The first 2020 Democratic debates, or Why the Trump campaign should be nervous…

With the first Democratic presidential nomination debates scheduled for Wednesday, June 26, 2019 and Thursday, June 27, 2019, here is an updated assessment of the relative position of the 25 declared candidates; former United States House of Representatives member (“Representative”) Joe Sestak of Pennsylvania declared his candidacy June 23, 2019). For details on how I calculate NSW-WAPA (national-and-state-weighted weighted-adjusted polling average), please see here[1].

Photograph from here.

The values in Table 1 were calculated from 93 national polls, 13 Iowa Caucuses polls, 14 New Hampshire Primary polls, 3 Nevada Caucuses polls, 12 South Carolina Primary polls and 41 polls of all subsequent nominating contests (n=16), for a total of 176 public polls released since January 1, 2019; no public polls included Sestak. Italics indicate that candidate has not yet been included in any public polls from that state; I assign a value of 0 to candidates excluded from a poll. Going forward, I will no longer present polling averages for former Georgia State House Minority Leader Stacey Abrams, who appears more likely to run again for governor in 2022 than for president in 2020.

Table 1: NSW-WAPA for declared 2020 Democratic presidential nomination candidates

Were the Democratic National Committee using these data to determine eligibility for the these first two debates, Montana Governor Steve Bullock and former United States Senator (“Senator”) Mike Gravel of Alaska would have been included, while New York City Mayor Bill de Blasio and author Marianne Williamson would not have been included. Still, these are differences of fractions of a 1/10 of a percentage point—a coin flip would be just as effective.

While fuller analysis will appear in the next regular monthly update, Massachusetts Senator Elizabeth Warren’s rise from 8.5% to 11.3% in three weeks is easily the largest shift in support.

Enjoy the debates!

**********

I will also update analyses of polling for matchups between President Donald J. Trump and hypothetical Democratic opponents in 2020, both nationally and in various states, in the next regular monthly update. Overall, state-level polling suggests that Democrats would win the national popular vote by between 2.6 (excluding former Vice President Joe Biden and Vermont Senator Bernie Sanders) and 5.9 (including Biden and Sanders) percentage points (“points”).

However, I will make some observations about how putatively undecided voters may cast their votes in 2020, at least nationally. Two pollsters, Emerson College and Harris X, have dominated much of the public polling of potential 2020 match-ups. This is one reason I aggregate polls two ways: 1) treating all polls as independent events, regardless of pollster, and 2) taking the average of individual pollster averages; I then present the average of the two averages.

Emerson College, rated B+ by FiveThirtyEight.com, has no undecided voters in its hypothetical national presidential polls; every respondent is urged to select either the listed 2020 Democratic presidential nominee or the Republican, President Donald J. Trump. HarrisX (C+), by contrast, does not appear to force such a choice at all, generally having >20% undecided in its national presidential polls. Given that both pollsters routinely sample ~1,000 registered voters (making for a more apples-to-apples comparison), averages can be compared across the two pollsters for the eight candidates (Biden; Sanders; Warren; South Bend, Indiana Mayor Pete Buttigieg; California Senator Kamala Harris; former Texas Representative Beto O’Rourkel New Jersey Senator Cory Booker; Minnesota Senator Amy Klobuchar) tested at least once in 2019 by both pollsters.

Table 2: Emerson College and HarrisX polling averages in hypothetical 2020 match-ups vs. Donald Trump for 8 Democratic presidential nomination candidates

Table 2 above contains the simple (i.e., not adjusted for time) averages of the percentages in these polls; weighting percentages by time did not materially affect the analysis.

The most striking result is that all eight tested Democratic presidential nominees defeat Trump when undecided voters are forced to decide between them, by an average of 51.8 to 48.2%; the margin widens slightly to 52.0-48.0% if the single February polls testing Booker and Klobuchar are excluded. By contrast, the HarrisX polls show Trump defeating these eight Democrats by an average 38.1-33.9% (only Biden and Sanders, the two best-known candidates, would hypothetically prevail). But these polls have an average of 28.0% undecided between the named Democrat and Trump (or would choose a third-party candidate); I estimate these voters would break roughly 7-4 in favor of the Democratic nominee. And if you exclude Biden and Sanders, the average percentage undecided increases to 30.0, and I estimate they would break nearly 2-1 (65.7-34.3%) for the Democrat. Overall, that is an average shift toward the Democrats of 7.8 percentage points.

Applying the 2-1 distribution of undecided voters to the other Democratic nomination candidates tested at least twice by HarrisX in 2019 produces the following changes:

• Entrepreneur Andrew Yang: -10.0 points to +1.6 points
• Hawaii Representative Tulsi Gabbard: -11.8 points to even
• Former HUD Secretary Julián Castro: -10.4 points to even
• New York Senator Kirsten Gillibrand: -4.0 points to +4.7 points
• Former Maryland Representative John Delaney: -10.4 points to +1.0 points
• Washington Governor Jay Inslee: -13.6 points to -1.6 points
• Former Colorado Governor John Hickenlooper:  -10.8 points to +0.8 points
• Ohio Representative Tim Ryan: -10.0 points to +1.0 points
• California Representative Eric Swalwell: -13.0 points to -1.4 points
• Gravel: -16.0 points to -1.2 points
• Williamson: -11.6 points to -0.2 points
• Miramar, Florida Mayor Wayne Messam:  -17.3 points to -4.8 points
• Massachusetts Representative Seth Moulton:  -15.3 points to -2.8 points

Thus, while the HarrisX polls have Trump beating these 13 potential Democratic presidential nominees by an average landslide margin of 11.9 points (12.5 points excluding Gillibrand), the distribution of undecided voters implied by the Emerson College polls brings these 13 Democrats to within 0.2 points of Trump, on average, essentially a tie. This is an average shift of an astonishing 11.7 percentage points in favor of the Democrats.

One other point about Table 2 is that the Trump percentages are remarkably consistent, ranging between 45.5 and 49.7 in the Emerson College polls and between 36.3 and 39.2 in the HarrisX polls. And, generally speaking, the better-known (and the higher the current ranking among Democrats) the proposed Democratic nominee, the lower the Trump percentage. This suggests that the president has not expanded his support much—if at all—beyond the 45.9% of the total national vote for president he received in 2016 (which translates to 48.9% of the votes cast only for him and 2016 Democratic presidential nominee Hillary Clinton) and/or that voters dissatisfied with the current administration seem prepared to cast their 2020 presidential ballots for nearly any Democrat.

The giant flashing neon sign caveat (besides the fact that there will be other general election candidates for president besides the Democratic nominee and Trump), of course, is that the 2020 presidential election is still more than 16 months hence; polls this early are of questionable value. Nonetheless, it should gravely concern the Trump campaign that when forced to decide, voters currently break heavily for every proposed 2020 Democratic presidential nominee.

Until next time…

[1] Essentially, polls are weighted within areal units (nation, state) by days to the nominating contest and pollster quality to form a unit-specific average, then a weighted average is taken across Iowa (weight=5), New Hampshire (5), Nevada (4), South Carolina (4), the time-weighted average of all subsequent contests (2) and nationwide.