Georgia senate seats runoff(s) megathread

[ProgressiveModerate]

What's going to be really annoying is if one side wins both of these, they're going to act like they have a mandate to either fully persue Biden's agenda or entirely block Biden's agenda, even though basically half of the people who voted voted against their party.

A split verdict would be hilarious; it would be interesting to see how the parties react, and it would just be a crazy end to a crazy election cycle.

If the verdict is split, then Susan Collins winning would have been a godsend. Had she lost, Republicans would have needed to win both Georgia races, now they only need one.

Also the Democrats nominating Cal Cunningham over Jeff Jackson or Anthony Fox in North Carolina would have amounted to be a godsend for the Republicans as well.

Yes. If anything, this was a much more winnable race for Democrats than Maine.

Also Bill Nelson failing to adequately reach out to Hispanic voters.

If Republicans lose the Senate, it'll be because of NH-Sen 2016, and MT and WV Sen 2018, not to mention Martha McSally. Both sides really had their fair share of missed opportunities and getting lucky and it all kind of cancels out.

Here were all the close calls that allowed both parties to get where they are right now.

[ProgressiveModerate]

We could, but we probably won’t.

[ProgressiveModerate]

VBM accepted
Nov GE: 1.32M (53.0% white, 30.3% black) ---> +22.7 white
...
1/1: 928K (54.0% white, 31.9% black) —> +22.1 white


IN PERSON EARLY votes
Nov GE: 2.69M (58.2% white, 26.4% black) ---> +31.8 white
...
1/1: 2.07M (56.7% white, 30.3% black) —> +26.4 white

I'm using super wide General Assumptions here.... but, if you suppose that the mail-in net votes is a wash (same Net votes as the General) for Dems ... and just look at the gain in net votes for Dems via In-Person EV (strictly using the gain in Black/White margin.... since race is the most consistent determination of votes in Georgia), you get:

536K Votes for Dems (Nov race breakdown, applied to 2.07M runoff votes)
626K Votes for Dems (Runoff race breakdown, applied to 2.07M runoff votes)

which is a +90K (Net) Gain for Dems.  So if the Election Day numbers were exactly the same as November:

Ossoff would win by 2,000 votes (since he lost by 88K votes in Nov).
---------
To me, this is in the ballpark of the best case scenario for Republicans- one reason of which, is it allows for the same number of people turning out on Election Day (as that total goes down, R's net less votes on Election Day, assuming similar R/D turnout model).

Based on the Nov breakdown (contained in the post ^2 above^ --- this hypo scenario does seem plausible as Republican's best case scenario, in which they would have an even shot of pulling out a narrow victory.  

Granted you would have to read many of the unknowns into R's favor. But this also requires the same level of turnout on Election Day as the General (this is where I think they will fall short, even in this scenario.  I do think the margin split will be similar to ED in Nov, And that the Trump rally/attention will cause turnout to be higher than expected next Tues, but- just not as high as Election Day).

I think part of the issue is you have so many overlapping unknowns that the range of plausible outcomes is still pretty wide; if the statistics we had were more finite, that would be true, but we're making a lot of assumptions to get those numbers. A +5, even +7/8 result for either side still can't be completely written off, though the median outcome is that they will be close. I think of the range of plausible outcomes in these races as like a t distribution instead of a normal distribution.

[ProgressiveModerate]

Yep.

What’s killing us is the lack of credible polling

I just don’t take Emerson, surveyusa, trafalgar, etc seriously

Gravis is ok I guess, but even they have pooped the bed at times in recent cycles

Where are fox, cnn, abc/WAPO, Monmouth, Sienna, etc?

Ah yes, the brilliant posters who respectively had Biden up by 16 nationally, up 17 in Wisconsin, Van Drew trailing by 5, and double digits undecided a couple days before the election.

No one had Biden up by 16 nationally. And had that one poll that had Biden up 17 was an outlier. The very pollster said so.

For the most part, the polls were spot on. They correctly predicted 48 states. The margins were off, but that’s called the margin of error.

The truth is, you can’t be 100% accurate when it comes to polls.

oh come on.

The polls were terrible. Every fair minded person admits that.

Yeah polls were pretty bad, especially in swing states. Ironically, the places where polling held up the best were safe blue states + GA. The polling map was eerily similar to the Trump 2016 vs Biden 2020 map, which makes me think polls do a better job at picking up D turnout than R turnout

[ProgressiveModerate]

One underdiscussed topic is what do folks think is the percent of the black vote that Ossoff / Warnock will get? Polls seemed to show Biden got 88-90% of the black vote, which is actually on the low end for a Democrat in a deep southern state like GA. I think there are some reasons to think Ossoff / Warnock will outperform this by a few percentage points.

If black voters make up 3% more of the electorate vs. November and give Ossoff / Warnock a 5% greater margin, that actually brings down the white share that Ossoff / Warnock need to receive from Biden's 30% to 26% or so.

My guess is 91%. While Biden did only probably win about 89% of the black vote in Georgia, the final 11% was probably something like 8-9% Trump and 2-3% 3rd party. These 2-3% 3rd party may have less of a reason to show up in a binary runoff but if they did, they would likely skew D. Furthermore, historically, black Democrats in the South have had a stronger showing in off years, so my gut says they would win about 91 or 92% of the black vote.

[ProgressiveModerate]

https://twitter.com/Nate_Cohn/status/1345765912037777411
https://twitter.com/Nate_Cohn/status/1345766439135948801
https://twitter.com/Nate_Cohn/status/1345767256358342657
https://twitter.com/Nate_Cohn/status/1345768351579123713
https://twitter.com/Nate_Cohn/status/1345769192637800452

I generally agree, but I think 350k seems on the high end of Ossoff's current lead. As I've said before, i think my model which shows him with a 213k is a slight undercount, and in reality it's prolly closer to 250k or so, but 350k gives the benefit of the doubt to Ossoff.

I think the one thing we can all agree on is that Perdue will likely have to net AT LEAST as many votes as he did on e-day in November.

[ProgressiveModerate]

Nate's characterization seems a bit misleading as he says: "Of course, the strong Dem turnout means there are more GOPers left now"

well yes i guess technically, but hasn't the whole "Dems are cannibalizing their E-day vote with the early vote" thing been debunked? So there are still a ton of possible Dem voters left to vote as well?

Yes, it has been mostly debunked, but it's unclear how many of the Republicans who early voted in Nov who have not yet voted in these runoffs will end up actually voting on e-day; it could be nearly all of them or maybe only 60%; we just don't know.

So I guess while Democrats aren't canabalizing their own vote, the GOP could be saving votes.

[ProgressiveModerate]

Counties with the highest drop-off in early voting compared to November; the lightest colors mean only a 10% drop off while dark red indicates up to 40%.

[ProgressiveModerate]

Update:

Democrat's estimated lead has grown slightly as expected; now up to + 7.4. They also widened the margin in terms of raw votes from 213k last time my model was updated to 224k. Not a huge increase but in a close race every vote matters. Again, based upon my methodology, I think 224k is a slight undercount of their true lead, maybe by about 40k or so, but that's speculation on my part.

If votes broke as they did in the GE, Republicans would need a turnout of about 1.1 million on e-day to win. The votes needed on e-day can change based upon how heavily the e-day vote breaks for the GOP.



As always, these numbers are estimates and not exact

[ProgressiveModerate]

I wonder if it would be best for the GOP to purposely make Loefeller have a lot of gaffes to increase their chances at saving the Senate by making every voter willing a split a ticket vote Warnock - Perdue. She would rile up the Trump base who show up to vote for her and Perdue while their at it, and Perdue runs as a more sensible canidate who garners crossover support. It's too late to change what as been done but their closing messages could have impact on who does and doesn't show up on e-day and how they vote.

[ProgressiveModerate]

Just saw a new (at least to me) DSCC ad during the NFL broadcast that showed Loeffler and Perdue as wealthy, out-of-touch elites and encouraged Republicans to not vote on Tuesday.  Haven't seen that suggestion before.

I'm glad they didn't use this attack too early on because if they did by now it would render useless. Obviously, the ad will likely change very few people's plans but a few people here and there can make all the difference

[ProgressiveModerate]

Another interesting thought I saw on twitter today:

A dem win in GA on tuesday will actually increase the chances the repubs take the house in 2022, because a 51-50 senate increases chances of dems passing a law that could end up being unpopular in certain purple areas.

I think in general, it just increases the range of outcomes for 2022; it gives Democrats the chance to be extremely unpopular but also extremely popular.

[ProgressiveModerate]

Seems like a lot isn't going well for the GOP right now, and they may very well still win these races, but if they lose both and therefore the Senate, they likely would've been very avoidable losses.

[ProgressiveModerate]

Sorry for the 3rd post, but does anyone know if we're getting an exit poll on Tuesday?

The NBC results page would suggest so:

https://www.nbcnews.com/politics/2020-elections/georgia-senate-runoff-results

Oh nice. Excited for that one.

Sorry if this was asked before, but what is the order in which votes are counted again? I forget what the trajectory was in Nov (besides Biden increasing towards the end)

Generally in person to early to absentee, and more urban counties tend to take a lot longer to count; the perfect formula for a pretty extreme Republican counting bias.

[ProgressiveModerate]

Sorry for the 3rd post, but does anyone know if we're getting an exit poll on Tuesday?

The NBC results page would suggest so:

https://www.nbcnews.com/politics/2020-elections/georgia-senate-runoff-results

Oh nice. Excited for that one.

Sorry if this was asked before, but what is the order in which votes are counted again? I forget what the trajectory was in Nov (besides Biden increasing towards the end)

Generally in person to early to absentee, and more urban counties tend to take a lot longer to count; the perfect formula for a pretty extreme Republican counting bias.

ah okay, so the initial totals will likely be at least somewhat GOP. though i guess we can always compare to how the first results were back in Nov too to see how it's going

Actually, in Nov, the first few % were very heavily Biden but that's just kind of random based on which counties start first.

[ProgressiveModerate]

I’ll play along with the fantasy that Dems win...,

What about PR statehood ?

(DC too... but that’ll be struck down as unconstitutional)

Why?

Presumably some fear-mongering about how the 5 conservatives will never allow the Democrats to pass anything ever again.

Quoting from an article on the matter:

The Founding Fathers wrote it into the Constitution. Article I, Section 8 provides explicitly for a national capital that would not be part of a state nor treated as a state, but rather a unique enclave under the exclusive authority of Congress — a neutral “district” in which representatives of all the states could meet on an equal footing to conduct the nation’s business.

Also DC was once part of Maryland and technically Marylanders could still reclaim it (like Virginia did in the 1840s), it could be a significant argument against statehood.

Except Maryland DOES NOT want to reaclaim it's portion of DC at all. I've been to the DC area quite a bit and DC doesn't want to join Maryland and vice-versa.

[ProgressiveModerate]

Small update:

Democrats estimated advantage up to +7.5 or a 228k vote lead. To make that up on e-day, if votes broke as they did in the GE, they would need a turnout of about 1.1 million.

Here's a county map, note I did increase the log scale on the gradient:



As I always say, these numbers are estimates based on available data and historical precedent. If I had to guess, 228k is on the lower end of possible Dem leads considering the data and how my methodology works, but that's the number my model outputs

[ProgressiveModerate]

IMO, these runoffs will be an interesting test of partisanship vs negative partisanship. Democrats are showing to vote FOR a Biden presidency whereas Republicans are showing up to vote against a Biden presidency.

[ProgressiveModerate]

IMO, these runoffs will be an interesting test of partisanship vs negative partisanship. Democrats are showing to vote FOR a Biden presidency whereas Republicans are showing up to vote against a Biden presidency.

You could argue that most Democrats are voting *against* Mitch McConnell at this point.

True, but in general, it seems like Warnock and Ossoff are the ones running a campaign for something rather than against something. Honestly, McConnell hasn't been talked about as much as I expected in these races.

[ProgressiveModerate]

https://twitter.com/Acyn/status/1344846289427599362

Glad it's Mitis Touch who ran the ad instead of some firm with an obviously D name. Democrats were smart to save these attacks till the very last second since Loeffler and Perdue now have very little time to refute them. Ultimately, I doubt that a bunch of Trump voters staying home will happen but a few here and a few there can make all the difference in a close race.

With that being sad, it's kind of sad how nasty these campaigns have gotten but what goes around comes around I guess.

[ProgressiveModerate]

Biden is speaking at a rally for Warnock and Ossoff in Atlanta:

https://twitter.com/daveweigel/status/1346212750620585986

538 doesn't think that the argument necessary help Democrats

https://fivethirtyeight.com/features/will-the-debate-over-2000-stimulus-checks-help-democrats-in-georgia/

It'll probably be close to a net wash, but I really can't see how it helps the GOP.

[ProgressiveModerate]

isnt there a good chance we wont even know who won the senate races tommorow?

I would say there’s a 50% chance we “know” but only a 10% chance the media calls it

[ProgressiveModerate]

My dad was going on today about how black farmers in South Georgia are going to make Perdue win handily.

I'll bite.  Why?

Maybe his dad saw this tweet

https://twitter.com/ewerickson/status/1345889577899356162?s=21

What evidence do we actually have that Perdue did well with black farmers in the South?

[ProgressiveModerate]

Somehow, mentally, I feel like Warnock - Perdue split is inevitable, even though my brain knows that isn't true.

[ProgressiveModerate]

My prediction for how tomorrow will go:

-There are a lot of reports about the in person vote throughout the day that a bunch of people post in this thread and latch on to too much

-Polls close, and no one has the clear advantage for the first hour

-Northern GA which has lagged in turnout ends up having medicore turnout for the GOP, but they’ve sustained their margins

-Atlanta takes a long time to count, but while the margins in some counties are underwhelming turnout is strong

-The “black belt” is notably strong for Democrats

-Democrats perform “eh” in a lot of the midsized cities.

-South rural white GA is just fine for the GOP; nothing notable

At first everyone says Perdue and Loefeller are favored, only to then say Warnock and Osoff are favored, and just willy nilly switch back and forth until someone is clearly favored. A bunch of notorious Atlas partisans will continue to make case for their canidates until it’s clear as day they have no chance of winning

We know by Wednesday morning who will win, even if the networks haven’t officially called the races

I currently think it’ll be Perdue + 4 and Loefeller + 2, which would probably make this a very boring and disappointing senate cycle, at least from the perspective of a Democrat, though that’s my gut, my brain and Excel say otherwise

While the races are tossup, I think whoever wins wins by slightly more than most of Atlas suspects