Steven Pinker recently wrote controversially of the victory of the better angels of our nature: we are, he claims, less likely to suffer violence at the hands of other people today than at any time in history.

But what of accidental death?  Almost fifty years ago, the historian Paul Hair attempted a similarly sweeping survey.  Using coroners’ records and other sources, he suggested that the likelihood that someone in England would meet a violent end has remained broadly constant for seven centuries.  ‘The axe of the drinking companion and the neighbour’s open well were regulated, to be replaced by unruly horses and unbridged streams; when these were brought under control it was the turn of unfenced industrial machinery and unsignalled locomotives:  today we battle with the drinking driver.’

I have often referred to Hair’s work in illustrating the problem economists call ‘moral hazard’.   If people are protected against risk, they will take more of it.  In his nice little book ‘Risk’, John Adams employed the metaphor of the risk thermostat.  We have a certain tolerance, and if the atmosphere cools we turn up the heat.

Yet if Hair were alive today (he died, peacefully, a decade ago) he might be surprised to revisit his statistics.  Since he wrote, the incidence of accidental death has fallen sharply – by almost half.  The principal reason is the decline in road fatalities.  The battle against the drinking driver has essentially been won, and compulsory seat belts and safer cars have also contributed.  The likelihood of being killed on British roads is barely a quarter of the level of fifty years ago.  Industrial accidents, which had become the major cause of violent death in Victorian times, have also been in steady decline (though most of this decline had occurred by the time Hair wrote).

  The principal causes of accidental death today are falls and suicides.  If you come to an unexpected violent end, it will probably be as a result of falling down the stairs at home.

Yet the metaphor of the risk thermostat remains relevant.  In modern Britain, drug abuse kills more people than road accidents. The residual problem of drunk driving mainly involves young men who kill themselves when plastered. Hair was right:  when people are no longer free to poison themselves at the neighbour’s open well, some will poison themselves anyway for the thrill of it.

There are two broad lessons.  One is that the issue of moral hazard is real. It is not just drug addicts who take risks because of the high it gives them. Other people take inappropriate risks in the deluded belief that ‘it will never happen to me’.  But the best corrective to that belief is the experience of seeing it happen to other people.  Fannie Mae and Freddie Mac would never have assembled the bloated balance sheets they did had their lenders not believed – correctly as it turned out – that the US government would make them whole.  The lesson of the Lehman bankruptcy was  appropriately salutary: the problem  was that the lesson to creditors was so salutary that it threatened to destroy the whole financial system.

The second lesson is that well designed public policy interventions can make a difference.  The dramatic reductions in industrial injuries and fatal road accidents have happened because government has effectively enlisted the support of employers and the affected public.  ‘Health and safety’ has become the butt of jokes in Britain – so much so that the Health and Safety Executive has established a ‘challenge panel’ to refute absurd claims about the obligations it imposes (consulting the list of cases referred to the panel by the public provides an amusing diversion).

Yet the policy is a demonstrable success.  The aggressive implementation by contractors of their own well-designed safety policies in high risk industries like construction has  substantially cut accidents.  And telling people to fasten their seat belts and not to drive when drinking goes with the grain of what people themselves know makes sense.

Policy towards risk in the financial sector needs to learn both lessons.  If creditors are protected from risk, the long run effect will be more risk in the system, not less.  And the policies that will prove effective in managing risks will be policies that financial services firms design for themselves.   They need much stronger incentives to do so.