On the face of it, using historical stock returns to forecast for the future seems unimpeachable. It is certainly common, and the method is simple and clear: Analyze stock returns over some long period, and assume future returns will follow suit.
While this method seems reasonable, it can produce strange results if an investor happens to choose the wrong start or end date in his study. The author gives an example of this — using particular extreme end points and then looking at returns over only the next few years — to make the point that forecasts derived from past averages are often backwards. That is, after periods of strong returns, expected returns may actually be lower for the future.
Many investors learned that lesson the hard way when the telecom-technology bubble burst in 2000. In this paper, the author explores some other lessons that investors learned — or what they should have learned — in that crash. Among them:
- Long-term investors should not be 100% in stocks
- International diversification is not a waste of time
- Dividends are good and for some surprising reasons
- Earnings do not grow at 10% a year
- Value wins in the long term
- You cannot trust Wall Street to compare apples to apples
- Timing the market is not all bad
- The general public is full of bored, innumerate gamblers
- Do-it-yourself trading is a bad idea
The author concludes: "As Warren Buffett says, investing is simple — but not easy."
Graham and Dodd Readers' Choice Award 2005