The US Government’s Semiconductor Solutions Seeking Problems

Editors’ Note: MarketMinder doesn’t make individual security recommendations. Any reference to an individual firm herein is solely to help illustrate a broader point. Additionally, MarketMinder favors no politician nor any political party, assessing developments solely for their economic and market impact, or lack thereof.

How do you know when governments are concerned about their popularity and staying power? They start trying to “fix” things. We saw that last week, when the governments of the US, UK, Japan, South Korea, India and China announced a coordinated release of strategic petroleum reserves in the name of taming oil and gas prices. Now, in the UK, energy regulator Ofgem is trying to “fix” the problem of power companies folding like dominoes by regulating them like banks, with capital requirements and tougher stress tests—never mind that these companies are failing because retail electricity prices are capped while wholesale prices aren’t, forcing them to operate at a loss. Last but not least, the Biden administration is peeved at Congress not because the clock is ticking on a potential government shutdown and a debt ceiling can-kick, but because the House has yet to pass a bill aimed at boosting semiconductor production and helping the US compete economically with China. We don’t think this bill has much market impact for good or ill, but we do think it is worth a look, as its fecklessness speaks volumes about how strong the semiconductor industry is.

The bill in question is the US Innovation and Competition Act (USICA), which contains a formerly standalone bill known as the Creating Helpful Incentives to Produce Semiconductors for America Act—CHIPS for short.^[i] The Senate passed them over the summer, but the House hasn’t taken them up, and the administration’s calls to do so are growing louder. The larger USICA is classic industrial policy of European proportions, built on the presumption that the mighty US private sector can’t compete with state-directed investment and five-year plans in China. I can’t help but laugh my head off at this, as I am old enough to remember when everyone said the same thing about Japan beating the US in the late 1980s. You might remember that thesis fizzled when the Nikkei bubble popped, the result of years of misdirected top-down investments that didn’t pan out. Not that China is on the verge of a similar bust, but it is a fallacy to presume that a heavy government hand in technological development is a sign of strength. Free markets and private-sector innovation and investment are the Goose That Laid the Golden Egg, in my view, and protecting them is the key.

But that is a tough sell politically these days. There is a shortage of chips globally, and it is hitting automakers particularly hard after they guessed—wrongly—that COVID would kill demand for new cars and canceled chip orders. Other industries quickly took their place at the feeding trough, leaving car companies scrambling. Meanwhile, traffic jams at ports are making it hard for overseas chip foundries to fulfill orders here in a timely fashion. But telling voters these issues are complex and outside governments’ control can be a nonstarter when inflation is up and midterms loom. So cue the government, umm, help.

That is where the CHIPS Act comes in. If passed in its present form, it would aim $52 billion of government funds at domestic semiconductor production, mostly in the form of tax breaks and subsidies … which would go to some of the world’s largest, most cash-rich, profitable companies. It would also establish a National Semiconductor Technology Center, which—as best I can tell—would be a state-run think tank and job creation program for bureaucrats.

It rather reminds me (and many others) of the last time the government tried to help the US win a global semiconductor arms race: the mid-1980s, when the US government partnered with over a dozen chip manufacturers to form the gloriously named Sematech—one of the all-time great government boondoggles. Despite buckets of money from the Department of Defense and Defense Advanced Research Projects Agency—and all the knowhow of Bob Noyce and others—it floundered. As Cypress Semiconductor founder T.J. Rodgers summed up in a delightful Wall Street Journal op-ed earlier this year: “Due to the relentless pace of progress under bare-knuckle competition, Sematech’s Austin, Texas, chip plant was obsolescent when it opened. The layoffs started in 2002.”^[ii] In the end, it was a hallmark example of inefficient government spending.

Any government-run efforts now probably go similarly, while that same bare-knuckle competition is already driving progress. Yes, the US produces a small share of the world’s semiconductors, but this isn’t a sign of US weakness—it is a sign that semiconductors are commodities. They are inexpensive and require low production costs to be profitable. The US has long since shifted from cheap manufacturing to complex, high-end production. Parallel to this, the US also leads the world in chip design. For eons, designing chips here and outsourcing the fabrication to Taiwan has just made sense for all involved—designers, fabricators and customers, who are keen to keep their own costs down.

When the cracks appeared in that business model earlier this year, companies didn’t sit around waiting for Congress. They didn’t have to! These are some of the world’s biggest growth-oriented firms, with some of the fattest gross operating profit margins in all the land. That lets them self-finance all sorts of major initiatives or, if they prefer, tap banks or capital markets for dirt-cheap funding.

Boy, have they been doing this! In March, Intel announced it would invest $20 billion in two new chip fabs in Arizona. TSMC shortly followed suit. More recently, Ford Motor decided to solve its chip woes by partnering on chip design and fabrication with GlobalFoundries. General Motors announced it is pursuing a similar strategy. Oh, and just last week, Samsung announced it will build a $17 billion chip plant near Austin. Yes, new chip foundries take years to come online, and these efforts won’t start adding to global supply until 2024 at the earliest (and some of the more cutting-edge facilities may take a decade to bear fruit). But any government-run help would probably take far longer, as always happens when political considerations, rather than pure competition, start directing traffic. Moreover, the private sector has already deployed capital far above and beyond what the government has earmarked.

In our view, the CHIPS Act is a solution seeking a problem. Yes, the chip shortage is real, but companies are already solving it on their own, with private capital.^[iii] Meanwhile, the US is already the global tech leader, and American chip design is already one of the world’s economic crown jewels. There is just no evidence the industry needs a government-run think tank to drive innovation—the profit motive is already doing a bang-up job at that. So if the CHIPS Act eventually passes, don’t get too excited. The impact on the Tech sector and companies that rely on semiconductors should be miniscule, with next to no impact on stocks. A government tech boondoggle is always entertaining,^[iv] but Sematech 2.0 won’t move the needle.