Emergent Wisdom

Thoughtful Observations from Business and LIfe

Why Polls and Election Results Are Sometimes So, So Wrong

I’m going to explain why pollsters got the result so wrong in the lead-up to the UK parliamentary general election in 2015, and simultaneously the real reason why parties, or candidates, win or lose. And use this as a basis for predicting the 2016 US presidential election.

Yet again, during the post-mortem of an election, politicians explain why they won or lost (primarily those who lost. Those who won are usually pleased to bask in the glory and relief), and pollsters are scratching their heads to explain why their polling turned out to be so incorrect.

The reasons both give for their failure - of either type - are usually blaming various factors supposedly outside their control - usually nefarious behaviour of the other party / candidate,  or broader circumstances over which they had no control. But if we take a look at both elections and polling predictions - both pre-voting and exit polls - superficially there doesn’t appear to be a credibly reliable explanation. The strangeness of polling errors is particularly interesting because at times the polls leading up to an election are wrong, and at times it’s the exit polls that are just as wrong. And it doesn’t happen just in the UK as the same kind of issues happen in the US.

Let’s take a look at some example elections: In 1992, the Conservatives had won three elections in a row under Margaret Thatcher, the economy had gone into recession, John Major, who had taken over from Margaret Thatcher, one of most intensely hated prime ministers in history, was hoping to win an unprecedented 4th general election for the Conservatives in a row, and all the polls indicated a Labour victory under Neil Kinnock. Even the exit polls indicated a strong Labour victory, but when the votes were counted it was an outright victory for the Conservatives. Had it been a third world country, allegations of vote rigging would have been made, and would have been seemed credible as the result was so different from everyone's expectations and ALL the polling. No-one credibly even suggested that is what happened in the UK as the fairness and transparency of the count is taken very seriously.

In the 2004 US presidential election, various exit polls indicated a significant victory for John Kerry over George W. Bush, which proved to be overestimates for John Kerry. Against many expectations, George W. Bush won again.

And in the 2015 UK general election, many polls agreed with each other - it was a dead heat between Conservative and Labour, and would almost certainly be a hung parliament again. But this time, the exit polls indicated a surprise jump in the Conservative vote. Former Liberal Democrat leader Lord Ashdown promised he’d eat his hat if the exit polls were correct. In the end, the Conservatives won a substantially unexpected overall majority of seats, higher than anybody’s expectations and surpassing all polling predictions, including the exit polls.

So what was going on then, that in 1992 the polls, and especially the exit polls, were hopelessly out, and in 2015 the polls except the exit polls were desperately wrong? And why did John Major win against Neil Kinnock anyway? And how did George W Bush beat John Kerry? And why was there such a difference in eventual voting between David Cameron and Ed Milliband?

A clue comes from the Liberal Democrat internal polling prior to the 2015 UK general election. Based on voters feelings for Nick Clegg, Lib-Dem leader and for the previous five years Deputy Prime Minister, the Lib-Dem vote looked very poor indeed. But when asked about their feelings for specific local candidates, their vote looked much stronger. This led Paddy Ashdown to make his infamous declaration about hat-eating, and Nick Clegg to say on the eve of the election that the Lib-Dem vote would be one of the surprises of the election. It was. Surprisingly bad to them, although broadly in line with polling expectations. So it appears that when people were asked about local candidates, their feelings were not a reliable indicator, yet when asked about the leaders of parties, their feelings did turn out to be reliable indicators. So from this survey, the leader made the difference that mattered.

Why would this be? To resort to my usual venue for these sorts of thought experiments, a stone-age village, lining up behind the right leader in a hostile world could be a matter of life and death. You only need to read a bit of the Old Testament to be aware of the often frequent hostility between different tribes, and the massive swings between the opportunity to take control of a village and its assets - including taking the men as slaves and the women as wives - or else the threat of it happening to you, to realise that the right leadership really could make the difference between living a life in luxury, misery, or death. So in that case, backing the right leader is likely not to be so much of an intellectual activity as a primal survival-based one, driven by primitive feelings about someone and their capability to lead you to glory or at least defend you from slavery and disaster.

When decisions are driven by primitive factors, what usually happens with human beings is that they rationalise with logical-sounding arguments what is actually a deeply emotional decision - a bit like the reason people pick a particular car. Often the list of reasons they’ll give will be rational ones, until the very last one, which will be the real reason. But sometimes people's intellectual preferences differ from their emotional drives.

This helps explain some election results. Around 20 years ago, Tony Robbins did his first UK ‘Unleash the Power Within’ seminar, which I attended, despite what is to British ears, a rather vomit-inducing title, given that we were not acclimatised to such grandiose claims. But during that seminar he pointed out in a number of UK and US elections some simple criteria that it was easy to correlate with election results. Essentially, he said the person who gets elected looks good, sounds good, and you believe they mean what they say. I’m going to amend that slightly, and say that in addition, if the mood of the nation is under threat, then it’s the person you most believe will protect you as an additional criteria, and if the nation feels safe, then it’s the person who you believe will most improve the economic situation, in addition to the criteria of looking good and sounding good. Who seems to be the most successful and powerful is an important leverage point in both of the protection / economic success factors, which is why people who are existing celebrities often do so well in elections. So despite Bill Clinton’s campaign claim that, “It’s the economy, stupid”, it might have been a factor at that time, but not always.

So lets take a look: In the 1992 Election, given that we’re looking at leaders now for election results and have swept aside other factors that determine the eventual outcome, it was John Major against Neil Kinnock. Let’s be honest, most politicians are not exactly beauties, but it’s fair to say that John Major was a lot more handsome than Neil Kinnock. You could say they were reasonably matched on their voices, John Major perhaps being easier to listen to for most people. Neil Kinnock had done some substantial leading of the Labour party away from some traditional and increasingly anachronistic hard-left positions, laying the ground for Tony Blair subsequently, but this was not really apparent to most of the public. John Major meanwhile, had seemed to be experiencing a meteoric rise, going from position to higher position and finally Conservative leader and therefore Prime Minister taking over from the increasingly unpopular Margaret Thatcher. It was as if he was leaping up the ladders on a snakes and ladders board like a circus acrobat, despite being, according to Rick Mayall’s Alan B’Stard of The New Statesman, “The only man known to have left the circus to join a firm of accountants”. He appeared to be indifferent to other people’s opinions, which itself is a sign of personal strength, while still genuinely caring about people,  in comparison especially with many politicians who seemed to be focussed more on ideology. At the time he looked and sounded like a success, So it’s not surprising that at the time he was chosen by the electorate, winning an overall majority. (that changed dramatically later, resulting in a Labour landslide when Tony Blair was swept to power). 

Much has been made of a rather foolish pre-election ‘victory rally’ that Labour held in 1992, basing their expectations of winning on the polls. That may have swayed some people, or caused some labour voters to stay at home, but that doesn’t explain the exit polls still pointing to a Labour victory that many people ignore in relation to that. In my view there’s another, more subtle, more primal factor that influenced the direction of people’s pencils when putting their X on the ballot paper: After the shock loss of the election, a devastated Neil Kinnock resigned. Not long after he became a European commissioner. When interviewed a few months later he seemed like a different person - happy, enjoying what he was doing, believing it was important, and critically, congruent. What it highlighted to me was by comparison how incongruent he had seemed as Labour leader. Not because he wasn’t happy being leader, but because he was essentially at that time forced to espouse policies he didn’t seem to personally believe in, even though he had dragged the Labour party away from some of its most extreme positions. So this incongruency seems to have weakened him in comparison with John Major even further, but perhaps at a more subliminal, emotional and less accessible level for people to explain why rationally. Just a feeling.

Why did George W. Bush defeat John Kerry in 2004? And how was he even elected in the first place? By the time of his re-election he was regarded as a buffoon in many circles, widely ridiculed for being stupid. Few politicians can have experienced having a documentary made about them as eviscerating as Michael Moore’s Fahrenheit 9/11, that achieved the distinction of being the most successful documentary of all time. And many questions were raised over his suitability prior to his even being elected in the first place.Yet he still won re-election as President of the US. But the difference I noted was that, while George W. Bush wasn’t particularly articulate (inventing words such as ‘misunderestimate’), he nevertheless seemed to be sincere in meaning what he said. He was congruent. Whereas John Kerry had taken at least one or two positions, such as opposing gay marriage, that he didn’t seem to personally believe in but seemed necessary to espouse for political reasons, and were out of step of being in favour of other protections and rights for LGBT people. In other words, being incongruent

And what about the 2015 UK general election? In 2010 Gordon Brown had made a significant number of gaffs and was just not good at connecting with the public that Blair had found so easy and comfortable. Yet David Cameron had not exactly seemed like Mr Charisma himself, either personally or in policy, espousing the rather nebulous 'big society' that no-one seemed to know what it actually meant. The result was the hung parliament that resulted in the Conservative coalition with the Liberal Democrats. In the next five years David Cameron had not made many colossal blunders, even though many people fiercely opposed the policy of austerity. Nevertheless, he stuck to it. By comparison, Ed Miliband was the surprise elected leader of the Labour Party, defeating even his own more well-known brother more widely regarded as a credible potential PM. The difference between them being for the ordinary voter, David Miliband seemed more comfortable in a public position whereas Ed Miliband seems a bit... well, sort of weird. He just doesn’t project ‘natural born leader’. In Private Eye magazine he is satirised in the Mr Milibean comic strip, a composite of himself and Mr Bean. It wasn’t always entirely clear what he was standing for, politically, whereas Cameron had been banging on about the same consistent message for the last five years. So the election result wasn’t decided in the last few weeks, it was the previous five years that mattered most.

As an example of time being important, the most high profile Labour politician to lose their seat was Ed Balls, the shadow chancellor (ie, the MP in the official ‘opposition’ party who is chosen by the official ‘leader of the opposition’ as the opposite number to the actual chancellor). He normally would have been regarded as having a ‘safe’ seat, with a substantial majority. But while he was busy being a high-profile shadow chancellor, the local Conservative candidate had spent the best part of the five years before the general election electioneering, eating away at his majority like woodworm.

And what of the Liberal-Democrats and the Scottish Nationalist Party? Some have credited the massive explosion of support for the SNP on a growth of nationalism in Scotland, and the collapse of the Lib-Dem vote on being punished for going into coalition with the Conservatives. There may be elements of truth in both. William Hague appears to have been one of the few astute enough to have seen at the onset of the coalition that by entering into coalition with the Conservatives with the inevitable substantial weakening of the Lib-Dem's goals and policy positions as the minor party, that it would ultimately nearly destroy the Lib-Dems, which on a parliamentary basis it nearly has. But commentators are not taking into account that the leader of the SNP, Nicola Sturgeon, had shown herself to be both competent and a confident communicator. Her performance in the leaders’ debates prior to the election, the first that most that the majority of English people had seen and heard much of her, had some English people asking how they could vote for her. And although the Lib-Dems had skilfully negotiated themselves a position in government out of all proportion to their political representation, their first time in proper power for 90 years, with Lib-Dems in some key ministerial positions, many of whom were perceived to have performed well, Nick Clegg himself appeared to be relatively weak and ineffective as deputy Prime Minister to the general public. At one point when David Cameron went on holiday, Nick Clegg was absent, apparently having forgotten that he was effectively Prime Minister while the PM was away and planned his holiday for the same time. That was a blunder of staggering proportions which substantially undermined his credibility, although it’s fair to say Lib-Dem support had already dropped substantially shortly after entering coalition - they had appeared to have substantially sold themselves out. Given that there were few realistic options, rationally the Lib-Dems maximised the leverage of their position and did effectively moderate the more extreme positions the Conservatives would otherwise have taken, but they were effectively doomed for having done so because voting is not an entirely rational act. In essence though, once again the apparent strength of the leader is consistent with the number of MPs in their party returned to Westminster.

But hold on - what about the polling errors? How can exit polls be so wrong, when they are supposed to represent what people have already done? As I have described, voting - ie, the actual moment that someone puts a cross in the box of their chosen candidate - is a primally-driven activity where they are actually voting for the leader, not the local candidate. The local candidate, in the number of cases that make the difference, is the means to the end of getting the right leader.

That we have a number of primal drives should not be a surprising notion - for example, eating, drinking, having sex (with the end point notionally being procreation, ie, survival of the genes, for most people). What can happen though is that there is sometimes an intellectual disagreement with the primal directives. For example, intellectually and consciously someone may be determined to have sex only with condoms in circumstances where they really wouldn’t want a pregnancy to result from sex on this occasion, for whatever reason. But what can happen is that at some point in the process, just in time, a biological program instigated by the primal drive to reproduce seizes the opportunity, takes control and directs the couple to have sex without condoms as if they have been taken over by some strange external force, until the point where the deed is done. It then relinquishes control and disappears out of sight, leaving the individuals suddenly bewildered and asking themselves, “What just happened? Why the heck did I do that???”. And what I’m suggesting is that sometimes this happens with voting. The person enters the voting booth, and as they are holding the pencil ready to put their mark, perhaps to support the local candidate of the party they have historically voted for on an ideological basis, they suddenly find the pencil veering in a different direction and their mark goes against a different candidate, to support the leader they feel more comfortable with at a primal level. Then afterwards, what they tell the exit pollsters depends on whether or not they felt guilty or bad about who they just voted for. If they feel ok about what they just did, they'll tell the truth, as appeared to have largely happened in the 2015 UK general election. But in 1992, many who voted for John Major (sorry - I mean technically they would have voted for the local conservative candidate) apparently didn't feel so good about it and didn't want to own up, so the exit polls still looked like a Labour Victory.

Polling prior to the election is more of an intellectual activity, unless the right questions are asked. Asked who people will vote for, they will declare based on different criteria to who their unconscious mind will cause them to vote for when they’re actually in the voting booth. These factors may coincide, but not always. If the pollsters don’t know which questions will reveal the actual truth, then they’ll predict the wrong results. And it’s apparent most pollsters don’t really know the right questions to elicit the actual results, either of who people will vote for, or who they just have voted for, reliably. So I suggest that the critical question revealing their actual voting behaviour is to score the leaders of each of the main parties. Best to bury that question amongst others of course, so it doesn’t specifically stand out, to get a more natural and truthful answer, but then the ratings of each party leader can be used as a predictor of eventual voting intentions, or who people actually voted for, regardless of who they say they will vote for or just did vote for.

So, applying that philosophy to the 2016 US presidential election, who will be elected? Well, I’m going to narrow it down to who, at this stage of the game, it appears the eventual candidates will be. I’m going to suggest it’s going to be Hilary Clinton V Jeb Bush, and here’s why: There is overwhelming support among hardcore Democrats for Hilary Clinton. She has almost been crowned already. And among the currently increasing field of Republican candidates, Jeb Bush will eventually prevail. If you ask yourself now who is ultimately more presidential, who you can actually realistically see in the white house, my perception is that Jeb Bush will keep chugging forward like a tractor while other candidates in the Republican field will fall away after briefly burning brightly.

And in the eventual beauty contest between Clinton and Bush? Let’s take a look: who looks good, sounds good, and seems as though they mean what they say? Who seems capable of protecting you in times of danger, yet helping the economy in times of economic possibility?

Leaving other questions aside for a moment, I’m going to suggest at this point that while Hilary Clinton is popular at some level with many people, a huge number of people have serious misgivings at a visceral level about her integrity, to a point where they may say they will vote for her, but when it comes to putting pencil to paper, or pressing the right button on the machine, their hand will stray to Jeb Bush. And that’s with only the issues that have come out of the woodwork so far. The thing with the Clintons - both of them - is that they seem to be constitutionally incapable of being consistently truthful and doing things with transparent integrity if there’s the opportunity to do it in a way that looks as though they’re covering something up that is actually for their own benefit. That means that between now and the actual election we can be sure that other things are going to come out of the woodwork. Yes, they’ll be explained away, but even among their supporters there will be that lingering feeling that it isn’t quite the innocent-seeming way they ‘splain it (let's be honest, they pretty much come as a pair).

There’s one other factor I haven’t mentioned, and this again goes back to primal issues that influence our actions. When I was at school and for some years afterwards I was kind of desperate to get off with girls but was mostly unsuccessful, apart from a couple of relationships I had with women as a young adult. I know now that part of it was a fear-driven desperation to prove my heterosexuality, whereas deep down I knew I was alwys much more attracted to guys. Yet, after I faced up to reality and let go of trying to date women, I found a lot more women coming on to me, and I finally twigged why: My ‘need’ had been driving them away. When I was relaxed and myself, I was more attractive. This was highlighted when I went to a bar with a straight friend who was specifically keen to meet a girl and have some fun, and I went just to be social. Despite the fact that I’m not exactly a male model and he was confident and attractive I couldn’t help noticing girls he approached were paying more attention to me than him, despite his frustration. And then I remembered that the guys who most often got the girls - plural - in school had always been the ones who seemed as though they couldn’t care less. Damn! If only I’d known at the time.

So how does this apply to elections? Let’s go back to the 2008 election when the battle was hotting up between Barack Obama and Hilary Clinton to be the Democratic nominee. From reading many accounts of that battle at the time, including accounts of behind-the-scenes events and the history of the two individuals and their families, there seemed to be a fundamental difference between the two: Barack Obama believed he would be a good president, and wanted to be president. Hilary Clinton seemed to need to be president, and that put people off. Another effect is that need-driven motivation is more like driving a car with the rev counter spending too much time in the red region. By the time he won the nomination, he looked great and on a roll, she was worn out to the point of being unwell.

If you read about Jeb Bush, he has hardly gone for any elected offices, and only when he believed he could do a better job. It feels as though he wants the job because he believes he’ll do a better job, whereas Hilary still feels to me as though she needs to be president. In not-fully-tangible ways this will affect her behaviour and what she says, her body language, and oh-so-many other subtle ways in which her true underlying values, beliefs and motivations are revealed to us at a subliminal level, regardless of the impression she actually intends to present. Indeed, in my view, this has been happening for quite some time already. And my belief is that, in the final analysis, this need-driven motivation will repel enough erstwhile Democratic voters to send Jeb Bush to the White House. And I don’t care what the polls say in advance, or even the exit polls on the night. And as usual the explanations of why it has turned out this way will be completely wrong.