MW If rising rates were enough to end a bull market, we'd have entered a bear market long ago

Mark Hulbert

The 'Fed Model' has turned bearish - should you care?

At the outset of the global financial crisis of 2007-08, the 10-year Treasury yield was lower than where it had been three months prior.

Interest rates more often than not are falling when bull markets approach their final top, not rising.

Many Wall Street analysts apparently are unaware of this, however, as in recent weeks they have been asking how high interest rates must rise high before they kill the bull market. History teaches us that they are asking the wrong question.

Consider the 14 bull markets over the last 50 years in the calendar maintained by Ned Davis Research. For eight of them, the Treasury's 10-year yield was lower on the day of the top than where it stood three months prior.

Take the top of the bull market in October 2007, just prior to the global financial crisis, the worst bear market since the Great Depression. The Treasury's 10-year yield on the day of the top was 36 basis points lower than where it had stood three months previously.

To be sure, there have been other bull markets that topped out as interest rates were rising. But on average, across all bull-market tops, rates were only modestly higher on the day of the top than in prior weeks and months.

Today's 10-year yield (TY00) has increased more than the historical average, as you can see in the accompanying chart.

If rising interest rates were enough to precipitate a bear market, the bull market would most likely have ended months ago.

This doesn't mean that the stock market doesn't care about rising rates. It just means that the relationship between interest rates and the stock market is more complex.

Stocks' complex relationship with interest rates

To appreciate just how complex that relationship can be, consider the obstacles you confront when trying to construct an econometric model to predict the stock market's future return. Regardless of what other indicators you include in your model, you often find that including interest rates in your model not only does not increase its explanatory power but actually reduces it.

Take what happens to the predictive power of the price-to-earnings, or P/E, ratio when you adjust it according to prevailing interest rates. This adjustment is often referred to as the "Fed Model," a valuation indicator that got much attention on Wall Street several decades ago. It compares the stock market's earnings yield (the inverse of the P/E ratio) with the 10-year Treasury yield; the indicator is considered bullish when the earnings yield is higher than the 10-year yield, and bearish otherwise.

The table below compares the model's predictive power with that of the simple E/P ratio, focusing on a statistic known as the r-squared over the last 50 years. Notice that over all three forecast horizons the Fed Model reduces explanatory power.

 
                           R-squared using earnings-to-price ratio  R-squared using Fed Model 
Predicting 1-yr. returns   0.02%                                    -0.16% 
Predicting 5-yr. returns   2.67%                                    -0.06% 
Predicting 10-yr. returns  22.39%                                   4.19% 

Where the Fed Model currently stands

The Fed Model currently is bearish, given that the 10-year yield (currently 4.70%) is higher than the S&P 500's SPX earnings yield (3.78% based on trailing 12-month earnings per share, or 4.67% based on projected next-12-months EPS).

Fortunately, the indicator's track record suggests that its bearish turn will not be what brings today's incredible bull market to an end.

-Mark Hulbert