Chipping Away

Intel’s August announcement of missed earnings, investment cutbacks, and widespread layoffs provided an ideal opportunity for critics of American industrial policy to embarrass themselves. The CHIPS and Science Act (“CHIPS Act”) allocates $52 billion to subsidize semiconductor R&D and construction of semiconductor manufacturing facilities (“fabs”) in the United States. In March, Intel announced it was going to receive up to $8.5 billion of it in direct funding to support expansion of its domestic chipmaking “capacity and capabilities.” The $10 billion cost-saving measures followed a few months later. Taxpayer money? Struggling company? Cue the op-eds.

“Intel was one of the biggest beneficiaries of the CHIPS Act,” wrote Stephen Moore and Phil Kerpen in the New York Post. “What did we get in exchange? Intel this week announced it was laying off 15% of its workforce — 15,000 positions.” At National Review, Dominic Pino smirked, “Government has picked a real winner here.”

But when Congress was debating the CHIPS Act two years ago, the argument from industrial policy opponents was exactly the opposite. “Taxpayers should at least know they’ll be subsidizing highly profitable companies that don’t need the help,” lectured the Wall Street Journal’s editorial page. “Semiconductor companies are highly profitable, growing rapidly, and able to actively compete with China without additional subsidies or new tax credits,” said Congressman French Hill, then the ranking member of the House Financial Services Committee’s subcommittee in charge of national security. If the companies are doing well, you see, that proves industrial policy is unwise. If the companies are not doing well, that proves the same thing.

In fact, Intel’s recent performance has little to do with the CHIPS Act except as a stark reminder of how badly tarnished this one-time crown jewel of American industry has become, and why industrial policy was necessary in the first place.

The measure of the CHIPS Act’s effect will not be corporate earnings released before CHIPS-funded fab comes online, but in the standing of the domestic industry at the decade’s end. Intel itself understands this; assuming its efforts to right itself succeed, the company asserts it remains on track to meet its goal of seizing the second place spot among global foundry companies by 2030. More broadly, Secretary of Commerce Gina Raimondo believes that the United States is “on track to produce roughly 20% of the world’s leading-edge logic chips by 2030, up from the zero percent we produce today.” The Semiconductor Industry Association, in a report released with Boston Consulting Group, shares that optimism, projecting U.S. fab capacity to triple, and advanced logic capacity to jump from its 0% global share today to 28% by 2032.

But much can be said now, on the law’s second anniversary, about both its successes and its challenges, and about lessons learned for future industrial policy efforts. None of that will satisfy ideological opponents determined to interpret whatever facts might emerge as confirming their market fundamentalism. But for leaders eager to update their policy agendas to the realities of the 2020s and determined to reindustrialize the American economy, a clear understanding of where we are now is invaluable for charting the course forward.

A Fork in the (Supply-Side) Road

Amidst the flotsam of pretextual anti-CHIPS Act talking points, one argument has proved quite helpful. Why support the semiconductor industry, opponents asked, when we could just cut taxes instead? “We aided all American industries and our competitiveness when we passed the Tax Cuts and Jobs Act – removing double taxation on earnings from abroad,” noted Congressman Hill. “Over $1 trillion in capital flowed back to the United States.”

Numerous commentators and lawmakers agreed. Senator Ted Cruz offered a classic of the genre: “I’m all for using the tax code to incentivize manufacturers to build semiconductors in America,” he said. “But when the federal government simply gives billions of taxpayer dollars directly to massive corporations, it invites cronyism and corruption.” If the incentive to chipmakers could be described as a tax cut, he was in favor. But if it was a subsidy, he was opposed.

What makes this argument so helpful is its testability. Large corporate tax cuts and targeted subsidies to critical industries are both policy tools that Congress might use to spur investment and expansion of domestic productive capacity. Indeed, they are both quintessential examples of “supply-side” policy, notwithstanding the view of many market fundamentalists that the former counts while the latter is merely socialist central planning. As I observed in “Rebuilding the Supply-Side Platform:”

The basic claim of supply-side theory is that policy should seek to induce greater productive investment. In Reagan’s time, high marginal tax rates and sclerotic regulatory regimes were obvious culprits to target. But you only get to reduce tax rates from 70% once. Subsequent efforts at repeating the trick have brought diminishing returns and rising costs.

Two years after passage of the Tax Cuts and Jobs Act (TCJA), economists eagerly pored over the investment data to find the positive impact of their plan. Two years after passage of the CHIPS Act, they can do the same here. Which supply-side policy has more effectively spurred domestic investment?

TCJA, unfortunately, failed on its own architects’ terms. As Oren Cass has noted, Kevin Hassett, chair of the White House Council of Economic Advisers at the time, took the stage at the American Economic Association’s annual meeting one year after TCJA’s passage and presented preliminary data showing impressive effects on investment and growth. But using those exact same metrics, updated to reflect the revised and final data, Hassett’s charts would show that TCJA had accomplished nothing.

Perhaps that’s why Congressman Hill, attempting to celebrate TCJA’s impact, referred only to “$1 trillion in capital flow[ing] back to the United States” without any mention of what that capital flowed into. Cutting corporate taxes indeed created an incentive for firms to generate and realize profits in the American market. It did not, however, yield what the American economy needed.

By contrast, the CHIPS Act immediately and dramatically altered global industrial investment. Taiwan Semiconductor Manufacturing Company (TSMC) has steadily increased its commitment to Arizona as federal subsidies came online, from an initial $12 billion (at the urging of the Trump Administration and in expectation of forthcoming financial support), to $40 billion in 2022 after enactment of the CHIPS Act, and finally to $65 billion in 2024 upon confirmation of $6.6 billion in government subsidies, $5 billion in loans, and construction cost tax credits of up to 25%. Samsung’s investments in Texas have followed a similar trajectory.  

To appreciate the scale of these investments, consider that over the ten years from TSMC’s first project announcement (in 2020) to the anticipated production start in its third and most advanced Arizona fab (“by the end of the decade”), its annual investment will average $6.5 billion. This is the average level of economy-wide greenfield foreign direct investment experienced in the two years after passage of TCJA.¹Based on annual BEA data. BEA’s annual New Foreign Direct Investment in the United States release defines greenfield investment expenditures as “expenditures to either establish a new U.S. business or to expand an existing foreign-owned U.S. business.” This figure was $7.7 billion in 2016, $6.4 billion in 2017, $9.1 billion in 2018, and $4.0 billion in 2019. See: BEA 2014–2016; BEA 2017; BEA 2018; BEA 2019. In total, investment in construction of manufacturing facilities for computing and electronic devices has increased 15-fold, according to American Enterprise Institute scholar Chris Miller. Semiconductor companies and related suppliers have announced over $300 billion in investment over the next decade. U.S. manufacturing construction is skyrocketing, driven almost entirely by investment in this sector, which has in turn been partially driven by the CHIPS Act—even before it became law, as industry anticipated its enactment.” 

A nation gets what it pays for. If policymakers reward corporate profits with tax cuts, they may well get more corporate profits. But if productive domestic investment and a revitalized semiconductor industry are policy priorities—as they should be—the CHIPS Act is already proving a better course is available.

Of course, many of the investments are yet to be made. Some will surely be stalled; others will prove ultimately unproductive. That is the nature of innovation. But that’s not a standard the tax-cutters have ever held themselves to when attempting to celebrate their own results. Does the profit-maximizing behavior pursued in response to a tax cut lead to productive new investment, or merely financial speculation and shareholder payouts? The Wall Street Journal editorial board is strangely uncurious. At a minimum, we can say the semiconductor fabs were not returning to our shores. Now, by all accounts, they are doing just that.

Bumps in the (Implementation) Road

The potential for the United States to use industrial policy effectively in the twenty-first century has been validated. How best to do it, however, is another matter. Criticism of whether to do a CHIPS Act has fallen flat, but much useful commentary has highlighted flaws in how implementation has occurred.

Republican attention has focused in particular on extraneous requirements imposed by the Biden administration in service to other priorities, generally standard Democratic Party fare. Senator Marco Rubio enumerated them in his recent National Affairs essay, “Industrial Policy, Right and Wrong:”

[One] problem with the Biden administration’s implementation of the CHIPS and Science Act is that it imposes burdensome regulations on chipmakers that directly contradict the law’s stated goal of building more factories in America. One of the CHIPS program office’s first actions was to identify the hoops companies must jump through to qualify for grant funding — many of which sound like parodies of left-wing identity politics.

For example, the rules state that companies must “develop an equity strategy” so that groups the left believes deserve special treatment — ”people of color,” “LGBTQ+ individuals,” and “justice-involved individuals” (i.e., convicted criminals) — do not face “systemic barriers” to employment. Companies also must establish inclusive hiring practices to ensure that enough of their contractors are owned by minorities, women, and veterans. And they must submit a “construction workforce plan” with a “detailed description” of how they will recruit female construction workers.

Other Republicans have been swift to blame these requirements for delays in successful progress. “DEI killed the CHIPS Act,” according to the headline of an op-ed from Matt Cole and Chris Nicholson, respectively the CEO and Head of Research at Strive Asset Management, Vivek Ramaswamy’s anti-woke investment firm. Senator Tommy Tuberville cited similar concerns, arguing that “this red tape has predictably caused significant delays in the implementation of the CHIPS program. In fact, these policies, and subsequent funding eligibility requirements, have slowed implementation of the bill to a near halt.”

These claims overstate the case. The CHIPS Act has been neither killed nor halted; the Department of Commerce has committed most of its $39 billion in grants and expects the remainder to be finalized this year. As noted above, real construction investment has exploded, and firms have announced hundreds of billions of dollars in investment. A good example of the distance from rhetoric to reality is Tuberville’s complaint that “the CHIPS Act established yet another foreign-aid slush fund” and “the Biden administration recently announced they would use some of the funding to produce semiconductors in Kenya.” Total funding? $1 million. This seems unlikely to be a major obstacle.

Still, the delays are real. TSMC has pushed back anticipated production start dates several times on more than one of its Arizona fabs, citing questions about federal support. Intel announced delays to its Ohio project in early 2024, also citing slowness in CHIPS funding. “Sluggish Chips Act payouts slam[med] the brakes on Samsung’s fab,” according to one report.

Whether requirements for equity plans and child care played a major role in delays is debatable, but also somewhat beside the point. Republicans would object to these requirements whether they delayed implementation or not. Mission creep in general is the more pressing problem, and one not just emphasized by Republicans. Ezra Klein famously called this tendency of Democratic lawmakers to pursue numerous social problems in every policy initiative, to the detriment of its primary purpose, “everything bagel liberalism.” Writing in 2023 about the CHIPS Act provisions, he observed:

But there is a cost to accumulation. How many goals and standards are too many? And why is subtraction so rare? It is impossible to read these bills and guidelines and not notice that the additions are rarely matched by deletions. Process is enthusiastically added but seldom lifted. The result is that public projects…aren’t cost competitive, and that makes them vulnerable….  when a bad economy hits or a new administration takes over and the government cuts its spending. Liberalism is much better at seeing where the government could spend more than at determining how it could make that spending go farther and faster….I’m not predicting failure for CHIPS because it includes a child care mandate. These are expensive factories that can figure out child care. But even if no single standard or mandate is decisive on its own, the accumulation of them, in an industry in which we’ve already fallen ruinously behind on cost, can do real damage.

The danger of this kind of diffuse focus is not only to the particular project, but also to the broader effort to build consensus around necessary economic policy. Klein is correct: Big chip companies did figure out how to navigate child care requirements. The real vulnerability is political durability, and whether the next industrial policy push will muster bipartisan support after the perception that the current one has been politicized in its implementation. DEI may not have killed the CHIPS Act, but the perception of “woke” implementation could certainly strangle the next important legislative effort in the cradle. Conversely, it is not hard to imagine a Republican Congress or White House seeking to impose conservative social priorities in a future round, thus losing progressive support.

American re-industrialization depends on a new economic consensus emerging around its importance. The potential for that consensus to solidify must be vigilantly protected. Future industrial policy should utilize clear legislative language stating the explicit purposes of the public funding and limiting requirements for program participation to those directly necessary to advance the law’s stated purpose. This approach would force greater clarity of purpose and priority at the outset, ensure smoother implementation, and strengthen the political coalition for supporting domestic industry.

Clearing the Real Roadblocks

Ask the government officials and industry participants about the experience of CHIPS Act implementation and they will credit the rapid progress while also highlighting several significant challenges. On the government side, building the “state capacity” to administer industrial policy has taken time. Hiring was slow and managing the application volume required new bureaucracy. The private sector likewise had to learn a new mode of interaction with the government and planning for massive capital projects. These growing pains are inevitable as the United States attempts to relearn the mechanics of the public-private partnerships necessary to industrial leadership in the globalized economy. Two challenges stand out, though, as particularly significant and, because they are general features of the American economic ecosystem, likely to persist as roadblocks to future efforts at industrial policy: environmental permitting and workforce development.

Fab construction delays due to environmental permitting requirements have been the subject of immense and bipartisan concern, starting well before the CHIPS Act passed and continuing now. In a 2023 Senate hearing, Secretary of Commerce Gina Raimondo warned that subjecting CHIPS Act projects to full National Environmental Policy Act (NEPA) review requirements could delay them “up to years,” and she has pushed Congress to exempt the projects from NEPA review. A bill to address the problem has support from 120 bipartisan members of the House and Senate as well as the Biden administration, and has passed the Senate twice. The House has not gone along.  

House Republican leadership appears to be opposing the NEPA exemption for fear of losing leverage to negotiate broader permitting reform later, an entirely correct and admirable goal. As the need for CHIPS Act exemptions only underscores, NEPA is a primary obstruction to re-industrialization efforts. But intentionally hampering the CHIPS Act in service to that broader agenda oddly mirrors the worst of the Biden administration’s excesses in advancing its own extraneous agenda. The point of the CHIPS Act is to jumpstart domestic semiconductor production as quickly as possible; by definition, its premise is that this particular set of projects is more important than others that might be undertaken.

And while proceeding with a CHIPS Act exemption may abandon short-term leverage for more extensive reform (which, regardless, seems unlikely to dislodge the environmental lobby’s opposition), its precedent has the potential to build positive momentum. Endorsement of the exemption would itself represent a powerful acknowledgment that NEPA review has substantial costs that can exceed its benefits. If NEPA-exempted projects move forward in an environmentally responsible manner and validate much leaner reviews, they will provide a template for broader reform.

A lack of qualified and willing workers is a more serious problem. After decades of inattention and atrophy, the American education and workforce development systems are not aligned with the needs of a resurgent domestic chips manufacturing sector. Chip companies have sought to address the problem, with some success, by negotiating with organized labor in worksite communities and by making use of foreign talent as needed. But individual companies are not equipped to solve a nationwide problem on their own, even with CHIPS Act help, and the competitiveness of the American economy in attracting these types of investments over the long-term depends on solving it for them broadly—as other nations take it upon themselves to do. Closing the gap between the needs of industries returning to American shores and the capabilities of the American workforce will be particularly important to the success of future industrial policy efforts. 

More than anything, American workforce development must reorient itself around meeting employer needs. Employers, as the ones who hire and deploy talent, understand best what their needs are. Especially when their growth is a national economic priority, rather than merely a private pecuniary interest, public policy should aim to help them. Legislation like the American Workforce Act, advanced by Senators Tom Cotton and JD Vance and Congressman Max Miller, would retool the U.S. education and workforce development system to enable a much wider range of on-the-job training opportunities as alternatives to college, prioritizing the training that employers want their workers to have. Their model is the right one, and would go a long way to fixing the problem.

It would also, however, take time. In the short-term, the CHIPS Act experience highlights an important point of potential convergence between advocates and critics of temporary guestworker visa programs. Even strong advocates of reforming or restricting guest-worker programs acknowledge that, in the context of industrial policy, short-term visas specific to those policy priorities would be constructive. When it comes to industrial policy, the entire premise is to encourage a set of investments that the market is not delivering on its own, and for which the market may not yet have demanded a workforce. Policymakers should eagerly ensure access to the labor that will allow those investments to proceed.

Thus, while most immigration policy remains highly polarizing, efforts to promote rapid industrial progress through smart immigration policy merit bipartisan attention. At the Economic Innovation Group, Adam Ozimek and Connor O’Brien have argued:

…an industry-agnostic [immigration] perspective tends to be most effective in stoking innovation. However, there are a handful of rare, notable exceptions where policy can be targeted at specific industries for geopolitical and national security reasons. Few industries meet this high bar, but Congress has, with justification, decided that the security of semiconductor supply chains is too important to be allowed to be concentrated in the hands of adversaries or geopolitically vulnerable regions. Given this policy stance, the industry-agnostic approach to immigration should be reconsidered in this context as well; CHIPS subsidies are headed out the door regardless, so it is imperative we do what we can to make them work.

At American Compass, meanwhile, Oren Cass has pointed out:

But note what’s different [about the CHIPS Act]: policymakers have chosen a specific need that the market was not meeting, chip capacity, and they are seeking to override price signals and private capital allocations. The unavailability of labor to support the objective should be neither surprising nor discouraging. The projects are ones that private actors did not consider economical and were not pursuing. That’s what the industrial policy is for.

Legislation along these lines would have a great deal of substantive and political promise.

On Guardrails

During congressional debate, Republican opposition to the CHIPS Act focused on two issues: the downsides of subsidies generally and the failure of the CHIPS Act, specifically, to prevent subsidies from supporting investment in China. Both were reasonable concerns, and ones that will be present in all deliberations over industrial policy. Neither was a reason not to proceed with the CHIPS Act, but both warrant continued attention from policymakers.

As a tool of industrial policy, subsidies are usually the fastest way to alter market behavior. But in the long run, critics have a point: A subsidy war with other nations equally determined to attract investment is likely to waste resources and yield overcapacity. As policymakers seek to mitigate the damage done by American deindustrialization, and to prevent it from happening again, they will need to consider what sticks should accompany the carrots of CHIPS Act–style policy. Indeed, it was partly the lack of such sticks that permitted the degradation of American industry in the first place. This is, in essence, about guardrails—about how the United States will need to bound its market with policy tools like tariffs and local content requirements (LCRs). As Michael Lind explains:

Global gluts are not possible if tariffs are combatting the subsidized dumping by surplus countries. The profit-seeking firms that compete for limited consumer and business demand in a tariff-protected market—including, in some cases, “transplants,” or local subsidiaries of foreign firms—have no incentive to produce more goods than can be sold in the protected market. 

While the preference for subsidies instead of tariffs is irrational from the perspective of sound strategic trade policy, it serves the interest of politically powerful special interests. Investors and corporations devoid of patriotic loyalties naturally prefer a system in which multiple jurisdictions—cities, states, nations, or blocs—compete with one another to offer the most generous incentives for investment with the fewest strings attached. Likewise, the same self-interested investors and corporations would oppose tariffs or local-content requirements that force investments to be made in a particular jurisdiction because this precludes a race to the bottom among competing governments. Even if the goal is the same—encouraging the building of semiconductor fabs in the U.S., for example—the power dynamics are quite different, depending on whether subsidies or tariffs are used. Tariffs and local content requirements shift power from the firm to the territorial state and the national community it represents.

Similar logic applies to restrictions on investment flows. In the CHIPS Act negotiations, Republicans sought strict guardrails to curtail the investments in China a firm could make while also receiving federal support. American companies like Intel—strong supporters of CHIPS Act subsidies, for obvious reasons—were outraged at the prospect of such constraints, and unleashed an intense lobbying campaign to fight it.

With respect to the CHIPS Act specifically, Intel had a point. Imposing restrictions only on companies that accepted funding to build domestic fabs would only discourage precisely that behavior which policymakers were seeking to spur. The correct approach was not to leave the multinational corporations alone, though, but to impose blanket bans on such investment in China regardless of whether they received public funds.

Congress should take that lesson seriously and continue pressing for broader restrictions on the investment decisions of American corporations that are plainly not in the national interest. It’s an encouraging sign that, in this case, the more principled Republican objections were to too little government regulation. This healthy turn in conservative policy thinking continues to be reflected, for example, in the growing Republican support for reviewing outbound American investment. But as for those Republicans eager to make such interventions in the CHIPS Act context only because the relevant firms were accepting supply-side incentives to invest domestically—well, in that case, perhaps they could also sign on to restrictions applying only to any firms that accept a lower corporate tax rate.

Paving the Way

Lessons learned from the CHIPS Act on mission clarity, economic obstacles, and the need for global guardrails will all apply to future industrial policy efforts. But there is another lesson to be learned, too: some things Congress is not best suited to do.

Over a year elapsed between introduction and enactment of legislation to authorize a program to promote domestic semiconductor manufacturing. Only after another year and a half of debate did Congress appropriate funding. This started the clock on implementation, which ticked for another year and a half before announcement of the first CHIPS Act grant. Operating on this timeline where other industries or projects may warrant comparable support will always be cumbersome and contentious, compounding all the vulnerabilities that critics are right to highlight in industrial policy. The United States needs a more agile approach for deploying public funding to leverage private investment in critical sectors.

Congress should establish a permanent federal entity for this purpose. A “national development bank” or “industrial finance corporation” could allocate public funding in support of critical private investment on an ongoing basis, addressing more quickly and across industries the kind of challenges the CHIPS Act tackled for semiconductors. The bank could also be equipped with a wider range of tools (e.g., taking equity stakes, issuing direct debt, offering credit and completion guarantees, and supplying technical assistance). A permanent bank would also be able to develop state capacity, building a base of processes, expertise, and staff that ad hoc legislative efforts cannot create.

If the United States is going to build on the initial success of the CHIPS Act and address the economy-wide damage done by decades of disinterest in the nation’s industrial base, the CHIPS Act must cease to be a special case, the exception that proves the neoliberal rule. Industrial policy—done well and aimed at the right problems—should become a normal, integrated aspect of American economic policymaking.