“National security” broadly describes what a nation does to protect its essential interests. This is often viewed in solely military terms, but the reality is that many things besides the military balance affect national security.
Viewed from that perspective on national security, we can see that supply chains are an integral part of national security, and not just with reference to items specific to the defense industrial base. Aspects of infrastructure and services are so fundamental to the functioning of society that they, too, should be considered national security issues. Secure food and energy supplies, for example. Or public safety. Or protection against environmental threats. In some cases, shortages resulting from supply chain disruptions can develop in commodities that a nation must have. These could include pharmaceuticals and personal protective equipment, energy, food, and raw materials used in manufacturing.
In recent decades, supply chains have become increasingly dispersed, crossing numerous national borders. Enabled by improved communication and transportation technology, economic actors—predominantly private companies—have located parts of their supply chains in places where the materials cost, labor cost and availability, and regulatory environments are most favorable. Private interest, in the form of efficient production and use of resources, has driven the creation of dispersed but highly interconnected global chains.
These highly interconnected supply chains are a fact of life, and in many ways beneficial. Efficient production leads to company profits, distribution of capital across markets, improved productivity, lower prices, wider availability of goods and a host of other benefits.
But, with benefit comes vulnerability. Dispersed supply chains develop because actors find it economically advantageous to seek the least expensive and most productive sources of supply. While this may be individually beneficial for the actors, actions taken by a company or even a government organization to protect its supply chains do not necessarily promote collective protection of national supply chains. A company might find that its most efficient supplier resides in a company with serious policy or diplomatic disagreements with the United States.
The fact that the U.S. and the supplying country now have an economic tie in common does not guarantee that the policy differences will disappear or even be mitigated. Indeed, such interdependence may greatly complicate responses to geopolitical challenges, creating costs and risks where none were evident before. One profound example of this potential for complication lies in the semiconductor supply chain.
The case of Taiwan and semiconductors
Semiconductors are present in effectively every sector of the U.S. economy, as well as in every other advanced economy. The People’s Republic of China (PRC) is not a major player in advanced chip manufacturing. Its “rogue province” Taiwan, however, is not just a major player but, in some parts of chip manufacturing, a dominant one. Taiwan does not possess anything like the overall economic power of China, but it has built up a near monopoly in the production of high-end (less than 10 nanometers) logic chip semiconductors, largely through the Taiwan Semiconductor Manufacturing Corporation (TSMC).
TSMC’s dominance over the advanced semiconductor market—producing 94% of the most advanced logic chips—results both from some unique market conditions and from its diligence and careful management. TSMC is a technically proficient company operating in a portion of the microelectronics supply chain that is very capital intensive—and thus unattractive to companies seeking an immediately high rate of return. It has also received direct support from the government of Taiwan, which has served to put this company in the center of a supply chain vital to the world. Finally, it pursued a “global foundry model” with multiple customers, as opposed to the vertically integrated model pursued by Intel. Its dominance is in many ways the natural culmination of market impulses.
Taiwan’s position as the home of a company with a near monopoly on key parts of the semiconductor supply chain would seem likely to strengthen Taiwan’s importance to the United States (and the rest of the world). But the most important national security implication might go from protecting Taiwan’s autonomy to protecting access to a key material resource. While it might seem like the need to protect the United States’ access to semiconductors would strengthen the country’s historical commitment to protecting Taiwan, that may not necessarily prove to be the case.
No good option
In June 2022, to explore the geopolitical implications of Taiwan’s semiconductor dominance, the RAND National Security Supply Chain Institute conducted a tabletop exercise (TTX) with representatives from the executive and legislative branches of the U.S. government and from a variety of industries that rely on semiconductors. No single TTX can give a complete answer as to policy outcomes. This TTX did, however, demonstrate that there are generally only bad options for responding to the PRC attempting to coerce Taiwan in current circumstances.
Scenarios are ways of presenting reality and illuminating choices. They do not represent reality but explore a reality that could plausibly occur. In this TTX, RAND presented the players with two different ones, both intended to illuminate the impact of semiconductor supply chain vulnerability. Both began with a common set of conditions in which the PRC, for geopolitical reasons, imposed a coercive quarantine on Taiwan, as outlined in a recent RAND report. The scenarios diverged in Taiwan’s response to the coercive quarantine.
In the first case, rather than continuing to resist, Taiwan capitulates to Chinese demands, and the United States is forced to deal with a PRC now in possession of a near monopoly on high-end semiconductor manufacturing and a healthy portion of other semiconductor manufacturing. In the second, Taiwan attempts to resist, resulting in the PRC taking actions that increasingly disrupt Taiwanese semiconductor production, and thus supply of high-end semiconductors to the U.S. and the world.
In the first scenario, U.S. industry players sought to continue business as usual, while legislative and executive participants sought paths to alternative supply. However, the actors generally did not view this as a catastrophic outcome. The attitude of many industry players was that U.S. industries routinely do business with Chinese suppliers and that while the dominance of the PRC over high-end semiconductors might result in complications, they would not necessarily imply any major change in existing trade or contractual relationships. Government players were more focused on intellectual property and security implications, but no group necessarily saw a change in the national ownership of TSMC’s semiconductor “fabs” as catastrophic.
In the second scenario, Taiwan resists the initial demand, and the PRC steadily increases pressure on Taiwan, beginning with a demand for a curtailment of exports from Taiwan to the U.S. and moving steadily upward toward increasing disruptions of semiconductor production. In making these demands, the PRC understood that it would be hurt economically to the same degree or greater than Taiwan’s partners, but it opted to continue with pressure to achieve a long-standing political end. Throughout the process, the PRC offered an immediate lessening of pressure in exchange for Taiwan accepting the political condition of unification. At no time did the PRC offer an armed intervention beyond the imposition of the blockade/quarantine that it had already initiated.
The U.S. teams found that they had few desirable choices as the pressure continued. The U.S. always had the option of trying to impel Taiwan toward a settlement that would preserve access to semiconductor chips even if at the expense of its autonomy. Without U.S. support and security guarantees, Taiwan would rapidly find itself isolated. Although no U.S. team advocated this, all understood that this could become a very real possibility.
A second option would be to attempt a radical decoupling from both Taiwan and the PRC and develop “friendshored” sources of supply. Such changes likely could not occur in the short term. They would take time, capital, an available workforce, and possibly changes in technology, on a timeline that would likely exceed the time Taiwan could reasonably be expected to withstand pressure. For example, we know that it would take the United States and allies two to five years to build and outfit sufficient fabrication capacity to offset the loss of Taiwan’s production. This timeline includes optimistic assumptions regarding tooling, permitting, and the labor market. Developing sources for other commodities more directly controlled by the PRC—such as processed minerals—would also take time, cooperation, resources, and possibly significant policy changes.
Friendshoring could also be coupled with imposing counter sanctions against the PRC—in hopes of creating costs the PRC would find difficult to bear—and providing incentives for manufacturing in friendly countries. As an autocratic society, however, the PRC might be better able to harness the whole of government and private economy to pursue objectives. It would certainly be hurt by efforts to exclude it from markets, possibly more than the U.S. and its allies, but the question then turns to how long the different societies could withstand the disruption. The TTX did not specifically examine this. However, we know from the response to COVID-19 that the PRC has considerable capability to lock down its population and accept diminished levels of economic production. Worth bearing in mind is that the timelines for Taiwan’s collapse in the face of pressure are considerably shorter than the timeline for creating greater levels of supply chain resilience in the rest of the world.
The TTX specifically took military actions out of play, but the game pointed to the challenge of having few options between acceptance of the PRC’s demand or responding with military force. “Tit for tat” responses proportional to the provocations generally were not available, largely because the consequences of supply chain disruptions were immediately dire for the global economy.
Next potential steps
The TTX was, as mentioned already, one representation with a set of assumptions about behavior that might not prove accurate in the real world. The fact that it highlighted difficult options does not imply that no effective action could ever be taken. But this TTX and other efforts strongly suggest that the U.S. and allies must form partnerships, partnerships that must include industry, to increase supply chain resiliency and offer leaders something other than poor choices. The following are a few preliminary steps:
- Both the public sector and the private sector should improve their analysis and understanding of the semiconductor supply chain specifically and the overall level of supply chain interdependence in general. From a geopolitical perspective, many of the planning scenarios that address how to handle a potential conflict over Taiwan’s autonomous status do not include the loss of Taiwanese semiconductor capacity as a likely consequence. This consequence deserves significant consideration.
- An immediate and concerted effort must be made to reduce the concentration of semiconductor production in Taiwan. This condition not only is dangerous to the world’s economic well-being, but it also actually increases Taiwan’s vulnerability. Reducing this concentration will take several years. The management of vulnerability is thus to a very large degree a matter of timing. There are several steps that should be taken:
- TSMC must be incentivized to relocate production out of Taiwan. This does not imply moving all production, nor does it necessarily imply transfer of ownership. It means geographic relocation of production to places without as much geopolitical significance as Taiwan. Reducing the risk of semiconductor disruption because of Chinese aggression would increase the willingness of the United States and allies to support Taiwan should aggression occur. This should be a powerful incentive for Taiwan.
- Irrespective of TSMC actions, governments should take action to strengthen domestic and/or allied semiconductor production. Action does not imply top-down direction for investment, at least not in every case. It does involve creating incentives for investment and creating opportunities for workforce training and/or liberalized immigration. It probably also involves management of intellectual property sharing with a clearer eye toward the security impacts of sharing designs, even those without an obvious defense tie. There may be designs that should only be accessible to producers inside the United States or preferred allies.
- Movement of facilities and equipment to the PRC should be specifically discouraged and heavily regulated. If markets are incentivized to invest in the PRC and/or sell Chinese companies advanced equipment, both are likely to occur. Eliminating such incentives is likely to require coordination with allies and does go against the normal imperatives of a market economy. Incentives need to be structured in ways that industry will see as effective.
- Collaborative relationships with allies, industries, and governments are essential, even if these appear counter to the normal impulse to separate sectors. The interdependencies created by supply chains are complicated and extensive, with individual and collective interests intertwining to a degree that neither market nor normal government decision-making will be sufficient. This complexity requires extensive consultation, to the point that the relationships may of necessity be “cozier” than most democratic governments or private industries would prefer. The relationship between public and private will require careful management, as will the relationship with allies who have their own private-public challenges. But the TTX reinforced that neat separations between private and public interest are not possible in this context.
