Earlier this month, urban policy researcher Todd Litman published a useful summary of some of his new research into the cost of sprawl:

Our analysis indicates that by increasing the distances between homes, businesses, services and jobs, sprawl raises the cost of providing infrastructure and public services by 10-40 percent. Using real world data about these costs, we calculate that the most sprawled quintile cities spend on average $750 annually per capita on public infrastructure, 50 percent more than the $500 in the smartest growth quintile cities. Similarly, sprawl typically increases per capita automobile ownership and use by 20-50 percent, and reduces walking, cycling and public transit use by 40-80 percent, compared with smart growth communities. The increased automobile travel increases direct transportation costs to users, such as vehicle and fuel expenditures, and external costs, such as the costs of building and maintaining roads and parking facilities, congestion, accident risk and pollution emissions.


We estimate that in total, sprawl costs the American economy more than $1 trillion annually, or more than $3,000 per capita, and that Americans living in sprawled communities directly bear $625 billion in extra costs, and impose more than $400 billion in additional external costs. This is economically inefficient and unfair: it wastes valuable resources and imposes costs on people who do not benefit from sprawl.

These findings should not be particularly surprising to regular readers of Transportblog – or, indeed, to anyone with an elementary understanding of geometry. (Serving dispersed suburbs with network infrastructure is more expensive.) But the magnitude of the costs is impressive.

I was particularly struck by the following chart, which illustrates the amount of space required for various different transport modes. Litman estimates that each automobile requires a total of 80 to 240 square metres, mostly for parking. By comparison, walking, cycling, and public transport require less than 20 square metres per passenger:


As I’ve written before, space is expensive in cities, which means that we must use it efficiently. Moreover, the cost of space for cars is rising rapidly, while demand for driving is levelling off. In this situation, devoting more space to roads and parking – or preventing the re-use of road space and parking lots for other purposes – may represent a significant misallocation of resources:

AHB traffic volumes and real house prices, 1961-2014

So, we might ask: How much space have we misallocated as a result of our bias towards building roads rather than public transport and cycling options? And what else could we be doing with this space instead?

First, some data. As we tend to build road networks for peak demands, they tend to have spare capacity during the middle of the day and evenings. Consequently, I’m going to focus on trips taken in the morning peak. This is a conservative view on the space required for a car-based transport system, as cars used during off-peak times still require lots of parking.

According to modelling results reported by Wallis and Lupton (2013), in 2006 there were around 450,000 vehicle trips taken during the morning peak. (See Table 4.1 in their report.) Most of these are trips in single-occupant vehicles. How much space do we need to accommodate all these vehicles?

Based on Litman’s figures, travelling by car requires an average of 150 square metres of space – around 40 square metres of roads per car when moving and 110 square metres for parking. This implies that the 450,000 vehicles moving around during the morning peak occupy 67.5 square kilometres of space, including (at minimum) 18 square kilometres of roads.

That’s a lot of land. Urban Auckland covers a total area of around 544 square kilometres, which suggests that we’re using up around 12.4% of the city’s land area simply to move single-occupant vehicles during the morning peak and warehouse them during the day.

If anything, this is probably an under-estimate of the spatial cost of Auckland’s car-based transport system. A 2013 UN-Habitat report on streets as public spaces and drivers of urban prosperity found that Auckland devotes 14% of its land area to roads alone. Parking is likely to cost us even more space, as this map of Manukau central shows. Everything that’s not coloured in red or green is a carpark or a road:


But regardless of whether we’re dealing with 12% of the city’s land area or 30%, we’re talking about a lot of expensive space devoted to moving or storing cars. Even a modest reduction in the share of people travelling by car in the morning peak would save us a large amount of land.

Let’s say, for example, that we’d invested in better transport choices that enabled 10% of the people in cars to shift to public transport or cycling, which require around 10 square metres of space per person. If we had done so, we would have had an extra 6.3 square kilometres of land that didn’t need to be used for roads and carparks. [6.3km2=45,000 vehicles*(150m2-10m2)]

This is valuable land. Assuming an average land price of around $500 per square metre, it’s worth $3.2 billion. And it could be used for so many more things – housing, businesses, public parks, schools, etc – if it hadn’t been gobbled up by our space-hungry transport system. If it had been developed to the same density as Auckland’s average neighbourhood – around 43 residents per hectare – instead, it could have housed 27,000 people.

In a city that’s struggling to find enough space to house it’s growing population, this amounts to a minor scandal. Since the 1950s, local and central governments have spent lavishly on roads and neglected public transport, walking, and cycling. Those decisions have inadvertently contributed to our housing woes today, as they’ve saddled us with a space-inefficient transport system and a shortage of developable land.

At the moment, local and central governments are looking at ways to get best us out of their own properties. Auckland Council’s setting up Development Auckland to manage and develop its substantial land-ownings. At this year’s Budget, the Government announced a more hastily-developed plan to sell off a significant chunk of the land that it owns in Auckland for development.

Given their interest in the subject, they should also be thinking hard about how to minimise the “opportunity cost” of Auckland’s space-hungry car-based transport system. Investing in public transport and safe walking and cycling can allow us to move more people without gobbling up more valuable land.

What do you think about the spatial cost of our car-based transport system?

Share this


  1. The roads would still be there whether 500 or 5000 cars travelled on them. Connectivity between geographical points would remain. Playing maths games with per vehicle / occupancy etc is great in theory – reality is that those links between suburbs, cities, etc would still be there regardless of traffic volumes. Admittedly maybe not so wide or the need for wider based on user volumes, but the distances would still be exactly the same. Parking issues can be solved efficiently by building zero running cost multistory above ground parks. Parking ‘waste’ in Auckland is mostly down to single level parking areas.

    1. “zero running cost multistory above ground park”. I think you’re dreaming. And that’s before we get to the capital cost of such beasts.

    2. No building is “zero running cost”, and I think you’re forgetting the high capital cost which doesn’t stack up against the revenue as well. See reply to J below.

    3. If you look at what I’ve done, I’ve actually only assumed a very marginal reduction in space used for roads. 14% of Auckland’s land area is used for roads – around 76 square kilometres. I’ve considered a reduction of less than 2 square kilometres, plus a reduction in parking area of around 4.5 square kilometres. This would leave 97% of the city’s road surfaces thoroughly intact.

      1. I have read somewhere that Auckland has a low percentage of land dedicated to roads i.e too many dead end cal-de-sacs and so on. I think roads will still be needed for Bertaud type connectivity reasons. In fact maybe more ‘roads’ in the form of dedicated ROWs for cycling and PT are needed.

        I think the real savings lie in reducing land dedicated to parking.

  2. The house prices and traffic volumes graph makes no sense and nothing useful can be derived from it. House prices in Auckland are influenced by factors totally non related to vehicle use. House prices in Auckland are influenced by immigration, Chinese buyup and cheap interest rates. The blue line is of interest, the red line irrelevant.

    1. Yes, I agree that house price trends are largely exogenous to vehicle volumes. That’s exactly the point. When land prices change, for reasons outside the control of transport agencies, it behooves the agencies to realise that this is happening and respond to it by changing their investment strategy.

  3. Once again I’m going on a tangent, but I think it’s within the topic ie the cost of space for cars:
    I’ve always been fascinated why AC (or is AT?) build parking spaces for their Park ‘N Ride facilities on flat land. Why not use up less space by building a four level car park? Oh, because it costs too much? Well, that’s been my other thought: the true cost of providing PT isn’t always factored into the stated cost of PT. But let’s not get into that now (since the true cost of most things is not declared for most everything) since I’d rather throw this suggestion into the mix: since AT are now finding that their Park ‘N Ride facilities are proving so popular (ie all car spaces filled by 7-8am) and they need to expand them why not get some PPP’s going. Let private investors/companies build a four level carpark on the existing space at their cost, in return for charging motorists to use the facility – with the exception of one level that remains free ie ratepayers are keeping the free parking spaces they paid for. I’m guessing the parking operators would allocate those free spaces on the rooftop since that would be uncovered. I’d guess also that the deal would be predicated on an X% return for the PPP and Z years ‘free’ before the parking structure is returned to AT/AC/Ratepayers. You could also peg the cost of the paid parking in some way so that it doesn’t become a cash cow for the operator and continues to encourage drivers to use PT.

    1. As per http://greaterakl.wpengine.com/2014/05/15/what-do-carparks-cost-to-build/, multistorey carparks would cost something like $40,000 each to build. Say the operator wants a 6% return, which is probably on the low side. Then they need to get a return of 6% x $40,000 or $2,400 a year. That’s $6.60 for each of the 365 days of the year. Now, taking out weekends, and allowing for operational costs – rates, security, maintenance, cleaning – you’d be needing at least $10 in revenue for each workday. And there aren’t many places in Auckland where that would work.
      So a PPP wouldn’t work. The only reason Park N Rides are full is because they charge a below-market price, and the market prices wouldn’t be high enough to justify building them anyway.

      1. Yes, but this goes to the thrust of the original post above. Why should x00 lucky car owners who are in first get free all day parking? Why not make ALL car users pay the true cost of their cars/space? $10 a day is peanuts – other countries charge that [some even double that] per hour just to park.
        The wider answer/solution of course is to build medium/higher density houses close to PT stations so that people can just walk there rather than having to drive there. Maybe another solution is to require people wanting to use the Park ‘N Ride facilities apply for a permit to use the car park. People who live within a reasonable walking distance or connecting bus route [mobility issues excepted] wouldn’t get one, thereby saving space for those that live further away and are more deserving.

        1. The immediate solution is to price the parking. If we decide we want to give some part of a PT journey away for free to encourage uptake then make the bus or train free but charge for the parking to more sensibly ration it, and to price the ‘bad’ not the ‘good’.

        2. “The immediate solution is to price the parking.” This is the crux of both increasing the AC revenue and reducing road congestion caused by one occupant cars. The charges for the parking spaces controlled by AC should be such that 15% of them are vacant at any one time.
          The charges for occupying parking spaces should be low for the first hour slightly higher for the second hour and then substantial for any period after that. Parking on the street should only be available when it does not impede traffic flow (Traffic includes cycles and pedestrians). All routes carrying public transport should be no parking. This should aim to have bus lanes where the bus stops in the lane (no need to merge with other traffic).

      2. I think your dollar values for car parking are a bit high
        In the 1990’s a central Hamilton car park was valued at $12,000 (not inflation adjusted), that decade they remove the free market approach of “property owners willingness to supply and car owners willingness to pay”, and forced all new investors to be investors in car parking with the introduction of MPRs in the Hamilton central area.
        To day the value of a car park in central Hamilton area is $11,000, cost about $19,000 to build.

  4. Christchurch is the poster child for car-driven [!] wasted land use. It was before the quakes and is of course worse now. This massive inefficiency will be an increasing drag on its economic performance. Once the exogenous rebuild money is no longer flowing in it is likely that the high costs of this totally auto-dependent form will show up as mediocre growth. Essentially it is being planned to avoid agglomeration; an un-urban urbs.

    Of course parking lots are potential building plots; but with such dispersal will the business case be there to develop them well once the outside once-off cash has been spent? Especially when so much is going on parking structures to ensure auto-dominance?

  5. If Litman’s graph is true, then it is a damning piece of evidence on how we ruin our cities and waste valuable land on things we really don’t need.

    1. I think it is more about all the parking space required to store vehicles that are not stored very efficiently and may never really be stored well.

  6. Not central to the discussion but motorcycles take half the space of cars for travel. In some instances (queues) they actually take up car space so they are even more efficient.

    1. Most of the space taken up by cars when travelling is the leading distance to the next car. Are motorcycles not required to follow the 2 second rule? And until we have a narrow motorcycle lane, they will still take up the full width of a car lane.

      1. Motorcyclists follow each other in a staggered pattern with the first rider on the right side of the lane, the second on the left and so on. So you end up following the rider two in front with a two second rule. So equal safety with half the lane space. That is at highway speeds. Around town the efficiency is better than double due to the size of the vehicles. Also the pickup with a motorbike (and the increased attentiveness of the riders) means they are far more efficient at clearing intersections from a queued start than cars.

  7. AC is able to govern the charges for it’s parking and I believe it has a mandate to do so (via the consultations held a short time ago). This would have a twofold effect of decreasing congestion and increasing revenue and should be implemented as soon as possible.

Leave a Reply

Your email address will not be published. Required fields are marked *