I believe semiconductor stocks are likely to decline significantly over the coming quarters for several reasons, including a) I believe most semiconductor stocks are not likely to benefit significantly from the growth of artificial intelligence (“AI”), b) I expect most semiconductor stocks to report weak earnings, c) US-China trade war risks, d) recession risk and e) high valuation levels.
To profit from this, I recommend Direxion Daily Semiconductor Bear 3X Shares ETF (NYSEARCA:SOXS) with a Buy, since it is constructed to return 3x the inverse of the daily return of the ICE Semiconductor Index. This could enable SOXS to be among the best-performing investments in a recession.
I recommend Direxion Daily Semiconductor Bear 3X Shares ETF with a Buy.
I believe a US recession is highly likely in the coming year, which makes semiconductor stocks susceptible to significant declines. This is particularly true given my belief that most semiconductor stocks are not likely to benefit significantly from AI and are likely to have weak earnings, as well as rising concerns about US-China trade war risks, recession risk and high valuation levels.
I believe one investment that could provide some of the highest returns possible in a recession is SOXS, since it is designed to produce three times the opposite of the daily return of the semiconductor index. Thus, if the ICE Semiconductor Index falls at least 30%, SOXS could double or more in price!
I believe buying an inverse ETF like SOXS is better than shorting an ETF like the iShares Semiconductor ETF (SOXX), since it requires no margin account, which means no borrowing, no interest expense and no margin calls. It is also more intuitive, since investors generate profits when SOXS rises in price and incur losses when it falls in price, just like with traditional stocks and ETFs.
One key issue to understand before buying any levered ETF such as SOXS is that they often underperform their benchmark by the leverage factor, which in the case of SOXS is -3x. This is primarily due to “beta slippage”, which occurs during periods of high daily volatility. For example, if the Index rises 10% one day and falls 10% the next day, it will be down 1% [(1 + 0.10) x (1 – 0.10) = 0.99] over that two-day period. But if a -3x levered ETF like SOXS rises 30% the first day and falls 30% the second day, it will be down 9% [(1 + 0.30) x (1 – 0.30) = 0.91] over that two-day period, instead of the 3% one would expect it to fall. This is due to the mechanics of daily leverage pricing.
The Direxion Daily Semiconductor Bear 3X Shares ETF seeks three times the inverse of the performance of the ICE Semiconductor Index (“the Index”) on a daily basis. That means if the Index rises 1% on a given day, then SOXS should fall about 3%. And if the Index falls 1% on a given day, then SOXS should rise about 3%.
SOXS achieves this daily performance goal primarily by using swap agreements with large financial institutions such as JP Morgan, Bank of America, Citibank and UBS, as well as futures contracts and other financial instruments.
Due to the compounding of daily returns, this triple inverse ETF is not likely to generate three times the inverse of the return of the Index cumulative return for periods lasting more than one day. In low-volatility markets that are primarily trending in one direction, the cumulative return for longer periods will typically be higher than -3x the return of the Index. But for high volatility markets without a strong trend, the cumulative return for longer periods will typically be lower than -3xthe return of the Index, as mentioned above.
SOXS has total assets of $1.05 billion. But it is larger than the $6.5 million competing ProShares UltraShort Semiconductors ETF (SSG). Note that SSG is a -2x levered ETF, which is less leverage than SOXS at -3x. SOXS average daily dollar volume is $621 million, which is much higher than the $0.8 million of SSG. The expense ratio is relatively high at 1.02%, which is higher than 0.95% expense ratio of SSG. The average bid-ask spread is relatively low at 0.09%, but it is lower than the 0.48% spread of SSG.
The ICE Semiconductor Index is comprised of 30 leading US semiconductor stocks. The top 10 stocks comprise about 57% of the Index and include companies such as NVIDIA, Broadcom, Advanced Micro Devices, Intel and Texas Instruments.
It is not possible to directly invest in the Index, but one can invest in traditional semiconductor ETFs such as the iShares Semiconductor ETF (SOXX).
High Risk In Semiconductor Stocks
I believe semiconductor stocks are at high risk now for five key reasons:
- I think most semiconductor stocks are not likely to benefit significantly from the growth of AI
- I expect most semiconductor stocks to have weak earnings
- US-China trade war risks
- recession risk
- high valuation levels
AI Mostly Benefits NVIDIA, Not The Industry
After falling nearly 50% from January to October 2022, SOXX has rallied an astounding 84% and is nearly back to its all-time highs. What has driven this dramatic turn of events?
I believe it has primarily been due to investor exciting over the rapid growth of AI technology, particularly the widespread adoption of ChatGPT and other new AI technologies. Leading tech companies like Alphabet, Microsoft, Amazon and Meta are very focused on AI, which requires lots of computing power and advanced microchips.
AI is a clear benefit for industry leader NVIDIA. As this article notes, NVIDIA “dominates the AI computing market with a market share of 80% to 95%, according to analysts.” In addition:
Nvidia has few competitors working at a large scale. While Intel Corp (INTC) and several startups such as Cerebras Systems and SambaNova Systems have competing products, Nvidia’s biggest sales threat so far is the internal chip efforts at Alphabet Inc’s Google (GOOGL) and Amazon.com’s (AMZN) cloud unit, both of which rent their custom chips to outside developers.
Nvidia’s lead there has come not only from its chips, but also from more than a decade of providing software tools to AI researchers and learning to anticipate what they will need in chips that take years to design.”
Unfortunately, NVDA is only about 8% of the semiconductor Index, so I believe the overwhelming majority of the Index is not likely to benefit significantly from the growth of AI.
Weak Industry Earnings Outlook
While AI is expected to drive incredibly strong EPS growth of nearly 140% for NVIDIA this year, I expect most other semiconductor stocks to have much more modest — or even declining — earnings growth.
For example, Wall Street analysts expect only three of the other top 10 stocks in the Index to grow earnings this year and their growth is likely to be much more modest. Expected EPS growth this year is +12% for Broadcom (AVGO), +10% for Analog Devices (ADI) and only +5% for Microchip Technology (MCHP).
The other six of the top 10 stocks are expected to report declining earnings this year, according to Wall Street analysts. EPS is expected to decline -77% at Intel (INTC), -23% at Texas Instruments (TXN), -18% at Advanced Micro Devices (AMD), -17% at KLA Corporation (KLAC), -9% at ON Semiconductor (ON) and -4% at NXP Semiconductor (NXPI).
Why buy the Index, which has such weak growth prospects, when you can just buy NVIDIA?
Of course, I believe one problem with buying NVIDIA now is its extremely high valuation, which I will discuss later.
US-China Trade Wars
Leading chip companies such as NVIDIA, Broadcom, Intel and AMD derive 20% or more of their revenues from China.
Unfortunately, the US government is considering limiting exports of AI chips to China as part of the ongoing US-China trade war. According to NVIDIA’s CFO, “Over the long-term, restrictions prohibiting the sale of our datacenter GPUs to China, if implemented, would result in a permanent loss of opportunities for US industry to compete and lead in one of the world’s largest markets and impact on our future business and financial results”.
In response, China is planning to limit exports of gallium and germanium, which are metals used to manufacture semiconductors and other advanced electronics. Since China produces 80% of the gallium and 60% of the germanium in the world, this would likely drive up production costs and lower profits for semiconductor manufacturers.
Another major wildcard for the semiconductor industry is the risk of a Chinese invasion of Taiwan, which could trigger a US-China war. This would have a significantly negative impact on the semiconductor industry (among other issues!), since Taiwan manufactures more than 60% of the world’s semiconductors, including more than 90% of the most advanced chips.
Semiconductor sales tend to fall sharply in a recession. As the chart below shows, chip sales fell over 40% in the early 2000s recession and over 30% in the 2008-2009 recession. Chip sales have already fallen 21% over the past year, before a recession has even started.
Based on various leading economic indicators, including the rapid rise short-term interest rates, falling money supply and inverted yield curve, I believe the US economy is highly likely to enter into a recession in the coming year.
One proven leading economic indicator is The Conference Board’s Leading Economic Index (“LEI”). It is highly regarded for its ability to predict recessions and recoveries. The following chart shows The Conference Board’s LEI has fallen to levels only seen in recessions.
According to Justyna Zabinska-La Monica, Senior Manager of Business Cycle Indicators at The Conference Board:
“The US LEI fell again in June, fueled by gloomier consumer expectations, weaker new orders, an increased number of initial claims for unemployment, and a reduction in housing construction. The Leading Index has been in decline for fifteen months—the longest streak of consecutive decreases since 2007-08, during the runup to the Great Recession. Taken together, June’s data suggests economic activity will continue to decelerate in the months ahead. We forecast that the US economy is likely to be in recession from Q3 2023 to Q1 2024. Elevated prices, tighter monetary policy, harder-to-get credit, and reduced government spending are poised to dampen economic growth further.”
While many economists and business cycle experts expect a recession, I do not believe that is priced into the market yet. Typically before a recession starts and even many months after a recession has started, most investors are not aware of the risks of a recession, since they tend to focus more on coincident or lagging indicators such as GDP and employment, as well as historical earnings reports, rather than leading economic indicators. It is usually only after unemployment starts rising materially and companies start missing or guiding down revenues significantly that investors begin to panic and price in recession risks.
It is a fact of life that stocks usually fall significantly during a recession. Since the late 1920s, there have been 12 recessions. The median decline of the S&P 500 during those recessions was -39%.
Thus, in a recession, I believe stocks are likely to fall significantly, even if the Fed is cutting rate. Recall that in most recessions, including the early 2000s and 2008-2009 recession, the Fed was cutting interest rates aggressively before the recessions even started.
I expect semiconductor stocks to fall even more than the S&P 500 during this recession, for the reasons cited above. Thus, I believe SOXS should be one of the best-performing investments if there is a US recession, since it could provide a positive return of about triple the decline in semiconductor stocks. It will be hard for any other investment to beat that in a recession, particularly since most investments will likely be declining.
As shown below, SOXX fell significantly and underperformed the S&P 500 in prior recessions.
During the covid stock market panic of February and March 2020, SOXX fell -38%, worse than the S&P 500’s -35% decline. SOXS rose +125% then. That was -3.3x the return of SOXX. In 2022, SOXX fell -48% from early January to mid-October, worse than the S&P 500’s -28% drop. SOXS rose +187% during that time, which was -3.9x the return of the SOXX.
Of course, with -3x levered ETFs, relative performance can vary significantly, particularly when prices are highly volatile. For example, SOXX fell nearly 5% from the end of March 2022 to the end of March 2023, but SOXS fell 56%, which is +11x the return of SOXX and not the -3x expected! Thus, it is very important for investors to recognize that SOXS performance can stray significantly from what is intended over medium to long periods of time.
High Stock Valuations
Investor excitement over the potential growth of artificial intelligence technology has driven semiconductor stocks to very high valuation levels, which I believe causes higher potential downside risk in the event of a recession.
The dividend yield of semiconductor stocks, as represented by SOXX, is only 0.88%. That is 38% lower than the S&P 500 dividend yield of 1.43%. The P/E ratio of SOXX is 27.6x, 30% higher than the S&P 500 P/E ratio of 21.2x.
NVIDIA’s (NVDA) forward P/E is very high at 57x. The chart below shows NVIDIA’s (NVDA) EV/EBITDA ratio is currently much higher than its historical levels.
Intel’s (INTC) forward P/E is also very high at 83x. Intel’s (INTC) EV/EBITDA ratio is also much higher than its historical levels.
Valuation levels for any investment are primarily useful for determining long-term return potential and are usually not indicative of likely short-term returns. However, I believe they are very relevant for stocks in a recession, since valuations tend to fall significantly when fundamentals are weak.
They are only suitable for aggressive investors who can actively manage their trading positions and tolerate large losses in a short period of time. They are not suitable for “buy and hold” long-term investing, since the stocks tend to rise over time. They should only be held during periods when an investor has a strong conviction that the stocks will be declining in a bear market. I have that conviction now.
If the Index is highly volatile, the return of SOXS can be negative, even if the Index falls. And if the Index rises more than 30% in one day — which I believe is highly unlikely — then the value of SOXS could fall to $0. However, the fund manager typically positions the fund so that it is not responsive to changes in the Index of more than 90% on a given trading day. This should limit the maximum daily gain or loss of SOXS to 90%.
There are also risks related to the derivatives that SOXS invests in, as well as the counterparty risk they undertake with those derivatives. If there is a major global financial crisis, which would typically lead to gains for SOXS, there could be a risk that the financial institutions that are counterparties to SOXS swap agreements and other derivatives are not able to meet their commitments.
Beyond the risks relating to the fund, the primary risk to SOXS is if there is no recession and AI growth benefits the industry more broadly.
To mitigate those risks, investors in SOXS should keep a close eye on leading economic indicators and semiconductor industry news.
In the shorter-term, there is always the risk of sharp bear market rallies during a recession.
I recommend SOXS ETF with a Buy due to my expectations of a significant decline in semiconductor stocks due to less AI benefit than investors expect, the weak earnings outlook of most semiconductor stocks, rising concerns over the ongoing US-China trade war, recession risk and high valuation levels.
SOXS is highly volatile and carries substantial risks, which I have discussed above, but it also provides the opportunity for aggressive investors to generate among the highest returns possible in a US recession.
If you found this recommendation helpful, please let me know in the comments below.