Like many others last month, I believed the UK would vote to stay in the EU. In fact, shortly before the referendum, I argued that the balance of power among MPs and within the government was such that the outcome seemed crystal clear. Here I ask whether there is any justification for basing an analysis on elites (in this case, MPs) when predicting the results of a referendum? And, second, why exactly was my prediction wrong?
There are obviously no names of candidates or parties on referendum ballots (as distinguished from election ballots) – and citizens express their opinion about a problem they may not understand. This creates a situation where the voter seeks some guidelines that will help him or her to understand the question and take what appears to be a conscious decision. Thus it becomes of increased importance for parties to wage campaigns in the run-up to the referendum.
Researchers (among others — M. Franklin, M. Marsh, L. McLaren 1994; S. B. Hobolt 2006; P. Sciarini and A. Tresch 2011) are of the unanimous opinion that the more monolithic the political class (so not only the ruling party but also the opposition, including non-parliamentary), the more guaranteed the result. A schism in the elite creates the conditions for an unexpected outcome. Where there is disagreement between parties, there is a high probability that a voter will vote on a single issue on this basis of his or her own opinion rather than out of solidarity with a party or for tactical considerations.
Thus, there is no doubt regarding the position of political parties as an indicator. And so using it as a basis for my hypothesis appeared justified. It was also important from the outset to determine the degree of cohesion in the elites and their level of voter support. To establish this, I did not use public opinion surveys because of their reputation for ambiguity. Instead, I used data on the alignment of forces within the elite (government and MPs) and the experience of previous referendums. Firstly, this included the BBC and ConservativeHome who provided regular reports on the stance of MPs throughout the campaign. It is obvious there that the ratio of those supporting the UK’s membership in the EU to the Eurosceptics was approximately 3:1 in parliament and 4:1 in government. This had persisted throughout the entire campaign. One cannot say this was a sudden “last moment” alignment.
Secondly, analysis of other recent referendums in Europe had revealed a pattern and made it possible to verify a hypothesis on the decisive role played by elites, to compare how politicians’ positions translate into some expected result. Moreover, the experience of most recent referendums in the United Kingdom in 2011 and 2014 gave weight to the idea that pro-EU supporters would outnumber their rivals decisively. In each of those cases, voters’ choice was in line with the government and a majority of MPs even though there was a loud and reasonable rival campaign. In addition, the June referendum in Switzerland is a convenient example for analysis. The large number of questions (five) and the parties’ varying positions also confirmed the thesis regarding the crucial role played by the government and politicians.
Thirdly and finally, it was important to understand how strong the ties between voters and parties were and establish whether there had been a growth of Euroscepticism. A General Election was held in the UK only in May 2015. Although won fairly decisively by the Conservative Party (11.3m votes) the election showed an impressive growth for the Eurosceptics in UKIP, who received nearly 4m votes. The local elections held in early May, a mere month and a half before the referendum and at the very moment when politicians were waging their campaigns full on, could have presented another, exceptional, opportunity to see the strength of anti-European sentiments. Here, we can see that the Conservatives and Labour lost seats in England while the UK Independence Party (UKIP) once again performed strongly, gaining many local council seats. However, the growth of the Eurosceptic UKIP was perfectly balanced by the distinctly pro-European Liberal Democrats. And the total figures regarding the number of seats and controlled councils did not change or only changed a little bit. The local elections did not reflect mass displeasure with the elites overall or pro-European parties in particular.
Thus I came to the conclusion that despite the rather sweeping Eurosceptical sentiments, the referendum would have to be the indisputable triumph of those advocating for the United Kingdom to remain a member of the European Union.
So why then did my prediction fail? A number of theories about why Remain lost emerged after the referendum. They included:
Some of these explanations place the method of elite analysis under suspicion, indicating the general dissatisfaction people have with the establishment as a whole or with certain leading politicians in particular. However, these suppositions were not, for some reason, corroborated in the 2015 general election or the 2016 local elections.
Instead of a broad lack of trust, it would be more appropriate to speak of dissatisfaction or lack of confidence specifically in connection with the issue of a united Europe. Then this referendum is the same in spirit as a whole series of past plebiscites: Denmark 1992, Ireland 2001, France and the Netherlands 2005. One can also add to this list the cancelled referendum in Ireland that was supposed to take place in 2005 or 2006. In each of these cases, the political elites had a favourable opinion towards change but the population clearly said “no”. And overall, according to the results of those referendums, many of which were “re-played” later, the idea was less about trust in elites and more about unwillingness to accept something “by default” or to agree to the “partial loss of sovereignty”.
These interpretations and precedents are very important for prognosis. So it was necessary not to analyse all the referendums but to identify those specifically relating to the pan-European constitutional problem. But there is nothing to be found here which answers the question of how such a gulf could emerge so suddenly between politicians and voters in this particular referendum. It is all the more strange considering the fact that in the United Kingdom the ties between voters and their MPs are traditionally strong while the latter, in turn, are actively engaged in their constituencies.
In an electoral campaign, candidates are normally using every form of communication with the voter – meetings, canvassing, distribution of leaflets, phone calls, social networking activity, and so on. One could hardly have assumed that voters would defiantly vote against the requests of their individual representatives on a fundamental issue. And the requests of the overwhelming majority of MPs and ministers were going to have to guarantee the sure victory of supporters of the status quo. Thus, taking into account the ties between politicians and voters in the United Kingdom and the results of the most recent elections, the only explanation that can be suggested here is that there was a failure in the local campaigns of each and every one of the pro-European politicians.
From the beginning of the year, the Universities of Nottingham and Kent maintained records on local events dedicated to the referendum. By early June, the researchers calculated that both camps had held over 3.5 thousand events and the pro-European politicians had significantly outstripped their competitors. However, the majority of events were held in the capital. “In British politics, you often hear the complaint that few people like to travel outside of London and our data supports this view,” wrote Matthew Goodwin, head of the group.
The visual aids they provide in map form show that intensive activity was undertaken only in those individual English constituencies that were considered key. In the rest of the constituencies, there was little or even no activity. And even if the pro-European politicians’ campaigns went all out in the final weeks, that would hardly have altered the moods left to drift for the previous five months in any fundamental way. (For comparison, see the map of referendum results.)
My prediction of the EU referendum result did not take two factors into account: the experience of referendums specifically on European constitutional issues, where citizen’s initial Euroscepticism is seen, and the particular characteristics of campaigns on the local level, where MPs simply did not consider it necessary to put all of their efforts into defending their point of view. It will be interesting to see if this correction to a prediction approach works for future referendums.
Andrey Kulikov is director and editor-in-chief at Europe Insight. He tweets @EI_lead.
