A Grim Viewpoint
I don’t think it takes a savant to figure out what’s going on here. The IHME model was just wrong, we were about to hit the number of deaths it predicted for three months from now, and it recognized its error by tweaking a dial. Though it did hold out that its projections “reflect the effect of premature relaxations on restrictions,” a convenient alibi for a model that was already out of date.
The CDC model is a more appropriate manifestation of what we’ve been talking about here. The country is pulling back on lockdowns while stuck at a peak on deaths and cases. In fact, because aggregating a large country with one curve is completely misleading, the pullback is happening while peaking is still going on. If you remove New York, New Jersey, and Connecticut, where cases spiked early, you see a curve that goes down slightly in the middle of April, causing everyone to relax, and then starts its way up again.
No other country has given the go sign to its population under these conditions. Countries around the world are starting to return and there appears to be some jealousy or mass delusion that we’re in the same position. We’re not and we never were. In fact Italy’s relaxation of the lockdown looks a lot like our lockdown. Having blown February by botching testing, we’re on the precipice of blowing March and April by re-opening far too soon. Even California, seen as the responsible actor in this play, is dropping restrictions (albeit not many), doing its part to put the country on a bad trajectory. The signals are all in the direction of placating protesters and putting America back in an incredibly dangerous position.
So the CDC is adapting to circumstances like the IHME model, though in a more responsible fashion. It shows a ramp-up about two weeks from now, matching the beginning of the lockdown relaxations. This isn’t a second wave but the second half of a first wave that merely plateaued but never ended.
The Washington Post talked to the creator of the model, who disavowed knowing how the data was being used, but who helpfully explained that he was basing his decisions in the model on “reopening scenarios” that “could get out of control very quickly.”
This likelihood, by the way, is the worst possible outcome. Reopening and then seeing cases and hospitalizations that will overwhelm the system again, leading to another closure, would be worse for public health and worse for the economy.
The White House is sticking with a model made by their economist Kevin Hassett, who once predicted the Dow Jones Industrial Average would reach 36,000. (It hasn’t.) He’s using a “cubic” model which might as well be called a “fit the numbers to Donald Trump’s re-election plans” model. We’re seeing in real time a battle between epidemiologists and economists over predicting the future. Epidemiologists are catching up to limited data and reworking on the fly; economists are lying to flatter their bosses and ignoring reality.
If you really play out what the CDC model is showing, it gets you to half a million deaths by the end of the summer. You have to conclude, because what’s happening is so obvious, that this is an acceptable level of death to the president and his staff. They are fine with mass death in America. They have always been fine with mass death in America; we let mass murderers get guns, after all. They would rather have mass death than pay for unemployment benefits for too long. Their entire strategy throughout April was to figure out the best way to allow mass death, in service to the economy. This was a false choice, incidentally; we will have mass death and an obliterated economy. And when that happens, the only straw left to grasp will be the culture war.
Anyone working in the federal government on pandemic response right now who doesn’t want to be known historically as a mass murderer should probably resign.
First Response
David Dayen
Everyone is trying to make sense of the leaked CDC model that shows coronavirus deaths peaking at 3,000 per day, 70 percent more than the current plateau, and a whopping 200,000 new cases every day as well. The noted IHME model out of the University of Washington doubled as well, now predicting more than 134,000 deaths by August after sitting on 70,000 for a while.I don’t think it takes a savant to figure out what’s going on here. The IHME model was just wrong, we were about to hit the number of deaths it predicted for three months from now, and it recognized its error by tweaking a dial. Though it did hold out that its projections “reflect the effect of premature relaxations on restrictions,” a convenient alibi for a model that was already out of date.
The CDC model is a more appropriate manifestation of what we’ve been talking about here. The country is pulling back on lockdowns while stuck at a peak on deaths and cases. In fact, because aggregating a large country with one curve is completely misleading, the pullback is happening while peaking is still going on. If you remove New York, New Jersey, and Connecticut, where cases spiked early, you see a curve that goes down slightly in the middle of April, causing everyone to relax, and then starts its way up again.
No other country has given the go sign to its population under these conditions. Countries around the world are starting to return and there appears to be some jealousy or mass delusion that we’re in the same position. We’re not and we never were. In fact Italy’s relaxation of the lockdown looks a lot like our lockdown. Having blown February by botching testing, we’re on the precipice of blowing March and April by re-opening far too soon. Even California, seen as the responsible actor in this play, is dropping restrictions (albeit not many), doing its part to put the country on a bad trajectory. The signals are all in the direction of placating protesters and putting America back in an incredibly dangerous position.
So the CDC is adapting to circumstances like the IHME model, though in a more responsible fashion. It shows a ramp-up about two weeks from now, matching the beginning of the lockdown relaxations. This isn’t a second wave but the second half of a first wave that merely plateaued but never ended.
The Washington Post talked to the creator of the model, who disavowed knowing how the data was being used, but who helpfully explained that he was basing his decisions in the model on “reopening scenarios” that “could get out of control very quickly.”
This likelihood, by the way, is the worst possible outcome. Reopening and then seeing cases and hospitalizations that will overwhelm the system again, leading to another closure, would be worse for public health and worse for the economy.
The White House is sticking with a model made by their economist Kevin Hassett, who once predicted the Dow Jones Industrial Average would reach 36,000. (It hasn’t.) He’s using a “cubic” model which might as well be called a “fit the numbers to Donald Trump’s re-election plans” model. We’re seeing in real time a battle between epidemiologists and economists over predicting the future. Epidemiologists are catching up to limited data and reworking on the fly; economists are lying to flatter their bosses and ignoring reality.
If you really play out what the CDC model is showing, it gets you to half a million deaths by the end of the summer. You have to conclude, because what’s happening is so obvious, that this is an acceptable level of death to the president and his staff. They are fine with mass death in America. They have always been fine with mass death in America; we let mass murderers get guns, after all. They would rather have mass death than pay for unemployment benefits for too long. Their entire strategy throughout April was to figure out the best way to allow mass death, in service to the economy. This was a false choice, incidentally; we will have mass death and an obliterated economy. And when that happens, the only straw left to grasp will be the culture war.
Anyone working in the federal government on pandemic response right now who doesn’t want to be known historically as a mass murderer should probably resign.
No comments:
Post a Comment