Monday, August 22, 2011

Hussman Weekly Market Comment: Whack-A-Mole

As of last week, the S&P 500 has declined to the point where we now expect 10-year total returns averaging about 5.7% annually on the index. This is certainly higher than the 3.4% prospective return we observed earlier this year, but is still a prospective return more characteristic of market peaks than of long-term buying opportunities. Wall Street analysts continue to characterize stocks as cheap on the basis of completely specious approaches like "forward operating earnings times arbitrary P/E multiple," or worse, "forward operating earnings yield divided by 10-year Treasury yield." Unfortunately, despite a few anecdotal successes, there is no correlation between "valuation" on these measures and actual subsequent market returns.

There are numerous reasons why these toy models based on forward operating earnings are misguided, but the four most important ones today are 1) forward operating earnings presently carry the embedded assumption that profit margins will achieve and indefinitely sustain the highest profit margins observed in U.S. history; 2) the duration of a 10-year Treasury bond is only about 8 years, while the duration of the S&P 500 is about 42, meaning that any given yield increase implies 5 times more loss for stocks than it does for bonds, and there is no reason in the world why investors should treat those risks as equivalent; 3) the current conformation of evidence strongly suggests the likelihood of an oncoming U.S. recession, and forward earnings expectations tend to be stunningly off-base in those instances, and; 4) the norms typically applied to forward operating earnings are artifacts of the recent period of bubble valuations, and use norms for "trailing net" as if they are equally applicable to "forward operating." In fact, the correlation between forward operating P/Es and other normalized P/Es having far longer history suggests that a forward multiple of even 12 is quite rich.

As it happens, forward operating earnings, when used properly, can actually be very informative about prospective market returns (see Valuing the S&P 500 Using Forward Operating Earnings ). However, the phrase "used properly" can't be emphasized enough. Here and now, our forward operating earnings model delivers nearly identical prospective return estimates for the S&P 500 as our standard methodology. Stocks are emphatically not undervalued here on any reasonably long-term horizon.