COVID-19 # 64

There's new research on social distancing (by a team of economists no less, using econometrics' techniques). And it works like a charm: they estimate the reduction in cases in the U.S. from March 1st to April 27th at 90%. Interestingly, the also provide simulations showing that social distancing did us almost nothing in March (because the outbreak was still small), but those small gains were amplified in April. They broke down social distancing into 4 categories. They found that banning large events, and closing schools did little or nothing. All the effects were from shelter-in-place-orders, and shutting down of bars, restaurants, and other entertainment-related businesses.

Of course, this means that it's crazy for New York City to have allowed bars to re-open this weekend.

It's a big mistake by @NYCMayor -- he ordered streets blocked off to allow safe distance strolling. But in parts of Brooklyn, bars & restaurants opened windows, selling booze & food along those streets. Result: giant mask-free parties.https://t.co/gFaAHEcnOl

via @GoogleNews

— Laurie Garrett (@Laurie_Garrett) May 17, 2020

I must have been prescient when I noted the other day that I'm starting to think DeBlasio and Cuomo are not heroes at all. ProPublica, an investigative website, published a long article noting  the different policy choices between the mayors of San Francisco and New York City, and between the governors of California and New York. While smaller, New York ended up with 10 times as many deaths, most of them in New York City.

Also, you may recall BD and I noted this it might be premature, but it was not unusual, for the mayor of San Francisco to declare an emergency before they had any cases. Now this seems ... enlightened. Do note that in that post I wrote that it is common to declare an emergency before the emergency hits, to free up funds earlier. I also wrote that this policy is "goofy". Poor word choice ... I meant that it's unexpected when you learn about it, rather than stupid.

There's been a super interesting break out of China. Jennifer Zeng (definitely anti-Xi) reports that Foreign Policy, a respected periodical, has obtained an internal database from China tallying cases and totals. It was gathered by academics for use by the military in providing medical support throughout China. Here's the thing: the database has 8 times as many records ("rows" of data sometimes detailing more than one case) than China's official number of cases. The periodical is keeping a lid on the database currently as they try to confirm its details.

Javier O, writing at Medium, does a lot of digging about early cases in various places around the globe (here, here, and here). He finds an interesting correlation: auto production plants near initial outbreaks in many countries. Why is this interesting? There were 7 major international auto shows in the fall, including 3 in Wuhan (the Detroit of China), running for 12 days between November 1st and 17th. Uh on.

I am guarded about this. But, from a Bayesian perspective, it's getting awfully hard not to conclude that that the virus escaped from a lab. I was quite dismissive of that conclusion in February, both on this blog while in class (sorry BD), and in the seminar I presented to faculty and students on the 27th. Then we found out:

• China has level 4 virology lab near Wuhan.
• That lab's safety was criticized not just by the U.S. government but China's as well.
• They studied bat coronoviruses. In fact, they imported live bats into the region.
• No animal host has ever been as strongly correlated with SARS-CoV-2 as bats.
• They had a satellite lab just around the block from the wet market, again with live bats.
• The lab did work on augmenting bat coronaviruses (basically, letting them evolve in a petri dish to better infect human cells, so we can figure out our own weak spots).
• The samples of the bat coronovirus closest to SARS-CoV-2 were destroyed (they kept the genome).
• Cellphone data shows that the most secure part of the main lab was shut down for over 2 weeks in October.
• It's been a month now that China has been saying that a researcher who went missing from that lab was just working elsewhere. In a month, you should be able to find them for a zoom conference.

Each point pushes me from near 0 probability that the virus came out of a lab, towards 100% probability that it did come out of a lab. I'm not sure where I am right now, but I'd say well over 50%.