a longwaves comment on p/e ratios
Thanks very much for calling this important update to our attention.
The current conclusion of Campbell and Shiller--that P/E ratios are
not predictive of future earnings growth rates-- is really implicit
in, and quite evident in, their first paper.
About a month ago I presented here a reference to the recent paper of Pu Shen, an economist with the Kansas City Federal Reserve Bank.
groups.yahoo.com
kc.frb.org
Shen reviewed the original 1998 Campbell and Shiller paper and came to the conclusion that their scattergram chart of P/E's and future earnings growth rates did *NOT* show a valid relationship.
Thus Campbell and Shiller have really been "under the gun" to revise their conclusions. This is important because some have attempted to use the original faulty Campbell Shiller conclusions as justification for currently projecting a major economic
depression going forward.
Our own Bob Bronson, who has done a study of historic P/E cycles, is one of those economic bears, based partly on the Campbell Shiller studies.
Bob had this to say, in part, about Shen's study:
"One thing that she does feature from the current literature is that high stock market P/E ratios predict lower earnings growth rates (as well as lower stock market prices). Of course, this is totally contradictory to those who "hope" that the current declining trend in earnings growth rates(see attached chart) will prove to be transitory. We doubt it because our forecasting models tell us that aggregate corporate earnings growth rates will under perform their recent past for many, many years."
First, this is now shown once more not to be true. Second, this will be a much harderbelief to sustain now that Campbell and Shiller have retracted their original conclusion.
My own belief is that Campbell and Shiller will eventually have to retract their other main conclusion: that future stock prices are predicted by current P/E ratios.
Just to take the scattergram examples relating to high P/E's (and shown in Shen's article), there were 14 years whose trailing ten year P/E ratios exceeded 20. In only 4 of those years
did the subsequent decade's stock price decline by more than 2%. There were six such years with P/E's over 20 whose next decade price *increased* by more than 2%.
If one looks at less extreme P/E's, the results show that there are very few years whose subsequent decade produce lower prices.
Thus the evidence suggests that P/E's are not very useful at all either for earnings or price projections. In fact P/E ratios are like book value per share and other legal and accounting measures which really have little relevance in comparing one era to another. Their relevance is a chesnut of 1930's and 1940's investment approaches which has not stood the test of time.
Tom Drake
Campbell Shiller Abstract:
NBER WORKING PAPER BIBLIOGRAPHIC ENTRY
Valuation Ratios and the Long-Run Stock Market Outlook: An Update
John Y. Campbell, Robert J. Shiller
NBER Working Paper No. W8221
Issued in April 2001
---- Abstract -----
The use of price earnings ratios and dividend-price ratios as forecasting variables for the stock market is examined using aggregate annual US data 1871 to 2000 and aggregate quarterly data for twelve countries since 1970. Various simple efficient- markets models of financial markets imply that these ratios should be useful in forecasting future dividend growth, future earnings growth, or future productivity growth. We conclude that, overall, the ratios do poorly in forecasting any of these. Rather, the ratios appear to be useful primarily in forecasting future stock price changes, contrary to the simple efficient-markets models. This paper is an update of our earlier paper (1998), to take account of the remarkable behavior of the stock market in the closing years of the twentieth century.
++++++++++++++++++++++++++++++++++++++++++++++++++
Date: Sat Apr 7, 2001 9:48am
Subject: Most interesting article by Shiller on PE ratios
papers.nber.org
TD@TenorioResearch.itgo.com
tenorioresearch.itgo.com |
