When Stats Bite Back

By Edward Tenner

Whatever the causes of the Washington DC Metro crash that took the lives of nine people and injured scores of others, our first thoughts will be with the victims' families, and our reaction will be Never Again.

But there is another way of looking at casualties, statistical and actuarial, a style of thinking prized and highly paid in the Capital and around the Beltway and in academia. To the humanist rhetoric of infinite human worth, the idea that when a person dies a world is destroyed, it opposes the economic logic of limited resources and the bureaucratic watchword of consistency.

At an academic lunch some time ago, after a report of the collapse of a local parking garage -- I don't remember whether there were casualties -- a colleague in the policy culture launched zestfully into the thesis that not enough structures were falling down, that it showed that they were being built too safely, misallocating resources. There seemed to be an optimal number of building collapses. Maybe he was just provoking the rest of us, but I recall him sounding like he meant it.

Stephen Colbert lampooned the Environmental Protection Agency's reduction of the value of a statistical life a year ago as a typical Bush tactic to subvert regulation:

The Colbert ReportMon - Thurs 11:30pm / 10:30c
The Word - Priceless
www.colbertnation.com
See web-only content:
http://www.theatlantic.com/technology/archive/2009/06/when-stats-bite-back/20047/
Colbert Report Full EpisodesPolitical HumorStephen Colbert in Iraq

The disturbing truth is that even at the old, higher number, the loss of 9 human lives would not be grounds for replacement of the older model cars offering less survivability. Even if all nine casualties could have been spared, the $888 million estimate cost of replacing 1970s cars newer, safer models would have been almost $100 million per life, more than twelve times the pre-2008 $8.04 million statistical value of life used by the EPA.

The Department of Transportation uses slightly different, and lower figures. From a 2002 memo, apparently still current, on its web site:
 

Based on our improved understanding of the academic research literature, we have determined that the best present estimate of the economic value of preventing a human fatality is $5.8 million. This value should be used, effective immediately, for analyses performed by DOT analysts. In addition, we will, for the first time, require supplementary analyses at values for a statistical life higher and lower than $5.8 million. Specifically, analysts will prepare estimates based on assumptions of $3.2 million and $8.4 million for the value associated with each life saved.

So when a Washington policy analyst leaves the above-ground air of the District and enters a subway train, his or her statistical worth (leaving aside the age-weighting that the regulatory czar nominee Cass Sunstein favors; see the bioethicist Wesley J. Smith's blog) drops from $7.22 million to $5.8 million, or almost 20 percent, and is restored again at the end of the journey, gaining well over a million dollars just by stepping off the escalator. F. Scott Fitzgerald didn't know how right he was when he observed that "the test of a first-rate intelligence is the ability to hold two opposed ideas in the mind at the same time, and still retain the ability to function."

This article available online at:

http://www.theatlantic.com/technology/archive/2009/06/when-stats-bite-back/20047/