“Among the 23 states that report those details, Black and Latino people received far smaller shares of the vaccine than their share of cases and deaths, and compared to their share of the states’ populations.” (
Guilty to not backing up my claims, Don. I'll try to be a bit more clear at least about the assumptions I was making. Thanks for prompting me to clarify them a bit in my own muddled mind:1) We're trying to minimize quality life-years lost.2) Quality is based on health, nuanced by personal subjective value3) wrt this metric, the current U.S. prioritization scheme is much better than a random ordering wrt each of its main components (occupation & age)4) it could be even better if we add in a component which attempts to measure said personal subjective value via a proxy such as willingness to sacrifice something of visible value5) it could be even better than *that* if the 'something of visible value' could itself further the pandemic mgt cause6) prioritization schemes incorporating politically non-viable elements aren't worth considering. This is a severe constraint. E.g. straight cash payments to reserve a place in line fall into this category (unless conceivably the price were *very* high, hence my waffling at the end of my last post)7) contributing one's time *is* a politically viable proxy8) the fact that I buy 4)-7) hopefully clarifies a bit the rationale for proposing personal time spent on pandemic-mgt tasks as such a proxy.
I'm admittedly still guilty of not justifying these claims. The one which appears most worth scrutinizing to me is the 'age' part of item 3). I was quite skeptical about that until I read (from memory, or please correct me) that COVID mortality of 65-75 year olds is 90x that of early-20s, and that of 55-65 year olds is 30x. That plus considerations of political viability and the notion that public health experts do know something left me inclined not to second guess them on this. Which is admittedly more an admission that I was satisficing than a great argument. So, educate me:-)
I do not think this would be much of a problem. Surprisingly few non-blacks claim to be black when applying to college even though it would increase their chance of getting into their top choice school.
Age and occupation are not the "wins" you imagine. What are you actually trying to optimize here? Presumably, total number of deaths, or maybe life-years lost. Pandemic logic is not simple. There's a reasonable argument that first vaccinating "superspreaders" (perhaps 20-year-olds that party, or business travellers) would have cut R-eff enough, early enough, to be a net win to society (e.g. avoiding the huge Dec/Jan surge in cases). It is not at all obvious that, just because old people die from covid more often, old people should be vaccinated first (in order to reduce total deaths or life years lost).
You are saying that the market solution "discards valuable info", but you neglected to do the work to show that using that info as the government actually did, in fact did result in a superior social outcome. There's a reasonable argument that it made things worse.
That’s interesting. I think of spreadsheets (at least to the extent they resemble Excel, the only one I’ve used much) as one of our most useful tools in *dealing* with complexity.
I don’t like an exclusively market based approach as it discards valuable info (e.g. age, occupation) bearing on societal and individual health impacts. But the market can be a proxy for personal utility info that is otherwise hard to get, so it seems worth including as a component.
Details on the mkt component: A money-based market has big equity issues, some real and some just perceived. A time- based market (in which people can spend their personal time doing e.g. Mechanical Turk tasks if they want to give their current age/professional/whatever priority a shot in the arm) seems likelier to gain broad acceptance.
The buy-in and social benefit would be further boosted by making the eligible task set be ones which help the global pandemic effort.
Then again, an extra $100 billion would likely help the cause. Maybe it would make sense to include a modest financial-mkt component, if it can be done without sinking the whole operation.
Robin Hanson, 100% right again /sigh. It's so frustrating to see someone consistently have objectively better answers than current policy which I know will not be implemented. It wouldn't hurt so much if you weren't so consistently good at being right.
Any other time I'd say that the heuristic "The best is the enemy of the good" is why the post is wrong, but in this case the government has already put in place so many complicated rules about vaccine priority, that it has pushed down the usage of delivered vaccine to >50%. How could another set of prioritizing complications make it any worse?
If I ever go crazy you'll find me on the streets with a sign that says "We've blown our complexity budget". Probably yelling something like "Repent and cast down your nightmare rectangles!!! Death to spreadsheets!!!" at passers by.
I thought that it would be good to say black males at say age 62 die at the same rate as white males at age 65 and so now that they are vaccinating everyone over 65, they could have set it so blacks 62yo and older and whites 65yo and older can be vaccinated.
Guilty to not backing up my claims, Don. I'll try to be a bit more clear at least about the assumptions I was making. Thanks for prompting me to clarify them a bit in my own muddled mind:1) We're trying to minimize quality life-years lost.2) Quality is based on health, nuanced by personal subjective value3) wrt this metric, the current U.S. prioritization scheme is much better than a random ordering wrt each of its main components (occupation & age)4) it could be even better if we add in a component which attempts to measure said personal subjective value via a proxy such as willingness to sacrifice something of visible value5) it could be even better than *that* if the 'something of visible value' could itself further the pandemic mgt cause6) prioritization schemes incorporating politically non-viable elements aren't worth considering. This is a severe constraint. E.g. straight cash payments to reserve a place in line fall into this category (unless conceivably the price were *very* high, hence my waffling at the end of my last post)7) contributing one's time *is* a politically viable proxy8) the fact that I buy 4)-7) hopefully clarifies a bit the rationale for proposing personal time spent on pandemic-mgt tasks as such a proxy.
I'm admittedly still guilty of not justifying these claims. The one which appears most worth scrutinizing to me is the 'age' part of item 3). I was quite skeptical about that until I read (from memory, or please correct me) that COVID mortality of 65-75 year olds is 90x that of early-20s, and that of 55-65 year olds is 30x. That plus considerations of political viability and the notion that public health experts do know something left me inclined not to second guess them on this. Which is admittedly more an admission that I was satisficing than a great argument. So, educate me:-)
I do not think this would be much of a problem. Surprisingly few non-blacks claim to be black when applying to college even though it would increase their chance of getting into their top choice school.
Ask to see their ID? - the same ID that is presumanly used to check people off whatever list being used for any queue.
Do US driver's licences have race on them?.
Thank you, Robin, for saying important but politically incorrect things.
Age and occupation are not the "wins" you imagine. What are you actually trying to optimize here? Presumably, total number of deaths, or maybe life-years lost. Pandemic logic is not simple. There's a reasonable argument that first vaccinating "superspreaders" (perhaps 20-year-olds that party, or business travellers) would have cut R-eff enough, early enough, to be a net win to society (e.g. avoiding the huge Dec/Jan surge in cases). It is not at all obvious that, just because old people die from covid more often, old people should be vaccinated first (in order to reduce total deaths or life years lost).
You are saying that the market solution "discards valuable info", but you neglected to do the work to show that using that info as the government actually did, in fact did result in a superior social outcome. There's a reasonable argument that it made things worse.
How do you prevent people from saying they are 90 years old to jump the queue?
That’s interesting. I think of spreadsheets (at least to the extent they resemble Excel, the only one I’ve used much) as one of our most useful tools in *dealing* with complexity.
How do you prevent people from saying they're Black in order to jump the queue?
I don’t like an exclusively market based approach as it discards valuable info (e.g. age, occupation) bearing on societal and individual health impacts. But the market can be a proxy for personal utility info that is otherwise hard to get, so it seems worth including as a component.
Details on the mkt component: A money-based market has big equity issues, some real and some just perceived. A time- based market (in which people can spend their personal time doing e.g. Mechanical Turk tasks if they want to give their current age/professional/whatever priority a shot in the arm) seems likelier to gain broad acceptance.
The buy-in and social benefit would be further boosted by making the eligible task set be ones which help the global pandemic effort.
Then again, an extra $100 billion would likely help the cause. Maybe it would make sense to include a modest financial-mkt component, if it can be done without sinking the whole operation.
Robin Hanson, 100% right again /sigh. It's so frustrating to see someone consistently have objectively better answers than current policy which I know will not be implemented. It wouldn't hurt so much if you weren't so consistently good at being right.
Any other time I'd say that the heuristic "The best is the enemy of the good" is why the post is wrong, but in this case the government has already put in place so many complicated rules about vaccine priority, that it has pushed down the usage of delivered vaccine to >50%. How could another set of prioritizing complications make it any worse?
If I ever go crazy you'll find me on the streets with a sign that says "We've blown our complexity budget". Probably yelling something like "Repent and cast down your nightmare rectangles!!! Death to spreadsheets!!!" at passers by.
That seems a more complex approach, with more ways it can go wrong.
I thought that it would be good to say black males at say age 62 die at the same rate as white males at age 65 and so now that they are vaccinating everyone over 65, they could have set it so blacks 62yo and older and whites 65yo and older can be vaccinated.