If the economy starts to rebound, investors can expect technology stocks and cyclical issues -- paper and steel makers, and chemical and energy producers, for example -- to lead the stock market higher.

At least that's the conventional wisdom. And certainly a bet on these sectors would have paid off as the market climbed from its Sept. 21 lows. The technology-dominated

Nasdaq

climbed 37% from that date through Dec. 31, and the Morgan Stanley Cyclical Index climbed 27% during the same period.

But that rally is now behind us, and stocks are a lot more expensive than they were in September. And recent earnings news from the likes of

Intel

(INTC) - Get Report

and

Juniper Networks

(JNPR) - Get Report

in technology and

Alcoa

(AA) - Get Report

and

Dow Chemical

(DOW) - Get Report

in the cyclical group certainly raise the possibility that any economic recovery will be weaker and later than investors had expected in December.

So what should investors do with the conventional wisdom this time?

Should we file it in the trash with other recent investment truths, such as "Don't fight the

Fed," which got me in trouble last year when I believed that interest rate cuts were a buy signal?

Or should we look past the current weakness -- maybe it is nothing but the usual correction after such a heated rally -- and use it as an opportunity to buy the next set of market leaders?

To decide, I think you've got to understand operating leverage, the reason why technology and cyclical stocks usually lead the market upward after a recession. That will give you a relatively simple three-part system for judging the potential risk and reward of any potential market leader in the next phase of this market.

In this column I'm going to show you how to apply this system to technology stocks, and give you three stocks that I'd look to buy if the technology sector should correct some more over the next few weeks. In my next column, I'll do the same for cyclical issues.

Operating leverage works like this: Companies with lots of money tied up in machinery, plants, real estate and distribution networks can't easily cut expenses when the economy turns sour. Profits don't just fall, they plummet. But when the economy turns, the fixed assets that were a drag on profits start coining money again. Earnings don't just move up, they soar. (Please note that this kind of leverage has absolutely nothing to do with financial leverage -- most technology companies, for example, have very low financial leverage because they carry no debt.)

The Case of Cymer

Let me apply that rather abstract description to a highly leveraged technology company such as

Cymer

(CYMI)

.

Cymer makes expensive lasers used by semiconductor chipmakers to provide the focused light source needed to draw the current generation of tiny circuits onto chips. That hasn't been a great business lately. Cymer estimates that the total market for the kinds of lasers it produces came to 517 units in 2000. In 2001, that total could drop to 427, according to Cymer, or as low as 363, according to Morgan Stanley Dean Witter. Revenue at Cymer, which was $98 million in the September quarter of 2000, fell to $53 million in the September quarter of 2001.

And as bad as a 50% decline in revenue is, remember that Cymer is a high-fixed-cost company unable to cut expenses rapidly as the market for its product contracts. Analysts project that the company will show a loss of 9 cents a share for the December 2001 quarter, when it reports on Feb. 5. That's a 113% drop from the 68 cents a share the company earned in the December quarter of 2000.

That's the bad news. When the economy contracts, Cymer gets hurt. Very hurt.

But when the cycle turns, Cymer will mint money. Analysts estimate that Cymer owns about 85% to 90% of the market for the current and next generation of laser light sources. In 2003, when Cymer estimates that the market for laser light sources will climb to 823 units, that market dominance will translate into sales of about 525 units for Cymer. That's way ahead of the 400 or so units the company sold in 2000 and slightly ahead of the company's estimate of the entire market: 517 units in 2000.

Not only will Cymer sell more units when the market rebounds, it will sell more units at a higher price, thanks to the very small number of companies capable of producing these machines. Morgan Stanley estimates that the average selling price at Cymer will go from $500,000 in 2000 to $750,000 by 2003 and then to $900,000 by 2005.

And earnings? How about $4.80 a share in 2005? That's Morgan Stanley's estimate after looking at the 20% net margin that Cymer has earned at the peak of its cycle in the past. Not bad for a company that's forecast to earn just 11 cents a share in 2002. Using the historical 30 times multiple on projected future earnings that Cymer has commanded in the market, Morgan Stanley notes that Cymer could easily be a $144 stock at the peak in the recovery cycle. That works out to a gain of about 300% from the stock's recent price of $34.51.

Why Not Buy Cymer Now?

OK, so why aren't I telling you to rush right out and buy this one?

The short answer is that all these numbers are only projections, and the highly leveraged turnaround in Cymer's earnings is only a potential turnaround. And lots can go wrong with anything that's merely potential.

Just last week, for example, Intel cut its capital spending budget for 2002 by 25%. That wasn't good news for the companies that make the equipment that Intel buys because Intel alone accounts for about 20% of the semiconductor industry's capital budget. Other chipmakers are expected to follow Intel's lead. On Jan. 18, for example, Wall Street was awash in rumors that

Taiwan Semiconductor Manufacturing

(TSM) - Get Report

, the largest contract maker of chips for other companies, would also cut its capital spending plans for 2002. All that could push any return to profitability at Cymer out past the middle of 2002, the date now assumed by many analysts.

In addition, nothing says that the recovery for chip equipment makers such as Cymer has to be as robust as previous recoveries. Credit Suisse First Boston, for example, believes that semiconductor capital equipment stocks could well lag behind the rest of the technology sector in 2002. Investors have already bid the stocks up to unrealistic levels because of a mistaken belief that the peak of the next earnings cycle will resemble the peak in 2000. No so. Many of the traditional leaders in the sector won't see a return to margins nearly as high as in 2000 because the economic recovery this time is likely to be relatively weak.

That leaves investors between a rock and a hard place. No one wants to pass up the chance to earn 300% on Cymer over the next two to three years. On the other hand, no investor wants to sit in dead money for a year or two, waiting for a recovery that's always just around the corner. And most of all, no one wants to buy Cymer at 314 times projected 2002 earnings, and then watch the stock fall when the market decides that the recovery in semiconductor equipment stocks is further away than anticipated.

I don't think any investor is about to find certainty in a situation like this. If it were certain that Cymer would earn $4.80 in 2005, the stock would trade at a lot higher price than $34.51 right now.

TheStreet Recommends

Handicapping the Field

But investors can understand the factors that produce that uncertainty, and that might reduce or increase risk and reward. And investors can then handicap the likelihood that those factors will play out for or against the stock. The best investment candidate is the stock that shows the lowest odds that potential bad things will happen and the highest odds that potential good things will happen.

A three-factor model can capture most of the uncertainty that investors should consider.

1. Time:

How likely is it that the recovery that you're counting on will happen sooner as opposed to later? This factor has a powerful effect on an investor's potential return from an investment. If you assume a best-case scenario with Cymer, for example, and calculate that a recovery will be clearly apparent by 2004 -- enough so that the stock's price anticipates the peak earnings of $4.80 and peak forward multiple by early in that year -- then an investor is looking at a gain of 300% in two years. Push out a full recovery to, say, 2007, and the gain is spread out over five years. That would be an annual gain of 33% -- certainly nothing to sneer at, but well short of the return pulled down over the shorter two-year period.

2. Leverage:

What could go wrong or right and change the leverage that an investor is counting on? Technology companies lose their technology edge, and one of Cymer's competitors could eat into its market share, reducing the upside to the stock. Global supply or demand shifts could reduce or even eliminate any increases in average selling price. Government intervention could delay or even prevent the elimination of weak competitors at the bottom of the economic cycle that is needed before survivors can get higher prices for their products. And on a macroeconomic level, any economic recovery can be more or less robust than recoveries in past cycles.

3. Valuation:

There is no doubt that stock markets as a whole and the prices of individual stocks anticipate future good news. The more prices anticipate future potential, however, the less room there is for prices to rise as that future approaches, and the more possibility that prices will fall if that future should prove even mildly disappointing. High valuations can be especially problematic if the rosy future is separated from the present by a period of declining revenues or profits. The reality of the near-term decline will be a persistent challenge to any increase in stock price based on a rosy future that follows the tough times.

I have some reservations about how well Cymer scores against the first and the third of these three yardsticks. The recent cut in capital spending by Intel and the likelihood that other chipmakers will follow suit does increase the odds that the recovery for semiconductor equipment makers is further out than investors had expected when they bid up the sector at the end of 2001. Cymer was up by 60% from Sept. 21 through Jan. 15 on those expectations.

Wall Street and the company itself show revenue and earnings continuing to decline in the June quarter, and I think investors are facing heavy odds in favor of a near-term disappointment in the stock. The long-term potential is real, I believe, but the near term is risky.

Three Potential Winners

I think you can find technology stocks that avoid that problem if you look for issues in sectors that are showing reasonable signs of recovering demand instead of signs of continued weakness. For example, consumer spending on technology continues strong.

Microsoft

(MSFT) - Get Report

reported stronger-than-expected sales for its Windows XP software and Xbox game player.

Dell

(DELL) - Get Report

just raised its projections for the quarter that ends in January 2002 by a penny a share.

Genesis Microchip

(GNSS)

reported 9 cents a share above the Wall Street consensus on strong demand for its chips that go into flat-panel displays, a hot technology among consumers.

Allegheny Technologies

(ATI) - Get Report

reported continued strong demand for its graphics chips and chipsets.

Of course, many of the stocks in the strongly growing parts of the technology sector held up relatively well during the decline in the rest of the sector, and then rebounded with extreme strength at the end of the year. Shares of ATI, for example, were up 157% since Sept. 21 going into the company's earnings report. That kind of appreciation doesn't get a stock a solid score against the third yardstick, valuation.

But the recent weakness in the technology sector gives me some hope that the market is about to fix that problem for, at least, a few technology stocks. A few have pulled back enough recently that they don't discount too much of any potential recovery. I'd put Genesis Microchip on that list -- especially if it drops below $60.

GlobespanVirata

(GSPN)

, meantime, is reaping the rewards of rapid digital subscriber line (DSL) deployments in Asia and is about to see its European market step up. I'd like to get Globespan at $14 or so -- about the price at which I sold it in Jubak's Picks in July 2000 -- on current weakness.

And then there's

Intersil

(ISIL)

, the dominant player in chipsets for the wireless local area network market. That technology is moving into the mainstream with wins at

Compaq

(CPQ)

,

IBM

(IBM) - Get Report

and Dell. The stock has pulled back on speculation that

Broadcom

will enter this market later in 2002. That is indeed likely, and I'd certainly hold off on the stock until after Broadcom makes its move and the market discounts this new competitor.

And, of course, if we get to March and the bad news from the semiconductor equipment makers and their customers has beaten up Cymer a bit, I'd certainly add that stock to this list.

Jim Jubak appears Wednesdays on CNBCs "Business Center" at 6 p.m. EST. At the time of publication, he owned or controlled shares in the following equities mentioned in this column: Dell, E*Trade, Intel and Microsoft.