Cars & traffic

Jul 31, 2014

Are government traffic congestion forecasts too high?

A draft Infrastructure Australia report says officially forecast traffic congestion is double what's actually happening; flawed methodology overstates the need for new urban roads

Alan Davies — Editor of The Urbanist

Alan Davies

Editor of The Urbanist

BITRE urban congestion growth forecasts for the decade to 2011-12 versus actual growth. (source: Infrastructure Australia draft report, Spend more waste more)

I don’t think I’ve ever seen a more provocative government report than Infrastructure Australia’s draft industry consultation paper on roads, Spend more, waste more (1). It was hastily withdrawn last week by the embarrassed agency after the Fairfax press reported on it under the headline, Billions spent on roads in “hideously inefficient” way.

Its overall message is there’s too little focus on efficiency and value in the way road investment is managed by all levels of government. (2)

It’s debatable if such an important message from a government agency is well served by dramatic phrases like “Australia has a true gambler’s addiction to roads: the money spent is not a rational investment”; or that the nation’s near $20 billion yearly spend on roads “can only be described as hideously inefficient”.

And it’s unusual, to say the least, for a document designed to promote consultation to state baldly that efforts to get better road funding outcomes “should bypass road agencies, which in most observed cases will only suffocate or over-complicate such opportunities if given carriage of them”.

But underneath the sometimes intemperate tone, there’s a veritable motherlode of fascinating material. There’s too much to cover in one sitting, so this time I want to present what it says about forecasts of road congestion in cities. This section is brief and the tone is relatively sober; here’s the money quote (emphasis is as per the original):

The Bureau of Transport Infrastructure and Regional Economics (BITRE) publishes growth forecasts for traffic congestion in capital cities. Its current projections were set several years ago and assume that urban congestion in major cities will continue to grow steadily.

In 2010, Infrastructure Australia pointed out to BITRE that the statistical approach to these projections appeared deeply flawed and that the assumed steady growth was not being witnessed in actual congestion results, which instead suggested urban congestion levels were growing at around half or less than half the forecast rate:

At the time – and a number of times since – Infrastructure Australia asked that these forecasts be reviewed and their implications for current and planned urban road projects re-examined. In 2014 – four years later – no such review has taken place. It is not clear to Infrastructure Australia why this has not been attended to, although revising forecasts down would almost certainly result in pressure to reduce budgets for urban road programs.

Accordingly, the road system’s flawed analytical assumptions about matters of national importance appear to be giving the wrong impression to governments of the relative importance of road congestion (and the need for road spending) and public transport projects.

That’s about as self-explanatory as you can get. Doubtless BITRE has another interpretation which it’s unlikely to share publicly but, as the exhibit from the report shows, Infrastructue Australia believes there’s an enormous disconnect between the level of congestion forecast by BITRE over the ten years to 2011/12 and what in fact happened.

The charge is plausible because the inability of public agencies to adapt to what is a widely recognised structural change in the level of travel growth isn’t confined to Australia; here’s an equally remarkable US example. What’s especially worrying in the Australian case, though, is the charge that the change is being ignored.


  1. The full title is Spend more, waste more. Australia’s roads in 2014: moving beyond gambling.
  2. One of the report’s key ideas is to promote increased private investment in strategic roads, particularly those used for transport of freight, through implementing more sophisticated user charging systems for large trucks.
(Visited 3 times, 1 visits today)


Leave a comment

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

10 thoughts on “Are government traffic congestion forecasts too high?

  1. David Penington

    Forecasts of time savings from new roads are ridiculous too. I recall the Melbourne Citylink promised to save me 30 minutes off what was then a 30 minute commute! In fact Citylink was no faster than the original off-Citylink trip.

  2. Teddy Bear

    Duke, AD,

    The point of the chart is test to the predictive ability of BITRE’s modelling, ie “Did they get it right?” It is simply not possible to tell whether or not the 2012 projections are “better”; we need to wait to 2022.

    In the same vein, there was an interesting report by the Victorian AG which found tolls roads typically over predict traffic volumes while freeways (ie Pakenham by-pass) often under predict. Cynics (ie people not contributing to this discussion) suggest that this shows a bias that justifies toll road construction on one hand while prompting congestion relief from freeways one the other.

    Of course there is nothing new in IA’s report. The question is how did the author manage to get the report written and type set without the usual organisation oversight? It is not the effort of one person. Do we see an organisation that is a) feed up with lip service being given to rational decision making and b) a sense there is nothing left to loose?

    IMO it would be a shame if we lost IA. It is one of a few opportunities to promote rational investment decisions across transport and other public utilities.

  3. hk

    The existing land-use and transport system for the Melbourne Statistical Division (MSD) operates at a very low Level of Service (LoS) at many locations several times a day on more than two hundred days a year. Clearly the demand to move freight and people exceeds the current urban space and capacity available. Without modification in an integrated land-use and transport system, for the widely accepted view that the MSD population will top 5 million by 2051, the LoS of the urban infrastructure system can be predicted to be even lower if current operation practices continue.
    Location specific capacity increases need to be addressed though policy, planning and investment decisions.
    To the layman in the road network, congestions begins to become an issue of irritating inconvenience when their vehicle is stopped in a queue equivalent to more than one traffic light change. This sort of avoidable travel time increase is perceived as negative as it means sitting in a vehicle longer than is necessary under ideal conditions to be found elsewhere and maybe at other times. The congestion causing the in-vehicle time increases unhealthy sedentary travel time as well as longer exposure times to in-vehicle cabin pollution.
    The wide variation in the forecasting of vehicle kilometres of travel into the future is certainly a cause for concern; however the prediction of more kilometres of travel is not the main issue for many in the lay community as they wait to get through the hot spots of congestion. Feedback indicates the issue of congestion management boils down to controlling the number of vehicles trying to get through a pinch point at peak times of the day. Most of these present day congestion red spots have been known for more than 20 to 30 years, and written about by the RACV, the TCPA MTAG and many more community advocate organizations. Social capital research needs to now focus more on understanding who participated in the past and why the possible cost effective technological solutions to very obvious and costly congestion problems are still being prevented.
    In itself the level of over estimation of the congestion seriousness increase by one capacity model or another, whether it is by VicRoads, IA, the BITRE, ARRB or another well qualified organization is a side issue.

  4. Tom the first and best


    I suspect that the reason that the Pakenham bypass overshot the traffic predictions is that the sift from the South Gippsland highway through Cranbourne was not predicted. A lot of South Gippsland Highway users now use Koo Wee Rup Rd to travel between the South Gippsland highway and the Pakenham bypass. This has caused the current state government to order the construction of a bypass of Koo Wee Rup.

  5. Austin M

    The issue I have with this is that some well publicised projects seem to underperform in regard to traffic numbers and some seem to over perform like Pakenham Bypass etc.
    The problem with traffic modelling on the large scale is that you can’t fundamentally predict how land use and population changes will occur with a project i.e. will it all change at once or over 50-100 years. Some projects might trigger land use change and growth whilst the same project with a slightly different set of circumstances may have a completely different result. Fundamentally I also think traffic models fail to take account of the human effect i.e. peak spreading, rat running, signal reviews, forgone opportunity etc. What often happens IMO is that the “cost of congestion” calculated at a later date may fail to adequately account for the “forgone opportunity costs” associated with a potential development or improvement.

    I also think that the BCRs for individual projects and the results they ascertain are fairly specific and measurable. This discussion really comes down to government’s decision making and if it’s driven by the fear of doing nothing or the opportunity in doing something. Given the short terms of our governments I think they are more driven by the short/medium term political opportunity associated with projects and are really not concerned about some distant congestion issue (i.e. the multi $billion congestion numbers are just another selling point).

    As for federal departments questioning the effectiveness of road agencies it’s just a bit of good old fashioned cross government mud slinging. Just like the states saying the feds are rubbish at their management of telcos or defence and the feds saying medical and education is a basket case because of the sates. The person watching always thinks they could do it better but I doubt the road system would be suddenly better if managed by the feds rather than by current states or local councils. It would just be pork barreling at a new level and less local attention/content/understanding (hardly a recipe for successful management of what can be diverse and complex issues).

    Getting stuck into BIRTE in the report and failing to acknowledge more recent work stinks of content bias along with what seems is unsubstantiated claims which is not really acceptable in what I would assume should be an impartial assessment to government.

  6. Socrates

    Re BITRE see

    Report 128 similarly projects continued long term increase in freight traffic. That is probably more correct.

  7. Socrates


    I disagree. If you are referring to BITRE report 127, it only allows for a short term drop in growth during the GFC, then assumes a return to “business as usual” traffic growth post 2014. I find this highly questionable given long term trends in declining per capita car usage. BITRE 127 is not based on any predictive model built with causal factors. It is essentially trend projection. IMO that is part of the problem.

    I think the IA report is accurate regarding waste in road projects. Where it is (ideologically?) weak is that there is also waste in public transport projects. Nor is private funding a panacea, as recent history has shown. The whole system needs to be overhauled.

  8. Alan Davies

    duke the lost engine #1:

    Good point. I wonder what Infrastructure Australia (IA) would say (or perhaps more likely, the author) to explain why it didn’t mention the BITRE report? It might say that it was too belated (2012); or that it doesn’t have forecasts for capital cities; or perhaps that it didn’t change BITRE’s official projections as communicated to IA. Whatever, the IA report should’ve acknowledged the BITRE work. IA’s report is important because it shows that at least some parts of the organisation are very frustrated with road agencies; but that doesn’t necessarily mean it’s right in all respects.

  9. Smith John

    So how long will it take the planners to come to grips with the new reality (assuming it does turn out to be a long term change)?

    My guess is: about 20 years.

    As they say: the dominant paradigm doesn’t change because people change their minds. It changes because old people die and younger people think differently.

  10. duke the lost engine

    surprisingly not mentioned are BITRE’s published revised forecasts in 2012, which explicitly take account of slowing traffic growth:

    i haven’t looked at the methodology, but the ‘base case’ forecast traffic growth averages to 1.5% per year, pretty consistent with the last 10 years growth rates reported by Infrastructure Australia

Share this article with a friend

Just fill out the fields below and we'll send your friend a link to this article along with a message from you.

Your details

Your friend's details