Friday, December 7, 2007

Fatal Truck Crashes: A Publishing Opportunity for Economists

From the Pump Handle comes this. I’m too busy to turn these findings into an economics journal article, but I’ll tell you how to do it so you can pad your own CV.


First you need a theoretical model. You can show that, with a fixed salary and positive barriers to mobility (for instance in a searching/matching model of the labor market), truck drivers are unable to optimize on their relative preferences for money income and safety. This welfare loss can be overcome by the payment of piece rates. Now each truck driver can locate his or her own personal optimum in wage/risk space. (We abstract from the welfare benefits/costs of the direct effects of amphetamine use.) But there is also a potential externality, in that other drivers or pedestrians may be killed or maimed in truck accidents. The solution is a nonlinear wage schedule that reduces effective hourly pay by the expected cost of the externality according to increasing effect of work hours on accident rates. If you want to get even fancier, you could throw in the principal-agent dimension and put your solution in the context of optimal contract theory.

For the empirical section you would need the raw trucker data. It would be very simple, really a spreadsheet exercise, to impute the marginal value of an extra hour’s work from the piece rate schedule, and to calculate the marginal increase in the probability of a fatal accident. From this you could determine the VSL (value of a statistical life). The required level of analytical foggery could be achieved by testing for baseline effects, income and substitution effects, lots of stuff.

One thing that would not go into the article would be the observation that all of the above smiles on personnel practices that kill truckers.

It’s really a shame that they don’t let me supervise Ph.D. theses.

1 comment:

Michael Perelman said...

Thanks for the interesting post.