The United States made hundreds of millions of doses of safe and effective COVID-19 vaccines available by early 2021, a little more than a year after the pandemic began. This impressive feat was largely achieved by Operation Warp Speed, the public-private partnership involving the Department of Defense and Department of Health and Human Services. The initiative—often by relying on the Defense Production Act (DPA) of 1950—accelerated the research, development through clinical trials, and early manufacturing of a diverse array of potential candidate vaccines.
But by early summer 2021, US production of COVID-19 vaccine doses mostly stopped growing. Production in the European Union surpassed the United States, as even Pfizer-BioNTech and Moderna—the two biggest vaccines being made in the United States—expanded operations more in Europe.
The missing US supplies aggravated shortages in a world still desperate for vaccines. Scarcities were especially acute for the highly effective, but technologically complex, messenger ribonucleic acid (mRNA) versions that Pfizer-BioNTech and Moderna were manufacturing in very few places and thus would reach most of the world only through international trade.
The reason why US production lagged after a strong start remains a mystery. But new research suggests that how the US government used the DPA to contract with vaccine manufacturers and some input providers to implement Operation Warp Speed may have been partly to blame.
Americans got doses early because the DPA "priority-rated" contracts agreed to in 2020 allowed the US government to take on some of the financial risk to encourage vaccine companies to invest in costly production facilities without knowing whether the Food and Drug Administration (FDA) would authorize their vaccines for use. The primary aim was to obtain enough doses as fast as possible to vaccinate the US population.
The use of the DPA in this way to implement Operation Warp Speed, however, had unintended consequences. DPA rules left unclear when US plants manufacturing those same vaccines would be permitted to export doses globally. This uncertainty partly prompted companies such as Pfizer and Moderna to expand production more elsewhere to meet global demand. To get additional vaccine doses quickly to a world in need, the US government would have needed to recalibrate and clearly articulate its policies in early 2021 to encourage those companies to further expand their US production capacity for overseas markets.
The good: Large and early US deliveries of COVID-19 vaccine doses
Operation Warp Speed was novel in its design. The US government invested in a diversified portfolio of vaccine candidates in the early months of the pandemic, in case any of them failed. It also subsidized many of the costly and lengthy Phase 3 clinical trials needed to determine whether the vaccines were effective and safe.
These US government incentives were essential for companies to build their manufacturing supply chains early and "at risk." Firms received funding to begin the manufacturing process simultaneously with Phase 3 trials, and thus before knowing whether the FDA would authorize their vaccine for use. The benefit of reducing some financial risk borne by the firms was that once the FDA approved the vaccines, companies could deliver doses almost immediately. (Had the government not stepped in this way, companies may have waited to construct their costly production processes until after the results of the Phase 3 trials, extending the period without vaccines another four to six months.)
As expected, some vaccine candidates failed in these trials. US funding in those instances did not pay off. But subsidies for candidates that passed clinical trials accelerated the delivery schedule of hundreds of millions of vaccine doses to the US public in early 2021, a rapid timetable that became the envy of much of the world.
According to data from Airfinity, a British-based global health intelligence firm, US plants delivered 103 million doses by the end of February 2021 (figure 1). In contrast, India provided 40 million doses and the EU only 27 million. By the end of March, the United States supplied nearly 200 million doses, compared with 117 million for India and 140 million for the EU.
The US success benefited from luck as well as planning, but even its planning was not perfect. Tens of thousands of lives and livelihoods were being lost each month, along with trillions of dollars of economic activity globally. Some public health experts argued that the US government should have been more ambitious. Ahuja et al. (2021), for example, maintained that in 2020 the US government should have funded 27 vaccine candidates (not six) and spent more than three times as much money (not $18 billion).
Operation Warp Speed was created to benefit Americans, but it also generated what economists call positive "externalities" for the rest of the world. Expedited clinical trials costing hundreds of millions of dollars, coordinated and paid for by US government subsidies, would not require repeating elsewhere. Instead, other governments could focus on subsidizing manufacturing capacity at risk. Nevertheless, most governments failed: Too few invested and at too low levels.
The main criticism directed by some experts at the US government was not focused on its vaccine manufacturing policy but on its unwillingness to share those early, US-produced doses internationally. That failure, coupled with multiple failures of other governments, led to delays and shortfalls of global vaccine supplies in 2021. Only exports from vaccine manufacturing countries, including the United States, could meet much of the global demand.
Yet, the very policies that led to early successes in US production, somewhat perversely, also contributed to limiting US vaccine exports.
The disappointment: Failing to expand US production more in 2021
In mid-2021, the tables turned as European and Indian COVID-19 vaccine production started exceeding US production levels (figure 2). There was no shortage of global demand. By the end of 2021, the EU supplied 160 percent more doses than the US plants, and even India produced over 60 percent more. The United States eventually did export more than 50 percent of its local production—most of it in the final six months of 2021—compared with over 60 percent for the EU and less than 5 percent for India. But because US production lagged, its exports were not enough to satiate global demand. Something on the US supply side had gone wrong.
Though not the best comparison for the United States, India's vaccine production and export record is an important cautionary tale of its own. The Serum Institute of India (SII), the largest vaccine manufacturer by volume in the world prior to the pandemic, dominated India's supplies. (SII manufactured 1.4 billion doses of the AstraZeneca-Oxford vaccine by the end of 2021.)
Despite this eventual record of production success, India contributed to the problem of late and insufficient global vaccine supplies in many ways. First, the Indian government failed to offer subsidies to its vaccine manufacturers until April 2021. That lack of at-risk and early public investment helps explain why India was not a large early supplier (see again figure 1). Second, the insufficient early supply led India to ban vaccine exports between April and October—in response to a massive wave of COVID-19 cases in India in the spring of 2021—further contributing to the global supply shortfall. This export ban cut off shipments to the COVID-19 Vaccines Global Access (COVAX) program—which, unlike the Indian government, had provided some at-risk funding in 2020 to SII (via Gavi and the Gates Foundation) to manufacture doses needed to vaccinate low-income countries in particular. (SII ultimately manufactured more than 600 million doses between April and October 2021 that could not be exported.)
The European Union offers a better comparison for the sputtering trajectory of US production because it also ended up manufacturing a similar portfolio of vaccines. The EU's superior supply performance by the end of 2021 is striking. To clarify, the European-US disparity is only partly explained by the EU manufacturing four regulator-authorized vaccines locally, compared with three in the United States. (The FDA never authorized AstraZeneca for use in the United States, so it was deprioritized from US production in mid-2021.) The discrepancy is also only partly explained by problems experienced by the Johnson & Johnson vaccine in early 2021. The Emergent BioSolutions plant in Maryland, which was charged by the US government in 2020 to manufacture both the AstraZeneca and Johnson & Johnson vaccines, failed to follow good manufacturing practices. As a result, roughly 300 million doses had to be destroyed due to cross-contamination, greatly wiping out potential US production. (The FDA was also forced to shut down the Maryland facility for four months.)
Even taking these problems into account, a major puzzle persists over why US production faltered after its early success. EU supplies of all three vaccines overlapping with US production—not just Johnson & Johnson, but also, more importantly, Pfizer-BioNTech and Moderna—ended up considerably higher than US supplies by the end of 2021.
According to data from Airfinity, European plants supplied 1.8 billion doses of Pfizer-BioNTech and 360 million doses of Moderna (see again figure 2, panel b). US facilities added only a little over 600 million doses of Pfizer-BioNTech and 300 million doses of Moderna to global supplies.
Why Pfizer-BioNTech and Moderna manufactured so many more doses in Europe than in the United States
Pfizer-BioNTech and Moderna announced increases in production scale in both the United States and European Union in early 2021, following regulatory authorization of their vaccines. Each added more plants or production lines to manufacture the vaccine's drug substance. Each also outsourced the "fill and finish" part of the manufacturing process—by which vaccines are put, assembly-line style, into millions of vials for distribution—to contract development and manufacturing organizations (CDMOs). Why then didn't Pfizer-BioNTech and Moderna increase their US production more?
One rough measure of capacity is monthly supplies. In Europe, monthly supplies grew steadily over 2021 (figure 3). Peak European monthly supplies of both Pfizer-BioNTech and Moderna in 2021 ended up at roughly twice their peak US monthly supplies, respectively.
The failure to expand US capacity was partially a legacy from the downsides of Operation Warp Speed's otherwise successful 2020 approach.
One problem was that vaccine manufacturers were uncertain over when they would be legally permitted to export from their US plants. Their US government contracts under the DPA had given them priority access to inputs to make vaccines. However, those same regulations prohibited the companies from using that priority access to acquire more inputs than the ones needed to fulfill their US government orders.
This uncertainty created problems. Vaccine manufacturing is known for the difficulty in forecasting production yields. (Even the big firms had little experience with priority-rated contracting under DPA.) What if inputs were "left over" after a company had fulfilled its US contract? Could they be used to make doses for export to satisfy demand from foreign governments? It was hard for companies to know or to plan. The lack of clarity was further complicated by the US government continuing to trigger options for more doses for boosters needed later in the year. The main result was that Pfizer reportedly only began exporting doses in April, with Moderna beginning in May.
The European Union took a different approach in 2020, leading to the opposite situation. It offered few subsidies to get companies to invest in vaccine manufacturing capacity at risk. Without European government support, US-based production beat EU-based production to 100 million, 200 million, and 250 million available doses by early 2021 (see again figure 1). The lack of access to early doses likely cost European lives and caused severe economic losses. The European Commission was widely criticized for those delays and almost got into a trade war over shortages of doses with the UK in late January 2021. (Unlike the EU, the UK government provided subsidies to AstraZeneca in 2020 for at-risk investment in manufacturing that gave it earlier access to doses.) Yet, even from the beginning, firms in Europe could export vaccine doses.
Eventually, production of vaccine doses in Europe did expand (see again figure 1), and not surprisingly its exports exploded. In contrast, US manufacturing capacity mostly stalled out (see again figure 3). Lacking clarity from the US government on when they would be legally allowed to export in early 2021, Pfizer-BioNTech and Moderna instead looked to Europe when making their marginal investment decisions about where to expand capacity to help meet global vaccine demand.
Things could have gone differently in 2021
Given booming global demand for vaccines, what would it have taken for the US government to incentivize the companies to significantly expand their US-based production capacity, in addition to what their production facilities were doing in Europe?
Additional legal clarity on when, and the conditions under which, firms could export from their US plants would have helped. But the US government could also have been even more aggressive in recalibrating its Operation Warp Speed template.
For example, in early 2021, the US government could have used similar contracting logic to what it had used in 2020 under the DPA, guaranteeing a market if the firms were willing to invest in creating additional US production capacity. Demand was not a problem. There were lots of orders from foreign governments and COVAX. The US government could have worked with such countries to creatively package the orders into a larger contract, but one that was also contingent on the companies agreeing to add production capacity. For example, that collective offering could have funded the costs of Pfizer adding production lines to supply, say, 100 million doses per month by the middle of 2021 instead of just 50 million (see again figure 3). The US government could have organized a larger contract to incentivize Moderna to hire additional CDMOs to get its US production to 60 million doses per month, instead of 30 million.
Such an approach would have meant re-contracting with firms to expand their capacity, not just providing doses from existing facilities. Contracting on capacity creates something socially useful—a positive externality—that also benefits the rest of the world during a pandemic. (It shortens the wait time for someone else.) Contracting on doses, which simply relies on existing capacity, does not. In fact, if a new order simply moves another country later in line, that contract can generate a negative externality.
The policy failure to adopt this sort of contracting approach in 2021 was partly a collective action problem. Government orders from foreign countries simply got into a queue, while companies acting in the absence of government guidance fulfilled contracts to meet their own objectives, including where and by how much to expand capacity.
Getting the best of both worlds
Producing more vaccine doses early is one of the biggest public health challenges during a pandemic. Many lessons—both positive and negative—can be learned from the US experience during COVID-19. There were other impediments to scaling up vaccine production, including shortages of critical inputs.
Speed and scale were needed to save lives, and vaccine manufacturers in the United States led the world in supplying a lot of doses early. But that achievement came with a tradeoff of laggard expansion afterwards. Production and supply chain decisions made by companies like Pfizer, BioNTech, and Moderna can help policymakers identify the policy mix needed to combine the best of both worlds from the response to COVID-19—getting those advantageous, large, early US doses as well as the massive ongoing expansion of European supplies.
1. In April 2021, Operation Warp Speed was renamed the COVID-19 Countermeasures Acceleration Group (CAG).
2. See Chad P. Bown. 2022. COVID-19 vaccine supply chains and the Defense Production Act. PIIE Working Paper 22-9, June.
3. Since the focus here is on US vaccine manufacturing, comparisons are limited to other countries' production of the same pool of vaccines. The comparisons thus exclude China, which manufactured roughly 5 billion doses of vaccines from Sinovac and Sinopharm by the end of 2021.
4. See proposals from experts convened under the Accelerating Health Technologies (AHT) group, including Susan Athey, Michael Kremer, Christopher Snyder, and Alex Tabarrok. 2020. "In the Race for a Coronavirus Vaccine, We Must Go Big. Really, Really Big." New York Times, May 4; Susan Athey, Juan Camilo Castillo, Esha Chaudhuri, Michael Kremer, Alexandre Simoes Gomes, and Christopher M. Snyder. 2022. "Expanding Capacity for Vaccines Against Covid-19 and Future Pandemics: A Review of Economic Issues." Oxford Review of Economic Policy, forthcoming; Amrita Ahuja, Susan Athey, Arthur Baker, Eric Budish, Juan Camilo Castillo, Rachel Glennerster, Scott Duke Kominers, Michael Kremer, Jean Lee, Canice Prendergast, Christopher Snyder, Alex Tabarrok, Brandon Joel Tan, and Witold Więcek. 2021. "Preparing for a Pandemic: Accelerating Vaccine Availability." American Economic Association Papers and Proceedings, vol. 111: 331–335; Juan Camilo Castillo, Amrita Ahuja, Susan Athey, Arthur Baker, Eric Budish, Tasneem Chipty, Rachel Glennerster, Scott Duke Kominers, Michael Kremer, Greg Larson, Jean Lee, Canice Prendergast, Christopher Snyder, Alex Tabarrok, Brandon Joel Tan, and Witold Więcek. 2021. "Market Design to Accelerate COVID-19 Vaccine Supply." Science 371, no. 6534: 1107–1109. See also Christopher M. Snyder, Kendall Hoyt, Dimitrios Gouglas, Thomas Johnston, and James Robinson. 2020. "Designing Pull Funding for a COVID-19 Vaccine." Health Affairs 39, no. 9: 1633–1642.
5 . See tables A.1 and A.2 in Chad P. Bown and Thomas J. Bollyky. 2022. How COVID-19 vaccine supply chains emerged in the midst of a pandemic. The World Economy 45, no. 2: 468-522.
6 . The Department of Health and Human Services' Health Resources Priority and Allocations System regulation implementing the DPA states: "Rated orders may not be used to obtain: (i) Delivery on a date earlier than needed; (ii) A greater quantity of the item or services than needed, except to obtain a minimum procurable quantity…" (45 C.F.R. § 101.38 (a) (2)).