Covid 9/3: Meet the New CDC

This week’s news all centers around policy decisions. The new data contains few important surprises, so attention shifts to what actions will be taken and how that will affect the path we follow going forward. The CDC’s fall and transformation into an arm of the White House reelection campaign is now complete. Others continue to come up with, suggest and criticize various policies. 

Before we get to all that, let’s run the numbers.

Positive Test Counts

July 9-July 15 108395 53229 250072 20276
July 16-July 22 117506 57797 265221 20917
July 23-July 29 110219 67903 240667 26008
July 30-Aug 5 91002 64462 212945 23784
Aug 6-Aug 12 93042 61931 188486 21569
Aug 13-Aug 19 80887 63384 156998 20857
Aug 20-Aug 26 67545 66540 132322 18707
Aug 7-Sep 2 55000 75401 127414 21056

Only the West’s number here is reassuring. The South’s number here is disappointing but reflects a rebound in the number of tests after a steep decline last week. The Midwest situation continues to get worse. The Northeast has some reason to worry, but the increase is mostly explained by increased testing.


June 25-July 1 858 658 1285 818
July 2-July 8 894 559 1503 761
July 9-July 15 1380 539 2278 650
July 16-July 22 1469 674 3106 524
July 23-July 29 1707 700 4443 568
July 30-Aug 5 1831 719 4379 365
Aug 6-Aug 12 1738 663 4554 453
Aug 13-Aug 19 1576 850 4264 422
Aug 20-Aug 26 1503 745 3876 375
Aug 27-Sep 2 1245 759 3631 334

The Midwest number is bad news, the West and Northeast numbers are excellent news. The South’s is an improvement, but less of an improvement than expected, so it counts as bad news. Deaths are on a clear downward trend in general and that should continue for at least several weeks, as the overall situation continues to improve right now. 

Positive Test Percentages by Region

The Covid Tracking Project’s data has a very strange and very negative number of positive tests from Massachusetts this week, which I’ve corrected to a reasonable number. 

Percentages Northeast Midwest South West
7/16 to 7/22 2.49% 5.13% 13.29% 8.56%
7/23 to 7/29 2.54% 5.51% 12.32% 7.99%
7/30 to 8/5 2.58% 7.26% 12.35% 6.68%
8/6 to 8/13 2.30% 5.67% 14.67% 6.98%
8/13 to 8/20 2.06% 5.62% 9.41% 6.47%
8/20 to 8/26 1.86% 5.78% 9.93% 5.88%
8/27 to 9/2 1.87% 6.37% 9.38% 4.78%

This makes it clear the Midwest is getting worse and not merely testing more, and the West is rapidly improving. The South’s situation remains ambiguous, but looking at the individual states makes it looks like things are indeed improving slowly. 

Test Counts

Date USA tests Positive % NY tests Positive % Cumulative Positives
June 25-July 1 4,352,981 7.1% 419,696 1.2% 0.82%
July 2-July 8 4,468,850 8.2% 429,804 1.1% 0.93%
July 9-July 15 5,209,243 8.4% 447,073 1.1% 1.06%
July 16-July 22 5,456,168 8.6% 450,115 1.1% 1.20%
July 17-July 29 5,746,056 7.9% 448,182 1.1% 1.34%
July 30-Aug 5 5,107,739 7.8% 479,613 1.0% 1.46%
Aug 6-Aug 12 5,121,011 7.3% 502,046 0.9% 1.58%
Aug 13-Aug 19 5,293,536 6.2% 543,922 0.8% 1.68%
Aug 20-Aug 26 4,785,056 6.0% 549,232 0.8% 1.77%
Aug 27-Sep 2 5,042,113 5.5% 606,842 0.8% 1.85%

New York’s positive percentage creeped up substantially this week while the test count continued to rise, especially in the last few days. I am definitely worried that something has gone wrong and we are no longer on a slowly but steadily improving path. If things are suddenly getting worse here now, presumably it is a school problem, and that does not at all bode well.

The national picture here however is quite good. Our test numbers creeped back up a bit and the positive percentage fell substantially. (Recorded) hospitalizations are down as well. Yesterday was the first day in a long time they didn’t decline day over day, but for now I’m treating that as a mere blip.

Center For Disease Control Sorta Partially Walks Back Its Opposition To Disease Control

After taking a pounding from all sides for several days, director Robert Redfield (who, alas, probably can’t be played by the newly retired Robert Redford in the inevitable HBO movie version, but I’m hoping he’ll make an exception because come on) ‘clarified’ the new guidelines that led to last week’s headline.

In a statement, Director Robert Redfield said those who come into contact with confirmed or probable COVID-19 patients could be tested themselves, even if they do not show symptoms of the virus.

“Testing is meant to drive actions and achieve specific public health objectives. Everyone who needs a COVID-19 test, can get a test. Everyone who wants a test does not necessarily need a test; the key is to engage the needed public health community in the decision with the appropriate follow-up action,” Redfield said.

So he allows for the possibility that people who come into contact with confirmed cases could be tested, in theory, I mean it’s a thing that happens from time to time. Very generous of him. And it’s great to hear that everyone who “needs” a test can get a test, especially considering the numerous reports that this is not the case for any meaningful value of getting a test, and the fact that this is not the case is the only good reason to revise the guidelines.

So… Do you feel clarified now? 

Me neither. This does not feel like a walk back to me. It feels like they’re doubling down.

Instead, it seems their strategy is to assert control over… evictions

I don’t want to get too deep into the economics of this move. I won’t discuss whether it is completely and totally insane, or how much it will permanently drive up rental costs since renting means the government might decide to seize your property outright and pay you nothing in return, while you maintain it under penalty of law at your own expense in the hopes that the government will one day give it back. 

I will instead say that this is completely and utterly unconstitutional and illegal and in no way something the CDC has any authority whatsoever to do. You are the Centers For Disease Control, not the Centers for Rent Control. 

So you know what? Fine. You did it. Congratulations. Burn it to the ground. CDC Delenda Est. 

Centers For Disease Control Advocates Disease Control

This just in: The CDC has also informed states that they should be ready to distribute one of two vaccine candidates by November 1.

Under normal circumstances this would be both the correct action and great news. It would mean that the two vaccine candidates have a substantial probability of being far enough along to be worth deploying soon, potentially heralding a swift end to the pandemic. Given “medical ethics” and the general overwhelming paranoia about deploying a vaccine by all Very Serious People, I have an extremely strong prior that any deployment would be too late rather than too early.

It’s certainly good news, even in these times, that they have the good sense to tell states to get ready to distribute whether or not there is any intent to actually distribute. We should get ready to distribute long before we expect to need distribution. Things will inevitably go wrong and cause delays, which we can address now before those delays cost lives.

Unfortunately on so many levels, these are not normal times. We have the president we have, who is facing a presidential election… on November 3, two days after the target date. That does not in any way feel like a coincidence.

I would be very surprised if this CDC announcement is not being made under, at a bare minimum, extreme pressure from the White House. This was a political decision, and together with other CDC news, it seems safe to respond as if the CDC is completely captured by the White House and is acting under its direct orders to serve the President’s political interests and whims, rather than as a center for the control of disease.

If we take as given that Trump is planning a big October Surprise, I’ll take ‘issues an order to distribute the vaccine early’ over every other alternative I can come up with, except for the possibility that it might actually work and win him the election.

The thing is, he’s right.

He’s not right for the right reasons. He’s not understanding the situation and doing the Bayesian calculus and realizing that early distribution of a known-to-be-safe vaccine is a huge net benefit to America and the world, and we should follow in the footsteps of China and Russia and get on that. Of course not. That’s not how he thinks. 

He will issue the order, if he issues it, because he thinks it will help him get reelected, full stop, without caring about whether it is a good idea.

That doesn’t make him wrong. If you think he’s wrong, as Tyler Cowen says, show your work.

And if and when he does issue that order, if you are Biden, how do you respond?

If Biden says ‘yes, that was the right thing to do’ then obviously it’s a huge Trump win (and also a win for the world, but in context neither side cares about that).

If Biden says ‘no, that’s not a responsible thing to do’ then Trump is the one who is doing the only action that matters to get us out of that, and Biden is the one not doing it because “medical ethics.” 

Thus, it would be a great play even if there were risks that made it a bad idea – it’s not like those risks could be properly communicated to the public. Nor could a lack of such risks be communicated to the public, especially over the objections of the Very Serious People, but also even with their full support. A huge percentage of Americans don’t want the vaccine, sight unseen, even under the best conditions. 

I wonder why the public has such distrust for public health authorities and doesn’t want to inject strange things into their bodies on such authorities’ say so. It’s not like they are constantly lying to us about pretty much everything.    

Health Experts Warn of Dangers of Ignoring Health Experts

What’s new with those vaccines in Russia and China? I can’t find any news on whether they’re working, but we do have news that the Very Serious People are Very Concerned

Whenever people who will always have objections object to something, it’s important to remember that you should not expect to update your beliefs in any particular direction. Health experts will warn about the dangers of doing the thing their ‘ethics’ say not to do, with whatever case they think is the strongest, whether or not they have a good case. So when you see them make their case, you should update based on whether their case is stronger or weaker than expected. If they make terrible arguments that are worse than you expect, you should update in favor of there not being good objections.

In this case, it seems there are two concerns.

The first concern is that the vaccine is based on the common cold. Therefore, those who have had the wrong common cold will already have an immune response ready, and the vaccine won’t work on those people. This might reduce how often the vaccine is effective. 

That’s a reasonably good objection. It’s a great objection if you’re choosing what approach to use. As an objection to deploying the vaccine versus doing nothing, though, it’s rather weak. If often the vaccine does nothing, then the calculus on whether the vaccine is a net benefit is unlikely to change much. Every extra immune person helps, and the costs of deployment are trivial relative to that benefit. What you’re looking for is active downsides, not reduced frequency of upside.

The second objection is that a previous HIV vaccine that used some similar characteristics in its delivery ended up making people more vulnerable to HIV, so they warn that this too could make people more vulnerable to HIV.

I know complete and utter BS when I see it. The previous HIV vaccine put people at risk for HIV because it was trying to be an HIV vaccine and messed up. Not because it so generically forked with the immune system that it happened to make HIV worse. This vaccine is trying to be a Covid-19 vaccine. It could plausibly make Covid-19 worse. But if Very Serious People are talking about HIV risk here, it means they have no cards to play. Update accordingly.

Arizona University Kind of Solves Covid-19

Seriously, it kind of did. Check this out.

It turns out, if you actually care about solving the problem, you can test waste water from each building, and then test everyone in the building when the water tests positive, thus catching cases before they have much chance to spread. Do that consistently, using the quick tests that are actually easy and dirt cheap, and it’s over. That doesn’t mean the University of Arizona is in the clear, because no one else is doing it and they therefore have to constantly worry about reintroduction. But if we all followed this procedure? It would all be over in a month.

This has been your periodic reminder of The Kinds of Things a Functional Civilization Would Do.

As opposed to, say, not telling people when a classmate tests positive

What About Those Reinfection Cases?

This week’s periodic panic about lack of immunity was unique because it had actual bad news to consider. Normally people don’t need actual bad news, and mumble something about how we can’t be sure how long things will last in order to sound serious. In the past, this has somehow kept happening while there were actual zero reports of reinfections.

Now there are a non-zero number of reports of reinfections, which led to a moderately larger amount of panic and fear mongering. It turns out that its frequency and intensity does respond somewhat to actual news. So how worried should we be about these new reports?

As usual, the news article starts out with the scariest take it’s willing to dish out, with bullet points like “These reinfection cases demonstrate how immunity to the novel coronavirus is somewhat transient, especially with mild infections.” But overall, I’m actually very happy with the lack of mongering going on here from Business Insider, so positive reinforcement to them. 

They get to the right answer here, which is definitely ‘not very worried.’ 

What these cases show is not that immunity is short lived. They show that a very small number of people don’t get complete immunity when they are infected. 

But that is neither surprising nor particularly impactful. A system of containment doesn’t care much about a 1% failure rate given how this virus works. With a total of 6 known cases worldwide and large incentives to find them, there’s no way the number of people who don’t regain full immunity is enough to be worth worrying about. It shouldn’t impact how anyone lives their life at least until after they have symptoms again. And in most of these cases, the secondary infections were mild anyway. 

What this definitely doesn’t mean is that we now have to suddenly worry about immunity fading quickly. In these cases, the second infection happened quickly, often within a month or so. We know for sure that immunity almost always lasts far longer than that. So this isn’t people who got immunity and then lost it, it’s people part of the small group who were never immune in the first place. Which we’d prefer didn’t happen, sure, but isn’t impactful. 

If we suddenly had six new cases, all of which had their first infection in February or March and their second one in August, then I’d be much more worried that five or six months was enough to start to meaningfully degrade immunity. That’s not what we saw, so six months is insufficient to do this. We can assume that for practical purposes immunity lasts a minimum of seven months, and then apply Lindy, and assume that the end of that is where things begin to be a problem. Which should be enough time to get the vaccine online. Excellent.

This was worse immunity news than I expected this week. But overall, does this week make us think immunity is shorter (because we found some reinfection cases) or longer (because almost everyone stayed immune one more week)? I don’t think that is clear.

Physical World Does Not Think Six Feet Is a Magic Distance

People claiming with presumably straight faces to be ‘researchers’ used that authority to get into the paper that perhaps the six foot rule could use a bit of nuance. That it matters how long you’re there for, indoors or outdoors, poorly or well ventilated, silent versus spoken versus shouting or singing, dense versus sparse crowd. If I had to choose three additional considerations when measuring risk and deciding how far to keep away and whether to require masks, then those are probably the correct variables to consider. And all their directional assessments seem right. So, good job, I guess. As far as it goes.

If it makes people actually think about their physical situations a bit and optimize somewhat, that would be great. Hopefully the nuance is net helpful. 

If you want a lot of nuance on what to be doing and how to measure risk, the microCOVID project is one option. I had the chance to comment on their document and models a bit. They didn’t take every suggestion I made, but they are definitely trying to come up with reasonable answers and provide practical help. If that seems interesting or valuable, check it out for another opinion. 

A note for those who try the microCOVID project is that their basic system of ‘use a budget to allocate risk’ originates in the need to find a policy that roommates can all live with and follow, without anyone feeling cheated or causing anything too perverse. If you have different binding constraints, different strategies will make sense for you.

Important Things Are More Important

Periodically we see outrage like this about the hypocrisy of letting Very Important People like celebrities or the rich get away with doing things that the rest of us are told not to do. It seems that while mostly not allowing concerts, New York allowed the Video Music Awards to completely break a lot of the rules. 


If anything, the report shows a decided shortage of such hypocrisy. The event had to be spread out throughout the city, extensive precautions were taken for spots that lasted only a few minutes. I am guessing that everyone involved was tested in advance, probably multiple times. And that was then shown to millions of people. Not my thing, but the same way that sports must go on, other things that bring joy to millions in exchange for the exposure of dozens or hundreds is obviously a trade-off that we want to make.

People are so against doing things that make sense, and so unwilling to deal with ‘hypocrisy’ or ‘inequality’ that they think that you not being allowed to have a private dance party means the VMAs should stop. That we shouldn’t look at the value of an activity in dollars or happiness, and compare it to the risks involved, when deciding what to do, to maybe help make this lockdown liveable for all and helping the economy survive. 

Or that we shouldn’t give extraordinary flexibility to those willing to take extraordinary precautions. If you have the time and money to test everyone and make something safe, I don’t care if it otherwise violates guidelines.

The key is that this needs consensus that the exception is a reasonable exception. That it involves minimal risk given the benefits involved, that precautions were taken, that it is an efficient allocation of risk with a solid story attached. Otherwise, even if it’s a good idea, it decays people’s willingness to follow the rules.

I would hope that the ‘it’s being broadcast to millions of people who want to see it’ rule together with the ‘it’s worth enough to spend what it takes to get everyone tested beforehand and take all the precautions’ rule would cover the right times to make an exception pretty well. 

If both of those apply, do it. If they don’t both apply, respect the rules.

Or, if there’s something you think is too important and has to be done anyway, understand that not doing so will undermine the rules themselves and decide whether it is worth it.

Contrast this with, say, Nancy Pelosi going to a hair salon and not taking precautions. There is zero excuse for that. The outrage is completely justified.

This entry was posted in Coronavirus. Bookmark the permalink.

19 Responses to Covid 9/3: Meet the New CDC

  1. Chris says:

    Pfizer announced this morning (?) that they expect safe/effective trials to be complete by the end of October.

    • TheZvi says:

      Thanks! Interesting. Two problems.

      One, SPY -0.87% on the day and Pfizer +0.3% or so. If this was a real update, markets would rally, and Pfizer would rally a lot. This can’t be big news.

      Two, he’s saying that “if there are enough events” between now and then he can do it. Which is trivially true, but doesn’t mean he has a high expectation that he can do it. My guess is that this is a make-nice announcement.

  2. myst_05 says:

    Would you personally agree to get vaccinated on November 1st, presuming the stars align just right and it becomes possible?

  3. Brett Bellmore says:

    I continue to think that, if you’re going to overlay multiple graphs, such as you do with deaths, they should all be normalized to population. Your deaths graph makes it look like the South, (Where I live.) is being hit as hard as NY was back in April. But, of course, we’re not remotely facing the same death rate per 100K as NY saw, we just have a much larger aggregate population.

    Failing to normalize to population is just bad data presentation practice.

    • PDV says:

      Depends on the goal. If the goal is to depict the scale of the overall problem, normalizing to population is reasonable. If the goal is to depict how well the problem is being controlled in various areas, normalizing to population is counterproductive. In fact, in that case, it would be better to anti-normalize, e.g. display it relative to a baseline of “how bad was it in this region six months ago”.

      • Brett Bellmore says:

        By normalizing to population, I mean graphing deaths per 100k population. So that a region with twice the population doesn’t end up looking twice as bad, even though exactly the same fraction of the population died.

        “The South” has a population of 125.58 million. New York has a population of 19.45 million, the North East minus New York, 36.53 million, the Midwest 68.33 million, the West 78.35 million. These regions have hugely different populations.

        Is it any wonder the South looks like we’re doing badly in that graph? We’ve got nearly twice the population of the Midwest or West, 3.4 times the population of the NE-NY, 6.5 times the population of New York.

        If that graph was displaying deaths per 100K, rather than total deaths, it would be clear that we’re not doing much worse than the West, and nobody else is doing remotely as bad as NY or the NE were doing in April.

      • PDV says:

        That is exactly what I assumed you meant and what my comment was assuming “normalize to population” means. The deaths per 100k is not actually the main thing we care about, because exponential functions are a bitch. We care about how well contained the pandemic is in those areas, and whether they’re losing ground or gaining it.

        The correct “normalization” is one that shows all the graphs at the same time. Because if one city has 1 per 1000 residents, and last week they had 1 case per 500, that is good. It is probably *better* than somewhere which has 1 case per 10,000 this week and had one case per 20,000 last week. Normalizing to population does not help us get at the important data.

  4. jacklecter says:

    I *was* angry about the VMAs, and this changed my mind about that.

    (I hate fake-rule-of-law, where the “rules” are not the rules, but that horse looks to have well and truly left the barn by now, and the additional factors you cite are what we’d write into the law if I had my druthers.)

  5. Anonymous Bosch says:

    The idea that Trump is “right for the wrong reasons” assumes that the downside risks of a politically rushed vaccine can’t possibly outweigh the positives, and I think that is probably true if you only consider it in terms of “COVID health risks” vs “vaccine health risks.” I think it’s implausible that even a relatively shitty vaccine would have a bad enough ratio of protective effects to side effects.

    But if the vaccine does turn out to be ineffective or comes with surprising side effects, the negative effects could go beyond the current pandemic by causing lasting damage to the concept of compulsory vaccination in the public mind. At best it would strengthen anti-vax movements, at worst it could turn into a fully partisan acid test (only trusting vaccines approved during your party’s administration, for instance).

    This is not merely a speculative consequence; a smaller-scale version of this tragedy played out when the Ford administration rushed a swine flu vaccination in 1976. It is difficult to quantify in the “show your work” sense because polling was not as ubiquitous and we don’t know how many people were discouraged from vaccinating by the controversy. But it’s not hard to make some assumptions where the downsides of “an extra month or two of COVID cases to ensure the correct choice of vacccine” could be outweighed by significantly elevated non-compliance rates during a future pandemic, or even a couple of bad flu seasons.

    • TheZvi says:

      Long term doesn’t matter to him. All he cares about is winning the election (or getting close enough to dispute/steal it, depending on what you believe on that front). So he doesn’t care if the vaccine works. He only needs to keep the balls in the air until November 3 – and if he announces the vaccine on e.g. November 1, by the time we learn if it works or if it’s really safe, it’ll be far too late. He can also use optionality – if he’s losing by a lot, throw a Hail Mary, if he’s winning then no need.

      • Anonymous Bosch says:

        I assumed from “he’s right” you were talking about YOUR view. That Trump is probably making the right choice to rush a vaccine even though he’s obviously doing it for the usual grubby, selfish political reasons. That’s the only reason I pointed this out.

      • TheZvi says:

        Ah, right. My dad was actually involved in the 1976 Flu discussions. I do know it is *possible* for deploying early to net backfire, but I’d be shocked if a reasonable calculus thought it was anywhere close.

  6. Error says:

    Request: Could you categorize or tag your covid posts? They currently show up as “Uncategorized”.

  7. Pingback: Covid 9/10: Vitamin D | Don't Worry About the Vase

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