As Sweden Sees More Immigration, The Far Right Sees Opportunity
Given increasing immigration to Europe, an important question is whether objective immigration increases the popularity of radical right parties. This question is especially relevant in Sweden, the country that accepted the largest per capita share of refugees during the so-called European Migrant Crisis of 2015. Sweden’s 2018 national election is only a few months away, and polls put the Sweden Democrats (SD) at around 20%, a clear increase from its 12.9% finish in the 2014 election.
There are two competing theories of native-born reactions to immigration. According to the first, known as “group threat theory”, increases in the size of an out-group (in this case foreign-born) leads members of an in-group (in this case native-born) to feel threatened or fear competition for economic resources, cultural hegemony, and/or political power. Thus, based on this account, immigration should lead to an increase in voting for parties that actively seek to restrict immigration. The second theory, known as the “contact hypothesis”, posits that an increase in immigration also increases opportunities for positive interaction between members of different ethnic or racial groups. According to this account, intergroup contact makes viewing an out-group as threatening less likely and, therefore, should limit support for parties that rhetorically identify immigrants as a threat to a nation-state.
Which theory does the evidence support? In Western Europe, the answer is not straightforward. A recently published meta-analysis of 48 peer-reviewed articles published between 1990 and 2017 shows that measures of objective immigration do not have a consistent effect on voting for radical right parties in Western European countries. The authors, Abdelkarim Amengay and Daniel Stockemer, find that immigration is positively correlated with voting for the radical right in only 38% of the 369 statistical models featured in these studies. In 14% of the models, immigration has the opposite effect. However, in 47% of the models, there is no statistically significant relationship.
What about in Sweden? The meta-analysis also indicates that previous research on the 2006 and 2010 Swedish national elections are also somewhat inconclusive. Immigration is not always related to voting patterns. To help make sense of what this could mean in the upcoming 2018 election, in the figure 1 below, I plot the size of the immigrant population against rates of voting for SD in 2006, 2010, and 2014. To do this, I use data from Statistics Sweden to illustrate the relationship between the share of immigrants at the municipality-level and share of votes for SD in three different elections (rows) in three election years (columns). From top to bottom, the elections are national, county, and municipality. From left to right, the election years are 2006, 2010, and 2014. Here we can see that the relationship between the percent of the population that is foreign-born and vote share is strongest in 2010 in all three elections. However, as SD becomes more popular in 2014, this relationship all but disappears in the national and county elections. The presence of foreigners at the local level does not appear to drive support for SD.
But what about increases in the size of the immigrant population? Figure 2 illustrates the between-election change in the percentage of the municipality population that is foreign-born and the change in percentage of votes for SD between two elections. The left plot illustrates change between the 2006 and 2010 elections and the plot on the right represents change between 2010 and 2014. Here we can see a clearer relationship between immigration and voting. Municipalities where the size of the foreign-born population increased also saw increases in the share of votes for SD. This pattern is similar for the county and national elections, but strongest at the municipality-level.
Taken together, these plots indicate that increases in the size of the immigrant population are more strongly associated with support for SD than the overall size of the foreign-born population. Thus, it appears that increases in immigration increase levels of threat, at least initially. By 2014, however, overall levels of diversity are weakly related to voting, suggesting that greater numbers of immigrants also increase positive experiences with diversity (such as the incidence of intergroup friendships) and reduce threat over time.
The aggregated data presented here tell us nothing about the individual characteristics of SD voters, only whether patterns exist across Sweden’s 290 municipalities. Previous research demonstrates that individual-level characteristics play an important role in voting behavior. Socioeconomic factors, such as years or level of education, are negatively correlated with support for radical right parties, but so are people’s own experiences with immigrants, which tend to be captured better by survey research and analyses of even smaller geographic units.
So while the Sweden Democrats are poised to increase its share of the electorate this September, it is unlikely that overall levels of regional diversity will determine from what parts of the country these votes will come. However, it is likely that the regions that saw an increase in immigration since 2014 will also see a boost in support for SD this autumn.
This article is brought to you by the Centre for Analysis of the Radical Right (CARR). Through their research, CARR intends to lead discussions on the development of radical right extremism around the world.