Is P/E Still A Very Useful Metric?

by: Braun Research

Through the kindness of a business associate, I was made aware and subscribed some time ago to a service called "Chart of the Day." Over time, I found it to be sending some very interesting long-term charts like the one I am reproducing below:

For full disclosure, I only subscribe to the free part of the service and the reason I mention it here is out of gratitude for ideas that the charts generate for me every now and then, like the comments I am making here about the usefulness of the P/E ratio in the practice of investing. And of course it is also a matter of respect for the ownership of their intellectual property.

Chart of the Day comments on this graph as follows:

Generally speaking, when the PE ratio is high, stocks are considered to be expensive. When the PE ratio is low, stocks are considered to be inexpensive. From 1900 into the mid-1990s, the PE ratio tended to peak in the low to mid-20s (red line) and trough somewhere around seven (green line). The price investors were willing to pay for a dollar of earnings increased during the dot-com boom (late 1990s), surged even higher during the dot-com bust (early 2000s), and spiked to extraordinary levels during the financial crisis (late 2000s). As a result of the continued surge in corporate earnings the PE ratio remains at a level not often seen since 1990 despite what has been a significant upward trend in stock prices so far this calendar year.

Fair enough… as far as it goes. The P/E is a well-worn "rough-and-dirty" measure for how expensive or cheap a stock may be and you will find it mentioned quite extensively by professional analysts, particularly on the sell side, as one of the key metrics they use to make buy or sell recommendations. The question to be asked is whether by any chance, P/E may not be too "rough" or too "dirty" for practical purposes in the current period. The chart above would indicate so, unless of course you are of the persuasion that the recent volatility of earnings that we experienced since the start of the century is a passing thing and that we are now back to the "normal" of the 20th century.

I happen to believe that we are in for some more rough sledding, albeit that does not mean at all that I suggest capitulating. It only gives me the impulse to look for more usable ways to judge the value of stocks and to evaluate the total environment for early enough signals indicating that any stock-by-stock careful valuation may be thrown to the winds by the general state of the economy. I expressed before some ideas in this direction (see here) but let me offer another opinion on why P/E may be a little too rough a measure for my taste.

To begin with, anyone who has worked with P/Es and tried to aggregate them, knows that if a company makes losses, the P/E is a negative number and becomes thus meaningless, impossible to aggregate for a "market P/E." No problem, say the analysts: we can use E/P and then aggregation becomes possible again. How can you not accept this?

But wait a moment. Take a company with problems, facing bankruptcy, whose earnings may not have turned negative (yet) but have rather become minuscule, say they went from $2/share to a cent or two. The price will certainly decline but not necessarily to the same extent, say from $20 to $5. The P/E will go from 10 to 500 or 250. Is this company comparable with Cisco whose P/E went to something like 200 during the craze? Cisco was definitely not in bankruptcy. Sure, some more analysis is required for such a company, same as when using any single metric anyway.

These two situations show that judging any company's value by its P/E assumes a continuity which does not exist. The example I chose is of course extreme and would raise a red flag to any thinking person. But at what P/E level does such a discontinuity start to matter? Where does the P/E become a misleading measure? Hard to say. One thing is clear though: high P/E firms and low P/E firms are not on a continuum: they are not comparable as they exist in completely different worlds. And this is similarly true for any measurement of "exposure" as Barra liked to call such ratios.

So what is a portfolio manager to do? One solution is to divide stocks into groups by some other parameter, like for example placing them into "bins" defined by the extent to which a large sales or earning surprise is likely to occur. If a negative surprise is unlikely to happen, i.e. a downgrade by the sell-side analysts is not probable within the next few weeks or months, that stock is a candidate for further analysis to be bought for a portfolio. Finding a good number of such stocks, enough to be left with a decent size portfolio, doing it quickly and frequently, with a high degree of confidence (say 70%-80% as 100% does not exist in this business), is not a simple matter; but it can be done and those who did it, are being rewarded with above-average results.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.