On June 8, 2017, British citizens will vote in general elections. Prime Minister Theresa May called snap elections with the explicit purpose of securing a strong Tory majority in Westminster after the political turmoil triggered by the outcome of the June 2016 referendum on the United Kingdom’s membership of the European Union. “Brexit” represents the first attempt by a long-standing member state to abandon the EU, and perhaps the most dramatic moment in the process of European integration so far.

By using data from the REScEU mass survey conducted in the United Kingdom in the fall of 2016, this article investigates some factors behind the outcome of the referendum, and respondents’ preferences about future UK-EU political and economic relationship. The article focuses in particular on the role played by immigration and Britons’ “welfare chauvinist” attitudes.

Explaining anti-EU attitudes

Several scholars explain voting behaviour in EU-related referendums relying on voters’ values and preferences for European integration. They take into account factors such as respondents’ cognitive skills (Gabel 1998) and post-materialist values (Inglehart 1997), self-motivated cost-benefit analysis (Anderson and Reichert 1995; Gabel 1998), culture and identity patterns (McLaren 2002; Hooghe and Marks 2004), and anti-immigrants feelings (de Vreese and Boomgardner 2005).

Another stream of literature argues that the key predictors of attitudes and behaviours related to the EU come from the domestic politics. Citizens tend to rely on cues from the domestic political realm, such as their satisfaction with the democratic process, their evaluation of the national economy, and their trust in parties and the main political institutions (Franklin et al. 1994; Franklin et al. 1995). An important role is also played by partisanship: parties act as a filter for the interpretation of political messages, and cues and help poorly informed voters to choose which outcome is closest to their ideal point (Hobolt 2006; 2007; 2009).

The relevance of immigration and welfare chauvinism

Infographics #9 #10 and #11 have already showed our sample’s opinions on the Brexit referendum and on post-referendum scenarios. Here, I empirically test what factors affect two different dependent variables related to Brexit. The first is the probability that respondents voted Leave in the referendum, while the second is the probability that they now prefer a “hard Brexit”, i.e. the idea according to which the United Kingdom should give up full access to the European single market and to the free movement principle.

The regression models includes several predictors related to the theoretical arguments illustrated above, such as the socio-economic status of respondents, their perception of material deprivation, their political sophistication and degree of cosmopolitanism, their support to the integration process and their attitudes toward the openness of the domestic welfare state to immigrants. The model also includes respondents’ satisfaction with the way democracy work in the UK, their evaluation of the national economy and the party for which they voted in the 2015 general election.

Respondents’ perception of a relative material deprivation, measured through the evaluation of their household financial situation in the last five years, has no significant effect on their voting choice in the referendum, but it positively affects their preference for a hard Brexit.

Figures 1 and 2 show, respectively, how much the probability of having voted Leave and the probability to prefer a hard Brexit changes according to a variation in the predictors that are significantly associated with the two dependent variables in the regression models.

Figure 1: The impact of statistically significant predictors on the probability that respondents voted Leave.

Figure 2: The impact of statistically significant predictors on the probability that respondents prefer a hard Brexit.

In line with some previous studies (e.g. Clarke et al. 2017), once one controls for the effect of attitudinal variables, age, gender, occupation and education are not significantly associated to “Brexit”. Respondents’ perception of a relative material deprivation, measured through the evaluation of their household financial situation in the last five years, has no significant effect on their voting choice in the referendum, but it positively affects their preference for a hard Brexit. Respondents who have visited another country for work, study or leisure have a 14% lower probability of having voted Leave. Furthermore, they have a 6% lower probability of preferring hard Brexit. Unsurprisingly, the predicted probability of having voted Leave by respondents who think that the integration process should be strengthened is 42% lower compared to those who think that integration has already gone too far. Furthermore, pro-integration respondents have an 11% lower probability of preferring a UK-EU agreement that restricts free movement.

For some respondents voting in the referendum represented an occasion to voice their protest against the EU.

However, the attitudinal variable that display the strongest impact on the Brexit vote is the respondents’ opinion on the openness of the domestic welfare state to immigrants. Compared to respondents who think that foreign citizens should have access to the same social security benefits as natives, those who opt for a restriction of welfare benefits only to citizens of other EU member states (“EU bonding”) show a 17% higher probability of having voted Leave, and a 10% higher probability to prefer a hard Brexit. Choosing to restrict welfare rights only to British citizens (“welfare chauvinism”) instead increases the predicted probability of having voted Leave by 43%, and the probability to prefer a hard Brexit by 19%.

In addition, partisan cues played an important role in shaping the outcome of the referendum. Compared to those who in 2015 elections voted for the Liberal Democrats—the party with the most Europeanist stance—Tory voters have a 25% higher probability of having voted Leave, while the same predicted probability increases by an impressive 53% for those who voted UKIP. Even those who casted a blank ballot in 2015 have a 25% higher probability of having voted Leave, signalling that for some respondents voting in the referendum represented an occasion to voice their protest against the EU.

What is behind welfare chauvinism?

So far, empirical results show that attitudes toward cross-borders welfare rights played a key role in explaining voting behaviour in the referendum and attitudes toward a total withdrawal of the UK from the EU. I took a further step by investigating what reasons lie at the basis of the association between welfare chauvinism and Brexit.

Scholars adopt two main logics to explain why citizens of Western democracies fear immigration and opt for “demarcation” (Kriesi et al. 2008; 2012). The “cultural approach” argues that immigration is mainly associated with cultural values that have more to do with conceptions of national identity rather than concerns about personal economic circumstances. Therefore, European integration bringing together people of different countries, regions and cultures and with different ethnicities and religions, threatens the cultural identity of resident populations who seek to defend their traditional way of life. The “economic competition” approach instead argues that massive immigration, especially in a period of economic stagnation and rising unemployment, may intensify competition over scares resources, such as jobs and social security benefits. This is particularly true for unskilled workers in those economic sectors more exposed to foreign markets, and for recipients of welfare benefits, such as the retired or the unemployed.

I tested these two arguments by making the variable measuring respondents’ attitudes toward cross-borders welfare rights interact with two exogenous variables measured at the regional level: the 2011-2015 net difference in foreign-born residents, and the difference in the unemployment rate in the same period. If Brexit has been mainly driven by the respondents’ perceived cultural threat we should expect a stronger effect of welfare chauvinism in those regions where citizens experienced a “cultural shock” determined by increasing migration flows (Ward et al. 2001). On the contrary, if Brexit was caused by respondents’ self-motivated cost-benefit calculations, we should expect a stronger impact of welfare chauvinism in regions where the unemployment rate decreased less from the peak of the economic crisis. These two hypotheses are not necessarily mutually exclusive.

This finding runs against one of the most common narratives of Brexit, according to which the latter is mainly driven by fears of cultural diversity associated with migration flows.

Figure 3 shows two graphs plotting the marginal effect of a discrete change in welfare chauvinist attitudes on the probability of voting Leave (first graph), and the probability of preferring a hard Brexit (second graph). The marginal effects are computed at changing values of the net difference in the share of foreign-born residents in British regions in the 2011-2015 period and while holding other covariates at their mean.

Figure 3: Marginal effect of a discrete change in welfare chauvinist attitudes on the probability of having voted Leave (graph above) and on the probability of preferring a hard Brexit (graph below) for different values of the 2011-2015 net difference in the share of foreigners.

The two graphs reported in Figure 3 clearly show that the impact of respondents’ welfare chauvinist attitudes on both the probability that they voted Leave and they prefer a hard Brexit does not get stronger in regions where the share of foreign-born residents increased between 2011 and 2015. This finding runs against one of the most common narratives of Brexit, according to which the latter is mainly driven by fears of cultural diversity associated with migration flows.

Figure 4 displays two graphs analogous to those in the previous figure, in which the marginal effect of welfare chauvinism on Brexit attitudes is computed for different values of the net difference in the unemployment rate between 2011 and 2015 in the forty British regions.

Figure 4: Marginal effect of a discrete change in welfare chauvinist attitudes on the probability of having voted Leave (graph above) and on the probability of preferring a hard Brexit (graph below) for different values of the 2011-2015 net difference in the unemployment rate.

As the two plots show, welfare chauvinism has a stronger impact on both the probability that respondents voted Leave and, above all, that they prefer a hard Brexit in more depressed regions , in which unemployment decreased less between 2011 and 2015. These results provide empirical evidence in support of the economic competition argument. Respondents’ desire for the United Kingdom to “take back control” of its borders seems to reflect their fear of losing jobs and social security benefits in favour of foreign citizens.

Political implications

The above findings confirm the salience of the issue of free movement and the relevance of Britons’ attitudes toward the openness of the welfare state to immigrants in explaining Brexit (Glencross 2016; Clarke et al. 2017). This article also underlines an important political implication for the current UK-EU negotiations. Even though Theresa May will be able to secure a strong majority for her party in Westminster, she will face an apparently insuperable challenge: she has to try conciliating UK participation in the EU single market, as requested by major British companies, with the spirit of the referendum outcome, which demonstrated deep popular dissatisfaction with the EU’s intrusion in British sovereignty, especially when it comes to free movement.

Photo Credits CC Ben Chapman