03Cities need to become denser to achieve net zero

Cities are likely to keep playing a disproportionate role in helping the UK hit its net-zero target, as their density means that they are places in which greater progress can be made in cutting transport and domestic emissions. But it will mean that cities and large towns will need to become denser to make this happen.

UK cities are not very dense by international standards

The UK is often thought of as one of the most densely populated places in Europe. When using a traditional definition of population density (total population relative to total land area), at 280 people per square kilometre, the UK is third in Europe behind the Netherlands (508) and Belgium (382).11

But this hides significant variation between and within cities, especially when using alternative measures of density, such as those where only the size of the built-up area is taken into account. Sometimes called ‘lived-density’, this better reflects the compact nature of an urban area.

UK cities, especially those outside London, are not particularly dense by international standards. Figure 7 shows the distribution of square kilometre units of land with a population of more than 10,000 for a selection of European countries. In the UK, 86 per cent of these areas have between 10,000 and 15,000 inhabitants, with no area having more than 25,000 per square kilometre. This contrasts strongly with places like Spain, Belgium and France. Even the UK capital ranks below a number of European cities – its most populated square kilometre, in Maida Vale, has just over 20,000 inhabitants, compared with more than 50,000 in cities like Barcelona and Paris.

Figure 7: Compared to other European countries, the UK is not very dense

Source: Eurostat, 2011.

London is also less dense than cities like Geneva, Bilbao or Naples; and many UK large cities such as Manchester or Sheffield rank below a number of European cities of similar size, particularly French and Spanish ones (Figure 8).

Figure 8: UK cities are not as dense as many of their European counterparts

Source: Eurostat, 2011.

Much development has run counter to achieving net zero

A lot of this has to do with the location, over many decades, of new residential developments in the UK, as they tend to be on greenfield land on the outskirts of cities rather than on brownfield land or in existing built-up areas. Such development increases car dependency.

This has historically been explained by the considerable expansion of the automobile industry in the 1960s, which slashed transportation costs and contributed to urban sprawl by increasing demand for housing in suburban areas. It has been widely documented in the United States, and holds true in the UK.12  This was most clearly characterised by new towns such as Milton Keynes, Telford and Warrington, which were designed as low-density, car-dependent urban developments.13

These patterns have continued more recently, despite the UK’s net zero goals. Work by the Foundation for Integrated Transport explored the characteristics of a selection of new-built developments across the country, including in cities like Peterborough, Leeds, Newcastle, Warrington and Swindon. It showed that, with a few exceptions, most were car-dependent, located away from jobs and public transport networks, and built at very low-density levels.14

This critique also holds for commercial space. The many Enterprise Zones and business parks that have been built in recent decades have encouraged car commutes because they are more difficult to serve by public transport (Box 4).15

Box 4: A example of out-of-town business park development: Newburn Riverside

The Newburn Riverside Business Park is located on the outskirts of Newcastle, by the River Tyne. It is now home to the Waterfront Estate, which combines industrial facilities and offices. When it was built in 2003, it was presented as a landmark example of a sustainable development, on a large decontaminated brownfield site using sustainable design and materials.

But the carbon footprint of this new development is not as good as these greener features may suggest. This is in large part because of its out-of-town location, more than three miles away from Newcastle city centre. The site is adjacent to the A1(M) motorway, and has very low public transport accessibility. This is likely to have encouraged car-based commuting: in 2011, around 85 per cent of workers used a car to get there, compared to an average of 58 per cent in Newcastle City Council area.16

These patterns have in part been shaped by the economics. Greenfield land on the outskirts of cities tends to be cheaper and less risky for developers to build on, while an out-of-town location is a better fit for some businesses than a more central one.17

That said, policy has had a role to play in further encouraging these patterns. The current planning system has encouraged the existing pattern of house building, as Box 5 explains.18 And successive subsidies for out-of-town employment space, such as Enterprise Zones, science parks and other public sector-led developments have encouraged low density, car-dependent development in fringe locations.

Box 5: The role of the planning system

In its current form, the planning system does not produce the most effective results in terms of the quantity of houses being built or the location of new developments.

It is a ‘plan-led’ system, where landowners come forward with areas for development as part of a ‘call for sites’ process. The land must then be assessed and accepted by the local planning authority, first to be part of the local plan and then to be given planning permission.

The land is rarely in existing built-up areas, mostly because development is easier where homeownership is simpler. Given that built-up suburbs tend to have high levels of homeownership, where plans for densification can be met by a lot of resistance, it is politically less risky to build on the outskirts of cities, for instance on farmland.

It is easier for developers too. Densifying existing built-up areas, usually on land that has not been allocated for development in the local plan (‘windfall sites’), presents more upfront costs and risks. It often requires developers to purchase and assemble small lots, which is a lengthy process and there is no certainty over the outcome, as the plan can be rejected on a case-by-case basis. This is a major barrier to scaling suburban densification.

This has shaped transport choices in cities

In low-density cities and large towns like Huddersfield and Telford, transport emissions per head are far higher than in cities such as Oxford, London and Brighton (Figure 9). In the latter group, more than half of the population commutes by walking, cycling or public transport.

Figure 9: Higher density leads to higher active travel usage, resulting in lower transport emissions

Source: BEIS, 2020. Census 2011. Centre for Cities’ own calculations. The bubble size here refers to the take up of public transport and active travel.

This relationship between density levels and transport emissions is clear: modelling by Centre for Cities shows that a 20 per cent increase in density would lead to a 10.4 per cent fall in transport emissions, based on current behaviours – and much of this would be explained by changes in public transport usage.19 In absolute terms, this reduction would be equivalent to four times Newcastle’s current total transport emissions. If density increased by 50 per cent, emissions would go down by 21.7 per cent, a reduction equivalent to the size of London’s current emissions.20

It means that a household moving from an area with Telford’s density levels (around 30 people per hectare) to somewhere as dense as Brighton (60 people per hectare), would expect to cut their transport emissions per head by 33 per cent on average – moving from 1.8 tonnes of CO2 to 1.2 tonnes of CO2.

This relationship between density and transport emissions can be seen in cities and large towns too. Research from abroad has shown that transport emissions per capita are lower in city centres than in suburbs.21 And although there is no sub-local authority emissions data in the UK, data on car ownership points to lower ownership (and so fewer emissions) in denser neighbourhoods.

While there is little difference in car ownership levels between cities and non-urban areas, the gap is much larger between city centres and suburbs. City centre residents own 0.2 cars, on average (defined here as per adult population), but that is doubled for residents of suburban areas (Table 1). In London, for example, there is a large gap between inner boroughs like Islington, Hackney, Camden and Tower Hamlets, where there are fewer than 0.2 cars per person, and outer London boroughs like Hillingdon and Bromley (more than 0.6 cars per person), as shown in Figure 10.

It is also in city centres that car ownership levels have decreased the most in the past 10 years (see Table 1), despite the fact that their populations grew by around 34 per cent around the same period. In Coventry and Oxford, as well as in larger city centres like those in London, Liverpool, Sheffield and Newcastle, car ownership levels went down by more than a third. This further illustrates the fact that people located in denser areas have less need for a car.

Table 1: Car ownership levels are higher in suburbs than in city centres, and higher again in rural areas

Area Cars per capita 2019 2010-2019 change Population change 2010-2019
City centres 0.23 -21% +34%
Suburbs 0.54 +5% +6%
Hinterland/outside 0.65 +6% +6%

Source: DfT, 2020.

Figure 10: Car ownership levels are much higher in outer London

Source: DfT, 2020

It has shaped domestic emissions too

In recent years, suburban sprawl and the construction of single-family, detached housing has negatively affected domestic emissions. Energy Performance Certificate (EPC) data shows new apartments emit 67 per cent less than new houses – flats built (or converted) in 2019 emitted 0.9 tonnes of carbon annually, while houses constructed in the same year were responsible for 1.5 tonnes.22 This gap has widened in recent years – between 2013 and 2019, emissions from new flats went down by 18 per cent, on average, compared with 11 per cent for new houses (Table 2).

Despite this, houses account for an increasing share of all new builds – in 2019, nearly 80 per cent were houses, up 12 percentage points since 2013 (Table 2). Additional data from the National House Building Council (NHBC) confirms that most new builds are large, detached houses that tend to have a higher carbon footprint. In 2019, this type of home accounted for 30 per cent of all new dwellings, compared with 22 per cent in 2013.23

Table 2: New houses emit 65 per cent more emissions than new flats

2013 2019
CO2 emissions (tonnes/year) Share of all new build completions (%) CO2 emissions (tonnes/year) Share of all new build completions (%)
New flats 1.1 33 0.9 22
New houses 1.7 67 1.5 78

Source: EPC domestic register, 2019.


  • 11 UN World Population Prospects, 2019. https://statisticstimes.com/demographics/countries-by-population-density.php
  • 12 See, for instance, Glaeser, E., Kahn, M. (2003), Sprawl and urban growth, National Bureau of Economic Research.
  • 13 Campaign for Better Transport (2014), car dependency scorecard.
  • 14 This is the case in London. Research published in 2020 by the Centre for London has shown that recent housing developments have increased car dependency instead of reducing it. They are more likely than pre-existing housing to feature car parking, and as a result the new residents are more likely to use their car on a daily basis. In addition to ‘pro-car’ principles, such as minimum parking requirements, part of the problem is the length and arbitrary nature of the planning system, and the fact that many development decisions do not take into account residents’ future travel choices. An example of that is the Battersea Power Station development, which was given permission in 2013 with higher car parking requirements than recommended, despite being located near the future extension of the Northern line. See: Centre for London (2020). Building for a new urban mobility.
  • 15 See Swinney,P. and Sivaev,D. (2013) Beyond the High Street. Why our city centres really matter, London: Centre for Cities.
  • 16 Census, 2011.
  • 17 Serwicka I and Swinney P (2016): Trading places: Why firms locate where they do, London: Centre for Cities
  • 18 See Breach, A (2019), Sleepy Suburbs, London: Centre for Cities.
  • 19 The modelling uses density (defined as the ratio between population and size of the built-up area) and transport emissions, controlled by the following variables: city centre density, average distance to work and public transport accessibility.
  • 20 This is echoed in a number of existing research.  In a US study, Glaeser and Kahn (2003) have found that “holding family income and size constant, gas consumption per family per year declines by 106 gallons as the number of residents per square mile doubles” (Glaeser, E. and Kahn,M (2003). Sprawl and Urban Growth. Handbook of Regional and Urban Economics).
  • 21 For instance, Glaeser (2003) found that in metropolitan areas like New York, the average urban family consumes more than 300 fewer gallons of gas per year than their suburban counterparts. Glaeser, E. and Kahn,M (2003). Sprawl and Urban Growth. Handbook of Regional and Urban Economics
  • 22 CO2 emissions estimates are based on standardised assumptions about dwelling occupancy and energy use.
  • 23 NHBC (2021), New Home Statistics Review.