Bringing you live news and features since 2013
Bringing you news, views and analysis since 2013


“Are you lonesome tonight?”


Our regular contributor, Randeep Grewal, calls on Elvis to examine so-called lonesome stocks and industries.

Are you lonesome tonight,
Do you miss me tonight?
Are you sorry we drifted apart?
Does your memory stray to a brighter sunny day?

Elvis Presley, 1 Nov 1960

Elvis Presley recorded ‘Are you lonesome tonight’ at the suggestion of his manager Colonel Tom Parker. The original song was written by Roy Turk and Lou Handman in 1926 and was a favourite of Marie Mott, the Colonel’s wife.

Recently I have been looking at ‘lonesome’ stocks and industries…

There are a number of industries which operate on a global basis but whose competitors in any one geographical market are limited. However much investment banks and brokerages claim to operate on a global basis, it is rare that analysts covering a sector really coordinate at an in-depth level across regions. As a result of limited, or indeed no comparable, local players these ‘lonesome’ stocks tend to be under-covered. (It is much more profitable for an analyst to cover an industry with dozens of players in his region as knowledge can be amortised over much more trading volume).

I thought I would illustrate, with an industry that is generally loathed, how digging deep can find some interesting attributes for investors in a lonesome industry.

The memory industry is an example of a lonesome industry: this comprises two main subsectors – DRAM which is used within computers for storing code and data for immediate execution; and flash memory which is used for longer term storage. Though the two types of memory are different, there is the ability for the large manufacturers to switch production lines, albeit with some considerable effort, between the two; thus investors need to consider both subsectors together.

In DRAM there are three players (Samsung, SK Hynix and Micron) who, between them, control over 90 per cent of the market. In contrast the flash memory market has six players (Samsung, Toshiba, Sandisk (part of Western Digital), Micron, SK Hynix and Intel) who control over 90 per cent of the market. It might appear that a US analyst would have a chance of understanding the flash industry, but for both US players (Western Digital and Intel) flash memory is only part of a much bigger business. The list of flash makers significantly understates the market consolidation as Sandisk and Toshiba manufacture via a series of joint ventures (known as ‘Flash Ventures’); similarly Micron and Intel have a joint venture (IMFT). 

Thus a diligent memory analyst in Korea, Japan or the USA in a regionally organised brokerage is likely to cover only part of the memory industry.

The memory industry is the archetypal highly cyclical commodity producer – at least in the minds of many investors; with low barriers to entry, intense competition and massive capital investment. The history of the industry, with multiple bankruptcies or closures, and poor capital returns also scars many investors and again contributes to its ‘lonesome’ status.

One of the reasons I look for ‘lonesome’ industries and companies is that there is a greater opportunity for a variant perception if the market view is based on impressions rather than analysis. For instance, industries tend to be cyclical if there are numerous players and the principal requirement to enter the industry is capital. As shown above, the memory industry is significantly consolidated; additionally patents and know-how required to manufacture memory are a considerable barrier to entry.

Micron was started by three design engineers in the basement of a dental practice; to keep costs down their first fab was based on the plans of a supermarket (the meat locker on the plan was enlarged and made into a clean room). The first fab was built on a ranch to reduce the land costs. As the site was short of water, one of the local businessmen who had invested used a witching stick to find a gusher. That first fab was built for USD7 million; and one of their major investors was a potato farmer.

It is unlikely that in the current day a new startup in memory would be competitive or obtain the necessary scale without spending over USD10 billion and there would still be the issue of obtaining the intellectual property and the knowhow to build competitive chips. The (Flash Ventures) Yokkaichi fab alone produces 50 variants of flash and uses 20,000 different processes in its manufacturing.

Another example of investor ‘impressions’ is that historically the memory market was highly dependent on PC sales; and the purchasing power of a limited number of PC manufacturers versus numerous memory suppliers ensured that the suppliers always had a weak hand. Nowadays, however, the limited number of suppliers, and the diverse areas of demand for memory – including mobile handsets, smart televisions, cloud infrastructure providers and automotive manufacturers – suggest that bargaining power has shifted to the suppliers.

The massive consolidation in the industry means that there is now a single site (Yokkaichi) which produces one third of the entire global flash supply. In such a situation, where three to five companies are dominant in the whole industry, one might wonder if cyclicality is going to be reduced going forward.

Investors will delight in discussing the latest variant of the iPhone. However they are often less interested in the various product lines in a ‘lonesome’ industry. Over time, DRAM has become specialised for specific applications, or designed for certain customers’ requirements (eg for specific mobile phone manufacturers); also, certain customers have very exacting standards (eg automakers require suppliers to be qualified). Thus, according to some estimates, only 25 per cent of DRAM is pure ‘commodity’ chips which are headed for the desktop PC market. The increased cost of switching helps reduce the commodity nature of the industry. I like finding industries whose products, over time, become more specialised and differentiated.

Generally, silicon chips benefit from scaling from Moore’s law – and this means that flash memory, in the form of Solid State Drives (SSDs), is penetrating the laptop PC market. Indeed, by the end of 2017 or early 2018 it is likely that over half of new laptops shipped will have SSDs rather than hard disk drives. Similarly, there are some server or cloud based applications where it is already more economic on a total cost of ownership basis to swap out hard disk drives for SSDs. As prices fall the addressable market increases rapidly. For instance if price was not an issue, who would not want a 2 TB SSD in a laptop; and 1TB of storage in a mobile phone? It is interesting that though mobile phone companies (especially Apple) are covered by enthusiastic analysts, the PC industry is another ‘lonesome’ industry (partially because there is only one major US listed PC manufacturer – HP Inc.) Often when the ‘customers’ of a lonesome industry are themselves poorly covered, investors miss significant trends such as the scale of replacement of hard disk drives by SSDs in laptops.

I have attended analyst meetings where there has been discussion about the most obscure items to do with eg Apple. However I have not met an analyst who will discuss with me if the new Intel mobile processors in the Kabylake (and subsequently in the Cannonlake) processor families will lead to adoption of LPDDR4 and hence allow laptops with more than 8-16GB of DRAM which would attract power users. Though it may be an obscure issue, the point is that potential future drivers are often neglected as far as lonesome stocks are concerned.

Historically memory was manufactured in 2D; currently the industry is undergoing a transition to 3D – with the leading edge chips comprising 64 layers (SK Hynix is working on a 72 layer process). In the short term this transition has also tightened supply until the process is perfected. Such transitions for a lonesome stock can be very profound – as it potentially changes the profitability for the whole industry but is poorly covered.

All the above suggests a ‘brighter sunny day’ for memory; however it would be remiss of me, or any analyst, not to also identify that there are a few potential clouds on the horizon. 

Firstly, will the transition to 64 layer 3D chips lead to oversupply? Secondly, there are significant questions over Toshiba as a result of its nuclear division; but for investors in the memory industry, this raises the issue as to whether the ‘New Fab 2’ will be on time or delayed and the impact on other players. Thirdly Intel and Micron have been working on a new memory technology known as 3D X-Point, which apparently has hybrid features of both DRAM and NAND; however, it is not entirely clear when it will be widely available or how it will fit into memory architectures. Fourthly Chinese state-owned Tsinghua Unigroup is planning on spending up to USD70 billion to enter the memory market. Their first significant production is expected in 2018 but it will take several years to ramp up; and it is not entirely clear whether they have the knowhow or access to the intellectual property portfolio to give them freedom to operate.

Interestingly RCA and the Colonel were unsure what to make of Elvis’ rendition of ‘Are You Lonesome Tonight’ and held off releasing it for several months. Yet when it was released on 1 Nov 1960 there were 900,000 orders in the first week alone. Nowadays it is considered one of his greatest works after his return from the army.

Of course, music (together with photos and video) is one of the big drivers of increasing demand for memory.

Similarly, despite the above listed concerns it is likely that 2017 will be considered a vintage year in the memory industry with supply tightness and growing demand leading to supra-normal profits. Hopefully, it will also be a profitable year for investments in lonesome industries. n

The Trium Multi-Strategy Fund had positions in some of the stocks mentioned at the time of writing. These positions may change without further notice.

Latest News

According to the latest ESG data from PwC Luxembourg finds that investment flows towards EU..
Solactive and private equity data provider CEPRES have established a new partnership for to introduce..
New research published today by the CFA Institute Research and Policy Centre analyses the many..

Related Articles

Pension funds
UK defined benefit (DB) pension plan sponsors could have access to GBP 1.2 trillion in surplus assets over the next decade, industry research reveals...
UK defined benefit (DB) pension plan sponsors could have access to GBP 1.2 trillion in surplus assets over the next..
Tim Crawmer, Payden & Rygel
Tim Crawmer and Frasat Shah of Payden & Rygel write that higher yields are attracting more demand from investors. Also, given that equities had a strong year last year, big funds have taken some chips off the table in equities and put them into fixed income...
Tim Crawmer and Frasat Shah of Payden & Rygel write that higher yields are attracting more demand from investors. Also,..
Lady justice
Top marks for the Pensions Regulator (TPR) whose efforts to improve resilience in the UK pension funds’ liability-driven investment (LDI) strategies received glowing commendations from the Bank of England in its March report...
Top marks for the Pensions Regulator (TPR) whose efforts to improve resilience in the UK pension funds’ liability-driven investment (LDI)..
Pension funds
Four potential operators of pensions dashboards (Just Group, Legal & General, Moneyhub and Standard Life, part of Phoenix Group) are coming together to instigate a new industry coalition...
Four potential operators of pensions dashboards (Just Group, Legal & General, Moneyhub and Standard Life, part of Phoenix Group) are..
Subscribe to the Institutional Asset Manager newsletter

Subscribe for access to our weekly newsletter, newsletter archive, updates on the site and exclusive email content.

Marketing by