A consolidated academic literature argues that the interplay between the consequences of globalisation and European integration has led to the emergence of a new political conflict, which polarised Western societies into winners and losers of globalisation (Kriesi et al. 2008; 2012; Teney et al. 2014). The former are often social classes that benefited from the political and economic integration of the European Union. The latter are citizens with a low level of education, with low skilled or temporary jobs, who feel threatened by processes such as the increasing number of migrants, the freedom of movement of workers and the delocalisation processes of manufacturing companies. The two groups of citizens support different positions with regards to European integration and the opening of borders, with losers in favour of more “demarcation” and the imposition of limits on free movement of workers and their access to social benefits. On the contrary, winners feel more at ease in an inclusive and globalized world and tend to express open-minded attitudes. Furthermore, while winners tend to vote for mainstream parties, losers tend to support new or “renovated” populist and Eurosceptic parties, especially from the far-right.

The latest national elections in Italy held on the 4th of March 2018, represent an interesting case to test the validity of the arguments put forth by the “globalisation losers theory”. Italy was strongly hit by the economic recession and sovereign debt crisis, which led to low levels of economic growth and a high unemployment rate. Moreover, electoral results marked a rise of national consent for Eurosceptic parties, such as the antisystem “Five Stars Movement” (M5S) and the far-right “League”. Are social groups defined as losers more sceptic towards the EU and the integration process, and more likely to vote for Eurosceptic parties? We carried out our analysis by using data taken from a pre-electoral national survey conducted by the Pastel2018 project of the University of Milan and IPSOS in collaboration with ITANES, and REScEU. The survey was conducted with a sample of 1,323 voters categorised by age, gender, and area of residence through the CAWI method, between the 24th of January and 13th of February 2018.

Defining losers

Following the seminal works of Kriesi et al. (2008; 2012), we have defined globalisation’s winners and losers mainly by looking at their socio-economic backgrounds. In particular, we emphasized two main factors polarising the electorate: education and employment status. Education levels capture both different citizens’ skills, as well as how much they value cultural liberalism. The employment status captures instead whether the economic sector, as well as a citizen’s status, is threatened by the pressures coming from economic integration. Therefore, we identified losers as respondents who are poorly educated (lower secondary attainment or less), manual workers, unemployed and those with a temporary job. All in all, losers constitute 32% of our sample. Unlike winners, losers are more concerned by problems such as unemployment and immigration, emphasize their economic and cultural vulnerability, and consider issues such as the economic growth and public debt less salient. With regards to the respondents’ political orientations, we have detected that losers are less prone to express left-wing sentiments than winners, while there are no significant differences for centre and right categories. Interestingly, the percentage of respondents that refused to locate themselves on the traditional left-right political spectrum increases from 25% among winners to 37% among losers.


Figure 1. Winners and losers self-location on a 0 (Left) – 10 (Right) scale

Attitudes towards the EU and the Euro

To test the potential association between losers and Eurosceptic attitudes, and the tendency to vote for anti-system parties, we have regressed a battery of dependent variables on education and occupational status. We have also included gender, age (18-44; 45-64; 65+), geographic area of residence (North; Centre; South) and economic status (live at ease; live without difficulties; live with some difficulties; live with a lot of difficulties) as controls.

First, we analysed respondents’ attitudes towards the European Union and the Euro collected during the electoral campaign. An extensive literature (Hooghe and Marks 2009) has emphasized the growing contestation of European integration in national political debates and we expect losers to be less supportive of both the EU and the Euro. Eurosceptic attitudes are measured through a couple of survey items, whereby we asked whether membership in the European Union/the Euro is “a good thing”, “a bad thing” or “neither a good nor a bad thing” for Italy. Our analyses suggest that there is a significant correlation between both education and occupation and attitudes toward the EU and the Euro, together with macro-region of residence, and subjective economic status. More precisely, respondents with a higher educational attainment tend to be more supportive of the EU and the Euro than people with the lowest educational level. Contrarily, manual workers tend to express negative attitudes towards the EU and the single currency if compared to the category grouping large employers and managers. People who cannot make ends meet are also associated with the highest levels of Euroscepticism, compared to more affluent groups. Finally, people living in the northern regions express a better opinion of the EU and the Euro than those living in the centre and the south of the country. These results are reinforced by the fact that, considering a question asking respondents’ vote choice in a potential referendum on Italy’s membership of the Eurozone, losers are also more prone to vote for the exit from the single currency.



Figure 2. Predicted probabilities (outcome = “-1 – A bad thing”) (with 95% CI) from ordinal logistic regressions. DVs are respectively membership to the EU and to the Euro (-1 – A bad thing; 0 – Neither a good nor a bad thing; 1 – A good thing). *Note: omitted base categories are “18-44” for Age, “Men” for Gender, “Primary education” for Education, North for Macro-region, “Live with ease” for Economic Status, and “Employers/Executives” for Employment status.

Immigration and Welfare Chauvinism

Secondly, we have investigated whether and how winners and losers have different policy preferences in relation to foreigner’s access to the Italian job market and social security benefits. Here we expect losers, feeling more economically and culturally threatened by immigration, to show more chauvinist attitudes than winners. As dependent variables in our regression models, we used two questions asking respondents whether all foreigners, only citizens of other EU member states or no foreigners, even if legally resident in Italy, should have the same rights of Italians to access the labour market and the social security system. Empirical results show that citizens with a university degree are less supportive of exclusionary measures on the side of foreign citizens than their low-educated counterparts. If tested alone (plus controls), the occupational class plays a significant role showing that manual workers tend to be more chauvinist than employers and managers. However, as shown in the graph, once controlled for education the impact of occupation on chauvinism turns out to be non-significant. This result seems to indicate that cultural aspects and values play a more substantive role in explaining chauvinist attitudes of losers than their economic vulnerability. Interestingly, results show a non-linear relationship between age and chauvinism. Respondents aged between 45 and 64 years, which constitute the lion’s share of the active and working population, express more negative attitudes towards the integration of foreigners than the younger generation.


Figure 3. Predictive probabilities (outcome = “”) (with 95% CI) from ordinal logistic regressions. DVs are respectively job-market and welfare chauvinism. *Note: omitted base categories are “18-44” for Age, “Men” for Gender, “Primary education” for Education, North for Macro-region, “Live with ease” for Economic Status, and “Employers/Executives” for Employment status.

Propensity to vote

Finally, we moved beyond the association between socioeconomic status and Eurosceptic attitudes by assessing if there are differences in the respondent’s propensity to vote for the three major parties that took part in the last Italian elections. If the analytical framework we adopted holds true, then we should expect globalisation’s losers to be mobilised by Eurosceptic and populist parties like M5S and League and winners to vote for mainstream parties, such as Democratic Party (PD).

In line with this argument, regression coefficients indicate that the propensity to vote for PD is positively associated with tertiary education and negatively associated with the category of manual workers. These results couple with a recent study of De Sio (2018), arguing that PD is becoming the party of the elite losing support of the working class. On the contrary, highly educated winners are significantly less prone to vote for M5S and League than low-educated losers are. However, there are no significant differences according to the occupational class of respondents; only students show a lower propensity to vote for either M5S or League than employers and managers. Important geographic differences also emerge. Holding other factors constant, citizens from the southern Italian regions are significantly more prone to vote for M5S and less prone to vote for League, given the historical northern connotation of the party founded by Umberto Bossi.


Figure 4. OLS regression coefficients (95% C.I.) of propensity to vote for PD/M5S/League. DVs are 0 (least likely) – 10 (most likely) scales. *Note: omitted base categories are “18-44” for Age, “Men” for Gender, “Primary education” for Education, North for Macro-region, “Live with ease” for Economic Status, and “Employers/Executives” for Employment status.

In conclusion, what globalisation’s losers theory hypothesized has found some empirical support during the last Italian national election. Negative externalities of globalisation and the EU integration process are unevenly distributed in the society. Losers – low-educated and low-skilled citizens – tend to be more hostile to the EU and the Euro and to express chauvinist attitudes, compared to winners. Moreover, they tend to be more prone to vote for Eurosceptic parties such as M5S and League.

However, several recent studies stressed that several other factors, such as citizen’s orientations, contingent situations and relations they experience, might play a relevant role in mediating the impact that the socio-economic status has on Eurosceptic attitudes and behaviours. While socio-economic status is resilient and hard to modify, experiences and their perceptions are more amenable to the influence of targeted interventions.

This article is the part of the series “Understanding the Italian vote” which investigates the results of the past Italian elections.

Photo Credits CC: Alex Blăjan on Unsplash

Download PDF

Also published on Medium.


Italian General elections 2018

Twitter live stream “#Italianelections2018”:

Leave a comment
  • Facebook