Safety is at the very heart of Network Rail's - and the Government's - justification for these masts. In the House of Commons on 21 May 2004, Keith Hill, Minister of State in the Office of the Deputy Prime Minister, described the masts as "Network Rail's new safety system", and went on to say that their introduction "is primarily safety driven. It will implement a key conclusion of the Cullen investigation into and report on the Ladbroke Grove rail accident...."

The Cullen Report's support for a nationwide communications system is the first plank in Network Rail's case for the masts. They call it "a key recommendation" - in fact, it's number 51 out of 89 recommendations, of which 41 were designated as 'key'.

However, if you read the Cullen Report itself, a rather different picture emerges from the one Network Rail paints.

First of all, all the trains involved in the Ladbroke Grove crash (which took place on 5 October 1999) already had in-cab radio systems.

In fact, the train that caused the crash was equipped with CSR, and Lord Cullen concluded that, if it had been used effectively, it would have prevented the accident! He blamed Railtrack's management for failing to train signalmen to use it properly in an emergency. His only criticism of it was that it was only installed in the south of England.

A similar combination of effective technology and incompetent usage occurred with the Southall crash - the train that caused that accident had, not one, but two safety systems installed: the basic Automatic Warning System (AWS), which sounds a buzzer in the cab when a warning signal is passed, and BR-ATP, Railtrack's version of the more advanced Automatic Train Protection (see below), which applies the brake if the driver fails to stop. They did not prevent the crash because neither was working at the time - the AWS because it was broken and the ATP because the driver was not trained to use it so it was switched off.

The Cullen Report also makes favourable comment about First Great Western's experiment with issuing conventional mobile phones to its drivers and conductors.

We have extracted the relevant passages. Alternatively, you can read the full report (PDF 3mbs).

It's important to remember that rail is already the safest form of transport by a wide margin. Read PDF article by Norman Bradbury (65kb) published in Railwatch, November 2002. It shows that (measured by fatalities per 100 million kilometres travelled) rail travel is over 27 times safer than road travel.

Here are four quotes from a report (808kb PDF) published by Railway Safety and the Strategic Rail Authority in April 2002, called ERTMS: Towards a Better, Safer Rail System, from the Strategic Rail Authority's site. (ERTMS stands for European Rail Traffic Management System, of which GSM-R is part, and ATP means Automatic Train Protection):

As TPWS [Train Protection and Warning System, which is already operational on all UK tracks] significantly mitigates ATP-preventable risks [preventing over 80% of them], the relatively small additional risk reduction achieved by ERTMS (once TPWS is installed) appears not to be justified purely as a safety investment.
ERTMS… will address the majority of ATP-preventable risk. This equates to a maximum estimated saving of 83 equivalent lives over the next 40 years. By comparison, UK road deaths are currently running at approximately 66 per week [our emphasis].
On this basis, the approximate capital cost per equivalent fatality avoided is between £75 million... and £45 million... over 40 years.
This means that there will be other potential safety investments, on and off the railway, which deliver much greater safety benefits [our emphasis].

However, the Health and Safety Executive, in Train Protection - Review of economic aspects of the work of the ERTMS Programme Team, (1.5mb PDF) published in 2003 , gives an even lower estimate of the safety benefit of Network Rail's preferred version of ERTMS: “Our own estimate is that the risk, at today's traffic levels, with TPWS and TPWS+ but without ERTMS, is one fatal ATP-preventable accident in about ten years [our emphasis]. With an estimated four fatalities per accident, this would work out at £267.5 million per fatality over a 40-year period. (TPWS+ is an up-rated version of TPWS, which works at higher speeds.)

The same report criticises the ERTMS Programme Team for ignoring one "politically important" risk: "namely, that of an ERTMS-induced multi-fatality accident in the early days of ERTMS".

If the money is to be spent directly on rail safety, it would be far better spent on track maintenance and staff training: of the four recent major crashes, two (Potter's Bar and Hatfield) were caused by poor track maintenance, and the other two (Southall and Ladbroke Grove) were primarily caused by poor training resulting in the failure to make use of the technology already available, which if it had been used would have averted both accidents.

If saving the most lives for the money spent is the objective, the £4.28 billion Network Rail is planning to spend on ERTMS and the GSM-R network (let alone the amount it would probably end up spending) would be far better spent on improving the reliability, punctuality, comfort, and convenience of services, and subsidising fares more heavily, in order to attract road-users back to rail transport (thus cutting their risk of death in transit by more than 95%). No doubt regular rail users would also appreciate more reliability, punctuality, etc, plus lower fares!

Unfortunately, Network Rail is squandering so much money on ERTMS and other grandiose schemes (all justified in the name of safety, though rail is by far the safest form of travel already) that money earmarked for road improvements (many of them genuinely safety-related) is no longer available. An article on the subject, called Roads facing axe to meet rail bill, was published in the Daily Telegraph (1 April 2004).

Yet road improvements give a far higher pay-back in terms of safety - in the report ERTMS: Towards a Better, Safer Rail System, Railway Safety and the Strategic Rail Authority state, in a footnote, that the DTLR's (Department of Transport for Local Government and Regions) 'appraisal value' for cost per road fatality avoided is £1.14 million, "although the actual spend per life saved on local authority highway safety schemes is understood to be very much lower".

In fact, according to the Commission for Integrated Transport, the amount spent on road safety is less than a tenth of this 'appraisal value':

The Ladbroke Grove Rail Inquiry, Part 2 Report includes a "Joint statement of experts on risk management" which records that several experts observed differences in government planned and contemplated expenditure on road and rail safety: the actual expenditure to prevent a road fatality is around £0.1 million [our emphasis]; and the cost per fatality avoided by fitting TPWS is about £10 million.

CfIT's initial assessment report on the 10 Year Transport Plan, Part 7, 22 May 2002

According to a news release by the European Road Assessment Programme, called Cutting road deaths is cheap - but we can't do it, say councils, the cost to prevent one road fatality in many areas of the UK is just half that figure, at £50,000 or less - but no money is available.

Network Rail is clearly aware of, but not in the least abashed by, this disparity. In fact, in a breathtaking display of spin, it uses road-users' far higher fatality rates to justify the adoption of ERTMS Level 2 (which requires GSM-R) rather than Level 1 (which the Uff/Cullen Report recommended): it argues that Level 2's (hypothetical) increase in capacity would allow it to attract more road-users back to the railway, thus saving between 300 and 700 road fatalities over 40 years. The Health and Safety Executive Report cited above disputes these figures: “Our own estimates suggest that the EPTFR (ERTMS Programme Team Final Report) overestimated the increase in road fatalities, for their assumed lost of rail capacity, by a factor of four or five.”

In any case, the 36,000 road-users' lives that could be saved by spending the money directly on road safety don't figure in the calculations at all.