|
I am sure that almost everyone is familiar with one of the great puzzles of all time - the Rubik's Cube. I don't know whether I could ever solve it, but one of my friends could hold it behind his back and in 30 seconds have it completed.
When you read annual reports (an accountant's Rubik Cube) or the mid-year hand outs from companies you are always quoted what the Price/Earnings ratio is at that time. It seems relatively simple - something anyone can comprehend. But is it?
Many of the great market analysts believe that high P/Es (20 and higher) represent the top of the stock market and will ultimately over time come down to low numbers again (5 to 8) although this can take years to occur in both directions. Another fact overlooked by analysts is that computation of the total market P/E leaves out all companies that have no earnings at all. Only companies with positive earning are reported.
The P/E for any particular company can be legally manipulated by very clever accounting practices yet stay within the rules set forth by the FASB, Financial Accounting Standards Board.
Research done on P/E ratios from 1950 to 2001 showed only a random pattern in relation to the ups and downs of the general market. A further study by Ken Fisher and Meir Statman covering 1871 to 1999 showed an unreliable relationship to forecasting future returns using published P/E returns. Just because the P/E was high did not mean the market would go down and similarly when the P/Es were low did not mean the market would go up.
Their studies took the assumption the market should go up or down the next year and failed to take into effect that momentum carries markets to extremes in both directions. Once a person (I hesitate to call him an investor) buys stocks or an index mutual fund he immediately forgets all the reasons he bought it and only cites these when it starts to go down to justify his mistake. The P/E has nothing to do with whether your stock rises or falls and P/E is not reliable as a predictive tool when taken alone. Most have a positive bias and that means a lower number.
It is my contention that P/E data should be viewed using some type of moving average such as 5 or 10 years. This would smooth the data, but probably not be any help for predictive purposes.
Keep in mind that company P/Es are based upon how the company accountant wants to manipulate the figures. To name just a few: leveraged borrowing, timing costs, write offs, inventory valuations, restructuring charges, capitalizing costs and many others. All this "stuff" gives the final P/E. This is Rubik Cube accounting and it takes a very clever accountant to understand what the other one did with the numbers. It is an analyst's nightmare.
Don't be puzzled by the Rubik Cube P/E trap as the P/E is almost meaningless.
Copyright 2004 Albert W. Thomas All rights reserved. Author of "If It Doesn't Go Up, Don't Buy It!" www.mutualfundmagic.com comments to al@mutualfundmagic
|
