Who will win the US presidential election?

Who will win the US presidential election?

This was originally posted on Medium on the 8th of November, 2016.

divided flag

How many words can be written about the 2016 United States presidential election that have not already been said? America’s penchant for long political campaigns often feels at least a little bit ridiculous, but conventional political wisdom has been totally thrown out the window over the past eighteen months. Within that time, billionaire businessman Donald Trump has gone from being ridiculed and joked about for running any kind of political campaign, to being within reach of being the next president of the United States. Hillary Clinton, on the other hand, has gone from being The Chosen One to looking seriously vulnerable. What we have now is a contest that is repeatedly being labelled ‘too close to call’, which is normally an annoying phrase that feeds the media narrative of a horse-race that doesn’t end up reflecting reality, but in this election may actually be true.

Such an election is worthy of a in-depth preview, and that is what I shall strive to give you here.

Historical parallels

“History never repeats itself, but the Kaleidoscopic combinations of the pictured present often seem to be constructed out of the broken fragments of antique legends.” — Mark Twain, ‘The Gilded Age’

While Hillary Clinton and Donald Trump are (like all of us) unique individuals, there are patterns in their personalities and policies that share a more-than-passing resemblance to American presidents of days gone by. In judging this election, it is therefore worthwhile to examine these similarities, to see whether we can gain any insight into their behaviour, and their chances of winning.

Personality

Hillary Clinton — Dwight D. Eisenhower
“And I tell you this: you do not lead by hitting people over the head. Any damn fool can do that, but it’s usually called ‘assault’ — not ‘leadership’…I’ll tell you what leadership is. It’s persuasion — and conciliation — and patience. It’s long, slow, tough work. That’s the only sort of leadership I know or believe in — or will practice.”

While Hillary Clinton has never been the head of an army, her chosen method of communicating strikes a similar note to that of Dwight Eisenhower. For him, patience was the key, knowing that if he stuck at it long enough, what he believed in would come to fruition. Clinton has been at this for a long time now, becoming First Lady in 1993, a Senator in 2001, and Secretary of State in 2009, along with two nomination campaigns — all with the goal of the presidency in mind. Conciliation through explaining what positions they hold, why they hold them, and why you should hold them as well — that is the method (and it is very much methodical) of Eisenhower and Clinton, and they are determined to stick to their guns, regardless of the criticism that comes their way.

This, for the most part, also keeps them out of trouble, designed with caution in mind. By planning things out well in advance, they can be prepared for trouble, even if the plan itself has to change. But when things go wrong, the ‘weather the storm’ approach relies on the storm dissipating. When it doesn’t, the result instead is that criticism builds and builds until there’s a breaking point. This is true of them personally as well — both Eisenhower and Clinton are known for patience, but also for exploding with an incredible temper when events get to be too much for them.

Donald Trump — Lyndon B. Johnson
“I’m just like a fox. I can see the jugular in any man and go for it, but I always keep myself in reign. I keep myself on a leash, just like you would an animal.”

Lyndon Johnson would have seen more than a little of himself in the way Donald Trump approached the Republican primaries. Trump managed to spot the jugular of each and every one of his opponents, but only went it once they became a threat to his chance of winning, unleashing a nickname upon them that stuck in the public’s mind and stopped them from beating him. Much like Johnson, Trump also doesn’t keep himself as ‘leashed’ as he says he does, and is, in fact, willing to do whatever it takes to get the job done. Their styles are different , but their methods are remarkably similar: both test for the weak spot, and then use it.

Johnson relied on intimidation (also known as ‘The Johnson Treatment), using a litany of research on you, piling it up and making sure he hit every weak spot until you give in. Rather than intimidating, Trump tends to use a sales pitch (as befitting a billionaire salesman) that entices you in by testing what gets the best reaction out of you, and then pushing that as far as it will go…until you give in. Persuasion is one of the most important tools any politician will ever have, and these two have it by the bucketful. But this way of doing things is an abrasive one, and leaves its user being liked and loathed in equal measure. If they’re convincing enough, they’ll win more people over than turn them off. If not, they’re in for a rough time.

Policies

Hillary Clinton — Bill Clinton
Bill and Hillary go hand in hand, literally and metaphorically. While the latter has posed herself as the continuation of Barack Obama’s presidency — which is, admittedly, quite true — it is actually the Democratic presidency prior to that that her own priorities most resemble. It is hardly surprising that a husband and wife should have very similar political aims, but to give you an indication of just how similar they are, here are some of the major policies that Bill Clinton got legislated, signed off on or pushed for while he was president: free trade, public health care, paid leave, foreign interventionism, voter registration, gun control, loosening corporation and banking laws, loosening abortion restrictions, climate change prevention and ‘don’t ask, don’t tell’.

Every single one of these is something that you could also say would be on Hillary Clinton’s agenda, and no-one would bat an eyelid at the suggestion. The only one that the Clinton camp would make noise about is free trade, due to her dubious claim to now be against the Trans-Pacific Partnership. Given she has been an advocate for free trade for decades, has a personality of not giving up her positions and pushing for them patiently, and the fact that she has stated that she has different public and private positions on policies, it is hard to imagine a Hillary Clinton presidency not involving free trade, along with all the aforementioned positions of Bill’s.

Donald Trump — Theodore Roosevelt
Not since Teddy fought Taft has the GOP had a crisis akin to the one it’s been going through, and the similarities between Roosevelt and Trump politically are worth noting (and incidentally, Ted could well have been slotted in place of LBJ earlier, as he too was a bold, brash, impetuous individual). The biggest similarity between the two is in their attacks on what they see as the collusion between corrupt politicians and big business, complete with their own phrases: ‘invisible government’ was Roosevelt’s, ‘special interests’ is Trump’s. Trump has singled out a number of large corporations for breaking up —media conglomerates being first and foremost, of course — channeling Roosevelt’s trust-busting in doing so.

Both place a primacy upon talking about America and its greatness, and not just in the usual way most candidates do — for them, it’s a tactic designed to create unity or, at the very least, loyalty, as well as being something they believe in very strongly. Roosevelt was highly critical of ‘hyphenated Americans’, who pledged allegiance to something other than America, and while Trump has not spoken particularly about this idea, identity politics has been more or less an enemy of his throughout his campaign. Both also talk very tough on immigration, with Trump’s ‘extreme vetting’ of Muslim immigration actually being softer than Roosevelt’s total ban on Chinese immigration. Roosevelt was also a big believer in negotiating, and was talented at it, earning himself a Nobel Peace Prize no less — music to the ears of Mr ‘Art of the Deal’, I’m sure.

Election

William McKinley (Rep) vs William Jennings Bryan (Dem), 1896
The election that brought about an era of Republican dominance is, in many respects, a precursor to the one happening 120 years later. William McKinley, previously Governor of Ohio, was a defender of the economic standards of the day: protectionism and the gold standard. He is the Clinton of this scenario. His opponent was a Nebraskan Representative by the name of William Jennings Bryan, a crusader for the working class, railing against the elites for what he believed were their deliberately corrupt practices, enriching themselves at the expense of the American worker. He is the Trump of this scenario.

Final result of the 1896 United States presidential election — McKinley won with 271 EVs to Bryan’s 176

As you can see, the electoral map is basically the opposite of what we would expect today, party-wise — GOP in the north-east and on the west coast, Democrats everywhere else — but it is almost exactly what the Clinton campaign will be hoping for candidate-wise, with Clinton holding onto the what is now the rust belt, but was at the time the centre of industry in the US.

So, what happened in 1986?

McKinley had a easy road to the nomination, working on the back of his governorship of Ohio and having his campaign managed by businessman Mark Hanna, who had earnt his wealth in the coal and steel industry, and was extremely well connected in business circles. By the time of the convention, McKinley’s campaign had vastly outspent and outworked any of his competitors, with Hanna funding it all, and the Republican delegates overwhelmingly voted him in.

Meanwhile, the Democrat nomination was as fraught as the Republican nomination was straightforward. The ongoing economic depression had created friction within the party, particularly on economic issues of money supply and tariffs. Outgoing Democrat president Grover Cleveland was a leader of the ‘gold faction’ of the party, and was on the outer. The dominant force was the ‘silver faction’, who believed that the gold standard was seriously harming the rural poor at the expense of wealthy bankers and industrialists, and thought that using a silver and gold standard would be beneficial, as it would cause inflation. Bryan had been pushing for bimetallism since he was elected, though his belief in the cause strengthened over time, and by 1894 he was already preparing to run for the nomination.

Bryan was well renowned for his oratorial skill, particularly on the silver issue, prior to the convention, but he was not the favourite for the nomination. Another bimetallist, Richard Bland, who had been in Congress for 23 years by then, had been asked to run and had done so, though he did not desire the position. The gold faction, meanwhile, was supporting Robert Pattison, and there were also another eleven candidates voted for on the first ballot. But Bryan, in the last speech prior to the ballot, delivered one of the most famous addresses in America’s history, that would deliver him the nomination: the Cross of Gold speech. In it, Bryan decried the gold standard and the arguments of its defenders with such skill that the audience more or less exploded. He had to stop his speech multiple times while waiting for the audience to calm down, and after it was finished he was carried around the floor on the shoulders of delegates.

Bryan asked that a ballot not be performed immediately, supposing that if his support couldn’t last more than a day, it wouldn’t last until the election. For the first three ballots, Bland led and Bryan was second, but Bland was unable to gain any more support and Bryan was able to gain support on every ballot. Upon realising this, Bland’s supporters switched to Bryan, and he was nominated. The gold faction Democrats had, by this point, left the convention, and went on to form their own National Democratic Party, running a ticket that many believed was designed to elect McKinley, and received campaign funds from Hanna. The silver faction Republicans, on the other hand, nominated Bryan as the candidate of their Silver Party, on the basis that he was the greatest advocate for the issue. The Populist Party, which had carried four states in the 1892 presidential election, felt compelled to nominate Bryan given his similar platform to their’s, lest they split their base and hand the presidency to McKinley.

Media reacton to McKinley’s nomination was far more favourable than it was to Bryan. The only northern Democratic newspapers that endorsed Bryan were those owned by William Randolph Hearst, on account of Hearst’s ownership of silver mines, as the rest criticised Bryan as a populist and a demagogue. Non-Democratic newspapers in the west, on the other hand, supported his candidacy, as did southern Democrats, due (at least in part) to many African Americans supporting McKinley. Later in the campaign, Bryan was painted as a religious fanatic, with policies that would ruin the economy. Republican newspapers, meanwhile, not only endorsed McKinley, many of their owners also donated to his campaign, as Hanna funded his campaign with a method used ever since: asking businesses for funding in return for favourable policy. For industrialists and other business owners living mostly in the north-east, the gold standard was a necessity for the continued well-being of their businesses. Bryan was, therefore, the enemy, and Hanna used this fear of bimetallism to gain their financial support for McKinley. Hanna’s prominence in the campaign was much criticised by Hearst’s newspapers, characterising him as the puppet-master, and McKinley as the clueless puppet.

Bryan also managed to start his own presidential campaign trend. He was the first presidential candidate to go wherever he wanted, whenever he wanted, making campaign stops in towns all around the country in order to get his message heard. After all, he could not afford to send pamphlets everywhere like McKinley. The Republican nominee took a more typical approach, in that nominees had historically tended not to campaign for votes far and wide. But he took it to its furthest limit, choosing not to leave his home at all, preferring to have people transported to hear him speak on his front porch, which they did in large numbers.

The fervour of support for Bryan tricked some election watchers into believing that he was the frontrunner to win, but what the silver issue actually did was polarise support for the two parties. In the rural south and west, particularly amongst farming communities, Bryan was immensely popular, and Democrat support signficantly increased. But in the industrial north, the Democrats actually suffered from the focus on silver, as skilled workers who had traditionally supported the party feared that inflation would lower their own wealth. Bryan failed to appeal to them, and the Republicans succesfully got them to support the gold standard instead. Ultimately, the election was won in the mid-west, which McKinley carried following a last-minute push. The urban working classes in these states, particularly from German backgrounds, did not support silver — that was enough to tip the scales red.

So, how many parallels did you spot? On one side, a politically experienced candidate, backed by big money, running a campaign built on fear of their opponent and the safety of continuing with the economic norms of the day, using their wealth to carpetbomb the constituency with information rather than relying on rallies. On the other, an inexperienced candidate, whose words bring the attention of the party to him, creating a great amount of fervour amongst his support, which his critics describe as demagoguery, and who relies on policies that crystallise pre-existing divisions within the US, while attempting to solve the economic and social despair of certain regions of the nation.

This is not to say that they are exactly the same. Trump’s oratorial skills are generally derided by his opponents rather than praised, and his coalition of support is more broad than Bryan’s was. But it is worth remembering what happened here, for history is a wonderful teacher.

Incidentally, the equivalent of the 1896 electoral college votes in today’s election would be Clinton winning with 278 votes, Trump 227, and 33 unassigned.

Factors in the 2016 election

We’ve looked at the past, so now we can turn our eyes to the present. Every election has its own factors, but this year seems to be turning the tables on some traditionally reliable sources of information. In predicting the results, we need to take a closer look at the factors that will matter, and why they will.

A Tale of Two Countries

What does this image tell you?

Results of the 2012 presidential election by county by size of winning margin, deeper shades represent bigger margins.

To talk about the United States being ‘polarised’ or ‘divided’ is a cliche. But it’s a cliche that works, because, well, it is a divided, polarised country. If you live in regional, rural United States, urban areas are something foreign, full of things you haven’t experienced and don’t understand. The same is true if you live in urban United States, in which case regional United States may as well be another country to you. The Republicans and the Democrats are the representative labels for this divide, essentially turning politics into purely a matter of identity, rather than ideology. The latter flows from the former.

But for many years now, Republicans voters have been poorly represented by the elites of their party. There has been a severe disconnect between the two. Into this equation steps Donald Trump, whose policies focus on areas those now attached to this Republican identity identify with and have been begging their party to address for so long. These aren’t the people who will change the overall result — because voting Republican is now part of their identity — but they are people who have felt that the Republican party doesn’t represent them, and have given the clearest indication possible that they want that to change, by making the party change.

What polarised America means at elections is that most states are not up for grabs, but specialised candidates like Trump who can hone in on an geographical area their opponent is meant to be stronger in, such as the rust belt, may be able to change things enough to win the day. After all, they have their half of polarised America to rely on — they don’t need to worry about winning them back over at every election.

The enthusiam gap

If elections were won on rally sizes and yard signs, Trump would be president of the universe. His rallies undoubtedly fall into ‘yuge’ territory, and anecdotal evidence from even the most blue states suggests that Clinton yard signs tend to only pop up in very selected communities, and even then it isn’t exactly en masse. Now, neither of these should be taken as evidence of anything much, but they are, in a sense, backing up something that most people have sensed throughout this campaign: people who support Donald Trump are far more enthusiastic about it than people who support Hillary Clinton. There is little-to-no zeal for Hillary, as what zeal exists tends to be more against Trump. Trump, on the other hand, has run his campaign on zeal from the get go, and is a big fan of calling it a ‘movement’.

Why does this matter? Well, it may not matter much at all, as the majority of rally-goers would vote regardless of the nominee, but it can matter if at least one of two things happens: first, that the enthusiasm translates to previous non-voters voting for Trump, particularly in areas that his message resonates strongly in. Second, that the lack of enthusiasm for Clinton translates to a drop in support in areas Obama got unprecedented support in, that she needs to win in. Turnout matters a great deal in deciding the outcome of these elections.

Independent’s Day

America’s unusual voter registration system, whereby people register as either Democrats, Republicans or independents, does at least help out pollsters and pundits trying to figure out what’s going. Once upon a time they made up a minority of registered voters — something like 10% — but they are now the biggest group of the three, accounting for approximately 40% of all registered voters, with the remainder being divided relatively evenly between the two, though there tends to be a slender lead for the Democrats.

This means that the votes of registered independents hold a huge amount of sway in who wins an election. Some have attempted to explain the increase in registered indepenents as being indicative of an increase in ‘swing voters’, but this seems at odds with another theory that has taken hold and seems more provable, which is that America has seen an increase in political polarisation. What this may actually mean is that while those who consider themselves Democrats or Republicans may be further removed from their opposite numbers than ever before, an increasing amount of American do not see either major party as being representative of them, though they likely do have more sympathy for one than the other. If they do not like either candidate, they don’t vote, or they vote with whichever party they sympathise with more.

The good news for Donald Trump is that he has a heavy lead amongst independents, and has done for a long time. This reflects the reality that he is very clearly not a typical major party candidate. In particular, he has significant support in areas that have a traditionally low voting rate. These areas have a lot in common with areas of the United Kingdom that voted ‘leave’ in their referendum earlier this year, and they are full of people who have felt disconnected from the political process and the political parties in their respective countries. Given the enthusiasm gap swings so heavily in his favour, Trump will be hoping that these areas do indeed turn out for him, as they will be pivotal to a Republican victory. It isn’t all bad news for Hillary Clinton, though. There is good reason to think that in some areas — specifically urban areas with high rates of ethnic diversity — independents will actually swing her way, rather than Trump’s. This could be crucial in holding on to a state like Florida.

Polls

Clinton, on the other hand, will be hoping for one thing: the polls being right. As of now, she hold a small-but-comfortable lead in most polls in swing states, and nationally. Trump and his campaign have denounced the polls as being bought and totally wrong throughout the campaign, except for when they go his way, in which case they are an example of how big his support is. While this is, at least in part, just his sales pitch (Americans love a winner and people love following the crowd, so Trump loved pointing out that he won every debate, even if he didn’t, and was always leading the polls, even if he wasn’t…and it worked!), it is worth scrutinising the polls, because they are a crutch for the tale the media tells, and most predictors (such as FiveThirtyEight and the Princeton Electoral Commission) rely on them in some way or another.

Most polls are done by phone, and consist of a random sample of the population, which are then recorded by response and by demographic. These responses are then occasionally adjusted according to the demographic measurements, but whether this is done and how it is done depends on the pollster. These polls can also be done either on a national or state level, and have different sample sizes. There are also two other variations: polling online instead of by phone, and polling the same population sample repeatedly rather than a new group each time. This all means that there a lot of different ways a pollster can get it wrong.

Phones, for example, are an increasingly poor method of polling, especially in mobiles are not called. Generally only a small percentage of calls get picked up, which means it is both costly and takes a lot of work to get a small sample. Furthermore, those that do pick up the phone have particular demographic skews, which is often reflected in the result. If a particular demographic supports one candidate heavily, and that demographic gets over-sampled, the poll becomes unreliable. Random sampling creates a margin of error, because the point of a poll is to gauge what the population thinks about candidates over time. If you ask ten people on the street one day, and then ten different people the next day, you may get the total opposite result. This also means random sampling is a bigger issue when the sample size is smaller.

Poll adjustment is also fraught with difficulty, as it is a result of the assumption that the raw numbers of the poll are wrong (not a good start), and that the adjustment that is made to it will be reflective of reality. This is a particularly big problem if there is a massive demographic skew in the raw numbers, and is also reliant on the demographic data for the state being accurate.

If that all sounds a bit too theoretical, let me give you an example of bad poll. This was done for the Arizona Republic by Cronkite News, and recorded a lead of five points for Hillary Clinton. The problem? Of the 713 voters asked, 413 were registered Democrats, 168 were registered Republicans and 132 were registered independents. Just under twice as many Democrats were asked compared to Republicans, and even less independents, even though total voter registration in Arizona has Republicans in the lead, then independents, then Democrats. There is absolutely no way this can be considered a reliable poll, and yet large mis-sampling (I make no comment on whether it is deliberate or accidental, though neither are a good look) seems to be all too common.

Now, this explanation regarding party identification may ring some bells to people who followed the 2012 election, where a similar explanation was used by some people — almost exclusively supporters of Mitt Romney — as proof that Obama’s poll numbers were higher than the reality, and that the polls needed to be ‘unskewed’. The problem they found was that on election day, the reverse was true — Democrat turnout was higher than anticipated in the polls, and Republican turnout lower, which made them all look a bit silly. However, the lesson that should be taken from this is not that the polls are right — as mentioned, most polls in 2012 actually underestimated Democrat turnout — but that US elections are almost always decided on turnout. Obama was able to get higher turnout amongst groups that the pollsters couldn’t spot. If either candidate can do something similar, they will probably win.

In fact, party identification is not a useful measure for the polls in and of itself. Usually, a bias in party identification in a poll is actually a sign of demographic biases in that poll, which just happens to come through in party identification (because of identity polarisation, as discussed earlier). This is why ‘unskewing’ is a near impossible task — it relies on party identification, and nothing else. For a poll to be reliable, it should actually avoid party identification, and try to accurately reflect the demographics of the electorate instead. For example, the Arizona poll heavily oversamples those who have completed college education, which is the actual reason for the discrepancy in the party registrations, and in the final result.

Does this mean that the polls should be ignored entirely? No. But it does mean that you must read them very carefully, which, unfortunately, is far too much effort to be practical for most people. If you do want to watch the polls, then do not pay attention to the headline figure. Pay attention instead to the numbers that make up the polls: age, location, income and the like. If they don’t match up with reality (as with the above poll from Arizona), then take it into account. If they don’t give you anything other than the headline figures (such as Rasmussen, for those that don’t pay for their service), then use them with the utmost caution.

Third party votes, undecideds and afterberners

One area that sticks out in the polls is that of support for third party candidates. We are told that the two major party candidates are so unpopular that the third party vote (for Libertarian Gary Johnson, and Green Jill Stein) will be much higher than ever before. I am almost certain that this will not turn out to be the case, particularly for Gary Johnson, who has managed to get up to double digits in some polls. In reality, these are safe zones for people to park their votes in the polls while another scandal rolls along. Once it is safe for them to return to the major party candidate of their choice, they do so. This is reflected in the drop that Johnson has had over the last few weeks, and it will be surprising if either gets more than 2% of the overall vote.

It is a similar story with undecideds, which are still quite high in a number of state polls, unusually so this late in the campaign. This is probably at least in part due to the anger still felt by a swathe of Bernie Sanders supporters, who have been suspicious of the Clinton campaign since the beginning of the primaries, and with each Wikileaks drop presumably become angrier. Some long ago made up their minds, either to bite their tongue and vote Clinton, bite the bullet and vote Trump, or to say to politics in general, ‘bite me’. But some undecideds (who aren’t just parking their votes there, and are genuinely making up their minds this late) that come from the Sanders camp may prove crucial. After all, Sanders won a number of key states: Michigan, Colorado, New Hampshire, Wisconsin and Minnesota. He was also screwed out of winning the Iowa and Nevada caucuses. If there’s a strong sentiment one way or another, that small percentage of the electorate could be enough to tip it. For those that are parking their votes in the undecided column, I would expect them to break similarly to that of the third-party votes, for much the same reason.

Early voting

Before we get to the main event, a quick word on the value of early voting for making predictions. In those states that allow early voting, the release of data is varied (some states do not release at all, some do it occasionally, some do it daily, some record party registration and others don’t), but there is one constant: we do not know the actual votes until voting day. Any predictions made by eyeballing the early voting figures are precarious at best, especially if they do not make a comparison to either 2012 and/or 2008 figures.

Predictions

This will be a state-by-state prediction in alphabetical order. States predicted to flip will be in bold. States are assigned electoral college votes according to their population, and a candidate must win at least 270 electoral college votes to win. Polls have been analysed beyond the headline figure where appropriate.

Run of the board

State (electoral college votes/last election)
Alabama (9/R): Has voted Republican since 1980.

Alaska (3/R): Has only voted Democrat once, and that was Johnson in 1964. There was a slightly wacky poll that had Trump up by only three points, but this is pretty safe territory for him.

Arizona (11/R): Bill Clinton won this in 1996 thanks to Ross Perot taking votes from Bob Dole, and nearly won it in 1992 for the same reason. Prior to him, the last Democrat to win AZ was Harry Truman in 1948. Dodgy polls from the Arizona Republic aside, Trump should win this state by a few percentage points.

Arkansas (6/R): Voted for every president from John F. Kennedy to George W. Bush, but went heavily for McCain and Romney instead of Obama, and Clinton probably only won in 1992 and ’96 due to his previous governorship of the state. Hillary won’t be winning this one.

California (55/D): Has changed a lot since it last voted Republican in 1988. Comfortably in the Democrat ledger.

Colorado (9/D): The first of our states to be a toss-up based on the urban-rural divide. Barack Obama won Colorado in 2012 by the margin that he won Denver county by: 150,000 votes. A high turnout in Denver and Boulder will more than likely mean a Clinton victory, but the rest of the state will turn up for Trump. I do not believe that there will be a significant enough turn around from 2012 for Trump to win here, unless the large, young support for Bernie Sanders that won him the state in the primaries decides to not turn out in Denver for Clinton.

Connecticut (7/D): Has voted Democrat since 1992.

Delaware (3/D): Has voted Democrat since 1992.

Florida (29/D): Florida has a pretty good record at picking presidents. Since 1912, it has only voted for the national loser three times (1920, 1960 and 1992). It is also the equal-third most valuable state in the Electoral College, behind California and Texas, and tied with New York. Florida is also, to put it simply, a forecaster’s nightmare. It is highly polarised, and can be divided into a number of totally different regions which have little in common bar their government. Anything even slightly misaligned in the polls can have a large impact on the headline figures. Even then, there are thing going on that we won’t know until after the result. For example: registered independents are up on previous years, and Trump has been doing better with independents to date, therefore it benefits Trump. But Hispanics account for a larger proportion of independents than average, therefore it benefits Clinton. But amongst these are spread significant Cuban support, which has been greater for Trump. But Cuban support for Trump is mostly only amongst older voters. But older voters are more likely to vote than younger ones. And all that is trying to account for just one piece of the Palm State puzzle. You could well just toss a coin and have a better likelihood of getting it right. I don’t want to cop-out of making a prediction here just because it’s difficult, therefore I believe Donald Trump will win this by a slim margin, for a reason that I will explain further on.

Georgia (16/R): The Peach State keeps getting talked up as a potential Clinton gain, but I do not see how. I am loathe to use racial demographics, but one that has stood out is that Barack Obama was unable to win Georgia in 2008 or 2012, despite Georgia having the third highest proportion of African Americans in the country. How, exactly, does Clinton overcome that?

Hawaii (4/D): Has voted Democrat all but twice since 1960.

Idaho (4/R): Has voted Republican since 1968.

Illinois (20/D): Much like Colorado, there is a strong urban-rural divide here. Much like Colorado, Obama won the state in 2012 by the margin he won its largest city by. Unlike Colorado, Trump has no chance of catching it back up, as the difference in Chicago is not 150,000 like it was in Denver, but 1,000,000. Clinton is also from Illinois, for whatever that is worth.

Indiana (11/R): Has only voted Democrat twice since 1940, but one of those was for Obama in 2008. This seems like an anomaly though, and IN should be very solid Trump rust belt territory.

Iowa (6/D): A small-but-fabled mid-west ‘swing state’ without a winning record like Florida’s, it’s voted Democrat for six of the past seven elections, but looks like switching to Trump at this election. There are two bases for this: the first is that Trump’s message naturally resonates in this state. The second is that the most accurate poll in America had him leading here by six points as of Sunday. The Des Moines Register poll is the one all the pollsters want, and all other polls want to be, with not only an excellent record of picking winners, but a good record with margins as well.

Kansas (6/R): Has voted Republican since 1968.

Kentucky (6/R): Has the same record as Arkansas since 1964, voting for the winner up to 2004, and heavily Republican since then.

Louisiana (8/R): Much the same as AR and KY, though George Wallace carried the state in 1968 as a third-party candidate.

Maine (2+1+1/D+D+D): One of two states to divide its electoral college votes between the state as a whole and each of its congressional districts, ME has its own urban-rural divide, despite its size. Clinton looks like winning the 1st district by enough to carry the statewide vote, but Trump is pushing very hard for the 2nd district, and I suspect he will manage to pull it off. It is only one vote, but they all count.

Maryland (10/D): Has voted Democrat since 1992.

Massachusetts (11/D): Has voted Democrat since 1988.

Michigan (16/D): It’s been a long time since Michigan was treated by presidential candidates as a state that could make or break their election. But as it stands right now, both candidates are absolutely determined to win MI, which doesn’t have early voting. How voters see their two options on election day will determine who wins here. Trump has made an enormous play for a state that hasn’t voted Republican since Reagan, and it’s not hard to see why. Nowhere can give us the visual of what ‘the rust belt’ means quite like Detroit. Manufacturing was (and remains) the cornerstone of the economy here, and blue-collar workers quite like the sound of the Republican candidate, who has spent a lot of time here. In fact, the group that Trump needs to convince is not the manufacturing workers that traditionally vote Democrat. It’s actually the suburban voters who traditionally vote Republican, who have been telling pollsters that they don’t like Trump as a candidate. If he can win them over, he will win the state.

Minnesota (10/D): Has voted Democrat since 1976, and was famously the only state to go blue in 1984. Polls here have been consistently mildly close, but not by enough to suggest Clinton won’t win it. Having said that, the Trump team seems bullish about their chances here, perhaps hoping the North Star State’s idiosyncratic political history will continue.

Mississippi (6/R): Has voted Republican since 1980.

Missouri (10/R): Has voted for every President since 1904, with three exceptions. One was in 1956, and the other two were in 2008 and 2012. Were it not for Ross Perot’s popularity, this would probably have been another state to have voted Republican since 1980, and it seems unlikely Clinton will be able to claw it back.

Montana (3/R): Has voted Republican since 1968, except for 1992 due to Perot.

Nebraska (2+1+1+1/R+R+R+R): The second state to divide its electoral votes by district, Nebraska has voted Republican since 1940, except for 1964. The electoral vote has only been split once, with McCain winning four votes and Obama one in 2008. It seems unlikely that Clinton will manage such a feat this year.

Nevada (6/D): Here comes the early voting boogeyman, and he’s coming to get Trump! Much is being made here of the lead in the early voting that the Democrats have, which is a testament to their voter drives. Polling in the state has been neck and neck, however this has been the case in previous years that have understated Democrat performance by at least a few percentage points, often by missing portions of the Hispanic population. It seems quite likely that this will be the case again, in which case this will be a solid Clinton hold.

New Hampshire (4/D): Once upon time solid Republican territory, recent decades have turned the Granite State into a swing state, and one that could prove enormously important despite its small size. The last time NH went Republican was 2000, when it helped George W. Bush crawl over the line against Al Gore. Clinton had been leading in the state for most of the year, but the mood seems to have totally flipped over the past two weeks, leaving Trump with a small but important lead in most state polls. Gary Johnson also continues to have very high polling numbers here, which is not surprising as the Libertarian Party has sought to make NH a stronghold of theirs for many years now. But given he received 1% in the state in 2012, it is hard to see him actually turning out significant numbers on election day, and I would expect them to break more for Trump than for Clinton.

New Jersey (14/D): Has voted Democrat since 1992. Any chance of a Trump upset disappeared with the recent conclusion of ‘Bridgegate’, which found former officials for Gov. Chris Christie guilty on all charges.

New Mexico (5/D): Has voted Democrat since 1992. As the name suggests, it has the highest proportion of Hispanics of any American state, which have been presumed to make this a comfortable Clinton victory. Polling, though, has been tighter than expected, but much like Minnesota, it’s unlikely to be enough to flip the state.

New York (29/D): Officially the home state of both candidates, NY is another state with an enormous difference between its major city and the rest of the state. Trump, who was born in the borough of Queens, is unlikely to find too much support in New York City, where the Democrats have been dominant for many years now. Upstate NY, on the other hand, is likely to swing his way — he has a long history in the region, and it is the eastern most part of the rust belt. However, the sheer size of NYC will be enough to ensure a Clinton carries the state.

North Carolina (15/R): Florida may be the eternal swing state, but I believe NC will be the place that shows us which way the contest will go. Demographic changes over the past decade have turned a state that was comfortably Republican, all the way from Nixon to Bush jnr, into a tight contest. Obama won by half a percent in 2008, and Romney by less than 2% in 2012. Polls have been bouncing between Trump and Clinton for months. Interestingly, there is an strong positive correlation between polls where Gary Johnson does better, and polls where Clinton is leading. Even where she is leading, it is by the margin of error. Furthermore, Trump is doing better than Clinton on early voting when compared to 2012. Assuming that turnout is at least as strong for the Republican candidate on election day as it was in 2012, this would mean a comfortable victory here for Trump.

North Dakota (3/R): Has voted Republican since 1968.

Ohio (18/D): In reality, Florida is not the real swing state. Florida’s presidential record can be put down the great variance that exists in the state from region to region, with support polarised one way or the other in most of them — hence the closeness of the usual margin there. No, the real swing state is Ohio. Only twice since 1896 has the Buckeye State got it wrong: 1944, and 1960. Right now, Trump is leading the polls there, and the Clinton campaign appears to be more interested in defending states deeper in their territory. While the major cities will go her way, she will need to hold on to the northern counties that Obama won, but also hope Trump does not improve on Romney’s position in the south and east — a tough ask. Not only is OH a good predictor of the presidency, it will also be a sign of what will be happening in the other rust-belt states.

Oklahoma (7/R): Has voted Republican since 1968.

Oregon (7/D): Has voted Democrat since 1988.

Pennsylvania (20/D): Republicans have long dreamed of getting Pennsylvania back, having lost it to Bill Clinton in 1992, but with Donald Trump they may genuinely have a chance of doing it. PA has restricted early voting, which has caused some concern amongst Democrats who believe that this will result in lower turnout than they need amongst key demographics of theirs. The most recent polls have shown the gap closing and closing, to what is now more or less a tied race. Philadelphia and its surrounds have been a haven for the Democrats in the state, with Obama winning around 90% of the vote there in 2012. The rest of the state, however, does not look anywhere near as promising for Clinton, even assuming she can match Obama in Philly. Western Pennsylvania, barring inner-city Pittsburgh, has been shifting Republican at each election, while the blue-collar south of the state will be turning out heavily for Trump after being unenthused about Romney.

Rhode Island (4/D): Has voted Democrat since 1988.

South Carolina (9/R): Has voted Republican since 1980.

South Dakota (3/R): Has voted Republican since 1968.

Tennesee (11/R): See Kentucky.

Texas (38/R): There have been a few wacky polls suggesting this will be close, or even a Clinton win. Texas has been Republican since 1980, and there have most certainly not been enough changes in the state recently to overturn the massive margins the Republicans have been winning by in recent elections.

Utah (6/R): This election’s third-party scare story, thanks to the Never Trump campaign. Evan McMullin has been up in a number of polls, particularly after the leaked ‘Trump tape’. But I suspect the reality of the situation is much like is with other third-party candidates in the polls: McMullin is a safe place for Trump voters to park their vote while there is a furore over him. McMullin will probably finish third with a decent amount of the vote, but Trump will win.

Vermont (3/D): Sanders territory that has voted Democrat since 1992.

Virginia (13/D): Between 1968 and 2004, Virginia only voted Republican in presidential election. Not any more. The border with Washington DC has meant that northern Virginia’s demographics have been turning it blue, and it may take a lot of effort for the Republicans to turn this around. Trump is likely to draw a lot of support in the south-west of the state, much of which is coal-mining territory. Indeed, in the primaries Trump won the state, but only by a few perecentage points over Marco Rubio, who was the preference of urban Virginians. But it is hard to see him increasing turnout in his stronger areas by enough to overcome the large urban advantage Clinton has.

Washington (12/D): Has voted Democrat since 1988.

Washington, DC (3/D): Has never voted Republican.

West Virginia (5/R): A state with a unusual voting history, but is now one of the safest Republican states in the country. Trump’s policies resonate here, and he may manage to improve on Romney’s 2012 result, which was +41.

Wisconsin (10/D): Has voted Democrat since 1988. The western side of the rust belt, Wisconsin has been the most overlooked of Trump’s targets throughout the campaign, and it also appears to be the toughest ask for him to win. Partly, this is because the Milwaukee suburbs which traditionally vote Republican are wavering on him as a candidate, even as he picks up votes amongst rural voters — much like Michigan, but to an even greater extent. If the suburban voters come out for him, and he keeps his lead with rural voters, he can win here.

Wyoming (3/R): Has voted Republican since 1968.

Final result

If the result in each state is as predicated, the electoral map will be as follows:

Now, Florida is the least certain of these states, and the reason I have given it to Trump is because I have predicted he will win the presidency via the rust belt. For while both Florida and Ohio have winning records when it comes to picking the president, they work in different ways. In essence, Ohio is one big sample that represenents the average American state fairly well, though with a demographic tilt mildly toward the Republicans. Therefore, you can be assured that whoever wins the state has likely also picked up states all around that average — enough to win the presidency. Florida, on the other hand, is like having four or five states lumped together, all of which are quite different from one another. It is therefore not predictive of the presidency, as (in theory) a candidate could win Florida by overwhelmingly winning one or two of these parts of Florida, but lose the election because the equivalents of those states were the only ones they could win.

However, the presidency is predictive of winning Florida, because you need to have one a variety of states with quite different demographics to win the presidency, which should follow by winning multiple Floridian regions. Likewise, winning the presidency is not predictive of winning Ohio, as you can theoretically win the presidency by winning the demographic extremes in both directions, but not Ohio itself or any similar states in the middle. To turn this into a maxim: If you win Ohio you win the presidency; if you win the presidency, you win Florida. I believe Trump will win the presidency because of Ohio, therefore he will win Florida.

Unfortunately for this maxim, it may not even last this election. It is entirely possible that Ohio and Florida are won by different candidates at this election, if Trump wins the rust-belt heavily but Clinton wins Florida due to a rise in Hispanic voting influence. It will then be up to the remainder of the rust-belt states to decide the election. Assuming he holds North Carolina, he must win Pennsylvania and Michigan, as well as New Hampshire if he doesn’t win Wisconsin.

If Hillary Clinton wins, it will be because her demographic coalition was able to overcome the white, blue-collar vote. If Donald Trump wins, it will be because he won over the rust-belt, and they really will call him ‘Mr Brexit’.

Leave a Reply

Your email address will not be published. Required fields are marked *