The Party as Pariah: The Exclusion of Anti-Immigration Parties and its Effect on their Ideological Positions

Abstract

During the past three decades, anti-immigration parties have emerged all over Western Europe. Some of them have been treated like any other party by their mainstream opponents and a few have even become members of governing coalitions (e.g., the Austrian FPÖ). Other such parties have been politically excluded: established parties have refrained from any cooperation with them and in some cases even refused to enter into a political debate with their politicians. This article investigates how the strategy of ostracising anti-immigration parties affects the internal dynamics within these parties. In particular, we assess whether these parties radicalise as a result of this strategy, and, conversely whether it has a moderating effect when these parties are approached more pragmatically. Our analyses, regarding ten parties at several moments in time, show that anti-immigration parties that were not ostracised became more moderate, whereas those that were treated as outcasts continued to be extremist.

Acknowledgements

An earlier version of this article was presented at the 2004 ECPR Joint Sessions of Workshops in Uppsala. We would like to thank the participants of Workshop 17, ‘Effects of Incumbency on Organization of Radical Rightwing Parties’, for their comments, and Bob Harmel (Texas A&M University) for organising the workshop. Special thanks to Karina Pedersen (University of Copenhagen) for providing detailed information on the Danish case. Wouter van der Brug thanks the NIAS for its hospitality and support, while finishing this article. We are also grateful for comments from two anonymous referees and the editor of this journal. We are responsible for any remaining mistakes.

Notes

1. In Belgium and France this strategy is referred to as the strategy of the cordon sanitaire, in German-speaking countries it is called Ausgrenzung.

2. Strategies of exclusion have not only been applied to anti-immigration parties, but to various kinds of parties in many established party systems. For example, communist parties were confronted with similar strategies during the cold war.

3. Rabinowitz and Macdonald (1989) coined the term ‘region of acceptability’ in the context of electoral research, where it refers to policies that are no longer acceptable to most voters. We use the term here in a different context.

4. Many anarchists refuse to vote for precisely this reason: those who are elected will become ‘encapsulated’ by the system.

5. Of course, asking to indicate the positions of political parties of ten years earlier, as Lubbers (2001) did, is likely to result in major misspecifications. This is not relevant for this paper, however. None of the conclusions of this paper would be different if the 1990 measurements collected by Lubbers were not taken into account. The point here is that there is a wide gap between anti-immigration and other parties, which justifies our classification of parties on the basis of party positioning on the immigration issue.

6. Note that the party ostracism dimension can apply not only to anti-immigration parties, but to any political party operating in any party system at a specific point in time.

7. Another possible critique of the approach is that it is not an interactive model, i.e., it does not account for interplay of actions of the one actor, and reactions of the other. The underlying assumption is that parties react to the existence of a specific (other) party, and following reactions of the ostracised party are of no or less importance. This is in accordance with the situation in Western Europe, however, where responses to anti-immigration parties seem to stem from political culture and tradition in the various countries. In any case, they can be safely assumed to be independent of the attitudes of the anti-immigration parties towards the other parties.

8. We restricted ourselves to assessing ‘party ostracism’ only for those party/year combinations for which we have also data on (one of) the dependent variables.

9. The observations are limited to the 28 party/year combinations for which we also have values of the dependent variable at our disposal (see below).

10. The category ‘moderate right’ in Castles and Mair (1984) ranged between 6.25 and 8.75 on a 0–10 left–right scale.

11. We tested our results by using data from a different kind of source: estimations by experts. We also performed the analyses with left–right party placements by experts, using the combined results derived from two different surveys, conducted at different points in time: Huber and Inglehart (1995) with data pertaining to 1994 and Lubbers (2001) on 1999. The conclusions of these analyses turned out to be similar to the findings on the basis of the voters' perceptions. The variance in left–right party placement was explained for 50% (adjusted R-square = 0.50) by party ostracism. Regarding 15 observations, the difference between the two categories was more than a full point (1.06). Our findings are also in accordance with left–right party placement by experts reported by Marks and Steenbergen (1999) and Benoit and Laver (2006).

12. With the possible exception of the Italian AN by 1999.