Return to Archive -by date - by topic - Archive 2002.
When you ask the general public for their views on safety the common sense displayed is encouraging
In these times of emotional correctness you would think that a body called the Parliamentary Advisory Council for Transport Safety (PACTS) would be fluffy in the extreme. And you would be wrong.
Under its Executive Director Robert Gifford, PACTS is a no nonsense organisation, securely located in the real world. Its most recent conference, held in London back in October, had the robust title ‘The price of a life'.
Regular readers will recall that those two old buffers Professor Uff and Lord Cullen found the concept of cost benefit analysis of safety expenditure based on the Value per Prevented Fatality (VPF) emotionally incorrect and fussed that it was not at all the same as putting a price on a life. So the title confirms that PACTs makes no concessions in its remit to cover transport safety as a whole.
During the course of the conference a number of ‘big lies' about railway safety were exposed as media hype. First up was Professor Michael Jones-Lee who is Professor of Economics and co-Director of the Centre for Analysis of Safety Policy and Attitudes to Risk (CASPAR) at the University of Newcastle upon Tyne .
Back in 1985, the then Department of Transport decided that if it was to allocate scarce resources for transport safety to best effect, it needed to have explicit costs for the risks of death and injury involved. The first approach was based on Gross Output: this combined future earnings lost by someone's death, the costs to society of medical treatment and ambulances plus an arbitrary allowance for the pain grief and suffering of the bereaved dependents.
At 1985 prices the gross output figure was £185,000 – say £450,000 today. But in parallel the alternative VPF approach evolved. VPF is based on the cost avoided so the Gross Output figure was never used.
When you are developing VPF one approach is to ask people how much they are willing to pay to reduce the risk of death and injury. In other words we are asked to balance our aversion to death and injury against our willingness to pay to reduce the risk of the same.
But asking someone how much they would be willing to pay to prevent their death in an accident in the next 12 months is meaningless. The answer would be ‘As much as it takes'
But if you ask 100,000 people how much they would individually be willing to pay to reduce the risk of death within the group by 1 in 100,000 – the Prevention of a Statistical Fatality (PASF) – you should get a much more disinterested answer.
If you then aggregate the answers, you get the value for PASF. This represents the aggregate willingness to pay for a very small decrease in risk and you have a number for your VPF.
If, for example, each person considers it worth £10 to reduce the risk, then the VPF is £1million. It is no longer, in Prof Lee-Jones words ‘compensation for the certainty of inevitable death'.
Now remember that the VPF represents the value put on the public's aversion to death and injury. The VPF in Railtrack's safety case is the same as that applied by the Department for Transport, Local Government & the Regions to the safety benefits of road schemes. Currently the figure is £1.24m. Note my customary caveat: the actual value for road schemes that have gone ahead in recent times is only £100,000.
So Rail is already being penalised in the real world by a factor of 10. But, because of the public's claimed aversion to multiple fatalities in rail accidents, where catastrophic risk is involved Railtrack's VPF is multiplied by a factor of 2.8.
But where did the Department get its VPF from?
When it abandoned Gross output in 1988, it commissioned a review of literature on VPF and picked a figure at the bottom of the range worth £500,000 at 1987 prices. Civil servants became uncomfortable with this arbitrary figure and in 1996, the Department and the Health & Safety Executive began a programme to re-calibrate the road VPF plus relative values for rail, domestic fires and fires in public buildings.
It was not easy because the public couldn't get its collective head round relative risks. For example, 40% of those consulted in the pilot scheme would spend the same amount to reduce a risk of 3 in 100,000 as they would for 1 in 100,000. Wealth/Risk trade-off was not an easily grasped concept.
Commendably, the sponsors didn't abort the exercise and in 1997 a new study began based on individual interviews with small focus groups, each of 150 people, at Newcastle , York , Bangor and Brighton .
These interviews gave a median VPF of £500,000 and a mean of £1-1.5million. The Study Report recommended a VPF of between £750,000 and £1.25million and in 1999, the Department increased its VPF by 10% to £1.04million in 1998 prices.
There are also similar costs for injuries, around £130,000 for serious and £10,000 for slight.
So that was road. What about rail? Well in terms of actual risk, the Wealth/Risk ratio should be much less and would probably disappear off the focus groups' collective radar if you worked in single lives.
So the question changed to if you had enough money to save 10 railway deaths or 15 road deaths which would you prioritise? If the answer was ‘rail' or ‘road', the proportions were changed until no preference emerged.
And the results showed that the VPF for rail was 0.834 that of road. For the record, Domestic and public fire risk VPFs relative to road were 0.926 and 0.923.
Now these focus groups included only small samples of regular rail users, as one would expect. Indeed, some of the Newcastle group never used rail.
People were also influenced by the sheer scale of the road death toll and assumed that road safety was more important and justified more expenditure. On top of which, there hadn't been a big railway accident since Clapham
But then came Southall and Ladbroke Grove and the media perception, fostered by the HSE, was that the ‘public' was ‘concerned' about rail safety. So a follow up was requested which took place in January and February 2000.
A cynic might argue that the Department and the HSE, in particular, were trying to reinforce their own touchy-feely prejudices. Because, first of all, three London commuting areas were added – St Albans , Reading and Guildford . Also, the make up of the focus groups was stipulated in terms of gender, age and class and at least 40% had to use rail at least three times a week.
So here we had the lab-rats pre-conditioned to produce the right result and guess what? In a blow for commonsense, a mighty rebuff for the safety hysterics in the HSE, the public wouldn't play.
After all the media coverage of the big rail accidents, the focus groups now thought that only slightly more should be spent to prevent a rail fatality. How slightly? Well 1.003 times. More significant, perhaps, public fire increased to 0.96 while domestic fire fell to 0.89. Perhaps the fire at Ladbroke Grove had influenced thinking.
But that 1.003 figure covered a wide range of rail users. If you went deeper into the raw data, the subjects with high rail mileages reckoned it was worth spending 1.157 times as much as road, but the low end users still thought road was more deserving with a figure of 0.993.
Overall, Professor Jones-Lee concluded that you can't ‘drive up the rail VPF above road. In other words, a life is a life. Where does this leave the 2.8 times figure for multiple fatalities in the Railtrack Safety Case? Certainly not justified on the basis of public attitudes.
But what about the multiple reflecting ‘societal risk'? ‘Societal risk equals media hype plus political over-reaction, it doesn't cover reality', Professor Jones-Lee told the PACTS conference. In his view, the three times multiple is applied ‘for job security reasons'.
There was support for large multiples came from the Bishop of Hereford who claimed that because you could theoretically make railways absolutely safe, you should, whereas you can't achieve absolute safety on roads. That sounds like a variation on the doctrine of free will.
From the floor Richard Hope pointed out that the VPF for European Train Control System, costed at around £3.5billion for network wide installation, would be around £100m after TPWS was installed. Professor Jones-Lee was temporarily rendered speechless, before commenting, ‘that disturbs me an awful lot'. In fact, installing ETCS on the East Coast Main Line by 2006, instead of when the existing signalling is due for renewal, has a VPF of £187m.
This is not to say that ETCS is an extravagance. It is the Eurostandard signalling system of the future and will replace current signalling as it becomes life expired. But premature installation simply to get ATP is a commercial nonsense.
While the Prof wouldn't ‘go to war' over three times the road figure, he was perturbed by the quoted VPFs of £10million for some rail schemes. As the 2002-03 Railway Group Safety Plan (See this issue) shows, you get a knock on effect.
Because the Railway Safety Regulations, mandating installation OF TPWS and Mk 1 stock replacement have an implied cost of averting a fatality in the order of £10m, this has set a precedent for using this VPF for schemes addressing similar risks, such as enhanced emergency braking (EEB).
Despite the DTLR's research quoted above, VPFs of £10million, a hundred times the real life cut off point for local authority road safety schemes, are now being applied to the fitment of on-train data recorders and enhanced emergency braking. This approach has been endorsed by the independent Safety Advisory Board.
Such VPF creep may be intellectually defensible, and EEB reduces other collision risks as well as SPADs but can the railway afford such navel gazing when out in the real world 90% of buyers of one make of popular car won't spend £500 on the anti-skid braking option?