In 2016, Hillary Clinton’s polling lead in Pennsylvania slipped away almost imperceptibly . t never fully disappeared, making her loss by 44,292 votes a shock to prognosticators. Collectively, more than 20 distinct polling operations surveyed a Clinton/Trump matchup in the state dozens of times. From the first poll in August 2015 to the last, released on Election Day, Trump led in only three, all of which were outliers at the time they were released.
Yet, the double-digit lead Clinton typically enjoyed in late-July and early August became a high single digit lead by late-August and early-September. After that, a the polls bounced in a narrow band until then-FBI director James Comey sent his infamous letter to Congress on October 28 announcing a reopening of the Clinton email case. Of the ten surveys i conducted entirely after the Comey letter, Clinton’s lead was three points or less in eight of them.
A May 2017 autopsy of the election by Nate Silver of FiveThirtyEight concluded that the Comey letter cost Clinton the election. Undecidedvoters broke heavily against her, Silver concluded. Still, on the morning of Election Day, FiveThirtyEight gave Trump a mere 11 percent chance of winning the popular vote nationwide and an 18 percent chance of winning the election. The final HuffPost forecast for Pennsylvania had Clinton leading narrowly, 45.5 to 41.4 percent.
As of the morning of September 29, 2020, the average of polls at FiveThirtyEighthas Joe Biden leading Trump in Pennsylvania by a similar 49.9 to 45.5 percent margin. Is Biden’s lead safe?
The short is answer is no. As Comey’s letter demonstrated, unexpected events late in the campaign can swing a lot of votes, and Tuesday night brings the first of four presidential and vice-presidential debates. Moreover, Clinton had a polling lead and still lost.
Yet, there are some reasons to believe that Biden’s lead is more sturdier than Clinton’s. Most obviously, Trump is no longer a hypothetical president but an incumbent with a four-year record voters can evaluate. Current surveys, including the bullish-for-Biden Washington Post-ABC News poll, show Pennsylvanians are unimpressed.
Biden’s support stands at 54 percent to Trump’s 45 percent among the Keystone State’s likely voters and 54 percent to 44 percent among its registered voters…
…Trump’s overall approval rating in the state among registered voters is 43 percent positive and 55 percent negative, with 49 percent saying they disapprove “strongly.”
An incumbent facing 49 percent strong disapproval is in real trouble.
Another troubling sign for the president is that his base isn’t quite as loyal as its #MAGA reputation. In the Washington Post-ABC poll, Pennsylvanians who voted for Trump in 2016 prefer him by 92 percent to 8 percent. But that pales in comparison to the transference of support from Clinton to Biden, which registers at 98 percent to one. The discrepancy between those two results accounts for almost the entirety of Biden’s nine-point lead with likely voters.
A closer look shows that Trump has lost support among key groups and in key regions. White Pennsylvanians without four-year college degrees went for Trump by more than 30 points in 2016 but his lead with them is now down to seventeen. Relatedly, Trump is severely underperforming in Western Pennsylvania, the region I pegged as most responsible for his surprise victory in my Washington Monthly piece How to Win Rural Voters Without Losing Liberal Values. Excluding Pittsburgh’s Allegheny County, Trump carried this region four years ago by 29 points. The Washington Post-ABC poll has him leading there by three.
Trump surely hopes he can bring some of these Western Pennsylvania non-college white voters many of who are hunters back into the fold as well as more affluent suburbanites. But they know what kind of president he is. He’s no longer just a guy they know from watching The Apprentice.
There was nothing in the 2016 election comparable to the COVID-19 pandemic, and Pennsylvanians prefer Biden 54 percent to 40 percent to tackle the issue going forward. Yet, there has been one benefit from the health crisis for the president: the Republicans have continued to knock on doors and register voters while the Democrats have not. This helps explain why the Republican Party has added 198,000 registered voters in the state since 2016 compared to only 29,000 for the Democrats.
The Washington Post-ABC News survey also gives Trump an enthusiasm advantage over Biden, which should have some impact on turnout. On the other hand, Clinton understandably had difficulty exciting the black community to the same degree as Barack Obama, but Biden relied on black support to win the nomination and chose Kamala Harris as his running mate. If higher black turnout results, it will help offset any enthusiasm gap.
Biden’s biggest advantage over Clinton, however, is that he’s much less unpopular. In 2016, voters who disliked both candidates went heavily for Trump, but the opposite outcome has been predicted in 2020 surveys going back to the Spring. There are many fewer undecided voters today than four years ago, but the chances are they will break heavily for Biden as they often do for challengers. This is especially true because the Democrats succeeded in having the Green Party candidate removed from the ballot in the Keystone State and other battlegrounds. In 2016, Jill Stein’s 49,941 votes exceeded the 44,292 vote separating Clinton and Trump.
In the aftermath of the 2016 debacle, Pennsylvania Democrats were split on the best approach for regaining the state in 2020. One camp believed the party needed to focus on turning out the urban/suburban base, and the other thought the losses in rural areas had to be held in check. So far, it appears that Biden is doing both at once. His strength in the affluent Philadelphia suburbs is remarkable but it also reflects nationwide Democratic strength among college educated whites:
Biden is strongest in southeastern Pennsylvania. The Post-ABC poll finds the former vice president leading Trump by more than 2 to 1 in the Philadelphia area, including its populous suburbs of Chester, Delaware and Montgomery counties. Clinton won these counties together by a 14-point margin in 2016, but they have grown more Democratic since then.
Biden’s average lead in polls (4.4 percent) in the Keystone State is not significantly larger than the lead (4.1 percent) Clinton enjoyed at the end of the 2016 race, but he looks to be in better shape.
Of course, if Biden wins Pennsylvanis we may not know that on Election Night. According to the Washington Post-ABC News poll, nearly a third of voters intend to vote either by mail or at an early voting site. They skew heavily Democratic. The mail ballots will not be tallied in advance, and there is litigation about how late these ballots can be received and still count. This means that when results first come in, Biden may be trailing but have a huge lead in the yet-to-be-counted vote, some of which may not even have arrived yet in the mail. Expect Trump to cry foul and try to stop the counting. Unlike 2016 that might be his best chance of winning.
I believe FiveThirtyEight had the Election Day 2016 odds of Trump winning the election at 29%, not 18%. Love your stuff, Booman.
When I feel inclined, I pray it doesn’t happen again. But then, I am not at all sure given his proclivity for escaping any sort of accountability or karma. This time though the world has seen something nasty about him just before the election. Unfortunately, the rubes will only believe what they choose to believe. The fight this time is going to be after election night. And we are dealing with a criminal.
Excellent analysis, as usual. Thanks.
I know a community organizer who has as much a claim to making the ACA a possibility as anyone else I know of; she did the organizing in Massachusetts that helped make “Romneycare” possible. For 2-3 years, any time anything happened related to healthcare reform in the Commonwealth, she made sure that leaders from her organization showed up and were in “the room where it happened”. One of her teachings was that real, substantive change *only* happens when you organize, and people show up “consistently and persistently”.
That’s how Biden beats Trump in Pennsylvania: people in Black neighborhoods in Philadelphia persuading a few more of their neighbors to vote; suburban middle-class moms making phone call after phone call to persuade their (formerly) Republican neighbors; Democratic stalwarts in northeastern PA and in rural counties in the central and western part of the commonwealth risking wrath and embarrassment to talk their neighbors into voting against the guy they voted for four years ago. Keep pounding the rock.
I agree with showing up. We had thousands of Trumpers driving around our ring and honking horns and making an awful traffic mess. We ran into a bunch of them as they were breaking up. Most had flags. Boy this could cause trouble on the other side of the election, maybe before.
5
I cannot get OVER him saying bad things happen in Philadelphia!
First of all — FUCK YOU!
SECOND OF ALL —
https://uploads.disquscdn.com/images/a01389907440a18005c4b062bef280feb2fb0e4c1e98ce0a4f8bec9eee816ec3.jpg?w=600&h=393
THIRD OF ALL — FUCK YOU!
Fourth — it’s a swing state Hillary lost largely because of lack of enthusiasm esp. in Philadelphia. What a bizarre rallying cry to give to people who are Trump adverse. And Steelers fans won’t care, they’re not as invested in the resentment culture as Philly. I say that lovingly.
Thanks for your comment. As to your last point, read the post-election analysis Martin did after 2016. Clinton’s vote margins coming out of Philadelphia and its suburbs *exceeded* Obama’s four years earlier (iirc). Where she lost Pennsylvania was in those dozens of rural and small town counties where Trump swung 10 – 20% of the voters his way, piling up huge margins of victory there.
All other things being equal, if Biden swings back even just half of those rural voters, he’ll win the state. Sensibly, he and the Democrats aren’t putting all their eggs in their one basket. At least from afar it looks like they’re also trying to run up their margins and vote totals in Philly and Pittsburgh, and in the Philly suburbs (which makes sense; it’s basically the local equivalent of Howard Dean’s “50 state strategy”).
she did very well in the suburbs but did not get as many Black city voters as Obama did.
https://www.buzzfeednews.com/article/johntemplon/this-chart-shows-philadelphia-black-voters-stayed-home-costi
People who sat it out and generally reliable voting blocks (or voting blocks the Dem party took for granted turnout-wise bc of Obama) certainly hurt her in PA. It was a multifaceted catastrophe. I still think it’s just beyond stupid to insult the biggest city in a swing state.
5
According to the link you posted, Blumenthal’s trend line had Clinton at +5 to +6 throughout that entire period. She had one double-digit polling advantage in a Marist poll conducted in April, but mostly polled around that average.
We should also point out that polling firms have adjusted their turnout models to reflect increased participation from non-college educated white folks (the ‘missing whites’ from Sean Trende’s infamous 2012 piece). So a +4 today is quite different from what it was four years ago.
Otherwise agree.
One other important thing is the poll Booman quoted has Biden OVER 50, which is much different to have a winning majority and be leading by 5 than be at 45% and lead by 5.
Oh, I feel fairly confident (from afar) that significantly more Pennsylvanians prefer Biden over Trumpolini, and even that many more of them will color little ovals by Biden’s name than by the dictator’s. The problem is that the Repub party is not trying to get Trump to “win” PA via democratic means (i.e more colored ovals); they are seeking to win via litigation means—by getting Trumpified “conservative” federal courts to order that millions of mailed in ballots arrived “too late” to be counted, or some such nonsense.
This issue is in the lap of Roberts’ Repubs right now. The PA Repub party has asked the Trump Court to stay the recent PA Supreme Court ruling laying out the ground rules for Pennsylvania’s Pandemic Election. The basis is that “conservatives” are now asking Roberts’ Repubs to create out of whole cloth a new “doctrine” that the US Supreme Court assert that IT has the power to review a state supreme court’s rulings on state election law, something the Supreme Court has never done before. This would stop supreme courts in Blue States from protecting the rights of voters in their state, and make all their rulings subject to reversal by Roberts’ Repubs. It will be a small miracle if Roberts’ Repub don’t take up the invitation.
A fascist movement doesn’t care too much about polls and voting and whether its assertion of power is legitimate. It looks to non-democratic methods to retain power, such as burning down the Reichstag. Trump (and McConnell’s minions) are now looking for their Reichstag Fire. It almost certainly will involve Trump’s democratically-illegitimate Supreme Court “conservative” majority, which I have been calling Roberts’ Repubs for years now. They will be adding another illegitimate Trumpite “justice” in the next 30 days, as we all know.
As we saw last night, it will take a supreme effort to remove this cancer from the WH. It has metastasized, and too many institutions have been thoroughly poisoned and corrupted by the gathering fascist movement.
It is unlikely that the polls are once again underestimating Trump voters in this election. They were pretty close in 2018 on average, although slightly bullish on Democrats in the Midwest, by a couple of points. This is not enough for Trump. People also talk about the enthusiasm gap, but conflate enthusiasm for voting itself with enthusiasm for voting FOR a candidate. Many might be not so enthusiastic voting FOR Biden, but would swim across a crocidile infested river to vote AGAINST Trump. Last night’s debate did nothing to help Trump on that account.
But we likely don’t have to wait for election day to figure out who will win this election (ignoring stealing, cheating, etc.): we have good evidence from early voting and mail-in voting numbers. Compare those numbers, with the percentage of people who say they are voting early or voting by mail, split by party. This gives us a good estimate of the voter turnout, which gets more and more accurate as we get closer to election day. I strongly believe that we will be able to call the election for Biden before election day, if the current trends continue. This happened in ’08 on the OpenLeft blog, and it will happen again this year – that’s my take.
4.5
5