Semiconductors: Not So Cyclical Anymore and Where to Find Them

17/2/2021

This week we continue with the thematic of mobility and look at another adjacent category, semiconductors. In particular where we see the market going and why it remains an important aspect to look at especially when one thinks off the EV market. For the more technically apt, please forgive if the following seems a tad layman-like, given that this author's particular forte is economics.

Context

Author: Sid Ruttala

Perhaps one of the most central aspects of modern day electronics is the use of semiconductors. Simply put, a semiconductor is any material that has an electrical conductivity between that of an insulator and that of most metals, enabled by the addition of an impurity, a process known as doping. The most important aspect of its manufacture, and one that we are more familiar with, is the integrated circuit (IC), which is crucial to everything from the computer or phone you are currently reading this article from to scanners.

So, what does this have to do with mobility? Think for a moment about EVs and their greatest hurdle, that of battery efficiency. The proliferation of the EV market has seen a resurgence in investment into new technologies, including the supply chain for the manufacture of semiconductors. An industry so convoluted that it would take me more than this short piece to explain, but below are a few key points:

As it currently stands, the industry is dominated by companies from the APAC region, namely China, Japan, South Korea, Taiwan and Singapore. This is the case due to historic reasons and the evolution of those particular nations in placing focus on manufacturing during their infancy in economic development. From the creation of the Keiretsu firms in Japan during reconstruction (helped along by the Americans) to the four Asian tigers, including South Korea, which followed the Japanese model by creating national champions in the space through the Chaebol firms.
The twenty year annual average growth rate of the sector stands at a staggering 13% per annum
The sector remains notoriously pro-cyclical and volatile due to its very nature. This should be intuitive since its revenues and fortunes are directly derived from consumer discretionary spending (electronics) and there happens to be a long leadtime from forecasting demand to the manufacture of inputs that go into final products. Remember this aspect because there might just be a catalyst via the EV market for this to change substantially.
The structure of the industry is based on a foundry model, which consists of 1) fabrication plants and 2) integrated circuit design operations. There has been intense consolidation with only three players globally currently able to manufacture the most advanced devices; Intel (INTC.NASDAQ) in the US, TSMC (2330.TPE) in Taiwan and Samsung (005930.KRX) in Taiwan. Intel is also currently considering outsourcing some of its production to TSMC.
Within this rather opaque structure there are niche players within adjacent categories that specialise in certain aspects with automobiles being dominated by STMicroeclectronics (STM.EPA), Infineon (IFX.ETR), Renesas Electronics (6723.TYO), Toshiba (6502.TYO), Texas Instruments (TXN.NASDAQ), Freescale (formally FSL.NYSE, now merged with NXP) and NXP Semiconductor (NXPI.NASDAQ).

Where to Next?

So with that context in mind, let's explore what the future might hold and why things might just be about to change as we continue in this revolution. The first point, geographic orientation and the heavy role played by APAC within the manufacture of this vital component. What began with the Trump trade war is set to continue even under the new administration. We are likely to see a relocalisation of supply chains and governments across the west and Asia promoting local manufacturing in this sensitive industry. And while it may be that, as it currently stands, many of the manufacturing units for those that dominate the space might reside in China, such as Infeneon or NXP, we are likely to see increased likelihood of government interference within the space. Don’t believe me? For those inclined, have a look at the proposed merger between QualComm and NXP; blocked by anti-trust regulators in China while allowing local players to consolidate (including state-owned enterprises). That’s not to say this is necessarily a bad thing for the investor though.

In a world where capital is able to move freely but trade isn’t, investors can take advantage of what is called rent-seeking in economics, as companies might ascertain greater profits than would be the case in a perfectly competitive market. Think Baidu and its market capitalisation, which would not have seen the success it did were it not for Alphabet being unable to compete in China. Or the proposed forced sale of TikTok to Oracle (from ByteDance) that was a result of the now previous administrations confrontation with China. If we didn’t have a change in administration Oracle may have just gotten itself a great deal, TikTok being one of the fastest growing brands globally in recent years, now with over 2bn downloads and 800m+ active users.

In a similar way we are likely to see the semiconductor industry disrupted but also in some ways advantaged by the changing geopolitical situation.

Coming to the second and third points about the growth rate and pro-cyclicality. Whether you believe in the reflation trade or not, which the current valuations are seemingly banking upon (i.e. bond yields and equities valuations), I believe we are at an inflection point. Historically, from a portfolio and asset allocators perspective, the semiconductor industry has been tied to underlying consumer demand but as we see the burgeoning of an arguably secular growth story (in many ways due to government intervention again) in the form of EV and autonomous vehicles, there is a case to be made that there may very well be a divergence from this historic correlation. Firstly, if current projections are taken as a base case then the automotive semiconductor market in particular is set to grow at circa. 17.5% p.a. between now and 2026. As nations adopt the technologies and, in many cases, mandate a shift out of fossil fuels based vehicles, this is one segment that will see continued investment as manufacturers scramble to lock in their sourcing.

In fact, the most recent signs of this has been the scramble by automakers, from the Spring of 2020 onwards, trying to lock in orders. A situation that eventuated in manufacturers sending a letter to the White House, asking the government to include “substantial funding for incentives in semiconductor manufacturing”, Biden promptly signed an executive order directing review of supply chains for critical goods. I make the call that even in the face of economic headwinds this is one industry that will receive some cushioning from policy makers, in much the same way as the 2008 crisis eventuating in the auto bailout. Another way of looking at this is that this aspect somewhat de-risks the allocation for investors should they wish to gain exposure to the space.

Coming to the fourth and fifth points, we will continue to see greater consolidation take place within the space, the most recent merger being that of Freescale and NXP. The Qualcomm merger, though shelved for now, might resurface in the future as the west tries to grapple with supply shortages. Similarly, incumbents such as Intel and Nvidia (whose niche relates to autonomy) continue to look at acquisitions to bolster their portfolio. Think Intel’s acquisition of Moovit and Nvidia currently in the process of acquiring ARM (a chip designer) for close to $40bn USD, though this has been vehemently protested by Tesla, Alphabet and Microsoft to name a few.

Best Bets?

I have previously written about NXP as the best potential bet not only because of its western footprint (de-risking somewhat geopolitically) but also on the other side with its joint venture in China with Datang Telecom. There is the risk of forced technology transfer, but we feel that this is very much in its infancy and NXP also has the European market to cater for. From an autonomy perspective Nvidia is a stellar player as always, closely followed by Intel whose acquisition of Moovit gives it an edge. For a purely automotive play, American-Irish company Aptiv (APTV.NYSE) and the Japanese company Denso Corp (6902.TYO) certainly look interesting. In terms of close seconds TSMC, given its scale and domicile in Taiwan, make it interesting. If you have issues with regards to their exports to China, I only have one question to ask, who do you think has been financing manufacturing in China for close to four decades? (Hint: it's certainly not the Cayman or Hong Kong, though the money seems to be coming from there). If it’s still not believable, feel free to type in Foxconn (2354.TPE) or Hon Hai Precision Electronics (2317.TPE), the actual manufacturer of the iPhone, it certainly is not a Chinese company. Another bet is Texas Instruments, whose edge comes from car safety in particular and is a member of ISO 26262 (safety standard in electrical and electronic systems in automobiles), in fact it leads the semiconductor subgroup within the overall working group.

On the other hand (if you equate biggest to best), Infineon is the largest having now overtaken NXP after its acquisition of Cypress.