnitude as the differences in car use; Australian cities have 18% fewer transport deaths per 100,000 people but 40% less car use per capita than US cities, European cities have 40% fewer deaths than US cities but 59% less car use and wealthy Asian cities have 55% fewer deaths but 86% less car use. As suggested above, there are therefore other factors at work which lend themselves to reducing transport deaths such as traffic engineering, management and education. However, there are enormous resources and human energy poured into road safety when by far the biggest gains would be made by shifting to other modes and reducing the overall level of car use. This approach is rarely mentioned in road safety discussions.

There are some exceptional cities in terms of the patterns of transportation-related deaths:

• Metro Toronto at 6.5 deaths per 100,000 has less than half the traffic fatalities found in US cities which suggests that a good transit system can have other flow-ons in terms of traffic safety, eg fewer teenagers need to drive. Metro Toronto's transport death rate seems to be reasonably consistent with its other features (eg 24% of total travel on transit, compared to only 4% in US cities).
• Amsterdam at 5.7 and Copenhagen at 7.5 deaths per 100,000 have among the lowest rates in Europe and also one of the highest usage rates of bicycles. This puts into perspective the perception that cycling is dangerous, perhaps indicating that the social patterns developed in a city to accommodate cyclists (such as giving priority to them at all intersections) can flow on to a generally safer road system. The case study on Copenhagen has managed to reduce its traffic accident rates through an emphasis on bicycling and a ‬ ܂culture of respect‬ "! for all non-motorised travellers.
• Tokyo and Hong Kong have among the best transport safety records at 5.3 and 5.7 per 100,000 due to their exceptional transit systems which appear to be far more important in determining overall transport safety levels than their congested major road systems.

Transport emissions

Carbon dioxide

Carbon dioxide is now a focus of international agreement on Greenhouse gas reduction strategies, with all developed cities needing to show how they are reducing CO₂. Many documents have been presented on the issue at international forums, but invariably the area that is seen to be the least amenable to reduction is transport CO₂ (OECD/ECMT,1996; McKenzie and Walsh, 1990). The data here give some idea as to how progress can be made.

First, it is not just a matter of making technological improvements, as has already been shown. More fuel-efficient vehicles can just be used more, particularly if road conditions are improved to create freer flowing traffic. An integrated transportation strategy is required which simultaneously improves technology, facilitates modal shifts and reduces the need for travel. That this is possible without harming city economies is clear. The large variation in US cities with respect to CO₂ generation rates shows some indication of this (total transportation CO₂ per capita varies from 3,778 kg per capita in the New York region up to 5,193 kg in Houston), but the fact that Toronto has 46% less CO₂ per capita than the average US city suggests that its CO₂ generation rate in transportation can serve as a best practice indicator in North American cities.

Toronto is providing transportation at a rate of 0.108 kg of CO₂ per dollar of GRP compared to 0.160 kg per dollar for US cities (48% higher than in Toronto). Australian cities can do much better as well with 0.141 kg of CO₂ per dollar of GRP. European and wealthy Asian cities may be approaching world best practice at 0.059 and 0.054 kg of CO₂ per dollar of GRP. The newly developing Asian cities at 0.317 kg of CO₂ per dollar of GRP need to do better, though their apparently high rate of CO₂ emissions per dollar of GRP is probably mostly due to their much lower wealth.

Figure 8 summarises CO₂ emissions per capita for the global cities in 1990, showing the contribution from private and public passenger transportation. In all cases CO₂ from transit is very small relative to that from automobiles.

Figure 8 Per capita CO₂ emissions from private and public transportation in global cities, 1990.

Regionally significant automotive emissions

The major automotive emissions of concern to health and regional air pollution, including photochemical smog precursors, presented in terms of NOx, SO₂, CO, volatile hydrocarbons (VHC) and particulates, follow the same patterns as car use, with a few interesting exceptions Figure 9):

Figure 9 Per capita emissions of smog-related air pollutants in global cities, 1990

‬ Australian cities are almost identical in per capita air pollutant emissions to US cities, despite having 40% less car use per capita. This is presumably because the vehicle fleet is very old due to lower wealth, very lax systems of vehicle inspections, and because there are lower emissions standards on new vehicles than in the US (see Newman et al, 1996). Policy debates continue to emphasise traffic management as a solution to air pollutant emissions. Australian urban traffic congestion is probably amongst the lowest in the world, as suggested by the data in Newman and Kenworthy (1989 and 1998) on average speeds; this shows how minimal is the factor of smooth traffic flow in reducing emissions, compared to the sheer amount of vehicle use and the state of the vehicles themselves. US cities have even higher average traffic speeds than in Australia, but with very high per capita transport emissions, again emphasising the futility of trying to tackle automotive air pollution through improvements to traffic flow.

‬ Toronto is low in CO₂ due to its transit system and integrated land use (see Kenworthy and Newman, 1994), but it is only an average North American city in other emissions. This is probably again due to a vehicle factor, as its fleet is older and it has the least fuel-efficient cars in North America at 4.38 MJ/pass km, compared to an average of 3.51 MJ/pass km for the US cities.

‬ European city air pollutant emissions are as expected much lower in general than in cities in North America and Australia (57% of the level of NOx per capita in North American cities, 36% of the CO, 52% of the VHC and 63% of particulates). SO₂ is 20% higher however, due presumably to the higher amount of electricity (and hence coal) used in powering transit and the higher share of diesel fuel in the transport system.

‬ Asian cities for the most part have the lowest per capita air pollutant emissions. The exception here is Bangkok, which for its relatively low level of motor vehicle use, has very high volatile hydrocarbons (23.2 kg per capita compared to similar levels in US and Australian cities with much higher vehicle use; Bangkok is also much higher than the European cities at 11.6 kg per capita). In addition, Bangkok has by far the highest particulates in the world (9.1 kg per capita compared to a little over 1 kg per capita in most other cities).

Both these pollutants are linked to health problems. VHC is primarily from very inefficient, poorly maintained vehicles which are often idling for hours in traffic jams, Bangkok being towards a global extreme in these problems. Particulates mainly come from poorly tuned diesel buses and trucks, as well as two-stroke motor cycles, and such vehicles are very common in Bangkok (they are also common in Jakarta and Surabaya where particulate emissions are also comparatively high). It is not surprising that Bangkok traffic police wear gas masks and that there are increasing air pollution-related health problems in this city (see Kenworthy, 1995).

Proportion of city wealth on transportation

A final parameter that in many ways brings together this perspective on the economics of automobile dependence is the percentage of GRP on transportation. It is a sum of all the direct costs attributable to private and public transportation which is then expressed as a proportion of the city‬ "!s wealth. It shows ‬ ܂how much‬ "! passenger transportation-related goods and services are as a proportion of total goods and services in the city.

The data are summarised in Figure 10 and show a similar perspective to that already shown: except it is perhaps even more extreme than what many would have expected.

Figure 10 The proportion of city wealth on transportation in global cities, 1990

It shows that those cities with the highest automobile dependence (Australian and US cities) have the least overall economic efficiency in their transport systems.

The cities (in the developed world) with the highest proportion of their wealth going into transportation are: Perth at 17%, Phoenix 16% and Adelaide, Detroit and Denver all at 15%.

The cities (in the developed world) with the least wealth going into transportation are the European and wealthy Asian cities (at 8% and 5%) with their stronger commitment to transit systems. The best North American and Australian cities - Toronto 7%, New York 10% and Sydney 10% - also indicate that transit-orientation is good for a city‬ "!s economy.

Understanding the patterns

The possible mechanism for the extra inefficiency associated with car dependent cities is suggested to be that car dependence:

• creates inefficiencies due to the extra land it consumes,
• the extra costs of infrastructure,
• the direct and indirect costs of the automobile
• the loss of investment associated with traffic dominated urban environments (compared to quality pedestrian friendly urban environments), and
• the opportunity cost due to loss of investment in productive industries instead of investment in unproductive suburb-building (see Newman et al 1997).

In terms of transportation policy the issues become very stark when the parameter of road provision is plotted against % wealth on transportation, and at the same time the amount of rail orientation is plotted against % wealth on transportation (Figures 11 and 12).

Figure 11

Figure 12

The correlations show that cities with the most roads are the least efficient, and the most rail-oriented cities are the most transport efficient. A similar graph can be done with walking/biking versus % GRP transport as in general the more transit there is in a city the more walking and biking there is.

Figure 13 demonstrates the significance of density in this relationship to city wealth (via the link to car dependence).

Figure 13 The total operating costs of passenger transport versus urban density in developed cities, 1990

.

Figure 13 shows that the lower the density of the city the more it wastes of its wealth on transport. And this is probably the basis of the reason why road-based cities are less efficient: providing roads and facilitating cars is the basic mechanism for sprawling a city and this is an expensive way to build a city. It is less efficient in terms of infrastructure and if a city is constantly sprawling rather than reurbanising then there is less capital available for investment in productive innovative aspects of the city (see Frost, 1991 and Jacobs 1984). Therefore in economic and environmental terms the sustainability agenda is clear - contain sprawl, reurbanise and go for sustainable transportation.

The cities that don‬ "!t follow this trend are the developing Asian cities (Bangkok, Jakarta) which are not car dependent, as they are still very dense, but they are car dominated. These cities are pouring their productive financial and human capital into auto-related activity but are not showing much benefit from it. The transit-oriented model of the wealthy Asian cities on the other hand appears to represent the world best practice on how to create wealth and not have car dependence problems.

Global information age processes and sustainable transportation

Finally, a word about the major economic process of our era - the information age. Original pundits when discussing the impacts of information age technology suggested it would help to disperse cities making them more car dependent (Webber, 1963, 1964, 1968). More sophisticated approaches recognised that information technology had the ability to reform cities based on the reduced need for face-to-face interchange in some activities, but the continuing need for some quality human interactions critical to economic and cultural processes (Castells, 1989; Castells and Hall, 1994,). Hall (1997) after several years of being very equivocal on this, now states:

"The new world will largely depend, as the old world did, on human creativity; and creativity flourishes where people come together face-to- face." (p89).

Others have emphasised that 'local milieus' will emerge (Willoughby, 1994) or that local culture will be strengthened as globalised information makes national borders less relevant (Ohmae, 1990; Naisbett, 1994; Sassen, 1991,1994) or that the importance of face-to-face contact will ensure centres emerge as critical nodes of information-oriented production (Winger, 1997).

All cities in our sample our reurbanising but the historic trend that our study has picked up is that all except two have reversed their trend downwards in density; measured as a combination of population and jobs there has been a substantial shift towards increasing the amount of urban activity per hectare (Newman and Kenworthy, 1998). In US cities the trend is to focus activity in outer suburbs through 'edge cities' and the inner areas continue to decline (apart from those cities where the social and racial issues have been resolved). Thus rather than dispersing urban activity the information-based technologies seem to be focussing it.

Reasons for this reurbanisation and concentration in relation to this information era economics would include the approaches outlined above. Thus it seems that for the global information-era in our cities:

1. Professionals require face-to-face interactions for creative project development work.
2. Community always needs face-to-face (youth culture especially is very urban).
3. Face-to-face meeting spaces are part of inner city design, but are lost in automobile city planning.
4. Reurbanisation around human-oriented city spaces is occurring in almost all cities.
5. Deindustrialisation of inner cities is making them even more attractive for human-based work locations.
6. Travel time budgets (1 hr/day) are being exceeded on fringe locations hence busy professionals are locating near work.

Thus the information age seems to be favouring a multi-nodal city where the sustainable transportation modes are increasingly important as they are more able to build the human-based centers critical for the new urban economy. The challenge will be to ensure that such sub-centers occur throughout the city, not just in wealthy enclaves; the role of light rail extensions into car dependent suburbs as a means of creating viable local employment and services centers, seems to be a growing agenda.

Conclusions

The patterns of automobile dependence, based on transportation, infrastructure and land use patterns have been shown to follow a consistent story on the global cities we have studied. Their economic and environmental costs show the same pattern. They suggest some overall conclusions.

1. Cities with substantial commitment to sustainable transportation are doing better economically as well as environmentally. There appears to be no obvious gain in economic efficiency from developing automobile dependence in cities, particularly as it is shown in US and Australian cities. There is no relative gain in GRP per capita or in the percentage of GRP spent on commuting, trip times to work are roughly the same everywhere, transit cost recovery is much worse and road expenditure is higher.
2. There are, on the other hand, significant losses in external costs due to automobile dependence which have clear implications for sustainability. There are much higher levels of per capita car use, energy, emissions, and transport deaths. As the global agenda is focussing increasingly on sustainability, there is an obvious need to address these differences by overcoming automobile dependence.
3. European and wealthy Asian cities appear to have both the most economically-efficient and sustainable transportation systems. However, they will all need to do better in terms of car use, which is growing in most of these cities as well.
4. Rapidly developing Asian cities have considerably less efficient and sustainable transportation systems than would be expected from their levels of wealth. The positive side however is that they still have strongly transit-oriented urban forms, which means that good electric rail systems and more provision for non-motorised transportation, have the potential to rapidly transform them into more sustainable patterns.
5. Rail transit systems, compared to all other motorised transportation, appear to have the best energy efficiency and greatest ability to attract people out of cars, they are the most important factor in the recovery of transit operating costs, they seem to be the catalyst for compact sub-centre development and they make a major contribution to sustainability on all indicators. Transforming cities towards efficiency in both economic and environmental terms would appear to involve good rail systems. This awareness is still not common in transportation practice and indeed is not even a priority in World Bank funding which has had a long history of anti-rail policies.
6. Non-motorised transportation is highly significant in both economic and environmental indicators. Cities which implement plans for improving the contribution of non-motorised transportation are likely to see immediate and long term benefits.

Overall there is little reason for sustainable transportation modes to be left out in the cold. They ought to be in the centre of transportation decision making. This is becoming even more so as the information age economy develops.

The challenge then is for professional practice to develop the techniques which can allow a greater orientation to sustainable transportation - new cost-benefit techniques and land use-transportation models, new funding mechanisms for rail using value capture techniques, new community sensitive processes... And the political process needs to make funding a more integrative, regional planning process with an eye to the long term.

It is time for sustainable transportation modes to come in out of the cold.

References

Castells, M. and Hall, P. (1994) Technopoles of the world. Routledge, London.

Castells, M. (1989) The informational city: Information technology, economic restructuring and the urban regional process. Blackwell, Oxford.

Frost, L. (1991) The new urban frontier: Urbanisation and city building in Australasia and the American West. University of N.S.W. Press, Sydney

Hall, P. (1997) Reflections past and future in planning cities. Australian Planner, 34(2), 83-89.

Hu, G. and Kenworthy, J. (1996) A preliminary study of land use and transportation patterns in Chinese cities: Caging the automobile dragon. Paper presented to the Asia Research Centre, Murdoch University, Western Australia.

Jacobs J Cities and the Wealth of Nations, Penguin, Harmondsworth (1984).

Kenworthy, J. (1995) Automobile dependence in Bangkok: An international comparison with implications for planning policies. World Transport Policy and Practice, 1(3), 31-41.

Kenworthy, J., Laube, F., Newman, P. and Barter, P. (1997) Indicators of transport efficiency in 37 cities. Report to World Bank, ISTP, Murdoch University, Western Australia.

McKenzie, J. J. and Walsh, M. P. (1990) Driving forces: Motor vehicle trends and their implications for global warming, energy strategies and transportation planning. World Resources Institute, Washington D.C.

Naisbett, J. (1994) Global paradox: The bigger the world economy, the more powerful its smaller players. Allen & Unwin, Sydney.

Newman, P. and Kenworthy, J. (1989) Cities and automobile dependence: An international sourcebook, Gower, Aldershot, England.

Newman, P. W. G. (1996) Transport. In UNCHS An urbanising world: Global Report on Human Settlements, UNCHS, Habitat and UNEP, Nairobi.

Newman, P. W. G., et al (1996) Human Settlements. In Australian State of the Environment Report, Department of Environment, Sport and Territories, Australian Government Publishing Service, Canberra.

Newman, P. W. G., Kenworthy, J. R. and Laube, F. (1997) The global city and sustainability. Fifth International Workshop on Technological Change and Urban Form, Jakarta, June.

Newman, P. W. G. and 14 authors (1997) Car-free Copenhagen: Perspectives and ideas for reducing car-dependence in Copenhagen. Royal Danish Academy of Fine Arts, Copenhagen.

Newman P. W. G. and Kenworthy J.R. (1998) Sustainability and Cities: Overcoming Automobile Dependence, Island Press, Washington DC.

OECD/EMT (1996) Urban travel and sustainable development. OECD, Paris.

Ohmae, K The Borderless World, Fontana, London (1990).

Rainbow, R. and Tan, H. (1993) Meeting the demand for mobility. Selected Papers, Shell International, London.

Sassen, S. (1994) Cities and the world economy. Pineforge Press, Thousand Oaks, California.

Serageldin, I. and Barrett, R. (1993) Environmentally sustainable urban transport: Defining a global policy. World Bank, Washington D.C.

Webber, M. (1963) Order in diversity: Community without Propinquity. In Wingo, L. (ed) Cities and space: The future use of urban land, John Hopkins Press, Baltimore.

Webber, M. (1964) The urban place and the non-place urban realm. In Explorations in Urban Structures, University of Pennsylvania Press, Philadelphia.

Webber, M. (1968) The post city age. Daedulus, 97(4), 1093-1099

Winger, A. R. (1997) Finally: A withering a ways of cities? Futures,

Willoughby, K. (1994) The 'local milieux' of knowledge based industries. In Brotchie, J., Newton, P., Hall, P., Blakeley, E., and Battie, M. (eds) Cities in competition, Cheshire, Melbourne.

World Bank (1996) Sustainable transport: Priorities for policy reform. Development in Practice, The World Bank, Washington, D.C.

Discussion questions

1. Is there any encouragement that sustainability in cities is feasible from the data gathered in this case study?
2. How do decisions made about transport become land use decisions?
3. Is automobile-dependence a sustainable approach to city development?