As the Brexit negotiations grind towards a conclusion, there is much talk of what it means to honour the 2016 referendum result, and of whether another referendum should be held once the Brexit terms are known. A new paper by Stephen Fisher and the Unit’s Alan Renwick sheds fresh light on these issues, examining what people thought they were voting for in 2016 and how that affected their vote choice. In this post, the authors summarise the findings and draw out lessons for today’s debates.

With increasing discussion of the possibility of the UK holding another referendum on its relationship with the EU, it is important to better understand what happened at the last one. Understanding how voters made up their minds in 2016 could provide insights into how another referendum might play out. Also, one of the key arguments against another referendum is to maintain respect for the outcome of the previous one. What it means to respect that outcome depends on understanding why the UK voted to leave the EU.

In our recently published paper in Acta Politica (available free-to-view here), we focus on the role of voter expectations of the consequences of leaving the EU. Following previous research by Sara Hobolt and John Curtice showing that attitudes to the EU, including expectations regarding Brexit, were the most powerful and proximate predictors of vote choice at the referendum, we wanted to investigate further how Brexit expectations mattered, and whether it made a difference if voters did not have clear expectations. In particular, we wondered whether, perhaps because of risk aversion, uncertainty about the implications of leaving the EU was associated with Remain voting.

In April 2016, before the referendum campaign, the British Election Study (BES) internet panel asked people what they thought would happen with respect to various different economic and political outcomes in the event of the UK leaving the EU. For most of the outcomes the modal response was to say that things would stay ‘about the same.’ These outcomes included the economy, unemployment, international trade, risk of terrorism, rights for British workers, personal finances, British influence abroad, and the risk of big business leaving the UK. There were just two exceptions. There was a slight tendency for people to think that Scottish independence would be more likely and a strong expectation that immigration would be lower after a Brexit.

These expectations were strongly correlated with how people said they would vote in the referendum. That held even after we controlled for a wide variety of powerful background factors, including how people had voted in past elections, what they had previously said about their referendum vote intention, whether they approved of the government, the strength of their national and European identities, and their attitudes to prominent politicians. Our analysis suggests that expectations about the consequences of Brexit were important drivers of whether people opted for Leave or Remain, not just reflections of prior Euroscepticism.

Expectations of economic outcomes were the most important, but other outcomes also mattered. In accordance with John Curtice’s work, we found that while, on balance, people expected Brexit to lead to worse economic outcomes, this was offset by a large majority wanting and expecting Brexit to reduce immigration. This helps explain why the referendum result was relatively close.

Another reason the result was relatively close is revealed by our finding that there would have been a similarly narrow Leave victory had everyone expected Brexit to produce little change for any of the outcomes asked about. In accordance with those who argue that Britain’s membership of the EU has never been based on a shared European identity and so fundamentally transactional, we found that British voters would on balance have preferred to leave if they had thought little would change.  

What of those who did not know what to expect? On average, responding ‘Don’t Know’ to a question about what would happen if the UK left the EU was associated with a greater chance of voting Remain. The BES also asked directly how certain respondents felt about what would happen in the event of a Remain or Leave vote. Those who expressed uncertainty over the effects of a decision to Remain were more likely to vote Leave than others – and vice versa. Voters are unsurprisingly reluctant to vote for something when they are unsure of its consequences.

Election and referendum campaigns should enlighten people about the choice they face. If so then people should have become less likely to report not knowing what would happen in the event of Brexit. There was some, but very little, sign of this in the BES panel. More generally there was very little change in Brexit expectations or vote intention over the course of the campaign in the BES panel. This is in line with the experience of the internet opinion polls, though not those conducted by telephone. Although modest, switching vote intention between Leave and Remain during the campaign was associated with corresponding changes in Brexit expectations.

Those who were still undecided in the final two weeks of the campaign eventually divided 53:47 for Remain if they did eventually vote. This split is fairly even, partly because most voters who were still undecided after an intense campaign were apparently more indifferent than uncertain. Among those who were more uncertain, in the sense of saying ‘Don’t Know’ in response to more questions about the consequences of Brexit, the eventual vote split was more favourable to Remain, but the numbers are small and statistically insignificant.

Our paper goes into more detail, but the main conclusions are that expectations and uncertainty about the consequences of Brexit appear to have affected how people voted, even though the eventual balance of considerations led to an overall outcome that was much the same as would have transpired if people had thought Brexit would not change much.

Understanding what Leave voters expected at the time of the referendum is important for the current debate about whether the UK should have a referendum on the terms of Brexit. Perhaps the most common argument against is that hosting another referendum would disrespect the outcome of the previous one. It is not quite clear whether this means that the UK must leave the EU regardless of the circumstances, or whether it means that the UK should get the kind of Brexit it apparently voted for in 2016.  

If respecting the referendum result means respecting what the Leave voters expected to get in areas where their expectations were relevant to their vote, then any Brexit settlement should primarily lead to a reduction in immigration and economic outcomes that are at least as good if not better than before. During the campaign 80% of Leave voters expected the former and 89% of them expected the latter, and both were highly pertinent to their choice.

These findings from the BES are in line with opinion polls before the referendum which showed a clear majority for Brexit based on the desire to limit immigration. Those polls also showed that if people at the time were forced to choose between accepting free movement as a price of free trade in the EU, or abandoning both, the contest would have been close. But many did not think they faced such a trade-off. One poll before the referendum showed 57% of Leave supporters believing that, if Britain left, ‘it would probably be possible for Britain to negotiate free trade with the rest of the EU, without having to allow EU citizens to live and work in Britain’.

If there is another referendum, and if the drivers of vote choice are similar to last time, then attempts to influence expectations about the consequences of the outcome will be vital to the result. However, given recent research showing that voters now think of themselves as either Leavers or Remainers, opinions about the costs and benefits of EU membership might now be relatively fixed and difficult to change.

About the authors

Stephen Fisher is associate professor of political sociology at the University of Oxford.

Alan Renwick is Deputy Director of the Constitution Unit, at University College London.