How to vaxx the world
COVID started a global conversation about vaccines. There's a bigger picture we've been missing.
In African countries, millions of people died of easily preventable diseases, most notably malaria and HIV. Of these, more than a majority were children. Almost 2.5% of all “health losses” in the world were due to these “tropical” diseases. And yet very little progress has been made in advancing a vaccine for any of them (except Ebola). For instance, decades of public health work managed to slash malaria deaths in half - but still over 400,000 people, most of them young children, die of the disease.
Additionally, vaccinating people is both cheap and effective in Africa- extending life by one year through vaccinations for the simplest diseases can cost as little as $16 per person per year of life. This is also more effective than curing the diseases themselves.
Before COVID-19, the record for the fastest discovery of a vaccine was measles: the agent linked to the disease was discovered in 1953, and a vaccine was ready in 1963. COVID-19 set the record, with just 10 months. There is a technical challenge for this, to be fair: tropical diseases tend to mutate rapidly and develop multiple variants. But that can’t explain why there has been scarce progress for nearly 60 years.
How vaccines work
Since COVID started, we’ve heard a lot about vaccines: first and second doses, patents, supply issues, development, costs, prizes, and a whole lot more. So I decided to look into the basics: how do vaccines work, and what fun economic phenomena lie behind them.
In a nutshell, a vaccine is like a training exercise for the immune system. Most vaccines work by inserting some sort of deactivated or weakened virus, bacteria, or fungus (called a pathogen) into the bloodstream, so the body can fight it off safely and learn all about how to prevent it. The way this works is, each pathogen requires a specific “weapon” to be contained, known as an antigen. Some newer vaccines, called MRNA vaccines, directly contain the instructions on how to make the antigens rather than a pathogen.
Regarding the COVID vaccines, there’s actually 260 different ones - 96 of which are being clinically tested. The higher profile ones (Pfizer and Moderna) are mRNA vaccines, while only the Chinese vaccines actually use “traditional” inactivated viruses. Another common approach, which is the base of the Johnson & Johnson, AstraZeneca, or Sputnik V vaccines (bizarrely, the V stands for vaccine, not five) is to use a harmless virus to insert the instructions to fight COVID. And there’s a bunch of other kinds: vaccines based on live viruses, vaccines based on proteins, vaccines that use “virus like particles”, and even whatever the hell a self replicating vaccine is. I really recommend looking at the Miliken Institute Vaccine Tracker (also linked above) if you have like four hours to read up on every single proposed COVID vaccine.
Developing a vaccine is a complicated, expensive, time-consuming process: most vaccine approvals take about a decade to complete, but governments (understandably) fast-tracked approving the vaccines for COVID. Basically, the process has four main stages: a pre-clinical phase where the vaccine is actually developed and analyzed for side effects and tested on animals, a phase one trial where efficacy is tested on a small group of healthy people, a phase two trial that mostly focuses on the hows and whens and uses hundreds of people, and a phase three trial that tries to whittle all potential problems down by testing thousands of people. Usually, we would have expected that the “delay” in producing a vaccine against the coronavirus would come from actually making it - but it didn’t. In fact, Moderna’s vaccine (approved first in November) was actually made in a single weekend in January, and then was in various clinical trials for the rest of the year.
Normally, supply of vaccines isn’t really a problem because, since approving them takes a decade or so, there’s plenty of time to scale up production. But the COVID vaccines were approved in under a year, so actually supplying them was a problem - and an optimal solution would have been governments agreeing to pay upfront for them even before approval, which was obviously politically unfeasible. So instead we got a demand glut and rich countries buying excessively large numbers of vaccines (Canada bought doses for multiple times its population), while poor countries mostly scrambled and settled for “second tier” vaccines (which are generally just as good - when clinical trial results were actually made public).
Not enough arms…
Basically, besides the individual benefit of “not dying”, there’s also a group benefit. A lot of people can’t actually get vaccinated: for example, people with compromised immune systems can’t actually get even the weaker viruses that make up the vaccines; and some vaccines aren’t recommended for people who are very young or old. Additionally, the big benefit is herd immunity: if enough people are immunized to a virus, then nobody gets it because it doesn’t have enough hosts that can spread it in the first place.
This is a classic example of an externality: being vaccinated has a private benefit (being immune), but there’s also an external benefit (herd immunity) that doesn’t get taken into account. As a result, if everyone is left to their own devices, then not enough people value being vaccinated enough, and an insufficient number of people get vaccinated. Given that the more contagious a disease, the more people need to get vaccinated, then the stage is set for big coordination failures. Plus, most of the people who benefit from herd immunity are children (often very young), so by definition there’s not much they can do to advocate for themselves. Lastly, there’s also the issue that people frequently undervalue the benefits of preventing vs curing a disease, for the painfully obvious reason that you can clearly see when you get better from a disease, but not getting it in the first place isn’t a tangible benefit.
Consequently, the government has a solid case to intervene: it can simply mandate people get vaccinated (children, who are extremely susceptible to disease, aren’t allowed into schools without all their shots up to date), it can cover the cost of vaccinations, or it could even pay people for vaccines. For example, many countries in Latin America cut off families from cash transfers if children aren’t vaccinated, and experiments in rural India showed that giving parents a bag of lentils worth 1 dollar raised vaccination rates of infants sixfold. Governments can also solve the issue of underconsumption of vaccines, which tend to be expensive (for reasons seen below) by buying lots of them at below-market rates and then distributing them cheaply or for free, which is benefitial for both parties because individuals value a vaccine less than the market price, and because profits are maintained by lower margins but larger volumes.
… and not enough shots
There’s also a significant number of market failures on the supply side of vaccines. First, as I said, making vaccines is incredibly risky and expensive. Secondly, vaccines are patented, which means that most of the benefits are recouped by charging monopoly prices over them. Thirdly, differing regulations and criteria means that simply blocking some countries from accessing vaccines is easy, which means price discrimination (i.e. charging, say, Luxembourg more than Nigeria) is possible.
Finally, vaccine markets have very the wrong incentives for research: since the richest countries are the ones that will pay the most for any pharmaceutical, then most R+D efforts will be made towards the kinds of diseases that affect them, while infectious diseases that mostly affect poor countries get ignored - only 5% of all private R+D health spending was on this type of disease, compared to 50% of all health spending; and f the 1233 drugs licensed wordlwide between 1975 and 1997, only one was a privately developed drug for tropical diseases. This pattern is especially troubling because the kind of disease that poor countries almost exclusively deal with is also much more dangerous.
The typical example for a neglected disease is malaria. Malaria kills an estimated 55,000 people every year, and it’s estimated to be the single largest cause of death in all of human history. And in fact, it was common basically everywhere around the world until the 1940s - for instance, Abraham Lincoln had malaria, and he lived in Illinois his whole life. The main causes of the decline of malaria cases in rich places were public health measures, economic development, and agricultural changes that led to swamp drainages.
A further problem is that the way vaccine markets work, themselves, makes research less desirable. The key issue here is time inconsistency: governments can commit to one price (say, the market value) before a vaccine exists and then demand a much lower one after it’s developed, because they have the strongest bargaining position (due to monopson, i.e. monopoly but for buying). Pharmaceutical companies can predict this will happen, so a reasonable response is simply to not research at all. Additionally, developing countries aren’t known for having the most stable or honest governments, which furthers the inconsistency issue - and, since they don’t have strong IP protections or don’t enforce them, local or international rivals could simply copy the vaccine, patent it within the country, and profit off it by undercutting the inventor firm.
A possible solution is government funded R+D, but it isn’t free of problems either. Research that unlocks immunity from diseases is a global public good, which means that all countries can benefit from it regardless of whether they pay the costs. As a result, nations have incentives to become “free riders” and not invest enough (or anything) because they assume others will pick up the slack. This problem is very noteworthy for small, poor countries - they can save each a lot of money by not researching enough, but their combined slacking means that barely any progress in things like malaria research gets done because rich countries don’t care much about it.
Pushing and pulling
There’s, broadly speaking, two kinds of programs. The first are “push” programs, that incentivize research in general in an area - say, more funding for malaria medications. The second ones are “pull” programs, that reward research when it yields a finished product - say, a prize for the first company to develop a malaria medication that’s more effective than existing ones. In the first world, a combination of both types of programs has been successful at tackling many illnesses - so it stands to reason that a (perhaps adjusted) mix of them would be successful at targeting the conditions that ail the third world.
Push programs have historically been the norm, but pull programs have two main advantages: they incentivize shooting for projects that are likely to succeed by carefully selecting them, and they incentivize focusing only on viable solutions to the proposed problem. On the contrary, push programs incentivize exaggerating the chances of actually achieving anything, and don’t actually provide any insights into what kind of projects to prioritize. This discrepancy is most pronounced for applied research, and least pronounced for basic research. Plus, government grants tend to favor large, established players in the industry who can effectively lobby to receive them- for instance, the US government may have awarded it to Merck’s failed COVID vaccine and shunned Moderna under this regime.
A good example of a pull-adjacent program is the US Ophan Drugs Act of 1983. In a nutshell, an orphan drug is a medicine for a very rare disease - say, Huntington’s, or ALS. Since very few people ever get them, developing and manufacturing these drugs is not profitable, so in consequence there are fewer of them - and at typically higher prices. The Ophan Drugs Act attempted to fix this by granting a variety of benefits, like subsidies for clinical trials or seven years of exclusivity to whoever made a drug for these rare diseases. In the decade before the act, there were 10 orphan drugs approved for use in the US; in the decades after, more than 200 were. There’s actually a very good episode of 99% invisible about this issue.
The pull program par excellence is the Advance Purchase Commitment or APC. An APC is a promise by a government or organization to buy the first vaccine or medicine to be approved for a given disease at a premium. The benefits are clear and are listed above: this promotes one single goal, has transparent criteria, and offers a fair shot at all participants. Plus, if no vaccine is found, no money is spent. Another, ex ante identical pull commitment is a cash award for whoever first discovers a vaccine, which should be identical if the amount awarded is equally large (say, if you buy a million vaccines one dollar above cost, or buy them at the cost value but pay a million dollars in prizes).
There is a counterargument to be made: maybe the problem is (in an eerie parallel of recent debates) intellectual property itself. If medical research weren’t done for profit, then information could be shared more effectively and cooperation, not competition, would result in this - all funded by the public sector. Of course, this might (or might not) be reasonable. To quote Kremer & Glennerster (2003) replying to criticism:
The R&D system for rich-country pharmaceuticals is imperfect, and debate over how the entire pharmaceutical R&D system should be structured is certainly useful. However, if we think we should move to a system akin to open source software for pharmaceuticals, why should we do so just for products for the poor? If the system is not good enough for rich countries, why is it good enough for poor countries?
But there is a stronger, much stronger, case against patents to be made. It’s not a case against the concept of patents itself, but rather, against the concept of private ownership of them in this area of the economy.
The problem patents were designed to solve is the underprovision of ideas. Basically, people won’t put resources into research if they pay all the costs and then their work is immediately stolen. So the way patents work is by providing the person or company that invents an idea with a monopoly on it for a certain period of time.
The problem is that monopoly prices are higher than competitive prices, so people get shut out of the good - for example, millions of cases of pediatric AIDS in Africa were caused by the high cost of AZT during the 80s and 90s. Plus, patents encourage investing in small improvements to existing patented products that wouldn’t improve outcomes as much as new designs, but would be far more profitable. Thirdly, low-end social returns to R+D investment are twice as high as the higher estimates for private returns - 50% vs 25%; this means that, since private companies don’t take this into account, research is always undersupplied. In consequence, as much as a quarter of all benefits from a patented product are lost because of the monopoly pricing.
What are the ways around it? Abolishing patents seems like it would cause more problems than it would solve. And government funded R+D, as stated above, has plenty of problems: too many decisions in the hands of uninformed bureaucrats, endless potential for rent-seeking, and bad choice of projects. The returns (social and private) for government funded research seem to be much lower than the ones on private projects. Of course, push and pull incentives like advance commitments or cash prizes could be an option.
But there’s another alternative to the status quo or to push and pull incentives: buying out (socially useful) patents. An ancient example of this exists: in 1837, the French government bought the patent for the daguerrotype (basically a very primitive photograhphy device) from Louis Daguerre and immediately released it into the common domain, allowing the technology to be disseminated across the globe. The same process (albeit much more messily) was used by the state governments of North Carolina and Tennessee to acquire the rights to the cotton gin. A big benefit, besides from pushing down costs, is that many techniques aren’t ever patented, but rather kept as trade secrets, which hurts knowledge spillovers generated by the invention - Daguerre’s discovery eventually led to the creation of photography.
The biggest risks here are two sides of a same coin: that rent seeking inventors would lobby government officials to buy out useless patents, or that unscrupulous government officials would rip off inventors by all but expropriating very lucrative patents. There’s also a more technical risk, that governments buy out early patents that are technically inferior; for example, at roughly the same time as Daguerre, and Englishman named Talbot developed a similar process but charged for teaching it, meaning it was less widely adopted - and we can’t actually know if Talbot’s process was much more advanced, or efficient, than Daguerre’s.
There’s two big problems to solve for patent buyouts. The first is at which price to sell the patent; this is a tricky, technical issue, and mostly revolves around designing an auction that incentivizes companies and inventors to reveal the value of the invention and then to flip a coin and either award it to one of them, or be bought out by the government at a markup representing social benefits. The second tricky issue is how to decide which patents to buy out; the optimal solution is to just do it for literally all patents and then simply let the inventor decide whether to take the offer made for their patent or to reject it and run the thing again. Now, I’d argue that the government should be able to willingly choose to overpay for certain types of patents that could have big benefits - like, say, vaccines for deadly diseases. But that’s a topic for people much smarter than me to argue about.
The endgame
So far in human history, only two diseases have ever been eradicated: smallpox, and rinderpest; and right now, we’re closing in on eradicating polio and guinea worm disease as well. The eradication of smallpox took nearly two centuries to complete since the vaccine was invented - from 1796 to 1977. Rinderpest was a rare disease transmitted by cattle, and it basically erradicated itself by the before the development of a vaccine. So far, most of the diseases we have eradicated or are on the way erradicate have followed successful global vaccination campaigns
We have a problem with two sides: on the one hand, not enough people might want to get a vaccine. Making it easily accessible, informing them of its benefits, and, if all else fails, paying people to get vaccinated (or offering them some other perk - like free donuts, discounts, or even lottery tickets) might do the trick.
The supply side issue with vaccines is much more pervasive. Research is undersupplied, and patents are making the problem worse. A three pronged approach is in order: governments across the globe should fund and subsidize R+D in important areas, offer awards and commit to advanced purchases for various milestones (most notably completion), and make generous enough offers to buy out the patents for major, lifesaving vaccines.
Let’s focus on one interesting case: malaria. Vaccines for malaria actually exist, but they’re not particularly effective - one proposed vaccine, tried in Malawi in 2019, is only 40% effective, vs 95% efficacy for the ones used for, say, polio or mumps. But this year, groundbreaking news were announced: using a variant of mRNA technology (the same one Pfizer and Moderna used for their COVID vaccines), researchers at Yale were able to patent a very promising new malaria vaccine. Clinical trials began in February.
As usual for posts that deal with issues of global extreme poverty, I will urge readers to contribute to alleviate extreme global poverty by making small donations to effective organizations (like GiveWell or any of the charities endorsed by The Life You Can Save). As put forth by Peter Singer:
If it is the case that we ought to do things that, predictably, most of us won't do, then let's face that fact head-on. Then, if we value the life of a child more than going to fancy restaurants, the next time we dine out we will know that we could have done something better with our money. If that makes living a morally decent life extremely arduous, well, then that is the way things are. If we don't do it, then we should at least know that we are failing to live a morally decent life -- not because it is good to wallow in guilt but because knowing where we should be going is the first step toward heading in that direction.
Sources
Max Roser and Hannah Ritchie (2016) - "Burden of Disease", OurWorldInData
Max Roser and Hannah Ritchie (2019) - "Malaria", OurWorldInData
Max Roser and Hannah Ritchie (2018) - "HIV / AIDS", OurWorldInData
Understanding vaccines
WHO (2020), "How do vaccines work?"
CDC (2021), “Understanding mRNA COVID-19 Vaccines”
Miliken Institute Vaccine Tracker (2021)
Demand for vaccines
JPAL (2011), “Incentives for immunization”
The Nobel Prize in Economic Sciences (2019), “Understanding Development And Poverty Alleviation”
Supply of vaccines
Kremer (2002), “Pharmaceuticals and the Developing World”
Kremer (2004), “On How to Improve World Health”
Kremer (2000), “Creating Markets for New Vaccines Part I: Rationale”
Farlow (2004), “Over the rainbow: the pot of gold for neglected diseases”
Kremer & Glennerster, (2005), “Incentives for research on neglected disease”
Kremer (1998), “Patent Buyouts: A Mechanism for Encouraging Innovation”
Conclusion
Samantha Vanderslott, Bernadeta Dadonaite and Max Roser (2013) - "Vaccination", OurWorldInData
Sophie Ochmann (2018), “Can the world eradicate another disease?”, OurWorldInData
Singer (1999), “The Singer Solution To World Poverty”, the New York Times