How the government’s pro-remain leaflet shaped the EU referendum

New research shows how the official government leaflet successfully changed voting behaviour in the referendum

Since the 2016 EU referendum, ‘who voted for Brexit?’ has quickly become a compelling area of study for academics and the media alike. In spite of the growing attention, the impact of the government’s pro-remain leaflet on individual voting intentions is unaddressed. My new research provides evidence that the leaflet did indeed have a significant impact on individual vote preference.

In the form of a glossy 16-page document, the government’s mailshot was sent to all UK households between the 11-13th April 2016. The document entitled ‘Why the government believes voting to remain in the European Union is the best decision for the UK’ ubiquitously made the case for why people should vote to remain in the EU. The leaflet’s primary focus was on three key areas: protecting jobs, a stronger economy and providing security. The leaflet also provided information on areas such as the cost of living; travel abroad; controlling immigration; and the importance of the referendum outcome.

The individual-level characteristics that are associated with voting leave are becoming well established in the growing Brexit literature. Empirical studies have found that areas which voted relatively highly for ‘leave’ closely map with support for UKIP (Goodwin and Heath, 2016). Research has also shown how it was fundamental characteristics of the voting population (such as low education, low income and areas of high unemployment) that were key drivers of the leave vote, rather than exposure to immigration and trade (Becker et al, 2017). The trope that the elderly caused Brexit has come under scrutiny recently as new studies have begun to present conflicting evidence (Liberini et al, 2017). Moreover, the same authors argue that dissatisfaction with one’s own financial situation, rather than a sense of general unhappiness, was associated with being more pro-Brexit.

From an underlying policy evaluation perspective and in light of growing calls for a second referendum, it is imperative to better our understanding about how and why the leave vote prevailed on 23rd June 2016 and what the government could have done to more effectively, in respect to their leaflet, support the remain campaign.

In my new research paper, using a dataset of voters from across the UK and exploiting the fact that not every person read the leaflet, I examine the causal effect of being exposed to the government’s leaflet on voting intention in the referendum. The data are gathered in a survey by the British Election Study in the weeks prior to the referendum and the main question of interest is whether a person has ‘received and read’ the leaflet. To explore what drove this effect, I split the respondents into two groups; those who were exposed to other referendum information (such as the televised debates and other sources of news) to a high degree or to a low degree.

My results show that exposure to the government’s leaflet lead to a lower probability of voting to leave the EU. The findings indicate that exposed voters were, on average, 3 percentage points less likely to vote leave relative to their matched control observations; that is, non-exposed individuals with similar observable characteristics as the exposed individuals. The effect is primarily driven by those individuals who read the leaflet and had a low degree of exposure to other referendum information. Specifically, in this subsample, the drop in the probability of voting leave is larger in demographics that are more liable to persuasion bias. That is, the groups of people that previous studies have shown are more likely to be affected to a greater extent. For instance, women and the risk adverse were on average 8.8 and 10.2 percentage points less likely to vote leave after leaflet exposure, respectively.

Furthermore, I find that Conservative party supporters who read the leaflet and were exposed to a high degree of other referendum information were around 6.2 percentage points less likely to vote to leave. It seems likely therefore that some Conservative supporters will have interpreted the leaflet as a signal from the Conservative government to vote remain. In both sample groups, exposure to the leaflet is shown to significantly increase one’s likelihood of changing their voting intention from leave to remain.

The drop in the probability of voting to leave is a result of leaflet readers becoming more informed about a particular scenario should a leave outcome prevail, which shows the persuasive nature of the leaflet. For example, the leaflet made the case that the economy would be worse off should the UK vote leave and exposed individuals were more likely to believe this. This ‘persuasion’ mechanism of impact is confirmed by correctly finding no effect of exposure for a number of placebo tests using information not contained in the leaflet, for instance, whether a person believes the EU has undermined UK parliamentary sovereignty.

In the wider context of this finding, the results are supported by previous work by Gerber et al. (2011). They examine the impact of a mailing on the 2006 attorney general election in Kansas and find a comparable positive effect for campaign mailing. Moreover, they argue that campaign mailing in a low information environment should have a larger impact, which is validated by my research. This ‘low information environment’ refers to a situation where voters are less informed or have had limited access to vote forming information.

The findings of this research have important policy implications for future political, and referendum, campaigns. Clear, realistic information sent to the electorate can be a viable persuasive tool for such campaigns. In particular, when such information is targeted towards demographics that are inherently more susceptible to persuasive information, the effects can be substantially larger.

In summary, the research has shown the bounded success of the government’s leaflet of changing voting behaviour in the EU referendum. In a referendum characterised by a huge amount of rhetoric, information and statistics -some of which was misused- the leaflet was most effective for those individuals with limited other referendum information exposure and Conservative supporters who observed their publicly split party.