US Election Predictions, or The Elephant (and the Donkey) in the Room

For those who know me, you will know that I Love American Things.

My love of American things began with the innocent trio of Friends, Baseball and The West Wing then blossomed when I studied US History at university and then travelled by rail around the great nation in the spring of 2008.

My love of US election nights dates back to Obama’s success in 2008. The election party of that year was a truly joyous occasion and the 2012 sequel, though more nervy and more tempered in expectation, was still celebrated as generally ‘a good thing’.

The 2016 election has not been ‘a good thing’ and the outcome will leave most feeling dirty, regardless of who emerges victorious. This may cause Americans to take a long, hard look in the mirror and choose a new, less divisive, less corrupt, less obnoxious political path; it may not.

Nevertheless, for all the uniquely distasteful elements on display in the 2016 election, when the fundamentals and likely outcome are considered it is actually shaping up to be quite a normal, predictable election night.

Who will win the presidential election is of course impossible to know with certainty and those who are paid to crunch the polling, demographic and turnout data and give a level of confidence generally predict a Clinton win in 65-90% of contests, i.e. were we to re-run the election multiple times she would win most, if not almost all times.

The best way I have found to navigate around this not altogether helpful statistic is to ponder (using supposedly authoritative, professional knowledge) what key strengths and weaknesses each candidate has, what electorate these attract/repel, and how likely each strength and weakness is to translate into hard votes. From this I can piece together who is likely to win, and most importantly why. So here goes…

Hillary Clinton: ‘I’m with Her’

Clinton’s campaign is better organised and better funded and as such has a better ‘ground game’ that can identify and target sympathetic voters and get them to the voting booth.

Clinton’s (and the Democratic Party’s) strongest demographic support comes from blacks, Hispanics, the young (18-30), women and the economically less well-off (lower/working classes). Together these make up the so-called ‘Obama coalition’ that delivered landslide Democratic victories in 2008 and 2012 (but failed during the mid-term elections of 2010 and 2014).

Clinton’s campaign has marketed itself as continuity Obama and tried to revive this Democratic coalition for one more big win. However, she has faced problems in motivating blacks (especially young blacks) to the same degree as Obama and has faced similar issues with ‘Millennial’ voters more generally who tended to support her rival Bernie Sanders in the competition to become the Democratic Party nominee.

However, Clinton’s weakness with these groups appears to be balanced by her strength in appealing to the increasingly large Hispanic population and to the votes of women – in this she was helped by Trump’s derogatory comments towards both Hispanics and women on multiple occasions during the campaign. And just as Obama sold himself as the historic first black president, so too should Clinton expect to benefit from attempting to be the first female president.

In summary, Clinton looks likely to do better than Obama with women and Hispanic voters, or at worst should match his support. However, she will struggle to motivate as many young and black voters. This appears to point to an overall neutral impact nationwide, with particular consequences in individual state races.

The Hispanic share of the population in states like Arizona, Nevada, Texas, New Mexico and Florida is significant and rising fast, so the Democratic margin for error has increased in these since 2012. Similarly, more women vote than men across the nation, so if Clinton is able to win a larger share of female voters than Obama then this has a disproportionate benefit in total voting figures.

However, states where urban Millennials and black voters were central to the ‘Obama coalition’ might at risk: Detroit (Michigan), Cleveland (Ohio), North Carolina and Philadelphia (Pennsylvania) were all key focal points of Democratic strength that now look distinctly less encouraging under Clinton.

The ‘Obama coalition’ has thus changed subtly under Clinton’s leadership but not necessarily to the electoral disadvantage of the Democrats nationwide.

However, the wildcard (and probable linchpin of the whole election) are white ‘educated’ and ‘non-educated’ voters (often understood as those with graduate degrees and those without). Educated, white male and female voters appear to have moved towards Clinton for this election, largely because of the character of her opponent, Trump. If properly managed and targeted this shift could pay dividends for Clinton in the traditionally Republican suburbs of major mid-western cities like Milwaukee (Wisconsin), Atlanta (Georgia), St. Louis (Missouri) and Denver (Colorado).

However, Clinton’s ability to keep ‘uneducated’, white, working-class, male voters in the Democratic camp is in serious question. There are indications that her appeal to white, male, non-graduates in the ‘Rust Belt’ (Wisconsin, Michigan, Indiana, western Pennsylvania and Ohio) is weaker than normal for Democratic candidates – the Party is traditionally associated with industrial workers and good union jobs. She is a defender of free trade deals signed by her husband and successive presidents and by standing for ‘continuity Obama’ is seen as accepting the continued decline of ‘old’ industries’ like steel-making, coal mining and even car manufacturing. The success of Bernie Sanders in places like Michigan on an anti-free trade platform underlines her potential weakness as a candidate in states like this.

The key, when thinking about white voters and especially white male voters, is will the number of affluent, independent and Republican-leaning graduates staying at home or switching to Clinton outweigh the effect of a loss in support from non-graduate, white, male voters? In Clinton’s favour is the fact that the more affluent and educated a person, the more likely they are to vote. Trump may be winning the support of people who are not regular voters and cannot be relied upon to actually register and turnout on the day, whilst losing some of the most reliable Republican voters.

Donald Trump: ‘Make America Great Again’

Which, leads me on to Trump’s strengths and weaknesses. Trump is not the favourite because he has not run a predictable, traditional or even well-organised Republican campaign. He has not worked hard to shore up the mid-western and southern, Christian base, has not tried to make inroads into black and Hispanic support and has not appealed consistently to white, suburban men and women whose primary concerns are the economy and taxation.

However, Trump remains strongest where the Republicans are strongest (South and Mid-West) and weakest where they are weakest (North East and West Coast). In that sense this is a very normal election.

However, Trump has alienated larger numbers of Hispanics and women than previous Republican nominees and so has enabled Clinton to capitalise and thereby expand the ‘Obama coalition’ in potentially fatal ways. Trump has also probably lost some ground amongst those who would consider themselves ‘moderate’ Republicans and conservative independents, particularly when considering Mitt Romney’s appeal was often very directly to these groups in 2012.

Trump’s great hope of overcoming the odds and winning the election lies with his appeal as an anti-establishment candidate, not as a generic Republican candidate. By attacking Clinton, the media and several high profile moderate Republicans, Trump has increased his appeal with non-graduate, white male voters who have previously voted Democrat or not voted in the last few elections. His slogan (the only memorable campaign slogan of the entire election): ‘Make America Great Again’ is a well-calculated appeal to those who feel bypassed or taken for granted by the ‘Obama coalition’ and ‘Obama economy’ since 2008.

The core question of the election is whether Trump’s campaign organisation is sufficient to translate this pool of sympathy amongst a large section of the American population into actual votes, particularly in states that could really hurt Clinton’s electoral college chances. In particular, Iowa, Michigan, Ohio, Wisconsin and Pennsylvania appear to be the best route Trump has to the White House, assuming Florida and North Carolina are won as well.

In this sense Trump’s campaign has more in common with Obama’s in 2008 than many would care to admit. His party base from prior elections is not enough to seal victory on its own. His success depends upon an ability to motivate a section of the electorate that his party has previously been unable to motivate in sufficient numbers to make an electoral college impact. Obama’s campaign was clearly better organised than Trump’s and he was all-round a more popular and polished politician but the pool of voters Trump is aiming at is much larger than the black Americans and young Americans of the ‘Obama coalition’.

Trump’s path to victory is more unlikely than Clinton’s (and certainly less likely than Obama’s in 2008 or 2012) but it does not make it unthinkable. A sizeable turnout of voters in the ‘Rust Belt’ who are traditionally non-voters and former Democrats, combined with an anti-establishment attack on Clinton that dissuades enough young and black voters to not vote for her or not vote at all and he can squeak over the line.


However, my prediction, for what it’s worth is that the shift of white, graduate, suburban, and particularly female voters in places like Virginia, Philadelphia (Pennsylvania) and Milwaukee (Wisconsin) towards Clinton (or at least away from Trump), combined with a rise in Hispanic turnout and Democratic allegiance will provide enough of a firewall in places like New Mexico, Nevada and Colorado to mask the frailty of the Obama coalition in Iowa, Ohio, Florida and North Carolina.

Effectively, if Florida and North Carolina go for Trump, all eyes turn to Pennsylvania, Michigan, Iowa, Ohio, Nevada and Colorado because Clinton will likely need 4/6 to survive.

If, on the other hand, Clinton wins two of Pennsylvania, Florida and North Carolina, it’s probably all over and you can get an early night.

If Trump wins Virginia, it’s going to be President Trump!

Useful websites for predictions and statistical analysis – The daddy of US political predictions sites. Its model famously predicted all 50 races in 2012. The highlight for me is the ‘winding path to 270 votes’ – The Snake. – Part of the 538 site. An incredible tool that allows you to see the practical effects of differential turnout from particular demographic blocs. Warning: highly addictive! – The New York Times has produced a very useful digest of all the major prediction sites which allows you to see which states you should look out for on the night. – very strong independent site that allows you to build your own map quickly and effectively and keep track of the latest polls and predictions. – A professional, yet Republican-leaning predictions site. – A professional, yet Democratic-leaning predictions site.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s