Technical Correction “The Numbers Guy” And Wall Street
Technical Correction “The Numbers Guy” And Wall Street
Hal Plotkin, Special to SF Gate
Tuesday, November 21, 2000
URL: http://www.sfgate.com/cgi-bin/article.cgi?file=/technology/archive/2000/11/21/peratio.dtl
The stock market’s resemblance to a casino is becoming increasingly hard to ignore. It’s common these days to see the stocks of respected, big-name firms such as Intel, Oracle and Apple Computer rise or fall 20 percent or more within a single week, sometimes even faster.
The turbulence can be mystifying. If you’re like millions of frustrated tech investors, you might be wondering why individual stocks in your portfolio often go down even when the companies themselves are enjoying unprecedented success.
Take Cisco Systems Inc., for example.
Earlier this month, Cisco announced it rang up $6.52 billion in sales last quarter, representing a sizzling 66 percent increase from the $3.92 billion it took in during the same three months last year. Profits for the quarter hit $1.36 billion, almost all of which can be used for more research, product development and strategic acquisitions. Sounds like the company is on a roll, no?
Believe it or not, Cisco’s stock sank on the news.
The culprit: Wall Street’s infatuation with misleading and unreliable stock valuation formulas. The formulas are so bad they often make it hard to separate good investments from bad ones, particularly in the fast-moving tech sector.
What’s worse, they can lure investors into a false sense of security that also causes them to take their eyes off company-specific events and larger trends that are usually far more important.
So what’s causing all these often-illogical gyrations in tech stock prices?
You can put a lot of the blame on the comparisons of price/earnings, or P/E ratios, that you hear so much about.
Individual investors are often told that P/E ratios should be one of a number of factors they consider when they do their investment planning. P/E ratio comparisons are even more important to mutual fund managers, who very often use those numbers as a way to decide where to park investor cash.
There is, however, one little problem.
P/E ratios don’t really predict what they claim to predict, which is how a stock will do over time. In fact, they don’t just fail at this; quite often, they fail in ways that impact the very stocks whose performance they’re trying to measure.
It’s the financial equivalent of the Heisenberg Uncertainty Principle, which holds that certain measurements change the things being measured.
The flaw in P/E ratio comparisons is very evident if you understand how they work.
They’ve long reminded me of the “numbers guy” who used to call Larry King’s all-night radio show back before King hit the big time at CNN.
He’d call in, for example, claiming to be able to predict, oh, say, the closing numbers for the next day’s Dow Jones Industrial average. He’d rapidly rattle off his remarkably incomprehensible calculations, explaining, for example, how you divide 7, since that represents the days of the week, by the number of times the Earth passes Jupiter in four years, by some other number, and so on, until you finally came up with an answer.
When his predictions didn’t pan out, as of course they rarely did, he’d call back with a similarly entertaining mathematically based excuse. Working backward, of course, he could always show how missing some rather minor computation had tripped him up.
P/E ratios rely on the same basic premise.
The problem has to do with two things the P/E statisticians call “comparables” and “multiples.”
But first, lets start by describing what a P/E is.
P/E ratios are the ratio of a company’s share price to its per-share earnings. A P/E ratio of 10, for example, means that the company had $1 of annual, per-share earnings for every $10 in share price. (You get the per-share earnings by dividing the company’s total earnings by the number of shares outstanding).
So a P/E ratio is a relative measure of how much a company is earning in relation to the number of shares it has sold to investors.
Now here is where this gets a little tricky.
The big Wall Street security firms then calculate their decisions regarding what an individual stock should be worth (and hence their recommendations) by multiplying a company’s earnings by some other often arbitrarily picked number.
They might say, for example, that a stock with earnings of 10 is worthy of a multiple of 15, which means a fair price for the stock should be somewhere around $150 (10 times 15).
But how do securities analysts arrive at the critical “multiple” number?
They look at supposedly “comparable” stocks, firms that might, for example, be in a similar line of business. Then, just like Larry King’s “numbers guy,” they figure backward based on the most recent closing prices for those supposedly comparable stocks. If Nortel’s stock had a multiple of 27 yesterday, and Cisco is in the same business, they reason, then it should have the same multiple today.
But is Cisco’s business really just exactly like Nortel’s? And can you really compare Apple Computer, which has its own very particular set of business conditions, with Dell Computer, which in my experience often appeals to an entirely different group of consumers?
It’s sort of like looking at a foot race and trying to determine who is winning by averaging everyone’s time together.
The big problem with using comparisons of P/E’s to set target values for stocks, of course, is that it is entirely self-referential. If the stock of one company in a set of “comparables” goes down, it puts downward pressure on the entire group’s multiple, which soon takes the other stocks in the group down, too.
That’s how you end up with Cisco’s stock going into the tank even though the company is generating record amounts of revenue and profits. The professional investors who make stocks go up and down have no doubt about the strength of Cisco’s business. But they are worried that its P/E may be out of line when compared with other investment opportunities.
Conversely, if one company is doing particularly well within a certain group, all the other companies in that group will often get a boost in their stock prices, too, whether deserved or not. (And, as we found out when the Internet bubble burst, very often, it isn’t.)
That’s how P/E-driven trading contributes to the volatility we’ve seen lately that pushes stocks way up one week and then way down the next. When stock prices are tethered together in some arbitrary formula, the formula becomes more important than what’s actually happening to any of those individual businesses.
So fasten your seatbelts.
It’s possible someone will come up with a better way to measure the relative merits of different investments someday.
But until then, the wild ride is likely to continue.
