Thursday, May 21, 2020

Coronavirus Confidence Interval by @pdxsag: 300k-1,000k

[Our correspondent pdxsag (@pdxsag) shares his thoughts on corona mortality.]

This week Lion of the Blogosphere (from CBS Links), one of the earliest bloggers to call attention to the gathering WuFlu storm, came out this week estimating 2.5 million Covid19 fatalities in the US. He based his estimate extrapolating NYC's results assuming an overall 1% IFR sub-divided by age cohorts and a 70% overall infection rate before herd immunity is achieved to yield:

Age IFR Herd IR Mortality
0-19 0.003% 70% 1,722
20-44 0.1% 70% 76,209
45-64 0.9% 70% 528,570
65-74 3% 70% 711,690
75+ 7.5% 70% 1,151,850
Total 2,470,041

I'm not sure where he got his population numbers from, but using the census bureau Age & Sex table for 2019, I was able to get close-enough approximations to verify his arithmetic. However, my contention is that unique environmental and policy factors caused NYC to be an outlier. I believe IFR's in the rest of the country will be considerably less, as well as the level required to achieve herd immunity, owing to better preparedness and treatments. This is my rebuttal:

Age IFR Herd IR Mortality
0-19 0.001% 50% 408
20-44 0.05% 50% 26k
45-64 0.3% 50% 124k
65-74 1% 50% 158k
75+ 5% 50% 533k
Total 840k

Frankly, close to one million deaths still seems like an incredible number. It is a worst-case that assumes the IFR's are not off by an order of magnitude, which assumes there were no early Covid19 transmissions – despite many anecdotes to the contrary – and the very low serology rates we are seeing today are true. (This would be the no tin-foil, textbook-epidemiology scenario, espoused by the always eminently rational Cochran and Sailer.)

On the other hand, for the working age population in this scenario has less than 150k total deaths. In the context of a 330M population with annual births of 3.7 million, Covid19 won't even make for a negative population growth year.

Interestingly, a modest baby boom of 10% owing to the lock-downs would result in as many Covid19 babies (370k) as Covid19 fatalities under 75 yo (310k). As pandemics go, and as I've been saying since March, this one is a big fat nothing burger.

Finally, there has been a lot of debate on what the true Covid19 mortality rate is. As many people have argued Covid19 deaths are being over-counted, as have argued they are being under-counted. One thing that cannot be faked, however, is the excess fatality rate. It's not a perfect measurement because the lock-downs have caused a decrease in accidents and homicides, but it is an objective measure.

In the 8 weeks since the infection rate started ramping up in March, we are at 110k excess deaths (@lymanstoneky). 110k is a long way from 1 million. At ~10k a week, we'll never get there.

It will require a Spanish Flu-like 2nd wave greater than the first wave. Given our better understanding of epidemiology, better medicine, and lacking deprivation of a World War, I don’t believe that is at all a possibility.

My rule of thumb: natural events are never as bad as predicted when we see them coming. This leads me to estimate based on excess deaths to date is 200k by Aug and 300k by the end of the year. I will allow that a bad 2nd wave could get us to 500k by the end of the year (adjusting week 12 to be 0 for covid).


League of Women Voters said...

It’s becoming increasingly clear that a big (maybe the biggest) risk factor for coronavirus transmission is speaking. Singing is even worse. The louder you speak or sing, the worse it gets.

Some confirmed early superspreader events were choirs. A lot of others were churches, where everyone gets together and sings hymns full-blast. This person’s explanation for the surprisingly low rate of subway-mediated transmission in Japan is that nobody talks on a Japanese subway.

All this makes sense. Coronavirus has mostly droplet transmission. There are three ways to get droplets: coughing, sneezing, or talking/singing. You do one of those about a thousand times more often than either of the others.

Stagflationary Mark said...


Seems like a reasonable worst case. The one thing that we can take some comfort in is that as the deaths go up, the social distancing, mask-wearing, and hand-washing should also go up.

Risk taking should also go down. The virus can obviously spread easier among those who take more risks. There are a lot of risks that we actually have control over. One is grocery shopping. All things being equal, the person who shops every day is taking roughly 7 times more risks than someone who goes once a week. I choose to not go at all. I am fortunate that I can have them delivered (and we take special care to disinfect and/or quarantine everything that we have delivered).

Am I being too paranoid? The death rate alone gives me pause, but the hospitalization rate seals the deal. Until proven otherwise, I assume that I have a 10% chance of being hospitalized if I catch it and a 1% chance of dying. There are two of us in this household. If one of us gets it, then the other one probably will too. Based on my assumptions, that would put the odds of at least one of us going to the hospital at about 1 in 5 and at least one of us dying at about 1 in 50. Not even remotely worth the risk to me.

Interestingly, a modest baby boom of 10% owing to the lock-downs would result in as many Covid19 babies (370k) as Covid19 fatalities under 75 yo (310k).

I would be surprised if we see much of a baby boom, if any. From what I’ve read, people don’t tend to have more babies when there is increased economic uncertainty. That’s especially true when unemployment is at historic levels. Anecdotally, I definitely wouldn’t. But I don’t fully discount the idea. There’s been a run on puppy adoptions. Do puppies count as a baby boom? Puppies don’t need to go to college someday and are generally cheaper to feed and clothe (depending on the puppy and owner, I guess ;)).

For what it’s worth, I adopted my first dog as an adult the first time I was laid off. It probably wasn’t the smartest thing I ever did from a risk management perspective to adopt a new puppy with no job, but it worked out fine. Could have been a financial disaster of epic proportions if I hadn’t found a new job within a few months though. To be that young and optimistic again. I’ve nearly forgotten the feeling. It’s all illusion of prosperity thinking now I guess, thanks to the apparent need of perma-ZIRP just to keep the party going and the naturally increasing cynicism as one ages.

Allan Folz said...

From what I’ve read, people don’t tend to have more babies when there is increased economic uncertainty.

No idea where you read that. Certainly does not match my loose recollection of history.

If we don't get even a tiny baby boom (+0.8 std dev above normal) from 2 months of forced lock-down Western Civ is probably too far gone to matter what the Fed & Wash DC does or does not do.

Stagflationary Mark said...


March 11, 2009
U.S. News & World Report: Why the Recession Might Cause a Baby Boom

One possible explanation for the findings is that when people are laid off or under-employed, they see it as an opportunity to have more children because they have more time.


September 30, 2014
The Atlantic: The Recession's Baby Bust

At least a half-million fewer Americans were born as a result of the 2008 crisis.

There's a reason why "boom times" and "baby boom" both contain an onomatopoeic signifier of the procreative act. In developed countries, fertility rates tend to go up and down with GDP.

The one thing that could make it different this time, but I still doubt it:

April 13, 2020
Business Insider: The coronavirus forced the world's largest condom maker to stop producing — and now there may be a shortage of 100 million condoms

Some will choose to postpone having a baby. Some will not want to risk having a baby delivered 9 months from now in a hospital. What if there’s a second wave? Some will accidentally have children during a pandemic with high unemployment and uncertainty. Some will definitely choose not to do it. I wouldn’t choose to do it. Would you? Surely there are better times to plan to have a child.

For what it is worth, I believe that if free time was all it takes to start a baby boom, then there would be a baby boom started every weekend.