We all know people who would not countenance capital punishment for any other crime but who, nonetheless, support it for the murderers of children. Those who believe that the state in no circumstances has the right to take life are in a minority - albeit an influential one - and beyond this argument. There are others who find no philosophical difficulty in having the state execute murderers, provided the law enabling execution is passed by democratic means, and provided those executed have been found guilty after a fair trial. However, many such people still withhold support for capital punishment because of the risk of the wrong man being hanged. It is they who most need to reconsider their view after the events of this week, particularly the apparently clinching evidence of Hanratty's guilt.

The question of whether all murders, or certain sorts of murder only, should merit capital punishment is a complex one. The death sentence used to be mandatory for all murders, though in 1957 a separate category of "capital murder" was created. This meant a criminal was liable to be hanged only for murders committed in the course or furtherance of theft, or by shooting or explosions, or for the murders of police or prison officers or those assisting them, or for two or more murders on separate occasions. A child killer, therefore, would not have been sentenced to death after 1957 unless he had murdered two or more children in separate incidents, or had shot them, or was committing a robbery. Even when all types of murder attracted the death sentence, fewer than 50 per cent of those condemned to death were actually hanged.

Although ignored in the framing of the 1957 Homicide Act - in that relatively innocent era before the Moors Murders changed the perspective on such crimes for all time - the case for capital punishment for those who kill children is peculiarly strong. Moreover, it is strong not simply for the emotional reasons patronised, ridiculed and despised by the liberal establishment. There are other, harder arguments that merit serious consideration.

First, child murderers tend to leave DNA evidence because of the sexual nature of the crime. There was, admittedly, no DNA evidence to convict Michael Stone, now serving life for the murders of Lin and Megan Russell in Kent, and for the attempted murder of Josie Russell. However, Stone struck with a blunt instrument, did not appear to engage physically with his victims, and did not carry out a sexual assault. In cases where there is physical contact or sexual assault there will normally be DNA evidence. Once the argument of mistaken identity is removed from the debate, the issue becomes far clearer.

Second, psychologists on both sides of the argument admit that the paedophiliac impulse is not one that, like the urge to steal cars or rob old ladies in the street, evaporates with age. Many paedophiles are still active in their seventies. Of course, not all paedophiles are murderers; but given the nature of their condition, it is hard to justify the release of those who have not killed. Even to contemplate releasing one who has committed murder is little short of diabolical.

However, to reduce the argument to one of economy - that public money can be better spent than on keeping child killers in custody, and execution is advantageous in this regard - is a contention considered by many to be in bad taste, even where these people are concerned.

The third and fundamental argument is far more powerful: that of "proportionality" - that the punishment should fit the crime. "Proportionality" is one of the key considerations in the current review of the sentencing framework being undertaken by the Home Office. If you have an objection in principle to capital punishment, then the proportionate sentence for child murder must be a life sentence, for it is the nearest to true proportionality allowed by law. If you have no principled objection, and once you have eliminated mistaken identity, then proportionality in cases such as Sarah Payne's can mean only one thing.

Many who instinctively feel that Sarah's murderer should be executed may, almost apologetically, claim they are motivated by righteous indignation and a desire, bred from it, for revenge. That is not how the spin-doctor for the cause of capital punishment would present the case: he would rather say that the wish to have this man executed was prompted by an instinctive sense of natural justice, and an adherence to the doctrine of proportionality.

That spin-doctor might also avoid, as I have done, getting into the murky waters of the argument about deterrence. Those who argue that capital punishment does not deter people from committing murder, even when the offence is narrowed down to specific sorts of murder, cannot prove their case. However, neither can those who argue to the contrary. The murder rate has trebled since 1964, but abolitionists argue that that is merely consistent with other sorts of crime, and attributable to social factors. Perhaps it would have trebled had hanging been retained: we shall never know, though many who believe in restoration have their own, strong views on the question.

Opponents of capital punishment cite not just Hanratty - though they may have to forget about him now - but also the Guildford Four, the Birmingham Six and assorted other people convicted of murder but subsequently released in the light of new evidence or proof of unsafe convictions. These are not clinching arguments, however. We do not know whether, had capital punishment still existed, the crimes for which they were wrongfully convicted would have been carried out in the first place. We do not know whether, if they had been, a jury would have convicted the accused, knowing they would suffer the penalty of death. Also, these miscarriages took place in the era before the Police and Criminal Evidence Act and before the existence of DNA testing.

There has been a huge leap in forensic techniques and police practice since the 1970s. It is not just that when the police seek to convict someone of murder they can provide far more watertight evidence than used to be the case: juries will simply not convict unless they can. In the case of paedophile murderers, that evidence should always be obtainable.

Those who have long believed that the sentence of death is the only appropriate punishment for paedophiles who kill children feel that Sarah Payne was an inevitable victim of a society where liberalism has become callousness and an indifference to the rights of children. Society did not kill her, but in some measure it is society's fault.

The exactitude with which we can identify a killer 39 years on creates a new philosophical and legislative option for those whose desire to protect children has, hitherto, been purely rhetorical. Will they take it, or is this harmless little girl's destruction still not enough?

• The writer, a Daily Mail columnist, is a member of the External Reference Group of the Home Office's Review of the Sentencing Framework

The case for capital punishment   -  The Telegraph - Simon Heffer  -  23 Nov 2005