All podcasts / People I (Mostly) Admire / Summary

166. The World's Most Effective Public Health Intervention Is Under Attack

2025-09-13 - source - Read full transcript

Steve Levitt (host)Seth BerkleyMorgan Levey

Key insights

Gavi delivers the full W.H.O.-recommended childhood vaccine schedule for about $24 per child, a 98 percent discount off the roughly $1,300 U.S. price for equivalent vaccines.

Berkley attributes the reduction to bulk buying and expanding the manufacturer base. The cost-benefit ratio is roughly $21 returned per dollar invested in narrow terms, rising to $54 per dollar when broader health and quality-of-life gains are included, which Berkley calls unmatched anywhere else in healthcare.

vaccine-economics

Vaccines are underinvested in specifically because pharma has little financial incentive to develop them for diseases concentrated in poor countries.

Levitt argues that if developing-world patients could be charged U.S.-level prices, companies would race to invest in R&D; because they can't, diseases like malaria went roughly 50 years without a vaccine despite killing 600,000 people annually. Gavi's pitch to manufacturers was that serving poor markets adds volume, cuts cost of goods for their whole product line, and generates real profit even at steep discounts.

vaccine-economics

Vaccine development is empirically bottlenecked, not just financially: it fails about 93 percent of the time and previously took roughly 10 years, with the fastest pre-Covid vaccine at four years.

Berkley says vaccines must ultimately be tested in humans because animal models are unreliable ('mice lie, monkeys don't always tell the truth'), and every batch has to be re-tested if any input changes because a vaccine, unlike a chemical drug, is a living, hard-to-standardize product.

vaccine-economics

The Covid vaccine's 327-day development time was possible only because of pre-existing SARS/MERS research plus mRNA technology, not because vaccine science had become generically fast.

Prior work on other coronaviruses meant scientists already knew the right antigen target (the spike protein) going in. Combined with mRNA's speed advantage over traditional platforms, the sequence-to-injectable timeline compressed to about 42 days for design and 63 days to first injection, an unprecedented pace Berkley says reflects a rare stacking of prior knowledge and new technology, not a repeatable baseline.

vaccine-economics

Preventive interventions structurally fight against human psychology because their success is invisible: nobody experiences the deaths that didn't happen.

Levitt connects this to his own son's death from pneumococcal meningitis just four months before the Prevnar vaccine was approved, arguing that a successful vaccine erases the very horror that would justify its existence in people's minds, breeding complacency and openness to anti-vax narratives.

science-communication

Vaccine hesitancy is often a causality-attribution error, not a rejection of evidence, made worse because vaccines are given during a developmental window when unrelated symptoms (e.g., autism) become visible.

Because nearly all children receive the same vaccines (e.g., MMR) around the same age, any subsequent bad outcome will correlate with recent vaccination by default, making it easy for parents to wrongly infer causation, particularly when they lack broad data analysis to compare against a proper baseline.

science-communication

Covid politicized public health messaging in a way that has had a lasting corrosive effect on scientific communication.

Berkley says he views public health as inherently non-partisan and evidence-driven, but partisan attacks changed how honest agencies were willing to be, since science requires an iterative, self-correcting process of updating recommendations as data comes in, and updating publicly gets read as prior incompetence rather than good practice.

science-communication

Gavi only exists as a functioning public-private alliance because it built a market: reliable purchase commitments let manufacturers justify investing in developing-country vaccine supply.

Before Gavi, fewer than 5 percent of the world's children received a single vaccine dose (1974 baseline), and U.N./W.H.O. agencies distrusted industry mutually. Gavi's founding innovation (2000) was creating dependable, multi-year purchase mechanisms so manufacturers, including new developing-world manufacturers, would commit to production.

global-health-institutions

Bill Gates's initial $750 million commitment (after reading the World Bank's cost-effectiveness case twice) was the singular catalyst that made Gavi possible, though he is now only about 15 percent of its funding base.

Berkley contrasts this to his prior role running the Rockefeller Foundation's global health budget of about $15 million worldwide, illustrating how far outside prior philanthropic scale Gates's initial investment was and how it built a broader donor consortium afterward.

global-health-institutions

Covax was a deliberate bet to buy vaccines that didn't yet exist with money Gavi didn't yet have, to avoid repeating past pandemics where rich countries bought up all available supply and left nothing for poorer nations.

Gavi and CEPI moved early (Davos, late 2019/early 2020) based on informal signals of person-to-person transmission, ultimately delivering 2 billion vaccine doses to the poorest countries, though the lowest-income countries still lagged badly behind high-income vaccination rates in 2021 due to vaccine nationalism and export bans.

pandemic-preparedness

The U.S. is currently underprepared for the next pandemic despite Covid demonstrating both the danger and the solution: rapid vaccine platforms like mRNA.

Berkley calls recent NIH cuts to mRNA vaccine research 'the stupidest science decision' a colleague has heard of, noting mRNA speed matters not just for pandemics but for personalized cancer immunotherapy. He cites modeling suggesting a 50 percent chance of a Covid-like pandemic in the next 25 years, and warns that defunding institutions now will also drive young scientists out of the field long-term.

pandemic-preparedness

In an all-pay (ascending) auction, every bidder pays their final bid regardless of winning, which creates an escalation trap with no natural stopping point once a bidder is 'underwater.'

Levitt illustrates with a $20 bill auction: once a bidder reaches $19-20, stopping guarantees a near-total loss while bidding one more dollar only slightly increases the loss if outbid, so rational bidders keep raising past the item's actual value. A listener who tried this within his own family found it failed as a demonstration because family dynamics (side conversations, refusal to play) override the auction psychology that works on strangers or students.

vaccine-economics

Books referenced

Fair Doses: An Insider's Story of the Pandemic and the Global Fight for Vaccine Equity - Seth Berkley - Berkley's forthcoming book (published October 28, 2025) on vaccine equity and the pandemic response, referenced repeatedly as the source for the Ebola stockpile and epidemiologist detective-work stories.

Media referenced

People I (Mostly) Admire episode with Moncef Slaoui - podcast - Earlier PIMA episode with the former head of Operation Warp Speed; Levitt compares Slaoui's account of betting on mRNA to Berkley's.
People I (Mostly) Admire episode with Sendhil Mullainathan (2021) - podcast - Origin of the $20 all-pay auction demonstration that listener Ian later tried on his family.

Companies

Gavi, the Vaccine Alliance - Public-private partnership Berkley ran for over a decade; has vaccinated 1.1 billion additional children and is credited with saving over 20 million lives.
CEPI (Coalition for Epidemic Preparedness Innovations) - Organization Berkley worked with at Davos in early 2020, led by Richard Hatchett, that helped seed what became Covax.
Rockefeller Foundation - Berkley previously ran its global health programs, with a budget of about $15 million worldwide, before joining Gavi.
Pfizer - Example of a manufacturer providing its pneumococcal vaccine at a 95-percent price reduction to developing countries while still profiting via economy of scale.
World Bank - Published the 'World Development Report' that Bill Gates read twice, which made the cost-effectiveness case for vaccines and helped catalyze Gavi's founding.

Techniques and frameworks

All-pay (ascending) auction / $20 bill auction - A classroom demonstration where every bidder pays their final bid, not just the winner, creating an escalation trap; discussed in the listener-question segment with producer Morgan Levey.
Fill-and-finish vaccine stockpiling - Keeping candidate vaccines in bulk storage rather than vials until an outbreak occurs, then rapidly vialing, testing, and distributing them, as Gavi did with Ebola.

Summary

Steve Levitt talks with Seth Berkley, who ran Gavi, the Vaccine Alliance, for over a decade, about why vaccines may be the single best return-on-investment intervention in public health and why that fact hasn't protected the field from political attack. Berkley opens with the numbers: Gavi has vaccinated 1.1 billion additional children, saved over 20 million lives, and delivers the full W.H.O.-recommended vaccine schedule for about $24 per child versus roughly $1,300 for the same vaccines in the U.S. The economics only work, he explains, because Gavi built a functioning market: reliable, multi-year purchase commitments let manufacturers justify investing in developing-world supply, turning what used to be a market failure (fewer than 5 percent of the world's children got even one vaccine dose in 1974) into a system serving 60 percent of the world's children today. Bill Gates's initial $750 million commitment, prompted by a World Bank report on vaccine cost-effectiveness, was the singular catalyst; Berkley's own prior global-health budget at the Rockefeller Foundation had been about $15 million for the entire world.

A recurring theme is why vaccine development is genuinely hard, not just underfunded. Roughly 93 percent of vaccine candidates fail, historically taking around a decade to develop, because vaccines are living products (unlike chemically pure drugs) that require expensive re-validation whenever any input changes, and ultimately must be tested in humans since animal models are unreliable. The Covid vaccine's stunning 327-day timeline, Berkley stresses, was possible only because prior SARS/MERS research had already identified the right target (the spike protein) and because mRNA is uniquely fast to redesign, not because vaccine science generally got faster. He walks through parallel case studies, malaria (50 years without a vaccine because there was no market incentive), Ebola (a stockpile built only because it briefly landed on a post-9/11 bioterrorism watch list), and HIV (still unsolved because the virus mutates faster within a single infected person than the immune system, or a vaccine, can track), to show how market incentives, not scientific ceiling, largely determine which diseases get solved.

The conversation turns pointed on vaccine hesitancy and Trump-administration policy. Levitt shares that his own son died of pneumococcal meningitis just four months before the vaccine against it was approved, and argues that successful prevention is psychologically self-defeating: nobody experiences the deaths that didn't happen, so the intervention erodes the very fear that justified it. Berkley adds that most hesitancy is a causal-attribution error rather than genuine skepticism of evidence, since nearly all children get the same vaccines around the same age that other developmental changes (like signs of autism) become visible, making false correlation almost automatic. Both agree Covid politicized public health messaging in a lasting, corrosive way. Berkley is unusually blunt about the current administration, calling recent NIH mRNA-research cuts "the stupidest science decision" a colleague has heard of and describing agency cuts as taking "a chainsaw" to institutions that need reform, not demolition, warning this will also drive young scientists out of the field for a generation.

On Covid specifically, Berkley describes Gavi's decision to help build Covax, essentially placing pre-orders for vaccines that didn't exist yet with money Gavi didn't yet have, as a calculated gamble to avoid repeating history, where rich countries historically bought up all available pandemic vaccine supply and left nothing for poorer countries. Covax ultimately delivered 2 billion doses to the poorest countries, though vaccine nationalism and export bans meant low-income countries still lagged far behind high-income ones through 2021. Berkley closes with a stark forward-looking note: models suggest a 50 percent chance of another Covid-like pandemic within 25 years, and current U.S. policy is doing "as little as possible" to prepare, despite Operation Warp Speed being, in his view, one of the first Trump administration's genuine successes.

The episode closes with the usual listener-question segment: producer Morgan Levey relays a story from a listener named Ian, who tried recreating the show's "$20 all-pay auction" demonstration (where every bidder pays their final bid, creating an escalating trap with no natural stopping point) on his own family in a pub, and it backfired completely because family dynamics don't follow the same psychological rules as classroom or stranger settings.

Notable Quotes

"These are amazing technologies that have lifted us up from a situation where people are losing a third of their kids routinely, to a place where people assume you have one child, that child's going to live and be healthy." - Seth Berkley

"The irony of this is that it is the most incredible intervention, and yet because there is so little money in it for the pharma companies... there's been a systematic underinvestment in vaccines in R&D over the last 50 or 60 years." - Steve Levitt

"Mice lie, monkeys don't always tell the truth, and therefore, you have to ultimately do it in humans to really know whether it's going to work." - Seth Berkley, quoting a vaccinologist colleague

"The decision about mRNA vaccines, one of the scientists I work with describe that as, The stupidest science decision he's ever heard in his career." - Seth Berkley

"Our failure to prepare for the next pandemic may be the single worst policy mistake of our time." - Steve Levitt