Wednesday 8 October 2014

Public health and market failures

I'm not particularly worried about any large-scale Ebola outbreaks in the developed world. It's easier in the developed world to run effective quarantine and to track down those who might have been exposed to an infected and contagious person.

That certainly doesn't mean Ebola isn't worth worrying about. The burden of Ebola in Africa looks likely to be huge. Even where the total number of deaths remains very low relative to other African diseases like malaria, Ebola has an incomparable potential to disrupt economic activity. I'd put even odds that economic disruption caused by completely understandable fear around Ebola will kill as many people as the disease itself. When taxicabs become ways of catching disease and hospitals terrifying, and when a worker with flu symptoms could well be somebody who'll kill all your staff if he comes in to work and touches people - it's really not good.

Pandemic protection is a strong public good. Any dollar's investment in pandemic protection protects everybody who could be at risk: it's non-rivalrous. Every dollar spend improving hospitals in Africa, training doctors, providing medical equipment, and developing emergency response protocols is a dollar that protects each and every person in Africa, and by extension the rest of the world where there's risk of pandemics' spreading (think Bangladesh, Pakistan and India, not Europe and America). We all benefit from it. There is then a strong case for government funding of pandemic prevention efforts. There's also a really good case for things like Kickstarter, or PledgeMe, running campaigns on it to also help out.

The World Health Organisation is the outfit that's supposed to have an eye on the ball on this stuff. Unfortunately, in my view, the WHO has gone astray. Or at astray from what it should have been.

In the public health world, the social welfare function is either minimising the number of deaths, minimising incidence of morbidity, or maximising the number of disability-adjusted life-years or quality-adjusted life years. In economics, it's maximising some weighted utility function. The two yield very different outcomes and are what drive my dissatisfaction with the WHO's prioritisation.

Suppose we have several kinds of initiatives that would increase DALYs and reduce the overall burden of sickness and disease. Some of these, like polio eradication, provide health benefits without any particularly large cost as viewed by the person receiving the polio vaccine.* Nobody enjoys having polio. Nobody would voluntarily seek to have polio. If polio ceased to exist, the world would be a better place. Others of these, like regulations on soda, tobacco, alcohol, fatty foods, and salt, can increase DALYs and reduce the overall burden of sickness and disease too, but they come at a cost in terms of private experienced consumption benefits: many people like soda, smoking, drinking, and eating tasty things.

In the public health world, if a dollar's worth of effort yields greater expected DALY benefits in dietary and lifestyle regulation than it does in pandemic avoidance, that dollar should go to dietary and lifestyle regulation. In the economics world, we start by looking for market failures and target funding where markets won't do well on their own. And when it comes to lifestyle regulation, we need a rather harder look at things than just DALYs where choosing agents might well prefer to consume unhealthy things, at DALY cost, in exchange for current consumption benefits.

If we had to rank-order things by a very rough guess as to where the market failures are worst, I'd rank them as follows:
  1. Pandemic readiness and response, especially in developing countries;
  2. Antibiotic resistance and superbug mutation;
  3. Development of new antibiotics;
  4. Contagious diseases affecting developing countries where there's no particular profit in pharmaceutical investment: malaria and a big pile of others;
  5. Other contagious disease investigation and response;
  6. Vaccination promotion;
  7. Poor information about risks of smoking, diet and alcohol in the third world
The first one is at the top of my list not on expected value terms but as insurance. Like global warming, bad pandemic outcomes have the potential - low probability but potential - to destroy civilisation. It's then worth making investments in those kinds of areas not just on expected value calculations but also on an insurance argument - like Weitzman's argument on global warming. I also think governments should be investing more in asteroid detection and response, on exactly the same kind of argument.  

Pandemic readiness provides a strong nonrivalrous benefit to everyone. Third world countries generally do not have anyone with a sufficient encompassing interest in their prevention to ensure that the job gets done, so other mechanisms are needed, whether multigovernmental investment through the WHO or philanthropic investment by individuals. Note that there are exceptions, when you have enough Foreign Direct Investment. 

Antibiotic resistance flows directly from the negative externality imposed when someone over-uses antibiotics or uses them improperly. Development of new antibiotics has little profit potential where they're sensibly restricted to very hard cases and where governments and insurers would refuse to pay the per-dose amount that would then be needed to cover development costs. Similarly, there is little expected profit in developing cures for diseases that mostly affect very poor people and where patent protection is unlikely to be enforced. In all of these cases, prizes seem the most obvious solution for new drug development, where the funds would be provided both by governments and by philanthropists. 

Other contagious diseases and vaccination are lower on the list. Every one of us has an interest in protecting our own health. But because others bear direct contagion costs of our personal underinvestment in efforts against contagious disease, we will have less vaccination than would be optimal and less control against contagious disease. Subsidising the development and deployment of vaccines is then worthwhile. I can even see reasonable arguments for making vaccination mandatory.

Down at the bottom of the list is information-provision on lifestyle issues. There are plausible market failures around people being poorly informed about the health and addiction risks of tobacco and alcohol, and about the health risks of unbalanced diets. But a lot of observed behaviour would simply drop out of expected low life expectancies in developing countries and lower incomes, both of which increase the expected willingness to bear risk in exchange either for income or for pleasure. People doing unhealthy things isn't itself necessarily evidence of irrationality, or bad information. 

So that's how at least this economist would rank things. But if we ranked them by "where does a $1 do the most to improve DALYs", we could get an entirely different outcome. It would be like an electrical engineer ranking metals by conductivity and then proposing that we use silver for our main electrical transmission lines. Sure, it's more conductive, but it also costs more. Conductivity isn't the best outcome measure. Neither are DALYs where some of the potential policies impose costs on consumers that are not measured or weighted or considered by the public health side. 

In the public health world, tobacco is one of the very most important things to be worried about because it ranks very highly in the DALY-reduction rankings. Tobacco is not contagious, does not have existential risk to civilisation, and in some but not all cases results from informed decisions to enjoy nicotine at the expense of health and life-expectancy. The external costs imposed by a smoker are very low relative to the external costs imposed by somebody with contagious disease. Similarly, eating a diet with too much salt mainly hurts the person choosing to eat the diet with too much salt. There could be some case for providing information to remedy any market failure based on people not knowing the full health costs of eating too much salt, but that's about it. 

There are very real and substantial market failures in public health that are underweighted where public health advocates take a DALY-based approach rather than starting with market failures. If we look at priorities by budgetary allocations, it's great to see the focus on polio eradication, and the other work on tuberculosis, malaria, and AIDS; the first two have strong market failure arguments justifying them. AIDS receives a lot of private funding in the developed world, but there are reasonable arguments for providing funding there in the developing world. But increasing funding allocations to lifestyle issues and to tobacco control? It's harder to find a market failure case there.

I'm not building a straw-man here. They take "minimise burden of disease" as objective function, where every potential health cost matters equally, with no weighing of offsetting consumption benefits that come from alcohol or salty foods as compared to Ebola.

It worries me that journalists cannot find an after-hours WHO spokesperson on Ebola when a Spanish nurse contracts it in a Spanish hospital, that WHO officials cite being overstretched as reason for their not being able to respond as effectively to a pandemic, and that the WHO is at the very same time putting resources into e-cigarette regulation. There are opportunity costs to spending $122 million on the line item including alcohol, tobacco, drugs, and unhealthy diets.

This isn't just a "get off my lawn and leave me alone" thing. It's "why are you spending a hundred million bucks protecting people against themselves when you're under-resourced for pandemics?!?" thing.

Shouldn't we first target the health priorities that stem from market failures and impose existential risk?

Update: I've had request for clarification of what market failure means. My encyclopaedia entry is here. Further discussion here.

* Or at least there wouldn't be costs to getting the polio vaccine if the CIA hadn't managed to convince everybody that vaccinations are spy mission.


  1. Eric, when I share posts from your blog on facebook, they generally go in with a tagline that says "As I randomly delete anonymous comments, please use a handle.....". It'd be nicer if the post text came first...

  2. Darrochs comments appear slightly conflated with distance between him and the disease in question. Of course ebola is only killing a few people at the moment. AIDS only killed a few people to start with. Perhaps he wants to go to Africa and tell them they shouldnt be whining about AIDS - stopping smoking related deaths over 20-30years of consumption is more important than the wholesale destruction of a large segment of hte population?

    Maybe he can immortalise the WHO a little more? place them a litlte higher on the pedestal.

  3. If it's a philosophical disagreement between different intellectual traditions, how can it be resolved? Is public health doomed to perpetrate suboptimal policies, reduce efficiency and constrain people's lives forever?

  4. Umm... I have no clue how to fix that. Probably time I got rid of that preface to the comments anyway, now that I have disqus.

  5. Yeah, not sure if that's all of it. The question is why something that's in the comments shows as the first part of the post, instead of the post itself. But maybe if it wasn't there that wouldn't happen.

  6. This comment has been removed by a blog administrator.

  7. Crud. I deleted that text from the Blogger back-end. Don't know why it's doing that.

  8. So the economist looks for market failure, where markets don't work on their own.

    Please reconcile that with the fact that markets don't work on their own. Markets operate within bounds set by society. (Except maybe in the ideological neoclassical world).

  9. Government intervention should start in and focus on the places where markets will fail to work well, while weighing that policy often fails to work well too.

  10. "In all of these cases, prizes seem the most obvious solution for new
    drug development, where the funds would be provided both by governments
    and by philanthropists."

    The most obvious, perhaps, but instead of prizes, why not prediction markets? One where short positions aren't made valuable by paying out $1 per contract at an arbitrary deadline, but by paying out interest at market rates indefinitely, until a drug or vaccine with specified properties is invented and made available, patent-free. Governments and philanthropists can contribute to the cause by buying up shorts at prices the probabilities mightn't support, creating more long contracts, and thus more incentive for holders of long positions to work on making the necessary discovery.

    People holding long positions don't get anything until the drug is available, but when it is, they're all paid $1 per contract, regardless of whether their lab got the credit for the discovery. This means that everyone holding a long position has the incentive to cooperate on making the discovery as soon as possible, and you don't get the tournament games that can be created by prizes and patents, or the decades-long artificial monopolies inherent in the latter.

    The scheme can even support payments for incremental work towards the goal: Buy up a long position, do what work you can towards making the goal more likely, and then, when your own expertise is exhausted, close off your position for a profit.

    It's been called The Wall Street Performer Protocol, and I've written about it at

  11. That's a really neat idea. I need to read up on it. Thanks.