Why P/E ratios are a poor predictor of market performance.

The Lex Column in the Financial Times today talks about why raw aggregate P/E ratios are poor predictors of stock performance.  Good stuff.  But the final part of the article caught my eye.

Likewise, history shows there to be no predictive power for stocks when comparing equity yields with bond yields. Why should there be? Dividends are risky and rise with inflation; coupons are risk free and do not. It is like buying apples because pears are cheap. There are good reasons why equities are due a bounce – flaky valuation metrics are not among them.

This is basically a take down of the so-called Fed Model (not endorsed by the Fed) that argues that comparing bond yields to earnings yields reveals information about the level of the stock market.  I posted on this a few days ago here.  With bond yields low and earnings yields high, numerous "experts" are claiming this is a signal that the market is undervalued.   They are wrong.  It is a signal that bond yields are low and earnings yields are high.  That is all.