Ideas percolate.

Two years ago, investment manager Clifford Asness circulated early drafts of his research on the prospects for U.S. equity returns over the next ten to twenty years. "Bubble Logic," the papers title, was grim reading. It methodically obliterated treasured assumptions about the returns investors could expect from stocks.

  Last week he presented his research at the fourth annual Wealth Management Symposium in Chicago. This meeting, the brainchild of Undiscovered Managers founder, Mark Hurley, gathers some of the nations best financial planners to hear some of the smartest researchers in portfolio management and behavioral finance. It's one of those meetings where compelling ideas start their long march to becoming central beliefs.

One thing is certain about Mr. Asness's ideas: they are not likely to be popular. They run against the wishful thinking to which we are all prone. They also run against the ongoing promotion from Wall Street, which regularly assures us that stocks will provide superior returns "in the long run" no matter how absurd their price today.

  Here, culled from his presentation notes, are his fundamental findings:

This Is Still a Dangerous Market. Mr. Asness believes that you can't just assume stocks will return 10 to 11 percent from any price level. His research shows that stocks provide high returns when they are purchased at low levels of valuation. Similarly, they will provide low returns when they are purchased at high levels of valuation.

He isn't alone in this conclusion. Clients of the Leuthold Group were reading similar examinations in the mid-nineties and economist Robert Schiller published similar findings in "Irrational Exuberance" in early 2000.

Using the trailing ten-year average of inflation-adjusted earnings for the S&P 500 to measure the price-to-earnings ratio Mr. Asness found that high P/E ratios matter.

How much do they matter?

I'm sorry you asked. From current levels, stocks can be expected to lose money.

High P/E Ratios Mean Low Future Returns
P/E Range Median Real 10 Year Annualized Worst Real Total Return
   5.2 to 10.1 10.6 percent   45.1 percent
10.1 to 11.7    9.8   36.6
11.7 to 14.1    9.7      4.1
14.2 to 16.7    8.5 (20.1)
16.7 to 19.0    5.3 (32.6)
19.0 to 31.7 (0.2) (36.1)
Source: Clifford Asness, AQR Capital Management, LLC

While seems logical and obvious, it is overlooked by virtually all of Wall Street. He shows, for instance, that Wall Street firms were forecasting large gains for the S&P 500 this year. Last December, a group of nine firms was forecasting median gains of 16.5 percent. At the top end, UBS forecast a fantastic 34 percent gain. Only one firm, J.P. Morgan, forecast a loss.

By his calculations, it would require a 6 percent annual real earnings growth rate for 20 years for stocks to achieve the historic 10 percent rate of return over that period. If earnings rise at their historic real rate of 2 percent, long-term investors can expect losses of 6 to 7 percent a year.

Yes, you read that right. Losses of 6 to 7 percent a year.

And that's the good news. If earnings multiples "revert to the mean" quickly, short term returns could be "much worse."

Mr. Asness believes we are facing two possible paths: if investors accept the idea of very low expected returns, high price/earnings ratios may still be supported and we would live through a long period of low returns. But if investors demand historic returns, stocks will be priced down. He estimates the potential decline at 50 percent.

Hiding Won't Be Easy.   Mr. Asness suggests investors should deal with lower expectations for large U.S. stocks by owning more bonds, including Treasury Inflation Protected Securities (TIPS). He also suggests a shift to small cap, international, and emerging market stocks.

One escape route he closes is investing in value stocks--- those selling at low multiples of earnings, book value, and cash flow. Value stocks were selling at bargain levels two years ago but he now finds they are selling at reasonable prices.

He points out that bonds have worked better than international stocks to protect investors from short-term declines--- such as 1987--- but a global portfolio has worked better to protect investors from long-term declines. In the decade ending September 1974--- a terrible period when U.S. market had a real return of minus 40 percent--- a global portfolio returned 12.3 percent. Similarly, in the decade ending September 1998--- a wrenching period for the Japanese stock market--- a global portfolio produced a real return of 130 percent.

Bottom line: U.S. large company stocks are falling from their pedestal.