Cities Total cost (CA$ billion)

Dispersed scenario

Road capital cost Transit capital

Water and wastewater Fire stations

Recreation centres Schools Total

17.6 6.8 5.5 0.5 1.1 3.0

34.5

Recommended direction

11.2 6.2 2.5 0.3 0.9 2.2

23.3

Difference Percent difference

6.4 0.6 3.0 0.2 0.2 0.9

11.2

Ta ble 1: Infrastructure costs for different development scenarios in Calgary Dispersed scenario: additional 46,000 ha; recommended direction: additional 21,000 ha Source: IBI Group (2009)

In developing countries, however, urbanisation may not provide the same kind of economic gains across cities and firms. For example, Brülhart and Sbergami (2009) find that within-country agglomeration boosts GDP growth only up to national income levels of US$ 10,000 per head. The main reason for this is that very rapid – and sometimes chaotic – urbanisation can outstrip national and city governments’ ability to provide adequate infrastructure and services (Cohen 2006). Congestion could eat up the benefits of higher density as in the case of cities like Shanghai, Bangkok, Manila and Mumbai (Rigg et al. 2009). Venables (2005) similarly suggests that “the presence of increasing returns to scale in [some developing country] cities leads to urban structures that are not optimally sized”.

Lower infrastructure and operating costs Densification reduces the capital and operating costs of infrastructure. Evidence suggests that linear infrastructure including streets, railways, water and sewage systems as well as other utilities come at considerably lower cost per unit the higher the urban density (Carruthers and

Transport Infrastructure Dual-lane highway

Urban street (car use only) Bike path (2m)

Pedestrian walkway / pavement (2m) Commuter Rail Metro Rail Light Rail

Bus Rapid Transit Bus Lane

Capacity [pers/h/d]

2,000 800

3,500 4,500

20,000 – 40,000 20,000 – 70,000 10,000 – 30,000 5,000 – 40,000 10,000

-36 -9

-54 -46 -19 -27 -33

Ulfarsson 2003). Comparing smart growth areas and dispersed, car-dependent developments, Todd Litman suggests direct cost savings between US$ 5,000 and US$ 75,000 for building road and utility infrastructure per household unit (Litman 2009a). A recent exercise for Calgary (IBI Group 2009) indicates cost savings beyond pure linear infrastructure but also for schools, fire stations and recreation centres (see Table 1). Similarly, a recent study of Tianjin concluded that infrastructure cost savings as a result of compact and densely clustered urban development reach 55 per cent compared with a dispersed scenario (Webster et al. 2010).

Figure 4: Private transport fuel expenditure and urban density of selected cities shows how urban density can be an essential measure for decreasing long-term operating costs. Critically, this relationship is made even stronger in the right-hand graph which standardises 2008 fuel prices at the EU average (US$ 1.41) – in other words, it assumes that all cities in the sample face the same fuel price. It is clear that EU cities tend to be denser than North American cities and significantly more effcient in terms of fuel consumption – citizens of more sprawling North American cities tend to travel further. But even with current US fuel prices, density pays back. In the case of New York City, CEO for Cities (2010) estimates that density-related cost savings through reduced expenditure on cars and petrol translates into a green dividend of US$ 19 billion annually.

While denser city strategies tend to promote

greater energy efficiency and cheaper infrastructure, promoting transport modal shifts can deliver higher lifecycle capacity and lower running costs (see Table 2: Capacity and infrastructure costs of different transport systems). The most significant cost saving is derived from a shift away from car infrastructure towards public transport, walking and cycling. For example, at similar capacity levels, bus rapid transit (BRT) offers significant costs savings compared to traditional metro

Capital costs [US$/km]

10m – 20m 2m – 5m 100,000 100,000

40m – 80m 40m – 350m 10m – 25m 1m – 10m 1m – 5m

Tabl e 2: Capacity and infrastructure costs of different transport systems Source: Rode and Gipp (2001), VTPI (2009), Wright (2002), Brilon (1994)

Capital costs/ capacity

5,000 – 10,000 2,500 – 7,000 30 20

2,000

2,000 – 5,000 800 – 1,000 200 – 250 300 – 500

467