As described, these assumptions are based on the commentary and, in some cases, explicit forecasts of Jeremy Grantham, Robert Shiller, Andrew Smithers, and other analysts. All three base market forecasts on two factors—corporate profits and stock prices—and all three assume that both will eventually regress to long-term means (bearing in mind that such means may change over time and that regression can take decades). Because profits vary widely from year to year, all three also look at long-term average earnings, arguing that the market's valuation relative to a single year of earnings—2004, say—is almost irrelevant. Using this method, Smithers and Grantham, at least, conclude that stocks are overvalued and that long-term performance will be poor.

Specifically, in these 7-year forecasts (updated every month), Grantham's firm, Grantham, Mayo, Van Otterloo, estimates that the median real return for U.S. large-capitalization stocks over the next seven years will be minus 1.7 percent per year (to which the firm expects to add 2.1 percent per year through effective stock-picking). This forecast excludes inflation, which the firm assumes will be 2.2 percent per year. Including inflation, therefore, GMO assumes that U.S. large-cap stocks will return a median of 0.5 percent per year. Importantly, the firm also estimates that the range of possible annual returns is 6.5 percent above or below this median (again, responsible projections are more like shotgun blasts than laser beams). Thus, in GMO's opinion, the average return from U.S. large-cap stocks will likely be between minus 6 percent and plus 7 percent per year for the next seven years.

For my assumptions, I have added a 3 percent annual inflation assumption back to the GMO forecasts and, in most cases, slightly increased the expected return to account for a longer horizon (10 years versus seven years). Using the same range of potential outcomes, this would suggest U.S. stock returns for the next seven years of between minus 5 percent per year and positive 9 percent per year, with a median of 2 percent. I have made similar assumptions with regard to bonds, cash, and funds-of-funds (the latter of which, in my opinion, is anyone's guess: hedge funds have exploded in popularity, and, eventually, this should weigh on the average return). In the case of T-bills, I also looked at yields and mean-regression over the past 50 years.

It is important to underscore that these assumptions are based on the market's tendency to overshoot and undershoot the long-term trend. Instead of delivering consistent performance decade-in and decade-out, the market has had 10-year to 20-year streaks where it has performed above the long-term average, followed by 10-year to 20-year streaks where it has performed below it. Jeremy Siegel recently observed that, thanks to the tepid performance of recent years (the Dow is now basically at the same level it was five years ago, in 1999), stocks have returned to their long-term trend. This observation alone would suggest that the market will deliver its long-term average return over the next 10 years (10 percent), and it certainly might. A look at the entire 20^th century, however, suggests that the recent 18-year period of above-trend stock performance will probably be followed by a similar-length period of below-trend performance. This assumes, of course, that past performance will be relevant to future results, which it might not be. ...