Restoring Leadership In Semiconductor Production

Illustration Computer Keybords, Digital Future. (Photo by Viennaslide/Construction … [+] Photography/Avalon/Getty Images)

Getty Images

Last week, in the first installment of the series of which the present column is second, I showed how strikingly similar the state of American industry is now to the state of American industry when the Second World War broke out. The Roosevelt Administration, in close collaboration with the leaders of the most important manufacturing industries of the day, converted us virtually overnight from a country with negligible aircraft, naval and cargo vessel, tank and other military vehicle, munitions, and strategic materials industries into the world’s great ‘Arsenal of Democracy.’

           What had enabled this, I indicated, was a unique institutional pairing that we had also used during the First World War, not to mention the nation’s early industrialization along Hamiltonian lines. This pairing has always involved an interagency coordination board comprising Cabinet-level executive agencies and private sector industry leaders on the one hand, and a federal financing bank that could finance all board-formulated strategic manufacturing plans on the other hand.

           I then showed how my InvestAmerica Plan, devised early last year and advocated by myself and my New Consensus colleagues ever since, replicates precisely this pairing so as to enable our country quickly to restore self-sufficiency, reclaim global industrial leadership, and prevent domestic price inflation by massively levering-up the nation’s productive capacity.

           Plank I of the Plan, a National Reconstruction and Development Council comprising Cabinet officials and strategic industry leaders, would be an updated version of the War Industries Board and War Production Board of the two World War eras. Plank II of the plan, upgrading the capacities of the already-existing Federal Financing Bank housed in Treasury, would similarly update the World War eras’ War Finance Corporation and Reconstruction Finance Corporation.  

           In order both (a) to show how the Plan works to ‘turbocharge’ key strategic American industries on the one hand, and (b) to roadmap actual revitalizations of the key strategic industries of today and tomorrow on the other hand, I am producing a series of ‘Case Studies,’ of which the present column is part. I’ll start with semiconductors (a.k.a. ‘microprocessors,’ ‘chips,’ or ‘microchips’) – an industry that the US invented and then led the world in for decades until recently. More than with any other industry, our offshoring of this one has led to the most pronounced supply chain disruptions, consumer price inflations, and strategic vulnerabilities that confront us right now.

           Off, then, to the proverbial races…      

           Semiconductors and the circuit boards that combine them are at least as critical an input to modern industry and military capacity as petroleum was from the 1950s through the 1990s. Indeed, more than petroleum, they are essential both to the machines with which we produce products from household consumer goods through automobiles and defense systems, and to the functioning of such products themselves. Bottlenecks in microchip and circuit board supply chains such as have recently plagued the US and world economies accordingly induce production slowdowns and shutdowns, goods shortages, and price inflations as surely as did oil embargoes in the 1970s.

           Unlike petroleum, however, microchip and circuit board supplies are not simply the product of geological fate. They are themselves manufactured goods, the domestic supplies of which nations determine through variably productive or unproductive industrial policy decisions and building projects. Unfortunately, it has been US policy in recent decades to cede global leadership almost entirely to other nations – notably China – in four of the five main constitutive ‘stages’ of semiconductor and circuit board manufacturing. Those are (1) materials gathering and refining, (2) circuit design, (3) production equipment manufacture, (4) chip fabrication, or ‘fabbing,’ and (5) circuit board ‘packaging.’

           In order to restore its lost self-sufficiency and global leadership in semiconductor production and circuit board assembly, the US public sector must collaborate with and heavily aid the US private sector at the front end – as China does its rapidly growing chip industry – at least until ‘critical mass’ in domestic production is restored. The reasons are several.

For one thing, US producers at present are heavily dependent on foreign demand – especially Chinese demand – in order for private investment in domestic capacity, a particularly costly proposition, to be profitable.

This is especially so in view of the second reason warranting public sector involvement – the daunting enormity of plant construction expense. A typical high-end chip fabrication plant now costs as much as $15 billion to construct. Yet foreign – especially Chinese – demand for US semiconductor inputs at volumes warranting such outlays by our domestic industry cannot be confidently expected to continue in the face of both (a) sharpening geopolitical competition between China and the US, and (b) the tendency of producers at one stage of semiconductor production to migrate toward locations where producers at other stages of production congregate – a phenomenon known as ‘colocation.’

Third, other nations – again especially China – are themselves massively subsidizing their own domestic semiconductor industries, in hopes of attaining first self-sufficiency and then world leadership (cf. China’s ‘Made in China 2025 plan) in this all-important industry. US firms enjoy little equivalent to that as of yet apart from the recently passed but woefully inadequate CHIPS Act, more on which below.

Fourth, continued access to material inputs to the semiconductor production process is anything but guaranteed to American producers, given increasingly aggressive Chinese efforts to secure exclusive access to the geographic locales from which many of these materials are extracted – a form of ‘preclusive buying’ of the kind pioneered by Roosevelt’s RFC in the early months of the Second World War.

And finally fifth, production in the semiconductor and circuit board industries is heavily reliant upon a well-educated and highly skilled workforce – a now steadily dwindling part of the US labor market.

Over none of these factors does private sector US industry have any control. American manufacturers cannot unilaterally determine demand for their products, the costs of supplies and facilities, the subsidies other nations strategically supply to their domestic producers, the efforts that competing nations make to secure preclusive access to material inputs, or (a) the level of education of, or (b) other supports, such as housing and child daycare facilities, for the American labor force.

Faced with such uncertainties in respect of supplies and demand on the one hand, and the high cost of plant construction and related machine capital investments on the other hand, shareholder-answerable private sector firms in the US are not acting irrationally in refraining from attempting singlehandedly to restore US self-sufficiency and world leadership in the production of semiconductors and circuit boards.

The fact that US firms are acting individually rationally in not leading the charge to national self-sufficiency on the one hand, combined with the fact that not restoring this self-sufficiency in light of the perils described above is collectively irrational on the other hand, lends US dependence on foreign-sourced semiconductor and circuit board manufacture the character of a classic collective action problem. For a collective action problem just is a choice situation in which multiple individually rational decisions aggregate into collectively irrational outcomes.

           The solution to collective action challenges, which are ubiquitous in modern macro-economies, is straightforward. What are required are well-considered and well-planned exercises of collective agency – that is, concerted and mindful public action. The InvestAmerica Plan summarized in the first installment of this series and reprised above is meant to enable precisely that – collective agency of a kind that renders individual private agency once again profitable and hence individually rational. In so doing, it updates the same collaborative public-private teamwork that enabled the US massively to out-produce its adversaries in the Second World War – again as elaborated in the previous installment of this series and reprised above.  

           Let us then turn to how the Plan does this in the realm of semiconductor and circuit board manufacture.

           The key to restoring US self-sufficiency and global leadership in semiconductor and circuit board production lies in publicly addressing the impediments to rational private investment enumerated above. The InvestAmerica Plan provides the platform on which to do this.

           To begin with, the new Cabinet-reconfiguring Council – again, Plank I of the Plan – will designate the semiconductor industry a strategically critical sector in need of significant front-end public sector support – the kind China now provides to its semiconductor industry with a view to achieving global dominance by 2025.

           The specific Cabinet-level federal agencies whose mandates are implicated by the project of restoring national self-sufficiency and global leadership will then, in collaboration with industry leaders, provide specific advice and supportive data where their specific competencies figure in to the revitalization process. In so doing, moreover, they will in effect not only be engaging in strategic planning, but will also be blueprinting the financing task that will be assigned to the upgraded FFB – again, Plank II of the Plan – in financing the full strategic plan.

           The Department of Commerce, for example, in Second World War style collaboration with industry, the Small Business Administration (SBA), and the Department of Defense (DOD), will likely be the agency best suited to determining how much capacity the US must build or rebuild at all stages of the semiconductor and circuit board production and assembly processes, from materials gathering through chip design and fabrication, to circuit board assembly, in order to restore national self-sufficiency and indeed global leadership in the field. In so doing, of course, it will have to take account of the scale economies of production at each stage.

           In light of those economies, in turn, Commerce will then estimate how many new facilities have to be built or converted, whether it be a few very large facilities or a somewhat greater number of somewhat smaller facilities. It will then estimate both the costs of building these facilities and the likely profits to be earned in the long term as production ramps up and the products are sold.

           Input from firms operating in and adjacent to the US semiconductor industry – firms such as Intel and Apple, for example, as well as US subsidiaries of Taiwan’s TSCM and South Korea’s Samsung planning to build in the US – will of course have to be solicited too, as occurred in connection with war production during the First and Second World Wars. And the Department of Justice (DOJ) and Federal Trade Commission (FTC) will weigh in on potential antitrust and market concentration concerns in the interest of promoting an optimal degree of efficiency-inducing competition in the markets that we will be turbocharging. 

The US Trade Representative (USTR) will likely have additional advice to offer in respect of the foregoing questions, in light of estimated global demand potential for the output of a revitalized domestic semiconductor and circuit board industry. Similarly, the aforementioned DOD, as well as other federal agencies on the Council that make use of equipment that uses semiconductors, will weigh in both on specialized semiconductor and circuit board needs in their particular fields of operation, and on how their procurement policies can be made to guarantee a significant source of reliable demand for US semiconductor output. In this connection, private sector defense contractors and other suppliers of advanced equipment to public sector agencies and instrumentalities will also be asked to weigh in on anticipated needs – again as was done in the First and Second World War mobilizations. 

The Department of Labor (DOL), meanwhile, together with the Commerce Department and the Department of Education (ED), will identify the labor needs of a revitalized semiconductor and circuit board industry, current and likely future labor supplies able to meet those needs, and means of addressing shortfalls of the kinds and quantities of labor needed. This will likely require the identification of new forms of, and both existing and needed new facilities for the provision of, such educational ‘upgrades’ as the US workforce will require. The Immigration and Naturalization Service (INS), meanwhile, will be an important contributor too – particularly as highly educated engineers, concerned about pandemic and an ever-more hostile China, find emigration to the US from Taiwan and elsewhere increasingly attractive.

Here too the examples of the First and Second World War mobilizations are instructive, as the welcoming of foreign-educated nuclear and other scientists, and the training of millions of new skilled workers, was required in order for America’s newly constructed or converted vehicle, aircraft, shipping, and munitions factories to operate at capacity. Here again, furthermore, both public and private sector educational institutions – including community colleges and vocational training schools – will be asked for the benefit of their expertise and capacities, while being financially aided in expanding those capacities.             

           The Department of Labor also, now in tandem with the Departments of Energy (DOE), Interior (DOI), and Transportation (DOT), will weigh in on optimal locations for new production sites as recommended by the Departments of Commerce and Defense. The factors that will bear upon these optimal siting decisions will be several.

           One factor will be the comparative needs of specific regions or demographics of the country – especially recently blighted areas – for new production and employment opportunities. Another factor will be proximity to sites at which semiconductor inputs additional to labor are abundant or readily procured, as well as sites to which newly produced chips and circuit boards will be transported. And yet another factor will be proximity either to existing communications and transport networks or to places where new such networks can be readily constructed. Here again, the example of the First and Second World War mobilizations is instructive, inasmuch as entire transportation networks were built to enable ready shipment in and out of productive inputs to and outputs from new manufacturing facilities, respectively.

           The Department of Housing and Urban Development (HUD), too, will likely weigh in on housing needs and home construction opportunities for newly enlarged labor forces in places where new productive facilities are to be built. Here again the example of the Second World War mobilization is instructive. For as noted in the first installment of the present series, entire neighborhoods and housing units had to be newly constructed to accommodate rapidly expanding workforces in area where new plants were constructed. Schools and daycare centers quickly followed, as did new power, water, and sewage facilities. The US Army Corps of Engineers – with 35,000 employees, the largest construction ‘company’ in the world – will also be an important source of information and advice, if not also actual construction, both in respect of these facilities and in respect of plant construction itself.   

           As the Council develops a strategic Semiconductor Industry Revitalization Plan along the lines just elaborated, its financing arm – the FFB, Plank II of the InvestAmerica Plan – will develop a corresponding financing plan. The specific forms of financing that will be optimal will undoubtedly vary according to which of the stages in the semiconductor and circuit board production process is under consideration. They also will vary across the complementary industrieshome building, school building, road paving, power and water routing, etc. – that rebuilding semiconductor and circuit board manufacturing capacity will complementarily necessitate.

The US still leads in semiconductor design, for example, and is accordingly more in need of maintaining or widening its lead than of jumpstarting a new industry. Similar observations hold of some, but far from all, materials gathering, as the US is self-sufficient with respect to some while being dependent on foreign sources for others. Where chip fabrication and circuit board assembly are concerned, by contrast, much site construction from scratch is likely going to be requisite. Meanwhile, some domestic and foreign firms – notably Intel, Taiwan’s TSMC, and South Korea’s Samsung – are planning to build at least some new fabrication capacity in the US, and might accordingly require less public encouragement than will domestic semiconductor firms not yet planning any such expansion at the requisite scale.

In this connection, the recently passed CHIPS Act, which earmarks some $52 billion to finance an expansion of US chip research, design, and production, is a helpful development – in effect enabling a rapid commencement of Council and FFB investment activity in this sector. Relative to the need, however, which industry analysts believe to be at least ten times what is now earmarked, the CHIPS Act can be considered no more than a down payment. The same holds of the FABS Act, which has yet to be passed and in any event would work only through tax credits – a method, under the guise of ‘accelerated amortization,’ we learned was insufficient on its own during the Second World War mobilization.  

In any event, the form of financing supplied will be determined according to two criteria: first, speed and efficiency in reaching national self-sufficiency; and second, assuring the profitability of FFB investments in the long run, the FFB’s predecessor the RFC having been the only federal agency in its day that added to government income rather than subtracting from it. Indeed, unlike during the Second World War when post-Depression banks were on the ropes, today the FFB can form consortia of combined public and yield-hungry private sector investors, including both public and private employee pension funds, to assure Americans a direct financial stake in the success of its own productive revitalization efforts.

           In sum, then, the situation that the US now faces where semiconductors and circuit boards are concerned is strikingly reminiscent of that it faced in the lead-up to the Second World War. There is both immediate existential need and spectacular longer-term ‘postwar’ opportunity. Owing to the impediments laid out above, however – themselves reminiscent of counterpart challenges that confronted the US at the time of the Pearl Harbor attack – the US private sector lacks both the resources and the rational profit incentives to restore US leadership on its own.

           Though the urgency of actual imminent war is, thankfully, lacking in the present instance unlike in the twentieth century, the need to restore US self-sufficiency and global leadership in semiconductors and circuit boards surely counts as what the great American philosopher William James would have called ‘the moral equivalent of war.’ And with a ‘new cold war’ between the US and some of its competitor nations already now regrettably looking possible, we appear to be reaching a point where we needn’t speak about mere moral equivalents.