Market Design to Accelerate Vaccine Supply

Market design to accelerate COVID-19 vaccine supply is my new paper in Science, co-authored with Camilo Castillo, Michael Kremer, Eric Budish, Susan Athey and others. We make three vital points. First, governments invested much less than our group advised. We spent trillions on fiscal support and maybe $20 billion or so on vaccines, far too little. Nevertheless, the 3bn courses we have (conservatively) in 2021 capacity is worth on the order of $17.4 trillion or $5800 per course. If advance market commitments moved us from 2 billion to 3 billion courses then they were worth 2.4 trillion dollars. I feel pretty good about the work we did to encourage Operation Warp Speed and other advance purchases.

Second, it’s not too late to do more. If we could get an additional billion courses in capacity online by July 2021 that would speed up vaccination in high-income countries by 1.4 months and in the world by 4.3 months. A few months might not seem like much but that speed-up is worth half a trillion to the world economy. If we could get additional capacity online by April it would be worth a trillion dollars.

You might think that getting more capacity online by April isn’t possible but you can do a lot for a trillion dollars. Moreover, we can increase capacity not just by building more factories but by using the doses we have now more wisely. Low-dose syringes, for example, can increase supplies by 20%. I think the health authorities know this now (although they should have been prepared) but even at this late stage almost everyone is under-estimating how much it would be worth spending to get 20% more vaccine capacity. Similarly, going to half-doses is equivalent to doubling the number of Moderna and Pfizer factories. Even if we did half doses for the young alone, that’s a big increase in supply. We calculate that additional capacity is worth $576 to $989 per annual course, far higher than the price.

Third, we also give advice on how to structure contracts. Buying doses isn’t optimal because companies can just agree and put you to the back of the queue. Optimal rewards and penalties are very difficult to implement, especially when optimal penalties could bankrupt firms many times over (because the social value of vaccines is much greater than the private value.) So it’s much better to subsidize capacity with an option to buy doses at a discount produced from that capacity–this is similar to what Operation Warp Speed did with Moderna and Novavax.

Finally, here’s a fourth important point I haven’t made earlier. We suggest procurement auctions to surface prices on necessary inputs. Ordinarily, an increase in demand to a final producer such as a vaccine manufacturer is transmitted along the entire supply chain through the signaling and incentive mechanism of prices. When final goods prices are limited socially or by law, however, the supply chain can become dis-coordinated. Capacity contracts could be fulfilled, for example, and the producer could yet claim an inability to produce because raw materials are in short supply. Thus, we need a mechanism to coordinate supply chains.The US Defense Production Act is one such mechanism. An alternative procedure that may work more quickly is to organize procurement auctions for all the inputs and complementary goods required for vaccine production. The advantage of a procurement auction is that it can attract and incentivize firms globally, firms that are well beyond the reach of the DPA.

Read the whole thing and the appendix with details on methods and calculations.