Encouraging the Development and Manufacturing of Vaccines and Diagnostics
Michael Kremer | Episode 12
Michael Kremer joined the Princeton Bendheim Center for Finance for a webinar on COVID-19. Kremer is a Professor of Developing Studies in the Department of Economics at Harvard University and a 2019 recipient of the Nobel Prize in Economics.
Watch the full presentation below and download the slides here.*
Highlights
[1:06] Markus’ introduction: Vaccine uptake generates positive infection externalities that are not fully internalized privately, implying underinvestment in both vaccination and vaccine R&D absent policy intervention through mandates, subsidies, or prize-like mechanisms.
[10:19] Beginning of Kremer’s presentation: COVID-19 is estimated by the IMF to cost about $9 trillion over two years, implying roughly $375 billion in monthly losses and very large social returns to accelerating vaccine availability even by a single month.
[13:40] Is it worth building large-scale capacity in parallel with trials? Large-scale vaccine capacity is normally installed only after trials because plants are complex and costly, but COVID externalities and monopsonistic public buyers mean private developers underinvest in early, parallel capacity from a social perspective.
[17:57] Back of the envelope calculations: Calibrating historical vaccine success rates by technology and stage, and allowing for correlated risks, implies that achieving a 90% chance of at least one successful COVID vaccine within 18 months requires advancing roughly 15–14 candidates in parallel.
[21:46] Calculating the optimal number of vaccine candidates: Equating expected social benefits from earlier vaccination (probability gain × $375 billion per month × roughly six months saved) with capacity costs suggests that investing in about 15 candidates yields benefits (~$1.6 trillion) that dwarf total manufacturing costs (~$90 billion).
[25:40] Structuring funding: Mechanism design must handle asymmetric information about success probabilities and costs in multiproduct firms, implying the need for firms to retain “skin in the game” rather than full cost reimbursement or pure command-and-control capacity mandates.
[28:27] Proposed funding structure: A scheme in which public funders cover up to 80% of capacity expenditures in exchange for an option to buy doses at preset two-part prices can induce self-selection of higher-quality projects while preserving incentives for efficiency and speed.
[32:39] Calculating required pull funding: Given assumed capacity costs, production costs, and an 11% marginal success probability sharing a one-third expected market, required pull funding per successful firm is around $40 billion, rising to roughly $50 billion when strengthening incentives and allowing for multiple vaccines.
[34:48] Pull funding: Two-part pricing with a high price (e.g., $35) for the first billion people vaccinated and a near-marginal-cost price (e.g., $5) for subsequent 3.5 billion supports both dynamic incentives and static allocative efficiency at scale.
[35:48] Supply chains: For want of a vial, the vaccine may be lost: Vaccine manufacturing is constrained by fragile supply chains—particularly fill-finish operations and glass vials—creating cross-firm externalities and a rationale for public coordination or targeted subsidies for critical inputs.
[37:09] Could international cooperation be incentive compatible? A global partnership where members contribute roughly 0.15% of GDP could finance enough diversified capacity to vaccinate all vulnerable populations within three months and around 4.5–6 billion people within a year, with membership rewarded by price caps and priority access.
[50:01] Innovations to speed vaccine trials: Trial innovations—including interim data access, multi-arm adaptive designs sharing control groups, and ethically debated human challenge trials—could sharply reduce sample size and duration, thereby lowering required capacity spending and shortening the pandemic.
[1:07:19] Conclusion: Actual national and multilateral initiatives (e.g., ACT Accelerator, “Warp Speed”) are moving toward larger vaccine investments, but current efforts still fall short of the portfolio size and manufacturing scale implied by the back-of-the-envelope global welfare calculations.
Summary produced with artificial intelligence.