Fed Model's New Math: Bonds Are Too High

While usually used for stock prices, the calculation can also be read to forecast higher yields.
Author:
Publish date:

Even after six weeks of travail in the bond market, a widely used model that compares corporate earnings to interest rates still suggests fixed-income securities have further to drop.

That's right. Although analysts typically look at the so-called "

Fed

model" to determine whether stocks are cheap in comparison to bonds --and that's really its main purpose -- some pundits now view the valuation tool from a different angle.

"Investors who believe that the Fed model currently reveals an anomaly in the relationship between stocks and bonds may be correct," said Mark Haefele, chief financial officer and chief operating officer of Sonic Capital and contributor to

Street Insight

,

TheStreet.com's

sister site. "Yet it might be that a fall in bond prices, rather than a rise in stock prices, is what restores equilibrium to their system."

Inverse Floater

The so-called Fed model compares the yield on the 10-year Treasury note to the earnings yield of the

S&P 500

. When the earnings yield, defined as the S&P's earnings divided by the price, is higher than the Treasury yield, stocks are considered undervalued. Conversely, when the earnings yield is lower, stocks are regarded as overvalued. While the Federal Reserve has never formally admitted to using the tool, a report to Congress in 1997 suggested that it was something the bank looked at.

Back in mid-March, some analysts were screaming at investors to buy stocks because the yield on the 10-year note was sitting at around 3.58% while the S&P's earnings yield was more than 6% based on 2003 earnings estimates. But in reality, the Fed model was implying one of two outcomes: Either stocks could move sharply higher or bond yields could move sharply higher.

Both eventually happened. Today, bond yields are sitting at 4.3% while the earnings yield on the S&P is around 5.2%. The S&P 500 has climbed 25% since mid-March and the yield on the 10-year note has jumped 121 basis points from its low in mid-June.

Ed Yardeni, chief investment strategist at Prudential Securities, said he believes stock and bond yields are heading toward convergence. But he noted that this "may require the bond yield to still go higher." Indeed, he believes interest rates could well move up to 5%.

Relative Worth

Jordan Kahn, a portfolio manager at Berger & Associates and contributor to

Street Insight

, said the Fed model reflects the relative valuations between asset classes, not just the valuation outlook for equities.

"You could look at the model and say the S&P is undervalued but it's also saying that bonds are overvalued. A spike up in rates is one way that could bring these models more in-line, it doesn't have to be surging stock prices," he said.

Of course, some analysts believe the Fed model is simply too flawed to be of any real use. One big complaint is that the earnings used to compute the S&P's yield aren't wholly reliable because they are based on forward-looking numbers.

"There are reasons not to trust the earnings numbers being used in some versions of the Fed model," said Haefele. "Operating earnings, otherwise known as pro forma earnings, are notoriously overstated relative to GAAP earnings."

In other words, it is hard to tell how undervalued or overvalued bonds and stocks are if the earnings yield is artificially high.

Jeff Bagley, portfolio manager for McCabe Capital Managers, said while there have been times when the Fed model has accurately predicted market trends there have also been times, particularly in recent years, when it has been a poor prognosticator.

In early 1999, for example, the Fed model was showing that stocks were too pricey relative to bonds -- or that bonds were too cheap compared to stocks -- yet the equity market continued to gain ground and bond yields marched higher.

Flawed Standard

"It's not a very good market timing tool at all, it's much better at alerting the investment manager to pay attention to these valuations and putting it on their radar," said Kahn. "Sometimes they can be overvalued (or undervalued) for months or years before they correct."

Kahn said he uses a modified version of the Fed model, using the five-year Treasury note, which he believes is closer to the length of time that investors typically hold stocks in their portfolios.

Still, Bagley questions the very logic underpinning the Fed model, which ultimately suggests that price-to-earnings ratios should rise when interest rates are low and vice versa.

He notes that if interest rates are falling because of very poor economic conditions, equities become riskier to own because earnings are often put in jeopardy. That's not a situation where P/Es should expand.

Likewise, if rates are rising because economic conditions are improving

as is believed to be the case today, P/Es shouldn't necessarily decline. Although P/Es generally move inversely with interest rates -- because stocks and bonds are competing asset classes -- Bagley said there are exceptions.

"I don't think the Fed model works too well as an absolute valuation purpose, it just helps to give you a general feeling about the direction," he said.