Remember the Fed Model?

If your memory needs refreshing, here’s a quick review. For years, there’s been a strong correlation between the earnings yield for the S&P 500 (inverse of the forward price-to-earnings ratio) and the yield on the 10-year Treasury note. The implication is that stocks are fairly valued when the two are equal. When the earnings yield for stocks is greater than the Treasury yield, stocks are undervalued. When the Treasury’s yield is higher, stocks are overvalued relative to bonds.

The last time I read a detailed write-up about the Fed Model was in 2001, right after the markets reopened in the wake of the Sept. 11 terrorist attacks. That weekend, Barron’s featured a rarely bullish cover headline: “Buy Stocks Now.” The article stated that stocks were then undervalued by about 17%.

The piece actually inspired a rush of institutional buying for about three months. The so-called Fed Model reading – and Barron’s subsequent bullishness – apparently converted many bearish investors who had dumped stocks immediately after the terrorist attacks. The major indices tacked on strong gains before rolling over and falling further in 2002.

Since then, I have read little about the Fed Model in the financial press, and I thought it would be a good time to update readers on the model – and its current bullish reading.

Dr. Ed Yardeni “discovered” this model buried in a July 1997 Federal Reserve Monetary Policy Report. The Fed has never officially endorsed it, but it’s fairly obvious from the report that the Fed regards the model as a good indicator of whether stocks are undervalued or overvalued.

I devoted an entire book to the subject (Stocks, Bonds and Greenspan’s Model) in 2002, and I found the model’s accuracy in predicting future stock price movements very helpful. In particular, it does a great job of helping investors decide whether they’re better off owning bonds or stocks at any given time. I’ll show you how you can calculate the reading on your own, and we’ll take a look at the model’s reading right now.

The Math Behind the Model

Here’s how to calculate the Fed Model reading.

• Divide 1 by the present 10-year Treasury note yield. That’s 1 divided by 4.6%, which equals 21.74.
• Multiply your answer by the 12-month forward earnings estimate for the S&P 500. Right now, the consensus has that number at $67. So you’d multiply 21.74 by 67, which gives you 1456. That product is the fair value for the S&P 500. (Consensus earnings estimates can be found in Chuck Hill’s weekly Market Commentary.)
• Subtract the current value of the S&P 500 (roughly 1140 as of Thursday) from the fair value. That’s 1456 minus 1140, or 316.
• Divide that answer by the fair-value number. So that would be 316 divided by 1456, or 21.7. Therefore, right now, the Fed Model reading states that stocks are undervalued by 21.7% relative to bonds.

I ran a ton of numbers going back to 1979 to gauge the Fed Model’s accuracy in predicting future stock price movements. My study covered a period when stocks were undervalued by about 40% in 1979; overvalued by about 40% just before the October 1987 crash; and overvalued by a whopping 70% in January 2000. Near the bottom of the 2000-2002 bear market, the model indicated stocks were undervalued by about 40%.

In particular, I looked at all the weekly data beginning in 1979 to see how good the model was at predicting stock prices in the subsequent 13-, 26- and 52-week periods. I’ll share some of the highlights from the 52-week period study:

• When stocks were undervalued by 15% or more, the subsequent 52-week return of the S&P 500 was about 16%. Positive returns were recorded almost 90% of the time.
• When stocks were within a 5% range on either side of undervaluation and overvaluation, the subsequent 52-week return of the S&P 500 was about 12%.
• When stocks were overvalued by 25% or more, the subsequent 52-week return was a negative 6%.

Acting on the Information

From what I learned, and if history is any guide, it generally makes sense to reduce your stock holdings when the reading reaches extreme overvaluation and to increase your holdings when it’s at an extreme level of undervaluation. And simply stay invested in stocks, up to whatever your own target allocation may be, during more normal or neutral readings.

The Fed Model is not a short-term, market-timing model, and I certainly wouldn’t suggest you use it to help you trade. Furthermore, it’s not infallible, and it’s only as good as the accuracy of the earnings estimates that are plugged into the formula.

Having said that, the model can give you a pretty good idea of whether you should favor stocks over bonds or vice versa. Today’s reading, which shows stocks to be undervalued by almost 22%, could encourage you to stay put in stocks for the time being. Obviously, if interest rates spike up and/or earnings estimates decline substantially, stocks will become much less attractive relative to bonds. But until that happens, stocks may remain much more attractive than bonds.

I doubt this weekend’s cover story in Barron’s will be titled “Buy Stocks Now, Again,” but it should be.