Perhaps the most common question in valuing a company is: "What's the P/E ratio?" Different time horizons for earnings will result in different answers. A trailing earnings value is different from the current fiscal year estimate, which, in turn, differs from the fiscal year estimate for the following year.
Using trailing earnings to calculate the price-to-earnings multiple is rarely a good idea. Because stocks are priced on future earnings rather than past earnings, the trailing P/E multiple will be useful only if trailing earnings are similar to estimated earnings.
A key question remains: How far forward should investors look in estimating earnings to arrive at the most useful P/E multiple for valuing companies and picking stocks? The answer is simple -- as well as troubling --
as far forward as possible
I'm not kidding. That's the correct time horizon.
This presents a dilemma for quantitative analysts, because a single time horizon is not appropriate for all stocks. Visibility in earnings differs among industry groups. A pharmaceutical company is most accurately priced on the earnings potential for its drug pipeline. A defense contractor is best priced on the value of its long-term contracts. A bank is best priced on 12-month forward estimates, adjusted for amortization costs.
One important point is that it never pays to cap your time horizon, saying, for instance, "we look two years forward and no further." As visibility for future earnings clears, the company's valuation on near-term earnings will become irrelevant. As
has pointed out, for example, Y2K stocks should be priced on their ability to sustain earnings after 2000, not on their windfall profits over the next two years.
As an example of how price-to-earnings ratios are related to earnings visibility, the table below compares the share-weighted P/E multiple for companies ranked according to historical stability in earnings growth. The companies with the most stable earnings growth are in decile 1, the companies with the most unstable historical earnings records are in decile 10. As you can see, the highest P/E multiples are accorded to the most stable growers, companies with the best earnings visibility.
Stability of Earnings vs. P/E Multiples
Russell 1000 stocks, Aggregated by Decile
The high price-to-earnings multiple for stocks in decile 10 is due to the number of entertainment and oil exploration companies in that decile, industries typically priced on cash flow rather than earnings.
The table above uses 12-month forward estimated earnings in calculating P/E multiples. The stability of historical earnings growth is calculated by comparing the growth and standard deviation of growth in 12-month forward historical earnings estimates over the last 10 years. The ratio is then broken into deciles to create the stable growth ranking.
In screening for companies, my default time horizon is 12 months forward. Previous tests I've done looking for excess return on low P/Es using time horizons from 12 months trailing to an estimate five years in the future peaked in a range of nine to 24 months forward. In following up on a quantitative screen with fundamental analysis, however, I keep the ideal time horizon firmly in mind: try to look as far forward as possible, because the market certainly will.
Ted Murphy ( firstname.lastname@example.org) operates the MarketPlayer Web site. Prior to MarketPlayer, he was a partner at
Equinox Capital Management.