How safe are our reactions?

What hope is there for a rational debate about environmental issues such as nuclear power?; Magnox nuclear plants are not dangerous. But all discussion of risk is beset by human illogicality

Yesterday Nuclear Electric presented its side of the case over the the accident at its Wylfa power plant in Anglesey, but it has pleaded guilty to the charges. Already, though, the jibe of the prosecution - that the nine-hour delay in shutting down the reactor was "a classic example of brains in neutral" - has struck home.

So Jack Cunningham, the shadow trade and industry secretary, was able to say that the proposed privatisation of the nuclear power industry "will exacerbate the conflict between safety, accountability, profits and dividends". And Greenpeace argued that this showed that all nuclear power was unsafe and that Magnox reactors in particular should be shut down as soon as possible.

This is a fascinating example of human attitude to risk, and sadly, shows how difficult it is to have a sensible discussion about the subject. For whatever one's views about nuclear power - and I happen to think it is an inappropriate technology that will gradually be superseded - this particular accident surely disproves both reactions. The problem with Mr Cunningham's view is that this case shows that public-sector managers have a cultural approach to safety that would seem very odd in the private sector. Nuclear power was excluded from the initial electricity privatisation five years ago mainly because the financial markets were unprepared to shoulder the contingent risks: the private sector in general is much more risk-averse than the public sector.

As for Greenpeace, this shows how inherently safe the technology is, if managers are able to ignore a problem for nine hours and still no one is hurt at the end. The old Magnox gas-cooled stations are uneconomic but they have proved pretty reliable and have the particular advantage that the reactors operate at a low pressure and are encased within a large concrete shell. Because the mass of the reactors is so great, if there is a failure in the cooling system they overheat much more slowly than the much more common pressurised water reactors (PWRs), which run at higher pressures and are contained within a smaller steel vessel.

In other words, when Magnox reactors go wrong they do so slowly, giving operators plenty of time to decide how they should respond.

By contrast to this relatively benign UK experience, the 1979 disaster at Three Mile Island in the United States showed how, when the operators of PWR technology make a mistake, there is much less time to correct it. And the lesson of Chernobyl is surely that the extreme public-sector culture of the Soviet Union relegated safety to the bottom of the scale of human values.

Anyone still unconvinced should reflect on a further question. Suppose a similar situation occurred in another industry - say, on a North Sea oil platform - there is no way the operators could put their brains in neutral for hours and not face a disaster.

And so to the core of the puzzle: one can demonstrate that this a safe technology until one is blue in the face and people simply will not believe it. Yet these same people will take enormous risks in their own lives, not just by known "risky" activities like skiing or riding a motorbike, but more simply by smoking or failing to take regular exercise. There are still people who are frightened of air travel, despite its outstanding safety record. Why are we so irrational?

There are, I suggest, two problems. The first is that we are innumerate and therefore cannot assess risk properly; the second, that even when we do know the odds, we deliberately kid ourselves about them.

The difficulty we have in interpreting data has been brilliantly illustrated in a new book by John A Paulos out in the US, A Mathematician Reads the Newspapers - yes, journalists are among the worst offenders. Take a very simple statistic, infant mortality, a measure commonly used to show the quality of health services in a country. The US does rather badly by the standards of developed nations and many have used this to suggest that US health care is failing. But the US has a large number of very low- birth-weight babies, which inevitably have far higher mortality than normal ones. Statistically these tend to be born to single mothers, and the US has a high proportion of these. The conclusion is not so much that there is a problem of the health care industry, but rather one of the States' fragmented society.

There are myriad other examples of confusion or, worse, deliberate distortion of statistics, but we find risk hardest of all to comprehend. People find it difficult to accept that our roads are, on most measures, the safest in the world and that the number of deaths from traffic accidents last year was the lowest since the 1920s. If they cannot accept that, how can they grasp that producing electricity from nuclear power is safer than doing it from coal, oil or gas?

As noted above, the people who influence the financial markets are just as prone to miscalculation as the rest of us: they will weigh the risks and provide finance for mining and companies, but not for nuclear power. Collectively we are as irrational as we are individually. Why then do we kid ourselves?

The most thorough, if controversial, discussion of risk is by John Adams in his simply titled Risk (UCL Press). His thesis is that we have an ambivalence about danger and that efforts to reduce risk are doomed to failure. His most famous example is compulsory seat belts, where he shows that accident rates have fallen faster in countries where their use is voluntary than where they are required by law.

Adams's idea is that drivers wearing seat-belts compensate for any additional protection they feel by driving more aggressively, thereby transferring the risk to others. (An extreme version of his thesis is the "Volvo drivers syndrome" noted by motoring journalists - they feel so safe that they drive on the assumption that everyone else will get out of their way.)

Of course one can counter the Adams argument by reductio ad absurdum: if we deliberately made our cars as dangerous as possible, fitting spikes inside so that we would skewer ourselves if we had a crash, then we would surely increase the death toll. But it is hard to counter the argument that it is something other than seat belts (road design, cultural attitudes, better driver training?) that has cut road deaths.

If we cannot have a sensible discussion about the risk of something as straightforward as road safety, what hope is there for a rational debate about grand environmental issues such as nuclear power? John Adams is glum: can we manage risk better? he asks, and his answer is that we cannot. We will continue to flounder.

But as far as nuclear power is concerned it may not make much odds, for the decision is down to harsh economics. Under any sensible assessment of the costs of waste disposal, the whole thing is too expensive. Other technologies are cheaper and more flexible. It will be the besuited accountants, not the woolly-jumpered "environmentalists", who succeed in engineering nuclear power's retreat.