The 2022 CHIPS and Science Act was the United States’ most ambitious foray into industrial policy in more than half a century. The bill included roughly $50 billion to revitalize the U.S. semiconductor industry, which had been hollowed out over decades as manufacturing migrated overseas. Industrial policy, long eschewed in policy circles, had come back in vogue as a way to strengthen supply chain resilience for industries critical to economic and national security. Moving the legislation through Congress required a multiyear process that involved lengthy negotiations and complex maneuvering. But passing the law was just the beginning. As soon as it was signed, the action moved from Congress to the Department of Commerce, which had to figure out, quickly and with little room for error, how to stand up a new office and infrastructure to deliver on its aims.
Washington was out of practice when it came to industrial policy, which the United States had largely abandoned since the Cold War. So there was no playbook available to the new CHIPS Program Office, which was established to administer the bill’s $39 billion in semiconductor manufacturing incentives (another $11 billion or so was for semiconductor-related research and development). Where possible, the office drew lessons from past large-scale grant and loan initiatives, such as the 2008 Troubled Asset Relief Program; the 2009 Presidential Task Force on the Auto Industry, which bailed out automakers; and the Department of Energy’s Loan Programs Office, which has financed more than $40 billion in clean energy and advanced transportation projects. But much of the time, the program was building the plane as it flew.
It’s too early to pass a final judgment on how well it did: the semiconductor industry is fast-moving and cyclical, and many projects are just getting underway. But the program has made real progress. In two and a half years, it has unlocked more than $450 billion in private investments, helped support the construction of 17 new semiconductor fabrication plants (known as fabs), and made the United States the only country to secure manufacturing commitments from the world’s five leading chip manufacturers. When the CHIPS Act was passed, the United States produced none of the most advanced logic or memory chips, the hardware used in smartphones, laptops, and powerful AI systems; it is now projected to produce 20 percent of the world’s leading-edge logic chips by 2030 and ten percent of its dynamic random-access memory chips by 2035.
CHIPS set itself up for success in part thanks to clever statutory drafting and early choices on hiring, goal setting, and structure. Amid bipartisan consensus to expand industrial strategy beyond CHIPS, future programs would do well to learn from its example. Of course, there is no one way to do industrial policy, and the details will vary by sector and circumstance, but one constant is that large federal programs require the government to work quickly, efficiently, and well. That requires establishing nimble and dynamic teams with sophisticated expertise; building productive and transparent relationships with industry, other governments, and the general public; and figuring out how to overcome sources of delay—within government and outside of it—that make it hard to build new things. Most important, industrial policy requires clarity of purpose: a concrete set of specific objectives to guide investments and against which to measure success.
The first days of any new program are when the most critical decisions are made. For CHIPS, having a director that did not require Senate confirmation was an early advantage, since it meant that Michael Schmidt—whom Commerce Secretary Gina Raimondo tapped to head the office—was able to get working right away. Requiring a Senate-confirmed leader for an industrial policy program would, in most cases, be a mistake: confirmation takes a long time and can dissuade people from taking on the role. During the Biden administration, confirmations took an average of 192 days from official submission—nearly three times as long as they took under President Ronald Reagan. That is time an industrial policy program cannot afford to waste.
Most government programs take months to get off the ground. In its first few months, however, the CHIPS team released its first funding opportunity for commercial chip fabs, published an investments strategy paper, and hired dozens of staff across investments, strategy, legal, and external affairs. Raimondo also appointed a chief investment officer with private-sector investment experience and credibility with large industry players to work with Schmidt. Hiring from the private sector was essential for an industrial policy program intended to go toe-to-toe with some of the biggest, most sophisticated companies in the world. At the same time, the point of industrial policy is to meet economic and national security objectives, not just earn market returns. So the CHIPS team also needed staff that could evaluate deals based not just on financial terms but also by judging whether they would improve supply chain resilience, meet the needs of the defense industrial base, and strengthen cybersecurity.
Industrial policy is not about onshoring all production to the United States.
Hiring for these roles is easier said than done. By broad consensus, federal hiring is a mess. In 2024, it took the government more than 100 days, on average, to hire new employees. But even that number understates the problem, as it includes various forms of expedited hiring that are often not possible. Typical government hiring, through competitive public job announcements, takes significantly longer. What’s more, federal offices responsible for screening résumés rarely have the financial and technical expertise to evaluate applicants for specialized jobs and tend to review résumés more for consistency with the wording of the job announcement than for the substantive qualifications needed for the role. That’s one reason that, according to Jennifer Pahlka, who helped found the U.S. Digital Service, roughly half of all competitive federal job announcements fail, resulting in no hire.
The Office of Personnel Management (OPM), the independent agency that oversees the federal workforce, should systematically overhaul and improve government hiring. But in the meantime, high-priority programs need to be creative to get good people in the door quickly. For CHIPS, bureau heads in the Commerce Department temporarily loaned staff to get the nascent office off the ground. To build up its full-time staff, meanwhile, the CHIPS Office sought permission from the OPM to use two expedited hiring authorities that were critical to the program’s success: direct hiring, which bypasses most HR screenings, and excepted service hiring, which allows managers to hire without HR involvement. CHIPS hired more than 150 people using expedited authorities. The CHIPS Act itself also included a special hiring pathway, allowing the program to recruit its most senior staff quickly and pay them above-standard government rates.
For industrial policy to work well, recruiting is essential. The government needs to draw sharp minds from finance, industry, and other federal agencies. For workers in the private sector, a federal job means much lower pay; it might also mean long commutes or relocation to Washington. During the early months of the CHIPS program, CHIPS relied on personal outreach from Raimondo and from office leadership. Their pitches were simple: get in on the ground floor of a once-in-a-generation experiment in industrial policy. When the ask was made directly, most people said yes.
Equally important to staffing is developing an investment strategy—figuring out what, exactly, the United States is buying with its congressional funding. This strategy should be clear, provide concrete goals and metrics, and be accessible to the public so that Americans can judge whether the program is succeeding on its own terms and whether those terms are the right ones. Such a strategy, though hardly unique to industrial policy, is critical for a program’s operational and political success. Not every worthy project can receive funding. By specifying clear criteria by which applications for funding would be evaluated and by publishing a vision of success that set specific production objectives across the semiconductor industry, the CHIPS office could manage expectations for stakeholders both within and outside of government.
Importantly, the vision for success served as a guide to actual investment decisions. CHIPS tasked teams to evaluate how every potential deal would advance the program’s aims. In working on individual deals, however, it’s easy to get caught up in negotiations with specific companies over minute details and lose sight of the bigger picture. To offset that effect, the office also established an investment committee, composed of CHIPS leadership and a few outside experts, whose job was to recommend deals to a committee of senior Commerce Department leaders based on a portfolio-level view of the office’s investments. To manage industry expectations and ensure that there was enough money for a broad range of projects, CHIPS made clear that successful applicants would generally receive grants worth no more than 15 percent of a project’s capital expenditures.
A successful industrial policy also must incorporate foreign, state, and local governments. Foreign governments matter because modern goods have complex global supply chains and are sold into global markets. Industrial policy is not about onshoring all production to the United States. Even had that been the goal for CHIPS, $39 billion would not have been enough. Instead, industrial policy should identify which components must be made in the United States and work with friendly partners to create robust and resilient global supply chains for those that cannot be. For CHIPS, those partners included Costa Rica, Europe, Japan, Mexico, South Korea, and Taiwan, many of which had their own chip subsidy programs. To avoid duplicative efforts and to prevent companies from playing countries off each other in search of the highest award, CHIPS hired an international team responsible for monitoring global subsidies, developing a global strategy, and working with the White House and State Department to communicate or reinforce specific requests of foreign partners. CHIPS also developed close relationships with state and local governments, which provided their own subsidies in addition to controlling site selection and various permitting levers.
The same political will that drove the passage of CHIPS—and which may drive future industrial policy bills—also created a demand for fast results. But government processes rarely move quickly, and there are lots of opportunities for outside players—other agencies, courts, and Congress—to slow momentum.
Government differs from the private sector in the sheer number of actors that can impede progress. To be sure, officials outside an industrial policy program—in a legislative affairs office, on a communications team, or at the National Security Council—can spot gaps or missing priorities. But although many of them know the importance of their functions, few are able to trade off between them—to decide, for example, when a national security imperative outweighs a labor concern, or vice versa. Resolving such disagreements can slow things down. CHIPS was largely able to avoid such delays thanks to a statute that granted substantial discretion to Commerce and a funding process that gave the CHIPS office the flexibility to make policy tradeoffs during negotiations.
Other programs may not be so lucky. It could help for them to have a full-time coordinator at the White House, as CHIPS did, tasked with moving the program forward and resolving conflicts before they resulted in significant delay. A White House coordinator can also help ensure a comprehensive approach to industrial policy that aligns funding with other tools, such as export controls, tariffs, and government procurement.
Designing an industrial policy involves difficult tradeoffs between speed and nuance.
Courts can serve as another source of delay. In theory, nearly any program decision can eventually be litigated in court, and even if no suit is ultimately filed, processes are designed in the shadow of potential litigation. Consider the National Environmental Policy Act, a law that requires the federal government to evaluate possible environmental impacts of proposed action. NEPA reviews come on top of ordinary permitting requirements and apply only to federal projects. Although well intentioned, such reviews can take years, and if an opponent of the project sues, even longer. The original version of CHIPS required NEPA reviews for CHIPS projects; to comply, the office built an environmental team and published a 200-page programmatic environmental assessment. This strategy would have sped up reviews, but it would still have been vulnerable to lawsuits and considerable delay. In the end, two years after passing the CHIPS Act, Congress exempted most CHIPS projects from NEPA, easing this burden. But including a NEPA waiver in the initial law, and broadening it to cover all projects, would have saved a lot of time and effort.
Congress can also act as a source of friction. Ideally, the statute that creates any new program will include sufficient money and authorities to achieve its goals, but that is not always the case. Asking Congress to address an urgent need can take months or even years, as was the case for the NEPA exemption for CHIPS projects. Careful legislative drafting—including waivers of other laws where necessary—can make a big difference. Appropriating the entire budget of the program up front can help a program make long-term plans without worrying about political disruptions.
Finally, state and local governments can create permitting bottlenecks, which lead to years of delay. States want to win big projects, however, so the best approach to such obstacles is to set up a competitive dynamic among them. CHIPS, for its part, explicitly rated applications based on how quickly they could be constructed, including permit timelines. This created incentives for interested states to strengthen their processes and adopt reforms, such as creating fast tracks and a single permitting point of contact for a project. CHIPS also engaged early and often with local governments to break through logjams when they arose.
When making difficult decisions, the U.S. government often justifies outcomes by reference to a scrupulously fair, diligent, and transparent process. Such processes often lead to the right outcomes, and strictly following them can help protect ambitious government programs from legal consequences and reputational harm when some investments inevitably fail. But they can also be inflexible, counterproductive, and time-consuming. A successful program thus hews to standard rules where appropriate but seeks workarounds where necessary.
Take, for example, the rules governing federal grants. The rules run 180 pages long with 12 appendices, and it can be difficult to parse what is mandatory and what is merely standard practice. Because many of these rules are ill suited to ambitious industrial policy, Commerce appealed to Congress to include in the bill a special exemption from grant rules known as “other transaction authority.” A typical government grant simply reimburses applicants for the cost of their activities, which means that negotiations over cost are limited if they happen at all. But CHIPS wanted to have holistic negotiations with companies over how many and what kind of fabs companies would build and how much funding the government would provide—an iterative process that was critical to ensuring projects met ambitious economic and national security goals. Unlike the traditional government grants process, in which applicants set project scope, most CHIPS negotiations involved pushing applicants to build more, which paid off handsomely when companies such as TSMC agreed to construct additional fabs in the United States.
Another source of procedural delay are statutes such as the Paperwork Reduction Act, which requires two separate comment periods to evaluate the necessity and burden of providing information to the public, making it difficult to quickly announce funding opportunities. Procurement laws, meanwhile, can require monthslong competitions among potential bidders before a program can purchase outside services. A critical job for early staff is to identify and, to the extent possible, work around such obstacles by leveraging statutory exceptions and other available “fast track” procedures.
Finally, designing an industrial policy involves difficult tradeoffs between speed and nuance. One example is the complex process the CHIPS office designed to determine grant subsidies, which included calculating the rate of return for proposed projects. To avoid wasting taxpayer dollars, CHIPS deal teams sought to bring the rate of return for CHIPS projects close to the company’s standard rate of return. But rate of return is a difficult number to model, and small changes in assumptions can lead to big changes in the result. A simpler approach might have opened negotiations over grant size at a fixed percentage of total costs, avoiding drawn-out debates over model assumptions. Such an approach, however, would have been less accurate and might have resulted in overpaying companies that could have executed projects with less government money. CHIPS calculated that a more involved process was worth it for larger fab projects, particularly to signal to the biggest companies that the team was ready for serious, financially sophisticated negotiations. When it came to designing a funding process for semiconductor suppliers, which are smaller and sometimes less well-resourced than other program applicants and for which long negotiations are impractical, CHIPS followed the simpler approach, fixing most awards at ten percent of the project’s capital expenditures.
CHIPS was not without its critics. The program faced early pushback, for example, for embodying “everything bagel liberalism”—an attempt to shoehorn too many social and economic goals into a national security program. When the first funding opportunity was published in February 2023, critics took issue with language that encouraged labor agreements and required applicants requesting more than $150 million to submit a plan to provide childcare to their workforce. There were no signs that the language deterred companies from participating, and many companies spoke publicly about the need for affordable childcare options to meet workforce demands. But the blowback sapped some of the bipartisan support for the program, and the Trump administration has resurfaced some of the criticisms. This is not to say industrial policy programs should never include additional policy requirements, but these should always be in service of the overall goal: onshoring sufficient production capacity for technologies critical to economic and national security.
In addition, getting early design choices right does not inoculate a program against issues down the line. This is especially true for programs, such as CHIPS, that target a cyclical and sometimes unpredictable industry. When Intel’s stock nosedived in 2024, for example, the CHIPS team had to scramble to finalize a deal with the company to secure important projects while safeguarding taxpayer money in the event of the company’s collapse. NEPA delays remain a concern for projects that were not exempted from its reviews. And some external statutory requirements, such as retroactive prevailing wage requirements under the Davis-Bacon Act, have slowed and complicated funding negotiations.
Doing industrial policy well is not easy. It requires scaling up production quickly and carving out a place in government that can act like a private-sector startup by staffing up, implementing new processes, and finding creative ways to avoid delays. The U.S. government, bogged down by rules and bureaucracy, is not the ideal home for a startup. But with the right playbook, it can become a more hospitable one. As CHIPS demonstrated, policymakers need to make government an attractive place to work for people from finance and industry, identify and mitigate sources of delay, and measure success against clear and definable metrics. At the same time, policymakers should push for structural reforms to make government work better. Under the best of circumstances, industrial policy is an uphill battle. Making smart early decisions can make the climb a little easier.
Loading...
Please enable JavaScript for this site to function properly.