Wednesday, November 25, 2020

Vaccines and externalities

 A lovely point from the always creative Tyler Cowen

Say, for the purposes of argument, that you had 20,000 vaccine doses to distribute. There are about 20,000 cities and towns in America. Would you send one dose to each location? That might sound fair, but such a distribution would limit the overall effect. Many of those 20,000 recipients would be safer, but your plan would not meaningfully reduce community transmission in any of those places, nor would it allow any public events to restart or schools to reopen.

Alternatively, say you chose one town or well-defined area and distributed all 20,000 doses there. Not only would you protect 20,000 people with the vaccine, but the surrounding area would be much safer, too. Children could go to school, for instance, knowing that most of the other people in the building had been vaccinated. Shopping and dining would boom as well.

All along our authorities have had trouble distinguishing public health from the treatment of individual patients. That's why testing has been such an unfulfilled promise.  "Who gets it first" is treated like who should the government send money to. The point of vaccine is not mainly to protect individuals, it is to stop the spread of a disease.  


  1. Not a comment, but please, please, please do an interview with Glenn Loury. He is a fabulous interviewer. He frames the opposing view of anything very well to allow moving ideas forward. I think you two would have a great conversation on anything, but wokeness or systemic racism would be great topics. Please. Thank you. Kurt.

  2. John, really, this is America. Of course the point of a vaccine is to protect the whole population by restricting the spread of a disease. But that is "socialism". And socialism does not go over well in America.

  3. Surely there is a game theory problem to be resolved here, or optimization model. Biden is left holding a very rapidly decaying bag, as the fanfare about a vaccine has excited the public imagination, he will be blamed for any continuing spread of the virus. It is one important reason he defeated Trump.

    The primary problem with "central planning" or any form of the "leadership" model for economic/human progress is that anyone can BLAME a visible authority for disliked outcomes from a leader, but nobody blames "God" as was normal before socialism became a fad.

  4. I think this is why the government is buying vaccines is lots of 50 million doses. 20,000 doses would be a problem in a country of 330,000,000 people. Fortunately the Federal government is aware of the size of the country is acting appropriately.

  5. Define "much safer". Specify the size of "the surrounding area".

    Cowen's gedankinexperiment fails the Sink Test. In metallurgical engineering practice, the Sink Test is used to determine the property of a specimen of steel taken from a random lot of deformed bars ("rebar", in the colloquial). In preparation for conducting the test, the engineering takes a calibrated galvanized steel bucket, fills said bucket to the calibrated depth with still water, and places said filled bucket on a level flat surface. Taking the sample of steel, the engineer carefully positions said steel bar exactly 0.200 inches above the surface of said still water in said galvanized steel bucket and then releases said steel specimen and observes the reaction of said steel specimen. If the specimen sinks to the bottom of said bucket, the engineer shall write "Pass" in the test log book; if the said test specimen floats partially immersed or fully immersed but does not settle on the bottom of said calibrated test bucket, the engineer must write "Fail" in the test log book.

  6. "The point of vaccine is not mainly to protect individuals, it is to stop the spread of a disease."

    That's overstating things. While vaccines do produce a positive externality of stopping the spread of disease, the *main* function is to protect the vaccinated individual. First, some vaccines are non-sterilizing, i.e., they prevent symptoms in the vaccinated individual but don't necessarily make the individual non-infectious. Second, I believe the measure of efficacy in the current Covid vaccine trials is whether they prevent *symptomatic* Covid-19. (That doesn't mean that some or many vaccine recipients won't also be non-infectious, but the main goal is to prevent symptoms or severe symptoms in the vaccinated individual.) Has there ever been a vaccine that was targeted towards preventing infectiousness of the vaccinated individual without preventing symptoms?

    I can't think of any medical treatments where the *main* benefit is to everyone else except the person receiving the treatment. If that were the case with vaccines, we would have to do a lot more than provide them at discounted or no cost to get people to take them. We would have to pay them. Also, it would raise serious ethical issues if doctors were advising their patients to take a vaccine not out of concern for their patients' health but out of concern for everyone else. A doctor has the same fiduciary duties to his patient that a lawyer has to his client and, really, advisors of all types have to their advisees.

    Now, government policy on *distributing and subsidizing* vaccines might prioritize the positive externalities in slowing spread over which individuals "need" protection the most. But, that doesn't change the fact that slowing spread is the external, not the main, benefit.

  7. "The point of vaccine is not mainly to protect individuals, it is to stop the spread of a disease". That is NOT the "main" point of a vaccine. Vaccines are tested for effectiveness (which is defined as protection from a specified disease in individuals). To reduce transmission of illness is a secondary (and likely) outcome, but it is not the main objective of a vaccine. You should watch: where Moderna's Chief Medical Officer clearly explains what a vaccine does (and is tested for) and what it may do (but there is no proof of it, initially).

  8. Cowen's article does not answer this question: what method of distribution minimizes the number of hospitalizations and deaths? Taking a pessimist viewpoint, I see the broad countryside dieing off and the islands of vaccinated people re-populating the nation.

    1. The ~0.5% (quite possibly less) fatalities among people infected with SARS-CoV-2 are not going to constitute anything like a "die-off" of the broad countryside. And, of course, the age-stratification of that unfortunate 0.5% is weighted massively toward people who are past their child-bearing years.

      What we're realistically talking about even with hypothetical "unconstrained spread" is a large number of individual tragedies as the overall age-adjusted mortality rate increases by probably about 10% to 20% for a period of 1-2 years. Phrases like "die off" are simply absurd in this context.


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.