I wrote a paper last summer, “The CAPE Ratio: A New Look.” I presented it at a Q-Group conference in October. Bob Shiller and I debated it the day after he won the Nobel Prize. (Because of all the media coverage he appeared by videoconference rather than in person.)
The bottom line of the paper is I have a lot of problems with the CAPE predictions using S&P data, but not with the CAPE methodology. The CAPE methodology is brilliant and works. The data are the problem. I showed that the S&P 500 earnings data over the last 15 to 20 years, particularly in recessions, have been much different that prior to that, primarily due to the change in accounting conventions and the forced write-downs of assets. When I looked back at the historical data, S&P earnings compared to the business cycle, it was like night and day. You are not dealing with the same series.
I used the corporate profits from the National Income and Products Account (NIPA) data and plugged them in. Guess what? The Shiller overvaluation almost completely disappeared. And the cyclical behavior of the NIPA data is very consistent with the business cycle over the last 85 years.
When you normalize for the data, the market is not overvalued by much at all. Because Shiller’s CAPE is based on the average of the last 10 years, the crash in earnings during those recessions really pumped up the CAPE ratio, and that is a major, major reason for today’s over-valuation.