Does higher density mean lower car use?

Travel outcomes (Barnes)

Not necessarily – in fact in the US, not even usually!

It’s a truism that denser, more concentrated cities tend to have higher public transport use. Various studies have confirmed this intuition but what is usually left unexamined is the implicit assumption that such cities consequently have lower car use.

This study of 31 of the largest cities in the US found that assumption is not correct. Higher public transport mode share does not translate on average to lower kilometres of travel by car, shorter commutes by car, or lower levels of traffic congestion.

The primary finding “is that land use, at least at the aggregate level studied here, is not a major leverage point in the determination of overall population travel choices”.

Undertaken by Gary Barnes from the Centre for Transportation Studies at the University of Minnesota, the research found that, if anything, “the higher densities that increase transit share tend to increase commute times and congestion levels”.

The main objective of the project was to identify the effect of land use on travel behaviour. Most studies concentrate on the effect of average density on one or two variables, usually transit share and sometimes total kilometres of car travel.

Barnes’ approach was much more extensive. For each urbanised area, he defined 15 descriptors of travel behaviour, 11 of land use and 15 other, mainly demographic, factors. Moreover, he employed the concept of ‘weighted density’ (he calls it ‘perceived density’) to more accurately describe the distribution of both population and employment in each city.

Barnes confirms that residential concentration increases transit’s share of travel, but he notes the effect is not large. Contrary to the underlying assumptions of much urban policy:

Even very large changes in land use have very little impact on travel behaviour, in good ways or in bad. Apparently the larger effects sometimes observed in neighborhood-scale studies are just that: neighbourhood-scale effects that do not extend their benefits to the larger urbanized area.

His analysis implies that increasing residential density by 100% would increase transit share by only 5-6%. To get a 1% increase in walking and cycling’s combined mode share would require an increase in residential density of 5,000 persons/mile2 (1,931/km2). Similarly, a 14% increase in density would only yield a 0.5% decrease in in-car travel time per person.

He therefore cautions against extrapolating from the experience of one or two cities to make general rules. As the accompanying chart shows, the most concentrated city in the US, New York, fits the paradigm of the sustainable city – it has comparatively high transit use, low car kilometres of travel and low traffic congestion. But that pattern is not true of most of the other 30 urbanised areas examined by Barnes.

For example, some cities with high transit use, like Baltimore and San Francisco, still have high car kilometres. San Francisco also has high traffic congestion. Conversely, some low transit use cities – for example, Buffalo and Miami – nevertheless have low car use. San Antonio has low transit use and high car kilometres, but low congestion.

If New York is the paragon of sustainability, Los Angeles fits the stereotype of environmental vandalism. It has the deadly triad – low transit use, high car kilometres and high congestion. Yet Dallas and Kansas City, which also have the low transit use and high car kilometres of the sprawled city stereotype, nonetheless have low congestion. And Los Angeles also has the highest (average) density of any large city in the US.

As Barnes emphasises, there is enormous variation across the 31 cities. Virtually any combination of high or low transit share, kilometres of travel, and congestion can be observed in some city, he says.

He offers two possible explanations for why high transit share isn’t necessarily – or even usually – associated with lower car use. One is the interconnectedness of individuals’ travel decisions. Time saved on one trip may simply be used for additional travel somewhere else. The other reason is the interconnectedness of travellers. For example, “if some people cut back on driving, then road capacity will be opened up that others may take advantage of”.

In my view, a more direct explanation is simply that there are only small parts of US cities that have the conditions which make cars uncompetitive against public transport.

This study should sound a warning for those who uncritically assume that the experience of what works in a city like New York – or doesn’t work in a city like Los Angeles – can be transferred holus bolus to Australian cities. It should also encourage policy makers to think more holistically and systematically – there’s more to making good policy than just a single ‘headline’ variable like mode share.

Barnes has the usual caveats, counselling caution at many points about the limitations of the data and what can and can’t confidently be inferred. But it’s reassuring when an author attaches a summary of the actual numbers he used and the regression coefficients.

This research was published in 2001 and is based on data from 1991 and 1995. That might seem a little old but the relationships it examines aren’t likely to have changed significantly since then (and remember that Newman and Kenworthy’s influential work on the relationship between density and energy use was published in 1989). It also relates only to US cities – that’s a restriction but I would argue that the experience of ‘new world’ countries is more relevant to Australia than that of Europe.

This is a relatively short report and its one of those rare ones that is really worth reading closely. There’s much more to it than I’ve covered here. There is, for example, an extended discussion of daily travel time budgets and some speculation that we might have persistent, life time travelling habits. I’ll write further about the implications of this research on another occasion.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s