Yesterday, I posted some information on the 10-year, inflation-adjusted price/earnings ratio of the U.S. stock market. Here is the best, quick way to get a sense of how valid, or invalid, the 10-year inflation-adjusted P/E ratio. Bottom line: it looks pretty valid as an indicator of future returns. Plenty of noise, but some signal in there, too.
You can also get similar statistics in a book by Robert Shiller, Irrational Exuberance. (Go to figure 10.2 in the 2005 paperback edition.) Shiller looks ahead to how the predictor worked over the next 10 years. In contrast, the 2008 article I linked to above, by Christopher D. Carroll, uses a slightly longer period, 12 years, on the theory that it is the smallest common multiple of the Congressional and Presidential election cycles.
POSTSCRIPT: I found a graph of Shiller's future return data online, in a 2001 paper, "Valuation Ratios and the Long-Run Stock Market Outlook: An Update," which he co-wrote with John Y. Campbell. It is as this link, and the most useful part would be the bottom panel of figure 6.