Pennsylvania is ground zero for the 2020 presidential election. Currently, Nate Silver’s FiveThirtyEight gives Donald Trump an 11 percent chance of winning the election, but during a Sunday appearance on ABC’s This Week, he said the president should be favored to win if he carries the Keystone State.

Supporters of Joe Biden can comfort themselves that Silver gives the Democratic nominee an 85 percent chance of carrying Pennsylvania. Still, Biden’s 50.2 to 45.5 percent average of polls lead is barely outside the margin of error for most surveys.

Four years ago, Hillary Clinton was favored to win in Pennsylvania and lost the state by just over 44,000 votes. With this history and this context, all eyes will be on Pennsylvania on Election Night, but how will we know if Biden has prevented Trump from pulling off a second straight upset?

If you want to know whether Biden will win Pennsylvania in 2020, you first have to understand why Clinton lost it in 2016. It wasn’t because the state’s two big cities, with their large minority populations, didn’t turn out for her. Her net advantage (108,137 votes) from Pittsburgh’s Allegheny County was 17,489 votes larger than Barack Obama’s had been in 2012, which more than canceled out the 17,062 fewer votes she netted out of Philadelphia. Considering that Obama won that election comfortably by a 52 to 47 percent margin, this should have been enough.

It also wasn’t a result of underperformance in Philadelphia’s suburban counties. She matched Obama’s 2012 returns in Bucks and Delaware counties and greatly surpassed them in Montgomery and Chester counties.

What cost Clinton was a disastrous collapse in rural counties and a drop of support in Democratic counties in the Northeast portion of the state. There were 21 counties where Clinton performed at least ten percentage points worse than Barack Obama in 2012. If Trump is going to reprise his stunning 2016 upset, this is where he’ll need to do it.

It’ll be tough. Most of the Republican gains from 2016 disappeared in the 2018 gubernatorial election when incumbent Democratic Governor Tom Wolf defeated Republican challenger Scott Wagner by a whopping 57.8 to 40.7 percent margin. Trump will have to count on a special bond with these voters.

Since mail ballots in Pennsylvania cannot be counted before Election Day and can be received up to three days late if postmarked by November 3, the Election Night totals will be incomplete and heavily tilted to in-person votes. Still, I’ll be examining these 21 counties to see if Trump has held his 2016 coalition together and has a real chance to carry the state.

I’ve sorted counties into regions to see trends in how they moved from 2012 to 2016, and then from 2016 to 2018.

In 1986, while working on Bob Casey Sr.’s successful run for Pennsylvania governor, James Carville famously quipped, “Between Paoli (in the Philly suburbs) and Penn Hills (in the Pittsburgh suburbs), Pennsylvania is Alabama without the blacks.” This central region of the state is alternatively referred to as Pennsyltucky (in a nod to its Appalachian culture) or “the T” in reference to its shape. The top of the ‘T’ runs along the border with New York State, and it’s here where the Democrats have seen the most slippage.

[For this analysis, all percentages are of the two-party vote (i.e., it excludes third parties) to make it easier to compare different election cycles.]

In the chart below, you see four “Top T” counties where Clinton did at least ten points worse than Obama’s 2012 performance. Trump held Clinton under 30 percent in all of them, and while Tom Wolf clawed back most of the lost support in 2018, he still couldn’t match the 2012 results. Susquehanna County has over a thousand active natural gas wells, making it a prime area for Trump’s pro-fracking pitch.

TOP T2016 PCT2016 vs. 20122018 PCT2018 vs. 2016NET GAIN/LOSS (2012-2018)

In the stem of “the T,” Clinton fared worst in an area I refer to here as the “Central West.” Except for Cambria County (home of Johnstown), these are very low population areas. Obama didn’t do well in this region, but he earned more than 40 percent in two counties. Clinton couldn’t get even a third of the vote in any of them.

CENTRAL WEST2016 PCT2016 vs. 20122018 PCT2018 vs. 2016NET GAIN/LOSS (2012-2018)

Unlike in the Top T counties, however, Tom Wolf fully erased the Republicans’ gains from 2012 in the Central West.

Outside of “the T” proper, Trump made significant gains in two northwestern and two southwestern counties. In 2018, Tom Wolf not only reversed those gains but improved on Obama’s 2012 numbers in three of the four, making them all competitive.

NORTHWEST2016 PCT2016 vs. 20122018 PCT2018 vs. 2016NET GAIN/LOSS (2012-2018)
SOUTHWEST2016 PCT2016 vs. 20122018 PCT2018 vs. 2016NET GAIN/LOSS (2012-2018)

Wolf also did a good job of patching up the damage Trump did in the Northeast. He restored Lackawanna County (home to Biden’s childhood home, Scranton) as a Democratic stronghold and flipped Luzerne back into the blue column.

NORTHEAST2016 PCT2016 vs. 20122018 PCT2018 vs. 2016NET GAIN/LOSS (2012-2018)

A straightforward way to apply these numbers on Election Night: Biden getting less than 30 percent of the two-party vote in any counties on the list spells trouble for the Democrat. The closer Biden comes to Wolf’s numbers, the better, and in the unlikely event that he exceeds Wolf’s performance in a county or two, that’s a sure sign of trouble for Trump.

The counties in the T and the west are important, but the Northeast has significantly more population, including some Manhattan commuter bedroom communities.  As the tables show, the Northeast also has more historic swing. Luzerne and Schuylkill counties will be particularly telling, as the first can flip from blue to red and back, and the second’s 33-point oscillation between 2016 and 2018 was the biggest in the state.

On the whole, Trump will probably need to do better in these 21 counties than in 2016 to offset the Republicans’ erosion in the Philly suburbs. Trump’s challenge becomes immediately apparent when looking at the comparison between the 2016 and 2018 elections.

PHILLY SUBURBS2016 PCT2016 vs. 20122018 PCT2018 vs. 2016NET GAIN/LOSS (2012-2018)

There’s a good chance Biden will also bank more votes out of Pittsburgh and Philadelphia than Clinton managed in 2016, which would make Trump’s task all that much harder.

CITY COUNTY2016 PCT2016 vs. 20122018 PCT2018 vs. 2016NET GAIN/LOSS (2012-2018)

If Biden can come close to Wolf’s numbers in the cities and suburbs, it will be light’s out for Trump’s chances of carrying the state. But if black turnout is disappointing or Latino voters show a little more support for Trump and Biden’s numbers there aren’t any better than Clinton’s from 2016, then it will come down to how the president is faring in his strongholds.

We’ll have to wait until all the mail ballots are counted to get a complete picture, and that may take a week or more. Yet, focusing on these specific counties will help us get a sense of where the final result is headed.

Martin Longman

Martin Longman is the web editor for the Washington Monthly. See all his writing at