I came across this neat concept today: the statistical person. Statistical people are like real people, but without the uncertainty; whereas a real person might, or might not, be killed by being run over by a bus, we say that several statistical people are killed by busses every year. Because they always get the average, statistical people are certain to be on the receiving end of everything that the rest of us only chance.

The poor statistical person is a second class citizen, however; unlike real people, who it is considered immoral to kill, with statistical people we are allowed to kill as many as we like, provided it can be justified by a cost-benefit analysis. The paper goes into some detail about the reasons why we real people don't care that much about statistical people (although "it won't happen to me" is probably 90% of the explanation).

It is an interesting way of looking at public policy problems. But I can't agree with the thrust of the paper, which seems to be aiming at statistical deaths ranking equally in a moral and criminal sense with real deaths. What's an appropriate punishment for causing 0.1 statistical deaths per year, I wonder?