Polarization is Overwhelming

Large events fail to move the political needle

(Photo by David Todd McCarty on Unsplash)

In the wake of this week’s political events in the United States (an assassination attempt, a four-day long convention by the Republicans, and the increasing likelihood of Joe Biden withdrawing from the 2024 presidential race, I have the following thoughts.

Political Polarization is Overwhelming

The first, and overarching, thought I have is that political polarization has created an electorate that is more rigid, and more impervious to being moved by events, than any I have ever seen.

In a “normal” political cycle, these kinds of events, and certainly the more “shock”-type of events like a near-miss assassination attempt, and the slow-motion withdrawal, in July¹, of a presidential candidate, would have some rather noticeable effect on the polls. And yet neither of these events, one of which is, of course, still ongoing, have had more than a very minor impact on the polling numbers of the two candidates.

While it’s true that polls in July often aren’t dispositive of election results in November, it’s also true that they normally are not this rigidly set in stone, either. In 2024, by contrast, it appears likely that the baseline polls will not move very much at all between now and November, no matter what happens, because really big stuff has already been happening, without substantial impact on the polls.

The reason for is that almost all of the electorate is fully decided on this race, barring a tiny number of undecided voters. Most everyone else is very “dug in” to their position.

Therefore, while two-thirds of rank and file democrats are said to want Biden to withdraw from the race, almost all of these have also said that they would vote for Biden/Harris over Trump/Vance, if Biden does stay in the race. What that means is that while Biden’s presence in the race does not impact these voters, they are concerned that it will impact the votes of others who are less committed than they are, and therefore cost the Democratic party the 2024 presidential election.

Why is this?

Polarization means that There Are Very Few Undecided Voters

That brings me to my second thought about the past week: the tremendous degree of polarization and rigidity in the current electorate means that this presidential election will be effectively decided by a tiny number of undecided voters living in a handful of states.

Under the current electoral map, with Biden as the Democratic candidate, this means effectively three states — Pennsylvania, Michigan and Wisconsin — because the remainder of “battleground states” have wider state polling spreads, and appear unlikely to swing to Biden unless one of these three “blue wall” states does. It’s possible that a replacement for Biden would result in a bigger map, but this seems unlikely due, again, to the rigidity created by the high degree of polarization — that is, already-committed voters in the other “battleground” states are not likely to switch their allegiances no matter what happens.

Note that this doesn’t mean that the three blue wall states are filled with undecided voters. In fact, there isn’t much evidence that there are any more undecided voters in these states than elsewhere, it’s simply that in these states the split between the two polarized groups is close enough to mean that the small group (likely 150,000 voters, or less, in total, out of a national electorate of ~160 million²) of undecided voters in each of them will likely determine the outcome of each of these states.

Now, that does not mean that the national polling, and the national trend, are therefore irrelevant, as some claim. To the contrary, the national polling trend is definitely one factor that will be influential on the decision-making of at least some of these voters. But in this rigidly polarized electorate, those national polls are unlikely to move very much at all, regardless of what happens between now and November.

Critically, almost all of these undecided voters will make their decisions, as they typically do, late in the election cycle — that is, after mid-September. This means that polling numbers now, both nationally and in the impacted states, are of only limited relevance for understanding how this small, yet critical, group of voters will behave during “crunch time”.

Rigid Polarization Means the Entire Election is About the Late Phase

That fact brings me to my third thought, which is that since the national polls will move very little due to the rigidity of most of the current electorate, the key contest in this election is which “team” can both appeal to a significant number of the remaining very small number of “available undecided votes” in a handful of critical states while at the same time mobilizing its core base.

That is, since almost all of the electorate is already very firmly decided, it is priority number one for both sides to get those voters to the polls — whoever “wins” the get-out-the-vote (GOTV) contest will have a substantial leg up in the ultimate result of the election, regardless of the polls.

It also means that outsized gains in terms of the result can be achieved by winning a significant proportion, but not necessarily a strong majority, of the remaining undecided votes in these three states, if that can be achieved without undermining the GOTV efforts relating to your base voters — that is, if you can convince these very few as-yet undecided voters to vote for your candidate without saying and doing things that alienate your own base, and thereby work to depress your GOTV efforts.

To some degree this is sequencing. That is, the remaining time in the election cycle can be broadly divided into three periods of roughly equal length: (1) the period between now and the end of the Democratic National Convention on August 22 (i.e., five weeks from today), (2) the period between the end of the conventions and late September (say the four weeks following the end of the DNC) and (3) the final stretch run of the six weeks between then and election day.

Importantly, most of the direct outreach to the remaining undecided voters will happen after the DNC, and to a large degree during the last six weeks when so many of these voters make their final decisions. And it is also the final six-week period when each side will concentrate their GOTV efforts increasingly intensively as election day gets closer.

Polarization Means Replacing Biden is Less Risky than it Normally Would Be

And this brings me to my final thought on the past week’s events: the Democrats have an opening here, in the weeks before the DNC, to change their candidate without this having a massively negative impact on their chances in November.

The reason for that is the outsized impact of the extremely polarized electorate. While in a normal election cycle, this kind of late substitution would be extremely disruptive, in a cycle where there are relatively few undecided voters, and which will therefore be decided by a combination of GOTV and some portion of the small number of undecided voters breaking your way in a few states very late in the election cycle, a substitution at this stage of the game is likely to have minimal negative impact unless it is done in a way that is perceived to be unfair.

This is because the voters who are already committed to voting Democratic are extremely unlikely to become less committed if Biden steps back — in fact, with 2/3 of Democratic voters wanting Biden to withdraw, the substitution will likely have the positive effect of uplifting the Democratic GOTV effort over what it would have been had Biden had stayed in the race, which was likely to continue to dampen the enthusiasm of at least some Democratic voters. And because the substitution is happening well before the small number of undecided voters generally make their decisions, it still leaves plenty of time for the replacement ticket to make its case to these undecideds well in advance of their decision time-frame. So the two critical elements of the election, tactically, from the Democratic perspective — GOTV and some outreach to the small number of undecided voters — are either positively impacted or at worst neutral.

Of course this doesn’t mean that replacing Biden will be risk-free.

If the replacement is done in a way that alienates core Democratic constituencies — in other words, if there is a perception among core elements of the Democratic electoral coalition that the replacement process has been done in a way that is either generally unfair or specifically unfavorable to their part of the coalition — the process could create more problems than it solves. And if the process results in an active campaign of Democrat vs. Democrat running up to and through the DNC, it’s at least possible that this process will be more damaging than galvanizing or energizing (as at least some Democrats appear to be hoping it would be).

For this reason, the process that the Democrats choose, and how it actually flows, will have a large impact on the efficacy and impact of the replacement. A poorly executed, or poorly perceived, process could do irreparable damage to the party’s prospects in November, but a well-designed and well-executed one could have minimal negative effects while having significant uplift in terms of GOTV efforts, all of which would marginally help the Democrats.

Polarization Means that the Republicans Run the Risk of Peaking Early in an Election Cycle that will be Decided Very Late

Finally, what does all of that mean to the Republicans?

Well, on the most basic level I think the overall strategy of the Republicans doesn’t change much, either way. The Republicans are also going to be focused on GOTV for their own base in the critical states, coupled with outreach to undecideds to grab as many of those few remaining votes that remain up for grabs as they can. So the basic thrust of what they are doing will not change.

What will change, of course, is the dynamic of personalities, and that is where the greatest risk lies for the Republicans, I think. The current optical/visceral dynamic of a relatively sclerotic Biden competing with a relatively (compared to Biden) energetic Trump will change a great deal if Trump’s Democratic opponent is in their 40s or 50s. Likely this would eliminate or at the very least greatly diminish any advantage the Republicans have in that optical/visceral area currently.

Yet the importance of this should not be overstated, either — not, again, in an election season which is as rigidly polarized as this one is. For while the issue of visceral optics has dominated the discussions of the chattering class in the United States since the first presidential debate took place on June 27, most of the concerns raised relating to Biden’s performance in that debate pertained not to committed voters (most of whom said they would vote for Biden regardless), but to the impact of this on GOTV and undecided outreach. In other words, Democratic media and opinion writers mounted a campaign to replace Biden largely because of fears that Biden remaining the candidate would depress GOTV efforts somewhat, and make outreach to undecided voters more challenging.

I think that is likely, but the key is that those impacts are not reflected in the polling. The number of undecided voters in the critical states is so small that it has only a marginal effect on the polling until election day in most cases, and in many cases is therefore subject to some doubt in terms of the actual election day vote in those states (resulting in election day surprises). Similarly GOTV efforts are poorly captured in polls as well, because these efforts reach their greatest intensity, and have their greatest effect, at the very end of the election cycle, which is extremely late in terms of the available polls — in effect, you generally learn the effectiveness of each party’s GOTV efforts only in the final vote on election day.

In summary, then, because of a very abnormally high degree of rigid polarization in the voter base (1) the Republicans only got a minimal boost from some events that in a normal year would have led to a significant boost in their favor while (2) the Democrats will be able to “get away with” changing horses fairly late in the game because most of their voters are already committed, either way, and the few undecided voters are less affected because they don’t decide until late September anyway.

In short, polarization means the polls don’t move much no matter what happens. The parties are instead fighting over who can best motivate their respective bases to vote while at the same time successfully harvesting enough of the tiny number of remaining available votes — both of which occur in the final phases of the election cycle. The rest of the polling, news cycling, and the like, is mostly noise, with minimal impact on the election result.

1. Lyndon Baines Johnson withdrew from the 1968 presidential race, but he did so on March 31 — a full seven months before the election in November. Biden, by contrast, has not yet withdrawn, although is reported in media as being likely to do so in the coming days, with only a little over three months remaining in the 2024 campaign.

2. In the 2020 presidential election, a total of 158,429,631 votes were cast. The FEC’s official tally is here.