Can We Actually Predict Earnings?
Ed Easterling of Crestmont Reseach offers us the following very important chart. It shows reported earnings compared to the historical trend line. Earnings, especially when seen from a valuation standpoint, are mean reverting. They will fluctuate around the long-term trend line. And interestingly, that long-term trend line is nominal GDP. (Nominal GDP includes the effects of inflation.)
Click to enlarge.
Total corporate earnings for any particular large country and stock market by definition cannot grow faster than nominal GDP (though individual stocks can). And since the S&P 500 is largely reflective of the US corporate world, earnings for the S&P 500 index will fluctuate around nominal GDP.
Notice how smooth that growth line for nominal GDP is? That will be important in a few paragraphs. But first, let's look at how well Easterling's historical trend line (which is nominal GDP) compares with Robert Shiller's 10-year smoothed earnings. Rather than use the earnings from any single year, he smoothes them by using a 10-year average.
Important: Notice how closely correlated the earnings for Crestmont's nominal GDP and Shiller's smoothed earnings are.
Click to enlarge.
Now, this is where it gets interesting. Shiller's data is not predictive. But remember how smooth the earnings trend line from Crestmont was? Ed contends that there is a predictive element when we use nominal GDP. In other words, at some point in the future, earnings will grow back to and then exceed the long-term trend in nominal GDP.
So, while we’re in the process of dropping below the mean or below the long-term trend line of earnings in terms of nominal GDP, we can be confident that at some point in the future those earnings will again revert above the mean. It seems to have been part of the economic laws since the time of the Medes and Persians.
This has important implications for future values. Let's look at the next graph, from Vitaliy Katsenelson. Vitaliy uses a 6% growth of earnings as his baseline (which is, not coincidentally, very close to the long-term rise in nominal GDP). Again, notice how earnings fluctuate around the mean.
Notice also the small box on the right that shows where earnings could actually fall if they drop by the same percentage as they did in the 2000-02 recession. That would suggest that earnings will drop below $40 from the currently projected $48. Remember, last year projections for 2008 were $82.
Click to enlarge.
Email
Print
Del.icio.us
Digg
Facebook
Subscribe
