Neil and Leigh do not make strong claims in this regard – it is a hard call since homicide rates and suicide rates have been trending down strongly through time anyway. The punch line in the author’s claims is that homicide rates are bounded below at zero so that recent substantial decrements in this rate, since 1996, are very significant since the rate itself cannot fall below the natural barrier of zero.
In Andrew’s blogpost, though not in his article, he calculates the cost of the buyback at $500 million and then aggregates the suicides and deaths together to deduce that a minimum total of 128 lives per year were lost. Taking recent estimates of the value of a human life at $2.5 million he calculates that the buyback paid for itself in two years.
That sounds right if the value of a suicide prevented is taken at $2.5 million. But people presumably kill themselves when they feel their own lives are not worth living. Their incomes may be low, their prospects poor or they may suffer from serious debilitating diseases. In short to follow the ‘optimal suicide’ literature (Hamermesh and Soss) ‘as soon as the terrors of life reach the point at which they outweigh the terrors of death a man will put an end to his life’ (Shopenhauer, On Suicide). Adopting this viewpoint, preventing a suicide may not increase society’s wealth – it may in fact decrease it if you respect individual preferences. Presumably this is the idea behind the case for voluntary euthanasia.
Even if you dislike the macabre notion of attaching a zero welfare gain to preventing someone from killing themselves I am unsure that preventing people from killing themselves with a gun substantially reduces the suicide rate. Are not things like sleeping pills and carbon monoxide relatively painless substitute ways of killing yourself? If this is so then gun-driven suicides may be replaced with other types of suicides.
In either case, the cost-benefit case as presented becomes weaker. Including only the murders and maintaining a zero discount rate the results would suggest a net gain from the gun control measure if account is made of murders saved over 100 hundred rather than 2 years if the lower bound on effects is taken.
Another puzzling point that Neil and Leigh do not deal with is that suicide rates apparently declined for different forms of suicide - not just those involving guns. Thus the decline in the suicide rate might be hard to attribute to gun control.
Lest I be misunderstood I am in no way arguing that the ‘gun buyback’ was not good policy. This is only a qualification on what seems to me a very interesting study. I have lived in a society where there was widespread gun availability – Thailand – and I think that the fears that are a consequence of widespread gun ownership outweigh any benefits. The murder rate in rural Thailand while I lived there in the 1980s was massive and much of it was associated with gun use.