Excerpt from fund manager John Hussman's weekly essay on the U.S. market, in which he attacks the notion that stocks are cheap based on the "Fed Model" (compare the yield on Treasuries to the earnings yield on stocks):
...the loose one-to-one correlation between the forward operating yield and the 10-year Treasury yield is largely an artifact of stocks having been deeply undervalued in the early 1980's and profoundly overvalued by the late 1990's. So yes, interest rates fell during that period, but stock yields fell far more than can be attributed to the decline in interest rates alone. To attribute the entire decline in yields to interest rates as if it is a “fair value” relationship is to introduce a profound “omitted variables” bias into the whole analysis, which is exactly what the Fed Model does.
One can quickly validate that criticism (and invalidate the Fed Model) by noting that there is nothing close to a one-to-one relationship between interest rates and earnings yields – normalized or not – in historical data prior to 1980.
In short, the valuation tools upon which Wall Street analysts increasingly base their analysis are, in fact, pure unadulterated garbage. Over time, investors will discover this along with a good deal of pain.