Some interesting news coming out of Australia today, with yet another transport public-private-partnership (PPP) on the brink of collapse, due to over-optimistic traffic forecasts. This time it’s the Airport Link toll road in Brisbane:

THE operators of Brisbane’s Airport Link have gone into a trading halt, amid increasing speculation about the company’s financial future following much lower than expected traffic volumes…

…It comes as the operator struggles to achieve even 50 per cent of its forecast traffic volumes of 135,000 vehicles a day.

Since a discounted toll took effect in late October, Airport Link carried 53,172 vehicles a day down from 85,000 in its first six weeks of operation when motorists were able to use the tunnels for free.

BrisConnections had projected a figure of 135,000 from the end of the toll free period, rising to 160,000 within 18 months of opening.

A retired Sydney academic notes that this failure is far from unusual and all comes back to that same vexed issue that we discuss so frequently in blog posts: overly optimistic traffic predictions:

Professor John Goldberg has written a complex 24-page analysis of the project and his findings are unswervingly grim. He says that Brisconnections, the listed company which oversaw the $4.8 billion project, faces “inevitable financial collapse”.

Skewed traffic forecasts, poor cash flow and unmanageable debt will prove its undoing, he believes.

Until recently, the listed company had insisted that everything was fine.

That reassurance came even though vehicle numbers have fallen far short of expectations since the 6.7km road opened in late July with an introductory free tolling period which ended last month.

But Brisconnections acknowledged on November 2 that its bankers had appointed insolvency and restructuring firm PPB Advisory to conduct an independent business review. Investors were warned that there may be “adverse implications”.

Is Brisconnections heading for the same fate as RiverCity Motorway Group, which collapsed about $1.4 billion in debt less than a year after the 2010 opening of Brisbane’s cross-river Clem7 tunnel?

That’s what Prof Goldberg thinks. Normally, such a dire forecast might be met with skepticism.

But Prof Goldberg, who taught at Sydney University and worked as a senior researcher at CSIRO for 30 years, has form.

He correctly predicted the failure of the companies operating the Cross City and Lane Cove tunnels in Sydney. That foresight earned him a place before a 2005 NSW Parliamentary inquiry.

“The public-private partnership concept has failed in Australia and should serve as a warning to superannuation funds of the high risk of investment in road infrastructure,” he writes in his current paper.

Investors have poured more than $23 billion into 11 toll roads across Australia since 1994 and the net return on equity has been small or negative in each case.

On the one hand I’m not fan of PPPs for transport because I think they’re just a form of “creative accounting” that benefits nobody but the lawyers drawing up the complex contracts. However, on the other hand if these high profile PPP failures in Australia had just been public sector roads we may well have never known about them, because the horribly inaccurate traffic predictions wouldn’t have been an issue. So it seems that – perhaps by learning the hard way – the PPP system might end up bringing a bit more rigour to the process of assessing whether projects are actually needed or not.

But that will be too late for the Airport Link, where it seems the process of working out how many vehicles were going to use the road each day was based on a rather convoluted process:

A common flaw in the failed tolls roads and, notably, Airport Link, is the use of a “work back” philosophy to forecasting traffic numbers, Prof Goldberg says.

The promised return on equity to investors is a starting point used to work back to how much revenue must be generated from the expected daily flow of vehicles, which has been inflated to wildly unrealistic targets, he says.

Brisconnections had forecast 135,000 vehicles a day would use Airport Link from the start and the numbers would eventually climb to 195,000 daily.

The most recent average traffic count showed a dip to just 66,203 a day in October, a period when the road’s use was still free for more than half the month.

The Clem7 traffic performance has been equally dismal, with an average of just 24,000 vehicles a day, less than a quarter of expectations.

The ‘work back’ philosophy seems utterly bizarre, to say that least.

Given the numerous failures of PPPs in Australia, for reasons which seem to be happening in New Zealand too, it is interesting that our government seems so keen on pursuing them for projects such as Transmission Gully and perhaps the future Puhoi-Wellsford road. The worry I have is that if private investors in PPPs are so burned from the Australian experiences then all the “demand side risk” (i.e. whether the road will have enough use and generate enough toll revenue to make it worthwhile) is likely to end up sitting with the public while the private investors make out like bandits through ‘creative accounting’. Pretty much the worst of both worlds.

Why do we never learn from the mistakes of others?

Share this

47 comments

  1. Don’t talk to me about “work back”. It starts at the simplest level when logistics workers check in freight: “How many have you counted?” “How many do you need?” Outside the CEO’s window.
    It progresses through projects that are designed to generate demand for a service that’s never been offered on the market: “How many units will we process in Y1?” “How many do we need?”
    And then when you’re Key Account Director for a major aviation IT service provider, your targets are based on….what your direct report needs to achieve his performance bonus.
    I went back to the strategy people and told them that if they were basing profit forecasts on what I was supposed to deliver ( in a post 9/11 market with clients who were bloated with our products and zero potential for investment in the near future), then Armageddon was looming.
    I’m happy to say that they were prepared to listen..

  2. And yet we are pushing towards PPPs, in a recent OIA request I got the agenda for the September and October NZTA board meetings, here is an a couple of items from September

    And this for October

    You can bet I will be doing an OIA on those documents.

    1. Funny that on a first skim reading of point 3.2 above I thought it read “finding a business case for Transmission Gully”

      Maybe it’s down the back of someone’s sofa.*

      * I don’t believe that for a second. It will have to be pulled out of a magician’s arse.

  3. A ‘work back’ philosophy is worse than just bizarre – it’s the absolute opposite of rational and verifiable decision-making. Major investments should be subjected to intense scrutiny over how they come up with their forecasts – predicting the future is difficult but you have to be able to justify how you come up with your figures. It’s actually quite a major allegation Professor Goldberg is making, that using ‘work back’ is a common approach. Certainly no consultancy would be seen dead using anything of the sort, as once you put your name on a report you’ve got to stand behind it. The forecasting process may never be as rigorous as we would like, but surely there’s got to be more to it than this.

  4. Is this a failure of PPP or of the road in the first place? PPP is just a funding mechanism. It would have failed whether it was a PPP or a fully funded government project. The benefit of the PPP is the Government doesn’t loose its money, the investor does.

    So the issue is the road. Build a new road (and I assume keep the existing free option available) and it is never going to hit targets. Take a few cars of the existing route and suddenly the existing route seems OK. The new route was only 6.8km so I wouldn’t expect much time saving once you remove a bit of traffic from the existing route.

    Question 1: If the Central Rail Loop was to be a PPP which enabled it to be built 10-20 years early than a fully funded (local/central) government project, would you say no because it is a PPP?

    Question 2: why isn’t the CRL being proposed as a PPP? Is that because even with over inflated targets that normally come with PPP (by advisors who take their fees upfront), it still doesn’t stack up?

    1. Don’t think you’re right about the first point. What happens when these PPP’s fail is the govt has to take over, and is then often in the can for a large amount of money. PPP’s usually totally fail in taking risk out of governments hand. Also there are many extra costs involved in a PPP, such as contract documents being measured in the metres(!!!), and the need for finance and success fees for consultants makes them more expensive. If was a government road traffic projections would be more transparent, and not drive by profit.
      As for Q1 and 2, a PPP has been looked at for the CBDRL. However would be totally different type of PPP. Roads can be built with expectation of fare revenue going towards the builder/operator. However this is not option for CBDRL, with Aucklands proposed fare zones. A PPP would be more a funding mechanism for the council, a type of loan, and the council would pay off the builder/operator over a 20 or 30 year lifespan rather than borrowing or fronting up with $2 billion.

      1. Good points.

        The government may need to take them over but at least they are only responsible for maintaining, not the original build cost. They also have the option to close down if viable (take the Eurotunnel for example. If no one wanted to fund, they could just concrete up each end.

        I worked on one PPP in the UK and agree that the consultants time can add up. Add in advisers and success fees etc and it must add an extra 10% onto the cost if done fully funded by government.

      2. I’m not so pessimistic about PPPs. I know quite a few have failed in Australia, but can you find one that has led to the government to take over, and bailing them out. To me PPPs are good as they transfer most of the risk from the govt to private investers.

    2. In order to get the funding for many of these PPPs there is usually requirements that the loan is guaranteed by the government so when the fall over the taxpayers who didn’t want to fund the thing in the first place get left paying for it.

      As for the CRL (and its not a loop), I would more likely have to be a rental agreement where the private company paid and built it and tax/ratepayers end up paying an access fee. It would be saving on construction costs but pay more for it over the long term. It would likely work out much cheaper paying for it out of government borrowing.

      1. “In order to get the funding for many of these PPPs there is usually requirements that the loan is guaranteed by the government”

        That’s just stupid.

        If investors won’t invest without a government guarantee that’s a pretty bad sign about the viability of the project. The solution is to not go ahead with that particular project, not to guarantee it.

      2. Nonsense Matt. Point to one of the “failed” Australian toll road PPPs where the loans were guaranteed by government. You can’t, because they weren’t. Taxpayers have benefitted from the infrastructure, private investors have worn the losses.

    3. So the issue is the road.

      Exactly! Or to be more specific, new roads. Because we, and our neighbours in Oz, already have all the roads we need. What is being built now are the roads we want. Now ask me if we are willing to pay for them and I’ll point you back to his article. If this was a govt project nobody would blink an eye. Instead the politicians would proudly pontificate abut their road building, congestion slaying prowess.

      As for PPPs in Auckland I got the feeling that the even if there had been some interest the current govt wouldn’t allow it as it would give too much autonomy to the city. So no PPPs, no local taxes, no nothing. Also, the CRL is a project that goes against the ideology of the people in charge so …

      The more I think about it the more I believe PPPs are a bad match for infrastructure buildouts. How many investors are willing to wait 15-20 years for a positive ROI with hundreds of millions in capex?

  5. Ive driven the new Brisbane Airport tunnel a few times and it is a reall nice road to use. It saves time for sure maybe half an hour at peak times and has a real purpose. i.e. the tunnel you use when driving to the airport. It is liked by drivers and is a massive imporvement on the crazy traffic congestion that used to exisit. The failure here is in the traffic predictions not the road.

    1. If the road is truly needed and the projections are at fault then there is a simple solution – traffic is 50% of what the investor needs so double the toll price. That way the road will be paid for by the actual people that need to use the road.

      Problem solved?

      1. If they increase the price less people will use it. It depends on how elastic the demand is when determining whether it is in their interest to increase or decrease prices. Your suggestion is assuming perfectly inelastic demand.

        1. That is why I used the word ‘need’. If people ‘need’ something then the demand is inelastic. If it’s not and people aren’t willing to pony up twice as much cash, then we are talking about ‘want’, not ‘need’.

          I was trying to point out that often people who defend large projects defend them because they ‘want’ them and don’t really comprehend how very, very expensive they are.

        2. But no-one needs to get between two places especially fast. It will always be a want and hence I conclude that toll roads with many other options like this will, for the vast majority of customers, have relatively elastic demand. In that case, higher prices mean less, not more revenue for the operator.

    2. Welll that’s a little silly isn’t it? The road was built on the back of the traffic projections. I’m sure everyone would love a new multi billion motorway tunnel that serves some purpose or other, and it would be a really nice road to use, liked by road users and avoid traffic congestion… but that can be said for any other billion dollar boondoggle that ends up costing the public purse a kings ransom for each person that actually uses the thing.

  6. This goes to show that roads can’t make a profit. Traffic modellers should be able to use the decline in traffic when a toll is applied to work out how much people truly value travel time savings.

  7. Brisbane Airport seems to be a bit of a graveyard for PPPs,
    The Airtrain rail line fairly much went bust in it first 3 years, and it is only profitable now after the debt of couple a hundred million was written off….

    Brisconnections was already an investor nightmare, with many of the original retail shareholders being wiped out when they were unable to stump up with the final instalment for the partly paid shares. The company is now 80% owned by Macquarie and Deutsche Bank, ( and the Qland gover investment fund has ~8%)

    Looking at some of the back story on this, the Qland government will not be on the hook if this all does go down the crapper, so from their point of view, they got a transport link with out stumping up any of the money – They didn’t bail Brisconnections in 2009 when it first hit the fan, and that involved a huge number of retail investors, I doubt they will do it now with two fat cat banks taking a financial bath….

      1. They are great at their jobs – they managed to get investors to believe them so their banks could earn huge fees. The banks themselves will probably have no more at risk than the fees they earned, their customers are not so lucky.

        1. thats not much of a model then, relies on borderline Securites fraud and misleading of investors to get the funding. Totally hopeless. Meanwhile Macquarie bank gets away laughing as it has earned its $110 million success fee!
          Remember in NZ the NZCID are a front group for PPP’s, their previous CEO was also head of Macqaurie bank in NZ. Is scary how a few million thrown our way gets them such influential coverage in NZ media, and makes them sound very independent.

          In general PPP’s are stupid for an body that has sufficient capital and cashflow, which the NZ govt/NZTA does have, or have a very high debt loading and can make debt profile look better by hiding it in a PPP. Maybe for an Auckland Council project it could work however need to be billion dollar projects for all the financial engineering to stack up. The only way could get a good deal if was Chinese funded and built as they have access for cheaper capital. Might be option if council are desparate.

          Another issue with PPP’s is that the winning tender is the one with the best finance deal, not the cheapest or most efficient contractor.
          Fletchers also don’t want a bar of them and they would be the only NZ bidder looking at the CBDRL.

        2. ok, that was from last year I remember Fletcher saying they didnt like them, but I guess if thats on offer then they’ll take it.
          Either way the way these consortiums are made up mean that may not be getting the best builder available, and best prison operator available but best finance deal. Or the best builder and best operator could end up in different consortiums. Not really ideal outcome.

    1. I’m fairly sure that the Airtrain never actually went belly up. AIUI there was a debt to equity restructure that basically wiped out the original equity investors. Not hard to imagine why it failed in 2000 – the feed was positively awful. It has improved now and Airtrain is getting near its original forecast patronage.

      I’m sure a significant amount of public money has already gone into the AirportLink road. Only question is how much?

      1. some good caution there for Aucklands airport line. Airtrain only captures 8% of airport passengers. I think the case for Aucklands airport line should rely on similarly low airport patronage, but capture most of its patronage from Mangere, and benefits of speed improvements to the Onehunga line.
        Does seem difficult to get people out of cars for airport journeys. The pick up and drop off of family/friends at airport is quite a strong instinct.
        Still agree it is next highest priority after CBDRL but needs to be carefully investigated, and very well planned and implemented to ensure its a success.

        1. Luke, Auckland Airport will be a through station, not a terminus like Brisbane. The idea of a Southern rapid transit corridor needs to be promoted more – running Onehunga – Airport – Manukau. The PT equivalent of SH20.

        2. It’s a pretty crappy service, running only half hourly outside of peaks which has only recently seen the advent of services after 8pm (now finishes at 10pm).

          When it first started PT use in Brisbane was very low so any trip requiring a connection was a pain.

          Sydney gets a higher mode share and has a better service, but that has the same price to the CBD for a much shorter trip.

  8. Here is how governments in English-speaking countries form their national budgeting strategies:
    1. Allocate everything necessary to allow road transport to continue and grow according to forecasts. (This allocation never to be scrutinised or questioned. It simply ‘must be’)
    2. Take whatever is left after 1, and divide amongst all other national spending requirements, rigorously applying normal budgeting principles to ensure optimum and transparent allocation of resources (slash and burn if necessary).
    3. Promote 1 and 2 to the public as a fiscally responsible spending policy.

  9. The PPP to look at as a example for Transmission Gully is Mornington Peninsula Link in Melbourne. The previous State Labour government signed up for a lane availability charge over 30 years, which means this road will get paid for ahead of all other road maintenance or road-building. It is just like public sector borrowing at private sector interest rates, which are usually 2% higher than if the government borrows directly, but this way the debt is off its own books. There are no tolls planned for this road, because much of the traffic comes from a toll road (Eastlink) that lobbied for the new road to be toll-free to increase its own traffic volumes.

    http://www.linkingmelbourne.vic.gov.au/pages/peninsula-link.asp

    The new State Liberal government has cut road rebuilding budget to the extent that road foundations are only rebuilt every 50 years. And this is before we start paying for the Mornington Peninsula Link.

    I expect the Transmission Gully PPP would be set up as a toll road with a public sector lane availability charge that varies with traffic volume. Low traffic volumes would mean a higher public contribution.

    1. Yes that sounds similar to what I’m hearing. The private company will finance/build it and the NZTA will effectively lease the road at interest costs + some operational/profit figure. The debt gets kept off the books but we pay much more for it over the long term. If the expected traffic volumes don’t materialise then it is taxpayers who have to keep stumping up the cash, but the private company is guaranteed an income.

    2. Malcolm/Matt, No, in an availability PPP the project is on government’s balance sheet, not off. At least in Australia (not sure if NZ will necessarily be so conservative in the accounting treatment). And you only pay “much more for it over the long-term” if you ignore risk and the time value of money. Also, the private company is not “guarnateed” an income – it still has to perform in terms of availability of the road, maintenance performance etc.

      1. The problem is while that might be the normal way of doing things, I’m hearing that the NZTA is effectively planning to bend over backwards to get a PPP under their belt to please the government so are looking at some pretty shonky deals.

      2. you do pay more for it if the private company gets a high rate of interest than the NZ govt which is almost certainly true. NZ’s debt problem not that bad. As for guaranteed income, maintenance and availability are pretty cheap and simple for roads, hard to muck up for experienced contractors. And costs of this will pale in comparison to the effective interest rate the government are paying, probably north of $50 million for 30 years.

  10. If the current government signs up lots of low-return RONS projects for a subsequent Labour/Green government, perhaps there could be a special RONS road user levy on heavy vehicles to pay for the dodgy commitments that extend beyond the current government.

  11. Let’s be clear here, the failure of two toll roads in Queensland is not because of the PPP model, and it isn’t a failure for the Queensland taxpayer. It is the failure of over-exuberant investors and demand modelling that wasn’t fit for purpose.

    Taxpayers in Queensland are the winners as the roads have been built, they aren’t going anywhere and in effect these large expensive pieces of infrastructure have come at no expense to the taxpayer – they are not subsidised, like Transmission Gully will have to be. So from a public policy point of view it isn’t a problem.

    The companies in trouble for this are AECOM and ARUP, the demand forecasters, although the investors and banks have swallowed massive writedowns in their investments.

    There have been great success stories of PPP toll roads in Australia, such as Melbourne Citylink and all but two of the Sydney ones (Cross City Tunnel and Lane Cove Tunnel).

    However, that doesn’t mean it will work in NZ. Transmission Gully at best might be 15% toll funded, if the future RUC/FED revenues on the road were capitalised it might bring that up to 50% over the depreciated life of the asset – but it isn’t a viable private project.

    So it shouldn’t proceed.

    1. I guess the issue is that now private investors have been burned repeatedly in Australia there ain’t a hope in hell of seeing a PPP in New Zealand which puts some of the forecasting risk onto the private sector, as per the original intention of the Australian PPPs.

Leave a Reply

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