Nate Silver

Nate Silver @NateSilver538

It's also very likely that greater polarization also decreases the incumbency advantage. If you imagine a generic incumbent president (not Trump) in an average economy, we'd expect them to win the popular vote by only ~3 points given polarization today. — PolitiTweet.org

Nate Silver @NateSilver538

One other thing on this... you do get better (backtested) results if you adjust for the degree of polarization. A projected ~6 point Trump deficit in the popular vote today would be more like 10 points in an era with average polarization. — PolitiTweet.org

Nate Silver @NateSilver538

@dmorey Well, *I* don't know personally ... but we use the SP500 and consensus economic surveys and they actually have some predictive power. Also, though, our model thinks economic uncertainty is about 2.5x higher than normal. — PolitiTweet.org

Nate Silver @NateSilver538

This does anticipate some degree of an economic rebound. Economic conditions as of *today* would point toward more like a 10-point popular vote loss (actually pretty close to current polls). — PolitiTweet.org

Nate Silver @NateSilver538

Been working hard on the economic component of our presidential model and FWIW, projected economic conditions by November would have Trump favored to lose the popular vote by ~6 points, although with a large amount of uncertainty. — PolitiTweet.org

Nate Silver @NateSilver538

RT @FiveThirtyEight: The winners and losers in our updated NBA forecast: https://t.co/X9tuFtgQq9 — PolitiTweet.org

Nate Silver @NateSilver538

All of these parameters are educated guesses, of course. There's a ton of uncertainty. There may be scenarios where deaths don't increase that much. But it probably means that the increase in new cases has to stop, like, really soon. — PolitiTweet.org

Nate Silver @NateSilver538

So if we held at "only" 200K cases per day, deaths would probably increase some, but not necessarily a ton. The problem is, his model expects infections to keep going up to 265K. So that would get you to 530-1,060. That to say is, an increase from where we are currently. — PolitiTweet.org

Nate Silver @NateSilver538

He estimates the current IFR (infection fatality rate) is 0.2%-0.4%. This has actually fallen some from ~1.0% in the early stages of the epidemic. An IFR of 0.2%-0.4% on 200K new cases per day would lead to 400-800 deaths per day. By comparison, the current 7-day average is 527. — PolitiTweet.org

Nate Silver @NateSilver538

Will all these new cases result in more deaths? Here's one view: The @youyanggu model estimates there are *actually* around 200K! new symptomatic COVID-19 infections per day in the US. We're only capturing around 1/4 of these through testing, he estimates. https://t.co/Qxu8rSPoV1 — PolitiTweet.org

Nate Silver @NateSilver538

The other metrics were bad, though. The second straight day above 50K new cases. And while there's a fair amount of testing, the positive test rate has climbed to 8% after having been around 4-5% at the low point several weeks ago. — PolitiTweet.org

Nate Silver @NateSilver538

† Last Thursday was weird because New Jersey decided to start counting probable deaths in its tally. Nothing wrong with that but it led to a large 1-day increase from deaths that mostly occurred some time ago. Ignoring that, deaths are roughly steady week-over-week (644 v. 647). — PolitiTweet.org

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking: Newly reported deaths Today: 644 Yesterday: 701 One week ago (6/25): 647 + 1854† Newly reported cases T: 53K Y: 53K 6/25: 39K Newly reported tests T: 635K Y: 621K 6/25: 638K Positive test rate T: 8.3% Y: 8.5% 6/25: 6.1% — PolitiTweet.org

Nate Silver @NateSilver538

RT @juruwolfe: This jobs day, @ameliatd, @Neil_Paine and I are here to throw a little cold water on a report with good news. It's great une… — PolitiTweet.org

Nate Silver @NateSilver538

There may not be a lot of enthusiasm for voting *for Biden*, but there's plenty of enthusiasm for voting *against Trump*, and it's likely that largely serves the same purpose. — PolitiTweet.org

Nate Silver @NateSilver538

@ForecasterEnten What you fail to point out is that Carter led in the most important factor: enthusiasm. — PolitiTweet.org

Nate Silver @NateSilver538

@mattyglesias I have no doubt (indoor) bars do very poorly by any sort of cost-benefit test. I also think at this point, we're working mostly from a credible hypothesis + intriguing anecdata—I haven't seen many attempts at rigorous analysis yet. — PolitiTweet.org

Nate Silver @NateSilver538

p.s. This above is not meant as any kind of a stance on re-opening policy. I do think states do need to think about what signals they send. Can you do things that nudge people away from the 3 C's? (closed spaces, crowded places, and close-contact settings)? — PolitiTweet.org

Nate Silver @NateSilver538

@BenjySarlin I know, but if states that began re-opening in late April and states that began re-opening in May/June both begin seeing a spike in mid-June, that suggests the story is more complicated than re-openings alone. — PolitiTweet.org

Nate Silver @NateSilver538

But the fact is the recent spikes don't actually line up all that well with re-openings, either timing-wise (since they began in April and have been gradual; not a sudden rush of re-openings on June 1) or geographically (e.g. also happening in CA, which has been quite cautious). — PolitiTweet.org

Nate Silver @NateSilver538

Sure, the signals that states send by what business are open/closed can matter. So, it's complicated. And also, what things are open/closed isn't an exogenous variable; it partly reflects political pressure, which in turn reflects people's subjective perceptions of safety. — PolitiTweet.org

Nate Silver @NateSilver538

Throughout this crisis, there's been evidence that people's behavior is only loosely tied to a state's formal degree of openness. So there should perhaps be more explanations rooted in "people thought it was safe/got bored/complacent"; maybe not as simple as "reopened too soon". — PolitiTweet.org

Nate Silver @NateSilver538

To some extent I just wonder if any state/metro that was largely spared the first time around is going to have problems. CA is one example, but Ohio—which by all accounts was proactive and smart and had relatively low numbers in March/April—is also seeing its metrics rise. https://t.co/WBWeaa9glV — PolitiTweet.org

Nate Silver @NateSilver538

Not a good situation as we enter July. Obviously the South, Southwest and California are the biggest issues right now. But also rising caseloads in some Midwestern states. https://t.co/LVX6BUuypq — PolitiTweet.org

Nate Silver @NateSilver538

UPDATED US daily numbers via @COVID19Tracking: Newly reported deaths Today: 697 Yesterday: 576 One week ago (6/24): 721 Newly reported cases T: 53K Y: 44K 6/24: 39K Newly reported tests T: 621K Y: 649K 6/24: 512K Positive test rate T: 8.5% Y: 6.8% 6/24: 7.6% — PolitiTweet.org

Nate Silver @NateSilver538

It looks like @COVID19Tracking has updated their numbers as California was late reporting ... and it turns out we did blow through our record for most cases in a day after all. Updated metrics in next tweet. — PolitiTweet.org

Nate Silver @NateSilver538

Deaths did decline again week over week, despite some individual states (notably, Arizona) posting high numbers. Obviously, it's uncertain how long that will continue. There's a bit of a whac-a-mole issue where a state will post OK numbers one day, and terrible ones the next day. — PolitiTweet.org

Nate Silver @NateSilver538

Not really any signs of a turnaround yet from the bad recent trends. In fact, this was the highest positive test rate (8.1%) since May 9. Maybe you can claim there's a plateau (5 of the past 6 days have been between 42K and 44K new cases) but even that seems generous for now. — PolitiTweet.org

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking: Newly reported deaths Today: 587 Yesterday: 576 One week ago (6/24): 721 Newly reported cases T: 43K Y: 44K 6/24: 39K Newly reported tests T: 534K Y: 649K 6/24: 512K Positive test rate T: 8.1% Y: 6.8% 6/24: 7.6% — PolitiTweet.org

Nate Silver @NateSilver538

RT @FiveThirtyEight: Why statistics don’t capture the full extent of the systemic bias in policing: https://t.co/KbZ0PHo1yP — PolitiTweet.org