Validity of forward earnings estimates:
Does anyone know if the validity of the forward earnings estimates has been validated using historical data?
If forward earnings estimates aren't very accurate then looking at values of P/Eforward may not be very informative.

To validate using historical data, one could plot the actual earnings at a point in time along with the forward estimate of earnings for that point in time. Could quantitate level of agreement as well of course.
Perhaps the Yardeni folks have done this, but I'm not very familiar with their voluminous data.

Separate topic:
I love the difference idea. Idle speculation: maybe it's subtracting out some in-common noise. Anyway, it makes one want to go on a differencing hunt.

Echoing others, many thanks for all the work you've put into this wonderful forum Manlobbi!