In recent days, I am increasingly being pressed by South Africans about whether or not Donald Trump will win the 2020 election – and who will be his Democratic Party opponent in that election. My protestations on both of these points, that it is still a way to early to say, keeps bumping into those who demand firm clarity about the future. That pushes me to explain “why” it is way too early to predict the result of this race, at least for now.

Of course, if in the next 14 months there is an actual shooting war with Iran, a violent confrontation with North Korea, a real economic slump, or some other horror, predictions about Donald Trump’s future will become that much easier, but we are not in those circumstances yet. Thankfully. Still, there are some things we can say to help understand how to read the tea leaves.

First of all, we need to put this whole elections thing into a perspective deeper than just this week’s “breaking news”. Exactly one year from this week, it will be the initial week of September 2020. That week includes the US holiday of Labor Day, the effective end of the summer vacation season, and the start of school terms. And that is the traditional start of the general election campaign for the 2020 campaign as the ordinary citizen begins to tune in more fully to their upcoming choice.

While that traditional starting point has become a bit less fixed in time than it used to be in the past – in part due to the era of the perpetual campaign and fundraising cycles – this time around there is also a presidential re-election campaign and he will have been running all out to reinforce his base support for more than a year. Still, the period after the Labor Day holiday until the first Tuesday in November will be when the majority of voters really start paying attention to the choice they face in the presidential campaign.

In the meantime, the Democratic Party still has a baker’s dozen (or more) of wannabe nominees, and the field has yet to narrow down to a truly manageable handful of candidates before they can coalesce around one candidate – if they can do it. All – repeat, all – of the nominating convention delegate-picking caucuses and primaries have yet to happen, from Iowa and New Hampshire that come up very early in 2020, until the process finally winds up, just before the Democratic Party’s nominating convention in Milwaukee, Wisconsin.

Accordingly, speculation on the eventual Democratic candidate remains rooted in the intangible question of “electability”, and its close relative, the feeling that “anybody who can beat Trump is the right choice”. That, in effect, has become the rock upon which Joe Biden and his people have founded their church. The other first-tier candidates have been pushing and shoving, as if to say, “my choices and policies – here, see them all – are best for the next generation, so pick me, back me, and send money”. Somebody will win, of course, but who that someone will be remains unknown, even though Biden is leading in almost every poll.

That naturally leads to the question of polling. Yes, the US is a very litigious society, but it is also one in which polling, by many organisations, takes place all the time, about every aspect of political life. Numerous bodies constantly track the fates of the many candidates, the issues people say they are most or least interested in, as well as all those deeper values motivating voter choices.

Some of these groups, such as independent bodies like Pew and the Norc, as well as various combinations of national print and electronic media, have good reputations for quality work and reliable results. Other bodies, even a few associated with universities, have more problematic histories, including some with known biases leaning one way or another politically.

One trend worth noting is the poll-of-polls phenomenon that bundles the results longitudinally to establish stronger data about trend lines and to create at least some predictive capability, but still with the usual caveats about coefficients of reliability and margins of error. Polls are obviously constructed to model the actual voting population, but any poll that consistently contacts fewer than about 1,500 people should be viewed with at least some suspicion.

(Polls are not meant to be strongly predictive. Rather, they are a snapshot in time. But, the more polls whose results are gathered together, the more likely there is some useful element of prediction therein. Bottom line? The best poll is the actual vote, the one that counts.) One further note is that, so far, most polling is still matching a generic Democrat candidate against Donald Trump, even though there is no candidate named “generic” on any ballot.

Even with all these caveats, political strategists on both sides as well as independent analysts are parsing the likely shape of the upcoming electoral contest to figure out where the battle will be most fiercely fought or where they can gain an advantage. Crucial too will be how the evolving demographics of the country, and most specifically in the key battleground states, will matter. It is important to remember that the US election is essentially carried out state by state, not nationally, for electoral votes. These electoral votes are largely apportioned by population.

Generally, if a candidate wins a state by a very small majority, they still get the full electoral weight of that entire state. For this reason, national polling fails to give a fully accurate picture of how the electoral vote will fall into place. A successful candidate needs 270 of those votes out of 538 total votes (435 from the total number of congressmen, 100 from the total numbers of senators, and three allocated for the District of Columbia – the national capital of Washington).

As a result, while the Democrats have an effective lock on states such as California, New York, Illinois, Maryland, New Jersey and the New England states that equal a large electoral bloc, the Republicans are presumed to hold firm in most of the south, much of the Midwest, the Plains states and the Rocky Mountain states. Going forward, that leaves Florida, Colorado, Arizona, Ohio, Pennsylvania, Michigan and Wisconsin, along with places like Virginia, Georgia, and even North Carolina, on the bubble.

There are also a few brave (and either farseeing or just plain foolhardy) strategists who say Texas is just about there on the cusp as well, given that state’s evolving demographics. Accordingly, national support as measured in polling can be misleading, just as in the case with the 2016 Trump-Hillary Clinton match-up. Clinton garnered several million more total votes than the eventual winner; but, in losing Wisconsin, Michigan, and Pennsylvania by a grand total of fewer than 78,000 votes, she lost the whole ball game. And this is where it gets interesting – and, frustrating, if one wants predictive certainty.

Washington Post senior writer Dan Balz notes that given the country’s demographics and Donald Trump’s apparent inability to expand his base of supporters (and indeed the ebbing of support from groups like white suburban women whose shift to Democrats in the congressional elections of 2018 led to the Democrats reclaiming the House of Representatives’ majority) means this 2020 election will primarily be fought in Wisconsin, Michigan, Pennsylvania – and Florida.

As Balz writes:

“In a politically divided nation, with attitudes among many voters hardened and resistant to changing, the 2020 general election could be contested on the narrowest electoral terrain in recent memory.

“Just four states are likely to determine the outcome in 2020. Each flipped to the Republicans in 2016, but President Trump won each by only a percentage point or less. The four are Pennsylvania, Michigan, Wisconsin and Florida. Many analysts point to Wisconsin as the single state upon which the election could turn.

“Shifting demographics, the growing urban-rural divide and the gap between white voters with and without college educations have helped to create an electoral map unlike those of the recent past. So too have Trump’s unique profile, messaging and appeal.

“ ‘Because of the partisanship of the country and the partisanship of the president, we are now looking at the smallest map in modern political history,’ ” said Jim Messina, who was the campaign manager for former president Barack Obama’s 2012 re-election campaign.

“Both Trump’s campaign and that of his eventual Democratic challenger will seek to put other states in play. But those opportunities are fewer than in past campaigns.”

Balz goes on to note that the obvious unknown factor in all this is who will be the Democratic nominee and that choice’s subsequent impact on the overall electoral narrative.

He asks: “Will that nominee be running on a platform that moderate voters seen as too far left? Will that nominee be able to energise the party’s woke base and still appeal to white working-class voters?”

Regardless of who that nominee is, the upcoming election will key in on several demographic groups. First will be the white working-class voters who were strongly in Trump’s corner in the earlier election, but with special concern for white women without university degrees. Second are college-educated suburban voters, most especially women, who shifted strongly towards Democrats in 2018. Then there are African American voters who are seen as firmly in the Democrats’ corner, but for whom turnout (and their enthusiasm or the lack of it for the eventual Democratic candidate) is uncertain.

Finally, there is the question of Hispanic voters. They will be central to the eventual result in Florida and states like Arizona, Colorado and potentially even Texas, although that may be one, or even two bridges too far for the coming election.

Still, as Jennifer Rubin, a neo-conservative Republican columnist turned opponent of the president, wrote on 2 September:

“The Texas blue-wave nightmare [otherwise known as a Democrats’ triumphal wave] gained momentum with three Republican retirements in competitive congressional districts. Democrats now are looking for pickups in the Texas 22nd (Rep. Pete Olson is departing), 23rd (Rep. Will Hurd is leaving) and 24th (Kenny Marchant is retiring as well).

“Republicans’ collapse in the Texas suburbs — which is a major factor in making the state competitive — will only accelerate with its defiant attitude on any gun-safety measure. (Of suburban voters in the recent Quinnipiac national poll, I noted, ‘Among those voters, 96 percent favour background checks, 62 percent favour an assault weapons ban, 82 percent favour red flag laws, and 85 percent support gun licensing.’) Republicans’ intransigence in the face of the state’s second mass shooting in less than a month will only make matters worse for a party struggling to avoid alienating a key component of winning Republican coalitions in the past.

“It’s not just guns. Texas Republicans have other serious problems, such as the party’s inhumane treatment of migrant children and families. In July 2018, a poll on family separation found, ‘Overall, 28 percent of Texas voters support the practice — 16 percent strongly so — while 57 percent oppose it — 44 percent strongly so. Here the party’s problem with women voters is stark… If you wanted a way to put off female suburban voters, you couldn’t be more effective than to take an unyielding, indignant stance in opposition to gun safety and to embrace snatching children from their parents and keeping them in squalid conditions.”

In sum, the Republicans obviously already have their candidate and they are well and truly stuck with him. If they hope to regain the voter population segments drifting away from them on issues, such as suburban women, their dilemma is cruel – change policies that are now central to the presidential identity or drive key voters further away.

The Democrats, meanwhile, have to find that golden mean of a candidate who can capitalise on such dissatisfactions, but without alienating the energies of those on the left who seek more overarching policy changes, even as they slug it out, reasonably politely, over who will become their standard-bearer. All of this will play out in an electoral map that will, in real terms, be limited to a handful of states.

Stay tuned. Much, much more to come. DM