CAN URBAN DENSITY EXPLAIN PERSONAL TRAVEL LEVELS?

This article was downloaded by: [Monash University] On: 21 October 2010 Access details: Access Details: [subscription number 922191591] Publisher Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 3741 Mortimer Street, London W1T 3JH, UK

Urban Policy and Research

Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713449094

CAN URBAN DENSITY EXPLAIN PERSONAL TRAVEL LEVELS? Patrick Moriarty

To cite this Article Moriarty, Patrick(1996) 'CAN URBAN DENSITY EXPLAIN PERSONAL TRAVEL LEVELS?', Urban

Policy and Research, 14: 2, 109 — 118 To link to this Article: DOI: 10.1080/08111149608551604 URL: http://dx.doi.org/10.1080/08111149608551604

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.

Research Articles CAN URBAN DENSITY EXPLAIN PERSONAL TRAVEL LEVELS? Patrick Moriarty

Downloaded By: [Monash University] At: 23:09 21 October 2010

T

his article examines the changes in personal travel levels, urban density, and other land use patterns over the period 1901 to 1991 in large Australian cities. It is shown that urban density changes are not the direct cause of the observed travel increases. Instead, the huge rise in travel convenience brought about largely by the shift from public transport to cars, best explains travel growth in the post-war era. Urban density, however, does affect travel convenience (especially average travel speeds and ease of parking), and thus indirectly, travel levels. Keywords: urban density; urban transport; land use; travel convenience.

Introduction Reductions in vehicular travel, especially by car, have often been suggested as a way of reducing the problems high levels of urban traffic cause (Cevero 1995; Moriarty 1994; Royal Commission on Enviromental Pollution 1994.) These problems include not only air and noise pollution, traffic intrusion and accidents, but also oil depletion and increasing concentrations of heat-trapping greenhouse gases.

(Newman and Kenworthy 1989) and Melbourne (Moriarty and Beed 1987), it has been found that personal travel levels increase with distance from the centre, whereas urban density decreases outwards from the centre. Other studies which support the use of land use policy to reduce car travel include Pushkarev and Zupan (1977) and Holtzclaw (1994), both for the U.S., and Sherlock (1991) for the U.K.

Many researchers argue that appropriate changes in land use, especially increases in urban density, are an important means of reducing urban travel, and thus the problems it entails. Newman and Kenworthy (1989) for example, in their analysis of 32 world cities, including Australia's five largest, showed that for 1980, both per capita transport energy and travel itself, increased exponentially as urban population density decreased, although Brindle (1994) argues that this result was inevitable, given the way the data were analysed. For individual cities such as Perth and New York

Other urban researchers are sceptical of the value of urban density increases. Ruth Steiner (1994), in her review article on the topic, points out that studies supporting the link fail to separate out several factors associated with high density residential areas that also lead to differences in the use of the automobile'. She further argues that 'density could be seen as a proxy for these other unmeasured variables' which include income, household size and lifecycle characteristics of the household members. Australian critics include Kirwan (1992) and Brindle (1994).

109 Urban Policy and Research Vol 14 No 2 1996

The present article aims to help resolve this controversy, for large Australian cities at least, by examining in detail the changes in personal travel levels, urban density, and other land use patterns over the period 1901 to 1991. I argue that urban density changes are not the direct cause of the observed travel increases. Instead, the huge boost to travel convenience brought about largely by the shift from public transport to cars, best explains travel growth in post-war era. Urban density does affect travel convenience (especially average travel speeds and ease of parking), and thus indirectly, travel levels.

Downloaded By: [Monash University] At: 23:09 21 October 2010

Twentieth century transport and land use patterns This section presents selected transport and land use data for the five large capital cities over the 20th century. Table 1 gives data for 1901, a census year, for Sydney and Melbourne only, because the other three capital cities were very small in 1901. Table 2 gives data for 1947, the year of the first post-war census, and Table 3 for 1991, the most recent census year. All data, as far as possible, are based on the metropolitan area boundaries as defined at each census. The important exception is the second last entry in each table, urban density, which is based on urban centre boundaries and population, as defined by the Australian Bureau of Statistics (ABS). Since these figures are not directly available for 1901 and 1947, I have used Manning's figures for Melbourne and Sydney. The values for the three smaller capitals in 1947

are my rough estimates, calculated like Manning's, on the basis of the ABS urban centre definition. Also included in each table is the cut-off density, calculated on a Local Government Area (LGA) basis, for the quintile of the urban population with highest density. Since at each date this quintile were invariably living in fully urbanised LGAs, the lack of data at the Collector District level does not lead to any inaccuracy. The two indices for population density give consistent results, except for Sydney in 1947. Given that Sydney had the highest urban density at all other dates, including 1960-80 (Newman and Kenworthy 1989), it is unlikely to have been overtaken by Melbourne in 1947. I thus assume that the rank order for urban density was: Sydney, Melbourne, Brisbane, Adelaide, Perth at all dates. Most of the data for 1901 and 1947 came from State and Commonwealth yearbooks of the period, and Manning (1984). Sometimes public transport patronage alone was available, so revenue figures and average fares per km were used to estimate passenger-km (Moriarty and Beed 1992). The 1991 demographic data were based on published and unpublished census material for population, household size and other data. The 1991 transport data were based mainly on the 1991 Motor Vehicle Usage Survey (ABS 1993) and the recent study on 'Urban Transport' by the Industry Commision (1994). In general, the population and related data are more accurate than the transport data. Also, the recent transport statistics are more accurate than those of 1947, or especially, 1901.

Table 1 Transport and land use data, 1901 Sydney Population (000s) Per capita public transport pass.-km Average household size Average distance of -the population from the CBD (km) Urban density (persons/km2) Cut-off density for top quintile (persons/km2)

Melbourne

482 1100 5.5

496 1000 5.0

4.7 4430

4.9 3570

7860

6180

Sources: ABS State and Commonwealth Yearbooks; Manning (1984).

110 Urban Policy and Research Vol 14 No 2 1996

Table 2 Transport and land use data, 1947 Sydney Population (000s) Cars per 1000 population Per capita veh.pass.-km — public transport — car Average household size Average distance of the pop. from the CBD Retail sales % in central LGA Urban density (persons/km2) Cut-off density for top quintile (persons/km2)

Melbourne

Brisbane

Adelaide

Perth

1632

1228

402

66

75

65

383 99

4190 3550

3630 2900

2780 2150

2510 1550

278 74 2320 1600

640 3.9

730 3.6

630

N.A.

960 3.8

720 4.1

10.4 50.3 2900

8.4

5.6

6.5

7.4

44.3 3210

N.A. 1700

61.4 1500

70.4 1300

7000

5390

3990

3110

1710

Sources: ABS State and Commonwealth Yearbooks; Manning (1984); ABS Census of Retail Establishments, 1947-48.

Downloaded By: [Monash University] At: 23:09 21 October 2010

Taken together, the tables enable the following comparisons to be made:

Tables 1 and 2 show that over the period 19011947, the large population increases in Melbourne and Sydney were accompanied by declining household size and decreasing population density. Per capita vehicular travel also rose appreciably over the period, and of course, this increase was mostly by public transport. We therefore have the unexpected result that declining population density was accompanied by increasing per capita public transport travel. Both declining household size and increasing populations lead to the average distance of the population from the CBD steadily increasing, not only in 1901-1947, but also 1947-1991.

(a) changes in both transport and land use over the periods 1901-1947 and 1947-1991. (b) comparison between the five cities in transport and land use for both 1947 and 1991. Although many trends have been continuous throughout the century, such as declining household size and population density, the decades following 1947 are very different from the earlier period for transport, with car travel replacing public transport as the dominant mode. This change in transport led to important changes in land use as well, as we shall see.

Table 2 shows that the two larger cities, Melbourne and Sydney, not only had higher urban densities

Table 3 Transport and land use data, 1991 Melbourne

Brisbane

Adelaide

Perth

3673

3157

1358

1057

1188

476

535

524

549

573

9645 1525 8120 2.81

10050

11085

10150

11075

800

885

700

485

9250 2.88

10200 2.83

9450 2.61

10590 2.75

24.9 2005

20.2 1680

16.7 1140

12.5 1430

14.4 1165

3220

2530

2140

1900

1500

Sydney Population (000s) Cars per 1000 population Per capita veh. pass.-km —public transport —car Average household size Average distance of the pop. from the CBD Urban density (persons/km2) Cut-off density for top quintile (persons/km2)

Sources: 1991 Census; ABS State and Commonwealth Yearbooks; Industry Commission 1994; INTSTAT 1988.

111 Urban Policy and Research Vol 14 No 2 1996

Downloaded By: [Monash University] At: 23:09 21 October 2010

than the three smaller cities, but also had much higher levels of per capita vehicular travel, chiefly because of higher public transport use. Again, the conventional wisdom would be that the cities with higher urban density would have lower per capita levels of vehicular travel. However, the higher density cities did have higher levels of per capita public transport travel, as expected. In the more recent period 1947-1991, urban densities in all cities continued to fall, largely because of declining household size (Table 3). These density declines were accompanied by very large increases in per capita travel, largely the result of the explosive growth in car ownership and use. The static comparison of transport and land-use data is also in accord with conventional theory: the three smaller cities have lower urban densities than Melbourne or Sydney, and have slightly higher levels of per capita travel. Overall there are anomalous results in the earlier public transport period, but in the more recent car era the results are apparently in agreement with the argument that lower urban density decreases public transport patronage, but increases overall levels of personal travel.

Analysis of transport and land use patterns The previous section demonstrated that the relationship between land-use and personal travel levels were different in the public transport era than in the present era. The following two subsections analyse in detail the different land use/transport interactions in each of the two periods. Public transport era

Consider a very simple model for a city, in which all activities requiring vehicular travel are located at the centre—workplaces, retail, sales, major entertainment and sports venues and so on. Clearly, total vehicular travel in such a city is the product of the average number of trips each person makes, weekly say, to the centre, and the average distance the population lives from the

CBD. If two cities have the same population density distribution then ceteris paribus the city with the higher population will have higher per capita travel simply because the average resident lives further from the CBD. There is some evidence that the five major cities approximated this model up to the beginning of the post-war era. For example, Table 2 shows that the central LGA (i.e. containing the CBD) had a share of total metropolian retail sales for 1947-48 varying from 44.3% for Melbourne to 70.4% for Perth. For jobs, the position was similar. In Melbourne in 1951 for example, the CBD alone still contained 28% of all metropolitan jobs, and the central six LGAs with an area of only 65.5 km2, nearly 60% of all workplaces (MMBW, 1954). Partial data suggest that the percentages for the other cities were even higher, as was the case for retail sales. Of course, not all trip destinations were in the central area; in fact most weren't. There was much local shopping, working, visiting friends and socialising, but the majority of this was done by non-motorised means, walking and cycling. But vehicular travel—especially on a passengerkm basis— was largely centralised, as might be expected given the radial nature of the fixed rail networks, and as evidenced by actual travel patterns for Melbourne in 1951 (MMBW, 1954). Vehicular travel was largely centralised, but nonmotorised travel was strongly decentralised. The travel data presented in Tables 1 and 2 lend some support to this simple urban model. Table 2 shows that the two larger cities (with greater average residential distance from the CBD) had greater average per capita vehicular travel than that for the smaller three cities. Further, the per capita travel increases found over the period 1901-1947 for Melbourne and Sydney, can be at least partly explained by the rising distance of the population from the CBD, particularly if nonvehicular travel is included. Such cheap travel modes were a far larger proportion of total travel in 1901 than in 1947 (Manning 1984). Average

112 Urban Policy and Research Vol 14 No 2 1996

Decreased urban density as a key explanatory factor for travel increase only makes sense if separation of activities has increased, if, for example, the average separation of residences and workplaces (or shops, or schools) has increased. But Moriarty and Beed (1992) in their analysis for the period 1947-1986 could find little evidence of any such increases for shopping, work or education trips. For example, they calculated the minimum work travel distance— the average airline distance each worker would need to commute for minimum city-wide work travel (Moriarty and Beed 1988). From 1961 to 1986 minimum commutes actually decreased from 5.3km to 4.8km for Melbourne, and from 5.9km to 5.4km for Sydney. In other words the need for travel, at best, has increased only marginally.

Downloaded By: [Monash University] At: 23:09 21 October 2010

distance of the population from the CBD is in turn influenced by two factors: population density and city population size. In general, two cities of equal population but different density will have different levels of personal travel, the higher density city having less travel because the residents live closer on average to the CBD. Average distance of the population from the CBD cannot be the sole explanatory variable, since Table 2 shows that in 1947, Perth, with a larger CBD distance than Brisbane, had lower levels of personal travel. Other factors affecting vehicular travel levels included the geography of the cities, for example the presence of wide rivers, estuaries, bays and hills. The extent of the fixed rail public transport network was also important: Brisbane in 1947 had a much more extensive network than Perth. Finally economic prosperity increased vehicular travel, and depressions lowered it. Thus both rail and tram patronage were hit by the depressions of the 1890s and 1930s. (See for example, the historical transport statistics in the state yearbooks). But at any one time, income differences between the five cities was small, so average resident distance to the CBD along with fixed rail network extent, best explain the observed travel levels.

A simple way of making the same point is to calculate the difference in the value of the distance the average resident lives from the CBD on the one hand, and the equivalent distance for retail sales expenditure, workplaces and so on, on the other. In Melbourne, for example, this difference for retail sales has decreased from 4.4km in 1951 to 3.1km in 1991/92; for workplaces the difference is today the same as in 1951 (about 5 km) although it peaked at 7km in 1961 (MMBW 1954; ABS 1993, ABS 1994). The minimum work travel distances given above are in good agreement with the difference between average resident and average workplace CBD distances, which suggests that it is indeed the case that suburbanisation of activities has largely offset the effects of decreasing urban density.

Car era The late 1940s saw not only the lifting of petrol rationing, but the full establishment of car manufacturing in Australia. The changes to transport, as Table 3 shows, and everybody knows, were profound. The data in Tables 2 and 3 show that over the period, as urban density continued its decrease in all five cities, overall vehicular travel increased per capita and public transport travel levels fell. In 1991, the smaller cities had lower densities, lower public transport share and higher levels of vehicular travel. In addition, the residents of the denser inner suburbs of larger cities have lower vehicular travel levels than the residents of the less dense outer suburbs (Moriarty and Beed 1987). Do differences in urban density provide an explanation for differences in levels of personal vehicular travel in the present car era?

A closer examination of the urban density-travel relationship in the car era shows some anomalies. In 1991 for example, urban density for the two larger cities was about 40% higher than for the three smaller capitals considered together, but per capita travel differences between the two groups was less than 5%. At the very least, this suggests that density changes are not a very effective policy instrument for reducing travel. Also, since 1981, the urban density of all five capitals has stabilised, or in the case of Sydney,

113 Urban Policy and Research Vol 14 No 2 1996

Downloaded By: [Monash University] At: 23:09 21 October 2010

has even risen, yet travel levels are greater than in 1981. Indeed, for eight of the 32 cities analysed by Newman and Kenworthy (1989), urban density either increased or remained constant over at least one of the two decades 1960-1980, yet per capita travel and fuel use continued to increase, just as in the other cities. Clearly, urban density cannot be the explanation for these results. Further, the implicit assumption is often made that decreasing urban population density means that a given urban area can support fewer services such as schools and shops. But because of the large increases in both per capita incomes and proportion of the relevant age groups in secondary and tertiary education, for example, the 'retail sales dollar density' and '15-24 years old student density' have actually increased in all cities over the period. It is still true, of course, that higher urban densities would have led to even higher values. The point is simply that other factors besides population density determine viable catchment areas for a given service. So what has the dominant driving force been for increased levels of personal travel in the postwar era? The answer, I will argue, is the great boost to travel convenience associated with the shift from public to private travel since the war. Travel convenience has two parts (Moriarty and Beed 1992; Moriarty 1994). The first, independent of traffic conditions, includes privacy, all-weather protection, and ease of transporting young children or goods, and varies little from city to city. With car air-conditioning and stereo systems, and now car phones, it is increasing over time. The other part is traffic-dependent, varying with traffic speeds and ease of parking, both of which determine, for example, door-to-door travel times. Travel convenience in this sense not only varies somewhat from city to city, but also over a given city, and by time of day. Travel convenience can readily explain the minor differences in travel levels between the two larger and the three smaller capitals in 1991. Average travel speeds are higher, and ease of parking,

especially in the CBD, is easier in the smaller cities. Similarly, travel convenience differences can explain the lower levels of travel by inner city, compared with outer city, residents. The huge boost to travel convenience resulting from rising car ownership can also explain why travel has risen so much since 1947. In contrast to the radially oriented fixed rail public transport network, the ever-expanding road network made non-radial travel easy. The progressive suburbanisation of activities meant that most trips were less and less oriented to the centre. Such a dispersion of destinations also meant that car travel convenience could be kept high even as car ownership rose to about one vehicle for every two residents. The result was a large rise in discretionary trip making. A factor related to travel convenience, psychological benefits of private travel (Moriarty 1994), also contributed to travel increases. Travel convenience can thus offer explanations of all the data presented here. Urban density can explain some, but not all, the results. Also, increased activity separation since 1947 has at best been modest. The only other plausible explanation, income differences, cannot explain why the three smaller cities today have higher levels of personal travel than the two larger cities, nor why inner city residents travel less than outer city residents. This is not to say that travel convenience is the sole reason for travel increases in all cities at all times. I have already shown that in the public transport era, travel convenience was a minor factor, along with income differences. If in future motoring costs rise dramatically, and income inequality increases, it is possible that income could be the major factor in explaining travel differences within a city. Finally, this analysis has focused on Australian cities. It is possible that urban density is more important in some overseas cities. It is relevant to ask what all this increased travel has been for. Nearly all the increase was for discretionary trip making, since per capita growth

114 Urban Policy and Research Vol 14 No 2 1996

increase linearly from the centre to twice this value 30 km out, after which they remain constant, as suggested by the data in Moriarty and Beed (1987),and the more recent 1991 census LGA car ownership figures. Again using 1991 census data, average personal travel levels in the urban centre area will only be 2.5% less than in the much bigger S.D. area.

Downloaded By: [Monash University] At: 23:09 21 October 2010

in work and education trip frequency (and average trip lengths) has been modest. One important consideration is the shift from non-motorised to vehicular travel. In 1947 the great majority of discretionary trips—for shopping, personal business, recreation and visiting friends—were by non-motorised means, and were necessarily of short length. With rising car ownership, this natural constraint on trip length was lifted, with the result that vehicular discretionary travel expanded rapidly.

Similar calculations show that variables such as average household size, and car ownership are all affected very little by boundary changes. The main reasons are that the much smaller area still includes about 90% of the S.D. population, and that the distinctive inner city area is only a small proportion of the total population. Average distance of the population (and workplaces, retail sales) from the CBD will be more affected, but the difference between these will again be very small. Population density can be measured in a variety of ways (for a comprehensive discussion, see Mees 1994), with each giving a different value. However, different methods, if applied consistently, should not affect the relative density ranking of the five cities. In summary, despite much intra-city variation, particularly in Sydney and Melbourne, data obtained at the city-wide level gives reliable results, even when somewhat arbitrary boundaries are inevitably used.

In summary, in the public transport era, urban densities were high, but activities were more concentrated in the centre. In the car era, urban densities were much lower than in the first era, but suburbanisation meant that residences, workplaces and services of all kinds were now intermingled. The potential for the low travel levels of 1947 was present, but the travel convenience of the car resulted in actual travel levels today being several times greater than in 1947.

Possible objections There are two main criticisms which have been, or could be, made against the arguments developed so far: (a) the arbitrary and changing nature of the urban boundaries, and a related point, the use of highly aggregated data for analysis. (b) the importance of other factors, especially income, but also household size, age distribution and so on, as an alternative explanation for travel increases.

As noted earler, some researchers think that the lower personal travel levels observed for inner city residents is less the result of higher densities than of differing incomes, household sizes and other social factors. Data from Melbourne (Moriarty and Beed 1987) suggest that this argument is not valid for modern Australian cities. Per capita income levels, population percentage in the 15-65 year age group and percentage economically active —all correlated with higher travel levels (INTSTAT 1988)—were all higher in the inner than the outer suburbs (ABS 1991). True, household sizes were smaller, but this has been allowed for by using per capita values of income and travel. I conclude that the travel differences observed are real.

Personal travel levels increase with resident distance from the CBD (Moriarty and Beed 1987). If the urban boundary is drawn close to the centre, then evidently average personal travel levels will be lower than if they are drawn further out. How important is this effect? Consider the difference in average travel that would occur if the 'urban centre' area, which is only 21% of the Statistical Division (S.D.) area, was used instead for Melbourne in 1991. Assume personal travel levels

115 Urban Policy and Research Vol14No2 1996

Downloaded By: [Monash University] At: 23:09 21 October 2010

Travel growth over time in a given city is another matter, as it was accompanied not only by falling densities, but also by rising incomes. How important was income growth compared with increasing distance from the centre in the prewar era, or the increasing convenience of vehicular travel, especially in the car era? First, it is clear that rising incomes enabled households to finance their growing travel, particularly in the present era when it involves purchasing cars. To the extent that the psychological benefits of car use increase overall travel (Moriarty 1994), income growth can. be linked to travel growth. Psychological benefits aside, the idea of travel as a largely derived demand argues against income growth as a major direct contributor to travel growth. For example, higher incomes mean more money available for shopping, but this does not necessarily mean that shopping trip frequency must increase. Nor, obviously, does work trip frequency. But some new expenditure made possible by higher incomes may require new trips to be made. Overall, income effects on travel are mainly indirect. As the Melbourne example shows, lower travel can occur—as in the inner suburbseven with higher per capita incomes. Implications for travel reduction policies I have shown that, historically, a different urban form was associated with each dominant transport mode. These forms were well matched to the capabilties of the transport modes: more centralised in the public transport era, more decentralised in the present car era. Again, the public transport decades had large household sizes, low per capita incomes, as well as strongly centralised cities. Such is not the case today. But this does not necessarily mean that increased public transport and much lower car travel levels are impossible. The fall in public transport's share of urban travel is not only the result of changing land use patterns, but is also heavily influenced by the superior convenience we have created for car travel. While many trips will always be unsuitable for public transport, there is still a

significant untapped section of the travel market where public transport access and convenience is not much inferior to that for car (Moriarty and Mees 1994). And as we have seen, suburbanisation has created the potential for travel reduction, even the car's convenience has actually led to travel increases. The scope for some shift to non-motorised travel is also good, since many car trips are less than one kilometre. Regardless of the main reason for travel increases, reductions can always be effected by sufficiently high price rises for motoring. Similarly, falling incomes for any groups will lower their travel demand. Such an approach would, however, be inequitable. Increasing urban density (which usually has the indirect effect of reducing car travel convenience) offers a more equitable approach, but the very large changes required for significant results would not only take decades to achieve, but would face much opposition. I have argued in this paper that the greater convenience of the car is the chief reason why urban personal travel has risen so much since the war. Lowering the traffic-related component of travel convenience is therefore a possible means of curbing urban travel and the problems it entails, should such reductions be thought necessary. This would involve such actions as street closures, an end to further urban arterial road building, tighter parking restrictions in the CBD, much more pedestrian and public transport priority, and big cuts in speed limits. A similar set of policies has been recently proposed for the U.K. by the Royal Commission on Environmental Pollution (1994). Clearly if urban road space was to be drastically reduced, with no other policy changes, increased congestion, and possibly fuel consumption, would result. But if initially the road system was left unchanged (with no additions or closures), car traffic would still drop, and road closures could follow without congestion increasing. No policies for reducing car travel will prove popular, but this approach should produce results in a timely and

116 Urban Policy and Research Vol 14 No 2 1996

References

equitable manner. Implementation, however, will probably have to wait until local oil depletion or obligations to cut greenhouse gas emissions force us to take action.

Australian Bureau of Statistics (ABS) (Various years) Commonwealth and State Yearbooks.

Conclusions ABS (1949) Census of Retail Establishments 1947-48, AGPS. Also later surveys.

Downloaded By: [Monash University] At: 23:09 21 October 2010

This paper has analysed personal travel and land use patterns in the five large Australian cities over the period 1901-1991. The main findings are as follows.

ABS (1991) published and unpublished census data.

Between 1901 and 1947, declining urban density was accompanied by rising levels of personal travel, as expected. But in 1947 the larger (and higher density) cities—Sydney and Melbournehad much higher personal travel levels than the three smaller capitals. An explanation which fits both observations is that since vehicular travel was much more centrally-oriented than it is today, the average residential distance from the CBD was the main determinant of urban travel levels. The different coverage and service frequency of public transport, and their increase over time in all cities, also affected travel levels to some extent, as did rising incomes.

ABS (1993) Survey of Motor Vehicle Use Australia 30th September 1991 (and previous surveys), AGPS, Canberra. ABS (1994) Retailing in Victoria 1991-92. Brindle, R. (1994) Lies, Damned Lies and 'Automobile Dependence' —some hyperbolic reflections, Proc. 19th ATRF, Sept, Lome. Cervero, R. (1995) Why go anywhere?, Scientific American, September, 92-93. Holtzclaw, J. (1994) Using Residential Patterns and Transit to Decrease Auto Dependence and Costs, Natural Resources Defence Council, San Francisco.

Since 1947, personal travel growth has accelerated and urban density has continued to decline. Today, however, residents of the lower density smaller capitals travel a little more than those of the larger two capitals. Superficially, density declines can explain this travel growth. In contrast to the public transport era, however, there is little evidence of increased separation over time of residences and workplaces, educational centres or other services. Suburbanisation has led to their progressive intermixing. It is argued that rising levels of car ownership gave a huge boost to the convenience of vehicular travel, leading to higher personal travel but shrinking public transport patronage. Urban density has an indirect influence on travel, however, by reducing travel convenience for motorists. Lowering travel convenience thus offers a direct approach for reducing travel, should environmental or resource constraints require it.

Industry Commission (1994) Urban Transport, Report No. 37, AGPS, Melbourne. INTSTAT Australia P/L (1988) Day to day travel in Australia 1985-86, Federal Office for Road Safety, Canberra. Kirwan, R. (1992) Urban form, energy and transport: a note on the Newman-Kenworthy thesis, Urban Policy and Research, 10(1), 6-22. Manning, I. (1984) Beyond Walking Distance, ANU Press, Canberra.

117 Urban Policy and Research Vol14No2 1996

Mees, P. (1994) Toronto: paradigm reexamined, Urban Policy and Research, 12(3), 146163. Melbourne and Metropolitan Board of Works (1954) Melbourne Metropolitan Planning Scheme 1954, MMBW, Melbourne. Moriarty, P. (1994) Urban transport: time for change, Australian Quarterly, 66(4), 105117.

Downloaded By: [Monash University] At: 23:09 21 October 2010

Moriarty, P. and Beed, C. (1987) Land-use modification to reduce travel in Melbourne, Proc. Int. Symposium on Transport, Communication and Urban Form, Vol. 2, 195-216, Monash University. Moriarty, P. and Beed, C. (1988) Transport characteristics and policy implications for four Australian cities, Urban Policy and Research, 6(4), 171-180. Moriarty, P. and Beed, C. (1992) Explanation of personal travel increases in Australian cities, Proc. 17th. ATRF, 7-9 Oct., Canberra.

Moriarty, P. and Mees, P. (1994) Counter reformation in urban transport—seeking 'win-win' solutions, Proc. 19th ATRF, Sept, Lome. Newman, P.and Kenworthy, J. (1989) Cities and Automobile Dependence: An International Sourcebook, Gower, UK. Pushkarev, B., and Zupan, J. (1977) Public Transport and Land Use Policy, Indiana Univ. Press, Bloomington, USA. Royal Commission on Environmental Pollution (1994), Transport and the Environment, Report 18, Cm 2674, HMSO, London, UK. Sherlock, H. (1991) Cities are Good for Us: The Case for Close-Knit Communities Local Shops and Public Transit, Paladin, London, UK. Steiner, R. (1994) Residential density and travel patterns: a review of the literature, Transportation Research Record, 1466, 3743.

118 Urban Policy and Research Vol14No2 1996