Most mathematical modelling used to guide our economy is simplified and only modified when it becomes so out of touch that it is dangerous. When the Atlantic cod fishery collapsed the model being used to set fishing quotas was still suggesting the fishery was healthy. In retrospect it seems a kind of madness to have kept using it.
Traffic modelling in Australia is now similarly out of touch. A recent study by the Bureau of Transport Industry and Resource Economics modelled the future of traffic in Australian cities. If it had been a mere academic study it would not be dangerous, but it is now being used to justify massive road spending.
The report suggests that travel times will blow out and road congestion will cost the economy more than A$50 billion. Congestion, it seems, is out of control and will engulf us all, with terrible economic consequences for Australian cities.
Modelling relies on heroic growth assumptions
In order to reach these numbers the report had to make assumptions about growth trends in car use per capita. The high-growth scenario is included in the red box in Figure 1 below.
As is common in such modelling, a low-growth scenario is also modelled and then the future is taken to be in the middle.
The high-growth scenario can be seen to reverse a trend in car use per capita that appeared to be well set after a peak in 2003-04. This peak car phenomenon has been found in all developed cities and has been analysed by many commentators including ourselves.
Understanding the peak car phenomenon would appear to be a basic requirement before attempting future scenarios that wipe it out completely. However, this report has not ventured into such questions. Instead, it continues the methodology used for the past 30 years.
Travel time reality means car use has peaked
The fundamental problem is that the model does not understand how cities work. Travel time is seen as something that just expands as our cities of car-dependent suburbs grow outwards.
However, the Marchetti travel time budget of just over an hour on average has been found to apply universally across all cities. Some people can go beyond an hour and some much less, but the average everywhere is an hour. This has been found over and over to apply in every city.
If people find it hard to live with so much time “wasted”, they move to somewhere more within their travel time budget. If the overall options in housing, jobs and travel become so dysfunctional that people are forced to travel beyond their time budget then the issue becomes highly political: elections are fought over infrastructure and housing options. Cities adjust; they don’t keep expanding travel time.
It is possible to understand the global peak car phenomenon in terms of cities hitting the Marchetti wall. In our data, cities everywhere began to grow in their traffic congestion whether or not they built freeways or extra road capacity. This was because these just filled very quickly.
Rail projects unclog the urban arteries
Rail projects, however, could go around, over or under the traffic. Hence, over the past 30 years, rail lines have become more travel-time-efficient than traffic arteries. Many people, especially the young and wealthy, began to move back into areas where public transport was well provided.
The rejuvenation of central and inner areas is not just because they are cool and trendy but because they offer reduced travel time. Car use per capita goes down when there is less need to travel, especially by car.
This model does not consider this to be significant enough to consider in forecasting scenarios for Australian cities. It crudely projects a totally car-based-and-growing set of futures. It barely considers the remarkable global turnaround in the building of rail systems and the rejuvenation of cities.
Scenarios where these urban trends are considered to continue (rather than suddenly dying as suggested by Figure 1, followed by inevitable growth in car use) would make much more sense. This is the way planning is being done in cities like London.
The evidence in London and other cities – with good public transport alternatives and competitive door-to-door journey times – is that with fixed road capacity traffic volumes stay constant (or slightly decline). All the growth goes on to public transport and the car-based share of travel falls – down from 46% for residents of London in the 1990s to 32% now.
The scenarios for London are therefore all about how best to build public transport and land-use options that can enable economic growth to occur sustainably, rather than trying to increase road capacity. Similar considerations can be seen to be happening elsewhere around the world.
Did an ‘Abbott effect’ skew the modelling?
So why have we refused to change our modelling and are still trying to justify massive road capacity increases?
Traffic modelling reports like this take about two years to produce. What was happening around two years ago was the election of Tony Abbott with his commitment to spend some A$40 billion on urban roads and nothing on urban rail.
Huge road projects in Melbourne (East-West Link), Sydney (West Connex), Brisbane (Gateway) and Perth (Freight Link) were all announced before any assessment. Their rationale was very shaky and most of these projects have now either failed or are failing.
Each road project also has large impacts on their urban economies. These are dangerous as they destroy so much of the urban fabric necessary for rejuvenation and take away the ability of governments to pay for the more important urban rail projects on their agendas.
The upturn in the graphs in Figure 1 seems to be an Abbott effect. It is rather amusing now that this report came out just as Malcolm Turnbull took over and began talking up urban rail and urban regeneration.
Such models should be put into the museum and we need to start creating a better sense of where our cities can be going. This report can be easily passed over, but if the thinking behind it stays the impact on our cities will be very damaging.