The Global Semiconductor Talent Crunch: How Protectionism Backfired

Yves here. This post illustrates a pet theme of ours: how in complex systems, trying to map a simple path through them often makes things worse. A big reason why is that the parties trying to hack their way through what amounts to a jungle don’t even have a map of the terrain. Protectionism is a case in point.

Development economists have found that erecting various market barriers and providing support to young but high-priority industries can indeed succeed in having them grow big enough to survive global competition. But the US is allergic to industrial strategy even though it wants dirigiste-looking results, namely rebuilding our hollowed-out industrial base and defending, better yet improving, our standing in advanced technologies. The fact that we’ve been unable to increase production of 155 mm shells, and are dependent (per Alexander Mercouris) on a single factory in Poland for TNT (later gen less nasty explosives have proven to be too difficult to produce to date) shows a remarkable failure of planning. Yet that seems to be endemic in a world of MBAs whose brains have been addled by PowerPoint.

By Mehmet Canayaz, Assistant Professor of Finance, Smeal College of Business, Penn State University; Isil Ere, David A. Rismiller Chair in Finance, The Ohio State University; and Umit Gurun, Stan Liebowitz Professor of Finance and Accounting, University Of Texas at Dallas. Originally published at VoxEU

The US aims to revitalise its domestic semiconductor manufacturing industry, but will there be enough skilled workers to meet the ambitious goals? This column presents evidence, based on a global dataset of 1.6 million employees with chip manufacturing skills, that US protectionist policies – increased tariffs and visa restrictions – implemented since 2018 may have undermined the very workforce the industry needs to thrive. The number of US students graduating with skills relevant to chip manufacturing has decreased, and US chip manufacturing firms reduced hiring, especially for entry-level and junior positions.

As the US aims to revitalise its domestic semiconductor manufacturing industry through initiatives like the 2022 CHIPS and Science Act, a critical question looms: will there be enough skilled workers to meet the ambitious goals? In a recent paper (Canayaz et al. 2024), we present new evidence that protectionist policies – increased tariffs and visa restrictions – implemented since 2018, ironically, may have undermined the very workforce the industry needs to thrive.

The semiconductor industry sits at the intersection of national security and economic competitiveness. Recognising this, policymakers have sought to bolster domestic chip production through measures like tariffs, immigration restrictions, and subsidies. However, our analysis reveals that these efforts have had unintended consequences on the industry’s most vital resource: its talent pool.

Building on earlier work examining the broader economic impacts of trade frictions (Fajgelbaum et al. 2020, Amiti et al. 2019), we focus specifically on how protectionist policies have shaped the semiconductor workforce. Our findings paint a concerning picture of declining domestic hiring, especially for entry-level positions, and a shift in career choices away from the chip industry.

Using a comprehensive dataset of 1.6 million employees with chip manufacturing skills worldwide, we employ a difference-in-differences methodology to analyse employment trends before and after the implementation of protectionist measures in 2018. Our study examines both US semiconductor firms and the educational and career trajectories of individuals with relevant skills.

The results are striking. We find that US chip manufacturing firms experienced a 9% reduction in hiring activities and a 3% decrease in overall workforce size for science and engineering positions compared to other job categories within the same firms. This translates to an annual loss of 2,285 science and engineering positions in the US chip manufacturing sector. Between 2019 and 2022, this amounts to a cumulative reduction of 9,140 jobs in an industry that employed 66,382 engineers and 9,768 scientists during this period.

The decline in hiring is particularly acute for entry-level and junior positions, indicating that protectionist policies have disproportionately affected those new to the workforce. This trend is especially concerning given the industry’s need for fresh talent to drive innovation and growth.

Furthermore, our analysis of educational cohorts reveals a significant shift away from chip manufacturing careers. We observe a marked decrease in the number of US students graduating with skills relevant to chip manufacturing. In 2017, there were 65,290 undergraduates and 39,019 postgraduate students in relevant programmes. By 2022, these numbers had plummeted to 12,311 and 20,503, respectively.

We find that US semiconductor firms have responded to these challenges by increasing their recruitment of experienced workers outside the US. There was a 3% increase in hiring for both junior and mid-senior roles in international segments of these companies. Countries benefiting from this shift include Canada, which strategically amended its immigration policies to welcome more foreign engineers and scientists, as well as European nations with established chip manufacturing industries, such as the Netherlands.

These findings have profound implications for the success of initiatives like the CHIPS Act. The Semiconductor Industry Association (2023) projects a need for 115,000 new semiconductor jobs in the US by 2030. Based on our estimates, it could take approximately 16 years to fill these positions at current graduation rates. This talent shortage could seriously impede the industry’s growth and the US’s ability to achieve semiconductor self-sufficiency.

Our research underscores the relationships between trade policies, immigration, and workforce development. While protectionist measures were intended to boost domestic manufacturing and employment, they appear to have had the opposite effect on the semiconductor industry’s talent pipeline. Recent research by Bosone et al. (2024) shows that geopolitics began significantly affecting global trade after 2018, aligning with the US-China tariff war and coinciding with the timeline of our observed talent crunch in the semiconductor industry. Their study also found evidence of ‘friend-shoring’ in trade patterns, suggesting that geopolitical considerations are reshaping not only supply chains but, potentially, talent flows as well, further complicating the semiconductor industry’s access to global skill pools. This highlights the need for a more thoughtful but fast approach to industrial policy that considers the global nature of the semiconductor workforce and the importance of maintaining open channels for talent acquisition and development.

To address these challenges, policymakers could consider several key actions:

  1. Re-evaluate immigration policies: Implement targeted visa programmes, such as the proposed ‘Chipmaker’s Visa’, to attract and retain international talent in the semiconductor industry.
  2. Boost domestic STEM education: Increase investments in educational programmes and initiatives that encourage more domestic students to pursue careers in chip manufacturing and related fields.
  3. Foster industry-academia partnerships: Encourage closer collaboration between semiconductor companies and universities to ensure curriculum alignment with industry needs and provide more internship and research opportunities.
  4. Develop retraining programmes: Create initiatives to help workers from other industries transition into semiconductor manufacturing roles, tapping into a broader pool of potential talent.
  5. Incentivise talent retention: Implement policies that make it more attractive for skilled workers to remain in or return to the US, such as tax incentives or student loan forgiveness programmes for those who commit to working in the domestic semiconductor industry.
  6. To bridge the talent gap and secure a long-lasting competitive edge in chip manufacturing, prioritise investments that integrate AI into this sector. Coordinate the proposed ‘Chipmaker’s Visa’ programme with supplementary initiatives that link chip manufacturing and AI development.

The success of America’s semiconductor renaissance hinges not just on building new fabrication plants (‘fabs’) and securing supply chains, but on cultivating and sustaining a skilled workforce. Our research serves as a cautionary tale about the unintended consequences of protectionist policies on talent development and retention. As the global competition for semiconductor supremacy intensifies, the nation that best nurtures and attracts top talent will likely emerge as the leader in this critical industry.

The path forward requires a delicate balance between promoting domestic capabilities and maintaining the global interconnectedness that has long been a hallmark of the semiconductor industry. By addressing the talent crunch head-on with targeted, forward-thinking policies, the US semiconductor industry can work towards realising the full potential of its ambitions and securing its technological leadership for decades to come.

See original post for references

Print Friendly, PDF & Email

9 comments

  1. TimH

    I’ve worked in semiconductor companies in USA for 25 years, and I have estimate that 80% of the fab staff (engineers, technicians) were not born in USA. The recent TSMC build in Phoenix failed because they couldn’t get the staff.

    The people proposing this strategy surely know this, so some other play is going on.

    Reply
  2. Susan the other

    does China’s layered 3-D chip approximate quantum computing abilities? Or is there additional technology required to somehow employ entanglement? where is this industry heading? And how fast is simply too fast in this otherwise plodding world?

    Reply
  3. Louis Fyne

    IMO, the larger issue is that STEM in the US, from K to 12, stinks at the national level.

    And a frontline role on the fab floor is not a what MIT grad is looking for.

    the directional land grant schools need to be pumping out more quality STEM grads, but can’t if 12th graders don’t have a STEm foundation.

    No royal road to geometry or in the sciences

    Reply
  4. vao

    The decline in hiring is particularly acute for entry-level and junior positions, indicating that protectionist policies have disproportionately affected those new to the workforce.

    This specific point is for me a puzzle. The authors provide no explanation as to the mechanism by which the introduction of tariff barriers on imported products would cause a hiring reduction of junior and entry-level personnel in the USA.

    After skimming through the complete paper, it looks as if its “insights” actually derive from a typical approach of quantitative economics: computing regression coefficients on a bunch of parameters — one of them corresponding to the introduction of protectionist measures. That chip manufacturers in the USA reduced their hiring for entry-level positions is an observed effect of the regression, but remains unexplained. In the absence of any plausible mechanism, it could also be that the introduction of protectionist policies and the reduction of hiring are correlated to a third, explanatory parameter which the authors missed.

    An interesting aspect regarding the reduction of young people entering electronics studies: it appears to be a worldwide phenomenon (p. 29):

    In stark contrast, 58.7% of countries in our sample witnessed a decline in classmates entering science and engineering roles each year.
    […]
    Among the significantly impacted, the U.S. and India stand out, with both experiencing reductions of more than 1000

    If this is true, India progressing in a development path comparable to China, and the USA rebuilding its electronics industry seem both severely compromised.

    While industrial policy is, as indicated in the introduction, a skein that is quite hard to untangle and with paradoxical aspects, this article confirms my accumulated impression on VoxEU: lots of effort on mathematics, weak on substantial insights.

    Reply
  5. Altandmain

    There are even bigger problems than the education system producing enough engineers nor enough immigrants with a background in semiconductors. That’s what the article seems to focus on.

    The education system is a problem, but I have to ask, are the rich willing to pay the higher taxes for a good public education system, both at a primary and university level? Thanks to neoliberal economics, the rich are the only people with the disposable income for this and most Americans live paycheck to paycheck.

    What’s the other big problem. The corporate culture of these companies is going to need to change. They spent a ton of money on buybacks of their stocks. They are profit oriented, not oriented to making the best quality products, and short-term profit oriented at that.

    https://www.nakedcapitalism.com/2021/07/how-intel-financialized-and-lost-leadership-in-semiconductor-fabrication.html

    US taxpayers may very well just end up subsidizing buybacks.

    https://www.nakedcapitalism.com/2021/10/does-america-want-a-chips-for-buybacks-act.html

    I’ve also heard from Intel employees (and disclosure, I know one colleague who works for Intel) that the company pays below market wages and does not have good work life balance. He only took the job because he was laid off and looking for work. It means that chip design and manufacturing has much less appeal as a career than say, management consulting, finance, or computer programming.

    It isn’t just intel – Nvidia last year bought back stock:

    https://www.reuters.com/technology/nvidias-25-billion-buyback-a-head-scratcher-some-shareholders-2023-08-25/

    In the advertising and software side, Google has been buying back shares, as has Apple (which is a fabless company now like Nvidia).

    Imagine this scenario:

    US taxpayer subsidize the chip making companies, only they don’t spend a high percentage of the subsidies on capital expenses, research, and higher wages for staff, while buying back shares. Congress is corrupt enough to sign off on that via lobbyists and campaign contributors. Meanwhile, they use PR to pretend like they are “doing something”. This sounds like the most likely outcome to me.

    This will in the short run lead to a bubble in the stock market, but in the long run, nations like China will overtake the US in advanced technology. That may lead to even that stock bubble collapsing, as US companies, both fabs and fabless companies, struggle to compete.

    Recently, there was an interview of TSMC’s former lead scientist, and the popular Twitter influencer Arnaud Bertrand noted that the Chinese have been surpassing the US in semiconductor research papers:

    https://x.com/RnaudBertrand/status/1793457614816494011

    China has also been spending a ton on capital investment, especially after the US sought to deny China the most advanced tools, most notably ASML’s EUV technology.

    This is going to backfire on the US as hard as their sanctions war against Russia. The US hasn’t even begun to look at the real root problems and the rich are too greedy to solve them.

    Reply
  6. TomDority

    The play is to get talent from abroad cheap and further promote privatized profit of public research.
    Project 2025 PRESIDENTIAL TRANSITION PROJECT
    “National Economic Council. The NEC was established in 1993 by executive” (Bill Clinton) “order” “Those who have served in the role have ranged from former CEOs of the nation’s largest investment firms to financial-services industry managers to seasoned congressional staffers who have managed the economic policy issues for top financial and tax-writing committees.”
    “The remaining 14 science and energy labs should focus on basic research projects; demonstration and deployment of technology should be left to the private sector.”
    “Unleash private-sector energy innovation by ending government interference in energy decisions.”
    “The agency is unnecessary, risks taxpayer dollars, and interferes with risk-benefit decisions that should be made by the private sector.”

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *