Five years ago today, the UK voted to leave the European Union. This was a significant political upset; Brexit had been officially opposed by the major parties in parliament, the Prime Minister, and a clear majority of scientific and economic experts. Added to this, the vote was close; slightly less than 52 per cent of voters were in favour of Brexit and slightly more than 48 per cent voted against. The result fell within the margin of error of many pre-referendum polls and contradicted many of them which had expected a narrow Remain win. The vote is also noted for the fractious nature of the debate that surrounded it, not least in being linked to a very significant rise in hate crimes, along with the murder of an MP. If the hope of then-Prime Minister David Cameron had been that the referendum would end the EU debate in the UK, he would be sorely disappointed. Almost immediately after the vote, there were those who said that the result was too close to allow for such a fundamental change. Echoing key Leave supporters who had previously argued that a 48-52 split in favour of Remain could justify a second referendum, a petition (ironically started before the vote by a Leave supporter) calling for another referendum in the case of a close result quickly gained over a million signatures.
Such debate about the legitimacy of political votes is very uncommon in the UK. Even in situations in which people see their preferred options lose, it is generally accepted that electoral decisions are legitimate and that the consequences that flow from them are therefore also legitimate. The standard narrative goes that this contestation over the Brexit referendum was a result of it being particularly divisive; there was therefore more disagreement about the legitimacy of the outcome – but this was an aberration, which does not reflect broader feelings about the system as a whole. However, this narrative downplays the serious issues that can be seen in that public’s faith in elections more generally.
After the 1997 general election, the national election study asked people whether they felt that the election had been conducted fairly or not on a five-point scale (from ‘conducted unfairly’ to ‘conducted fairly’). Recoding responses onto a -1 to +1 scale, where -1 is conducted unfairly and +1 is conducted fairly, the average response in the UK was 0.67 – a reasonable but not overwhelming endorsement of the fairness of the election, and about average for a range of democratic countries (for a full comparison of 17 countries, see Losers’ Consent by Chris Anderson and colleagues, 2005: 40). Versions of this question have been repeated in the British Election Study Internet Panel since 2014, sometimes as a prospective question about an upcoming election and sometimes as a retrospective question about a just-passed election. In general, around 20,000-30,000 people are questioned* – selected in such a way as to give an approximately representative national survey. In May-June 2014 we asked a prospective version of the question (how fair do you expect the 2015 general election to be). The average score was 0.38 – already way down on the 1997 score. We asked another prospective version of the question ahead of the 2017 general election (in late April to 3 May 2017) and the average had fallen sharply again to 0.22. In early June 2017 we asked a retrospective version of the question (how fair was the 2017 general election) – directly equivalent to the 1997 question – and found the average response was down to 0.21. Returning to a prospective question in early November 2019, we find an average score of 0.22. In mid-December 2019, almost immediately after the 2019 election, the average response had fallen to 0.19, a new low for UK general elections.
When we compare this to the EU referendum, it is clear that it was seen as being less fair across the board – but not by as much as might be expected. In April-May 2016, shortly before the referendum, the prospective evaluation of how fair the referendum would be was 0.15. Immediately after the referendum (24 June-4 July 2016) this had fallen to 0.06. by November-December 2016 it was 0.07, but fell even further by April-May 2017 to stand at 0.05. Interestingly, in March 2019 a prospective question was asked about how fair people thought a potential second referendum on the UK’s EU membership would be. The answer here stood at -0.05, which means that on balance the UK public would expect the vote to be unfair as a starting position.
Clearly beliefs about the fairness of the EU referendum are very concerning – evaluations of the 2016 vote fall to a point where there is only a slight majority of opinion viewing the vote as fair. But when seen in context of perceptions about recent general elections, it is hard to maintain that the EU vote was an aberration. Instead, we have seen a collapse in positive beliefs about the fairness of elections across the population.
When we analyse the figures further, the results become more concerning still. In the literature there is a well known ‘winner-loser’ effect following elections. In short, on a wide range of measures people who have just seen their preferred party win the election will report being more satisfied with the process than those who saw their side lose. This may partly be due to a simple psychological effect – winning makes people happy, losing makes them sad; happy people are more likely to respond positively to questions, sad people are less likely. But when the effect is large, and when the issue at hand is a core aspect of the legitimacy of the vote, discrepancies are more concerning. Yet as Cees van der Eijk and I have argued in a recent paper (which can be found here) this is exactly what we see in the case of the EU referendum. These effects are very large – across the whole population the changes between winners and losers in beliefs about the fairness of the Brexit referendum represents a shift of over 40 per cent of the range of the variable (with Remain voters getting much more negative and Leave voters getting much more positive) – and beliefs about the fairness of the vote are a core metric of democratic health.
We see a similar pattern for the 2017 and 2019 general elections, although the specific effects are interesting. The 2017 general election was called by an incumbent Conservative administration, and the Conservatives won the election in the sense that they again formed the government, and the Prime Minister remained the same person. Yet the Conservatives notably underperformed relative to expectations and had to rely on support from the DUP to get their agenda through. At the same time, many other parties performed better than expected, including the Labour party under Jeremy Corbyn, which lost by less than people had assumed it would. While the literature consistently says that those who form the national government are the ‘winners’, evidently this upset left many Conservative voters not feeling like they had won. As such, we still see significant winner-loser effects in beliefs about the fairness of the election, but this time with Conservative voters actually becoming more negative (with a change of about 16 per cent of the range of the variable).
The 2019 general election was again called by a Conservative administration (now under Boris Johnson), returned a Conservative government, and the Prime Minister remained the same person. Yet in this election the Conservatives scored a significant victory, not only delivering a large majority but also capturing seats that had been held by Labour for generations. Perhaps unsurprisingly, Conservative voters now felt like winners. Particularly relative to Labour supporters, Conservative supporters following the election became significantly more positive about the fairness of the election – by a little over 11 per cent of the range of the variable.
On the one hand, we may be tempted to try to take comfort from the fact that the sizes of the effects are lower in general elections now than they were for the EU referendum. But this only makes sense in a world in which we are happy with significant differences in beliefs about the fairness of elections dependent on who ends up ‘winning’ (or feeling like they won). In 1997 there was no difference at all in evaluations of the fairness of the election between Labour and Conservative supporters, notwithstanding that Tony Blair delivered a Labour majority of unprecedented size in the post-war period.
Clearly there has been a massive political shift in attitudes to democratic elections since 1997. This seems to have happened largely without comment, and without a concrete plan to address the issue. Something is rotten in the state of Britain; but like the first sighting of a ghost haunting Denmark’s royal palace, at present we can see the effects but do not yet have a clear understanding of the cause or how to address it.
*Note that the May-June 2014 sample was only 5193 people, because the question was asked to a random subset of the full sample.
Dr Jonathan Rose is Associate Professor in Politics and Research Methodology at De Montfort University.
The Constitution Society is committed to the promotion of informed debate and is politically impartial. Any views expressed in this article are the personal views of the author and not those of The Constitution Society.