MU (MU, Financial) is a manufacturer of semiconductor memory chips (DRAM and NAND flash). This isn't our first go-round with MU; it was a large short position from January 2001 to February 2005. Back then, DRAM was a lousy industry with too many competitors selling an undifferentiated product, often below cost. In the first quarter of 2001 when the shares were trading in the low $40s we wrote:
MU is valued at 6.5x current run-rate revenues and, today, generates no profits. In its best year ever (fiscal 2000), MU recorded $2.52 per share of earnings, making the current price 17x the peak earnings of a cyclical, commodity manufacturer. In the previous two years, MU lost money.
At the time, the valuation was kept aloft by the hopes and dreams of sell-side analysts. In our next letter we shared the following anecdote:
In an exchange of e-mails with a leading sell-side analyst who recommends purchase of MU with a $70 per share target, we solicited his justification for the current $24 billion market capitalization (let alone the $40 billion suggested by his target.) Our analyst friend explained he tried to use cyclical valuation methodologies to come up with a rationale for buying the stock but failed because such an approach suggests "the stock should trade in the teens." However, he maintains, should we have a good pricing environment next year, "people will treat the stock the same way [as they have] and take it much higher than they should." Lest we be unclear about his raison d'être, he added he "could just perennially stamp an underperform on MU because he can't justify the $24 billion, but that would be boring." He need not worry; we are fans of boring.
This sort of unchecked cheerleading among sell-side analysts is by no means gone. Today, they spin different fables to justify otherwise inexplicable valuations for the latest flavor-of-the-month stocks. As for MU, a decade of poor results exposed every flaw in the business and killed any love for the stock. The sell-side groupthink has reversed: the mostly bearish analysts now contort themselves to justify earnings estimates that are too low, price targets that are too pessimistic, and stock ratings that are too negative.
We established a position in MU at an average price of $16.49, marking the first time we have taken a long position in a company in which we once had a material short position. The industry has changed and so has MU. Its purchase of Elpida Memory out of bankruptcy in August 2013 marks the end of a decade of consolidation from roughly a dozen major DRAM players down to just three. Technological advances and locked-up intellectual property have made it unlikely that any new players will enter the industry in the intermediate term. MU and its competitors have signaled that they will refrain from adding capacity and will instead prioritize economic value-add. For the first time in memory, MU intends to use its excess cash flow to shrink the outstanding share count rather than build new factories. We believe the company will approach $4 per share of earnings and free cash flow in calendar 2014, and should enjoy a better multiple as investors begin to appreciate the new dynamic. The shares ended the quarter at $21.75.
Â
From David Einhorn (Trades, Portfolio)'s fourth quarter 2013 commentary.
