Portfolio managers interviewed for the story generally like the idea of a p/e ratio using 10-year trailing earnings. But they pointed out that using just that measure to value stocks could cause them to inaccurately judge a stock of a newly public company that sees strong growth in its early years, and then tapers off or falls once it loses momentum.
"There's some good thought in that 10-year trailing earnings would reflect smoother earnings over the trailing four quarters," said Fred Dopfel, a senior strategist at Barclays Global Investors, San Francisco. "One possible bias in that valuation is that companies that have a very rapid growth rate would show an exaggerated p/e ratio."
Kevin Divney, managing director and co-chief investment officer of midcap growth strategies at Putnam Investments, Boston, said that in practice, investors would be wise to use a blend of valuation methods.
"I like the idea of using 10-year earnings," he said. "We're always trying to extend the horizon of our investment views, but you always have to balance that out with the volatility of short-term returns."
"The question is, how much recent history do you incorporate? I think long-term valuation at the aggregate level wins, but if you do this on a company-by-company basis, it might not work for every company. You might have a case where a company is in an industry in transition, and its assets are higher today because of practices such as ‘just-in-time' inventory and sales over the Internet. The bottom line is, it's hard to determine what's normal."
"We would blend (looking at) 10-year trailing p/es with forward-looking p/es and other methods."
Richard Pzena, president and chief investment officer at Pzena Asset Management, New York, said, "I can understand why looking at 10-year earnings would serve to smooth out non-recurring events. But the biggest drawback is when you are evaluating growth stocks. Since growth stocks of established companies produce steady growth over a long period of time, this method would not give you a very good valuation."
Mr. Asness acknowledged that while there are potential drawbacks to examining 10-year trailing earnings, it still might be the best measure. The portfolio managers interviewed for this story "are all 100% right," Mr. Asness said in an interview. "You'd be hard pressed to find any individual stock whose situation has been stable enough for 10 years to make the 10-year number meaningful."
"(In the paper,) we're just looking for some version of steady state trend earnings. While the last 10 years might've been better or worse than normal, the swings are going to be far less dangerously misleading for decades of the market as a whole than for most or all individual stocks," he said.
"Note, we are not bothered by the fact that 10-year trailing earnings will, on average, understate current earnings (as earnings do grow), as we're only comparing the ratio to itself through time, and not to other methods. In summary, I wouldn't argue with them at all, but I would stress this technique is really just useful for the market as a whole."