This is an appendix for our earlier post, “What Can Valuation, Debt and the Fed Tell Us about the Next Bear Market.”
The earnings data used for the Shiller P/Es and trend line P/Es are reported (not operating) earnings converted to 2013 dollars using CPI inflation.
Trend line P/Es
Here’s a chart showing that earnings revert to a rising long-term trend line:
The trend line is fitted to the full earnings history for U.S. large cap stocks (according to data supplied by Robert Shiller). We didn’t use it to calculate the P/Es in the main article, though, because we wanted measures that could be calculated at each point in time with data that was known at that time. Therefore, we used trend lines based on trailing 40-year periods. Here’s a chart that shows the trend earnings that were used to calculate the trend line P/Es in the main article (as in the chart above, we’re using exponential trend lines):
As you roll the 40-year periods through time, trend line slopes can be unstable, which is why the normalized earnings estimates shown above sometimes drift apart from trailing 4 quarter earnings for a few cycles in a row. This doesn’t appear to be a problem for the post-1950s period we used in our August article, but it becomes more of a hindrance when trend lines are extended through the world war periods.
On the other hand, the trend line approach mitigates the end point sensitivity of Shiller’s P/Es, which are based on 10-year earnings averages. These averages change rapidly when recessions roll into or out of the 10 year periods. Neither measure is perfect, but they improve upon traditional P/Es. We suggest looking at both of them.
P/Es based on earnings forecasts
You may ask why we don’t rely on earnings forecasts. We have three reasons:
- In economic expansions, the consensus always calls for increasing earnings. Forecasters may expect a short-term dip if company guidance has been weak, but the consensus projection invariably reverses the dip a quarter or two out.
- Based on (1), the consensus never anticipates a prolonged drop in earnings.
- History shows that earnings often defy the consensus and fall further and longer than expected, reverting to a rising trend line (as discussed above).
In short, we don’t believe that forecast P/Es add anything.