13 dezembro 2007

Preço sobre vendas

Esta reportagem da Forbes informa que o indicador de preço sobre vendas é melhor que o tradicional P/E

The Magic Metric; Forget P/E. James O'Shaughnessy says price/sales is the best way to find good, cheap stocks to buy.
Daniel Fisher
Forbes -10/12/2007 - 116 - Volume 180 Issue 12

What is the best measure of whether a stock is a bargain? The traditional starting point of stock research has been the price/earnings ratio. Money manager James O'Shaughnessy, who oversees $11 billion from an office in Stamford, Conn., says he has a better measure: Compare a company's market value with its revenue. This price/sales ratio should be your starting point in screening stocks.

It has worked for him. Over the past decade, he says, his small-cap growth accounts have averaged a 13.6% annual return after fees, seven points better than the market.

If you have $250,000, you can avail yourself of the talents of O'Shaughnessy Asset Management through Bear Stearns and other brokers for fees ranging from 2.5% to 1% annually, depending on asset size. If you don't, use this strategy in your stock picking. To be purely mechanical about it, find the 50 stocks in your database that are cheapest by P/S and buy them all. O'Shaughnessy leans to smaller companies, but his research suggests that the approach would work against a universe that includes big companies.

O'Shaughnessy, in fact, uses P/S only as a first cut; he does other kinds of analysis to make his final selections. But the validity of low P/S investing, he says, is borne out by a study he did of Compustat data, stretching from 1951 to 2003. To avoid contaminating his list with illiquid minnow-size issues, he chopped out everything with a market value below an inflation-adjusted $185 million. In this back test he reshuffled every year to keep the roster current.

Conclusion: His 50 low P/S collection beat the market by an annualized 16% to 13% over that half-century. He also tried out a hypothetical low price/earnings strategy. That one averaged only 14%. Buying cheap stocks as measured by price/book, another well-known metric, tied low P/S over the full period but wasn't as consistent. The P/S strategy beat the market in 88% of the ten-year periods, the P/B strategy in only 72%.

What all these strategies have in common is a value flavor. Meaning: You buy boring banks, utilities and steel companies, while passing up software, the Internet and biotech. And why does value work on Wall Street?

A low price/sales ratio might be just part of a package of indicators that a stock is relatively cheap and primed for higher returns, says Kenneth R. French, a professor at Dartmouth's Tuck School of Business. With his colleague Eugene Fama at the University of Chicago, French explained in 1996 that ratios such as price/earnings and price/sales are all proxies for the same attribute. French isn't sure just what that attribute is, but there is something about second-rate companies and industries that makes investors shy away.

Shy away more than they should, that is. A rational market will of course have Google's capitalization higher than General Motors'. The question is how much higher it should be. The low P/S theory says that the market tends to overprice the Googles relative to the GMs.

One explanation for why this might happen is psychological. Perhaps investors have vivid memories of Google and Amgen, fading memories of Interwoven and Enzo Biochem. So they overpay for growth and glamour. They also underestimate the tendency of businesses to regress to average profitability. A company with a fat profit margin (and thus, in all likelihood, a high P/S) will attract competition; a company with weak profitability will eventually replace its inept managers. The low P/S strategy capitalizes on this phenomenon.

To test P/S' market-beating abilities, we asked Wilshire Analytics researchers to replicate the O'Shaughnessy study for the past quarter-century, up through the end of September. Sure enough, over the past 25 years, the lowest-valued 50 (again minus very small issues) bested the S&P 500 handily, 19.3% annually versus 13.6%.

Some market savants used to adore P/S and have soured on it. Money manager and FORBES columnist Ken Fisher first popularized the price/sales ratio with his 1984 book, Super Stocks, but grew less enthusiastic by the early 1990s after further research showed the concept works best with tiny, illiquid issues--and when value stocks are rebounding anyway.

All that said, O'Shaughnessy, 47, is a guy worth listening to. He's well known as a proponent of the "dogs of the Dow" strategy, which focuses on the highest-yielding (and usually unglamorous) stocks of the Dow Jones industrial average. He's the author of several bestselling books, including What Works on Wall Street (1996) and Predicting the Markets of Tomorrow (2006).

O'Shaughnessy and other P/S partisans certainly don't feel this is the perfect metric, just the best of the lot. When he does his P/S screen, he doesn't want any stock with a P/S multiple of more than 1.5, which is the average for the S&P 500. Then he eyes this initial cut to see if the selections are indeed on the upswing. Earnings should be up over the past year and the stock price up over the past two quarters. When he gets done, he has a portfolio with a little more of a growth flavor than a pure low P/S portfolio (see table). Also, O'Shaughnessy leans toward smaller companies, like Geo Group and PriceSmart.Subscribe to Forbes and Save. Click Here. at 2.5 times revenue? Out. Google at 13 times revenue? Forget about it. Even Johnson & Johnson and Procter & Gamble don't qualify.

Declining revenue for a low P/S stock, of course, usually means that the stock is deservedly cheap because it stinks. Radio broadcaster Westwood One, at 0.3 of sales, is a case in point. It expects sales to be off in 2007. O'Shaughnessy doesn't want to own it.

What does this guy have against the classic P/E ratio? Two things. One is that the earnings figure is more easily manipulated than the sales one. The other is that companies on the rebound may have earnings that are temporarily depressed, distorting the P/E multiple. O'Shaughnessy likes, which has had its problems yet now is showing double-digit revenue growth. Its P/E is scary-looking at 39 (versus the market's 18). On P/S, however, it's cheap at 0.7.

And what's wrong with price/book? Fans of this ratio argue, as P/S fans argue about sales, that book value is hard to fake. The problem is that P/B works best for hard assets like factories and equipment, says Charles Mulford, accounting professor at Georgia Tech College of Management. Book value falls short for high-tech firms, whose assets tend to be intangibles like research and development. Ditto for consumer products companies such as Coca-Cola and their precious brands. As for the success of P/B in O'Shaughnessy's look-back study, it's quite possible that price/book would have worked well beginning in 1951, when assets were measured in smokestacks, but won't work in this century, when they are measured in lines of code.

One seeming weakness in the P/S calculation is that it ignores debt. Take two companies, each with $1 billion in sales, one that has a market value of $1 billion and no debt, the other that has $100 million of stock and $900 million of bonds outstanding. To an acquirer the price tag on either is $1 billion--the so-called enterprise value. To a fan of raw P/S ratios, however, the debt-laden company seems to be ten times as cheap as the other one. Why doesn't O'Shaughnessy use enterprise value instead of market value in his numerator? He tested it, and it did worse than the other three metrics.

For O'Shaughnessy, the key to getting good results is discipline, especially during manic stretches like the late 1990s, when people clutching low P/S stocks seemed to have "loser" stamped on their T shirts. In 1999, as the technology-dominated Nasdaq index soared 86%, the low P/S portfolio was up a measly 3.2%.

O'Shaughnessy's method can lead in some unpredictable directions. After the dot-com collapse in 2001, for example, his screens picked up a collection of steel companies such as Cleveland-Cliffs, Metals USA, Oregon Steel Mills and Quanex. "We scratched our heads," he recalls.

He bought them anyway, just in time to catch a Chinese-driven boom and a slew of acquisitions by foreign steelmakers like India's Mittal Steel. Many doubled in price between 2002 and 2005. He caught a similar wave when he bought Nortel Networks, a stock he'd avoided during the telecom bubble. He got it in January 2003 at $2, or 0.73 times sales; he sold it for a 76% profit 11 months later.

Nenhum comentário:

Postar um comentário