From Briefing.com: Ahead of the Curve: The Semiconductor Industry: Q3 Projections 09-Sep-03 09:59 ET
[BRIEFING.COM - Robert V. Green] Revenue and earnings projections for the semiconductor industry for Q3 are quite astounding. It is no wonder that many investors are excited about the prospects. On a stock by stock basis the prospects seem to indicate that a "rebound" of the entire sector is right around the corner, with an implication that all technology stocks will follow. Adding up all the projections for the entire industry tells a more interesting story, however.
Stocks Included
The object of this study is to compare projections for the coming quarter to the trend of the past five quarters.
For this study, we looked at all semiconductor stocks that met the following criteria:
* Trade on the NYSE, AMEX, or NASDAQ exchange
* Have both at least one revenue and earnings projection available
* Have at least 0.5% of the the industry's total market capitalization or total revenue
There are total of 193 stocks in the semiconductor industry (excluding bulletin board stocks). These parameters selected 36 stocks, but those represent approximately 90% of the entire industry market capitalization and 85% of the total revenue in the industry.
Once the stocks were selected, a spreadsheet was created capturing the quarterly revenue and net income for all of the stocks. By including the projected revenue and earnings for Q3, a "forty-thousand-foot" view of the overall industry can be seen.
For all of the Q3 estimates, we used the consensus revenue and earnings forecasts as collected by Multex.
Revenue Projections - Entire Industry
The following chart shows the revenue trend for the selected 36 semiconductor stocks
This following table summarizes the sequential revenue change for each quarter.
02 Q2 02 Q3 02 Q4 03 Q1 03 Q2 03 Q3 Est
Quarterly Revenue (B) $ 23.7 $ 27.8 $29.4 $26.9 $27.9 $30.1
Sequential Change --- 17.6 % 5.6 % -8.3 % 3.7 % 7.8 %
The total revenue projections are robust, but certainly reasonable and believable. That does not mean it will all happen, of course, but it certainly is not so strong as to require a "leap of faith."
Earnings Projections - Entire Industry
The following chart shows the earnings projections for the entire industry as total net income.
Unlike the revenue chart, this chart is simply extremely hard to swallow at first glance.
This chart is essentially a "pulling a rabbit from the hat" type of scenario. The earnings for Q4 of 2002 include extraordinary charges, which somewhat exaggerates the loss in Q4. (Accounting rules changed in 2003, incenting many companies to take charges in Q4.) We believe it is reasonable to include extraordinary charges when looking at trends on a scale larger than one year, but even backing them out does not change the picture much, as the total extraordinary charges in Q4 are less than half of the loss shown above.
The following table captures the data for the projected sequential increase in Q3 for entire semiconductor industry.
Total Industry, $US, Millions 03 Q2 03 Q3 Est Sequential Increase
Revenue $ 27,917 $ 30,089 $ 2,172
Net Income $ -309 $ 2,561 $ 2,870
In short, earnings projections for the entire semiconductor industry make the assumption that the industry, as a whole, will increase earnings more than revenue increases. If true, it would be an astonishing business feat. Some improvement in operational margins is to be expected, as the industry has shed many jobs over the past year, but to expand net income more than revenue increases is unusual.
We are not willing to say this projection is impossible, but in order for the scenario to play out, for the entire industry, it will require that most companies improve both their gross margins and their operational margins substantially. There is not much precedence for something like, even in cyclical industries.
Revenue Projection - Industry Minus Top 5
A slightly different picture emerges, however, when you remove the top five (by market capitalization) companies from the analysis. These companies are: Intel (INTC), Taiwan Semiconductor (TSM), Texas Instrument (TXN), Applied Material (AMAT), and STMicroelectronics (STM). These top 5 companies account for 52% of the total market capitalization in the industry and 36% of the total revenue.
Here is the total revenue chart for the "bottom half" of the industry.
This is still a very believable projection.
Earnings Projection - Industry Minus Top 5
Here is the total earnings chart for the "bottom half" of the industry.
This is a much more believable projection, based on past trends, but we actually think the "bottom half" of the industry might be where the disappointments occur.
Market Conclusions
Projections for the semiconductor industry call for an extremely healthy rebound in Q3.
Much of this sentiment comes from the extremely positive guidance given by Intel, who have said they expect to be in the top part of the estimate ranges for both revenue and earnings. It is very likely that Intel will fulfill these expectations.
However, Intel almost stands alone in providing strong guidance. The only other stocks in this group to provide upward guidance for Q3 have been Cypress (CY) and On Semiconductor (ONNN), much smaller players.
A major assumption has been made by most analysts, we believe, that "what is good for Intel, is good for the industry." This has been true for nearly 15 years, but may not be true this time.
The following table breaks out Intel's portion of the total industry
03 Q3 Sequential Increases Revenue, estimated increase Earnings, estimated increase
Total Industry $ 2,172 $ 2,870
Intel $ 878.7 $ 601.3
Intel Percent of Industry Increases 40 % 21 %
Intel's projected earnings increase is less than its projected revenue increase, a far more normal business scenario than that projected for the rest of the industry.
We view the technology market as maturing. A strong performance in Q3 by Intel and other top players, coupled with disappointing performances by the "bottom-half" players will only confirm that view. That is what happens when markets mature. The biggest players with strongest market share begin to squeeze out everyone else.
At one point in the auto industry, there were hundreds of auto manufacturers. Today, there are only three. Today there are 193 semiconductors (not counting those on the bulletin board.) How long will it be until there are only ten? It is time to start asking that question.
Investment Conclusions
There are two conclusions relative to semiconductor investments now.
* Investments in the bottom half companies, particularly if valuations are high, have become risky. An acquisition investment premise is the only one that should be used.
* Investments in the top half companies should be examined for valuation. While the underlying business may do well, a growth valuation may not be sustained for a long term investment premise.
As markets mature, growth valuations become replaced by mature market valuations. That process is likely to occur over the next two years, making it hard to come up with two to three year investment premises for most of the semiconductor stocks. Short term investment premises, fueled by momentum and the still strong idea that a "rebound is coming" may work well, but take your profits when you get them.
Comments may be emailed to the author, Robert V. Green, at rvgreen@briefing.com
Spreadsheet Available
If you are a direct subscriber to Briefing.com, you may obtain the Excel spreadsheet used in this analysis by sending me an email at rvgreen@briefing.com. Please put "Semiconductor Spreadsheet" in the subject line and your Briefing.com user name (not password) in the body of the email. Please also allow 24 hours for delivery.
PS: Notes on the Analysis
The following notes detail some aspects of this study, but, in general, none of them have a material affect on the conclusions.
Note: published earnings projections rarely include the total share outstanding number for individual stocks. When available, we used the average total net income estimates as the projected net income for a stock. When a total net income number was not available and only an earnings per share (EPS) number was published in an analysis, we used the consensus EPS number and multiplied by the current shares outstanding number to obtain total net income for a stock.
Note: we have used net income for this analysis, which analyzes the entire business, rather than earnings available for common stock (from which EPS is derived). If a stock pays a preferred dividend, the dividend is subtracted from net income before calculating earnings per share. Although a few of these stocks pay a preferred dividend, this distinction makes no material difference in the analysis.
Note: Kyocera (KYO), which constitutes 1.9% of the industry market capitalization and 6% of the total revenue, was omitted from this study because no revenue or earnings projection were available individually for Q3.
Back to Top
Copyright � 2003 Briefing.com, Inc. All rights reserved. |
