TRADE, INVESTMENT AND COMPETITIVENESS TRADE, INVESTMENT AND COMPETITIVENESS EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT Why the World Trade Organization is critical for vaccine supply chain resilience during a pandemic May 2022 Author: Chad P. Bown © 2022 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy, completeness, or currency of the data included in this work and does not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Cover design and layout: Diego Catto / www.diegocatto.com >>> Abstract Cross-border supply chains and international trade played a critical role in vaccinating much of the world to address the COVID-19 pandemic. Considering that experience, this note describes the changes needed to make the World Trade Organization (WTO) a more useful institution during such a public health emergency. It begins by describing the market failures confronting vaccines—especially on the supply side—to introduce the domestic subsidies and contracting arrangements needed to accelerate vaccine research and development, and to increase the scale and speed of vaccine production during a pandemic. As an application, it relies on illustrative examples of US subsidies that emerged during COVID-19. However, the challenge confronting policymakers is exacerbated in an environment characterized by cross-border supply chains, making input shortage problems impacting production even worse. Thus, the note highlights the need for new forms of international policy coordination, including initiatives on supply chain transparency, as well as agreements to increase subsidies across countries to jointly scale up vaccine output—and input—production capacity along the entire supply chain. It concludes that while the WTO was mostly absent this time around, it remains the best-positioned international organization to facilitate these novel forms of international economic policy cooperation. Chad P. Bown Peterson Institute for International Economics & CEPR Bown: Peterson Institute for International Economics, 1750 Massachusetts Avenue, NW, Washington, DC 20036. Tel: +1.202.454.1306, email: cbown@piie.com, web: www.chadpbown.com. Contents 1. Introduction 5 2. The Emergence of Covid-19 Vaccine Supply Chains 8 3. Vaccine Market Failures and Government Policy 12 4. US Subsidies for Vaccine Capacity in Response to Covid-19, International 15 Reaction, and Implications for the Wto 4.1. US subsidies to accelerate vaccine development and manufacturing capacity 15 under Operation Warp Speed 4.2. The international reaction to the US failing to export 17 4.3. What if the WTO banned all arrangements that did not immediately result in 17 exports? 4.4. Additional issues 18 4.5. Subsidies that accelerate and expand vaccine production capacity: 19 Implications for the WTO 5. Input Shortages for Vaccine Production During Covid-19 21 5.1. US policy to address input shortages: subsidies and rationing 21 5.2. Reallocating scare resources upon revelation of new information 23 5.3. Input shortages: implications for the WTO 23 6. Other Multilateral Policy Issues for Vaccines 25 7. Conclusions 27 References 28 1. Introduction Upon the emergence of a pandemic, new and life-saving products like vaccines are critical. So are the size and form of government subsidies needed to accelerate their research and development (R&D); the establishment of supply chains; and to increase their scale of production. In the case of COVID-19 vaccines, trade would turn out to play two essential roles. First, cross-border supply chains would emerge, and imported inputs were needed to manufacture COVID-19 vaccines. Second, most countries would have no local vaccine production, relying entirely on imports to gain access to these brand-new products. The existence of that trade, as well as a number of other pandemic factors, made enhanced policy cooperation more essential. And yet, much of the demand for more international coordination went unmet, though not for traditional reasons. Governments intervened too little into markets, not too much. In response to COVID-19, the US government was an outlier, allocating $18 billion in “subsidies” through Operation Warp Speed beginning in early 2020 to accelerate research, development through clinical trials, and sometimes even the creation of manufacturing capacity, for an initial portfolio of seven vaccine candidates. Despite these enormous subsidies, economists have almost universally criticized US subsidy efforts as being orders of magnitude too small. The human and economic costs of the ongoing pandemic ran into the trillions of dollars.1 In their modeling, for example, Ahuja et al. (2021) suggest that the United States should have spent more than three times that amount and diversified across 27 different vaccine candidates in 2020.2 (See also Athey et al. 2022.) The motto has become “spend billions to save trillions.” Yet, where was the rest of the world? The United States was, in fact, one of the few to subsidize the acceleration of this process much at all. The explanation was not that the United States had a monopoly on COVID-19 vaccine intellectual property or technology: the Pfizer-BioNTech vaccine was invented in Germany, the AstraZeneca vaccine in the UK, and the Johnson & Johnson vaccine was co-invented in the Netherlands. Indeed, plants to manufacture these and other COVID-19 vaccines emerged all over the world, with vaccine technology licensed to multiple firms in India and other developing countries even by the summer and fall of 2020. While the United States may have been slow to export finished vaccines, its subsidies to accelerate and pay for clinical trials generated global externalities for vaccines subsequently manufactured in facilities located around the world. 1 See, for example, Cutler and Summers (2020) and Agarwal and Gopinath (2021). 2 These and other economists had made such policy recommendations early and throughout the pandemic—see Athey et al. (2020), Snyder et al. (2020), Castillo et al. (2021), and Kominers (2021). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 5 The European Union (EU) and its member states offered very The WTO is also the best-positioned international organization few subsidies to accelerate the process, focusing instead on to provide a forum for policymakers to convene upon the allocating authority to the European Commission to negotiate emergence of a pandemic to address the commercial problem procurement contracts and advanced purchase agreements of input shortages that will inevitably arise at the beginning to order COVID-19 vaccine doses at low prices.3 Criticized stages of new vaccine production. Here, there are three for its slow, miniscule and poorly targeted policy response distinct needs. First, in terms of pandemic preparedness, the to COVID-19, the EU has since developed a major new WTO can help identify what critical inputs can be stockpiled in initiative—the European Health Emergency Preparedness advance—the WTO’s considerable expertise on transparency and Response Authority (HERA)—potentially modeled similar and diagnosing input-output relationships in supply chains to the Biomedical Advanced Research and Development can help organize such essential information. However, not Authority (BARDA) in the United States. India, whose firms everything can be stockpiled or worked out ex ante through made it the largest vaccine-producing nation prior to COVID- a pandemic preparedness playbook; a new vaccine is likely 19,4 did not offer subsidies to vaccine manufacturers to expand to require at least some new specialized inputs. As such, production capacity until April 2021, more than 14 months there will be an inevitable period of input shortages. Thus, into the pandemic. The Coalition for Epidemic Preparedness second, the WTO’s ability to quickly gain insight into cross- Innovations (CEPI), attempted a subsidy approach with some border supply chains can help diagnose short-run input qualitative similarities to the US government, but its efforts scarcity problems and assist policymakers as they facilitate were constrained.5 the rationing of those inputs to their best (inclusive of public health) use. Third, the WTO can use that information about World Trade Organization (WTO) subsidy rules may have input shortages to help domestic policymakers coordinate implicitly contributed to the problem. The WTO’s Agreement capacity-enhancing subsidies to input suppliers to ensure the on Subsidies and Countervailing Measures has a bias against full, cross-border supply chain is scaled up. government subsidies. Trade ministers and their staffs are trained in these rules; staff operating in vaccine-producing The fact that the WTO is a permanent body already providing countries were unlikely to contribute usefully to the domestic a forum for high-level government officials to continuously policymaking process that should have sought to structure interact over trade, supply chains, and periodically even subsidies to accelerate vaccine development and to increase scientific issues—and where governments sometimes both the speed and scale of production capacity. The WTO missed dispute and negotiate mutually agreeable solutions—makes it an opportunity to contribute meaningfully and showcase why capable of facilitating the additional international cooperation enhanced policy cooperation was also economically essential, and policy actions required to accelerate the creation of new given that the manufacturing supply chains to emerge were vaccines and the expansion of production capacity during a cross-border in nature. pandemic. The WTO also has a history of working with other major international organizations in good faith to bring in Vaccines alone showcase the outsized role for the WTO during required scientific and technical expertise when needed.6 a pandemic. Cross-border supply chains and foreign inputs are needed to manufacture vaccines. Trade is essential for Unfortunately, but predictably, the world resorted to “vaccine delivering finished doses globally. Having the most experience nationalism” of various forms during COVID-19. In part, this with subsidy rules, the WTO should be at the forefront arguing was because countries had not developed a global framework that laissez-faire and markets alone are not the appropriate that would allow them to share risk, pool resources, and rely policy response during a pandemic. There should be a basic on supply chain interdependence to manufacture and deliver pandemic playbook for subsidies, describing in advance the vaccines globally during a pandemic.7 In the absence of a funding, contracting frameworks, and “best practices” for guidebook, some governments acted unilaterally.8 However, domestic policymakers to use when confronted with the desire it is important in retrospect to better understand the various to accelerate the vaccine R&D and manufacturing process in forms of “vaccine nationalism” that emerged if there is hope such an emergency. to create new rules or norms for government behavior and 3 Though relatively small in scale, exceptions include Germany’s subsidies to BioNTech and CureVac, and UK at-risk subsidies to multiple candidates. See Bown and Bollyky (2022, Table 5 and Table 6). 4 See CEPI. 2020. Manufacturing Survey Results Analysis, June 29. 5 CEPI had fewer financial resources, more legal constraints, and lacked the political mandate. 6 Examples abound, but consider Codex Alimentarius, the World Organization for Animal Health, etc.—whether for issues arising under the Sanitary and Phytosanitary Measures Committee or in panels under WTO dispute settlement. 7 See Bollyky and Bown (2020a). 8 The notable and important exception was the emergence of COVID-19 Vaccines Global Access (COVAX), which was not allowed to succeed because it was not provided political support or with manufactured doses by the major economies where vaccines would be produced—that is, the United States, India, the EU, and China. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 6 international cooperation during a future health emergency. the supply side and the need for special forms of subsidies This time, while trade flows were mostly kept open, avoiding and contracting to accelerate and commit to production the most disastrous outcomes, the WTO played a very minor capacity enhancement. Section 4 introduces the specific role overall. And it played virtually no proactive role during and illustrative example of US subsidies to emerge during the critical period of 2020 when most of the essential policy COVID-19, highlighting parallels to the proposed structure decisions were being made on subsidies for the acceleration (if not scale) identified in Section 3, as well as implications of COVID-19 vaccines that would determine the speed and for international policy cooperation and trade rules. Section 5 scale of production, and influence longer-term production and tackles the problem of input shortages. Sections 4 and 5 both trade patterns. emphasize why the WTO is the best-positioned international organization to play a leading role in obtaining the international This note begins with Section 2 describing how COVID-19 cooperation needed to accelerate vaccine production in the vaccine supply chains emerged from scratch during this presence of cross-border supply chains and inevitable input pandemic. Ultimately, more than 11 billion doses were shortages. Section 6 describes other areas of concern for manufactured by the end of 2021, with 40 percent of those policy coordination to emerge for vaccine production and being exported. Section 3 then introduces the economics distribution. Section 7 concludes. of the market failures for vaccines, with a special focus on EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 7 2. The Emergence of Covid-19 Vaccine Supply Chains Getting a new vaccine from beginning to end requires investment in a number of sizeable sunk costs (figure 1). This includes the scientific research to invent the vaccine, the clinical trials to develop and check that it is effective and safe, the creation of a dedicated manufacturing facility with specialized equipment needed to produce the vaccine’s drug substance, and a separate manufacturing facility to formulate the drug substance into drug product for fill and finish, assembly-line style, into hundreds of thousands of tiny vials, for distribution. Pre-pandemic, the world manufactured roughly 1.5 billion doses of vaccines annually. With COVID-19, the global pharmaceutical industry was tasked with reallocating production facilities, establishing new supply chains, and creating new input streams to suddenly manufacture roughly 11 billion additional doses of new vaccines—mostly a two-dose regime aiming to inoculate 70 percent of the world’s 7.8 billion people. Even once COVID-19 vaccines had been invented and successfully passed clinical trials, such an effort would require a tremendous increase in dedicated production lines, as well as inputs into a sophisticated and highly regulated manufacturing process. This section summarizes the results of Bown and Bollyky (2022), which describes the details behind the manufacturing supply chains that emerged from scratch over 2020-21 for six different vaccine candidates: Pfizer-BioNTech, Moderna, AstraZeneca, Johnson & Johnson, Novavax, and CureVac.9 The focus here will mostly be on the first four, since Novavax was not authorized for use by anyone until late 2021 and CureVac was not authorized by anyone.10 Overall, dozens of different firms at nearly 100 different geographical facilities became part of even just the two most basic elements—that is, steps 3 and 4—of those vaccine manufacturing supply chains.11 Contract development and manufacturing organizations (CDMOs) played an important role in manufacturing almost every vaccine. 9 Bown and Bollyky (2022) also includes a database mapping vaccine sponsors to the manufacturing facilities as well as the announced timing of those matches. Unfortu- nately, that exercise did not tackle the question of how supply chains emerged for Chinese firms Sinovac and Sinopharm, except to note that most of the production facilities appear to have been located within China. 10 Nevertheless, their supply chain formation strategies and preparation were informative. In broad terms, Novavax followed the AstraZeneca model of using CDMOs to establish region-specific supply chains—that is, sticking to one plant covering step 3 and another plant covering step 4 in the same region. CureVac’s supply chain was entirely concentrated in Europe, and while it hired a number of different contractors to do its manufacturing, it was interesting that some of those arrangements were with large, global pharmaceutical companies—for example, GSK, Bayer, and Novartis—some of which may have found it worthwhile to team with CureVac to learn more about its mRNA technology, potentially to apply it to other pharmaceutical settings in the future. 11 This does not include the key input providers that fed into these manufacturing facilities, as will be described below. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 8 Here are some of the key characteristics of the supply chains All of the vaccine candidates set up parallel manufacturing that were announced as forming over this period. supply chains across different geographies. Each had at least a US-based and Europe-based supply chain for steps 3 and 4, First, none of the supply chains to emerge were in place prior for example. Pfizer-BioNTech would export to other countries to 2020. These all arose from scratch, and they almost all from those supply chains, as would Moderna.13 relied heavily on CDMOs. Even Pfizer, whose US-supply chain remained mostly in house through 2021, stated in May 2020 However, in addition to its US and European supply chains, that it would turn to CDMOs to outsource production of their AstraZeneca set up additional parallel supply chains in India, other (non-COVID-19 vaccine) pharmaceutical products.12 South America, the UK, South Asia, and elsewhere (Figure While some partnerships formed for COVID-19 vaccine 5). Indeed, most of the total production of the AstraZeneca production had had prior commercial relationships before vaccine would take place at the Serum Institute of India. COVID-19, many were new, including between firms that (Novavax, which started contributing to COVID-19 vaccine might otherwise be “rivals” for other pharmaceutical products. supplies late in 2022, adopted a similar supply chain model to AstraZeneca.) Production for every vaccine was fragmented across multiple facilitiesfor example, steps 3 and 4 were always done at Johnson & Johnson also subcontracted fill and finish facilities separate plants. Even for the complex Pfizer-BioNTech mRNA with a firm in South Africa (Aspen Pharmacare) that it was vaccine that is, in the United States, this was the closest anticipated would receive drug substance imported from thing to an integrated production process since steps 3 and 4 its plant in the Netherlands or from a contractor making its remained in house (that is, not outsourced)—production took vaccine in the United States; Johnson & Johnson also shared place across three different US facilities—that is, Pfizer plants its technology with a firm in India (Biological E.) in the summer in Missouri and Massachusetts combined to cover step 3, and of 2020 that, as of this writing, still had not come online. a Pfizer fill-and-finish facility in Michigan contributed step 4 (Figure 2). (The vaccine patent itself, of course, had been Each vaccine sponsor faced challenges even as it contracted offshored to those US Pfizer plants under an arrangement with with partners to add production facilities. Companies like BioNTech, a German biotech, where it had been invented.) Moderna and BioNTech complained about the shortage of And in the European supply chain for the Pfizer-BioNTech facilities and firms with the technological know-how with which vaccine, the process was even more fragmented. to partner for their brand-new mRNA vaccines, including for fill and finish. In some instances, the shortage of plants Each vaccine was manufactured by at least some CDMOs caused CDMOs to break pre-pandemic contracts with other from the beginning, with the exception of Pfizer-BioNTech. pharmaceutical companies—with which they had been (However, by mid-2021, even for the Pfizer-BioNTech supply scheduled to manufacture other products—to create the chain in Europe, production needs expanded beyond available emergency space needed for COVID-19 vaccine production, capacity at Pfizer or BioNTech facilities so they brought some especially for fill and finish. CDMOs into the manufacturing process, especially for step 4.) At one extreme was Moderna, a biotech with no commercial Next, there were also critical inputs, sometimes feeding scale manufacturing facilities prior to the pandemic, which in through mini supply chains, coming into steps 3 and 4 outsourced every part of the production process to different (manufacturing) as well as step 5 (distribution). CDMOs (Figure 3). But even Johnson & Johnson, a large, global, integrated pharmaceutical company, only relied on one For example, all vaccines required specialized inputs such of its own facilitiesthat is, some of its drug substance was to be as capital equipment (bioreactors) in addition to “single-use” manufactured at its own plant in the Netherlands (Figure 4). or “consumables” (for example, bioreactor bags, filters) that Fill and finish for Johnson & Johnson was done by contractors would need a continual stream from input providers feeding elsewhere, as was production of its drug substance to be into step 3. Over the course of scaling up production, there done in the US market, first by Emergent BioSolutions and were shortages of such inputs—by early 2021, virtually all ultimately also by Merck. of the vaccine sponsors were publicly complaining about insufficient quantities of available inputs, indicating this was 12 Carl O’Donnell and Michael Erman, “Pfizer to Outsource Some Drug Production, Focus on Coronavirus Vaccine,” Reuters, May 9, 2020. 13 Moderna did eventually sign up SK bioscience in South Korea to fill and finish some of its production as well. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 9 holding up their ability to meet production and delivery targets. At the other extreme was the United States. Even though Indeed, in such a highly regulated and complex production US government subsidies (described below) funded seven process, missing one input could have devastating impacts different vaccine candidates, including five that have now for vaccine output. (While a convenient excuse, some of it been authorized by some regulator worldwide, the US Food may be standard slippage for scaling up a new product. Some and Drug Administration (FDA) only authorized three for workarounds were found and some of the complaints were use in the United States by the end of 2021. Of those, total by companies whose vaccine was never authorized or was US production in 2021 ended up being mostly Pfizer and authorized much later, in which case them not getting access Moderna, as the third vaccine authorized for use—Johnson to inputs during a period of shortage and rationing may have & Johnson—had contamination problems at the primary been efficient.) drug substance manufacturing facility in the United States in March 2021 and had to be shut down for about four months. However, some of the vaccines also required specialized Thus, even though the US government helped set up and inputs that may not have previously manufactured at volumes fund supply chains and capacity that was then reserved for needed for commercial scale, let alone pandemic-level Johnson & Johnson, AstraZeneca, and Novavax, most of demand. Lipid nanoparticles (LNPs) were needed for the that would sit idle, not in commercial use, even though doses mRNA vaccines of Pfizer-BioNTech and Moderna, and each manufactured at those facilities could have been put to use brought in a suite of contractors to supply it (see Figures 2 in other countries where the vaccine had been authorized. and 3), before Pfizer eventually also brought some production While the United States was arguably the early success story in house. The Novavax vaccine would have needed access of 2021—in terms of its speed and initial scale of production, to a highly specialized adjuvant derived from the Chilean partially for the reasons described in Section 4—by the end soapbark tree. of 2021, US total production significantly lagged both the EU and India. All of the vaccines would also need relatively homogenous ancillary inputs like vials and glass stoppers (for packaging Most of India’s output was the Serum Institute of India’s in step 4), as well as syringes and needles (for administering production of the AstraZeneca vaccine. While enormous in in step 5). Because of cold-chain requirements for the mRNA scale, the problem was that very few of its doses were exported vaccines, refrigeration would also be needed for transport in 2021 due to an Indian government export clampdown in April from the plant through to the point at which final doses would that mandated its output be kept local due to a devastating be administered to people. outbreak taking place in the country. The concern was that COVAX was relying on doses from the Serum Institute to By the end of 2021, estimates are that 11.5 billion COVID-19 allocate to low-income countries, which dried up. vaccine doses had been supplied globally, over 90 percent of that by China, the EU, India, and the United States (Figure 6). By the end of 2021, the broad story was fourfold. First, it was an (The Chinese firms deserve a lot of credit, but for transparency amazing accomplishment to invent, get through clinical trials, reasons, much less is known about their supply chains, as and manufacture 11.5 billion doses of COVID-19 vaccines in well as the long-term effectiveness of their vaccines.) Nearly two short years. Few would have anticipated this outcome at 40 percent of that global production was exported. the outset of the pandemic in early 2020. The EU manufactured the largest portfolio of different vaccines Second, the doses did not arrive quickly enough. Accelerating for internal use, authorizing each of the big four—Pfizer- their arrival by 1, 2, or 3 months was estimated to save BioNTech, AstraZeneca, Moderna, and Johnson & Johnson— hundreds of thousands of lives and the equivalent of trillions by March 2021, with Novavax being authorized in December of dollars of economic activity (Athey et al. 2022). 2021. All of these vaccines had local supply chains in the EU (drug substance manufacturing for Moderna’s European Third, the distribution of COVID-19 vaccines was biased supply took place in Switzerland but fill and finish was done toward the countries and regions in which the doses were in Spain and France). This helps explain why the EU’s total manufactured. While more supply chains were established COVID-19 vaccine production was so large. by early 2020 than may have been capitalized on through production—often due to events outside of anyone’s control, EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 10 such as input shortages or the failure of a vaccine to be waste in the system of unused or expired doses put the overall authorized by regulators—the geographic concentration of demand for capacity at much higher than 11.5 billion doses. production had this implication. Thus, it is understandable that there have been calls for post-pandemic efforts to diversify These outcomes beg the question of why government policy in vaccine manufacturing capacity globally to better prepare for COVID-19 failed to address the market failures for COVID-19 future global health emergencies. vaccines. The next sections explore the policies that did emerge, the lack of international policy cooperation, and why Fourth, the scale of production was also too small. The lack of the WTO is needed to help overcome challenges involving sharing meant less than 10 percent of the population in poor international externalities brought on by a world characterized countries went inoculated by the end of 2021. Combined with by cross-border supply chains for vaccine manufacturing. the new demand for boosters (a third dose) as well as (normal) EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 11 3. Vaccine Market Failures and Government Policy How those 11.5 billion doses went from scientific ideas for a COVID-19 vaccine to the point at which they were injected into arms was not only left to markets. It was also the result of considerable financial support from the public sector. For decades, governments have struggled with the policy question of how to tackle vaccine “subsidies.” The environment is made difficult by a complex web of failures arising on both the demand and supply side of the market. On the demand side, the consumption of vaccines enjoys positive externalities. An individual that takes a vaccine provides additional, uncompensated benefits to society by breaking disease transmission. Without government policy intervention, purely market-based outcomes would result in too few individuals consuming vaccines relative to the social optimum. To help overcome this market failure, governments typically procure vaccines from companies and distribute them to consumers either for free or at highly subsidized prices. Indeed, this is what took place during COVID-19.14 Tackling supply-side market failures for vaccines often proved harder. In the 1990s and early 2000s, the inability to invent vaccines to tackle diseases arising primarily in poor countries, such as malaria, tuberculosis, and certain strains of HIV prevalent in Africa, became increasingly apparent. Decades of failure were linked back to how policymakers were struggling to create the right incentives by focusing on subsidizing inputs (R&D), with informational asymmetries ultimately resulting in governments picking the “wrong” entities to subsidize.15 The 2019 Nobel Prize winner Michael Kremer (2001a,b) and others have pioneered explorations into creatively solving the incentivization problem, such as by using advance market commitments (AMCs) in an effort to strike the right balance between incentivizing research and accelerating vaccine development and production. The basics of the AMC approach provide a useful template to understand the key issues that emerged during the COVID-19 pandemic. 14 There are important issues involving procurement that will mostly go unaddressed here, which focuses on government support to address supply-side market failures. 15 For a discussion, see Kremer and Glennerster (2004). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 12 On the supply side, one traditional market failure for vaccines any vaccine, the economic harm from the pandemic in this can be characterized by a hold-up problem. Suppose a firm example is $24 trillion, or the area given by the rectangle with is contemplating sinking hundreds of millions of dollars into base 24 months and height $1 trillion per month. The exercise R&D and the creation of a vaccine-specific manufacturing examines how different policy choices reduce different-sized supply chain. Once those costs are sunk, the purchaser (for areas of economic harm the pandemic imposes on society example, a government, since most vaccine demand has not captured by Figure 7. been left to markets) has an incentive to offer a price only just above the firm’s marginal cost of production. Using backward Suppose first that the firm waits to install capacity until it is induction, firms recognize the future difficulty they would face certain that it will receive regulatory approval. Then suppose in recouping the sunk costs of R&D and vaccine-specific administering the vaccine reduces harm linearly—for example, manufacturing, and they under-invest in the first place. In this assume for simplicity, doses were randomly allocated—and outcome, too few vaccines are invented and produced. the pandemic ends 21 months later, at month 24. In this scenario, the vaccine reduces economic costs by area “A.” To help solve this potential hold-up problem, beginning in the 1990s, experts suggested policymakers provide firms Next, suppose policymakers can better allocate vaccines by with advance market commitments and other contracting targeting priority groups. This could be through public health guarantees to incentivize them to make the investments and economic considerations—for example, providing early necessary to invent and deliver viable vaccines. The practical vaccine access to hospital personnel and other front-line idea was that companies could be incentivized if governments, workers, vulnerable populations with co-morbidities, or even international organizations, or even well-endowed foundations workers in important economic sectors (to reduce economic could provide a legal guarantee that a future market would losses). Better targeting alone avoids even more losses given exist for, say, 100 million doses of a new vaccine to combat an by “B.” (The convexity of the new border results from initial under-addressed disease.16 vaccinations having a greater marginal reduction of harm than later doses given to lower priority populations.) In the case of a global pandemic such as COVID-19, another layer of this supply-side problem arose—the need for speed Now consider what happens when governments also have in accelerating the process of getting vaccines manufactured. policy instruments to incentivize when and how the firm sets A group of experts proposed early and often that this problem, up its vaccine production facility. too, could be tackled by creative contracting and subsidy incentives with insights from the AMC approach, as well as First, suppose the government can convince the firm to establish forms of “push” and “pull” funding.17 its capacity earlier, so that it is ready to begin distributing doses the moment the vaccine is granted regulatory approval Consider the following stylized example, represented in Figure (and not three months later). In this scenario, the policy shifts 7, taken from Athey et al. (2022), but also substantiated through the risk of the vaccine failing in a Phase 3 trial from the firm to their formal modeling results. During COVID-19, estimates are the government. But doing so generates the societal benefit that the global economy was losing the equivalent of about (of an additional harm reduction) given by area “C.” (Despite $1 trillion per month in terms of economic output, morbidity, its unusual shape, C is equivalent to the $3 trillion gain from and mortality. Suppose that eventually a vaccine could be accelerating the end of the pandemic by three months, costing invented, make it through clinical trials, manufactured, and the world $1 trillion per month.) distributed to a sufficient number of people—say 70 percent of the global population, enough to achieve herd immunity—to Second, suppose governments also have a policy instrument end the pandemic. Suppose the vaccine passes clinical trials to be used to convince the firm to double the size of its at month zero. In their stylized example, it takes three months production capacity. This allows for the end of the pandemic for the firm to install vaccine manufacturing capacity to begin 10.5 months after doses have begun being administered producing doses for global distribution, and then another rather than 21 months later. This benefit is the reduction of an 21 months to vaccinate the world, ending the pandemic two additional area of economic harm given by area “D.” years after the vaccine receives approval. In the absence of 16 For an application of the approach to tackling the need for a vaccine for pneumococcus, see Kremer, Levin, and Snyder (2020). 17 See Athey et al. (2020), Snyder et al. (2020), Castillo et al. (2021), Ahuja et al. (2021), Budish and Snyder (2021), and Kominers (2021). The approach described here follows Athey et al. (2022). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 13 The policy problem here is now clear.18 There are large In the case of COVID-19, experts recommended policymakers societal gains to be made by convincing the private sector to subsidize by focusing on a mix of “push” (subsidizing inputs to both (i) accelerate manufacturing capacity investments at risk, expand capacity, regardless of whether the vaccine candidate and (ii) significantly expand that capacity beyond what may be proved successful) and “pull” (rewarding expedited delivery of in its private, commercial interests. Left to its own devices, a doses of approved vaccines) contracts with firms. Thus, not profit-maximizing firm would wait until after resolution of the only should governments provide subsidies across multiple uncertainty associated with the lengthy Phase 3 clinical trial. vaccine sponsors to diversify scientific risk, but they should do Furthermore, if prices for vaccines were relatively fixed—as so at risk so that firms could begin building their manufacturing they mostly were during COVID-19, in the range of $6–$40 infrastructure in parallel with the Phase 3 clinical trials. They per course, well below their estimated social value of $5,800 should also directly contract on production capacity and on a per course (Castillo et al. 2021)—the firm has little private specific delivery schedule—for example, a specific number of incentive to substantially increase production capacity to doses delivered by a specific date —and not simply contract fill orders more quickly, especially because there was little on doses that would allow the firm to simply put their order concern for market entry (or business stealing) by competitors into a queue. if it delayed, given the lack of other options for consumers (Budish and Snyder 2021). How could governments convince firms to start investing in manufacturing capacity in advance of—or in parallel with— regulatory approval? How could they convince firms to install additional capacity? 18 Ahuja et al. (2021) model these scenarios for the United States and globally to suggest the size and scope of supply-side policy interventions during COVID-19. Their results suggest even the United States—which subsidized the most and across the largest portfolio of vaccine candidates in 2020—should have spent more than three times what it subsidized in reality and should have diversified across 27 different vaccine candidates. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 14 4. US Subsidies for Vaccine Capacity in Response to Covid-19, International Reaction, and Implications for the WTO When COVID-19 hit, the US government implemented some elements of this proposed approach of subsidizing at risk and on some capacity to accelerate the development and manufacturing of vaccines, especially in 2020.19 Yet, because the United States did not export doses in the first few months after commercial production emerged, the world complained that these contracts were equivalent to the United States imposing export restrictions. This section re-examines these considerations in light of the types of subsidies and contracts the United States used, and then considers the implications for the WTO. 4.1 US subsidies to accelerate vaccine development and manufacturing capacity under Operation Warp Speed To start, two of the most important features of US vaccine subsidies were that they emerged early and there was some diversification. The subsidies began right away, some of them arrived as early as February and March 2020 (Table 1). On diversification, the US government started by providing “subsidies” or contracts for seven different vaccine candidates. For five of those vaccine candidates—that is, Moderna, Johnson & Johnson, AstraZeneca-Oxford, Novavax, and Sanofi-GSK—it developed explicit elements of the at-risk subsidies and contracting on capacity suggested in Section 3. For a sixth candidate from Pfizer, it wrote an early procurement-only contract, as discussed below. A seventh from Merck-IAVI was given $38 million in April 2020, but the candidate did not pass through initial clinical trials and was abandoned relatively early. For some of these vaccine candidates, the early subsidies helped facilitate completion of clinical trials, including the expensive, data-intensive, and time-consuming Phase 3 trials requiring 30,000 or more trial participants—sometimes with subsidies of $400–$500 million dollars per vaccine candidate. Furthermore, for biotechs like Moderna and Novavax, which did not have their own manufacturing facilities to make even enough doses for Phase 3 trials, subsidies also covered the costs of outsourcing production to CDMOs at that stage to acquire doses for trials.20 19 For a discussion, see Bown (2022). 20 See also AstraZeneca (Table 1). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 15 Then, once five of these promising candidates reached the and not many months later (after creation of the supply chain), stage of starting advanced clinical trials in the summer and fall saving lives and tens if not hundreds of billions of dollars of of 2020, the United States contracted with firms for over $1 economic output. At least intuitively, a number of contracts billion each, including some funding so that they would begin were designed to help capture some of area “C” in Figure 7 (if to install capacity to manufacture 100 million or more doses.21 not part of “D”—that is, depending on whether the contracts The guaranteed funding allowed companies to begin to sink encouraged them to invest in more capacity than they would investments and establish vaccine-specific manufacturing have otherwise, irrespective of timing). facilities and supply chains in advance of the FDA’s regulatory approval that, realistically (it turns out) were 4-6 months off Third, the US government contracting on capacity (and into the future at best.22 Some contracts also included pull-like “owning” the initial output) for the firm’s first 100 million doses incentives for speed. Moderna, for example, had a contractual would potentially generate positive externalities for other term promising a higher price per unit for deliveries if the FDA buying countries. That is, a buyer that contracts on capacity granted the company emergency use authorization (EUA) for creates something that can be switched over to other buyers its vaccine by January 31, 2021. after its order has been fulfilled, whereas a buyer that contracts on doses only lengthens the vaccine queue, imposing a Finally, and as expected, the uncertainty of the science meant negative externality on others (Athey et al. 2022). the FDA ultimately did not authorize three of those vaccine candidates. For those candidates, the US government Fourth, the major vaccine sponsors created multiple, parallel received little payoff for its investment. However, for three of manufacturing supply chains, in part because it would be the six candidates, the basic process worked. Overall, for the impossible for US plants to export at first because of those United States, Operation Warp Speed was a relative success, contracts (Bown and Bollyky 2022). Paradoxically, the US in that it resulted in multiple viable COVID-19 vaccines and contracts that incentivized the same vaccine sponsor firms millions of doses produced (at risk) early, to be distributed to install additional capacity in other countries may have immediately upon authorization, saving lives and helping to generated positive externalities for the rest of the world in the limit the economic losses. Nevertheless, the United States form of that forced diversification. There was also some gain clearly could also have done better (Athey et al. 2022), and to additional geographical diversification, should future shocks its relatively positive outcome also involved considerable luck. arise (export controls, manufacturing problems, and so on) However, for the purposes of cooperation of international that might be geographic- or plant-specific.23 One tradeoff, trade and commercial policy, there were a number of important however, was that forcing the creation of that extra capacity practical implications of these early, at-risk US subsidies: to be done in foreign countries may have been inefficient from a global perspective—that is, investing in additional capacity First, this funding model shifted substantial financial risk onto expansion at US plants could have led to more learning-by- the US government and away from private firms. doing or other local spillovers increasing production yields. That may have led to more doses produced more quickly, Second, since firm investment in manufacturing capacity globally, than creating multiple supply chains for the same and the creation of a brand-new supply chain could take vaccine. place simultaneously with Phase 3 trials, vaccines that were granted regulatory approval would have millions of doses Fifth, there were contractual limitations on the US government available for distribution almost immediately upon approval, preventing it from exporting the vaccine doses it was buying.24 21 Specific details of the contracts are unknown, as the publicly available versions are highly redacted. However, see GAO (2021), which reviewed the contracts and inter- viewed the firms for confirmation. Note also it was 200 million doses for the US government contract with AstraZeneca. 22 The exception in the United States was the initial contract with Pfizer, which negotiated a contract of $19.50 per dose for 100 million doses but which was not guaranteed and thus which Pfizer retained the risk if the vaccine failed in clinical trials. Furthermore, of the initial six US government contracts for $1 billion or more signed with vaccine sponsors in 2020, Pfizer’s was the only not to be given a priority-rated contract under the DPA, which would have given it priority access to US-based input providers (Bown 2022b). 23 For the importance of geographic diversification to make medical supply chains more resilient, see Grossman, Helpman, and Lhuillier (2021). See also Miroudot (2020a,b), WTO (2021), and Baldwin and Freeman (2021). 24 The US government’s July 2020 contract with Pfizer, for example, stated “The Government agrees that it will not resell any of the deliverables to any third party.” EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 16 4.2 The international reaction The biggest concern with such a rule goes back to the impact on firm behavior in the absence of an AMC guaranteeing the to the US failing to export government’s purchases. Firms bearing all of the risk would likely have meant many delaying sunk investment costs Much of the reaction by the trade community to the US of manufacturing until after they had received regulatory government approach was negative. For example, some have authorization for their vaccine; this would have delayed scaling referenced the US government “owning” those first 100 million up production, learning by doing, and the delivery of millions doses of production under such contracts as the same as if of doses by many months.27 Furthermore, without AMCs, firms it had imposed an export restriction on vaccines.25 While the may have had to worry about subsequently being held up— contracts did result in firms not being allowed to export doses that is, government procurement agencies only offering prices until those initial orders were fulfilled, it is worth exploring the high enough to cover the marginal cost of manufacturing and argument against characterizing those contracts as export not the costs of research, clinical trials, and setting up a supply restrictions. At least in this instance, that is because 100 chain for the new product in the first place. million doses would not have been available during that time horizon but for the US government investments in advance Some companies may have taken the risk. Pfizer took on of regulatory approval.26 That is, claims of export restrictions the risk of investment prior to regulatory approval, but its ignore the timing of the sunk costs, many of which were paid compensation came in the form of a higher price ($19.50 per for by subsidies (at risk) by the US government to accelerate dose) than other vaccines in the US market, some of which the timetable of production by 4-6 months. also received grants covering the costs of their clinical trials. Importantly, Pfizer negotiated that contract in July 2020, before Establishing the right comparison to make is nontrivial. Ideally completion of its Phase 3 trial (the FDA granted emergency use it would take into consideration the market failures and subsidy for its vaccine in December), thus it was not held up on price. policies. Put differently, the US policy decision to invest at risk Pfizer may have had access to internal financial resources gave it access to vaccines 4-6 months earlier than it might have and access to its own facilities to overcome financial problems otherwise. Thus, a better comparison would be to examine the if its vaccine had failed in Phase 3 trials. US export decision 4-6 months after the vaccines had been granted emergency use. However, even that comparison Yet, many of the other vaccine sponsors, even in the United would miss out on the fact that many foreign countries were States, may not have had the private resources to take on that only able to accelerate their vaccine production by benefiting risk. Moderna and Novavax had no production facilities of their from US subsidies to R&D, especially the considerable US own. Even Johnson & Johnson and AstraZeneca chose mostly subsidies that allowed for expedited Phase 3 trials that would to outsource COVID-19 vaccine production to CDMOs rather not require repeating elsewhere. than use their own US facilities. Convincing CDMOs to invest to convert their plants to produce a new COVID-19 vaccine would have required financing, which may not have been 4.3 What if the WTO banned all available without government support, given the uncertainty arrangements that did not involved before regulators had granted EUA for the vaccines. immediately result in exports? Furthermore, India’s experience is also suggestive in this light. The failure of the Indian government to offer any subsidies on capacity—it did not subsidize Indian vaccine manufacturers Suppose the WTO had a rule that prohibited export restrictions until April 2021—meant Indian firms delayed expanding and that was interpreted as de facto also prohibiting a country production. The Serum Institute of India was only able to pre- from writing an AMC, because the AMC (by definition) resulted invest in some capacity in 2020 by relying on its own resources28 in zero exports, at least for an initial period of time while the and through funding from CEPI under an agreement to provide first orders were being filled to satisfy the government’s at-risk doses to COVAX; it subsequently reneged on that agreement investment on capacity. when a wave of cases hit India and the Indian government 25 See, for example, Bown and Keynes (2021). 26 The exception was the initial contract in July 2020 signed by Pfizer, which was a procurement contract, and for which the US government took on less financial risk if the Pfizer vaccine was not granted regulatory approval (GAO 2021). Nevertheless, Pfizer’s contract with the US still took on an aggressive timetable for regulatory approval and for the initial manufacturing and delivery of 100 million doses, and for which the US government held options. Furthermore, Pfizer ran into trouble acquiring the inputs needed to meet production goals and requested a second US government contract in December 2022, which was a priority-rated contract under the DPA. 27 Firms may have been wary of investing at risk, given the experience of prior pandemics. Evenett et al. (2021) suggest that as the H1N1 pandemic of 2009–2010 waned, some governments pulled funding and certain companies could not recoup the costs of their investments. 28 Jeffrey Gettleman, Indian Billionaires Bet Big on Head Start in Coronavirus Vaccine Race. New York Times, August 1, 2020. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 17 imposed implicit export restrictions in the spring of 2021. (In a Nevertheless, new concerns could arise if the US government May 2021 interview with the Financial Times, the firm’s CEO were to have planned to share those initial doses. The US blamed the Indian government for not pre-ordering doses, government may not have subsidized as much in 2020 if it was stating he had decided against investing in capacity expansion going to subsequently share the doses because it would have earlier because “there were no orders, we did not think we captured fewer of the positive externality benefits locally that needed to make more than [1 billion] doses a year.”29) And arose through a more quickly inoculated American population. other Indian firms, including Biological E., which had licensed Smaller US subsidies could have been in the form of smaller the Johnson & Johnson vaccine in August 2020, were not able contracts to each sponsor (for example, enough initial capacity to expand capacity quickly at all. for only 50 million doses), contracts to fewer vaccine sponsors (for example, three instead of six), or contracts to sponsors Finally, lack of clarity in Europe about who had contracted granted closer to the FDA’s emergency use authorization on capacity (for example, the UK) and who had negotiated to reduce financial risk (for example, shortening the at-risk procurement contracts on doses (for example, the EU) led investment window to, say, two months lead time). Smaller to political and legal disputes that could have turned out initial capacity expansion would have then led to longer queues disastrous. A dispute between the EU and the UK over doses for other buyers, independent of the US sharing some initial of the AstraZeneca vaccine even resulted in the EU temporarily doses. Fewer contracts would have increased the chance invoking Article 16 of the Northern Ireland Protocol for a few that none of the vaccines received regulatory approval—for hours in late January 2021.30 The EU shutting off exports of example, what if the three it had chosen were Sanofi-GSK, vaccines to the UK could have resulted in the UK retaliating by Novavax, and AstraZeneca? Either way, in expectation, firms shutting off exports of vaccine inputs needed to manufacture would have been even slower at sinking investments into vaccines in the EU—such as lipid nanoparticles essential for capacity expansion, and fewer total doses in aggregate would the Pfizer-BioNTech vaccine—hurting both sides (Bown and have been delivered. Bollyky 2022, Figure 3). From this perspective, there were two bigger problems. Additional clarity is needed on these types of subsidies and contracts, in terms of where orders will be placed in a queue. As previously noted, one was that other countries with However, a rule designed to stop export restrictions written manufacturing facilities mostly failed to provide similar in a way that would prevent policymakers from contracting subsidies at risk to convince firms to build capacity earlier and on capacity that guarantees demand and shifts risk onto the at greater scale. government (and away from firms) would create additional and potentially bigger problems. It would eliminate an important However, countries without their own manufacturing facilities mechanism needed to incentivize firms to scale up production failed to coordinate with the US government to write at risk, larger and more quickly—satisfying the public health interest— capacity-enhancing contracts to further expand production than on their own profit-maximizing timetables. taking place at plants within the United States. Suppose all vaccine manufacturing capacity is located in only 4.4 Additional issues one country, for example, the United States. Optimal global policy would have been for other countries to collectively write From that perspective, any “export restriction” in the US case contracts—say, through a multilateral entity like COVAX—with was less the result of the US government subsidizing the vaccine manufacturers that would guarantee purchases of installation of capacity and committing in advance to purchase doses for the commitment that companies invest in additional all 100 million doses of the initial COVID-19 vaccine production production capacity at risk, so that they could deliver more than its subsequent failure to allocate some of those doses doses more quickly. The failure to act collectively meant internationally after having taken delivery. However, the US companies simply allocated buyers to a longer queue and government would have needed to write different contracts to fulfilled orders (and expanded capacity) on their own timelines allow it to subsequently redistribute doses—for example, as and according to their own profit-maximizing objectives, and through COVAX—to prioritize global public health needs, and not the global wellbeing. not only the needs of American citizens.31 29 Stephanie Findlay, India’s vaccine shortage will last months, biggest manufacturer warns. Financial Times, May 2, 2021. 30 George Parker, Jasmine Cameron-Chile, and Michael Peel. 2021. EU pledges vaccine controls will not hit UK supplies. Financial Times, January 30. 31 Eventually, in July 2021, the US government did write such a contract with Pfizer ($3.5 billion for 500 million doses) to procure vaccines that would then be allocated to COVAX https://www.hhs.gov/sites/default/files/vaccine-donation-contract-with-pfizer.pdf. However, its initial contract prohibited the US government from selling doses to third parties. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 18 Finally, consider potential contractual limitations on the United 4.5 Subsidies that accelerate States exporting doses that it had procured to third countries, and why this too must be considered in advance. Suppose and expand vaccine production there were no global contracting arrangement, and so the capacity: Implications for United States was potentially interested in allocating doses globally from its orders from US facilities, as ultimately played the WTO out during COVID-19. Why is this an issue for the WTO as opposed to some other One concern involves whether the foreign country recipient international organization?34 One fundamental reason is has not waived potential liability for a given vaccine. This because the WTO is not simply a neutral, innocent bystander raises the question of who would bear the legal liability—the when it comes to subsidies that will impact industrial production. US government or the vaccine company—if there were to be The WTO’s Agreement on Subsidies and Countervailing a problem. This issue reportedly held up potential deliveries, Measures (SCM) not only exists, but it has a general anti- including to India of the Pfizer vaccine in the summer of 2021.32 subsidy bias. While there may be nothing explicit in the SCM Agreement prohibiting subsidies in response to a global health A second potential issue is whether allowing for such emergency, trade ministers and their staffs are trained in these unexpected exports might inadvertently make matters worse rules; their intuition can influence domestic policy. Yet, staff for poorer countries over the long run. For example, Kremer operating in vaccine-producing countries were unlikely to and Glennerster (2004, p. 35) explain what happened in the contribute usefully to the domestic policymaking process— early 1990s, when American politicians complained about one that would have benefited from structuring subsidies to US vaccine companies charging lower prices for children’s accelerate vaccine development and to increase the scale of vaccines allocated to poor countries than they were charging production capacity and it being made at risk, along the lines at home. The response of the companies was not to continue outlined in Section 3. to supply vaccines to poor countries and lower the vaccine price in the US market to match the foreign price; doing so The WTO needs to play a proactive role by offering subsidies would have eroded the profits needed to recoup their R&D guidance and not simply meet such a situation with silence. costs. Instead, the companies simply stopped submitting bids to UNICEF to provide any vaccine in the poorer foreign The WTO could assist by facilitating agreements—on the market. The implication is that tiered-pricing (international accelerated installation of capacity at risk—between vaccine price discrimination) can be welfare-improving. Thus, any sponsors and a multilateral entity that would also have pre- unexpected reallocation of supply—via, say, the United States committed to handle global distribution.35 Such an approach re-selling or donating doses to foreign governments—must would internalize many of the externalities on both the take into consideration the broader, long-run implications on demand and supply side and head off many of the subsequent incentives for supply availability elsewhere. problems that arose in the context of COVID-19. Ultimately, in July 2021, the United States did negotiate a However, if contracting takes too long, or a multilateral entity contract with Pfizer to buy doses that would then be exported lacks sufficient enforcement powers with respect to firms in to COVAX for distribution worldwide. The US government paid a domestic contracting environment (as arose in the United the equivalent of $7.50 per dose ($3.5 billion for 500 million States under Defense Production Act [DPA]), or multilateral doses), much less than the $19.50 it paid for doses initially to cooperation is politically impossible, it may be infeasible to inoculate Americans signed at risk in July 2020.33 However, if coordinate all of these sufficiently quickly. events had evolved differently (and as they had in the 1990s), political pressure may have resulted in international price In the absence of a first-best global subsidy and allocation discrimination being made more difficult, reducing COVID-19 approach, the world must then account for governments vaccine availability to poor countries. taking on some of the risk by pursuing their own capacity- 32 See Neha Arora and Rupam Jain. “India close to giving indemnity to foreign vaccine makers like Pfizer—sources,” Reuters, June 10, 2021. 33 https://www.hhs.gov/sites/default/files/vaccine-donation-contract-with-pfizer.pdf 34 Indeed, CEPI attempted to handle many of the issues raised here during the pandemic. A legitimate question is why an international organization like the WTO is needed. 35 Again, CEPI (and COVAX) developed a framework for much of this, but was not given the financial resources, as well as the political and legal authority to pull it off. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 19 installation subsidies with firms to commit them to install well-founded for public health reasons, given the wave of additional production capacity at risk to facilitate the quicker cases facing India at that moment in early 2021. What was and larger scaling up of production than they would arise in a problematic was that someone else—in this case CEPI—had purely market-driven outcome. subsidized the production of those COVID-19 vaccine doses at risk, and they were destined for COVAX to be distributed to The subsidies needed in response to COVID-19 should have other low-income countries. been earlier, focused on capacity expansion, and orders of magnitude larger than those that emerged. Furthermore, it Rules and enforcement procedures are needed to prevent is paramount for small countries and those without their own what India did from happening again. India did the opposite of manufacturing facilities to subsidize to organize collectively— the United States—it failed to plan ahead, and it expropriated so that firms expand productive capacity overall rather than the at-risk capacity investment of someone else. Without allocate their individual orders into a queue that they fill on such rules and enforcement, entities will be discouraged from their own timeline with existing capacity. coordinating the cross-country subsidies that are needed to increase capacity at risk in the presence of cross-border The WTO should help governments facilitate those subsidies supply chains, especially as described next. and contracts. The EU set up an export monitoring system in early 2021 to Certainly, there is also a role for the WTO to discourage the use track the destination of vaccine shipments manufactured in of implicit or explicit export restrictions. However, identifying the EU to countries outside of the bloc. In March 2021, Italy the difference between a policy that imposes international actually blocked exports of 250,000 doses of the AstraZeneca externalities and one that does not require more information vaccine from a Catalent plant in Italy destined for Australia. than simply examining data on whether or not exports exist. While this singular action was problematic, it was small in Again, arguably the suite of US subsidy policies—taken at scale. The broader action to set up the export monitoring risk, and to expand and accelerate production capacity—were system was also motivated by the lack of transparency by not what was worrisome. (Paradoxically, the US government suppliers and about the queuing of contracts. purchasing all of the initial doses may have actually generated some international positive externalities by triggering additional Additional transparency is recommended, and that should help capacity expansion elsewhere.) The worrisome US policy was reduce uncertainty and limit demands for export restrictions. not its vaccine contracts, but its failure to share the output The use of export restrictions ultimately creates the incentive arising under those contracts. for firms to invest elsewhere. However, and while export restrictions are not condoned, the larger problem is the India’s suite of policies was quite different and was worrisome. lack of coordinated, at-risk public investment on capacity It suddenly imposed export restrictions on the vaccine expansion which drove the shortages and highlighted the lack output of the Serum Institute. In isolation, the decision to of cooperation. prioritize doses toward the Indian population may have been EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 20 5. Input Shortages for Vaccine Production During Covid-19 Virtually all of the vaccine sponsors complained about input shortages at the early stages of COVID-19 vaccine production.36 This included big items, such as a lack of manufacturing facilities and contracting companies with which to partner, as well as shortages of skilled workers to whom the technology could be transferred, and a lack of specialized variable inputs, such as lipid nanoparticles, disposable bioreactor bags, filters, and more (see again Figure 1).37 There are numerous contributing explanations for the input shortages. The first and most important was the massive and sudden increase in demand for inputs by numerous vaccine sponsors all seeking to create new production facilities quickly and simultaneously for their brand-new product. A second was a potential concentration of demand into only a select few input providers, given that some vaccine sponsors were seeking to replicate the exact production process across potentially 10 or more different manufacturing plants globally to obtain a consistent drug product (Pall 2021)—see AstraZeneca’s supply chains in Figure 5. 5.1 US policy to address input shortages: subsidies and rationing Suppose that price signals from vaccine manufacturers to input providers—to convince them of the demand surge to scale up production capacity—were either slow, incomplete, or legally incapable of being sent.38 Or suppose that input-providing companies did not have incentives to invest in additional capacity because it would become unproductive once the pandemic was over and sit idle, preventing the firms from recouping their investment costs. The failure to invest in new capacity meant input providers allocated orders for their inputs to companies in a queue that they would service on their own timetables. The failure of input providers to add capacity would have thus served as a bottleneck that slowed the expansion of downstream vaccine manufacturers. 36 See the vaccine-specific examples and discussion in Bown and Bollyky (2022). 37 Bown and Bollyky (2022) provide an extensive discussion of the various shortages reported by these firms. 38 On the latter, Bown (2022b) describes one of the side effects of the priority-rated contracts the US government wrote with vaccine sponsors under the DPA was that they could demand inputs to fulfill those orders from suppliers at fixed prices. The inability to send higher price signals may have worked as a disincentive for input suppliers to add capacity. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 21 In this environment, there is an incentive for governments agencies tasked with implementing Operation Warp Speed, to subsidize vaccine input providers, in addition to vaccine which got access to information on supply chain relationships manufacturers, through the same sort of push contracts on and input needs through the priority-rated contracts signed capacity installation (which generates positive externalities) under the DPA. During this period, for example, the CEO and not simply the delivery of an input (which generates of MilliporeSigma—a major vaccine input provider whose negative externalities by allocating other buyers to later in the supplies were being rationed—said his company was “in queue). This would certainly apply for capacity into equipment nearly daily communication with ‘colonels and majors,’ the that might be pandemic-specific, in terms of scale. pharmaceutical companies and their contract manufacturers,” and that they were collectively forced “to start making trade- In 2020, the US government used Operation Warp Speed to offs when you’ve got limited supply and limited capacity to also subsidize the additional capacity expansion for some focus on the need of the moment.’’40 Informed rationing would input providers located in the United States. It agreed to push- require knowledge about the regulatory status of a particular style contracts to selected companies manufacturing relatively vaccine candidate—that is, the emergence of new information homogeneous inputs, such as glass tubing and vials for fill that a vaccine candidate was unlikely to be authorized by and finish facilities, as well syringes and needles that would regulators ideally would not have been prioritized for scarce be needed to administer doses of each vaccine (see Table 2). inputs. (However, given that the FDA was independent from Operation Warp Speed officials, it is unclear how efficient Subsidizing the expansion of production capacity would even this rationing process could be.) Also note that in the US prove harder for specialized inputs needed by each of the case, rationing would also be needed because price signals six vaccine candidates contracted under Operation Warp could not be relied upon to allocate inputs since input prices Speed. In October 2020, the US government also subsidized were fixed in the short run for priority-rated contracts signed Cytiva—a recent spinoff from GE Healthcare—for $32 million under the DPA (Bown 2022). dollars to expand capacity “for vaccine-related consumable products, such as liquid and dry powder cell culture media, Not only were the COVID-19 vaccine manufacturing supply cell culture buffers and mixer bags, as well as hardware chains to emerge in 2020-21 cross-border, but some of the including XDR bioreactors.”39 Cytiva was described by critical inputs needed by vaccine manufacturers globally may US policymakers as “the primary supplier to many of the have only been available from US suppliers. Rather than focus companies currently working with the U.S. government to on the input shortage problem, however, companies operating develop COVID-19 vaccines.” plants in India and Europe alleged the United States’ use of the DPA, which was rationing scarce inputs to manufacturers However, at the time the contract was signed, the US located within the United States, was creating an artificial government was unlikely to be aware of the full range of input shortage problem by banning exports from leaving the United shortages that would arise, even for US manufacturing plants, States. While an export embargo was subsequently refuted given that it would take time for the firms to fully develop by the data, this episode created unnecessary diplomatic their supply chains. Even though the US government got and political problems, and even resulted in French President access to knowledge about most of those supply chains—and Emanuel Macron accusing the Biden administration of input-output supplier relationships through the priority-rated imposing an export ban on vaccine inputs.41 Eventually US contracts it wrote and had to administer under the DPA—it policymakers began liaising, at least informally, with their was unlikely to have learned details of Pfizer’s input needs, counterparts in the European Commission as well as with the for example, until Pfizer signed its first priority-rated contract Indian government, to ration scarce inputs to some of these under the DPA in December 2020. foreign manufacturing facilities as well. In the very short run in 2020, US policymakers had to ration These responses to the input shortages revealed at least two scarce inputs deriving from a number of suppliers, spreading fundamental problems. First, the rationing that was taking them across a variety of manufacturing facilities (see Figure 8). place in the United States was likely to have been inefficient The rationing itself was apparently assisted by logistics experts from a global perspective. For example, the data subsequently from the Department of Defense, one of the government revealed the Serum Institute of India was sourcing inputs 39 BARDA. 2020. Trump Administration expands manufacturing capacity with Cytiva for components of COVID-19 vaccines. News Release, October 13. 40 Riley Griffin, “A Cold War-Era Law and Vaccines,” Bloomberg, January 2, 2021. 41 See Bown and Rogers (2021), Bollyky and Bown (2021b), and Sam Fleming, Jim Brunsden, Mehreen Khan, Michael Peel, and Guy Chazan. EU leaders confront US over vaccine patent waiver demands. Financial Times, May 8. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 22 during this period from a roster of vaccine supplier companies CEPI ultimately created “COVAX Marketplace” to help make in the United States that included MilliporeSigma, Thermo such matches for equipment and other variable inputs in July Fisher, ABEC, and Sartorius, in addition to Cytiva/Pall (Bown 2021, provided that the firm seeking to acquire the inputs was and Rogers 2021). US policymakers were unlikely to have providing COVID-19 vaccines to COVAX for distribution.44 insight into the input needs of foreign manufacturers, some of which may have been producing higher-priority vaccines than some of the ones in the United States. 5.3 Input shortages: Second, with this missing information, any US subsidy policy implications for the WTO designed to tackle the input shortage problem was likely both too small and incorrectly targeted from a global perspective.42 Given the nature of cross-border supply chains, there are Nevertheless, while input shortages were ubiquitous, it is four complementary roles for the WTO to play in helping to also worth reiterating that there is little evidence that any address the inevitable problem of input shortages that arise entities aside from the United States subsidized input capacity for vaccine production during a pandemic. expansion. It is possible, though highly unlikely, that this is because input providers were located exclusively in the First, the WTO could be the international organization tasked United States. with helping to diagnose where input shortages will arise. This would require liaising with vaccine manufacturers to survey Thus, there is a need to incentivize and coordinate subsidies them and then their input providers in their supply chain across countries along the full vaccine supply chain. In this (see again Figure 8). The WTO has the capacity to provide case, that would have meant other governments (for example, a secure platform for sharing and protecting confidential in Europe and India) subsidizing US input supply companies business information. It has done so in the past in the to expand capacity to get sufficient inputs to their firms quickly. context of WTO litigation—see, for example, the confidential information in the lengthy and politically contentious Boeing- Airbus disputes between the United States and the EU that 5.2 Reallocating scare resources the WTO Secretariat was trusted to keep secure. upon revelation of new Second, the WTO could be the entity that helps coordinate information rationing of scarce inputs to their most beneficial use during the very short run when dire shortages are inevitable. During the period of input shortages, there was a role for a secondary market to help redistribute newly freed up inputs Third, the WTO could provide the forum for policymakers in in high demand elsewhere. Newly available inputs would be key countries agree to coordinate implementing subsidies— expected to emerge, for example, when a vaccine candidate including contracting on the capacity expansion of input realized poor Phase 3 results. Many such candidates had providers—across the full cross-border supply chain to take been preparing facilities at risk, in advance of regulatory advantage of positive externalities and ensure input shortages authorization—that is, establishing supply chains and do not hold back vaccine production. accumulating inputs that would be freed up and could potentially be put to use elsewhere. Fourth, the WTO could work to reallocate resources when some are unexpectedly made available because of newly For example, in June 2021, CureVac revealed the disappointing realized information from Phase 3 trials. CEPI had been doing results of its Phase 3 trial for its mRNA-based vaccine some of this during COVID-19, but the WTO would be better candidate, freeing up resources to equip a supply chain positioned to make it part of their contribution. of multiple CDMOs it had lined up to manufacture 1 billion doses by the end of 2022. Yet, no one took advantage of the opportunity to quickly repurpose those facilities to engage in the manufacturing of other vaccines.43 42 In a (non-pandemic) model of trade in specialized inputs and offshoring, Antràs and Staiger (2012) find the lock-in effect results in equilibrium imported inputs being ineffi- ciently low, with subsidies to inputs being the optimal policy. 43 See Bown (2021). 44 CEPI. CEPI launches COVAX Marketplace to match buyers and sellers of critical manufacturing supplies and speed up global access to COVID-19 vaccines through COVAX, July 15, 2021. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 23 During COVID-19, each of these functions was done piecemeal those scarce inputs). In normal times and for normal goods, and incompletely. The WTO could be the entity needed to this is largely achieved through the price mechanism, markets, ensure the full suite of policy cooperation happens. For and trade. In a pandemic, some of those price incentives were example, incentive compatibility may require compensating severed and markets disappeared, especially in the very short a country that gives up scarce inputs (through rationing run, requiring other institutions instead to help intermediate to facilities located abroad) with guarantees of some of the and facilitate those benefits arising through trade. resulting benefit (a share of the doses manufactured from EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 24 6. Other Multilateral Policy Issues for Vaccines Insufficient attention was paid to two substantial problems holding back earlier and larger vaccine production: subsidizing vaccine capacity directly and at risk as well as input shortages. Holding back vaccine production held back vaccine trade and thus access for countries that would only acquire doses through imports. Nevertheless, other issues emerged impacting international commercial cooperation and trade. One was regulatory cooperation. Regulatory oversight is critical for both the clinical trials— ensuring that vaccines given to otherwise healthy people are effective and safe—as well as the production process. A number of issues arose in light of the COVID-19 experience. On clinical trials, there was considerable coordination globally between the relevant regulators to help limit the need for redundant clinical trials, especially the lengthy and costly Phase 3 trials. Nevertheless, a number of vaccines were ultimately accepted in some markets but not others; this has implications for output coming from unused supply chains. As examples, take the AstraZeneca and Johnson & Johnson vaccines. The fact that the FDA did not approve the AstraZeneca vaccine for the US market ultimately wasted valuable resources given that a US supply chain had been established at risk to manufacture 200 million doses of the vaccine. In theory, output from that US supply chain could have been used to inoculate people in other countries that had authorized that vaccine for use, even if the United States did not choose it for the American population.45 India only approved the Johnson & Johnson vaccine in August 2021—it had been approved in the US, the EU, and by the World Health Organization in February and March, respectively—potentially holding back domestic production by Biological E., that had licensed the technology in August 2020. Second, there are numerous examples confirming the difficulty in manufacturing COVID-19 vaccines and thus the need for regulatory oversight. One was the contamination arising at the Emergent BioSolutions plant impacting US supply chains and severely limiting the output of the Johnson & Johnson (as well as the AstraZeneca) vaccine. Another was that, in January 2021, a fire at the Serum Institute of India—the world’s largest vaccine manufacturer—impacted production to such an extent that the company reportedly used it to declare force majeure and 45 This example was admittedly complicated by one of the US facilities established to manufacture the vaccine—an Emergent BioSolutions plant in Maryland —failed to follow good manufacturing practices and contaminated doses, forcing the FDA to shut it down for four months. However, if the United States had not had other manufacturing capacity—via the Pfizer and Moderna supply chains—for authorized vaccine supplies available, an open question is whether policymakers may have faced pressure to act more quickly to resolve problems at that plant. Furthermore, the fill and finish capacity reserved for the AstraZeneca vaccine in the US market was also left unused. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 25 get out of contracts to supply doses to Brazil, Morocco, and that firms did not face an incentive to expand production Saudi Arabia.46 Administering the Pfizer-BioNTech vaccine in capacity more quickly to prevent “business stealing” from Hong Kong was also temporarily halted when defective vials other suppliers. and packaging were found in vaccines bottled in Germany but transported, stored, and distributed by BioNTech’s Chinese However, the additional entrance by more firms could have partner, Fosun Pharma.47 Finally, Moderna had to recall resulted in additional challenges. It would have meant more COVID-19 doses after stainless steel contaminants were found demand for inputs, potentially further exacerbating the problem in vials destined for Japan after being filled and finished at its of shortages. Finally, there was unused capacity that could CDMO partner Rovi’s plant in Spain.48 The implication is that have been reallocated but was not (for example, CureVac, regulatory oversight is needed to ensure the manufacturing Novavax, and so on). process delivers a safe and effective vaccine, and this may have been a (necessary) barrier to entry for firms even had Thus, even eliminating constraints on intellectual property the issue of patent protection allowed for more firm entry into protection, such as might have arisen had patents been vaccine production. waived, as was proposed by India and South Africa, the result could have been the creation of new problems. In addition to geographic concentration, there was also the issue of market concentration. Despite 11.5 billion Finally, trade facilitation was another important area during the manufactured doses, these were dominated by relatively pandemic. The need for cold-chain transportation and storage few vaccines on the market even by the end of 2021. Early for the mRNA vaccines of Pfizer-BioNTech and Moderna, evidence suggests the price implications of this concentration in particular, meant trade logistics could have operated as were not enormous—economists have estimated the social a bottleneck preventing some individuals from being able value of a course of vaccines to be $5,800, and the social to access vaccines, this was especially a problem in lower- and political pressure managed to keep prices relatively low, income countries. at $3–$40 per course. However, the lack of competition meant 46 Indrani Bagchi, “SII Fails to Deliver, New Delhi’s Vaccine Diplomacy Hits Hurdle,” Times of India, March 21, 2021. 47 Sui-Lee Wee, Alexandra Stevenson and Tiffany May, “Hong Kong Halts Use of Pfizer-BioNTech Vaccine Over Packaging Defects,” New York Times, March 24, 2021. 48 Rocky Swift and Carl O’Donnell, “Moderna to recall COVID-19 doses in Japan after stainless steel contaminants found,” Reuters, September 1, 2021. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 26 7. Conclusions By the end of 2021, more than 11 billion doses of COVID-19 vaccines had been manufactured globally. Yet, more international policy cooperation on vaccines and supply chains was needed, and one lesson is the numerous areas for the WTO to play a more active role. It was surely unprepared heading into the pandemic. But some of the WTO’s disappointing engagement during the first year of COVID-19 especially may have been due to a leadership vacuum. In May 2020, the WTO Director General (DG) suddenly announced plans to depart the organization that summer, more than a year earlier than the August 2021 expiration of his term.49 Installing a new DG amid an ongoing pandemic was a challenge in its own right and was further held up for months by an intransigent US administration. Finally, as a member-driven organization, the lack of any positive engagement by one of its major members—the United States—was also likely partially responsible. To their credit, when the new DG Ngozi Okonjo-Iweala and her new team arrived in February 2021, they prioritized addressing the pandemic and access to COVID-19 vaccines and served to facilitate useful dialogue between the private sector and policymakers. Unfortunately, the world was more than a year into the pandemic by then, already well onto the vaccine delivery path set by policymaker and commercial decisions made much earlier in 2020. By then, it was too late. Thus, the lack of much WTO engagement on a positive agenda in 2020 meant the only policy issue on the docket was a proposed waiver for intellectual property rights protection, a politically divisive item between members, and yet one that was unlikely to have much positive impact in the short term, given the complex and highly sophisticated nature of vaccine production.50 The WTO has long been understood as an institution that helps facilitate cooperative policymaking, improving global economic well-being by providing a forum and mechanisms that allow governments to coordinate their actions to minimize their negative international externalities.51v In a world now characterized by cross-border supply chains, there was an unmet need to help coordinate WTO member subsidies and contracts to expand vaccine supply chain capacity at risk, as well as to address the inevitable input shortages that arise early in a pandemic. The WTO would have been the logical institution to take on such a role. 49 Bryce Baschuk and Jenny Leonard. Azevedo Stepping Down Early From a WTO Already on the Brink. Bloomberg. May 13, 2020. 50 See Bollyky and Bown (2020b). The only other initiative was the Trade and Health Initiative, a modest proposal highlighted by actions on trade facilitation and addressing export restrictions. 51 See Bagwell and Staiger (1999, 2002) and for evidence, see Bagwell, Bown, and Staiger (2016). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 27 References Agarwal, Ruchir and Gita Gopinath. 2021. A Proposal to End the COVID-19 Pandemic. IMF Staff Discussion Note 2021/04, May. Agarwal, Ruchir and Tristan Reed. 2021. How to End the COVID-19 Pandemic by March 2022. World Bank Policy Research Working Paper 9632, April. Ahuja, Amrita, 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 Wiecek. 2021. Preparing for a Pandemic: Accelerating Vaccine Availability. American Economic Association Papers and Proceedings, 111: 331-335. Antràs, Pol and Robert W. Staiger. 2012. Offshoring and the Role of Trade Agreements. American Economic Review 102, no. 7: 3140-83. Athey, Susan, 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. Dartmouth College working paper, January. Athey, Susan, Michael Kremer, Christopher Snyder and Alex Tabarrok. 2020. In the Race for a Coronavirus Vaccine, We Must Go Big. Really, Really Big. The New York Times, May 4. Bagwell, Kyle, and Robert W. Staiger. An Economic Theory of GATT. American Economic Review 89, no. 1: 215–48. Bagwell, Kyle, and Robert W. Staiger. 2002. The Economics of the World Trading System. Cambridge, MA: MIT Press. Bagwell, Kyle, Chad P. Bown, and Robert W. Staiger. 2016. Is the WTO Passé? Journal of Economic Literature 54, no. 4: 1125–31. Baldwin, Richard and Rebecca Freeman. 2021. Risks and global supply chains: What we know and what we need to know. NBER Working Paper No. 29,444, October. Bollyky, Thomas J. and Chad P. Bown. 2020a. The Tragedy of Vaccine Nationalism: Only Cooperation Can End the Pandemic, Foreign Affairs v99, n5 (September/October): 96-109. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 28 Bollyky, Thomas J. and Chad P. Bown. 2020b. Vaccine Nationalism Will Prolong the Pandemic. A Global Problem Calls for Collective Action. Foreign Affairs, December 29. Bollyky, Thomas J. and Chad P. Bown. 2021. The Real Vaccine Procurement Problem: Why America Should Make Its Supply Chain More Transparent. Foreign Affairs, June 24. Bown, Chad P. 2022. COVID-19 vaccine supply chains and the Defense Production Act. PIIE working paper, May. Bown, Chad P. 2021. Don’t let CureVac’s COVID-19 vaccine supply chain go to waste. PIIE Realtime Economic Issues Watch, August 9. Bown, Chad P. and Thomas J. Bollyky. 2021. Here’s how to get billions of COVID-19 vaccine doses to the world. PIIE Trade and Investment Policy Watch, March 18. Bown, Chad P. 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. Bown, Chad P. and Soumaya Keynes. 2021. The murky world of export restrictions for COVID-19 vaccines, Trade Talks, Podcast episode 152, March 21. Bown, Chad P. and Chris Rogers. 2021. The US did not ban exports of vaccine supplies. But more help is needed. PIIE Trade and Investment Policy Watch, June 8. Budish, Eric and Christopher Snyder. 2021. Bigger is better when it comes to vaccine production. Wall Street Journal, March 17. Castillo, Juan Camilo, Amrita Ahuja, Susan Athey, Arthur Baker, Eric Budish, Rachel Glennerster, Scott Duke Kominers, Michael Kremer, Greg Larson, Jean Lee, Canice Prendergast, Christopher Snyder, Alex Tabarrok, Brandon Joel Tan, and Witold Wiecek. 2021. Market Design to Accelerate COVID-19 Vaccine Supply. Science 371 (6534): 1107-1109. Cutler, David M. and Lawrence H. Summers. 2020. The COVID-19 Pandemic and the $16 Trillion Virus. Journal of the American Medical Association 324: 1495–1496. Evenett, Simon, Bernard Hoekman, Nadia Rocha, and Michele Ruta. 2021. The Covid-19 Vaccine Production Club: Will Value Chains Temper Nationalism? European University Institute Working Paper, March. GAO (Government Accountability Office). 2021. COVID-19: Efforts to Increase Vaccine Availability and Perspectives on Initial Implementation. GAO Report to Congressional Addressees GAO-21- 443, April. Grossman, Gene M. Elhanan Helpman and Hugo Lhuillier. 2021. Supply Chain Resilience: Should Policy Promote Diversification or Reshoring? NBER Working Paper No. 29330, October. Kominers, Scott Duke. 2021. World Can Have Covid Boosters and Its First Doses, Too. Bloomberg, September 28. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 29 Kremer, Michael. 2001a. “Creating Markets for New Vaccines Part I: Rationale,” in Adam B. Jaffe, Josh Lerner, and Scott Stern (eds.) Innovation Policy and the Economy, Volume 1, pp. 35–72, Cambridge, MA: MIT Press. Kremer, Michael. 2001b. “Creating Markets for New Vaccines Part II: Design Issues,” in Adam B. Jaffe, Josh Lerner, and Scott Stern (eds.) Innovation Policy and the Economy, Volume 1, pp. 73–118, Cambridge, MA: MIT Press. Kremer Michael and Rachel Glennerster. 2004. Strong Medicine: Creating Incentives for Pharmaceutical Research on Neglected Diseases. Princeton, NJ: Princeton University Press. Kremer, Michael, Jonathan D. Levin and Christopher M. Snyder. 2020. “Advance Market Commitments: Insights from Theory and Experience,” NBER Working Paper No. 26775, February. Miroudot, Sébastien. 2020a. Reshaping the policy debate on the implications of COVID-19 for global supply chains. Journal of International Business Policy 3: 430–442. Miroudot, Sébastien. 2020b. Resilience versus robustness in global value chains: Some policy implications. in Richard E. Baldwin and Simon J. Evenett (eds.), COVID-19 and Trade Policy: Why Turning Inward Won’t Work, London, UK: CEPR Press. Pall. 2021. Supporting Rapid Development with a Standardized Single-Use Manifold and AAV Platform Process. Available at https://www.pall.com/en/biotech/blog/technologies-supporting- covid19-vaccine.html, February 3. Snyder, Christopher M., Kendall Hoyt, Dimitrios Gouglas, Thomas Johnston, and James Robinson. 2020. Designing Pull Funding for A COVID-19 Vaccine. Health Affairs v39 n9: 1633- 1642. WTO. 2021. World Trade Report 2021: Economic resilience and trade. Geneva, World Trade Organization. WTO-IMF. 2022. WTO-IMF COVID-19 Vaccine Trade Tracker, last accessed January 18. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 30 > > > FIGURE 1. The vaccine value chain Vaccine origination and development 1 Research and development Send vaccine 2 Clinical trials formula for testing License formula to Vaccine manufacturing Drug substance Capital Raw and single-use Other pharmaceutical 3 and drug product formulation equipment materials ingredients • Bioreactors • Bioreactor bags • Adjuvants Send drug Inputs • Pumps • Cellular material • Lipids product to • Filtration units • Filters • Preservatives • Excipients Capital Other 4 Fill and finish equipment inputs • Vial-filling • Glass vials Send vaccine Inputs equipment • Stoppers dose to • Refrigeration Delivery 5 Distribution Inputs Equipment • Needles • Syringes • Diluents • Antiseptic wipes Note: Stages and inputs depicted illustrate general vaccine production process and are not comprehensive Source: Bown and Bollyky (2022, Figure 1). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 31 > > > F I G U R E 2 . The Pfizer-BioNTech COVID-19 vaccine manufacturing supply chain a. Partners and facilities involved in Pfizer/BioNTech vaccine production as of December 31, 2020 Lipid Drug substance and drug Fill and production product formulation finish Delivery US Firms US Firms US Firms Distribution • Avanti Polar Lipids - • Pfizer - Missouri • Pfizer - Michigan Alabama • Pfizer - Michigan • Pfizer - Massachusetts European firms European firms • Pfizer - Puurs, Belgium Croda - Snaith, UK European firms • BioNTech - Mainz, Germany • Pfizer - Puurs, Belgium (Firms handle different parts of production process) b. Partners and facilities involved in Pfizer/BioNTech vaccine production as of June 30, 2021 Lipid Drug substance and drug Fill and production product formulation finish Delivery US Firms US Firms US Firms Distribution • Avanti Polar Lipids - • Pfizer - Missouri • Pfizer - Michigan Alabama • Pfizer - Michigan • Pfizer - Kansas • AMRI - New York • Pfizer - Massachusetts • Pfizer - Connecticut • Exelead - Indiana European firms • Pfizer - Puurs, Belgium European firms European firms • Siegfried - Hameln, • Croda - Snaith, UK • BioNTech - Mainz, Germany • Polymun - Klosterneuburg Germany • Depharm - Saint-Rémy, Austria • BioNTech - Marburg, France • Evonik - Hanau, Germany Germany • Sanofi - Frankfurt, • Evonik - Dossenheim, • Pfizer - Puurs, Belgium Germany Germany • Pfizer - Dublin, Ireland • Novartis - Stein, • Merck KGaA - Darmstadt, • Dermapharm - Brehna, Switzerland Germany Germany • Therm Fisher - Monza, • Dermapharm - Reinbek, Italy Germany • AGC Biologics - Heidel- berg, Germany (Firms handle different parts of production process) Source: Bown and Bollyky (2022, Figure 2). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 32 > > > F I G U R E 3 . Moderna’s COVID-19 vaccine manufacturing supply chain a. Partners and facilities involved in Moderna vaccine production as of December 31, 2020 Lipid Drug substance and drug Fill and production product formulation finish Delivery • CordenPharma - US Firms US Firms Distribution Colorado, US • Lonza - New Hampshire • Catalent - Indiana • CordenPharma - Liestal, Switzerland European firms European firms • CordenPahrma - • Lonza - Visp, Switzerland • Rovi - Madrid, Spain Chenôve, France b. Partners and facilities involved in Moderna vaccine production as of June 30, 2021 Lipid Drug substance and drug Fill and production product formulation finish Delivery • CordenPharma - US Firms US Firms Distribution Colorado, US • Lonza - New Hampshire • Catalent - Indiana • CordenPharma - • Moderna - Massachusetts • Baxter - Indiana Liestal, Switzerland • Aldevron - North Dakota • Sanofi - New Jersey • CordenPahrma - • Thermo Fisher - North Chenôve, France European firms Carolina • Lonza - Visp, Switzerland • Rovi - Granada, Spain European firms • Lonza - Geleen, • Rovi - Madrid, Spain Netherlands • Recipharm, Monts, France Korean firms • Samsung, Biologics - Incheon, South Korea Source: Bown and Bollyky (2022, Figure 2). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 33 > > > F I G U R E 4 . Johnson & Johnson’s COVID-19 vaccine manufacturing supply chain Partners and facilities involved in Pfizer/BioNTech vaccine production Drug substance and drug product Fill and formulation Finish Delivery European firms US Firms Distribution • Johnson & Johnson - Leiden, • Catalent - Indiana Netherlands • Grand River Septic Manufacturing - Michigan US firms • Merck - Pennsylvania • Emergent Biosolutions - Maryland • Merk - North Carolina European firms • Catalent - Anagni, Italy Asian firms • Reig Jofre - Barcelona, Spain • Biological E - Paonta Sahib, • Senofi Pasteur - Mercy l’Etoile, Himachal Predesh, India France • IDT Biologika - Dessau, Germany Firms in rest of the world • Biological E - Paonta Sahib, Him- achal Pradesh, India • Aspen Pharmacare - Gqeberha, South Africa Source: Bown and Bollyky (2022, Figure 8). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 34 > > > F I G U R E 5 . AstraZeneca’s COVID-19 vaccine manufacturing supply chain Partners and facilities involved in Oxford/AstraZeneca vaccine production Drug substance and drug Fill and product formulation Finish Delivery UK supply chain • Oxford Biomedica - Oxford • CP Pharmaceuticals (Wockhardt) Distribution • Cobra Biologics UK, Keele - Wrexham US supply chain • Catalent - Maryland • AstraZeneca - Ohio • Emergent BioSolutions - Maryland • (Production ended in April 2021) European supply chain • Novasep/Thermo Fisher • Catalent - Anagni, Italy - Seneffe, Belgium • IDT Biologika - Dessau, Germany • Halix - Leiden, Netherlands Indian supply chain • Serum Institute of India (SII) - Pune, India Australian supply chain • CSL - Broadmeadows, Australia • CSL - Parkville, Australia Japanese supply chain • KM Biologics - Kumamoto • JCR Pharmaceuticals - Kobe, Japan prefecture, Japan • Daiichi Sankyo - Japan Brazilian supply chain • Serum Institute of India (SII) - Pune, • Fiocruz Institute - Rio de Janeiro, India (Initially) Brazil Southeast supply chain • Siam Bioscience - Bangkok, Thailand Source: Bown and Bollyky (2022, Figure 8). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 35 > > > F I G U R E 6 . COVID-19 vaccine doses supplied by country and by company, through 2021 Vaccine doses (billions) supplied through 2021, by country 6,000 5,000 4,000 3,000 2,000 1,000 0 China European Union India United States Other Vaccine doses (billions) supplied through 2021, by country 3,000 2,500 2,000 1,500 1,000 500 0 AstraZeneca Pfizer Moderna Johnson & Sinovac Sinopharm Other Johnson Source: WTO-IMF COVID-19 Vaccine Trade Tracker, as of December 31, 2021. > > > F I G U R E 7. Example illustrating benefits from accelerating and expanding vaccine capacity Vaccine doses (billions) supplied through 2021, by country 1.0 Social harm (trillion $) 0.8 C 0.6 0.4 A B 0.2 D 0.0 0 6 12 18 24 EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 36 > > > F I G U R E 8 . How to use supply chain transparency to minimize COVID-19 vaccine input shortages Five policy steps 1 Survey vaccine production facilities about their inputs to establish what they need, where they source from, and on what schedule. 2 Aggregate the information by input-supplying firm to determine the volume that each firm needs to provide. 3 Survey each input supplier to cross-check the data and determine if their existing capacity can meet demand. 4 Identify input shortages. 5 Shortages of customized inputs Short-therm solution: For general inputs Use data accumulated in step 3 to identify Increase production at existing facilities, alternate suppliers with spare capacity e.g., incentivize addition of second, third, and weekend shifts. Long-therm solution: Incentivize investment to expand capacity. Use subsidies when there is insufficient private (market) incentives. If input shortfalls still arise, policymakers can help ration limited suppliers. Source: Bown and Rogers (2021, Figure 6). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 37 > > > TA B L E 1 . US contracts to COVID-19 vaccine sponsors, February 11, 2020–October 22, 2021 Amount Company (Millions Date Task (DPA priority rating) of dollars) 21 February 11, 2020 Support nonclinical studies and a Phase 1 trial 436 March 27, 2020 Contract amendment 1,002 August 5, 2020 Demonstrate large-scale manufacturing (100 million doses) Johnson & 85 August 21, 2020 Unknown Johnson 0 September 21, 2020 Post-award modifications, including award of priority rating for contracts (DO) (Janssen) 454 November 13, 2020 Support Phase 3 trial (contract amendment) Collaboration with Merck to repurpose its facilities for drug substance and fill and 269a March 2, 2021 finish, DPA invoked (priority rating unknown) 32 March 25, 2021 Expand Phase 2a trial for adolescent population 31 April 10, 2020 Accelerate nonclinical studies and Phase 1 trial Conduct Phase 3 trial, support manufacturing demonstration project for 100 million 2,042 July 30, 2020 Sanofi-GSK doses 0 June 4, 2021 Priority-rating clause of US government contract removed 6 August 6, 2021 Unknown Merck and 38 April 15, 2020 Accelerate development of vaccine candidate IAVI 430 April 16, 2020 Accelerate development of vaccine candidate 53 May 24, 2020 Expand manufacturing capacity 472 July 25, 2020 Support Phase 3 trial 1,525 August 11, 2020 Support manufacturing of 100 million doses, with option for 400 million more Contract amendment to give Health Resources Priority and Allocations System 0 September 8, 2020 (HRPAS) priority rating (DO) Moderna 1,667 December 11, 2020 Purchase another 100 million doses 1,750 February 11, 2021 Purchase another 100 million doses 63 March 12, 2021 Support Phases 2 and 3 of adolescent study and booster for adults 236 April 18, 2021 Support for clinical studies (cost increase) 144 June 15, 2021 Support Phase 2 and 3 trials for children six months to 12 years old 3,304 June 15, 2021 Purchase another 200 million doses 413 May 20, 2020 Support clinical development and manufacturing AstraZeneca Accelerate development and manufacturing to begin Phase 3 trial and make avail- 1,200 October 28, 2020 (Oxford) able 300 million doses (DO) 0 June 4, 2021 Priority-rating clause of US government contract removed 60 June 4, 2020 Manufacture components for use in Phase 2 and 3 trials 1,600 July 6, 2020 Demonstrate commercial-scale manufacturing for 100 million doses Novavax Contract modification awarding priority rating for procurement of raw materials, con- 0 September 10, 2020 sumables, repair parts, and major end item assemblies (DO) 0 June 4, 2021 Priority-rating clause of US government contract removed 1,950 July 21, 2020 Purchase 100 million doses 2,011 December 22, 2020 Purchase another 100 million doses, with option for 400 million more (DO) 2,011 February 11, 2021 Pick up option to purchase 100 million doses Pfizer 4,870 July 21, 2021 Pick up option to purchase 200 million doses (BioNTech) Purchase 500 million doses for donation to COVID-19 Vaccines Global Access (CO- 3,500 July 30, 2021 VAX) Purchase 50 million pediatric doses (age 5–11), one third the strength of those 1,230 October 22, 2021 intended for 12 years and up Note: a. Payment to Merck for the collaboration. Sources: Compiled by the author from Biomedical Advanced Research and Development Authority, 2021, BARDA’s Rapidly Expanding COVID-19 Medical Countermeasure Portfolio, BARDA’s COVID-19 Domestic Manufacturing & Infrastructure Investments, and publicly available firm contracts. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 38 > > > TA B L E 1 . US federal subsidies or contracts to COVID-19 vaccine input suppliers, February 11, 2020–June 30, 2021 Amount Company (Millions Date Task (DPA priority rating) of dollars) SiO2 Materials 143 June 5, 2020 Establish US-based production for glass tubing and vials Science 204 June 5, 2020 Corning Expand capacity for glass tubing and vials 57 March 23, 2021 Becton, Dickinson 42 July 1, 2020 Expand capacity for syringes and needles and Co. Retractable Technolo- 54 July 1, 2020 Expand capacity for syringes and needles gies Smiths 21 July 11, 2020 Expand capacity for syringes and needles Medical Cytiva 32 October 13, 2020 Expand capacity for cellular material, mixer bags, and bioreactors ApiJect 590a November 19, 2020 Expand capacity for prefilled, single-dose injectors Systems Meissner Filtration 13 April 1, 2021 Expand capacity for filtration products for vaccine manufacturing Products a. Loan to finance 75 percent of project’s capital costs. Sources: Bown and Bollyky (2022, Table 4), compiled from Biomedical Advanced Research and Development Authority, 2021, BARDA’s Rapidly Expanding COVID-19 Medical Countermeasure Portfolio and BARDA’s COVID-19 Domestic Manufacturing & Infrastructure Investments; and US International Development Finance Corporation. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 39