SanDisk Corporation (SNDK) is a global leader in flash memory storage products. The company sells a variety of such items, including USB flash drives and other portable media [e.g. SD, MicroSD, Compact Flash), embedded memory for mobile phones and tablets as well as a range of high end solid state drives (SSD)] used in laptops, desktop PCs and network servers. The company pioneered (solely or in partnership with other market leaders) many of these technologies and has a sizable patent portfolio that has generated $1.15 billion in licensing revenue over the last three years (covering much of the $1.35 billion the company spent on R&D over the period!).
There are a lot of things to like about SNDK. Perhaps most notable is its solid capital structure with $3.84 billion in net cash which comprises 45% of its market cap, resulting in an EV of about $5 billion. On this basis, the company has generated average free cash flow over the last three years of around $800 million, for a yield greater than 16%. Furthermore, the company has experienced significant growth, with 2011 revenues 45% higher than five years prior (7.7% CAGR: not bad for a recession!).
At the time of writing, the company’s stock is down 26% YTD as a result of a weak first quarter and poor forecast for the second quarter. When I see a company with solid historical performance collapse in price as a result of one bad quarter, I investigate to see whether this is a temporary anomaly that might present a good value opportunity.
So what happened in Q1? Checking the company’s earnings press release, we see a slight decline in revenues and pretty significant gross margin contraction. The president and CEO had this to say on the conference call:
We now believe the industry experienced an imbalance between supply and demand in the first quarter, and we believe this imbalance is likely to continue through the second quarter of 2012. Industry demand has been weaker than we had expected, and we believe supplies have increased more than we had estimated at the beginning of the year. This led to greater-than-expected price decline, particularly in the second half of the first quarter, impacting our revenue and gross margin. …
For SanDisk, our first quarter results were impacted by 2 key factors: greater-than-expected price decline; and slower-than-expected demand for mobile cards from certain OEM customers with whom we have high shares.
Let’s take the second factor first. SNDK has little to no exposure to Apple (NASDAQ:AAPL), and Samsung (OTC:SSNLF) (the largest mobile manufacturer) only accounts for 10% of SNDK’s sales. So the bulk of SNDK’s revenues from mobile OEMs are from players outside the top tier. Industry dynamics have been forcing the (distant) second tier to cut prices dramatically in order to compete with the top tier. The result is that these manufacturers are no longer bundling microSD cards with their phones, which means consumers then have to buy cards on their own. These sales would logically be higher than the bulk sales to OEMs, but not all of these purchases will buy microSD cards and many may buy from SNDK’s competitors (from my check of various retailer websites, SNDK appeared to be priced significantly higher than its competitors). Note also that SNDK executives stressed in the conference call that the company has several products being tested by the top OEMs (read AAPL) for implementation later this year which should help SNDK’s mobile sales.
The bigger issue is the first: a “greater-than-expected price decline.” In some industries, this is the result of a one-time phenomenon that leads to short-term pain as companies digest oversupply and atone for overly optimistic past projections. Unfortunately, this is not one of those industries. You see, SNDK and its competitors are trapped on a rapidly moving treadmill that requires all parties to bring their costs down at an ever increasing rate in order to maintain profitability (hence the “expected” decline). From the company’s 10-k:
We believe the markets for flash storage are generally price elastic … In order to profitably capitalize on this price elasticity, we must reduce our cost per gigabyte at a rate similar to the decrease in selling price per gigabyte, while at the same time increasing the average capacity and/or the number of product units enough to offset price declines. We continually seek to achieve these cost reductions through technology improvements, primarily by increasing the amount of memory stored in a given area of silicon.
On one hand, the competition to bring prices down is good in that it will drive the transition from traditional HDDs to SSDs and open up the industry to a much larger audience. On the other hand, a bet on SNDK (or any of its competitors) is really a bet on its ability to reduce its costs per gigabyte faster than the price decline per gigabyte. In a perfectly competitive environment, the company with the lowest marginal production cost sets the price and drives down the profitability of all the other players (before we rush to assume perfect competition, note that industry participants are subject to a number of ongoing class action lawsuits alleging price fixing – see Note 16. Litigation). So SNDK is constantly striving to have a lower cost base than its competitors, which will allow it to set the price.
There are a few things to be worried about here. First, whenever industry competition is based largely on price there is a significant risk created as marginal competitors are driven from the industry. These competitors dump their products and try all sorts of wild and unpredictable things as they attempt to stay alive, often upsetting (even if just for a short time) industry economics.
Second, the effort to reduce costs is expensive in itself. Here we see that SNDK has spent 20% of its Enterprise Value over the last three years just on R&D in an attempt to achieve a lower cost per gigabyte than its competitors.
Third, the drive to lower costs often leads to commitments that would not normally be desirable. For example, the surest way to lower costs is to work toward a more fixed cost basis and then attempt to ramp up units sold, achieving increasing economies of scale. This is known as operating leverage and increasing it has been SNDK’s strategy:
Our strategy of investing in captive manufacturing sources could harm us if our competitors are able to produce products at lower cost or if industry supply exceeds demand. We secure captive sources of NAND through our significant investments in manufacturing capacity. We believe that by investing in captive sources of NAND, we are able to develop and obtain supply at the lowest cost and access supply during periods of high demand.
While the company does have some non-captive manufacturing in which it has some swing capacity via third party manufacturers, the bulk of its product is done in-house through a joint venture with Toshiba.
It is important to note that operating leverage, like financial leverage, cuts both ways, so as demand declines those fixed costs remain the same and losses begin to mount. So SNDK’s strategy of increasing operating leverage in order to achieve greater marginal profitability is increasingly risky as the company is relatively worse off during times of lower demand. From a review of the company’s recent 10-Ks, we see that it has been continuously increasing its investment in these joint ventures in order to maintain its cost advantage. Like a company that continuously increases its financial leverage, investors should find a company that continuously increases its operating leverage to be highly risky and thus unattractive.
There are other negative consequences of this strategy. As I mentioned, the company has a joint venture with Toshiba for manufacturing in Japan. The arrangement is that both SNDK and Toshiba purchase 50% of the manufacturing output of the JV at the cost of production plus a given markup. The problem is that sales are largely in USD while manufacturing costs are incurred in Japanese Yen, which has gained 46% over the last five years. This acts as a drag on the company’s performance (the company does utilize hedges, but this only delays the ultimate reckoning when there is a clear long-term trend in currency exchange rates). Unfortunately, this is not the biggest problem.
The larger issue with these joint venture arrangements it that they are structured as variable interest entities (VIEs) in which the company owns just a hair under 50%, giving Toshiba control. First, as a VIE, SNDK has been able to push obligations it has on behalf of the VIE off its balance sheet, understating its liabilities. Here’s what the company says about these guarantees (emphasis added):
For semiconductor manufacturing equipment that is leased by Flash Ventures, we and Toshiba jointly guarantee on an unsecured and several basis, 50% of the outstanding Flash Ventures’ lease obligations under original master lease agreements entered into from March 2007 through November 2011 and refinanced master lease agreements entered into from April 2010 through November 2011. These master lease obligations are denominated in Japanese yen and are noncancelable. Our total master lease obligation guarantee as of January 1, 2012 was 56.5 billion Japanese yen, or approximately $732 million based upon the exchange rate at January 1, 2012.
Thus, investors need to account for the fact that there are significant off balance sheet liabilities related to its manufacturing operations.
The other issue with this structure is that, when a company does not have control over its subsidiary (either owning more than 50% or via contractual arrangement), it accounts for the subsidiary via the equity method rather than the consolidation method. Under the consolidation method, we would see all of the JV’s liabilities, assets, revenues and expenses show up on SNDK’s financial statements, with a minority interest line deducting the portion attributable to Toshiba. This would more accurately represent the true cost structure of SNDK. Instead, SNDK uses the equity method, which means that only its share of the (meager) profits of the JV shows up on its income statement and only its portion of the equity in the JV shows up on the balance sheet. The effect of this is that profitability appears greater and leverage appears lower.
So there are a few things here that leave me unhappy with SNDK’s decisions regarding how it reports financials, all stemming from its JV with Toshiba. On its own, I wouldn’t be as worried because we can adjust the financial statements to be more representative. But the larger issue is that SNDK operates in an industry with unattractive economics that create incentives to increase operating leverage, which increases risk.
I’ve tried to make the case here that, while the company has performed well over the last few years thanks to its captive manufacturing and economies of scale, investors today are reliant on the good times continuing. If declining demand for the company’s products persists, the double edged sword of operating leverage will come back to sting investors. The rapid decline in gross margins this past quarter is a testament to how quickly things can change. Consequently, I would have to see the share price fall much further in order to get interested.
Oh, one more thing. The company has a shareholder rights plan to entrench management and directors at the expense of shareholders. I hate that, and so should you.