One part of this is easy.
PE is industry jargon. It’s a shorthand way of expressing the value of an individual stock, an industry group or a whole stock market, in terms of how many times one year’s earnings we’d be willing to pay to own whichever it is.
On the face of it, a low PE, say, 4x, would seem to be good; a high PE, say, 50x, bad.
But how do we know? What factors enter into determining a PE?
A point that I’m maybe too fond of making is that, strictly speaking, there’s no demand for stocks. There is demand for liquid investments, though (for most people in the US, it’s so they can save to send their children to college and to retire). Bonds and cash are the other two big categories of liquid investments. The apparent hair-splitting distinction is important, however, because each fixed income markets is much larger in size than stocks. They’re also less risky. The potential returns on these alternatives have a deep influence on what people are willing to pay for stocks. In fact, academics turn the PE upside down (1/PE) to get what they call the earnings yield on stocks. If you make the assumption that $1 in earnings in the hands of company management is more or less the equivalent as $1 in interest paid to you by the US Treasury, then the yield on Treasury bonds should (and virtually always does) have a powerful influence on what the earnings yield, and PE of stocks should be. Why pay 20x for stocks if bonds are yielding 10%?
As I’m writing this, the 10-year Treasury bond is yielding 2.68%. That’s the equivalent of a PE of 37. This compares with a PE of 26 on the S&P 500, based on current earnings. So either stocks are cheap or bonds are overpriced.
Today’s situation is very unusual, given that the financial meltdown in 2007-08 compelled the Federal Reserve to push interest rates down to intensive-care lows. The consensus judgment of financial professionals, which I think is correct, is that bonds are unusually expensive today, not that stocks are dirt cheap. If the 10-year is on the way to a 3.5% yield as the Fed returns rates to normal over the next year or two, then the equivalent PE on the S&P would be 28.5x.
That’s about where US stocks are now priced vs. bonds, which suggests that stocks are fully valued if we factor in the likely course of the Fed. This suggests that only new positive information will move the overall market higher.
Now the going gets harder.
The second important point is the the stock market is a futures market. That is, it is always pricing in tomorrow’s prospects as well as current earnings. Willingness to pay for future earnings ebbs and flows with the business cycle, however. During recessions, investors play their cards very close to the chest and look at most a few months into the future when pricing stocks. In normal times, the market begins to price in the following year’s earnings in June or July. In a very buoyant market, investors may pay for earnings two or three years into the future.
A third consideration, related to the second, and applying to individual stocks, is the rate of earnings growth. The importance of this factor changes from time to time. But a useful general rule is that the PE based on this year’s earnings can reach as high as the long-term growth rate of the company. In other words, if earnings per share are growing at a 50% annual clip–and will likely continue to do so for the next several years (or at least there’s no easily visible bar to growth like this)–then the PE can be as high as 50x.
Generally speaking, the US economy can probably grow at about 4%-5% a year in nominal terms (meaning, including inflation). If so, publicly traded companies, which are arguably the cream of the crop, will grow earnings per share by about 8% – 10% annually. All other things being equal, this latter figure should be the trend growth for stocks in general.Put a different way, a company that can sustain growth of 50% a year in an economic environment like this must have something extra special going for it.
This rule of thumb doesn’t work for many “value” stocks, since no growth/earnings declines would imply a zero multiple–which in most (academics would say “all”) cases is clearly wrong (Academics say every stock retains at least a kind of option value).