Inflection Point

 President Biden (praise be!) would typically cite the phrase in speeches in discussion of the US' place in history and of the status of elections in our constitutional republic.  I think that, typically, the phrase was not really getting across to the public in that fashion, and especially to those swing voters he was trying to attract.  

An inflection point can be described as a change in the direction of a graphical curve. In a broader sense, it means a turning point. Biden believes that 2024's election is such a turning point.  I suggested something similar in my rant about the "precipice" at the beginning of the year.  This election can clearly be seen as the climax of something, this battle of our political norms vs. Trumpism. On behalf of all, even the Trumpists, there is a belief that this election will be decisive.  

I can clearly see an inflection point, and it is not on Election Day (at least not yet, it could still happen) but it already happened, on July 21, when Biden chose to announce he would end his campaign for re-election and suggest Vice-President Harris for the party's nomination. Before that, Biden's standing v. Trump was trending downward, slowly, but seemingly without any positive inflection in sight (first derivative negative, second derivative=0).  Since then, a month with a favorable sequence of events, all evidence shows steady improvement in the public's view of the Democrats' ticket.

Where does Kamala's trend go from here?   My estimate of the gain since 7/21 is 3-5%, from down 1-2% (Biden or Harris vs. Trump) to up 2-3%.  One change that contributed to the gain since 7/21 was the collapse of RFKJR's candidacy, which, along with Harris replacing Biden, has made this much more a head-to-head, two-person race. Roughly speaking, the popular vote now looks to be about 48-46 Harris-Walz over Trump, with some 3% insisting on another individual and 3% undecided.

Those 3% are probably more like "don't know", meaning that they don't know anything about it that they will admit to knowing.  Those particular folks may not end up voting, unless something really grabs their attention (like a war threat).  Regardless, some combination of changes will bring the numbers up to 100% total (with the 3% 'other').  There isn't more than that amount (3%) of downside to the current percentages I suggested, for either candidate, if that much.  Trump's popular vote floor I would say is 43%, and Harris' 45%, which is where Biden was when he stopped.  I will say, though, that a 3% downward move for either from the current level would absolutely lead to a landslide.  So, the time is one of vulnerability, avoiding harm, even as Harris may gain even a bit more during her campaign's post-convention afterglow. Another debate looms; another inflection point is likely before November, though that could be an acceleration (less likely), a reversal--let's hope not, a plateauing (most likely). Or maybe not, just a steady increase, all the way. Sounds good.

 

Graph of the Predictit market for Democrats winning Presidential election by 35-64 Electoral Votes, from May 26 to August 25, 2024.

Is National Popular Vote Irrelevant?

Clearly I don't think so, as that's what I've been discussing--using its estimator, the polling averages.

I often see it suggested as being meaningless by Republicans who are on the defensive because they would lose by that measure every time these days.  (2004 the only recent exception)  I would argue that, instead, it is a measure for which there is ample data and one that, broadly speaking, can be used to provide a key measure useful to predict outcome in the critical states. .     

If we look at the percentage vote for one of the parties in the Presidential elections within these states, it is not at all static.  It rises up and down with the relative fortunes of the party.  However, there is less movement relative to the overall popular vote, and even moreso if one looks at the order of these states on that measure.  

To take an example, North Carolina's popular vote percentage margin for Democratic presidential candidates, as compared to the overall US popular vote margin, resulted the following in the last four elections:  -7.0 (2008), -5.9 (2012), -5.8 (2016), and -5.9 (2020).  (Note:  2008, the worst of the four data points, was the one when the Democrats won the state!)  So, you could estimate that the US Democratic margin nationally should be above 6% to favor their chances. 

Let's look at that measure, which also shows some distinctive trends in individual states vs. the national voting, and some conclusions that may be drawn from it: 

 

State	2008	2012	2016	2020	Comment
Nevada	+5.2	+2.8	+0.3	-2.1	Negative trend; projects to -4
Michigan	+9.1	+5.6	-2.3	-1.7	Projects to -2
Pennsylvania	+3.0	+1.5	-3.1	-3.3	Projects to -3
Wisconsin	+6.6	+3.0	-2.9	-3.9	Projects to -3 or -4
Arizona	-15.8	-12.9	-5.6	-4.2	Improved rapidly; projects to -3
Georgia	-12.5	-11.7	-7.2	-4.3	Changing rapidly toward D; projects -3 to -1
Florida	-4.5	-3.0	-3.3	-7.7	Not out of reach if Dems +4; 2020 looks like an outlier

In terms of rank order of this measure, GA, AZ, and NC have consistently been lowest, MI at or near the highest, and PA right in the middle.  With all these swing states below the national average in 2020--a high turnout year like 2008, and like 2024 is likely to be, it is clear that those who say Harris-Walz will need a clear lead in the popular vote, like 3 or 4%, have the recent data on their side.  At the same time, though, a lead of 5-6% might lead to an Electoral Vote "landslide" of 100 EV or more, and much more if FL flips. 

This model doesn't take into account specific variations and their possible causes; I would suggest the outlook in Wisconsin looks more favorable than this because of the Walz effect.  Turnout nationally, and its effect on the US national popular vote, will be up for Democrats because New York and California are so critical for regaining control of the House and will get extra attention, which will boost the Democratic vote margin, without doing much in some of the swing states. 

With a 50-45 popular vote margin, which I think is quite likely, this projects to a 297-219 lead for Harris-Walz, with Nevada and North Carolina too close to call and Georgia, once again, somehow on the Democrats' side.  On the other hand, if it finishes 49-47,by this model Harris would lose Pennsylvania, where Democratic turnout totally matters, and most likely the election as well.  The margins are fine. 

Unfortunately, though exit polls might help, the popular vote takes days to assume its final shape.  So we are left to lean upon the unwieldy support of the polling averages. That is, if we don't want to go by the even more inconsistent measure of polling averages in the several states.