The daily news is full of presidential election stories that harken back to 2016.  One theme is that 2016 polling was wrong.  It showed Clinton ahead of Trump in 2016, but Trump won.  Biden may be leading now, but will Trump come back from behind and win again with a smaller percentage of the popular vote?  There are two differences this year that make this outcome less likely.  Biden is more popular now than Clinton was then, and the polling data is less volatile. Let’s look at the data, and then speculate on the next 60 days.

Figure 1 shows polling results for Clinton and Trump in 2016.  The data is from RealClearPolitics.com, and the two dashed lines are the rolling average of the latest ten polls from May through the election.  Clinton led most of the time, but Trump occasionally got close.  The last 10-day rolling average saw Clinton at 47.1% and Trump at 43.7%.  Clinton was ahead by 3.4%, but the gap was closing.  The final vote count was Clinton with 48.2% and Trump with 46.1%.  This 2.1% difference was well within the polling error band.  The national popular vote polls were not wrong, but as discussed below the Electoral College vote can run counter to the popular vote when popular vote margins are small.

Figure 1 2016 Polling

Notable is how much each candidate’s polling numbers changed over time.  From the beginning of May through early September, Clinton’s numbers spanned 5.9% while Trump’s spanned 7.1%.  Figure 2 overlays Biden’s and Trump’s polling averages for this year.  The span for Biden is 4.0% while the span for Trump this year is just 3.7%, nearly half.  Much less variability so far this year, especially for Trump.  Unlike 2016, Trump has never closed the gap.  He was within a tenth of a point of Clinton in late July of 2016, but the closest that he has come this year is 4.1%.

Figure 2 Comparison of 2016 and 2020 Polling

On average, Clinton led Trump by 4.4% through early September in 2016, while Biden has led Trump by an average of 7.5%.   The difference is due to Biden having higher averages than Clinton, 49.1% versus 45.3%.  Trump’s averages are nearly the same, 41.6% this year versus 40.9% in 2016.

Figure 3 depicts the polling edge that Biden and Clinton have had on Trump, and the figure reinforces that in 2020, Biden has a larger and steadier lead over Trump than Clinton did in 2016.

Figure 3 Difference Between Democratic and Republican Candidate Polling

What about the next 60 days?  Also shown in Figures 2 and 3 are several events in 2016 that shaped the election.  The three debates coincided with Clinton separating herself from Trump.  It didn’t help Trump that the Access Hollywood tape came out in the middle of the debates.  What appears to have really hurt Clinton though was Comey’s letter to Congress in late October.  Clinton’s averages crashed and Trump’s soared after that.  The 2020 debates are still ahead of us, and who knows if any material disclosures will occur.

Is Biden’s 7.5% average popular vote lead enough to win the Electoral College?  Going back to 1864 (here), Figure 4 plots the Electoral College vote percentage versus the popular vote margin of victory.  No one has lost in the Electoral College with a 7.5% margin of victory in the popular vote, but there are four instances (1876, 1888, 2000 and 2016) where small popular vote margins of victory led to an Electoral College loss.  If the race this year tightens and Biden’s lead drops by half or more, there is an historical 1-in-3 chance that he would lose the Electoral College.  While possible due to a poor debate performance or some Black Swan event, the magnitude of Biden’s lead and the relative stability of the polling numbers this year argues against that happening.

Figure 4 Electoral College Vote Percentage versus Popular Vote Margin of Victory for Presidential Elections Since 1864