This is a copy of the full text of Professor Phil Goodwin's comment piece in Local Transport Today, published on 13th April 2012. You can read it on LTT's Transport Xtra website here, with a subcription: Due Diligence, traffic forecasts and pensions.
Keith Buchan, Chair of the Transport Planning Society, last month wrote that the Department for Transport’s traffic forecasts were “now so far from reality that there must be an urgent review” of the National Transport Model (DfT’s London’s traffic forecasts are implausible, Letters, Local Transport Today, 30 March 2012). As with all such reviews, the outcome depends on the thoroughness, independence, excellence and credibility of the technical work.
The suggestion is given added salience by David Cameron’s speech at the Institution of Civil Engineers where he suggested the need for private investment in the road network, because “we do not have enough capacity in places of key demand”, and the announcement by the Treasury of a Pension Infrastructure Platform run by UK pension funds, designed to encourage them to invest in the country’s infrastructure.
The pension industry has rather welcomed the idea, albeit with caveats. One interesting comment, from Mathias Burghardt, of AXA Private Equity, says infrastructure investments of this form are “less exposed to sovereign risk as revenue streams are derived from millions of end users paying for essential services, rather than a limited numbers of public entities. This means it is a natural fit with the long-term liabilities of many pension plans.
However, pension funds – indeed, any private investment whether on behalf of pensioners, shareholders, or trust beneficiaries – have legal obligations to spend their stakeholders’ money on sound financial principles. If they do not do so, their decisions can be challenged not just in political hurly-burley, but in the courts, with real money at stake. As with a share flotation, the key issue will be the Prospectus, the formal analysis of an offer which has sound expectations of financial return, and careful consideration of risks, especially downside risks – what are the chances of a worse return than expected, and who bears the risk if so? The prospectus methodologies may (or may not) be the same as those used for public sector forecasts such as the National Road Traffic Forecasts, but the application, responsibility for error, focus, authority and implied power are different. In other words, due diligence means that Keith’s suggestion of a review is not just a good idea, but is certain to happen, in one form or another, and with legal consequences.
This seems to me to be one of those cases where declared policies, almost from the day of their launch, must almost inevitably evolve into something different, whose form can be seen, albeit darkly, from the beginning.
The first thing such a review will need to do is consider the track record of the currently authorised forecasting procedures. Conveniently, Kit Mitchell, for many years head of social research at the TRRL, and now contributing to the increasingly impressive website providing excel worksheets (http://www.iammotoringfacts.co.uk/ ) from the Institute of Advanced Motoring, has been compiling the information for this.
"anybody, just anybody, looking at this graph is going to think that there is a downside risk of the long term traffic flows being substantially less than the forecasts, as they have continually been for at least the last quarter of a century."
In a flurry of activity over the Easter Weekend he, Gordon Stokes and I managed to take this back to 1989 (and even 1973 – see below), but we have not yet tracked down the original documents of the 1980 and 1984 forecasts. (Perhaps LTT readers can help?). The figure you see above is the result so far, for car traffic, showing successive downwards revision of the forecasts as for 25 years car traffic stubbornly refused to behave according to expectations. The revisions were of the form ‘growth later’, not ‘less growth’.
So let us do a role-playing game. Suppose you are the investment manager for XYZ Pension Fund, considering whether to invest in the M999 bridge and motorway widening programme, vitally necessary, you are told, because it is already operating to capacity and the traffic will increase by 50% over the next 25 years. Sounds good, you think, and wonder whether to opt for a real charging scheme, taking in the fivers from a million motorists, or a shadow scheme, paid by the Government in relation to future traffic. So you look at the forecasts, and the forecasting record. Now correct me if you don’t agree with my next step, but I don’t think it is just my subjective judgement: anybody, just anybody, looking at this graph is going to think that there is a downside risk of the long term traffic flows being substantially less than the forecasts, as they have continually been for at least the last quarter of a century. In that case, an income depending on real charged prices is going to be less profitable than an income stream guaranteed by the Government based on the Government’s own forecasts. So you will ask for a guarantee. But that’s hardly attractive to the Treasury. The downside risk for Government is paying a lot of money, not underpinned by buoyant tax revenue, in respect of traffic flows which under-perform, for a project which for that reason turns out to be a lot less necessary anyway.
But consider the opposite outcome: suppose that the traffic forecasts do turn out to be accurate, or even underestimates. Then the cash flows are more robust but there is a danger of significant reputational damage, since congestion will actually be getting worse, not better (as demonstrated by the DfT’s own calculations), and the investing agency will be taking a substantial, highly visible, controversial income, in respect of a worsening quality of service.
These twin fault-lines - financial risk if the forecasts are overestimates, and reputational damage if they are correct or underestimates – suggest, it seems to me, that the ideas will evolve over the next year or so (or maybe more swiftly) into something else. The ‘something else’ could be in one of two pro-active forms. First, it could be a real road pricing scheme with a much greater public, rather than private, focus, for the traditional reasons of tax revenue and travel demand management, rather than road expansion. The second possibility would be to evolve into more and more extravagant guarantees, ending in a PFI-like scheme which risks paying substantially too much to the private providers. Both options are currently very unattractive politically. So a third, passive scenario could then emerge from the gloom – well, since traffic is rather stable, maybe it is better just to let the issue lie for a while.
I mentioned the 1973 forecasts, made at that time by TRRL (Tulpule, Report LR543, on a method initiated by John Tanner). They suggested a slower rate of growth with a distinct levelling off in the early 2000s, and are quite extraordinarily accurate now, 40 years later. An academic alternative model by Romilly et al (Journal of Transport Economics and Policy 1998) also suggested that a small negative general trend for car ownership was already starting to kick in, and would lead to a stabilisation of ownership by the period 2000 to 2010. That doesn’t prove their models were inherently better of course, but how intriguing that they were right, but dismissed.
I’d certainly want my pension fund to look hard before betting my money on traffic growth.