It is said that the pharma industry spends more money on researching cures for baldness than it does for researching cures for malaria. This is, I think, not true. I say this with great sadness, because Bill Gates repeated the claim, and Bill Gates will, I hope, be remembered as one of the great heroes of the 21st century: someone who will have saved literally millions of lives, through his work on malaria and, lately, on funding Covid-19 vaccines. But even heroes can be wrong.
Still: even though Gates was wrong on the specific case, he is bang on when he suggests that there is a market failure. The pharma industry has done many wonderful things. To pick one example that is very salient to me, metastatic melanoma was essentially a death sentence 20 years ago; now I have two relatives who are alive and healthy, several years after diagnosis, because immunotherapy has made it manageable, if not curable.
Like what you’re reading? Get the free UnHerd daily email
Already registered? Sign in
But there are real problems with the pharma industry. While it is amazing at developing drugs for chronic conditions for rich people in rich countries, it is less effective at producing medicines — notably, but not limited to, vaccines and antibiotics — which are not profitable in the short term but which are socially beneficial in the long term. It’s also not well set up to produce cheap drugs for poorer countries, even if those drugs are relatively easy to make and distribute. People are dying of treatable diseases because it’s not profitable to cure them.
This is the backdrop to an article by Owen Jones in the Guardian last week, in which he pointed out that Pfizer will stand to make a profit of £9.8 billion from their Covid-19 vaccine. That is, he says, unethical. He quotes activists saying that Pfizer’s patent on its vaccine – and, I assume although he does not say, other firms’ patents on their vaccines – should be suspended, so that cheap, generic versions can be produced and rolled out to the developing world. As it is, biotech patents stand for 20 years; until then, firms essentially have monopolistic power, and can set prices how they like.
I have a lot of sympathy for Jones’s sentiment. He has identified a real problem. But his proposed solution is a bad one, and I worry that it would lead to disaster. You want pharma companies to distribute their products in as equitable a way as possible, but you also want them to continue to make those products. And taking away their profits when they make a good one will not encourage them to do the same next time. So what is the best course of action?
First, it’s worth noting that £9.8 billion is not in fact very much money. Obviously it’s quite a lot of money, on the human scale. But in terms of the global economy it’s peanuts. A paper by the economist Larry Summers estimates that even if we control the virus by the end of 2021, it will still cost the US economy alone about $16 trillion (£12 trillion), if you account for the deaths and misery as well as its simple economic costs. The IMF is more optimistic, putting it at a mere $9 trillion worldwide, although that doesn’t take into account the non-financial costs.
Take the latter, lower figure. If Pfizer making a profit reduces the impact of the virus by one-tenth of one percent, then it will pay for itself, even before we start thinking about the lives lost. I think it is a safe bet that the vaccine will have a much greater effect than that; Pfizer will only see a tiny fraction of the total benefits the vaccine will bring the world.
(It’s also worth noting that AstraZeneca have said that they won’t make a profit from their vaccine “during the pandemic”, although they reserve the right to say when “the pandemic” is over.)
But the main point is that there is a real, and to some extent unresolvable, tension in what it is that we want pharma companies to do. On the one hand, we want them to produce new drugs that save our lives. On the other, we want them to manufacture those drugs cheaply and get them to as many people as possible.
Imagine that a drug firm produces a vaccine for some deadly disease that is affecting people in both the developed and the developing world. Roughly speaking, a new vaccine costs about a billion dollars to get to market, after putting it through animal trials, Phase I safety trials and Phase II and Phase III efficacy trials, applying for licensing, and so on. Let’s imagine it costs about a dollar a dose to make. So if you only made enough of the vaccine to dose one person, the cost of that dose would be $1,000,000,001.
But if you make two doses, the average cost per dose is $500,000,001. If you make four, it’s $250,000,001. If you make a billion, it’s $2. The more doses you make and sell, the closer your average cost per dose gets to your marginal cost of making a single dose.
The socially optimal outcome is that we charge everyone the marginal cost, $1, so it can reach the maximum number of people. But if we do that, the pharma company will never cover its fixed costs: it will be $1 billion in the red, and probably won’t want to make any more vaccines.
On the other hand, if you let pharma companies charge the market value — that is, the amount that people are willing to pay — then it will cover their losses, because people in rich countries want to be vaccinated against this disease. But poorer countries will not be able to afford it. People who would be willing to pay the marginal cost are priced out of a vaccine.
The cost that will encourage industry investment R&D, and the cost that maximises the number of people covered by the vaccine, are different. It’s a simple tradeoff. “It’s obvious that, if your financial model is to sell to rich consumers at above cost price, you’re going to do R&D on products that you can do that for,” says Owen Barder, a development economist who has worked on pharma incentives. “The patent model suffers a really big problem.”
One obvious solution would be to take the whole business out of pharma companies’ hands, and let governments and philanthropic organisations run the whole thing. But the profit motive is very good at one particular thing: making pharma companies drop a product that isn’t working. “I think the key skill of a pharma company is knowing when to exit,” says Barder. “And, to a gross generalisation, governments are rubbish at exit.”
Most candidate drugs don’t work, or don’t work very well. So most of the time you have to abandon research, after spending a lot of money on them. But it is very hard for a government to say “This thing that we’ve spent £100 million on doesn’t work, and we’re going to throw all that money away.” It’s easier to throw another £100 million in and avoid the question. Drug companies, we all agree, are extremely good at making money. And part of being good at making money is not throwing good money after bad. Publicly funded research is great for blue-skies, open-ended questions that help us understand things, but it is less effective than private firms at developing specific products, because of this exit problem.
Instead, says Barder, what you need to do is separate the fixed cost from the marginal cost. You need to find some way of encouraging pharma companies to sell their products at a low cost to poor countries, while also incentivising them to keep putting money into R&D.
There are several ways of doing this. One example is a financial prize: we (governments, philanthropic agencies, etc) will give whoever comes up with the first vaccine some large amount of money: perhaps $1 billion, on the condition that they then agree to make it in large quantities and sell it at close to marginal cost to the developing world.
That’s a useful model, but it has a disadvantage. In the last two weeks, trial results for two Covid vaccines have been announced: the Pfizer one and, this week, the Moderna one. Initial signs suggest that Moderna’s one is a bit better than Pfizer’s: it appears slightly more effective, and it can be stored at normal freezer temperatures for months, while Pfizer’s needs to be kept at -80°C.
If we had put in place a prize system, so that the first group to develop a working vaccine got a billion dollars, then (assuming Pfizer’s vaccine gets licensed first) Pfizer would get the whole lot, even though it appears Moderna’s product is slightly superior. Depending on what costs remained, it might well make financial sense for Moderna to abandon the project, because they would not get the reimbursement that would make it viable. But, obviously, if a superior product comes out a week later, you still want to be able to buy it. (And you may want lots of different vaccines anyway, because each will have different advantages and disadvantages.)
But there is a better way. The pneumococcus bacterium causes meningitis, pneumonia and septicaemia; this Lancet paper estimates that it killed about 800,000 children in the year 2000 alone. The pharmaceutical industry — Pfizer, as it happens — developed a vaccine, Prevnar. But it was only suitable for the strains of the disease found in Western countries, not in the developing world. Creating a version that worked for the strain found in poorer countries was perfectly possible, but the financial incentives just weren’t there.
To solve the problem, in 2004, the economist Michael Kremer — he of deworming fame — and the British economist Rachel Glennerster proposed something called an “advanced market commitment”, or AMC. Five national governments and the Gates Foundation put aside $1.5 billion dollars. They promised the pharma industry that, if they developed a new vaccine for the developing world, then for every dose that the companies provided at the marginal cost of $3.50 a dose, they would reimburse them a further $7.
Unlike a prize, an AMC rewards both speed of development and quality of product. If your vaccine is first to market, you can sell it and make money without any competition; but if you know your vaccine is better than the one that’s first to market, you can carry on developing it without worrying that it will be rejected because all the money has gone to the earlier, but inferior, product.
Better yet, it keeps all the risk on pharma companies. If they know a vaccine isn’t likely to work, then they will abandon it, because it won’t make them money — it won’t meet the requirements. If governments or philanthropic agencies were funding the research directly, they’d have the problem of exiting after spending lots of money on it. The funding bodies never have to pay for a product that doesn’t work (although they do have to pay a premium for the ones that do).
With pneumococcus, it was a roaring success. By 2016 the vaccine was being distributed in 60 countries, and more than 150 million children had been vaccinated; Gavi, the body that oversaw the AMC, estimates that it will have saved 700,000 lives by the end of the year.
To be clear, the AMC model won’t work for everything. In the case of Covid, for instance, the problem wasn’t that companies weren’t investing in R&D — they knew that there was a market. But building a vaccine factory takes months and millions of dollars, and pharma companies, understandably, don’t want to invest millions of dollars in a product that very possibly won’t work; they’d rather build it at small scale until they’ve done the testing, then scale it up afterwards.
Most of the time that works fine. But in the Covid case, the costs to society of making a few factories that ended up not being used were negligible compared to the cost of waiting even a few more weeks for a vaccine. So it became worthwhile for the UK government (for instance) to start paying for manufacture of the Oxford-AstraZeneca vaccine even before it had released its results, and why Bill Gates (again: one of history’s greatest heroes) invested billions of dollars building seven factories for different vaccine types, knowing that some, or even most, would not work.
And AMCs wouldn’t work for antibiotics. Antibiotics are a special case: we need lots of new antibiotics so that we can combat antibiotic-resistant microbes, such as MRSA. But we don’t really want people to use them, because then microbes will become resistant to the new ones, as well. Most of the time, if a pharma company brings a new drug to market, it will win a significant chunk of that market — if you make a new cancer drug or antidepressant that works better on some group than the existing ones, then it will be prescribed for that group. But with antibiotics, most of the time, doctors will continue to prescribe doxycycline or methicillin, and will save your new drug for when they encounter drug-resistant bacteria.
So there is very little incentive for pharma companies to invest in antibiotics, because — essentially — your new antibiotic will rarely be prescribed, so it will rarely be sold, so you’ll make very little money. According to Jeremy Knox, policy lead in antibiotic resistance at the Wellcome Trust, there’s an additional problem that, for hard-to-explain reasons, antibiotics command low prices anyway. “So you have this toxic combination of very unpredictable small volume markets and very low prices,” he says: “and both of those things combined means that if you’re a producer you’ll probably forecast making a loss.” An AMC that paid them for each dose sold wouldn’t help, because not many doses will be sold even at low cost.
There are models in place that get around this. Jim O’Neill, who was the antibiotics tsar under David Cameron’s government, suggested a lump sum given to each pharma company that brings a novel antibiotic to market. Wellcome and the British Government are also looking at what Knox describes as “the Netflix model”, where bodies like the NHS agree to pay a fixed sum a year to the manufacturers, and are then given as much as they need. So if you only need to use five doses all year, you pay X million; but if you need 5,000 doses, or 5 million, you still pay X million.
It means that pharma companies aren’t incentivised to sell as many doses as they can, and the decision to prescribe can stay in the hands of doctors. (It is worth noting that the UK Government, and the health secretary Matt Hancock in particular, have been quick to take up this model; I rarely have nice things to say about Matt Hancock, so I will flag this now.)
There is also a funding scheme, CARB-X, which supports the R&D of new antibiotics. It’s not enough — Knox thinks we need about 10 new antibiotics every decade, and that at the moment we might be lucky to see two or three. But it’s an improvement.
Coming back to Owen Jones’s article, it’s worth noting what I think is a factual error. It’s a minor one, but I think it’s instructive. He says that vaccine companies “abandoned” research into a vaccine for Sars in 2003 because it was not “immediately profitable”. But as I understand it, the reason that the pharma industry stopped work on a Sars vaccine was because Sars (somewhat mysteriously) disappeared. If people weren’t getting infected, then you couldn’t tell if your vaccine stopped infections. In fact, Mike Osterholm, the US epidemiologist, says that the pharma industry invested hundreds of millions of dollars in Sars vaccine research, and then when it went away, the US government and philanthropic agencies left them “holding the bag”, and wary about future investments.
I think it’s instructive because most of the scientists working at pharma companies want to save and improve lives, just as scientists do at universities or other bodies. But just like the scientists at those other bodies, they have to navigate the incentive structure that surrounds them. For university scientists, it’s publish-or-perish – your work is useless if it doesn’t get you journal articles and citations, regardless of how true it is. For industry scientists, it’s profit.
The trick in both cases is to help the incentives align with what we want out of the research they do. In the case of the pharma industry, we want them to keep producing new drugs and vaccines, but also to provide them at a low cost to the people who need them. If we remove the pharma companies’ patents, and stop them making profits, then we might be able to provide this drug at a low cost, but we’ll be less likely to see the next one. Instead, we need to come up with clever ways of incentivising them to do what we want, so that we continue to spend more on curing malaria than we do on curing baldness.