As Micron Technology (MU +9.01%) continues to slide, now down 46% year-to-date, value investors are licking their chops, pointing to the strong cash flow and cheap valuation metrics. But these investors are either failing to look ahead and incorporate the cyclical nature of the business in their analysis, or they are erroneously believing this time is different.
Why the memory industry is like the oil industry
What do the DRAM and NAND memory makers have in common with oil producers? They both produce commodities. DRAM, the working memory for computers and mobile devices, is Micron's largest business, making up approximately two-thirds of revenue. The other third of Micron's business is NAND flash memory, the type of memory used in mobile devices as well as solid-state hard drives, which are becoming popular in laptops and other computers.
Both of these types of memory require expensive equipment to produce and highly developed technical knowledge. But they're still commodities and virtually identical to what the other big memory producers make. That means prices of DRAM and NAND are driven by industry supply and demand, and are just as prone to cycles as the oil market.
Why the cyclicality?
When demand is high, as we saw when demand for smartphones and tablets was exploding, and the industry is undersupplied, prices remain high, creating huge gross margins and high profitability for all of the industry players. But as prices remain high, the memory producers have an incentive to ramp up supply by running their fabrication plants at higher capacity or buying more equipment.
As supply ramps up or demand dampens, the industry becomes oversupplied and prices drop. Additionally, because of the rapid change in technology, inventory can't be held for too long because of the risk of obsolescence. It then becomes a race to sell as much product as possible while the prices are still above cost. But eventually prices decline so much that gross margins shrink, and it's not uncommon for operating margins to go severely negative, as Micron experienced in 2008, 2009 and 2012.
Eventually, capital expenditures on equipment is reduced and plants are run at lower capacities. This scenario reduces supply, and prices start to improve. The cycle then repeats.
Why doesn't the industry learn its lesson?
One reason the industry can't do much about this cycle is that future supply and demand isn't known ahead of time. Memory producers must hazard a best guess as to what demand and prices will be in the future. For example, they're fairly certain demand will be higher, so they order more equipment and build new fabrication plants. But they won't know if their judgment was correct until a year or more later, once the plant is complete. By then it's too late if it turns out they were wrong.
Second, there's the element of what economists call "game theory," which is the study of how players make decisions based on what they believe their competitor will choose. As with other commodities, producers would all be better off if they restrained supply, keeping prices high. But if they did so, each producer would have an incentive to then increase its own supply to take advantage of the high prices. Once the suppliers recognize this to be the case, they realize they'll only be losing market share if they don't increase their supply as well. In the end, they all choose to produce more than the optimal amount.
This time is not different
Bulls for Micron believe this time is different because the industry has been consolidated to three major players. But economics tells us that even with two or three players the incentives remain the same, and it's unlikely the cycle will be eliminated. Others believe the shift to mobile DRAM or 3D NAND will help differentiate their products, but there's little evidence to support that so far.
It's therefore erroneous to be looking at Micron's current cash flow and valuation metrics and judging the stock as "cheap" or a deep-value buy, because that's looking backwards at its recent and most profitable phase of the cycle. Yes, Micron is currently trading around a low price-to-cash flow ratio of 7, but at the bottom of its last cycle it got as low as 2 to 3 times price-to-cash flow.
So how should investors evaluate Micron? The key is to monitor the big-picture cycle of the memory industry, just as an energy investor would monitor the price of oil. However, the prices of DRAM and NAND aren't readily available the way the price of oil is, and knowing the prices usually requires a subscription to a data service such as Bloomberg, or an industry-specific source such as DRAMeXchange.
However, individual and retail investors can usually garner the trend of memory prices from various news sites. For example, DigiTimes recently featured an article analyzing the latest moves in DRAM pricing. According to the article, DRAM prices are expected to continue to fall through the end of 2015. The article also shows how Micron's stock price of is highly correlated to memory prices.
It appears the memory industry is still completing a regular cycle, and I expect Micron's margins will contract in the next few quarters as memory prices continue to fall. Also, unlike many oil companies, Micron doesn't pay a dividend while you wait for the cycle to reverse. It's therefore best for investors try to find other opportunities and stay away from Micron for the time.
