Mike Hanley discusses the cost benefit problems of increasing the safety of the railways.

Mike Hanley discusses the cost benefit problems of increasing the safety of the railways.

If you want to assure people in South Africa that a particular event is no big deal, tell them: "It is not a train smash". Rail crashes are generally considered so horrendous, that in that particular country they use them as a metaphor for catastrophe. Everywhere in the world, striking pictures of twisted, smoking metal and the sheer number of casualties help to focus the mind of governments and rail companies on the devil's trade off between money and life. In the UK, a debate is currently raging about the implementation of railway safety

equipment. Since the Ladbroke Grove rail crash in October last year, attention has focused on the failure to implement ATP, a system designed automatically to stop trains which run through red lights. Had it been in place, ATP would have prevented the accident at Ladbroke Grove. The decision not to install the system was taken by the Government for economic reasons - it concluded that £11 million was too much to pay to save each human life.

Professor Andrew Evans is the London Transport professor of Transport Safety at the University of London's Centre for Transport Studies, and was a key witness in the Ladbroke Grove accident inquiries. He explained some of the issues involved to Stated cRifc.

Willingness to pay
According to Professor Evans, economic decisions about investment in safety are informed by academic research which sensitively asks the general public how much they would be willing to pay to save a life - a method known as "willingness-to-pay" valuation.

Eliciting a sensible answer to such a question is, of course, easier said than done, and is a major research project. Questions are asked indirectly about individuals' preferences for particular types of risk as against particular types of safety. Preferences are probed through in-depth surveys carried out over a number of days.

Problems arise because individuals have extreme difficulty interpreting small probabilities - and it is in the small tails of the distribution where much incremental safety improvement takes place. Generally, willingness-to-pay studies demonstrate that different socio-economic groups have marked distinctions. Rich people are generally more generous than the less well off when it comes to purchasing a marginal unit of safety. This is reflected in a lower willingness to pay for safety in developing countries compared to richer economies.

The main rival to the willingness-to-pay method for putting a price on life is to value the lost output of the individual over the time of his or her expected life. This method can be used in cases where compensation is being sought for injury or loss of life. But it produces some inequalities - not least the tendency for millionaires to be valued much more highly than the rest of us. Willingness-to-pay studies get over this problem by assessing a common figure that is aggregated across a large group, giving one price for each potential life saved.

This price is then compared with the value of the risk reduction. This is determined by the expected number of lives saved by the investment in safety, divided by the cost of the safety measures. In the transport industry, the use of willingness-to-pay studies was first pioneered by the UK Department of Environment, Transport and the Regions (DETR) in assessing the value of marginal safety benefits when investing in new roads. When assessing the costs and the benefits of new roads, the DETR developed the willingness-to-pay methodology, and produced a figure for use in all investment calculations. This figure is then indexed each year to reflect inflation and other economic factors.

The latest published figure for the price of a life on the roads is £1.05 million. The figure for the year 2000, is expected to be around £1.15 million. Having said that, this figure is only of limited usefulness because local authorities have not got anything like that much money to spend on marginal road safety - even if studies say that that is what people would be prepared to spend. The amount that local authorities have to spend on marginal safety is only about £100,000 per life.

On the railways, the situation is somewhat different. Studies show that people may be willing to pay slightly more for rail safety than for road safety, perhaps because rail crashes are particularly horrendous, and because people expect railways to be almost completely safe, while individuals are prepared to take some responsibility for personal safety on the roads.

Equating cost and safety
Before Ladbroke Grove, the last serious rail crash in Britain occurred at Clapham Junction in 1988. The Clapham Junction crash was caused by a signal that had been installed wrongly - the signal had shown green when it should not have, and the driver was not at fault. An ATP system would not have prevented this accident. However, during the inquiry headed by Sir Anthony Hidden following the report, two further accidents occurred, killing five and two people respectively. Both these accidents were caused by driver error, and would have been prevented had ATP been installed. As a result of these crashes, the judge's remit was extended to consider the installation of ATP and he duly did so.

judge Hidden recommended that British Rail (BR), as it was at the time, implement ATP. He estimated the cost of implementation as £140 million. At the same time, however, he also recommended that BR should work out the costs and benefits of any safety system that was to be installed. So BR had ATP technology tested on the Chiltern and Great Western lines and determined that it would work throughout the system. Had the technology been implemented there and then, the Ladbroke Grove crash would never have happened. However, while the technology was being tested, BR also worked out what a system-wide roll out of ATP would cost. At £750 million, it turned out to be a great deal more than Judge Hidden had estimated.

BR's best guess at the number of lives that ATP technology would have saved was approximately 50 lives -

which implied an investment of about £14 million per life saved. This compared at the time with a willingness to pay for safety on the roads at about £700,000 per life. Even given a greater willingness to pay of about £2 million per life on the railways, the equation for ATP looked pretty bad. British Rail reluctantly concluded that the implementation of ATP was not economically viable. But the final decision was up to the Secretary of State of the time, Brian Mawhinney. Mawhinney asked the Health and Safety Executive (HSE), the final regulator of safety issues, to look at BR's estimates and conclusions. The HSE concluded that perhaps the estimate of the number of lives that would be saved was too conservative - up to 25% more lives would be saved if ATP was implemented. Even this only brought the estimate of price per life for ATP down to £11 million. Mawhinney recently defended his decision not to implement the system.

"At the time, the rail industry showed us tests which were done on the ATP. These highlighted technical difficulties. Our decision was not taken in isolation but based on the agreement of many sources, in the light of the test results. But this should not stop this government from proceeding, because I think it is time to re-examine this decision. They have no choice.

"There were also other considerations at the time, such as the good safety record, and the cost, which I recall was around £1 billion then. But after Southhall it was clear that the government would have to look at this again."

It is hard to argue that the decision not to implement ATP was taken improperly, says Professor Evans. Even so, the railways are at fault for the poor safety record because, while ATP was being considered from a technical and economic perspective, there was no alternative being considered or developed. "It is a case of the best being the enemy of the good," he says. ATP was the best system available, and the consideration of the best system precluded (for the railways) the development of something that would have been good enough to prevent the deaths at Ladbroke Grove.

What is the risk management lesson of this? As one senior executive at Railtrak told Strategic Risk, "We could have a perfectly safe rail system, but nobody would be prepared to pay for it." The cool, or perhaps morbid calculation that economists have to make about such investments revolves around the question of what price to put on each life saved. That is as it should be.

More importantly, as every risk manager knows, there are a myriad of small, inexpensive things that can develop a culture of safety that can be just as effective at saving lives as the latest technological safety innovation. In his report following the Clapham disaster, Judge Hidden recommended better training of drivers as well as the implementation of ATP. Perhaps if BR and the subsequently privatised rail companies had taken that recommendation as seriously as the economic equation against ATP, the series of accidents which seem to have plagued the railways might not have happened.
--Mike Hanley is an independent risk management journalist and consultant