Politico's Steven Shepard tells us that President Trump's chances of winning "hinge on a polling screw-up way worse than 2016." At The New York Times, The Upshot runs the numbers and says that Biden will win Wisconsin, Michigan, Pennsylvania, Arizona, Florida, and Georgia even if the polls are off by as much as they were in 2016. At The Atlantic, Derek Thompson insists that "the 2020 election won’t be a 2016 sequel" because, among other things, pollsters have learned to weight their surveys by education, which makes an undercount of non-college-educated white Trump voters less likely; because there are fewer undecided voters than in 2016; and because the polls have been stable for months.
That's the good news. Here's the dishearening news. First, from Politico's Marc Caputo and Matt Dixon:
Democrats are sounding the alarm about weak voter turnout rates in Florida’s biggest county, Miami-Dade, where a strong Republican showing is endangering Joe Biden’s chances in the nation’s biggest swing state.And Bloomberg's Tyler Pager has more bad news:
... Democrats are turning out at lower rates than Republicans and at lower rates than at this point in 2016, when Hillary Clinton won by 29 percentage points here and still lost the state to Donald Trump.
One particular area of concern is the relative share of ballots cast by young voters of color and less-reliable Democratic voters....
To date, Republicans have turned out 59 percent of their voters in Miami-Dade and Democrats have turned out 53 percent, a 6-point margin. That’s twice the margin Republicans had at this point in 2016.
Among Hispanic voters, who make up nearly 70 percent of the county’s population, the deficit is even bigger — 9 points.
In Arizona, two-thirds of Latino registered voters have not yet cast a ballot. In Florida, half of Latino and Black registered voters have not yet voted but more than half of White voters have cast ballots, according to data from Catalist, a Democratic data firm. In Pennsylvania, nearly 75% of registered Black voters have not yet voted, the data shows.But Tom Bonier of the Democratic data firm TargetSmart says we shouldn't worry:
The firm’s analysis of early vote numbers also show a surge of non-college educated White voters, who largely back President Donald Trump, compared to voters of color, who overwhelmingly support Biden.
Meanwhile non-college white vote share is down 4.4 pts. So what we're seeing is massive increases in black and latino turnout, but not quite as massive as the increases among white college educated and Asian voters.
— Tom Bonier (@tbonier) October 31, 2020
And in other states:
As I've noted before, Asian voters in Texas are already at 130% of their final '16 turnout. White college educated voters have increased their vote share by 4.6 pts while the non-college white share has decreased by 5.3%. Black and Latino voters are close to surpassing '16 levels pic.twitter.com/X6qnZWqGFZ
— Tom Bonier (@tbonier) October 30, 2020
7,956,095 voters under the age of 30 have already voted in this election. That's 2,344,721 more votes than that age group cast in total in the 2016 early vote.
— Tom Bonier (@tbonier) October 29, 2020
2,311,993 of these voters had never voted before they cast a ballot in this election.
One more note to provide context for the incredible turnout surge we're seeing in Arizona - more registered Democrats over the age of 65 have already voted in Arizona than voted in the entirety of the 2016 general election. They're at 112% of 2016's total vote, and rising. https://t.co/5UPmXxqKpw
— Tom Bonier (@tbonier) October 29, 2020
So are Democrats on track to win, or should we be bracing ourselves for another Trump upset?
I suspect that the polling margin really is too large for Trump to overcome. It's possible that Trump will win some close states where Biden has a polling lead -- I assume he'll win Florida, because Democrats routinely screw up there -- and there's a risk that the president might be one or two states away from a legitimate Electoral College win, at which point mail-ballot problems and partisan decisions by Republican judges (which is what we should always call them) could tip the election to Trump.
Or, more likely, Biden will win, but by a narrower margin than the polls predict. The betting market PredictIt is projecting a Biden electoral vote total of 305, a number that sometimes drops to 290 when bettors decide that Trump will win North Carolina. That seems much more likely to me than a Biden Electoral College win in the mid-300s.
If that's what happens, it will be because Biden didn't find a way to persuade both college-educated white voters and non-white voters to turn out in massive numbers for him. He's spent much of the campaign talking about the pandemic (which is helping him with older voters) and touting endorsements from Republicans (which might be helping him with moderate white suburbanites) -- but he hasn't done the same level of outreach to non-white voters (or, for that matter, young voters). Maybe it takes a Barack Obama to put together a coalition broad enough for a victory as large as the one he had in 2008, and resilient enough for the win he had in 2012.
But it shouldn't be that hard. Republicans are awful, and Democrats have more popular ideas. Democrats need to find a way to hold and expand on their base. I think Biden has done enough to win, but he might not have done enough to get the big victory Democrats deserve.