Tuesday, May 26, 1998
David Dreman has never tried to find the next Dell or the next Microsoft. The omission may have reduced the story content of the portfolios he has managed but it hasnt dampened their performance.
While Michael Murphy and others invent new ways to value esoteric technologies, Mr. Dreman ignores technology and new valuation methods. Instead, he is the Maharajah of the Multiples, searching for stocks that sell for low multiples of earnings, cash flow, and book value. Indeed, his most consistent interest hasnt been valuation methods but investor psychology— how amatuer and professional investors regularly make the poor decisions that cause them to under perform the market. Starting with Psychology and The Stock Market in 1977, Mr. Dreman has written a series of books around his findings.
Last week I was as eager to listen to him when he visited to talk about his new book— Contrarian Investment Stratagies: The Next Generation— as I was after reading Psychology and The Stock Market twenty years ago. Listen:
"When I wrote back in the seventies and eighties there was only one contrarian strategy— low p/e. Now we have four: low p/e, low price to cash flow, low price to book value, and high yield. We continue to find that the widows and orphans stocks beat the market.
"Weve also found a whole new contrarian strategy. We asked the question: Does it work to buy the low p/e stocks in an industry? We found that it did.
"This opens the entire market place and makes it possible to have a diversified portfolio." (One of the limits of an absolute low p/e strategy is that it will routinely show that stocks in a single industry hold the low p/e position.)
"We found that whether you look at p/e or price to book, the return on the lowest multiples is better. We also think this method is a lot more reliable than using analyst forecasts. With better data bases now available, we researched analysts forecasts— some 80,000 consensus forecasts— and found that the average annualized error was 44 percent! Think about that. If a company reports earnings that are 2 percent off forecast, its a stock market disaster. Yet the average forecast error is 44 percent.
"Let me put it another way. The chance of getting an (earnings) estimate that is within plus or minus 5 percent of actual results for 20 quarters is one in 50 billion. Thats not very good odds."
What his research shows is that historical evidence is a better decision making tool than forecasting. Investing in the lowest quintile stocks in each industry group, he found that all four of his tools beat the market.
From 1970 through 1996 the market returned 14.9 percent. But high relative dividend stocks provided a return of 17.0 percent; low relative price to book value stocks returned 17.8 percent; low relative price/earnings stocks returned 17.7 percent; and low relative price to cash flow stocks returned 18.4 percent. (In each case the stocks were ranked relative to their specific industry.)
That small difference in percentages turns into a massive difference in dollars. While $10,000 invested in the market in 1970 grew to $289,000 it grew to $572,000 in low p/e stocks and $626,000 in low price to cash flow stocks.
In fact, there is a direct relationship between returns and multiples: the higher the multiple, the lower the return. The table below shows the industry relative p/e approach for two, three, five, and eight year periods from 1970 through 1996:
|P/E Quintiles||2 years||3 years||5 years||8 years|
Source: Contrarian Investment Strategies: The Next Generation, pg. 197
Mr. Dreman also examined the same strategy in bull and bear markets. The low multiple measures did better than the market going both ways— they tended to lose less in bear markets and gain more in bull markets.
"This is also a very tax efficient strategy. When we buy and hold we still get the rate of return we get in year one. Holding means we reduce turnover and taxes."
Next Tuesday: Dreman on academic risk versus real risk.