I'm firmly convinced that an unhealthy fraction of the financial innovations that experts share with audiences exist mostly to give them something to talk about (there isn't always something to talk about or something to recommend. When there is, I'll say so. When there isn't, I'll complain about people who say otherwise). One of the worst examples of this is the popularity of the PEG ratio: Price/Earnings to projected earnings growth.
It's rarely as hard to spot stupidity as it is to explain why something is so stupid -- logic just isn't at home when it's surrounded by pure unreason. The best I can do is demonstrate that PEG ratios give rise to all kinds of nonsense. They start from the generally intelligent proposition that it makes sense to pay a premium for high growth -- the question of how much of a premium is probably the central question that financial markets exist to answer (so the claim that you can get that answer without knowing anything more complicated than long division is a little suspicious). But let's give it the benefit of the doubt: Let's see what happens when you make decisions based on the PEG ratio.
I used my TI-83 graphing calculator (costs $100; in terms of practical problem-solving potential, is probably more valuable than an MBA) to examine what kind of returns you'd get from buying stocks with a PEG of 1, if their growth rates varied from 0% to 100%, and if, after five years, they all traded at fifteen times earnings. The last assumption might sound peculiar, but it's vaguely the basis for discounted cash-flow analysis, which assumes that a stock will revert to the mean after a few years, and asks what it's worth if it has some extraordinary characteristics before then -- the fact that DCF basically says that a steel mill is a steel mill and Microsoft is Microsoft and, five years later, they're identical, is bizarre but happens to work pretty well.
The results are pretty interesting: If you bought a stock growing at 1% per year for 1X earnings, you'd make a healthy 70% annual return over five years. If you bought a 10% grower at 10 times earnings, you'd make about 20%. Returns bottom out at around 13% per year for growth rates in the 20-30% range, and start to rise comfortably thereafter (a 100% grower at 100 times earnings returns a pretty spectacular 37% per year).
What do these results tell us? First, if you can predict earnings five years out and can buy a stock with a PEG ratio of one, you'll do pretty well -- which is about as useful as pointing out that you can make a lot of money betting on horse races if you know which one will win. But it also shows that PEG ratios just don't predict much about performance. The same PEG ratio can mean wildly different things if it's attached to a mediocre company or a brilliant startup -- and it's an utterly useless number to analyze shrinking companies. Surely the value of a stream of profits that gracefully shrinks a few percent each year is a positive number. Really, all a PEG ratio can tell you is that if you have two stocks with identical growth, you should buy the cheaper one -- and if you have two identically cheap stocks, you'd probably prefer the fast grower. In short, all it does is misleadingly codify some common sense.
There are a few pretty good reasons for the PEG to persist in being popular. One is that it's easy to learn; one number you can get the newspaper and one number you can get from an analyst report, plus a pocket calculator, and you're good to go. Another reason is that it provides a (false) basis for comparing wildly different companies -- don't know whether to buy shares of that cheap, stolid real estate trust or a frenetically expanding software company? Check the PEG (and end up either wrong or lucky). But the best reason is that it provides The Number and The Answer, and even though the number is meaningless and the answer is probably wrong, it provides a pretty comfortable illusion of certainty. But markets aren't that simple: You can be certain, you can be right, but you can never count on being both.