This rail safety work is essential, but it should have been done long ago

Better fed up than dead: that is the only consolation available to the millions of rail passengers who find themselves delayed or stranded this morning. There is no doubt that the safety work now being carried out is essential. To that extent, complaints about the lack of warning or Railtrack's inability to carry out the work more gradually, causing less disruption to services, are misplaced. The only justifiable response to the four deaths in the Hatfield derailment is to do everything physically possible to make sure the tracks are safe.

Better fed up than dead: that is the only consolation available to the millions of rail passengers who find themselves delayed or stranded this morning. There is no doubt that the safety work now being carried out is essential. To that extent, complaints about the lack of warning or Railtrack's inability to carry out the work more gradually, causing less disruption to services, are misplaced. The only justifiable response to the four deaths in the Hatfield derailment is to do everything physically possible to make sure the tracks are safe.

This is not like the debate after the Paddington disaster, in which 31 people died, over the most effective and most cost-effective way of preventing trains crashing into each other. In that case, there are complex arguments over signalling technology and expensive automatic systems which detect hazards and apply the brakes. Although we will have to wait for the Cullen report into that accident, it seems likely that the risk of collision can never be completely eliminated and that it is better to focus money and energy on getting the basics right.

The issue of cracked and broken rails, however, is much simpler. Railtrack had been warned repeatedly about the dangerous condition of much of its network, and had indeed started to make significant inroads into the backlog of repairs over the past two years. But as Hatfield and the dramatic response to it over the past two weeks have demonstrated, this was not nearly enough to achieve a minimum acceptable level of safety.

It is a disgrace that it should take four deaths to force Railtrack to do what it should have been doing already. Equally disturbing is the new evidence emerging that essential safety work was postponed because it was "not cost-effective". Of course there has to be a balance between cost and risk, but we are not talking here about spending billions on high-tech, untested, computerised train control systems. This is a matter of spending a few million on the proven technology of ensuring that two steel rails are the right distance apart, and robust enough to carry predictable loads at predictable speeds.

The poor state of the nation's track was not caused by privatisation. On the contrary, it is the legacy of years of under-investment under public ownership, and it was getting better under Railtrack.

But the latest evidence does suggest that there is a short-term conflict between profit and safety. For that to be resolved, there must be a fundamental reform of Railtrack's management, ownership structure and the financial incentives under which it operates. It cannot be right that Railtrack was effectively penalised, by losing track-access charges from the train operators, if it closed sections of track for safety work. The regulator and the Strategic Rail Authority need to look at how to increase the rewards for safety and the penalties for accidents, which inevitably means less emphasis on speed, punctuality and volume of traffic.

If the current Railtrack management and its regulators cannot come up with a new regime that restores public confidence in the safety of the network, maybe Robert Kiley can. He turned the New York Subway from one of the worst transport systems in the world to one of the best, and has just been hired by Ken Livingstone to knock London Transport into shape. He will be watched with interest.