Many times in the last few years we have highlighted a ‘flat-lining’ or at least slowing of growth in car travel across New Zealand. The same trends have been seen in many overseas cities and countries – with the slowing in the UK dating back at least 20 years now. Yet for some bizarre reason this change hasn’t filtered through to those making projections about future traffic growth. In the UK we have seen projection after projection forecasting significant growth – even though consistently it doesn’t happen.

The same process has happened in Washington state, where the Department of Transportation has ignored the flat-lining of traffic growth and continued to forecast significant increases – despite all evidence suggesting they need to change:

Lance Wiggs has picked up on a recent report released by Treasury that looks into the evidence behind the key elements of the transport sector. Towards the end of the report Treasury analyses some projections of future transport demand prepared by NZTA and NZIER. Let’s let Lance pick up the story here:

This line records use of a certain item by New Zealand population since 2000. Where do you think it will be in say 20 years time?

The statistic went up, and then down, and so the best estimate to me would be a flat or downward trend. But rather strangely the authors of this chart determined that all of their estimates would be up, and that the lowest change would be a substantial increase. Here it is, with my added red line eyeball trend:

The chart, of course, is estimated vehicle use in New Zealand, expressed in millions of kilometres travelled. It’s sourced from the “ National Long-Term Land Transport Demand Model, NZIER and NZTA (2013)”, and that’s a critical model as it feeds into all sorts of cost-benefit analysis and policy for transport in New Zealand.

The data is sourced from this report, which appears to make the same mistakes as transport modelling projections in the UK and Washington state have done for the past 20 years – by refusing to believe that anything ever changes about how people travel. Even though evidence to the contrary is absolutely everywhere. Similar bizarre conclusions are made about future levels of public transport use. Back to Lance:

The bias towards cars is also reflected in this chart of forecast public transport statistics – which doesn’t pass the giggle test either. Witness the trend and the projections below (the red is my version of the trend-line):

So for some unfathomable reason the growth in the use of public transport is forecast to immediately and dramatically fall, while the growth in the use of cars will immediately and dramatically rise. It’s ludicrous.

What’s interesting is that a footnote in the Treasury paper notes that these projected trends are completely at odds with what has been happening in recent years. It says:

We note that, although the NZIER / NZTA model predicts a gradual decline in the future, the public transport share of passenger kilometres has shown a steady growth trend over the past decade. This may merit further monitoring and consideration as time progresses.

It seems that at least the public transport projections are a little bit too insane for even Treasury to believe.

In any other area of government activity or a business if the computer models were projecting the complete opposite of what’s happening, what’s been happening for quite some time and what’s happening in cities and countries all over the world, we’d chuck the models out and start again. Yet for some reason we keep believing these illogical outputs and use them to determine where and how to spend billions of dollars of public money. It’s quite disgraceful really.

Share this


  1. Are there any models out there (by universities, consultancies, PT operators..) that have produced better fits with either car or PT use over the last couple of decades? If so, why haven’t these gained traction?

    1. Or are the forecasting models themselves really the problem? Because if the (road traffic) models were the problem, why would they always overstate traffic? The error smacks of deception or lying or both.

      So here’s an interesting read.. it’s in the context of Australian toll roads from 2011, but may have broader applicability, since it references large European studies (Flyvbjerg and others) that also looked at rail projection errors.. which states the problem in quite nice technical terms “optimism bias and strategic misrepresentation” and helpfully adds a categorisation of “actors” identifying who the likely culprits are:

      – local transport authorities
      – local politicians
      – local economic interests
      – local civil servants
      – consultancy companies
      – individual MPs
      – PPP bidders (where they both the builder and operator)

      That probably covers most of you out there…

      On the other side of the ledger, “having a direct interest in avoiding optimism bias” is the ministry of finance, PPP auditors and (in the UK) the department of transport.. here I guess you might put the latter on the other side.

      Having identified the culprits, the paper lists 21 ways that road traffic projections can get inflated. Numbers 5 and 8 seem particularly apt: “the future will look exactly like the past” and “remove inconvenient truth (removal of true observations as outliers)”

      Unfortunately, one of the remedies looks quite hard: “institutional change with a focus on transparency and accountability”.

    1. Come on Tim, why would they when there is so much real news out there?

      I am sure there is some little, old lady who will be “forced” into an apartment by the Unitary Plan to interview – you know, the kind of hard hitting investigative journalism NZ is famous for.

  2. I sat through a presentation from an economist from NZIER at last year’s transport conference making this prediction. It made no sense then and when I asked him how he explains the pattern of the last eight years he said it was a ‘shock’. In other words an inexplicable temporary break in an otherwise permanent rise. He apparently could not or would entertain even the possibility that the has been a change in the underlying pattern of behaviour. Again this makes no sense. I would like to ask him now exactly how long does it have to go on be before he would entertain the possibility that it represents the new normal. Given that his talk involved lots of pretty charts and convoluted readings to try to explain away the data, all full of disclaimers about how hard it is to make predictions, I suspect he may take this ‘shock’ theory to the grave rather change his assumptions.

    In essence this guy was saying that the world has been ‘wrong’ and must (any minute now!) get back to its rightful role of following what the model says. The kindest way to describe this kind of delusion is cognitive dissonance. Interesting to see just how widespread it is world wide, but also important to note that it does of course make a curiously perfect fit with the current governments stated policy.

      1. of course, he is an economist. Economists only know charts that go up. Why is it that we always listen to economists even if we know they are mostly wrong or biased?

    1. That kind of cognitive dissonance (ie relying on computer models rather than real world experience/measurements)occurs in other areas of science too, unfortunately. I guess it has something to do with patch protection.

    2. Isn’t that attitude also called “Group think”, when the individual warnings are ignored and overruled by the group.

      Which supposedly was given a bad rap/shown the door after the Challenger disaster in 1986 when previous O-ring failures in the boosters were ignored as a safety issue, and then again after the next Shuttle disaster when ice falling from the shuttle at launch time and hitting the wings was also ignored as a safety issue.

      In both cases the prevailing attitude of the Engineers at NASA was “well its always been like this/done this” (and never caused a problem…) – famous last words indeed.

      Of course, your NZIER economist, is really saying that he believes the last 8 years are really an “outlier” – both here in NZ and also around the world of transit. So thats why he is so shocked.
      And I think he believes that its better (for his job security) to overbuild the roads, than underbuild them.

      It reminds me of the old IT mantra that used to be around about “no one ever getting fired for buying IBM”
      Of course times change, and plenty of people have got fired for doing exactly that since and maybe its time for some economists and traffic planners to be shown the door as well.

    3. “…In essence this guy was saying that the world has been ‘wrong’ and must (any minute now!) get back to its rightful role of following what the model says. ..”

      Actually, I think you’ve distilled the essence of economic science in that sentence.

  3. It would be really interesting to see details of the models used by these planners along with the inputs and assumptions the models need. Given that all the examples above produce the same outputs, I suspect that there may only be one or two models used by planners around the world.

  4. At least as interesting is the graph at the end of Lance’s article which shows delays in the morning peak for traffic reducing despite the growing population. As he says, this can only be attributed to more PT use, higher density housing and perhaps some cycling/walking growth.

    It will be interesting to see what happens when the Waterview connection is built and the motorway is “finished”.

    So I guess if people ask an example of a city that has reduced its congestion through higher density and investment in PT, a good answer would be – Auckland!

  5. Good Article. The predictions are probably made when AT looks out their window on the carpark and counts what percentage of employees drove to work despite being attached to a train/bus station. The AT carpark is huge and should be sold and turned into further commercial buildings and apartments. And the staff should be taking public transport. And the money from the sale could go to more investment in trains and fixing the ticketing stations at the rail stops. If the rest of Auckland is supposed to squeeze up into higher density more walkable liveable city lets start with the land the public owns.

    1. +1

      And if the AT staff “need” a car – the Council could park small, fuel efficient (BEV??) cars in one of their MANY half empty parking buildings for shared use by staff. Electric bikes could be supplied for shorter journeys.

      Lead by example AT.

  6. Yes, it would be nice if AT could lead the way and get rid of their car parks, even if it were only their management showing the way..

    1. A good chunk of AT’s staff now sit in the city centre, though the “acres of car parking” situation still exists at Smales Farm offices, and to a lesser degree, at their Waitakere offices.

  7. Wellington public transport trend is bucking the global trend towards PT higher patronage!

    Another PT fare rise to keep our farebox recovery at the top end of the country, is helping to make NZTA pessimistic PT projections come right in Wellington! International trends are for declining vehicle kilometers and rising public transport patronage, yet Wellington PT numbers are flat lining, and overseas consultant ARUP, says that PT patronage will decline further once all the new RoNS roads are built. A Wellington Regional Transport subcommittee has also recommended Bus Rapid Transit instead of light rail, limiting future PT capacity on the spine and through the Golden Mile. This could also result in swopping electric trolley buses for noisy diesel ones!.

    Well, it is all about giving people poor choices!!

  8. It’s all tied in with the global oil industry, whose value is based on future projections of consumption, and projects in support of that, particularly road building. They control most western governments, and determine transport policy.

  9. Isn’t Wellington’s problem that employment is falling? The Government’s sinking lid policy, and companies moving their HQ to Auckland? John Key said it was dying, and he should know, he’s wielding the scythe…

    1. Not really. Overall number of jobs in Wellington is about 237k and it’s been bouncing around by about 1k either side of that. It is down on the peak of 244k in 2008 but is still higher than any time prior to 2006.

      Year – Auckland – Wellington
      2000 – 521,170 – 206,030
      2001 – 522,830 – 209,030
      2002 – 534,760 – 210,540
      2003 – 556,560 – 213,740
      2004 – 581,400 – 215,870
      2005 – 601,920 – 226,380
      2006 – 616,090 – 233,380
      2007 – 629,770 – 238,480
      2008 – 646,160 – 244,210
      2009 – 624,560 – 239,540
      2010 – 616,740 – 238,360
      2011 – 628,620 – 236,100
      2012 – 644,150 – 235,880
      2013 – 650,430 – 236,840

  10. Don’t you essentially rely on those same projections when you criticise the Council’s transport investment plans from the Auckland Plan by saying that despite the investment congestion only gets worse? The main reason the models show such poor results is because of the underlying assumptions around increasing per capita vehicle use. It seems that all sides of the transport debate are happy to use models when they help their arguments but then quick to discredit them when they don’t.

    1. No we don’t assume rising per capita vehicle use because there is none. But new road projects can still induce driving in a declining or flat gross VKT environment. All projects are local after all.

    2. When the model used to determine project need also shows that the projects will make things worse, you don’t think it’s in the least bit valid to criticise the projects that are proposed? If the projects can’t even make the flawed model produce positive results, the projects are also obviously flawed.

    3. Actually I’d be quite happy to drop all transport models and build infrastructure based on it delivering the the kind of city we want rather than what some computer tells us we need. As for your comment about using both sides of the model argument, I disagree and think we are still being consistent. I don’t think that anyone disagrees transport demand will increase over time which will simply be due to population increases, not necessarily per capita increases. For congestion we already see that people who can will change their behaviour to avoid it by travelling at a different time or with a different mode.
      1. If congestion model is wrong (which I think it is) then it is likely because more people are using a different mode in which case the CFN is still very valid and needed for capacity.
      2. If congestion model is right then CFN is still likely to be needed to offer relief to those that don’t want to be part of it

      1. drop all transport models and build infrastructure based on it delivering the the kind of city we want rather than what some computer tells us we need.

        Those words should be the entire Government Policy Statement on Land Transport Funding.

        +infinity bazillion

        1. Well at least it would be being honest about what’s happening and we could have a contest of ideas for what’s built rather than pretending it’s all about the numbers like we do now.

  11. On your own congestion free network web pages you use modelling that relies on the same dubious projections about per capita vehicle use to make the case that congestion is set only to get worse in spite of proposed investment. Unfortunately, it is impossible to say what the effects of the proposed projects or induced demand are if you don’t know what effect those projections have on the results.

    1. From what I have observed over the past 20 years is year on year (apart from a slight dip post RWC) increase in PT passenger trips. In fact, after the Britomart opening, rail passenger numbers have surged. Likewise the Northern Busway. These trends show us that people will use PT, as a viable alternative to cars, if we provide good frequency and timely journeys. Based on passenger numbers over the past 10 years, you’d have to agree that carrying on with providing this kind of quality infrastructure is paying dividends.

    2. Perhaps also you need to be aware that when you (as a traffic planner) expect a normal incoming tide of 500mm per hour and you see a falling tide instead in the order of 2 metres in 10 minutes, you need to wake up and maybe suspect, that there is perhaps a Tsunami coming towards you and the status quo is now the “old normal” and about to replaced by something different?

      As for traffic predictions – the absolute number of vehicles on the road will increase over time by just about any measure you make, but per-capita the rise may not be the same it was historically and will go downwards over time. So it is also fair to say that without good PT options people will stick to cars as they are not given a choice. But those PT options have to prioritise the PT over normal traffic or the PT will get stuck in traffic too, and be no better off.

      So as Bryce says – build proper PT and they will use it.
      Really its the always building it only after the roads are wide enough thats the problem here
      – the roads can never be made wide enough fast enough for a car only transport system as LA found in the 70s and the rest of the world has found out since.

      While we can’t isolate induced demand from other factors with precision, we can say that getting out of the current congestion requires a different solution to the ones that got us into the current mess.

      The CFN shows that the current thinking of “spending the proposed $s on more of the same” will not work – which is why the CFN is proposing (for once) to prioritise the spending to ensure that a roll-out of good high frequency PT is done over the continued (and currently planned) option of spending those $s on more of the same – i.e. “cars for Africa.”

      While it may not be the best solution, it sure won’t be worse than the utterly predictable results shown by the projections of congestion is we continue as we have done.

  12. Greg, I’m not convinced that any transport mode solves congestion. When you shift people from a motorway to a train or bus, the unintended result is usually that more people drive, taking the place on the motorway vacated by those people who moved to the train or bus. The problem with this, is that planners still see the same congested road, so still want to widen it even though a lot of people moved to the train or bus.

    So while PT can grow, it doesn’t seem to reduce use of roads.

        1. You are both right!

          PT improvement, like Geoff, notes means more people can travel at the same time, often faster than driving. End result is increased mobility and accessibility, economic performance etc. So thousands more on the new rail line while the motorway stays the same is indeed no reduction in car trips.

          Yet that is also consistent with what Patrick says. The motorway staying the same is in fact a decline relative to population growth over time.

          So indeed PT can grow without *reducing* the actual number of trips on road, yet if road use stays the same it in the context of growing population and growing overall transport mobility, it is effectively a reduction.

          That all just supports what Jarrett Walker is fond of saying “you get exactly as much congestion as you provide for”, and what Patrick likes to say “what you feed grows”.

          If we kept our vehicle trips at the same level, but grew to a population of two million people we’d be better off. No reduction in driving, but no 33% increase either.

          I’m pretty confident the only way to actually reduce congestion is to reduce driving, which basically means restricting access the road network through constraints. I think people drive less per capita these days due to all sorts of factors, but driving less in total in the context of strong population growth would take something else.

          I don’t think there is much chance to reduce congestion by providing alternatives, more like you just move more people on the alternatives and the traffic stays about the same. If you go really crazy then you might make it so much faster and easier on the alternatives that driving just can no longer compete, but that would be pretty intense levels of alternatives.

          My conclusion is you just accept traffic congestion happens in busy cities at peak times. The question is how much you want and where, and what alternatives you should provide.

          1. “So indeed PT can grow without *reducing* the actual number of trips on road, yet if road use stays the same it in the context of growing population and growing overall transport mobility, it is effectively a reduction.”

            Yep, but the flaw in this process is when that level of road use that is staying the same, happens to be at a level deemed undesirable. They add more lanes, therefore undermining the process.

    1. Forget about reduced use of roads, and instead focus on increase mobility.

      The best way to manage roading demand is probably to only expand capacity in strategic areas of the network, and let congestion discourage others from making journeys in single occupant vehicles.

      But as in so many other ways, NZ tends to operate in its own opaque bubble of reality.

    2. I quote “roads can never be made wide enough fast enough for a car only transport system as LA found in the 70s and the rest of the world has found out since.”

      What is it about that statement you do not understand or disagree with?
      I think its pretty obvious to everyone (except NZTA, the MoT and the MOT it seems) that simply building more roads is not the answer and never has been.

      Maximising the way we use the roads we already have now (without spending huge $$s on making them bigger/wider/faster – and for what exactly?) offers the best outcome in a Auckland context.
      That means giving the right vehicles priority on existing roads and using dedicated corridors (Rail or Busway) for PT when possible, and bus lanes and other priority measures when you can’t.

      It is possible that once you have put your bus lanes in on the road that there is “spare” road space for more cars, and yes that will fill up with traffic in short order.

      But does that mean you need to widen the road? No it does not.

      As I’ve said before, the best car you can have, is the one you never need to own – because the PT system enables you to do everything you might need a car for around Auckland and you can rent/borrow or whatever one when you do.

      Which is what A Hooper said above – forget modes and roads, and think mobility instead and the best answer will become obvious.

    3. When you “shift people from a motorway to a train or bus”, you haven’t solved congestion for anyone still driving. But if the train or bus have a dedicated right of way, you’ve now “solved congestion” for all the people who are now no longer in congestion, because they are riding said train or bus.

      An interesting read arguing that we should just give up on trying to get rid of car traffic congestion:

    4. Geoff, congestion is not the only metric when talking about PT. Quality PT offers options. Just yesterday I drove to Newmarket to get my car serviced. I then hopped on my bike to Nuffield St, had a bite to eat and then onto a train to Henderson. Hopped on my bike again to visit someone at the other end of Lincoln Rd. Returned to Hendo on my bike, bought something from PB Tech, back on the train to Newmarket (must have been at least 1/2 full. Not bad for middle of the day) lunched back in Nuffield St, rode down to Osborne Lane and then back to pick my car up. Multi modal and had zero to do with congestion.

    1. Nice try but that wasn’t a long term forecast. That was a graph showing what rail patronage would have to do look like to meet the governments patronage target by 2020

          1. Indeed it is. As you guys have pointed out previously; the 2011 RWC spike, the various summer shutdowns for upgrades, and introducing a more honesty-based ticketing system have all no doubt had an impact on patronage figures over the last couple of years that may help partially explain the recent dip.

  13. The british traffic forecasts are top down regressions where the govt inputs forecast GDP and gets traffic growth as an output. I worked there on highways and then rail planning in the early 90’s. The weakness is that the government always forecasts growth because growth is a policy. That meant when we used lower forecasts based on our own count history the Dept of Transport made us redo using their high forecasts even though they didn’t believe it either. I see the NZTA report (p63 and 64) predicts PT use to decline as a result of incomes growing. They have assumed it is an inferior good. They attribute the recent increase in PT to the reduction in real income. Anyone who has ever had to forecast ends up humbled by time. As for the importance of the national km travelled model. I have never seen it used for B/C analysis. Growth is still 4 counts in five years or 8 in 10 years with the trend being whatever it actually is. If you are right and traffic is going to go down permanently then it will show in B/C analysis within five years of the change point.

  14. I don’t think it’s set in stone that any projection is “right” or “wrong”. If we trash PT, double down on motorways, and get rid of petrol taxes, driving would explode, far beyond those predictions. Whereas if we stop spending on new roads and put all our money into PT, PT would explode. Build a nationwide cycling network, and cycling would take off.

    People don’t have an inherent personal demand for one mode or the other – they use whatever’s convenient to them personally. So planners don’t need to predict the future – they need to choose the future.

    1. Yes! This and what Nick R said above at February 19, 2014 at 6:22 pm

      “I don’t think there is much chance to reduce congestion by providing alternatives, more like you just move more people on the alternatives and the traffic stays about the same. If you go really crazy then you might make it so much faster and easier on the alternatives that driving just can no longer compete, but that would be pretty intense levels of alternatives.

      My conclusion is you just accept traffic congestion happens in busy cities at peak times. The question is how much you want and where, and what alternatives you should provide.”

      should be tattooed on Brownlee and Joyce’s foreheads to remind them of it every morning in the mirror.

      So two simple things:

      1. Congestion is not a problem, it is a sign of a healthy city. Congestion in Detroit is very low.
      2. We get the transport system we build.

      Not rocket science, more like woodwork.

      1. You are right.
        And for an interesting example of this from the past,
        See this link ( ) tweeted yesterday by the Auckland Mayoral office relating to a “Pictorial Parade” TV item about Auckland motorways which screened in Feb 21 1964 (thats 50 years ago today).

        Its shown under the title of “Back in the Day” which is some TVNZ attempt to re-imagine the past as some fictional place that never really existed.

        In every scene of this short piece you see congested roads in the CBD, (cars, trucks, buses) and its always the case – even now after decades of “taking the pills and ointments” continuously prescribed by the Traffic Planners masquerading as the “Traffic Doctors” – who all know best. So you have to think that yes some congestion is always present.

        This item (as well as the others in the Pictorial Parade series on Auckland show) are always an interesting lot as they give a lot of background info not usually well known now on why things are the way they are.

        A voice over comment is given on how the Auckland Harbour Bridge traffic in 1964 had already exceeded the traffic projections not supposed to be reach until 1977 – some 13 years into the future.
        And another part shows the results of a traffic surveys undertaken how only 10% of the traffic in Auckland then is through traffic (i.e. over the bridge), 25% is to the CDB and the rest is cross town traffic going elsewhere (e.g. out west).

        There is also an explanation of the motorway moat (they called it the CBD ring-road bypass – ring roads were a very fashionable thing to do at the time), which puts the CBD firmly in its place as a castle surrounded by a moat of motorways/arterials – and we now suffer the results of this.

        You can also see mentions of the Newmarket viaduct (then in the process of process of foundations for it being readied/built) and also related demolition work to enable motorway approaches either side of it.

        Keen watchers will also spot the (brand new) Market Road off ramp and the long panning shot shows the recently built flyover for the motorway over Victoria park in all its brand new concrete colours.

        And in all cases, they show brand new 4 lane motorways with not many cars on them as the way of the future and sprawly suburbs right next to them so easily get to using these roads as the answer to every problem.

        And of course even more relevant/poignant is no one there seems to realise that the dream they’re being sold on is simply going to turn into a bigger version of their current congested nightmare in short order.

Leave a Reply