Covid 12/10: Vaccine Approval Day in America

Today, the FDA is meeting to discuss Pfizer’s Covid-19 vaccine. By the time many of you read this, they will have hopefully given the vaccine, and perhaps Moderna’s as well, emergency use authorization. If that happens, distributions to the states can begin, and some people will be vaccinated as early as tomorrow.

That is excellent news. Alas, supplies remain highly limited and the vaccine takes a month to work. States are getting smaller allocations of the vaccine than they expected. The situation for the holiday season remains unchanged, and most of us will likely not become immune until some time around May. 

Two weeks ago, I noted data was about to get weird. A week ago I observed that the data was weird. What I didn’t note was that the Covid Tracking Project had written a post about exactly in what ways they expected it to be weird, which were different in places from the ways I expected. Lots of good observations in there. The post makes a convincing case that I was underestimating expected delays and bottlenecks.

Despite that, we saw a big jump in positive test percentage on Tuesday, December 2, exactly when I’d expected one. That could mostly be a coincidence. Perhaps that was more of a shift in timing of reporting, with the real surge from Thanksgiving coming later. I certainly buy the augment that such super-spreader events infect people in multiple waves, as those infected take the virus home, and we should expect an unusually high effective R0 in the week after Thanksgiving. 

That will not give us much time to recover before Christmas and New Year’s Eve give their one-two punch. It seems unlikely we will turn the corner until after those get to play out. Despite previous holidays not showing up much in the data, these big three seem like a huge deal to me, especially now. People are sick of it all.

People are also sick, period. Death rates are up. Positive test rates are up. Cases are up. 

This week’s post attempts to limit Twitter links. In cases where the link is serving as a source but is otherwise inessential, I’m going to link it explicitly and purely as a source, so there’s no need to click unless you want to verify it. Any other Twitter links didn’t have a good substitute I could find easily.

Let’s run the numbers.

The Numbers


Last week’s prediction: My unconfident prediction is a 14.8% positive rate on 10 million tests, and an average of 2,200 deaths to account for at least some amount of catching up in reporting. But, again, wider than usual error bars all around. 

Results: We got a 13.7% positive rate on 10.4 million tests, and an average of 2,276 deaths. So somewhat better positive rate than I expected, but not as good as one might hope.

Next week: My baseline prediction is a 14.3% positive rate on 11 million tests, and an average of 2,550 deaths per day. 


Oct 8-Oct 1478212172366436
Oct 15-Oct 2180415912370523
Oct 22-Oct 2889517012208612
Oct 29-Nov 495619772309613
Nov 5-Nov 11108927122535870
Nov 12-Nov 181255293428181127
Nov 19-Nov 251761416933961714
Nov 26-Dec 21628381427421939
Dec 3-Dec 92437550842862744

If anything, deaths continue to lag infections, likely due to holiday effects, so this is likely an underestimate of how bad things are out there. Even at face value, things are clearly quite bad. The Midwest number is higher than any previous peak, the South is not far behind and the total death count is at an all time high, with numbers going up dramatically everywhere. 

Positive Tests

Oct 8-Oct 146828412574411799538918
Oct 15-Oct 217557114985113323843325
Oct 22-Oct 289498318188115812357420
Oct 29-Nov 411268425291716709870166
Nov 5-Nov 11157495387071206380108581
Nov 12-Nov 18211222452265255637150724
Nov 19-Nov 25269230435688294230170595
Nov 26-Dec 2256629357102294734185087
Dec 3-Dec 9354397379823368596263886

Outside of the Midwest, it seems clear things continue to rapidly get worse, and last week was a data hiccup as expected. In the Midwest, it looks plausible from this graph that things have peaked, but positive test percentages will be the judge of that.

Test Counts

DateUSA testsPositive %NY testsPositive %Cumulative Positives
Oct 1-Oct 76,009,9235.2%763,9351.3%2.28%
Oct 8-Oct 146,322,3265.7%850,2231.1%2.39%
Oct 15-Oct 216,440,1506.5%865,8901.2%2.52%
Oct 22-Oct 286,933,1557.5%890,1851.4%2.67%
Oct 29-Nov 47,246,0688.6%973,7771.6%2.86%
Nov 5-Nov 118,285,59810.6%1,059,5592.4%3.13%
Nov 12-Nov 188,917,70112.4%1,155,6702.9%3.47%
Nov 19-Nov 2510,430,05511.6%1,373,7512.9%3.83%
Nov 26-Dec 29,788,44611.5%1,287,0104.0%4.18%
Dec 3-Dec 910,458,81313.7%1,411,1424.9%4.61%

Test counts have recovered back to their peak and should presumably resume going up now. That positive rate is one ugly number, and it looks like New York is rapidly ceasing to be that much safer than the rest of the country, although it does run way more tests than most places. 

The cumulative positives number is getting quite substantial, especially given how many infections are being missed, and that those infected are not a random sample. Without immunity from those already infected, we would likely be seeing growth now like we saw back in March.

Positive Test Percentages

10/8 to 10/142.57%8.14%7.09%4.75%
10/15 to 10/222.95%8.70%7.85%5.36%
10/22 to 10/283.68%9.87%8.58%6.46%
10/29 to 11/44.28%12.79%8.86%7.04%
11/5 to 11/115.56%17.51%9.89%8.31%
11/12 to 11/186.99%18.90%11.64%10.66%
11/19 to 11/257.00%16.62%10.41%11.75%
11/26 to 12/28.38%17.90%12.45%12.79%
12/3 to 12/910.47%17.94%13.70%12.76%

I am ready to believe the story that the Midwest is more or less peaking now. Their test percentages have been in a narrow range for the last five weeks, and their immunity effects are rapidly increasing. The West’s number is encouraging, but it is on vastly more tests. That many more tests should drive down percentages enough to account for this, or alternatively disproportionally many negative tests got reported late and are warping the curve there a bit. Optimism seems premature. 

In the Northeast and South, trends do not look good, and the Northeast especially looks quite bad, as if a combination of holidays and colder weather has broken the control system, and things are likely to get substantially worse before they get better. Time to bunker down.

My best guess is that the likely overall true peak is after the effects of Christmas and New Year’s, so the measured peak of infections should be some time in mid-January, with deaths closer to the end of January. Early vaccinations will assist a little with this, but not much, as they are going to the vulnerable and deserving rather than to those most likely to spread the disease, and there are not that many doses. 

Note that I continue to not list hospitalizations in these charts. They continue to rise, but that seems to me to still say more about how hospitals are handling things than about how many are infected or sick. Even if things were improving and there were plenty of beds, hospitalizations would still be rising now, because there would be more people getting sick now than are ready to leave the hospital, as discharges are going to lag cases by quite a bit. So while it’s certainly something I look at as a sanity check, I don’t know what it can teach us.

Covid Machine Learning Projections

For those interested, here’s a closer look at the Covid-19 Projections website’s nowcast model for true prevalence. Layman summary by him on Twitter here. Can be compared to his earlier write-up here. Seems intuitively like reasonable approximations. What it doesn’t look like is the output of a machine learning algorithm, which is interesting. It looks a lot like something I would create if I was trying to plausibly fit a bunch of curves and generate reasonable claims in a smooth consistent manner, except that this uses a suspiciously large number of round numbers. I have some experience with this sort of thing. 

The machine learning project doesn’t know about Thanksgiving and its warping effects on data, so it still thinks there was a brief downward blip in infections a few weeks ago, although it now recognizes that things resumed getting worse shortly thereafter. If you smooth that out, answers here continue to seem sane and they see us as having had 630k infections per day two weeks ago, with 2.6% infected at that time and 17.3% of the population having been infected at some point, or over 20% by today. I continue to think that answer is on the very low end of plausible.


Germany continues to be stable. Many others have seen dramatic declines from their peaks, and the United States once again has a far worse Covid problem than Europe. 

My news sources have not seen much happening in Europe this week, and the charts seem straightforward. Countermeasures have worked in at least containing things and preventing them from getting worse, but no one is aiming at suppression, instead they are waiting for the vaccine. 

No, Not That Pandemic Vaccine, That Other Pandemic Vaccine

Team behind Oxford Covid jab start final stage of malaria vaccine trials. This is excellent news. Also a reminder that the world is smaller than it appears, as the same team behind one of three Covid-19 vaccines also produced this one. Given how most of the time and effort involved in vaccines seems to involve testing them and overcoming regulatory barriers, it should not be surprising that teams that make one vaccine can also make more. 

Malaria is kind of a big deal, killing over 400,000 people a year. Bed nets are one of the prototypical Effective Altruist causes. Whether or not the best strategy would be to instead eradicate malaria in Africa the way it was eradicated in America, that is not happening at the moment. Barring that, a cheap and widely available vaccine seems like an excellent way to go.

“A lot more people will die in Africa this year from malaria than will die from Covid. I don’t mean twice as many, probably 10 times,” Hill said. The vaccine “is going to be available in very large amounts, it works pretty well. And it’s going to be very low priced.”

The vaccine could be in use by 2024 if the final human trials are successful, he said.

So we are now entering the final stage of trials, which means we are only three years away from starting vaccinations. Some time next year, 4,800 children will be vaccinated, then we will wait a very long time, likely in part because we only vaccinated 4,800 children. This is what people mean when they say what a great achievement the Covid-19 vaccines have been in terms of speed. Here we have a vaccine that likely could save over a million people if we sped up its release and distribution by three years, and we are not doing that because of reasons. 

All of you Effective Altruists who were already working hard on malaria, is there any way to push on this and make it go faster? Perhaps you can arrange for a larger trial that will gather more data faster? Perhaps you can fund scaling up production or distribution faster in advance? What about Bill Gates, who is very involved with malaria and funded pre-production of Covid-19 vaccines? By the end of 2021, we should have sufficient doses of Covid-19 vaccines, which leaves years to handle production on this one even if they are in conflict. It does not appear this is getting the level of urgency it deserves.

All I Want For Christmas Are a Covid Vaccine and a PS5, But They Underpriced Them And Now They’re All Sold Out 

What does everyone else want for Christmas?

A PS5, yes. And thanks to my good friend Seth Burn, I got one! Section title will have to change.

For a slim majority, also a Covid vaccine. 

There was a Pew Research survey this week on who intends to get the vaccine. 

For those who want to guess who wants it more before looking, the categories are Male, Female, Postgrad, College, HS or less, Upper income, Middle income, Lower income, Over 65, Under 30, White, Black, Hispanic and Asian. 

Answers are at the end of the section via screenshot (or Twitter source here).

To help convince the hesitant ones to change their minds, at least some skin in the game is officially here: Former Presidents Obama, Bush and Clinton volunteer to get coronavirus vaccine publicly to prove it’s safe. I would assume Biden as well. Trump already had Covid-19, but if the goal is to show it is safe rather than effective, it would be great if he joined them anyway. I haven’t seen this actually suggested anywhere? Will be a while before it matters much, given the severe shortage of doses.

You know who should likely get the Covid vaccine early? Airline pilots. 

So naturally, the FAA is currently telling them that if they get their vaccine, they will lose their medicals, and not be allowed to fly. They are monitoring the FDA’s decisions closely, but pilots are not allowed to get vaccinations that don’t have full approval. I would never have predicted this particular plot point, but on a meta-level I find it entirely unsurprising. It makes sense that there would be government regulations that ban the vaccine from getting to some of the people who need it most. That the best we can hope for is that some such authorities try to kill us less aggressively some of the time.  

Scream If You Wanna Go Faster

The vaccine even works faster than we thought! From the FDA report: Looks like efficacy even for one dose might be 88.9%. Here’s the graph.

That doesn’t mean we can do only one dose per person for a while until we have enough doses, because we don’t know the consequences of a longer wait or whether the immunity from only one dose would last, but it’s still great news. It also puts people in a profoundly weird spot, since they’re mostly immune now, but will be much more immune soon. How safe do you play it during that period?

Business insider asks why the United States is behind the United Kingdom in approving the vaccine. Here is the functional explanation they give:

The regulatory process involves experts reviewing “thousands of pages” of technical information, an FDA representative told Axios on Tuesday.

The agency also has to evaluate the manufacturing process, check statistical analyses, and look at the effect of the vaccine on groups of people with high risk of side effects, the representative said.

The FDA and the Medicines and Healthcare products Regulatory Agency, the UK regulator, have different processes to review vaccines, Dr. Penny Ward, a visiting professor in pharmaceutical medicine at King’s College London, told The New York Times.

The FDA asks drugmakers for raw data, which it reanalyzes, while British regulators rely more heavily on reports produced by the drugmakers.

An alternative hypothesis is that the FDA is delaying the process intentionally. With the cover story that this is being done ‘in order to be careful’ so as to prevent fear that the vaccine was rushed. I strongly agree that there are approximately zero people who are going to look at the gap between application for approval and when the FDA approves the application, and use that as a true objection that justifies not taking the vaccine. To the extent that this group has non-zero members, the group Tyler points out, that takes the delay as worrisome and uses that as its true objection, is not obviously smaller. 

The simpler explanation is that there was already a meeting scheduled for December 8-10, people had holiday plans, and the FDA simply did not care enough about doing this faster to move the meeting.

In other Tyler modeling news, while the cause of faster vaccine approvals is clearly righteous, I do not think we can model a day’s delay in approval as delaying vaccine production by one day. The story flat out makes no sense to me. Pfizer is not going to produce vaccine slower, at any point, because the FDA gives formal approval on December 10 instead of November 30. The vaccine is already approved in the UK, so there is a guaranteed place to sell more doses than they can possibly produce this month, in the worst case scenario. The probability of approval by the FDA isn’t 100%, but given their knowledge of the data, it’s close. Every sign points to the approval being a formality. So I hold strongly that the damage done is that current doses are shipped with a delay, not that all doses are delayed. 

Still, that’s pretty bad. It looks plausibly like a third of all deaths are of nursing home residents that could be reached in the first wave. So they all are getting vaccinated two weeks late, which means the cost here is something like 1,000 additional deaths per day of delay from that alone. Then add in the healthcare workers, and their new ability to actually do their jobs fully and also avoid turning hospitals into outbreaks. So yeah, it’s a lot of blood, but giving them every death between application and approval isn’t fair.

After I wrote that section, Tyler wrote a post even more explicitly disagreeing with this, essentially claiming that yes approval would accelerate other tasks substantially, which I continue to mostly not buy. I think his argument holds when approval is truly uncertain, but not when it is a formality. I do understand and share the frustration with those who use constraint X to justify not working to remove constraint Y, which we don’t work on because constraint Z, which we don’t work on because constraint X. Yes, we should be working on the other constraints. But that doesn’t change the answer to the question.  

Switzerland is going even slower, making the usual noises about the need for ‘caution.’ My read is that this is because they did not order vaccine doses early enough, and now they are all sold out, so why not spin that by not approving the vaccine for a while and calling it ‘caution?’ 

As we know from last week, the Europeans are taking it relatively slow, despite knowing the answer already:

The impression it gives to me is not that officials were taken by surprise. It is that they had a choice to make, and made it.

Cell asks how we might speed up future vaccine development in case of a new pandemic. The strategy is based on it not being that hard to predict where future pandemics come from. If most are going to be variations on a few existing virus families, we can do a lot of the work in advance. They suggest that for only a few billion dollars, we can do advance Phase I and Phase II trials for vaccines designed for each virus family that could plausibly cause a future pandemic, stockpile doses, and be ready to proceed to Phase III right away. 

If we had the vaccine in question ready to go on day one, then the phase III could begin the equivalent of February, then all we’d have to do is wait for enough infections. 

What they do not say is that we could do much better than this by using human challenge trials. If we were willing to do such trials, and had already gathered safety data, there is no reason we could not know the vaccine’s effectiveness and safety within a few weeks of identifying the new pathogen. Remember, Moderna created their vaccine in two days. Then all that is left is to scale production. 

Cell does not suggest this, instead taking the need for a regular Phase III as a fixed law of nature that cannot be changed even when preparing in advance for a future threat. 

And that’s valuable! I am glad that people are taking seriously the question of “if the people and procedures in charge of slowing down and stopping development must be accommodated, how can we do that as quickly as possible?” Very good question. I only wish we could also fully ask an even better one.

The vaccine is the ultimate case of More Dakka. However much we are spending on it, it won’t be enough. We should spend more. That was true before. It’s still true now. The case for going big remains overwhelming. This whole time, we could have built far more capacity in exchange for (in context) trivial amounts of money. Pfizer offered us 200 million doses of vaccine in July, but we balked at the price tag, and only ordered 100 million doses. Then this week the Trump administration tried to buy more, but Pfizer informed us that other countries bought those other doses instead, and they won’t have capacity for more until next June. Whoops. Then Trump issued some sort of executive order demanding that the doses go to us instead, because America, to which the chief scientist of Operation Warp Speed has said, “I literally don’t know” what it means. And in an evergreen statement, “The White House is doing what the White House is doing.” 

I don’t think this is going to slow us down much once Moderna is added to the picture, but if this mistake costs us all an extra month or two of our lives, it would be one of the worst times someone has walked away from an obviously great deal. If I was told my available dose was in June rather than April, I don’t think I personally take that much extra risk given how perilous a state the hospital situation might be in for the next few months, but it definitely feels like a plausible last straw for many. In particular, being given one timeline, then having a delay for such a stupid reason, seems very much a ‘screw it I’m done with this’ moment. Again, I’m not doing that if this happens. But if others do that, I totally get it.

The Money Illusion has thoughts, speculating that bureaucracies are simply incapable of adjusting on the fly when the things that change matter. In Sumner’s proposed model, what you enter the crisis with is mostly what you are stuck with unless you force their hand, and we forced it but not enough.   

The Grumpy Economist wants to go fastest of all. He suggests the obviously correct answer, which is to allow free market vaccines and to distribute them by price, starting in February for those willing to buy it under those conditions. He claims this would have prevented the pandemic, which is going too far, since scaling up production and convincing enough people to take it would have taken months. 

If we wanted to prevent the pandemic in the first place, that requires going much further. We would have needed to allow free market testing.

Let us be clear. Government bans on private action are entirely responsible for the entire pandemic in the West. Period.

Meanwhile in India, they rejected the emergency use authorization for AstraZeneca’s vaccine. Which seems entirely unsurprising, given that there are all sorts of problems with the studies so far, which is why they didn’t even attempt such a thing in the West. But of course, it’s also, from an expected value position, completely insane. The vaccine is super cheap and the safety data seems ironclad. So what if we don’t know how effective it is? It’s weird to pick on India here. If anything, they deserve praise for making it worthwhile to ask the question at all. They might have gotten the right answer, even though they didn’t.

For Me But Not For Thee

Austin Mayor Steve Adler (source):

Late Show Host Steven Colbert (source):

The Ramsey Tax Is Too Damn High

I think this calculation by Tyler Cowen, viewing Covid-19 as a “Ramsey tax problem,” is not only importantly wrong but exactly backwards and rather perverse, in ways worth exploring. 

The whole point of saying we are overreacting, to say that the costs of Covid-19 are not that high, is exactly to argue that the costs borne have been extraordinarily high. The claim is that the costs from infection are low, and that we are making a mistake to bear such high costs, such huge deadweight losses. There hasn’t been a huge non-health loss as a law of nature. There’s been a huge non-health loss because we made choices. Why must we accept those choices as given, and thus assume that the non-health costs must be so high relative to counterfactual health costs, if that is the current situation? And why are such people being accused of making moral claims when they are making efficiency arguments? 

In particular, why is “talking people out of their high elasticity” equivalent to the statement “Don’t leave NYC just because the taxes are going up!  It will wreck the city.”? Tyler’s sentence is calling for sacrifice by the individual on behalf of the group. Whereas talking people out of their high elasticity is more like “you would be better off living your life mostly as you lived it before, and accepting that you are likely to get Covid-19 at some point, because life is short.” If anything it’s the opposite of the group sacrifice. What is going on here?

The better educated tend to be staying safer than the lower educated, as Tyler says. That seems like a distinct argument. 

He then goes on to note this:

As a side point, note that in the 1968/1957 pandemics elasticities of adjustment were way lower, because you couldn’t switch things to Zoom, Amazon, and so on.  So those pandemics were closer to being a “lump sum tax” on human life and thus they were cheaper, and had lower deadweight loss, probably in per capita terms as well.  From the framework on welfare economics, that is.  The value of human lives was lower then too.

That does not seem like a side point! It seems like it’s going even farther, and saying that we are worse off because we have the choice to use Zoom and Amazon, and choices are bad. In this case, in part because the choice creates a Copenhagen effect. It seems clear to me that in general having the option to overpay for safety is quite bad. The option to have us collectively do this is even worse. If Tyler is explicitly agreeing that if we lacked this choice, we would have all been better off, then what is he actually arguing for?

There is a distinct argument offered that those with better educations tend to be taking more precautions. To some extent this is obviously unfair, since those with higher educations are often those better able to work from home, and those with better financial cushions. They are also those who tend to listen more to authority, and worry more about looking responsible and safe, and they tend to be members of the blue tribe. The idea that they are doing this because they are smarter or make better decisions is… one hypothesis, sure. Definitely not the whole story.

Also worth saying that an individual cannot decide to not pay the Ramsey tax here, and accept the Covid risk. That choice is unavailable at this time. The world around you is not going to let you do the old normal things. Thus, one can realize that given one’s remaining options it makes sense to play it safe, while also preferring the world in which no one played it safe and you did not either. 

It is also perfectly fine to do any or all of these things, while also preferring the world in which we did proper suppression.

If we knew at the beginning what we know now but were restricted to the choices of “pay more costs” or “pay less costs” a lot of people would choose “pay less costs,” both as a collective and as individuals. 

Both for society in general, and for themselves in particular. 

All Masks Are Not Created Equal

Robin Hanson asks why we treat all masks as the same, simply saying ‘wear a mask,’ when it is very clear that N-95 and surgical masks are more effective than cloth ones:

In my model, all three of these explanations play a meaningful part in the debate. 

There is still a huge lack of consensus on how much better surgical is than cloth. That’s what happens when you ban experiments. It makes it hard to learn things. 

There is clearly very limited tolerance for debating Covid policy, and even less for advocating such policy to the public, or even getting them the message. 

That’s one of two central things going on, in my mind. You Have About Five Words. We can maybe tell people “Social distance, outside, wear masks.” We don’t want to burn another word here. 

There’s also what happens when you make that higher ask of people. If you count cloth masks that one can reuse as needed and are in plentiful supply, and that in my experience are relatively comfortable and look relatively normal, you have made a highly reasonable ask, which a lot of people still treat as unreasonable. You can even customize your cloth mask as a fashion statement or other signal of affiliation. 

If you instead were to say, it needs to be N-95 or surgical (and if you allow both, then it’s not only one word of your five that you’re burning here), then there would be a lot of ‘nope, not reasonable, not doing that.’ Then they’ll often wear no mask at all.

Or they’ll wear a mask, but less often, especially when it actually matters. For the most part, that trade-off isn’t worth it.

The public doesn’t have the ability to process ‘sort of works’ or ‘this works, but this works better.’ They can’t handle ‘this is somewhat safe, but this other thing would be safer.’ You get to say, these things are safe, do those, these other things are unsafe, don’t do those, with clear markings. That’s it. Try to be detailed and there will be a huge demand for “Do cloth masks work, or not? Well? Stupid so-called ‘experts.’” Or they’ll think asking that question owned the libs.

They certainly would not take kindly to being told the most important thing is a cloth mask, then the moment there are enough other masks, changing tunes (after initially lying about masks working at all) and demanding better. 

So instead we wisely took the option for a clean, simple message. Wear a mask, period. If you’re on the ball enough to upgrade to N-95, great, that’s even better.

This is a whole section because this logic extends to all our other strategies and policies, including outside of Covid-19. 

In Other Covid News

CDC recommends this week that Americans wear masks indoors when not inside their own homes. Let that sink in. Until last week, the CDC had not recommended this. Still, better ludicruously late than never, even for the single most important thing to be doing. Thank you, CDC.

Taleb notes: Masks work, says Danish study people described as not saying that masks work

New York City vaguely aware sewage data might be helpful, but not so aware as to release said data anywhere I’m aware of, or to make any effort to isolate locations or otherwise use it to contain the virus. 

Police were sent to seize the laptop and phone of Rebekah Jones. That was her account via Twitter, alternatively two news report are here and here. She was responsible for Florida’s Covid-19 database before she was fired. She claims (credibly, as far as I can tell) that she was fired by Florida for refusing to falsify Covid-19 statistics early in the pandemic. She has spent the time since then trying to provide accurate data and expose corruption. For this, the police pointed a gun in her face and the face of her children, and seized her devices, claiming it was about a security breach.  Which, one way or another, I suppose it was. In particular, they cite this particular breach, which was accomplished by ‘no one bothered to change the login credentials’ (direct source):

If you’d like to get involved, the thread linked to has details on how you can help.

Yet another cheap, fast, easy and seemingly reliable method of Covid-19 testing that, if it works and was implemented, would end the pandemic. This one uses the microscope on your smartphone and can even give you viral load information. So of course I barely registered this news at all because I assumed there was almost no chance the test would be allowed. FDA delenda est. 

There’s also this method of testing (source):

Seems great and I have little doubt we will never do anything of the kind.

A rather complete takedown, as far as I can tell, of an often-cited paper that concluded lockdowns are highly effective and saved three million lives in Europe. It’s a lot of words that mostly aren’t necessary, because it suffices early in the takedown to point out that the paper assumes its conclusion. If you presume that the only things that ever change transmission rates are non-pharmaceutical government interventions, then the alternative to such interventions is that almost everyone quickly gets infected while hospitals are completely overwhelmed, and that’s how you get three million lives saved. So, sure. If there are no control systems and people don’t adjust behavior except when the government issues an order, at which point they instantly adjust to the new transmission rate (which is what the paper assumes), then of course you are going to conclude that all credit must go to the only thing that ever changes anything. This is the level of scientific literacy and interaction with physical reality that even relatively competent Western governments and policy makers (e.g. European ones) seem to be employing these days.

California is locking down hard. 

Here’s the order for Los Angeles. Everyone must stay in their homes. Among other details, podcasts are explicitly exempt, religious services must be outdoors but are otherwise all right, and protests are explicitly allowed. Malls open, playgrounds closed. I think I sense a protest coming on

It’s unclear to me how this is supposed to work out. Do they think people are going to obey rules where they can’t meet a single other person while socially distanced, unmasked and outside? Do they not understand that if you push unreasonably far, people stop listening to you? How would one even begin to enforce such a rule? How long do they intend to maintain things this severely? Conditions elsewhere are probably not going to be much better until at least February or so at the earliest. The full turnaround will likely take months beyond that. Is the plan to shut down the city completely that whole time, on top of what has already been done and not worked? 

Ohio High School Athletic Association coronavirus rules: Students can wrestle, but can’t shake hands. On the one hand this is obviously absurd. If shaking hands is not safe, what exactly do the decision makers think is happening when kids wrestle? On the other hand, shaking hands is a purely symbolic action that you need to stop across the school, whereas wrestling is a compact activity providing value. 

Paper claims that a region in Brazil has 75% overall infection rate. About half by the epidemic peak, the rest after. Raw test numbers were not that high, but they feel confident correcting for waning antibody counts, and the implied IFRs are not unreasonable. They have some speculations on how the numbers got this high, but nothing satisfying. My best guess is quite simple, that the region was unusually vulnerable and also mostly young and taking few if any precautions, and it got hit hard enough to far overshoot its immunity threshold.

Winter is Coming

In other More Dakka news, we want to meet everyone outdoors and go on walks, but baby it’s cold outside. Sounds like a problem. In a move I delayed for too long, I’ve ordered the warmest coat known to man. Should have done it last month, but better late than never. I was going to go with the third warmest coat known to man because for my needs that was definitely enough dakka, but in a running theme these days, they underpriced them and now they’re all sold out, so instead I am ready, if necessary, for a cozy winter in Winnipeg. Which was the motivation behind my friend Seth Burn finding out about this coat in the first place, which he loves. I’ll report back when it arrives. Not cheap. But if it works, totally Worth It.

Until next week.

This entry was posted in Uncategorized. Bookmark the permalink.

32 Responses to Covid 12/10: Vaccine Approval Day in America

  1. Just for the record, as a person who hates using twitter even a little bit, if you’re trying to minimize the necessity of twitter to fully grok these posts I am all in favor.

  2. Liam R says:

    >> Today, the FDA is meeting to discuss Pfizer’s Covid-19 vaccine. By the time many of you read this, they will have hopefully given the vaccine, and perhaps Moderna’s as well, emergency use authorization

    I suppose I’m early to this post myself, but it sounds like an authorization won’t happen until tonight at the very earliest:

    The Vaccines and Related Biological Products Advisory Committee meeting is streaming live now, starting about 45 minutes ago:

    >> Looks like efficacy even for one dose might be 88.9%.

    I was worried Alex Tabarrok’s would be misinterpreted when I read he posted that. The 88.9% figure represents the final efficacy number for everyone who has had *at least* one dose, even though 99% of them went on to get a second dose.

    We only have limited data on what happened in the 28 day period between dose 1 and dose 2, enough to perhaps estimate that the efficacy for just one dose and one dose only, at least for a few weeks, is somewhere between 52% and that number.

    See the analysis around the 20-minute mark here:

    • TheZvi says:

      A naive reading of the graph says that it takes some amount of time to become mostly immune, before which the lines move together, then the treatment line is mostly horizontal starting around day 10, and that horizontal stage comes before the second dose, which would have a delayed effect. My instinctive analysis was “yep, dunno if it’s 88% exactly but looks plausible on this chart.”

      It certainly looks substantially over 50% which is good enough to raise the question of whether to double-dose people.

      • Tobias says:

        Maybe the best solution is one dose now, a second dose when scarcity is no longer a problem? That way more people get at least some immunity, although contacting people and scheduling the second dose 3-6 months from now will be a pain, and many people won’t bother.

  3. Anonymous-backtick says:

    So I understand why American authorities banned testing–all their actions this year were designed to maximize suffering in the hopes it would be blamed on Trump.

    What I’m missing is why, say, Japan or Taiwan didn’t run all the proposed experiments.

    • TheZvi says:

      Not only Japan or Taiwan, you also need to explain Europe and almost everywhere else.

      Here’s a question. If the reason American authorities are maximizing suffering is to get Trump, then have they stopped doing that now that he’s lost the election? Do you expect them to stop doing that on January 20?

      • Anonymous-backtick says:

        In the counterfactual world where Biden peacefully enters the White House I absolutely do think they would let up on covid-related dystopia in the short term.

  4. Anonymous-backtick says:

    *banned experiments

  5. Mike B says:

    I know the winter in Winnipeg was a joke, but it’s definitely not a place you want to be right now! The entire Canadian prairies learned that no, we are not special snowflakes who handled this uniquely well in the first place – we just got exceptionally lucky and locked down fast enough to mostly miss the first wave. But now winter has arrived, and it’s grim:

  6. myst_05 says:

    >>I think this calculation by Tyler Cowen, viewing Covid-19 as a “Ramsey tax problem,” is not only importantly wrong but exactly backwards and rather perverse, in ways worth exploring.

    I think that his primary fallacy is that he’s confounding voluntary behavior and government action, implying that 100% of the lockdown costs are due to people choosing to stay home. Interestingly, the article talking about how lockdowns have saved 3 million people goes in the other direction – as you correctly pointed out they believe that people have perfect adherence to whatever the government does and thus the government is 100% responsible for all the lockdown costs (or benefits, in this case).

    We would probably be arguing about this question for decades to come…

    >> That’s one of two central things going on, in my mind. You Have About Five Words. We can maybe tell people “Social distance, outside, wear masks.”

    Even “outside” seems to be missed by many, at least in Seattle where I live. I like to go for a walk late at night (think 12am) and even at that hour I often see people jogging/walking alone with a mask on. I mean, I don’t object to them wearing anything they want, but I’m perplexed as to why people feel the need to put on a mask when the closest human is at least 50 feet away. At the same time (as Robin correctly points out) even those hyper-careful people usually have a cotton mask on, even though they could definitely afford an N95/KN95. Same deal in Whole Foods – some folks are happily buying $100 wine while wearing a cheap cotton mask.

    >> Should have done it last month, but better late than never. I was going to go with the third warmest coat known to man

    Strongly off-topic: is there a website that lets you find stores that sell high-quality clothing like this? I’m happy to spend $1000 on a jacket, but I want it to be a very high-quality jacket, not a jacket that costs $1000 because its made by a famous designer or because of the brand name alone or because its made from some exotic fabric that isn’t actually any better than good old cotton.

  7. h6 says:

    With such long posts, would it be possible to make subheadings linkable? It’s a hassle to send the link to an acquaintance and say “search for these keywords to find the part I recommend most”

    Thanks for good posts

  8. Sniffnoy says:

    Also in non-covid vaccine news, here’s an article about a potential universal flu vaccine (i.e., it would immunize against the flu in general, rather than specific strains like we currently do).

  9. Data_Schmatta says:

    Do you have any thoughts on these masks from a company called Sonovia claiming their masks eliminate over 99% of the Covid virus within 30 minutes of contact?

    Link to their test results page:

  10. Alsadius says:

    Is it just me, or does Rebekah Jones seem she’s going down the path of the grifter? I mean, yes, the cops overreacted by drawing guns, even after she took like twenty minutes to let them in. But when your fundraiser is asking for a million bucks? That seems excessive, tbh. And while it’s bad security that they didn’t change the login, that doesn’t give her permission to misuse the emergency alerts account. You’re not allowed to rob a house just because it’s unlocked, either.

    And it’s a bit odd to suggest that she was fired for not faking data, when the data she was lauded for posting after being fired was just a re-skin of the official Florida government stats. If she thought that their data was good enough to use in her dashboard, even after they fired her, presumably she doesn’t think it was actually fake. See, for example.

    Also, it looks as though the initial firing was a series of fairly typical human errors, on both her part and her employer’s part, that just spiraled a bit out of control. (tl;dr, she posted some data a bit early, the epidemiologist asked her to pull it down for a review, Jones thought it was bad for transparency and tried to block said epidemiologist, got re-assigned away from her dashboard job for screwing with the work of doctors, thought she was being censored and sent a cranky email, got called out by her boss, and sent a rather sheepish follow-up with the subject line “umh…uh oh?”. Looks to me like severe miscommunication all around, but not malice.)

    • myst_05 says:

      You’re not the only one. Quite frankly I don’t understand the significance of the data that was apparently fabricated. As in – in a parallel universe where the data was NOT fabricated, what would actually be different? Less dead people in Florida? Less cases? More mask mandates? Or is it just data for the sake of data?

      • Alsadius says:

        If data was actually fabricated, that’s bad. Decision-makers need good data. And since it’s a democracy, we’re all decision-makers. Making a stink about fake data is a worthy cause, and I’m cool with that in principle.

        The thing is, in the case of Florida, I can’t find much evidence that it actually was faked. Whenever I dug into it, there was a 90% chance of some journalist being bad at math, and a 10% chance of a harmless difference in methodology.

        For example, Jones’ private dashboard included both positive PCR test results and positive antibody test results as “known cases”, while the official Florida stats counted them in separate categories, and only called the PCR results “known cases”. Neither one is really wrong, and there wasn’t a true discrepancy – just a presentation difference. (With the state government site using the standard method that all the other states used, and the private site being innovative and throwing every scrap of data they could find on the screen – no surprises with either of those.)

    • remizidae says:

      Jones has a substantial criminal record. Technically no conviction yet, but she’s been charged with robbery and cyberstalking an ex. And those charges were before the firing, too. She seems crazy to me. Of course, crazy people can be right sometimes.

  11. cakridge2 says:

    I still must disagree that proper testing would end the pandemic – it’s one thing to know that, say, 150 million people are infected, it is something else to get them all to stay home for 2 to 3 weeks.

    • TheZvi says:

      The thing is that it’s *close* already. R0 is clearly between 0.9 and 1.3. So if we don’t need everyone, we only need (at most) a third of everyone currently going around infected to stop doing that, and things end very quickly.

      • Alsadius says:

        Yes, but people’s behaviour would change as well. I suspect we’d get vastly softer lockdowns (both in terms of private behaviours and in terms of laws), and R0 would stay nearly constant.

        Of course, this would mean substantially less economic damage, and a massive increase to collective mental health, so it’s still a win. But I don’t think it’d be that particular win.

  12. timunderwood9 says:

    On the EU dragging its feet, here in Hungary, Orban, probably as part of the negotiations over the EU budget was talking up the idea of leaving the EU as being potentially beneficial by pointing out, among other things, how Britain was able to immediately approve the vaccine while we had to wait.

    • Alsadius says:

      I am strongly in favour of Hungxit. I’m no big fan of the EU, but it’s supposed to be a group of democracies, and Hungary no longer qualifies. Since they can’t be expelled without unanimous agreement from the other members, and last I heard the Poles are protecting them, them leaving on their own solves a bunch of issues very neatly.

  13. johfst says:

    Great post as always. If you want to link to twitter posts without linking to twitter itself, you can convert them into links by just appending the stuff after the “” in the URL (for example, -> I’ve been using this proxy for reading twitter content on my laptop for the past few months, and other than the occasional hiccup it’s pretty good.

  14. Ray Taylor says:

    I completely agree that Challenge Trials would have speeded things up by many weeks, maybe months. We need a new form of specialised pandemic ethics committees, now!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s