Scientific rambling
At the heart of Biden's economic agenda is the underlying tension between industrial policy and competition policy. Efforts to reshore manufacturing can lead to the protection of national leading enterprises. Some policymakers support this for the purpose of improving efficiency, but concentrated industrial producers in the U.S. can exercise market power without dramatically increasing the "resilience" of the chain, which is the main principle cited by Biden in his reshoring plan. For other potential drawbacks, just think about Boeing's current engineering issues.
This threat hangs over the implementation of industrial policy programs, particularly the $53 billion earmarked in the CHIPS and Science Act for grants, loans, and other incentives to support U.S. semiconductor production. The companies selected by the Ministry of Commerce for these programs, along with the conditions attached, will structure the domestic market to support scale or resilience.
But so far, apart from a very general document released last year called "Vision for Success," the CHIPS Program Office (CPO) has awarded only a small grant and issued very few directives on what to decide on the department's review process. One specific proposal is to create geographic clusters of chipmakers. But recently the U.S. Commerce Secretary said more subsidies would be released within eight weeks.
With Biden eager to win before ***, a major round of subsidies is expected in the coming weeks. Recent reports suggest that the subsidy will require several billions of dollars in money that could go to two of the largest semiconductor makers: Intel and TSMC, the most dominant Taiwanese semiconductor giant in the logic chip space. TSMC holds a 57% market share of global foundry revenue, almost five times that of its closest competitor, Samsung, and an even higher profit pool of around 80%.
Intel is one of the few remaining chipmakers in the U.S., has one of the top 10 foundries in the world, and maintains an effective duopoly with AMD on CPUs, a specific niche for logic chips.
These huge allocations are somewhat unavoidable. The industry is already highly concentrated in most of the logic chip market segments, with high barriers to entry. If the goal is to fundamentally reshape the market in the short term, it is unlikely that all of this funding will be used to incubate start-ups. But the new funding still raises the question of whether the CHIPS Act will succeed in transforming the U.S. semiconductor industry, or whether it will merely amount to corporate welfare that establishes centralized power.
The timely release of the U.S. Economic Freedom Project addresses this challenge head-on and provides a roadmap for lawmakers on how to address the future of the chip market. The focus is not only on reshoring as an alternative to the ** chain shortages that the United States experienced during the pandemic, but actually on restructuring the market in order to address these shortages from the system.
* It is believed that the goal of U.S. semiconductor policy should be to break the overseas outsourcing "fabless" model favored by Wall Street and break TSMC's shackles on foundries. It also needs to "strengthen" the market for the best merchants with operations in the United States.
First, the authors distinguish between the economic structure of the frontier processing chip market and the mature node market. Leading-edge logic chips are state-of-the-art chips used for high-tech and operate in a highly concentrated market. Mature node markets continue to be the most commonly used markets for consumer electronic devices, with more fragmented markets and overproduction due to foreign dumping practices and commoditization.
Similar historical forces have driven the development of both markets. The United States led semiconductor technology in the mid-20th century, thanks in large part to R&D programs at centers such as Bell Labs, which were funded. As a result, U.S. chipmakers both provide a boost to the tech industry and must comply with competition regulations, such as open licensing requirements that allow new entrants.
At that time, chipmakers were almost vertically integrated, from chip design to production in-house.
But the agreement opens up the global economy to foreign competition. Under pressure, chipmakers both outsource production and expand by acquiring competitors, which has been made possible by the lack of antitrust enforcement and intellectual property protection reforms.
By the outbreak of the epidemic in 2020, 60% of the world's semiconductor production and 90% of the most advanced chips were concentrated in Taiwan, which brought great pressure to the ** chain.
The main legacy of this structural shift in semiconductor production is the fabless model. It explains why companies like Nvidia, Qualcomm, Broadcom and even Apple are the highest-value chip makers, even though they are not actually in the foundry business. This "capital-light" model means they only design the chips, own the intellectual property, and then outsource the actual production to overseas foundries, where it is cheaper to manufacture.
Wall Street prefers this kind of company because of the method of earning high profit margins without having to make an actual capital investment, all of which can generate unusually high returns for shareholders through dividends and buybacks.
According to the calculations, nine of the top 10 U.S. semiconductor companies, plus Apple, have spent $698 billion in investor spending over the past five years, equivalent to 14 times the total spending of chips. If businesses had been more sensible in investing their earnings in capital over the past decade, they wouldn't have had to step in.
Both ends of this fabless model are monopolized. Fabless companies are concentrated, and most of the production is carried out overseas by TSMC and a few other companies.
The number of the top U.S. chipmakers has fallen by 44% since 2010, in part due to a period of intra-industry mergers and acquisitions, most of which were approved by federal regulators, the ** found. "Both foundry and fabless companies have strong market power, so there is a real cost issue beyond the threat to resiliency from outsourced manufacturing," the authors wrote. ”
Fabless production primarily impacts the frontier logic market. In contrast, mature node chips do not have the same monopoly problem, but face bottom-to-bottom competition from excessive foreign competition.
While logic chips are highly profitable, U.S. mature node manufacturers have very small margins, so much so that in many cases the cost of investment can outweigh the returns. This is partly due to foreign dumping practices.
Due to the high level of competition, most established node producers, such as Micron or Intel, opt for more vertical integration in order to make as much profit as possible. The authors argue that while there are many issues that policymakers need to address, competition in this market sector is instructive.
A particular focus of this is that Apple, as the largest single buyer of all its iPhones and Macs, has a huge influence over the entire semiconductor market. The tech giant's exclusive arrangement with TSMC to produce all of its state-of-the-art logic chips is a prime example of how disruptive the fabless model can be.
The deal freed Apple from being a competitor's buyer and gained first-tier status thanks to TSMC's special discounted pricing, saving Apple billions of dollars. This also limits production capacity, as a large part of TSMC's production has been locked in by Apple.
While the deal has provided significant financial tailwinds for both companies, it has accelerated the offshoring of U.S. production to Taiwan and severely damaged rival chipmakers.
In recent years, more and more tech companies have followed Apple's lead and replicated this fabless model. For example, Amazon acquired an Israeli semiconductor startup, removing a promising new player from the market. As the newspaper put it, "Apple will decide the success or failure of the CHIPS Act." ”
The document lays out several specific categories of reforms to supplement CHIPS Act funding, some of which are already within the Commerce Department's implementing authority. An overarching theme is the need for the Department of Commerce to work with antitrust regulators and authorities in a holistic manner.
The document sets a central goal for the Commerce Department's reshoring efforts: to incubate at least four independent leading logic chip manufacturers to ensure a "thicker" and more resilient market that breaks down concentration. This may require strengthening Intel, Samsung, and potentially AMD to make them challengers against TSMC.
"CPOs should prioritize providing more funding to second-tier foundries," the authors suggest. ”
But to ensure that the business environment is sustainable for competitors, policymakers need to take additional action on both the supply and demand sides.
On the one hand, U.S. chipmakers need to have reliable long-term demand for their products, especially in highly volatile mature node markets.
As a market buyer, intervention may be a way to manage and maintain a stable level. As the authors point out, this arrangement is similar to decades of intervention to stabilize markets for important commodities prone to booms and busts, such as agriculture and oil.
The authors also recommend legislation to force fabless companies to dual-source from multiple foundries, creating a competitive opportunity for U.S. chipmakers. This could break the monopoly of Apple and TSMC on the market. Another potential remedy to stop the fabless model is to require the opening up of IP licensing practices, as federal authorities have done in the past, and to prevent patent abuse. A buyback limit based on the floating size associated with the investment may also be beneficial, and the Department of Commerce can impose such a limit on recipients of CHIPS Act funds.
To further discourage offshoring, the authors suggest a series of actions that incentivize both U.S. production on the ** side and on the demand side to buy U.S.-made chips.
Another related solution would be to impose high fees on semiconductor buyers if these companies do not source at least 30% of the total amount of chip components from the United States. As a carrot, tax incentives may also be offered for these buyers using domestic sourcing, similar to the Inflation Reduction Act. However, both proposals require legislation to be passed by Congress.
Restoring U.S. power in the semiconductor industry will require a series of reforms, not just funding through the CHIPS Act. Failure to do so will not materially reverse business dynamics.
The authors conclude: "The CHIPS Act would be meaningless if the market power of leading fabless companies were not restricted. ”