Supply Chains, National Security, and the Limits of Industrial Policy
America has real strategic vulnerabilities. Our scattershot industrial policy is failing to address them.
The debate over trade and industrial policy has increasingly been framed as a choice between two camps. On one side are defenders of free trade and free markets who are broadly correct on the economics, and who rightly point out that many efforts at industrial policy are wasteful and driven by special interests. On the other side are those seemingly prepared to classify half the economy as “strategic” or “vital to national security,” and to use those labels to justify tariffs, subsidies, and government favoritism on an almost limitless scale.
Neither side has the right focus. The actual national security problems are serious enough without exaggeration, and free traders need to acknowledge these threats as well. China has the ability to choke off key inputs that sit deep inside U.S. military supply chains. But that is a much narrower and more concrete problem than what much of today’s industrial policy seems to be targeting. Strategic vulnerabilities are real but they do not form a general case for protectionism or for government steering the economy. Misdiagnosing the problem has already led to scattershot policies that are too broad, too blunt, and in many cases self-defeating.
So what industrial policy is actually justified? When we use the words “industrial policy,” we mean deliberate government efforts to favor particular industries, technologies, or production capabilities rather than letting market forces alone determine what gets built and where. This could take the form of subsidies, tariffs, procurement preferences, tax incentives, regulatory carveouts, or other policies designed to tilt investment and production toward favored domestic sectors.
Where Free Trade Breaks Down
Standard free trade theory is broadly correct and has been essential for economic growth. Specialization and free and open exchange raise living standards, while government intervention generally distorts that logic and tends to produce waste. Still, every model rests on a series of assumptions. We argue that three key assumptions behind conventional trade theory break down in important ways when it pertains to China and industries that concern U.S. national defense.
The first assumption is that trading partners are largely neutral actors who will not weaponize supply chains. China is not a neutral actor. Its leadership has repeatedly demonstrated a willingness to weaponize supply chains against its adversaries. In 2010, China cut off rare earth exports to Japan during a maritime dispute. Similarly, between 2024 and 2025, Beijing imposed export controls against the U.S. on gallium, germanium, antimony, and rare earth elements, which are precisely the inputs that go into our semiconductors, drone motors, missile guidance systems, and radar electronics. The controls are no longer at their 2025 peak; however, the architecture remains in place. When your trading partner retains the capability to cut you off at the worst possible moment, the expected value of that dependency shifts dramatically.
The second assumption is that comparative advantage reflects genuine country differences (e.g., geography, culture, resources) and not strategic attempts at control. China’s dominance in rare earth refining did not come as a consequence of its unique qualities that made it predisposed to being a better producer. Instead, it was manufactured over decades through state subsidies, deliberately lax environmental enforcement, and systematic limit pricing designed to eliminate Western competition whenever and wherever it emerged. When a market outcome is engineered in this way in part in order to achieve dependency of other countries and strategic dominance, the standard argument for accepting it wholesale weakens considerably.
The third assumption is that disrupted supply chains can easily be rebuilt. This assumption falls apart with a mere cursory look at the current U.S. market. The average U.S. mine-to-production timeline is 29 years, second longest only to Zambia. If a conflict with China disrupted critical material flows tomorrow, domestic production capacity could not reasonably compensate on any militarily relevant timeline.
The consequences of these dependencies and existing policies are stark. Data from Govini, which is an analytics firm that works directly with Department of Defense procurement records, show that over 40 percent of the semiconductors sustaining DoD weapons systems depend on Chinese suppliers. 78 percent of U.S. weapons systems contain components with gallium, germanium, antimony, tungsten, or tellurium (i.e., minerals all subject to Chinese export controls between 2024 and 2025). Roughly one in ten Tier 1 defense subcontractors is a Chinese firm, and the number of Chinese suppliers in the defense industrial base has quadrupled since 2005.
These percentages are not merely abstract supply chain risks. Among the weapons systems with Chinese semiconductor dependencies are the B-2 Spirit Stealth Bomber and the Patriot Missile Defense System. These weapon systems are absolutely central to American strategic deterrence and air defense. Thus, the supply chains that we would rely upon in a hypothetical conflict with China run through China.
The specific mechanism is worth understanding clearly. The United States is the world’s second-largest miner of rare earth ore. That initially sounds reassuring until you learn that we export over 95 percent of that ore to China for processing since we have almost no domestic refining capacity. In 2024, the entire United States had one operating rare earth refinery in Texas, which produced just 1.3 kilotons of alloy. To put into perspective how paltry that number actually is, consider that in that same year, China produced 240 kilotons of finished rare earth magnets. These are the same magnets that go into F-35 guidance systems, Patriot interceptors, and Virginia-class submarines. China currently controls 92 percent of global rare earth refining.
The bottleneck has nothing to do with our supply of ore or insufficient capital in our financial markets. Domestically, we have plenty of ore, and capital markets are fully capable of funding refinery construction when the risk profile is acceptable. The problem is regulatory. The average U.S. permitting timeline for a rare earth refinery is 7-10 years. U.S. allies Canada and Australia manage to do the process in roughly 2 years. There is no physical or technical reason our process needs to take as long as it currently does. This is purely an issue of our own government getting in its own way and operating against its own national security interests.
The policy response to date has gone well beyond what the evidence supports or what can be meaningfully defined as national defense. Defense-grade semiconductors, rare earth magnets for guidance systems, gallium and germanium for radar electronics are the obvious places that warrant national security concern. However, we are currently applying Section 232 national security tariffs to all sorts of goods that under no reasonable definition should apply. These include upholstered furniture, kitchen cabinets, softwood lumber, and passenger vehicles. No American should care if China makes cheap patio chairs or even passenger EVs. The correct response to cheap consumer goods from some state-subsidized competitor is not a national security tariff. It is to require safety standards, software audits where relevant, and accept the gains from trade wherever genuine national security is not actually at stake.
If a country wants to subsidize an inefficient market that ultimately benefits Americans in the form of cheaper prices, then so be it, but more importantly, if everything is a national security issue, then nothing truly is.
Addressing the Threats While Stopping the Waste
So, what is the right response? Four interventions stand out, and three of them cost little or nothing. Only one of our proposed interventions is really industrial policy in the classic sense. The others are better understood as procurement reform, transparency, and removing domestic barriers to production.
First, permitting reform for critical mineral refining. Streamlining NEPA review, consolidating federal processes, and creating pre-approved site categories for refining hubs are administrative actions. If our allies are permitting comparable facilities in a fraction of the time, the gap is procedural, not technical. This is a form of industrial policy only in the mildest sense: not government trying to pick winners through subsidies, but government getting out of the way in targeted ways where its own rules are blocking strategically important production. No legislation is required to optimize federal processes, althogh legislation may be necessary to prevent states from blocking the expansion of strategically important industries.
Second, we should extend supply chain audits through the lower tiers. Country-of-origin certification requirements currently stop at just Tiers 1 and 2 in defense procurement. The Chinese semiconductor exposure in the B-2 and Patriot production is happening at Tiers 3 and 4, which are the suppliers to component level subcontractors and suppliers of raw materials and commodities. Requiring prime contractors to certify the provenance of what their suppliers’ suppliers are sourcing is a procurement rule change, not a spending bill, although it would increase the costs of inputs that no longer will be sourced from China. This is not industrial policy in the classic subsidy/tariff sense, but it is a state-directed intervention in industrial organization and supply chains.
Third, we should use multi-year purchase commitments allowed under Defense Production Act Title III to de-risk private capital investment in rare earth refining. The market is not building this capacity because the demand signal is simply not there. A long-term government offtake commitment (e.g., the DoD’s existing arrangement with MP Materials, guaranteeing 7,000 metric tons per year of magnet purchases over a decade) gives private investors the contracted cash flows to justify billion-dollar capital expenditure. The DoD has already committed over $400 million under DPA Title III to rare earth supply chains since 2020, but this effort will need to be scaled.
Finally, there is also a structural fix worth considering to reduce the likelihood of having non-essential industries classified as protected under a national security pretext. If a tariff is going to be classified as a national security measure, it should require the Secretary of Defense to certify a specific military interest. That single procedural requirement would have preserved all of the interventions that actually matter while preventing the patio chair tariffs that do not.
The United States has a real problem with its strategic supply chains, but it also has a tractable set of solutions at its fingertips. The most important step is deciding to solve the actual problem rather than using national security concerns as cover for a much broader protectionist agenda. Precision does not just matter in weapon systems—it matters in policy as well.