books about online stock trading, forex, futures, stock investing, market, trading systems
Advanced Micro Devices
Back to contents page

Now, it's TIME for the fun part. You now have an investment philosophy, you know what economic moats are, and you know how to read financial statements and evaluate companies. Let's put all this knowledge to "work by analyzing two real-world companies: chipmaker Advanced Micro Devices (AMD) and Biomet, a medical device firm.

(Note: All of the financial data in this chapter Is available free on Mornlngstar.com in the same format that you see it here.)

Advanced Micro Devices

At first, AMD might look like an attractive investment. It's one of only two companies that manufacture microprocessors—the brains in a PC or server— and computers are pretty ubiquitous devices. AMD also makes flash memory chips that are used in a variety of devices that should have solid long-term demand, such as mobile phones and network routers. In the microprocessor market, AMD has caught up to archrival Intel on the technological front over the past several years, and for a brief spell in the late 1990s, the firm was selling chips that were arguably faster and better than those being sold by Intel.

This is the AMD story that you might know if you followed the news a bit and casually flipped through the company's Web site and its recent annual reports. However, it's not enough information to make a sound investment decision, so let's approach the firm systematically to see just how solid an investment AMD really is.

Economic Moat

First, look for evidence of an economic moat. As we discussed in Chapter 3, "we can do this by examining how profitable AMD has been in the past by analyzing free cash flow, margins, return on equity, and return on assets (see Figure 11.1).

It looks like AMD has had a pretty spotty history of generating free cash flow. After a few good years in the early 1990s, free cash flow turned negative as the firm heavily increased capital spending, and AMD didn't throw off any free cash again until the technology boom in 2OOO. That's not a good sign that the firm has much of an economic moat. However, some rapidly growing firms spend years plowing all of their money back into capital spending, which means their free cash flow is negative because they're still building their economic moat. (Starbucks was a great example of this during the 1990s.) Even these kinds of firms should have solid profits, however, so let's turn to AMD's margins and returns on capital (see Figure 11,2).

Looking at operating and net margins, we see that although AMD scores better on this front than it does on free cash flow, it still lost money for 6 of the past 10 years. The trend doesn't indicate that these were simply cyclical

Figure 11.1 AMD's free cash flow.

Figure 11.2 AMD's profitability numbers.

losses—the firm lost money for four straight years between 1996 and 1999, even though the U.S. economy was g om g gangbusters.

ROE and ROA tell a similar story. After some salad days in the early 1990s—ROEs In the high teens are pretty good—AMDs performance went into a big slump until the tech boom in 2OOO. We can also see that asset turnover has declined, which means that AMD has become less efficient and financial leverage has gradually edged upward, indicating that AMD has probably taken on more debt.

AMD has had one great year recently—ROE, free cash flow, and margins were all excellent in 2OOO—and it's worth our time to find out why so we can determine whether it was an aberration or a sign of better times to come. After reading through some recent annual reports, we find that AMD made money in 2OOO the old-fashioned way: It rolled out a chip that was faster and cheaper than anything Intel had at the moment. This made the company a ton of money until 2OOI, when the demand for PCs slowed down and Intel rolled out a competing chip.

So, AMD's one big recent success was due to a product with superior technology, which "we know from Chapter 3 is the least sustainable source of an economic moat. It shouldn't come as a surprise to us that Intel quickly used its superior size and cash hoard to accelerate the rollout of a competing chip and regain the market share it briefly lost to AMD.

Overall, there's not much evidence of a sustainable economic moat in AMD's historical financials. It's not difficult to understand why: Intel dominates the microprocessor market, and AMD's market share has typically been around IJ percent for the past several years. Intel's size lets it spend four times as much on research and development as AMD does, which is a big advan­tage when you consider how fast semiconductor technology changes. Intel's size has also allowed it to attain much greater economies of scale, because it can spread the fixed costs of its manufacturing plants across a much larger volume of chips.

AMD's lack of an economic moat means we'd want a big margin of safety If we wanted to buy the stock, but even no-moat companies can sometimes be decent investments if the fundamentals aren't too shaky and the stock is cheap enough. Let's complete our analysis by looking at the five areas I dis­cussed in Chapter 6 —growth, profitability, financial health, risks, and management—and then by doing a rough valuation of the stock.

Growth

Figure ii.3 shows that revenue growth has been somewhat volatile for AMD over the past decade. Overall, it's not been terrible, but a 6 percent average annual growth rate over a decade "when PC demand was pretty hot is nothing to crow about.

It's tough to say much that's meaningful about growth in profits because AMD had so many money-losing years during the 1990s. About the best "we can do is note that AMD made a lot more money in its most recent profitable year (2000) than it did in the previous profitable year (1995), but it also lost much more money in 2OO2 than it ever had before—a spotty track record, at best (see Figure 11.4).

Profitability

We looked at free cash flow and returns on capital "when "we "were assessing whether AMD had an economic moat, and the verdict was pretty negative. Let's dig a bit deeper to see what else we can understand about how AMD

AMD Revenue Growth 93 94 95 96 97 98 99 00 01 02

Figure 11.3 AMD's revenue growth.

94 95 96 97 98 99 00 01 02

Net income ($mil) 218.4 294.9 300.5 -69.0 -21.1 -104.0 -88.9 983.0 -60.6-1,303.0 Figure 11.4 AMD's net income.

makes (and loses) money. Look at Figure II, J, which is a common size income statement. Common size statements are great tools for evaluating companies because they put every line item in context by looking at each of them as a percentage of sales.

These numbers show some disturbing trends. In the early 1990s, gross margins were around 50 percent, and they've steadily slid to between 20 percent and 30 percent in 2OOI and 2OO2. Spending on overhead (SG&A) has been pretty steady at 16 percent to 19 percent of sales, while R&D spending has increased dramatically. The big increase in R&D spending helps support our previous intuition that chip companies need to spend large amounts on R&D to stay competitive. It also correlates with the big increase in capital spending we saw on the statement of cash flows—AMD was spending boatloads of money on expanding its manufacturing capabilities.

Overall, we have a company making less money per chip, not becoming much more efficient in terms of overhead spending, and having to increase research spending in a big way. Couple these trends with highly variable sales—see the revenue growth line in Figure 11.3—and you have a pretty dismal profitability picture.

Figure 11.5 AMD's common size income statement, representing each line item as a percent of revenue.

Unfortunately, things don't look much better when we examine AMD's financial health.

At the end of 2OO2, the firm had Si.9 billion in debt and $2.5 billion in shareholders' equity. The resulting debt-to-equity ratio of 0.7 isn't terribly high relative to the market, but it's not great for a firm that has as much trouble generating profits as AMD does. AMD's current ratio—which, remember, is simply current assets divided by current liabilities—was about 1.5. Again, not terrible, but not comforting either, given AMD's second-tier industry position and spotty history of profitability.

Finally, we learn by digging into AMD's IO-K filing that the firm has some big loans outstanding with banks in Germany , relating to a large manufacturing facility that the firm is building there. Adding these loans to other contractual obligations, "we see that AMD will need to pay out more than $950 million to various parties between 2004 and 2006. That's a huge amount of money for a company that has generated only $250 million in net operating income since 1993 and has bled about $2.7 billion in free cash flow over the same period.

(You can calculate the cumulative operating income and free cash flow amounts yourself from the previous figures.) I'd say AMD doesn't score very "well on the financial health front.

The Bear Case

Although I normally recommend developing a strong bear case for any company you analyze, we've uncovered many more negatives than positives about AMD so far, so let's move on.

Management

Let's look at some of AMD's proxy statements to see what we can find out about management. For one thing, they're paid pretty well: Outgoing CEO Jerry Sanders pulled in about $1 million in salary every year between 1997 and 2OO2, an additional $400,000 in "deferred retirement compensation" for each of those years, and bonuses ranging from zero to $5.1 million. He also received a good-sized amount of additional "in-kind" compensation from use of company vehicles and the company plane. In 2OO2, for example, he re­ceived about $184,000 worth of company-provided vehicle services. That's either an expensive car or a well-paid chauffeur.

Other top executives also did well in 2OO2, with the top five individuals receiving salaries of between $450,000 and $900,000, and three of the five receiving hefty bonuses despite AMD's Si. 2 billion loss that year. Incoming CEO Hector Ruiz didn't get a cash bonus, but don't feel too sorry for him—he received 1.2 million options (about 10 percent of the total granted that year) instead. Depending on the assumptions used, Ruiz's options will be worth between $12 and $30 million over the next decade.

Speaking of options, it looks to me as though AMD has been giving away the store to its employees and diluting shareholder value as a result. From 2OOO through 2003, AMD issued about 46 million stock options to its em­ployees and officers, increasing the number of shares outstanding by 15 percent. (You can find the total number of options granted each year in that year's proxy statement.) Therefore, anyone who bought AMD shares in 2OOO and held them through 2003 saw their stake in the company shrink by 15 per­cent after three years, simply because the firm gave away so many options to its employees. This kind of egregious options granting tells me that manage­ment cares little for outside shareholders.

Overall, I'd say that AMD's management is overpaid, and they are not the type of folks to whom I'd entrust my money.

Valuation

Valuation is tough, because AMD has lost money for the past two years. On a price-to-sales basis, AMD was trading at about 1.5 times sales in September, 2003, which was much cheaper than the seven times sales valuation of the chip industry average and in line with AMD's own 1.6 times sales average valuation over the past five years—not too bad.

We can't use a P/E because AMD was forecasted to lose about $0.30 per share in 2004. In any event, it would be tough to have much confidence in earnings estimates for a company that's been as volatile as AMD—for example, 2004 earnings estimates for AMD ranged from a low of—$0.85 per share to a high of $O.2O per share in September, 2003. In other words, no one has the faintest clue how much money AMD will be making (if any) in the near future, which tells us that we should demand a big margin of safety from our val­uation we arrive at.

Because AMD is highly cyclical, we could also attempt to predict its earnings at the peak of the next cycle and value the company based on the resulting P/E ratio. Figure 11.6 shows that AMD's last peak earnings per share was almost $3 in 2OOO. However, "we'd be "wise to take this figure "with a grain of salt because the earnings occurred during a technology boom that's unlikely to repeat and during a rare time when AMD was able to catch Intel napping. Still, even if we assume that AMD will earn only $i per share during its next cyclical peak, the stock doesn't look too expensive—it was trading at $12 in September 2003, "which is 12 times peak earnings.

The problem is that we have no idea when (or if) AMD's next cyclical peak will arrive. After all, the company has a weak competitive position and a troublesome balance sheet. Thus, I don't think valuing it on peak earnings makes much sense.

Finally, "we can attempt to forecast cash flows and use a DCF approach. This is also tough because AMD has generated positive annual free cash flow only three times in the past decade. Let's give it a "whirl, though.

If we conservatively estimate that AMD returns to positive free cash flow in 2005, generates $200 million that year, and increases free cash flow at 5 percent annually for the next 10 years, it will generate about $2.2 billion in free cash over the next decade. Discounted back to the present at 14 percent, that's about $1.0 billion. (I used 14 percent because AMD had a poor track record of profitability, had considerable balance sheet risk, and operated from a weak competitive position in a cyclical industry.) Add in our perpetuity value of $750 million, and AMD is worth about $1.8 billion, or about $5 per share (see Figure 11.7).

AMD Net Income 93 94 95 96 97 98 99 00 01 02

Figure 11.6 AMD's net income and shares outstanding.

With the stock trading at $12 as of this writing, I'd take a pass. With such an uncertain future, this is a stock that I would buy only at a big discount to a conservative estimate of intrinsic value—and perhaps not even then, given how many strikes the firm has against it. The intrinsic value could be much lower if AMD fails to generate free cash flow in the near future or runs into liquidity troubles, or it could be much higher if AMD gains some kind of competitive edge against Intel. In particular, the next-generation "Hammer" family of chips AMD has rolled out could give the firm a boost. Given the firm's poor track record of holding its own against Intel, though, that's not a bet I'd want to make. The stock could move higher if the market gets enthusiastic about AMD's prospects—it tends to be pretty volatile—but it seems like a poor long- term investment.

Smarter trading The art of day trading Trading Chaos Sane Investing In An Insane World
Beat The Odds In Forex Trading