THREAD: I’m still very interested in Trump’s remarkable success in Miami-Dade and Rio Grande Valley/Texas border. The explanations given in the press...
for his gains include the appeal of Trump’s hyper-masculine persona among Latino males, fear of Biden’s position on oil and gas, fear of Biden’s “socialism”, Democrats "forgetting" how to “talk” to Latinos, Democrats taking the Hispanic vote for granted, etc.
If reasons given are accurate, wouldn’t we expect a grand defection from the Biden ticket similar to that of the Reagan Democrats in Macomb County, Michigan in 1984? There, Mondale dropped to 97K votes from Carter’s 120K in 1980. Meanwhile Reagan increased from 154K to 194K.
However, what happened in Miami-Dade and RGV is very different. Biden basically matched or barely exceeded Clinton’s numbers. There was no grand defection. Instead what we saw was a spectacular increase in new voters, and all of them seemingly voted for Trump.
In county after county, we find Biden meeting Clinton’s vote levels and then stopping cold, with Trump receiving almost 100% of a very sizable number of new voters.
I don't mean every new voter literally voted for Trump, but in every county Trump’s increased vote total was roughly equal to the number of new voters.
I don’t know if this is common historically, and this would be an interesting project. I also don’t know how to interpret this pattern. A benign explanation is that the Trump team put together stunningly effective GOTV effort. I’m not sure if that stretches credulity or not.
Below I aggregate just six counties in the RGV (you can find many more at the link from my first pass at this issue) and then I look at Miami Dade. The six counties are all on the border: Cameron, Hidalgo, Starr, Zapata, Webb and Maverick. https://twitter.com/DavidManel/status/1323921182622732288?s=20
Comparing ’16-’20 - 6 border counties
Clinton 231870-Biden 252186 (+9919)
’16 Trump 97130-’20 Trump 181794 (+84664)
Tot. voters 350891-438148 (+87257 new voters)
Registered Voters - 729977-823416 (+93439 newly registered)
Trump share of new voters /newly registered = 97% / 90.6%
As we can see, Biden didn’t lose votes. He did just as well as Clinton. There was not a massive shift away from the Democratic candidate. Rather, what seemed to happen is new voters rushed to Trump. We see the exact same thing in Miami-Dade...
Miami-Dade
Clinton 624146 - Biden 617647 (-6499)
’16 Trump 333999 - ’20 Trump 532734 (+198735)
Tot. Votes 998608 - 1165883 (+167278)
Registered Voters 1379248 - 1563572 (+184324)
Trump share of New Voters / Newly registered = 118% / 108%
Again, Biden didn’t lose many votes. The Dems didn’t seem to suffer defections. Trump just seemed to net thousands and thousands of new voters.

Now, six counties plus Miami-Dade
Six counties plus Miami-Dade
+254535 New Voters
+277763 New registered voters
+283399 New Trump votes.
+3420 Democratic voters
Trump share of new voters / new registered” = 111% / 102%
Before the election, Kellyanne Conway was quoted as saying: “The person who coined the term, ‘hidden, undercover Trump voter in 2016,’ there are even more of them, and they're even more committed now, and they're going to surprise you as to who they are this time.”
The Trump team did spring a big surprise on election day. They did indeed have thousands of undercover voters. And, it is surprising who they were.
Again, it doesn’t appear that Biden lost thousands of voters in Miami/RGV, instead, the Trump campaign appears to have found tens of thousands of sympathetic voters who were undercover and successfully got them to polls.
A first look suggests a similar pattern in some counties in South Carolina, New Mexico and North Carolina. Very often they are poorer counties with substantial non-white populations. I may look at those more closely in the future.
-Mercifully, The End
Webb County, Texas
Clinton 42307 - Biden 41820
'16 Trump 12947 - '20 Trump 25898
+12464 Turnout; +12951 Trump Vote
You can follow @DavidManel.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: