By Rob Bennett
The Wall Street Journal recently published an article raising an interesting concern - Robert Shiller's P/E10 valuation metric. It is titled Bull Market Has Long Memory of Bearish Times.
The article observes that today's P/E10 level is 31, a level almost as high as the level that applied just prior to the onset of the Great Depression. It notes that most commentary on the metric focuses on the numerator because it is changes in prices that concern the sorts of investors who consider valuation metrics when making decisions about stock investing. However, the denominator also of course affects the calculation. Richard Barley, author of the article, raises a concern that the denominator may be a bit off today, causing the metric to sound a not entirely justified warning bell.
He writes that the strength of the P/E10 metric is that, because it considers 10 years of earnings rather than a single year, it generally offers a better indication of the extent of overvaluation in the market at a given point in time. However, at this particular moment in time, the 10 years of earnings determining the P/E10 value includes the years of the global financial crisis. "Outliers can skew a calculation of the mean," according to Barley.
He suggests using the median of 10-year earnings rather than the mean. This would lower the P/E10 value to 28. The article explains that: "In the next couple of years, that big hit to earnings will drop out of the calculation; all else being equal, that would lower the Shiller PE." The suggestion is that investors who believe that valuations matter should not feel as much alarm as the P/E10 reading of 31 might otherwise provoke in them.
I think the point being made is legitimate. However, I agree with a point made in one of the comments to the article that switching to use of the median of earnings rather than the mean because the number generated by the conventional calculation is too high for our liking risks "playing with figures to make them justify our actions." The benefit of turning to the historical return data to guide investment decisions is that data is objective. When we give in to temptations to adjust the numbers that don't support our intuitive preferences, we give up that benefit.
More importantly, I don't think that Shiller would say that the P/E10 metric offers precise guidance. Using the median for earnings rather than the mean pulls the number from the very, very scary high of 31 down to just the very scary high of 28. That's not much of an improvement. The 32 reading is especially alarming because it is so close to the reading that applied in the days just prior to the Great Depression. But the full reality is that there has never been a time when the P/E10 rose to 25 when we did not suffer an economic crisis in the days that followed. The mean-adjusted reading of 28 is plenty dangerous itself.
The biggest obstacle to effective stock investing is that the act of going with a high stock allocation causes a bias in favor of stocks. Once our stock allocations go above 50 percent, it becomes almost impossible for us to think clearly about the dangers associated with stock investing. We rationalize away all warning signs. The smartest among us are the most effective at the rationalization process.
The P/E10 reading of 31 does not provide a perfect indication of the level of risk present in stock investing today. But it is probably the best tool available to us for obtaining a general sense of where things stand. The market is dangerous. That's the message delivered by a P/E10 value of 31 and that's the message delivered by an adjusted P/E10 value of 28 too.
We may not see a price crash next week or next month or next year. No one can say. All that we can say is that the risk of a crash is much greater than it would be if prices were at fair-value levels. We all should take that reality into consideration when deciding how to invest the money we hope will remain available to us to finance our old-age retirements.