Sunday, April 22, 2007

Destroy 600,000 guns , save 1,000+ lives

This story in tomorrow's CT derives from this research.

Andrew Leigh himself explains the modern way it came about:

I never would have gone to read the paper in question if it wasn’t for my father picking up the phone to say ‘you should have a look at this study and see if it’s right’. That led to a couple of blog entries, after which commenter Christine Neill and I (egged on by Justin Wolfers, who has done some great work re-analysing dodgy death penalty research), decided to write up a brief comment paper.

New research suggests that the gun-buyback introduced in the wake of the 1996 Port Arthur massacre has saved between 1,150 and 2,500 Australian lives.

The research, to be published today by economists Andrew Leigh of the Australian National University and Christine Neill of the Wilfrid Laurier University in Canada reinterprets the data used in a widely-publicised released last year by the British Journal of Criminology.

That study found that Australia’s gun-buyback had had no statistically significant impact on Australian homicides...

On the morning of April 28 1996 a lone gunman Martin Bryant went on a shooting spree at the site of Tasmania’s historic Port Arthur prison killing 35 people and injuring many more. It was Australia’s worst mass killing and worse than last week’s Virginia Tech Massacre in the United States.

In response the newly-elected Prime Minister John Howard pushed through tight restrictions on who could own guns and removed 600,000 guns from a population of 20 million.

Last year’s British Journal of Criminology study found that although Australian gun deaths fell after the buyback this was part of a pre-existing trend. It found no evidence that the buyback accelerated the change.

Dr Leigh and Dr Neill discovered that the British study examined only a 25-year time period even though it could have examined the sweep of gun deaths over a century. The time period began at the start of the 1980’s – a time Leigh and Neill say was unusual because gun deaths were historically high and diving. The British study projected that trend forward and concluded that gun deaths would have fallen dramatically with or without the buyback.

But Leigh and Neill find that the maths was flawed. It projects negative deaths by 2010, something they refer to as a “resurrection problem”.

“An effective modeling strategy should place a zero probability on the occurrence of a logically impossible event,” they conclude.

Using a modeling strategy that does not predict resurrections and using all the data available Leigh and Neill find that without the 1996-97 National Firearms Agreement many more Australians would have died by gunfire in the following decade than actually did.

Whereas the British Journal of Criminology study found that the buyback prevented a statistically insignificant number of homicides and prevented 126 deaths per year from suicide, Leigh and Neill find that it prevented between 14 and 35 gun homicides per year and between 142 and 233 gun suicides.

“We find reductions in both gun homicide and gun suicide rates that are statistically significant, meaning that they are larger than would have been expected by mere chance,” Dr Leigh said. “Our best estimates are that the gun buyback has saved a total of between 128 and 282 lives per year.”

Economists typically place a financial value of each life saved of around $2.5 million. Leigh and Neil say in their study that that suggests the $500 million spent on the gun buyback was a good use of public money.