Christian Religiosity and Voting for West European Radical Right Parties

Abstract

This article examines the relationship between Christian religiosity and the support for radical right parties in Western Europe. Drawing on theories of electoral choice and on socio-psychological literature largely ignored by scholars of electoral behaviour, it suggests and tests a number of competing hypotheses. The findings demonstrate that while religiosity has few direct effects, and while religious people are neither more nor less hostile towards ethnic minorities and thereby neither more nor less prone to vote for a radical right party, they are not ‘available’ to these parties because they are still firmly attached to Christian Democratic or conservative parties. However, given increasing de-alignment, this ‘vaccine effect’ is likely to become weaker with time.

Acknowledgements

We would like to thank John Bartle, Thomas Poguntke, Elinor Scarbrough and Jack Veugelers for their valuable comments and suggestions on an earlier version of this article. We are also grateful to two anonymous reviewers and the editor of this journal for their helpful comments. Of course, the usual disclaimer applies.

Notes

1. A number of authors have reported that the confessional influence on electoral behaviour varied over time and according to place. For example, Childers (1983: 188) argues that in the mid-1920s the NSDAP actually gained support from Catholics and Protestants in equal proportions, and that it was only from 1930 that it made the greatest gains among Protestants. Falter (1991: 61), on the other hand, notes that there were numerous villages in Catholic regions of Germany where the NSDAP recorded high electoral results. That said, he concurs with the finding that Catholics were much less likely to vote for the Nazis than other social groups because of their firm attachment to the Zentrum party and because the Church frequently spoke out against the Nazis in the period before 1933 and forbade church officials from publically supporting or joining the NSDAP. On the role of intermediate groups in mass and totalitarian society see among others Arendt (1951); Kornhauser (1959); but see also Berman (1997) for a more recent critique.

2. For example, the Austrian FPÖ devoted only a very short chapter of its main platform to the subject of ‘Christianity as the root of European culture’. Moreover, a large section of this chapter was concerned with the need to retain the separation of church and state: see http://www.fpoe.at/index.php?id=459 (accessed 1 May 2007). Similarly, the Danish People's Party devoted only three sentences of its basic manifesto to Christianity and the role of the state church of Denmark: see http://www.danskfolkeparti.dk/sw/frontend/show.asp?parent=19185&menu_parent=22669&layout=0 (accessed 1 May 2007), and the Norwegian Progress Party made just one reference to the ‘Christian Ethos’ as something positive in its official statement of its ‘Principles’: see http://english.frp.no/Innhold/FrP/Temasider/Flere_sprak/ (accessed 1 May 2007).

3. See, for instance, the ‘Vienna Declaration’ that pits ‘inalienable Christian values’ (without saying what these values are and what they imply) against ‘aggressive Islamism’: see www.ots.at/presseaussendung.php?schluessel=OTS_20051114_OTS0051&ch=politik (accessed 21 November 2005).

4. Many studies have also found that more women than men tend to be religious (Miller and Stark 2002; Walter and Davie 1998). That said, women have been shown to be less likely to vote for a radical right party than men (Betz 1994; Lubbers et al. 2002). One of the key benefits of our analysis is that we can disentangle these effects and are thereby able to resolve such apparent contradictions.

5. While in an ideal world we would have included more political attitudes in our model such as people's evaluation of candidates, their stances on particular policies, and their ideological distance from the parties, such indicators were not available in the dataset we used. However, this is not a significant problem as stances on policies and ideological location are highly correlated with radical right attitudes anyway.

6. Cronbach's alpha exceeds 0.7 (the conventional threshold for a ‘good’ reliability) in all countries but Denmark, which is a borderline case (0.68).

7. The correlation between church attendance and our latent variable (between 0.72 in Denmark and 0.86 in Austria) shows that the common practice of using church attendance as an indicator of religiosity is justified to a certain degree but that it is preferable to employ multi-indicator variables if possible. Since we employ an extension of the Structural Equation Modelling (SEM) framework (see Data and Methodology section), our latent variables are generated from their indicators by a variant of factor analysis. More specifically, SEM iteratively constructs a combination of latent variables that are most likely to have given rise to the empirical covariances between the observed variables (given the restrictions implied by the structure of the model). In a single equation setup, one could closely replicate the construction of the religiosity variable by first extracting a single factor or principal component from the indicators and then entering this into the regression as an independent variable. The MPlus scripts can be found in the replication archive for this article at http://hdl.handle.net/1902.1/12312.

8. Cronbach's alpha varies between 0.81 (Italy) and 0.91 (France), indicating that our measure is very reliable.

9. 46.2 per cent of Dutch respondents were Catholic, while 37.5 per cent were Protestant. In Switzerland 49.5 per cent of respondents were Catholic and 41.5 per cent were Protestant. Elsewhere there was much less variation: Catholics made up 87.4 per cent of the respondents in Austria, 90.0 per cent in Belgium, 86.2 per cent in France and 98.4 per cent in Italy. In Denmark 93.5 per cent of respondents were Protestant, as were 89.4 per cent in Norway.

10. The SVP is perhaps a borderline case for inclusion in the radical right party family. That said, Christoph Blocher's domination of the party since the late 1990s has led a number of observers to argue that the SVP has now evolved into a radical right party (McGann and Kitschelt 2005).

11. The exact numbers were 2,033 for Austria; 1,634 for Belgium; 1,388 for Denmark; 1,374 for France; 1,156 for Italy; 2,197 for the Netherlands; 1,889 for Norway; and 1,689 for Switzerland.

12. In some countries the main right-wing competitor is a liberal party rather than a Christian Democratic or conservative one – e.g. the VVD in the Netherlands. We have not included such parties in our model because we are concerned with examining the extent to which an identification with a Christian Democratic or conservative party encapsulates religious voters, and in so doing, decreases their likelihood of voting for a party of the radical right.

13. For each country, we created five imputations that were analysed in turn using Royston's (2005) implementation of MICE in Stata. Since MICE is a stochastic procedure (see van Buuren and Oudshoorn 1999), these datasets differ slightly from each other, reflecting the amount of uncertainty about the imputed values. Results from the separate analyses of these datasets were combined according to the rules outlined by Rubin (1987), resulting in approximately unbiased parameter estimates and conservative standard errors that take the amount of missing data into account. The tables in the next section refer to these combined results only.

14. The Root Mean Square Error of Approximation for each country is as follows: Austria: 0.069; Belgium: 0.063; Denmark: 0.051; France: 0.055; Italy: 0.066; Netherlands: 0.059; Norway: 0.061; and Switzerland: 0.056.

15. This might be either a life-cycle effect or a generational pattern.

16. Specifying a curvilinear relationship (in line with some of the early findings on the effect of religion) does not substantively alter this result.

17. The few exceptions to this overwhelming trend are i) Norway, where age does have a significant effect on the probability of a radical right vote, with voters aged 30 or over showing less likelihood of voting for the radical right than their younger compatriots, and ii) France, Norway and Switzerland where, when other variables are held constant, male voters are roughly twice more likely to cast a ballot for a party of the radical right than female ones. In these countries the oft-noted gender gap in radical right voting therefore persists even when we control for other variables. Removing the path from religiosity to voting behaviour does not substantially alter these results (not shown as a table).

18. As in Switzerland, in the Netherlands the effect of religiosity on the radical right vote is virtually identical for Catholics and Protestants (not shown as a table).

19. Amongst these, Belgium is a borderline case (p = 0.055).