I periodically make efforts to get people to use the geometric mean of income per worker, but then I lose energy and slack off...
>.... I think that e(0) is a bad number to choose because the law of
>iterated expectations doesn't apply -- e(0) != e0(e(10)). And the
>distortion is worse becasue you're taking the expectation too near the
>point in life when most fatalities occur. I'd choose e(n) where n was the
>modal age at death.
Touche. That's a good idea...