Do real-world developments predict far-right news on immigration?

In recent years, the political right became almost perfectly aligned with a more exclusionary stance on immigration. I analyze the emergence of exclusionary discourses based on immigration news published in one of Germany’s most popular right-wing media outlets over a period of more than two decades (Jan 1998 to Dec 2019). In particular, I test the influence of three real-world developments that are closely connected to the idea of group threat: immigration numbers, terror attacks, and crime by foreign suspects. I analyze data of Germany’s largest right-wing print outlet and the fifth largest weekly newspaper overall (Junge Freiheit, JF). Ideologically, the JF is located between the conservative and the far-right, often bridging democratic conservatism and the radical right. I trained a Naive Bayes classifier to identify articles about immigration based on hand-coded training data. Then, I used structural topic modeling to estimate the term probability for each topic and the topic probability for each article. I added the number of immigrants to Germany, the number of Islamist attacks, and reported crimes conducted by non-German suspects. I coded the topics based on their most probable terms as well as articles (a more detailed discussion of this coding will be available in the full article). I define six topics as relating to immigration (German immigration policy, EU immigration policy, Integration/citizenship, Asylum, Illegal immigration, Demography), one topic is addressing Terrorism and two topics relate to security and crime (Crime and Ethnic violence). In addition to these topics, there are four topics that do not directly relate to the real-world developments but are theoretically interesting because they correspond to cultural (Religion, National identity) or economic (Economy, Welfare state) threat. Firstly, the two topics that most strongly correlate with real-world immigration are Asylum and German immigration policy. For both topics, comparing times of lowest immigration with those of highest immigration implies an increase in topic probability by about six percentage points. With predicted probabilities of over eight percent, both topics dominate the JF output when immigration is high. Themes on cultural, economic, and security threat are work largely independent of world immigration developments. Secondly, attacks in Western countries lead to an increase of the Terrorism topic by about 1.2 percentage points. While this increase is statistically significant, it’s not very large. In contrast, the Terrorism topic increased strongly by about 6.7 percentage points after attacks in Germany. Interestingly, although all the investigated attacks were motivated by religion, the attacks did not change the presence of the Religion topic at all according to the model. Thirdly, reporting on crime is largely unconnected to real-world crime developments. While there is a positive effect on the Crime topic, it is limited in size. Comparing the lowest with the highest crime rates implies an increase of the predicted probability of the Crime topic by about one percentage point (from 3.9 to 4.8 percent). Ethnic violence, another topic that could potentially be connected to crime, is even negatively associated with actual crime.