You get some pretty funny numbers these days if you try to value technology stocks using earnings per share and price-to-earnings ratios.

For example, take a look at the value that

MSN MoneyCentral's

Research Wizard puts on semiconductor maker

PMC-Sierra

(PMCS)

. By multiplying the stock's current P/E ratio by the earnings projections of Wall Street analysts, you discover that, as of December 2001, the stock will be worth between $20.62 a share (about a 50% loss from the recent price of $41) and $791 a share, an 1829% gain (all data as of May 9).

We're all friends here, so why not just split the difference, and say, as the Research Wizard does, that the stock will be worth $165 a share, a 300% gain. Should you hold it until December 2002? Sure. By then the stock will be worth somewhere between $151 and $1,272 a share.

I'm not telling you anything you don't know if I say that a target price like this, or the $880 to $941 December 2001 target the Research Wizard puts on

Qualcomm

(QCOM) - Get Report

, is garbage. And I certainly understand the impulse to just throw out the traditional, basic methods for valuing a growth stock -- multiplying earnings per share by some price-to-earnings ratio, for example -- when it produces numbers like that. What's the good of knowing that PMC-Sierra should trade somewhere between $21 and $791 a share next December?

The problem isn't traditional valuation methods or the tools, like Research Wizard, that are built around them, but rather the quality of the numbers that are being plugged into these formulas right now. At a time when earnings visibility is severely limited, when hundreds of technology companies are reporting huge swings from black to red ink and when future swings in earnings on the upside are likely to be just as pronounced, investors can't simply plug numbers as given into the usual formulas and expect that anything other than rubbish will come out. This is one of those times in the market's cycles when investors have to manipulate the numbers to find their true meaning.

And the most important manipulation to undertake these days is one that turns current reported and projected earnings into something called "normalized" earnings. Once you've done that, you can go back to the traditional valuation formulas with some hope of producing meaningful target prices for technology stocks.

Huge Swings in Estimates

Let me show you what I mean by normalized earnings and then give you 10 names from the Fantastic Future 50 portfolio that look like interesting "new money" buys, according to this way of looking at the world.

To show you how normalized earnings work, I will use as examples two stocks that I have written about before:

RF Micro Devices

(RFMD)

and

EMC

(EMC)

. I've written elsewhere that RF Micro Devices was a decent buy -- in fact, I even raised my target price for the stock -- but that EMC was too expensive to buy now. On the surface, with unnormalized earnings, that opinion doesn't make much sense. After all, RF Micro Devices recently traded at a trailing 12-month P/E ratio of 162. And EMC sports a trailing 12-month P/E ratio of 50. But EMC is expensive at a P/E of 50. The picture doesn't change much if you use consensus

projected

earnings either. RF Micro Devices sells at about 104 times the Wall Street estimate for the fiscal year that ends in March 2003. EMC sells at 39 times projected calendar 2002 earnings.

OK, to work. I'll start with RF Micro Devices. The first thing to notice is how extremely lumpy the company's earnings stream looks right now, both within an individual quarter and over time. For example, analysts project that the company will earn as much as 33 cents a share in the fiscal year that ends in March 2002 or lose as much as 7 cents a share. That's a huge swing of 40 cents a share, and a good indication that the projections themselves are subject to a high degree of uncertainty. You can use the 2 cents a share median in your calculations, of course, but I don't think that number captures the volatility or uncertainty in the current situation very well. If sales and earnings recover a quarter earlier than expected at RF Micro Devices, for example, the change in fiscal-year earnings could be huge. On May 4,

J.P. Morgan Chase

changed its estimates for exactly that reason. Citing the likelihood of a better-than-expected June quarter, the firm pushed its estimates for fiscal 2002 to 5 cents from a penny. That's a 400% change.

The consensus estimates for fiscal 2003 are almost as uncertain -- they range from 23 cents to 43 cents -- and are about as subject to change. But what really strikes me when I look at fiscal 2001, 2002 and 2003 earnings estimates is the size of the swing from year to year. Earnings per share came in at 18 cents in fiscal 2001. They're projected to drop to 2 cents a share in the fiscal year that ends in March 2002 and then to rebound to 31 cents a share in fiscal 2003.

So what's the real long-term growth rate of this company? It's clearly not the negative 90% from 2001 to 2002, and it's just as surely not the positive 1,900% growth from 2002 to 2003.

Smoothing the Trend

What I'd like to do is take out the effects of the slump in growth this year -- a process called normalizing the earnings. Not because that slump didn't happen or because it was insignificant, but because mathematically, it distorts the trend. As an investor considering buying RF Micro Devices now, I'm certainly not willing to pay for long-term 1,900% growth because it isn't a meaningful long-term trend. And neither is the likely 90% drop in earnings this year.

You can stay very simple or get as fancy as you want in your normalizing process. At the extremely simple end, you can just use the 5-year earnings growth rate projected by Wall Street analysts. MoneyCentral lists that projection

under Analyst info, Estimates, Earnings growth rates. For RF Micro Devices, the consensus estimate is 35% annually over the next five years. At the fancy end, you could do your own estimates from scratch.

I like to begin with the consensus estimate and then see where modification is merited. For example, right now analysts are including historically low gross margins of less than 40% in their projections, largely because RF Micro Devices is experiencing relatively low manufacturing yields on its new global system for mobile communications, or GSM, wireless-phone power amplifier modules. That's typical of a new semiconductor product being produced in limited runs, and I think those yields will rise as the company gains experience and builds volume.

In addition, the current five-year earnings growth projection is well below the 100% earnings growth the company achieved from fiscal 1999 to 2000. I understand that growth in the wireless-phone handset market may not be about to return to the heady earlier years, but with RF Micro Devices just now entering the global GSM market, I think the company can grow faster than the wireless market as a whole. GSM was just 28% of revenue for RF Micro Devices in the March quarter. All in all, I think the 35% growth estimate is low and 45% is a more accurate normalized growth rate for the company.

Now I'll turn that into a target price:

  • I start with my own estimate for fiscal 2003 earnings, which looks like it will be the first full "business-as-usual" year for RF Micro Devices. The analysts' consensus is 33 cents. My estimate of 41 cents is closer to the high end of the analyst range at 46 cents.
  • Then I apply my 45% normalized growth rate to that figure, resulting in earnings projections of 59 cents for fiscal 2004 and 86 cents for fiscal 2005.
  • By that time the normalized trend should be clearly established for investors, and I think RF Micro Devices would carry a P/E of two times its forward growth rate, or 90. That gives me a target price of $78 a share for March 2005.

Now, discount that backward to see what you'd be willing to pay for that stock today. For a stock with RF Micro's risk profile -- good visibility and a first-class list of customers add up to a moderate risk in my book -- I'd like to see the potential for a 20% annual return on my money. Using that discount rate and working back toward the present, I get a target price of $37.62 in May 2001. (In other words, I'd be willing to pay that much now for a stock with this risk and this return potential.) And that's why I raised my target price to $37 on May 8.

Why EMC Is Fully Priced

You can do this same kind of normalization with any technology stock. For EMC, I'd start with the $1.05 Wall Street consensus for 2002 earnings per share and whack off a few cents to 98 cents a share because of the increasing competition I see for EMC from

IBM

(IBM) - Get Report

,

Hitachi

(HIT)

and other storage suppliers. Then, I'd adjust that number to $1.04 to put my estimate on the same March 2003 year as I used for RF Micro Devices. I'd also use a slightly lower long-term growth rate than Wall Street now estimates -- 26% annually over five years instead of 28% for the same reason. Taking that out to March 2005 gives me earnings of $1.65 a share. Using a multiple of two times the growth rate of 26%, or 52, I get a target price of $86 a share for March 2005.

At this point in the economic recovery of the technology sector, I think EMC is actually slightly more risky than RF Micro Devices. In the case of RF Micro Devices, investors can see that the revenue recovery is in place for the June quarter; the company has orders in hand to guarantee a 20% quarter-to-quarter increase in revenue. In the case of EMC, management is talking about a "firming up" of corporate information technology budgets in the U.S. that should translate into increased orders in the second half of 2001. So I'd use a slightly higher discount rate for EMC than for RF Micro Devices at this juncture. Maybe 22% instead of 20%. Discounting backward to the present at 22% gives me a target price of $39 a share. With the stock trading near $41 recently, it seems fully priced.

What would make EMC a buy instead of a hold? An increase in the certainty of revenue gains that cuts the risk on the stock would do. At a discount of 15%, the stock becomes a buy. So would anything convince me that margins at EMC weren't going to suffer going forward from increased competition? Absent conviction on either of those fronts, I won't be looking to add to my position in EMC.

I also find this framework useful in pointing me toward potential buys in the technology sectors. What does it tell me to look for? Stocks that are showing diminished risk because of rising business visibility. And stocks that will ride major new product initiatives or major gains in market share to better-than-expected growth.

Let me give you 10 potential "new money" buys from the Future 50 that I think fit either one or both of those descriptions. Among current new money buys I'd list are:

AOL Time Warner

(AOL)

,

Applied Materials

(AMAT) - Get Report

,

E*Trade

(EXDS)

,

Exodus

(EXDS)

,

JDS Uniphase

(JDSU)

,

Metromedia Fiber Network

(MFNX)

, RF Micro Devices and

Yahoo!

(YHOO)

. I'd give new, new money buy ratings to

i2 Technologies

(ITWO)

and

Qualcomm

(QCOM) - Get Report

. To make room for those stocks, I'm dropping my new money buy ratings on

Intel

(INTC) - Get Report

and

Openwave

(OPWV)

. I think the uncertainty is still rising for those two companies, and I'd prefer to wait until business trends are clearer.

At the time of publication, Jim Jubak owned or controlled shares in the following equities mentioned in this column: AOL Time Warner, Applied Materials, EMC, E*Trade, Exodus Communications, Metromedia Fiber Network and RF Micro Devices.

More from

MSN MoneyCentral