VSL at the bottom tail of the distribution

Economists measure the value of a statistical life by looking at the compensation required to induce somebody to take a known risk of death. American estimates typically range around $7 million.

Radio New Zealand reports that some Kiwis are risking death stealing copper from live power lines for sale on the scrap metal market.

Copper has been stolen from live power lines in Whangarei six times in the last fortnight and Northpower says the thieves are risking their lives.

Network manager Graham Dawson says the thieves might get only $5 - $10 for the metal from a scrap dealer - yet they are putting their own and the lives of others at risk.

I'd be curious if Northpower has any estimates on the fatality risk accruing to those stealing copper in this way. All we'd really need are numbers on reported thefts and mortality events. It's easy to Google up cases of people dying doing this; it's also easy to find some less than appealing pictures of the results, so be warned.

Would you expect a 50/50 chance of dying if you did this 20 times? 100 times? The current NZ VSL measure used by MoT is $3.67m*. If the average copper heist from power lines yields $100**, then a thief who values his own life at the MoT estimated VSL expects a risk of death of 1/36,700. That seems very low. I've not been able to find any stats on number of thefts, or on aggregate deaths, but this story has 4 deaths in the first half of 2010 in the Appalachian region; I can't believe that there were close to 150,000 incidents of theft of copper from power lines over that six month period. It's rather more plausible that the risk is much higher and that individual VSL values are more heterogeneous than the average figure used for policy purposes.

It's worth keeping in mind strong heterogeneity in individual willingness to accept risk when thinking about regulations around safety.

* It's a measure derived from surveys of what people would be willing to pay for a roading improvement that would save on average one life, not from individual real-world choices over risks. We might expect a NZ estimate based on risk preferences to give a number somewhere around $5m if we work backwards from US figures and apply estimates of the elasticity of VSL with respect to income.

** Assuming that Northpower is lowballing the scrap prices to try to deter thefts. Scrap copper pays about $7/kg; even if stolen copper trades at a discount, would a copper theft really yield only a kilo or two?