Final Vote Simulations for AZ, FL, PA and TX
Updated 10/30/20: added Pennsylvania; revised topper language for clarity
Last month, Equis Research partnered with People for the American Way on a series of vote simulations to pinpoint the levels of Latino support and turnout Democrats need in order to win in key battleground states.
To help reporters and advocates contextualize the Latino vote in these last days of the election, we’ve compiled some recent data, as a resource, and plugged them into a simplified version of our simulations.
As we talk about the Latino vote, it’s worth noting that the story of the Tea Party and Trump eras is that of a surge in support and turnout from white voters propping up the fortunes of the GOP. To compensate, Black, AAPI and Latino voters have had to shore up Democrats at high, sometimes unsustainable levels, often in spite of lackluster investment or focus. To prevent a winning Democratic coalition, Trump has tried to shave off small numbers of non-white voters, while simultaneously attempting to suppress the vote of people of color at a mass scale.
As such, we don’t look at the Latino vote in isolation, but rather position it in comparison to shifts in white voting — to understand how much of a Latino crutch the Democrats need.
For each state below, we present four sets of data:
Historic range of Latino share of the vote
Estimates on the racial composition of the electorate can vary widely. We provide a range of estimates on the Latino share of votes cast, pulled from several sources for recent elections in the state:
- Exit polling
- Catalist estimates based on individual-level vote history records, precinct results, and pre-election data.
- Center for American Progress (CAP) analysis based on data from census, ACES, and CCES, along with CAP surveys.
- UCLA Latino Politics and Policy Initiative (LPPI) projections on turnout for 2020 based on survey data, exit poll data, and official vote histories.
Current Latino support for Biden
We present both the most recent Equis polling and an average of other recent state polls that break down their results by race/ethnicity or interview Latinos exclusively
Current white support for Biden
We averaged the numbers from all publicly available October polls that break down their results by race/ethnicity.
Simulation results
We held fixed Biden’s current level of Latinx support, as well as his performance with other racial/ethnic minorities in the electorate, to show what level of Latino turnout, as a share of all votes cast, he needs to win the state.
ARIZONA
1. Historic range of Latinx share of the vote
2. Current Latino support for Biden
According to Equis’ most recent poll of registered Latino voters in Arizona (9/24–10/1), the race stands at 65–26, which translates into 71% 2-way support for Biden.
This is slightly ahead of his average (68% 2-way) in other recent state polls that break down their results by race/ethnicity or interview Latinos exclusively.
Historical context: 71% support among Latinos would put Biden in the same territory as Sinema in 2018 (70% according to the exit polls), and well ahead of Clinton’s 66% (according to the 2016 exit polls).
3. Current white support for Biden
To estimate current white support for Biden, we averaged the numbers from all publicly available October polls that break down their results by race/ethnicity.
Average 2-way support = 46%
Historical context: 46% support among white voters puts Biden ahead of Clinton’s 43% (according to the exit polls) and on track to equal Sinema’s performance in 2018 (46% in the exit polls).
4. Arizona simulation results
If we hold fixed Biden’s current level of Latinx support, as well as his performance with other racial/ethnic minorities in the electorate, what level of Latino turnout, as a share of all votes cast, does he need to win the state?
If Biden holds on to his current level of white support (or increases it), our simulations show him winning the state even if Latinx turnout is significantly below 2016 levels. If white support drops by one point, we still show Biden winning if he can equal Clinton’s 2016 Latino turnout share. However if white support drops by 2 or more points, Hispanic vote share will need to be significantly above the 2016 level in order for Biden to win.
TEXAS
1. Historic range of Latino share of the vote
2. Current Latino support for Biden
According to Equis’ most recent poll of registered Latino voters in Texas (9/24–10/1), the race stands at 52–31, which translates into 63% 2-way support for Biden.
This is equal to Biden’s average 2-way support in other recent state polls that break down their results by race/ethnicity or interview Latinos exclusively.
Historical context: At 63%, Biden is running close to Clinton in 2016 (64% 2-way in the exit polls) and O’Rourke in 2018 (65% 2-way in the exit polls). But he has yet to catch up to the level of support for a generic Democratic candidate (69% 2-way) in Equis’ February tracking poll.
3. Current white support for Biden
To estimate current white support for Biden, we averaged the numbers from all publicly available October polls that break down their results by race/ethnicity.
Average 2-way support = 35%
Historical context: 35% support among white voters is just ahead of what O’Rourke achieved in 2018 (34% according to the exit polls) and well ahead of Clinton’s 27% in the 2016 exit polls.
4. Texas simulation results
If we hold fixed Biden’s current level of Latinx support (as well as his performance with other racial/ethnic minorities in the electorate) what level of Latinx turnout does he need to win the state?
*In the table below we report results for two different scenarios for combined Black & AAPI vote share (low = 13%, high = 16%) to take into account the importance of these groups in the Texas electorate.
In the high Black/AAPI turnout scenario:
At current white support, our simulations show Biden winning Texas if Latinx turnout share improves a few points over 2016. If he can gain a few points of white support, even somewhat lower Latinx turnout compared to 2016 is enough for the win. However, if he loses a few points of white support, our simulations show him losing unless Latinx vote share reaches truly unprecedented levels.
In the low Black/AAPI vote share scenario:
Even at current white support, our simulations show victory out of reach for Biden given any realistic expectation about the kind of increase in Latinx vote share we might see this year. However, if Biden can improve his white support by several points we show him winning with levels of Latinx turnout share just a few points higher than 2016.
FLORIDA
1. Historic range of Latino share of the vote
2. Current Latino support for Biden
According to Equis’ most recent poll of registered Latino voters in Florida (9/29–10/4), the race stands at 52–38, which translates into 58% 2-way support for Biden.
This is slightly ahead of his average (56% 2-way) in other recent state polls that break down their results by race/ethnicity or interview Latinos exclusively.
Historical context: 58% support among Latinos would put Biden closer to Obama’s performance in 2012 (61% 2-way according to exit polling) than to Clinton’s in 2016 (64% 2-way according to exit polling).
3. Current white support for Biden
To estimate current white support for Biden, we averaged the numbers from all publicly available October polls that break down their results by race/ethnicity.
Average 2-way support = 42%
Historical context: 42% support among white voters puts Biden ahead of Nelson’s 40% in 2018, Gillum’s 39% in 2018, and Clinton’s 33% in 2016 (all based on exit polls).
4. Florida simulation results
If we hold fixed Biden’s current level of Latino support (as well as his performance with other racial/ethnic minorities in the electorate) what level of Hispanic turnout, as a share of votes cast, does he need to win the state?
If Biden holds on to his current level of white support (or increases it), our simulations show him winning the state even if Hispanic turnout is significantly below 2016 levels. This also holds true if white support drops by one point. However if white support drops by 2 or more points, Biden will need Hispanic turnout above the 2016 level in order to win.
For other recent data on Florida, see our post on the Latin American vote.
PENNSYLVANIA
1. Historic range of Latino share of the vote
2. Current Latino support for Biden
According to Equis’ most recent poll of registered Latino voters in Pennsylvania (10/1–10/13), the race stands at 68–21, which translates into 76% 2-way support for Biden.
This is slightly ahead of Univision’s recent poll of Latinos in Pennsylvania (10/25), which has Biden at 73% (2-way).
Historical context: 76% support among Latinos puts Biden on track to match Clinton (77% 2-way according to the exit polls), and is well ahead of both Wolf (69% 2-way according to the exit polls) and Casey (71% 2-way according to the exit polls) in the 2018 midterms.
3. Current white support for Biden
To estimate current white support for Biden, we averaged the numbers from all publicly available polls taken in the past two weeks that break down their results by race/ethnicity.
Average 2-way support = 47%
Historical context: 47% support among white voters puts Biden well ahead of both Clinton (42% according to the exit polls) and Obama in 2012 (also 42% according to the exit polls), but behind Wolf’s 52% in 2018.
4. Pennsylvania simulation results
If we hold fixed Biden’s current level of Latinx support, as well as his performance with other racial/ethnic minorities in the electorate, what level of Latino turnout share does he need to win the state?
In the table below we report results for two different scenarios — one with high Black vote share (12.5%) and one in which Black vote share falls back to 2016 levels (10%) — in order to take into account the unique importance of this group in the Pennsylvania electorate.
A note on interpretation: 0% in this table indicates that our simulations show Biden has enough votes to win even if Latinx vote share drops to zero.
In the lower Black vote share scenario:
At his current level of white support (or above), our simulations show Biden winning the state regardless of the level of Latino vote share. White support would need to drop by 4 points or more before he finds himself in need of historically high Latino turnout in order to make up the difference. (That situation remains within the realm of possibility given that the average state polling error in 2016 was about 5 points according to FiveThirtyEight.)
In the higher Black vote share scenario:
If Black vote share increases over 2016, our simulations show Biden winning the state regardless of Latino vote share as long as white support does not drop by more than three points. However, a drop of 5 or more points puts Biden in a world where he needs Latino vote share to increase significantly over 2016 in order to have a chance of winning.
That’s all for now
Our full simulations are available here.
For additional insights in the last days of the election, follow us and our co-founders on Twitter for occasional updates: @EquisResearch, @stephanievalenc & @carlosodio.