The End of Automobile Dependence: How Cities are Moving Beyond Car-Based Planning

By Peter Newman and Jeffrey Kenworthy

Cities will continue to accommodate the automobile, but when cities are built around them, the quality of human and natural life declines. Current trends show great promise for future urban mobility systems that enable freedom and connection, but not dependence. We are experiencing the phenomenon of peak car use in many global cities at the same time that urban rail is thriving, central cities are revitalizing, and suburban sprawl is reversing. Walking and cycling are growing in many cities, along with ubiquitous bike sharing schemes, which have contributed to new investment and vitality in central cities including Melbourne, Seattle, Chicago, and New York.
 
We are thus in a new era that has come much faster than global transportation experts Peter Newman and Jeffrey Kenworthy had predicted: the end of automobile dependence. In The End of Automobile Dependence, Newman and Kenworthy look at how we can accelerate a planning approach to designing urban environments that can function reliably and conveniently on alternative modes, with a refined and more civilized automobile playing a very much reduced and manageable role in urban transportation. The authors examine the rise and fall of automobile dependence using updated data on 44 global cities to better understand how to facilitate and guide cities to the most productive and sustainable outcomes.
 
This is the final volume in a trilogy by Newman and Kenworthy on automobile dependence (Cities and Automobile Dependence in 1989 and Sustainability and Cities: Overcoming Automobile Dependence in 1999). Like all good trilogies this one shows the rise of an empire, in this case that of the automobile, the peak of its power, and the decline of that empire.

• Language: English
• Publisher: Island Press
• Release date: Aug 11, 2015
• ISBN: 9781610916134

Peter Newman

Peter Newman lives in Somerset with his wife and son. Growing up in and around London, Peter studied Drama and Education at the Central School of Speech and Drama, going on to work as a secondary school drama teacher. He now works as a trainer and Firewalking Instructor. He sometimes pretends to be a butler for the Tea and Jeopardy podcast, which he co-writes, and which has been shortlisted for a Hugo Award.

Book preview

structure.

1

The Rise and Fall of Automobile Dependence

In the early twentieth century, automobiles began to be used in cities as a convenient replacement for horse-drawn carriages. Henry Ford’s mass-produced Model T, the most influential automobile in history, was sold between 1908 and 1927. With the assistance of a rapidly growing and ever more powerful consortium of automobile interests such as General Motors and Firestone Tires, American cities as early as the 1920s began to tear out their streetcar (tram) systems, thus sowing the seeds for the automobile’s usurping of local transportation. Through this process, New York had lost most of its extensive surface streetcar system by 1926 (Klein & Olson 1996). However, car usage was never really central to city shaping in any urban area until after the 1940s, when major freeway and parking infrastructure began to be built entirely for the automobile. In American cities this process accelerated after 1956 with the establishment of the Highway Trust Fund, which used a dedicated gasoline tax to accumulate prodigious sums of money for freeway building in order to facilitate their vast car-dependent suburbs. A similar process occurred in those Australian cities that, in our data, have developed most closely to the American model.

For the previous hundred years in all cities of the now developed world, trains and trams were the dominant city-shaping transport system, and before that it was walking. During World War II, urban tram systems in particular were still overwhelmingly popular (due, in part, to gasoline rationing), but they fell into a rapid decline thereafter as more and more systems were torn up throughout the world. (See chapter 4 for an expansion of this as the theory of urban fabrics.)

The development of automobile dependence in cities is a complex process, enacted over decades of land-use and infrastructure development linked to the dominant economic waves of innovation (Freeman 1996; Hargroves & Smith 2005; Newman et al. 2009). Over many years much has been written on this by the big thinkers on cities, such as Mumford (1961), Jacobs (1961), and Schneider (1979), who have all traced these processes in different ways and have all shown the problems that result from reshaping cities totally around cars. Thomson (1977) developed typologies for cities based on their dominant urban form and transportation systems; one such type was the automobile city, which he termed full motorization. No matter how they are described, cities built around the car have many serious problems related to the costs of sprawl and to the costs of transport, including the oil vulnerability issues that seem to have been a major factor in causing the global financial crisis, or GFC (Newman & Kenworthy 1999a, 2011a).

The environmental and social impacts from automobile dependence have been the major focus of twentieth-century urban transport writing, with an assumption that economic outcomes were favored by the car but these externalities should be considered. However, the twenty-first century is showing a very different perspective, with the environmental and social impacts worsening (congestion and carbon emissions as well as a growing list of health impacts, including obesity and depression in sprawling suburbs), and the economic aspects of automobile dependence are now tipping toward redevelopment and sustainable transport modes (Newman & Matan 2012b; Trubka et al. 2010a, 2010b; Glaeser 2011). New data will be used in this book to address these matters, especially the economic perspective, as cities are now competing on how quickly they can reduce their automobile dependence.

Although they did attempt to bring a little data to support their ideas about the way cities evolve and function from a transportation perspective, mid-twentieth-century contributions from urban scholars remained largely conceptual on issues related to car-based planning. There remained a need for sound and systematic data on cities to provide an adequate perspective on the evolution of their transportation systems. In fact, such data are rare until around 1960, when transportation engineers in urban governments began to collect data for the science of transportation planning and modeling.

Our own data-collection work commenced in the late 1970s, when we attempted to collect systematic data back to 1960 on transportation, land use, infrastructure, and energy (see Kenworthy et al. 1999 and Newman & Kenworthy 1999a for a complete perspective on our Global Cities Database). Using a highly standardized process, we have continued to collect data on a wide variety of global cities, though in many cities and regions the data-gathering process continues to be too difficult. For that reason, there are, for example, no Latin American, African, or low-income Asian cities included in our comparison of world cities in this book, although some new data on Sao Paulo and Taipei are presented in chapter 3.

Figure 1-1. Trends in car vehicle kilometers-traveled and public transit passenger-kilometers per capita in 26 cities, by region, from 1960 to 2000. Source: Authors’ own data

Using the Global Cities Database, we have now pieced together the most comprehensive data set possible from 1960 to 2000 on 26 cities across the world.¹ The data, set out in the table in the appendix, are provided to show how universal was the growth in vehicle-kilometers-traveled (VKT) across different cities and regions. The first signs of a plateau appear in the last decade of the twentieth century (see figure 1-1 for this perspective from our urban data, and figure 1-3 for this in the USA as a whole). Canadian and European cities plateaued, Australian and US cities slowed. US cities grew 2,200 kilometers 1980 to 1990, but only 1,000 kilometers 1990 to 2000, with San Francisco, Los Angeles, and Phoenix showing no growth. (The peak was starting.)

We examined this 40 year trend of growth in VKT to determine the factors that best explain why it occurred.² In the transportation planning community there is some disagreement about the causes of VKT growth, and analysis of this unique database provides a chance to resolve the matter. There is no doubt that cultural changes and motoring costs are contributors, but the biggest points of contention seem to be whether transit services or density are significant factors, and how the provision of road space contributes to the trends (Kirwan 1992; Breheny 1995; Boarnet & Crane 2001; Mindali et al. 2004; Coevering & Schwanen 2006; Gordon & Richardson 2007; Mees 2009a, 2009b). The explanatory data we have assembled across this 40-year period to explain automobile VKT per capita cover population size, centralization (percentage of metropolitan jobs in the central business district, or CBD), urban density, public transit service levels per capita and the proportion of that service representing rail modes, public transit use per capita, parking availability in the CBD, cars per 1,000 persons, and road length per capita.

In this chapter and throughout the book we examine the data on recent trends in car use and rail use in the world’s cities, as well as density changes. However, these trends are chosen from each country’s own data and, though they are likely to be reliable in themselves, they are not all on the same scale and comparable as in our original Global Cities Database (Newman & Kenworthy 1999a). Thus in chapter 2 we look at a smaller sample of cities that have been carefully placed on the same data scales in order to achieve a more scientifically correct understanding of the trends, as shown in the 1960–2000 data analysis above and below.

The Causes of the Twentieth-Century Rise and Plateau in Automobile Dependence

The data in the table in the appendix address the question of the underlying causes of automobile dependence (for details see McIntosh et al. 2014). Findings suggest that the two most significant indicators of levels of car use are urban density and transit service levels, with a minor role played by road length per capita (see figure 1-2). In chapter 6 we provide a further discussion of the issue of urban density versus transit service levels in trying to understand transportation patterns, as this is a major debate in the literature that needs to be resolved (for example, see Mees 2009a, 2009b; Newman & Kenworthy 2011c).

The data also very clearly demonstrate that rail-based transit services are the most strikingly linked to reduction in car VKT per capita. Thus, cities seeking to limit car dependence should be investing in quality rail systems and building up urban densities around them, rather than increasing road capacity—just as Gordon and Richardson (1989) suspected but found abhorrent. The data suggest that growth in car use occurred in the midcentury period because transit services and density were allowed to drop, but by the end of the century car use began to plateau as transit services and density began growing again. These two factors continue to be critical to our story on car dependence, as they have been for the past 30 years.

Figure 1-2. Determinants of car VKT per capita in cities based on an analysis of urban data from 1960 to 2000, showing the significance of public transit services and density. Source: McIntosh et al. (2014).

The Fall of Automobile Dependence in the Twenty-First Century

In 2009, the Brookings Institution in Washington, D.C., was the first to recognize a new phenomenon in the world’s developed cities: declines in car use (Puentes & Tomer 2009). Peak car use suggests that we have reached the peak of growth in car use and the end of building cities around cars as the primary goal of planning—at least in the developed world. The twenty-first century is witnessing the end of automobile dependence in the same countries and cities that gave birth to it.

Puentes and Tomer (2009) showed that between 2004 and 2010 an absolute decline in car use was evident. Figure 1-3 shows this peaking in the US vehicle–miles of travel. Sivak (2015) found that these trends continued into 2013.

Stanley and Barrett (2010) found a similar trend in Australian cities at a similar time: car use peaked in 2004. We have since mapped this in all Australian cities, including small ones (Canberra) where congestion is no issue (figure 1-4). Car use per capita continued to trend down (Newman & Kenworthy 2011b).

Figure 1-3. Peaking of US vehicle–miles of travel (VMT). Source: Compiled from US Department of Transportation data.

Millard-Ball and Schipper (2010) examined the trends in eight industrialized countries that demonstrate what they call peak travel. Figure 1-5 shows this for six of the countries.

Millard-Ball and Schipper conclude:

Despite the substantial cross-national differences, one striking commonality emerges: travel activity has reached a plateau in all eight countries in this analysis. The plateau is even more pronounced when considering only private vehicle use, which has declined in recent years in most of the eight countries. . . . Most aggregate energy forecasts and many regional travel demand models are based on the core assumption that travel demand will continue to rise in line with income. As we have shown in the paper, this assumption is one that planners and policy makers should treat with extreme caution. (p. 372)

It is possible to see similar data on many countries worldwide. Gargett (2012) provides a comprehensive examination of a time series of car-use data, which shows a similar peaking of car use in a number of nations. This has become accepted by some economists, who see it as a part of decoupling GDP from car use in more-developed nations; certainly it is obvious in the very wealthy centers of big cities like New York and London. The Times reported that in Central London traffic fell by 19 percent between 2000 and 2009, and further noted: Supporters of ‘Peak-Car’ theory see a future in which the inner cities are given over to pedestrians, cyclists, and public transport, and café culture replaces car culture. . . .³ A PricewaterhouseCoopers real estate report suggests that there has been a structural change in Canadian cities: Canadians are moving back into cities in growing numbers in search of a live/work/play urban environment rather than car-based commutes stuck in traffic.⁴ In Toronto, the central city has added 300,000 people to the population and has 150 high-rise towers about to be built.⁵

Figure 1-4. Peaking of car use in Australian cities. Source: Newman and Kenworthy (2011).

Figure 1-5. Peak travel in six industrialized nations. Source: Millard-Ball and Schipper (2010).

In chapter 3 we will show evidence that the peak in car use is not necessarily occurring just in wealthy cities; indeed, where policies to improve transit services and increase densities are made, then cities change quite rapidly in their car dependence, even in the emerging cities of the world. In the United States there have been many newspaper articles and websites discussing the peak car-use phenomenon and especially the cultural issues associated with it. The PIRG Report (Davis et al. 2012) that interviewed young people from across America found that they valued their mobile phones more than cars. An article titled Cars Are So Yesterday: Young and Rich Leave Guzzlers Behind showed that from 2001 to 2009, car use by 16- to 34-year-olds decreased from 10,300 miles to 7,900 miles per capita—a drop of 23 percent.⁶ Modal share (the choice of transportation type) among this age group increased 100 percent in public transit, 122 percent in biking, and 37 percent in walking. More importantly, there was a reduction in the need to travel long distances by younger people moving into more urban locations. A special 2013 issue of the journal Transport Reviews was focused on the peak car-use phenomenon with data from the world’s cities suggesting that a structural change was underway.

In summary, evidence from numerous sources shows that peak car use does appear to be occurring. It is a major historical discontinuity that was largely unpredicted by most urban professionals and academics. At the same time, urban rail is on the rise and this, too, is happening at unprecedented rates.

The Rise and Rise of Urban Rail

There is now a major worldwide rail revival, including light rail (modern trams that can run on streets or dedicated rights-of-way), metro rail (modern subways), heavy rail (usually suburban or commuter rail), and high-speed rail (fast trains that run between cities). This reflects growing concerns by municipal, regional, and national governments about the need to make their transportation systems more sustainable, their cities more livable, and their economies more resilient to future shocks from oil vulnerability and from the need to reduce CO2 emissions in the face of global warming (Newman et al. 2009).

The multiple advantages of modern urban rail are clearly attractive to policy makers. Hass-Klau et al. (2003, 2004), Litman (2004), and Kenworthy (2008) have summarized the major advantages of rail, especially its ability to use much less space than car-based infrastructure (a rail line can carry up to 20 times as many people as a freeway lane), its comparatively low impact (if well designed), and its potential to generate funding opportunities through land-value capture (covered in detail in chapter 6) and hence ease its