I didn’t plan an update until Saturday...but some news from the north seemed like it warranted an update. We got some counts in for CA-21, TJ Cox (D) vs Valadao...and these numbers aren’t what I expected to see. They’re better for us that I thought...like considerably.
A CA-21 refresher course (for those new to my CA reports)
The district sits in four counties, and every county is currently working through their uncounted ballots. Here’s a brief description of these portions of the district, as well as the percent of the House race’s vote that came from each county (in other words, 33% means that a third of the total vote for the CA-21 race came from that county):
Kern (37%) Very blue, young, hispanic area that went Clinton 70/30.1
Fresno (34.5%) A mix of open rural areas, but a lot of suburban population...and some urban. This portion of the district went for Clinton 55/45, but has traditionally voted GOP in the House race.
Kings/Tulare* (28.5%) Pretty conservative area that went Trump 56/44.
*Lumped together because the Tulare portion is tiny, and it matches the Kings area demographically.
1. Note on Kern: some commenters have pointed out that Kern County is NOT liberal and votes pretty Republican. This is true, but I am referring to the portion of Kern that sits inside the 21st district. The CA Sec of State provides voting data by district, and we can see that this part of the county votes very blue. Kern as a whole went 40/53 for Trump, but the portion of Kern in the 21st voted 70/29 for Clinton. This is why we are pinning our hopes on the liberalness of a pretty conservative county.
What has happened so far...
Our updated count over time chart now looks like this:
CA-21 |
Cox (D) |
Valadao |
Margin |
|
6-Nov |
33844 |
39475 |
-5631 |
Election Day Count |
11-Nov |
41739 |
44138 |
-2399 |
Kern counts half their ballots |
14-Nov |
46279 |
48260 |
-1981 |
Fresno/Kings counts half
|
On election night, this looked like a strong GOP win. The AP and networks called it.
Then last Friday, Kern County blew through half of their uncounted ballots, processing most of their mail ballots (but no provisional ballots yet). The Kern County results were pretty stunning...they broke much harder D than Kern had on election night. The race still looked like a long shot, but not a GOP blowout.
What happened today...
Here’s what I expected to occur next, but it did not. I expected Kings to issue an update with a big GOP margin, and for Fresno to issue an update with a modest GOP margin. I expected Cox to lose ground today. That would match what Election Day totals indicated, and what polls told us. But look at what happened...Fresno and Kings counted half of their outstanding ballots and Cox PICKED UP 418 VOTES.
I did not see that coming. At all.
Valadao really needed to see his margins increase when these counties came in. That shores him up against the ground we expect he’ll lose in the final Kern votes. But if he’s losing ground in both Kern and Fresno...this race could get really interesting.
There are several unknowns that make it hard to predict if this might flip in our favor:
- We know how many ballots are left to be counted in each County, but we don’t know how many of those ballots have the CA21 race on them. Fresno, Kern and Tulare all sit in multiple House districts.
- We know how Election Day ballots broke, and we know how mail ballots are breaking so far. We do not know how provisional ballots will break. They tend to skew left because they tend to be younger voters, minorities, people who move around, etc.
My Rough Math
Is this race in reach??? This is either a fun one to analyze, or torture. (If you’re either candidate...probably torture.) So here’s what I think.
These counties have processed about 55% of the uncounted ballots. Of these, 21,220 votes belonged in this district. If the same proportion of uncounted ballots belong to the 21st, this indicates that there are possibly 17,361 votes outstanding in this race.
Can there really be this many ballots left?
I’ve just spent an hour going insane over charts of past election data, lol, and I think there could be this many ballots outstanding in this race. Why?
Every one of the four counties is reflecting total votes in this election that are running 83-85% of the 2016 total. (Meaning if a county reported 100,000 ballots in 2016, they are reporting 85k cast in 2018).
To get to 85% of the total CA-21 votes in 2016, this race needs 18,000 more votes. So this is a second data point that suggests the estimate is reasonable.
1. The known outstanding ballots suggest another 17.5k votes based on the count and proportions so far.
2. The turnout projections based on a known number of ballots cast in 2016 and 2018 also suggest another 18k for this race.
Yeah...I think maybe there are this many ballots left to count in this race.
What does Cox need to win this one?
This is why tonight’s count update is so exciting — we just passed a threshold where all Cox needs to do is maintain his margin on the late count ballots to pull ahead.
In other words, so far in ballots counted since Election Day, he’s won about 58%. If there are 17,361 more votes, he needs 56% — then he’d close the gap by 2,083 votes and leap ahead.
In other words...ELECTION NIGHT CONTINUES AND THIS ONE COULD BE CLOSE!
Risk factors — why we may fall short
I would categorize this race as “lean GOP”, down from “likely GOP” this weekend and on election night “certain GOP if you work for the AP and don’t understand Californian elections” (ha). We might lose if:
- We are running out the count. If there are only 15k ballots left...or 12k...or less...we may run out of ballots before Cox can pull ahead.
- Maybe the ballots remaining are in bad areas for us geographically. Maybe the precincts in Fresno farm country drove their ballot trays to the county office and those ended up in the back of the line, and the rural mail votes still need to be counted.
But why we might win
Both factors above can be reversed and turn in our favor.
One, maybe there are more ballots that the data suggests. This was one of the highest profile races in any of the four counties, after all. Maybe a disproportionate number of the remaining votes belong to this race.
Second, maybe the remaining votes are more urban. This sees more likely then their being rural, honestly. Maybe the Fresno ballots left are all from the cold, hard streets of downtown “Eff No.” (For the record, I don’t know any actual Fresno slang.)
But the main reason we might have a shot here — we are moving into provisional ballot counting. That’s about 75% of the remaining ballots. These usually skew distinctly left compared to mail ballots and regular in-person ballots. Will they be one point better for us? Ten? Worse? I have no idea. If they’re very close to the mail ballots counted so far, man is this one going to be close.
But here’s the thing. The national youth vote broke records in this cycle — 31% turnout, beating the previous record of 21%. And the Hispanic vote soared. Experts are still tabulating...the DNC says they think hispanic turnout was up 174% over 2014.
So sheesh, fellow Kossacks.
We have a pile of provisional ballots that we know traditionally skew young and minority. We have a district that is 71% hispanic. We know hispanic voters tend to be younger. We know that hispanic voters went 70/30 for Dems in Congressional races nationwide.We know the youth vote and the hispanic vote skyrocketed in this cycle. And we know that Cox was working hard to engage these voters aggressively with a strong GOTV effort on Election Day.
Are those provisional ballot trays full of votes by young hispanic urban/suburban voters? Man, I sure hope so. And they very well could be.
As Rachel Maddow says, “watch this space.”