[ad_1]
Last year, I published a detailed bearish outlook for the semiconductor sector, regarding the ETF (NASDAQ:SOXX), in “SOXX: The Global Semiconductor Shortage Is Not A Buying Opportunity.” While the ETF rose initially after that article was published, it is now down by around 28%, following a staggering 43% decline this year. In 2021, few analysts had a sharply negative outlook regarding the semiconductor sector. Chief reasons for my view included overestimating the 5G or IoT growth potential, and the negative impact higher chip prices would have on consumer demand. The sector also faced overvaluation and significant potential for an economic growth reversal.
In 2022, much of that thesis has come to fruition. The economic growth outlook has declined considerably, with semiconductor sales turning over following a sharp rise in production. While global production is still weak, declines in demand have led to immense growth in most chipmakers’ inventory levels, signaling an end to the semiconductor shortage. Those most highly valued semiconductor stocks, such as Nvidia (NVDA) and AMD (AMD), have undergone massive revaluation, falling around 60% this year, as growth expectations are re-assessed.
The semiconductor stocks in SOXX have changed dramatically since last year. The ETF now posts a much lower weighted-average “P/E” valuation of 15X. Most investors’ growth expectations have moderated as the reality of economic strain hampers chip demand. Additionally, US regulations such as the recent CHIPS ACT seeks to improve US chip manufacturing potential through industry subsidies. SOXX has lost nearly half its peak value and is trading at almost the same price it was two years ago. With this in mind, many investors are likely looking to purchase SOXX or its constituents as a “discount opportunity.” Of course, it is often futile in markets to try to catch falling knives.
The Economic Situation Facing Semiconductors
The semiconductor industry has had a turbulent path since 2018. At that time, semiconductor sales and production boomed as prices reached record-low levels, arguably creating a market glut. The glut quickly vanished in 2020 as global lockdowns forced extreme declines in the worldwide production of chips and, more importantly, the output of chipmaking equipment. While many foundries restored production from late 2020 to 2021, the lack of chipmaking equipment slowed these efforts, creating a persistent decline in global output.
Semiconductor prices have declined dramatically over past decades as innovation and capacity growth have led to lower production costs. This trend arguably broke in 2020 as the sharp decline in output forced sales lower while demand remained high, leading to a material rise in semiconductor prices. See below:
Since 2021, higher semiconductor prices have reversed gains in US manufacturer capacity utilization and, more recently, a decline in global sales. While higher prices have likely hampered sales, waning total economic business activity is likely to be the primary culprit. The US manufacturing PMI, a leading indicator of GDP growth, has fallen sharply this year. Semiconductor capacity utilization is historically correlated to the index; both have fallen together this year. See below:
The manufacturing PMI is declining for many reasons, including weak consumer sentiment, an inverted yield curve, and falling real wages. As consumer prices have risen sharply in the US and worldwide, many people and businesses are forced to reduce spending on non-necessary items. In most cases, semiconductors are not an essential good for most people as one can often delay the purchase of a new phone, computer, or vehicle until economic conditions improve.
Indeed, we must consider the possibility that the semiconductor industry is fundamentally shifting. For decades, it has benefited from “Moore’s law,” which suggested semiconductors would become exponentially more advanced and cheaper over time. However, in recent years, leading innovators such as Intel (INTC) and Nvidia have pushed the atomic boundary of potential semiconductor density, meaning cooling needs are becoming far more extreme for any continued density growth. Moreover, computers, smartphones, and related technology are becoming ubiquitous worldwide, hampering their growth potential.
In my view, the industry follows the same growth pattern as steel in the 1900s – rapid growth and adoption, followed by perfection, maturity, and nearly stagnant demand. Fundamentally, the technology cannot be improved much more without exacerbating energy needs, and it seems most people and businesses do not need even faster processors. Combined with the rise in prices and the decline in general economic growth (a “recession” under traditional measures), the immediate and long-term outlook for the industry appears lackluster. This is not to say that demand for chips will decline substantially, but that investors should not endanger growth given shifting economic patterns and growing constraints within semiconductor production and innovation.
A Closer Look At SOXX’s Holdings
The “P/E” valuations of most of the firms within SOXX have declined since I covered the fund last. That said, SOXX is still up by around 85% over the past five years, and many of its holdings may see EPS decline over the coming year. This negative trend is relatively straightforward by looking at the sales growth levels of the top ten firms in the ETF compared with the ISM PMI index and semiconductor manufacturer capacity utilization. See below:
The sales growth rate of these ten firms is tightly correlated to the ISM manufacturing PMI and semiconductor manufacturing capacity utilization. The sales growth rate of these ten firms is 25% YoY, but I expect it will decline toward zero over the coming quarters, as implied by the negative trends in the economic indicators. Few signs indicate a bottom in the GDP’s decline as persistently high inflation combined with rising interest rates fuels more significant pressure on economic demand. With this in mind, I believe we will continue to see deterioration in macroeconomic demand for semiconductors as households and companies avoid or delay large purchases.
At the same time, these companies have seen significant, and often extreme, growth in inventory levels. See below:
Nvidia’s inventory growth is the highest over recent years and signals risk in the company’s sales prices. The company has stated that its excess inventory of certain products has dragged down its balance sheet this year. While Nvidia’s situation is the most extreme, other large semiconductor companies appear to be in a similar condition. If production levels generally remain high while demand continues to wane, further inventory buildup will likely push the market into a deep glut, causing prices to reverse sharply.
Overall, I believe there are clear indications that most companies in SOXX will see EPS decline over the coming 12-24 months as the market cycles back into glut dynamics. Given growing signs of macroeconomic deterioration in the US and other key markets like China and Europe, I suspect semiconductor demand will wane for a very prolonged period – likely longer than it did after 2008. Key reasons for this include the extent to which inflation harms economic demand and the apparent end to secular growth within the semiconductor industry.
The Bottom Line
While it is true that the semiconductor industry has grown constantly throughout almost all people’s lives who are alive today, I see little reason it should continue to grow over the coming decade. Semiconductor technology can improve further, but not nearly as it did in decades past and at a much higher price in my opinion. There may be demand for improved chips in specific segments, such as gaming, data centers, and the sciences. Still, for the bulk of chip uses (computers, smartphones, vehicles, etc.), it seems current processors achieve the necessary functionality. As exemplified in bitcoin mining, rising electricity costs and a weak energy supply may also lower demand growth for high-performance semiconductors. Overall, I believe this situation suggests SOXX should not be given the considerable “growth” valuation premium it has held in years past.
Most stocks historically trade at a “P/E” of 15X, with valuations higher when real interest rates are low and valuations lower when real interest rates are high. Real interest rates are slightly high today, dragging down the “fair” valuation multiples. Further, I suspect many stocks in SOXX will see their EPS decline over the coming one to two years, particularly if we adjust for inflation. With this in mind, I would personally not be willing to buy SOXX until its TTM weighted-average “P/E” is in the 10-12X range, as I give no long-term growth premium to the stocks. This valuation equates to a price range for SOXX of $205 to $250 or 20% to 33% lower than it is today. That said, I would not bet against the ETF as it does not appear as overvalued as the technology sector at large.
[ad_2]
Source links Google News