The Federal Reserve publishes a well-known valuation model that signals whether to buy or sell stocks based on bond yields, comparing the earnings yield of stocks to long-term U.S. government bonds.

Advocates of the Fed's valuation model believe that when the S&P 500's earnings yield (measured as earnings per share, or EPS, divided by the S&P 500's price) is higher than the 10-year Treasury yield, it's good for stocks, and vice versa.

Here’s what’s happening now: The S&P 500’s earnings yield, based on earnings per share over the past 12 months, is 3.90%, while the 10-year Treasury yield is 4.46%, more than half a percentage point higher.

As shown in the chart below, before the Fed's valuation model recently fell into negative territory, the last time it recorded a significantly negative value was during the 2008-09 financial crisis, suggesting that U.S. stocks may once again be extremely overvalued.

While the similarities are ominous, regular MarketWatch contributor Mark Hulbert believes that the Fed’s valuation model shouldn’t be taken too seriously: Its long record is poor.

To demonstrate this, Hulbert analyzed the Fed valuation model since 1871, using data from Robert Shiller of Yale University. He compared the earnings yield of the S&P 500 index with the performance of the Fed valuation model to predict the stock market's inflation-adjusted total return over subsequent one-, five-, and ten-year periods.

Below is his findings. For each indicator and time period, the table lists r-squared, which measures the ability of one data series (in this case, earnings returns or the Fed valuation model) to explain or predict another data series (in this case, the stock market). In each case, earnings returns alone have stronger predictive power than the Fed valuation model.

Many people were surprised to find that the Fed valuation model did not perform better than the earnings yield. Isn’t it obvious that stocks should be a better choice when interest rates are relatively low rather than relatively high? Why doesn’t history prove this out?

The answer is that the Fed valuation model is really comparing apples to oranges. The stock market's earnings yield is a real yield because historically corporate earnings have grown faster when inflation is higher. However, the 10-year Treasury yield is a nominal yield that does not fluctuate with inflation. So it's not surprising that the Fed valuation model, which draws its conclusions by comparing real and nominal yields, doesn't tell us much.

Cliff Asness, founder of AQR Capital Management, published a paper twenty years ago titled (Anti-Fed Model), which is probably the most authoritative theoretical and empirical paper against the Fed model. He concluded in the paper:

“The Fed’s valuation model appears to make common sense, but it is not true. The appeal of this common sense has convinced many Wall Street strategists and media pundits, but this common sense is mostly misleading, probably because it confuses real returns with nominal returns (i.e., the money illusion).”

This conclusion does not mean that the stock market is not overvalued, and investors may have many other reasons to be concerned about the outlook for stocks in the coming months and years. But this article emphasizes that the current state of the Fed's valuation model is not a valid reason for increased concern.

Article forwarded from: Jinshi Data