Policy Congruence and Style of Representation: Party Voters and Political Parties

Abstract

This article focuses on whether styles of representation influence policy congruence. Style of representation is defined at the party level as the proportion of representatives within parties who are partisans, delegates or trustees. Policy congruence refers to how close on the left–right scale the mean position of a party as placed by its candidates is compared to that of party voters. The article concludes that where there are higher proportions of trustees within parties, there is a greater degree of policy congruence, whereas a higher proportion of partisans results in less policy congruence. The proportion of delegates has no significant impact on congruence after taking account of other party and country measures. This indicates that party constraints on representatives are applied at the cost of congruence with voters, and that when representatives enjoy more flexibility to follow their own opinions, the party displays greater congruence with its own voters.

Acknowledgements

Drafts of this article have been presented at the 2nd CCS Plenary Conference Mannheim, 27–29 January 2012, ECPR joint sessions in Antwerp, 10–15 April, in the panel ‘Perceptions of Representation: A Cross Analysis of Citizens’ and MPs’ Views’, CDSS workshop for doctoral students at Mannheim University, spring 2012 and 22nd ECPR Standing Group Summer School ‘Political Parties in Modern Democracies’, Brussels, 10–22 September 2012.

Many people have offered substantial comments on the work in process. These include my supervisors, Hermann Schmitt and Olafur Th. Hardarson, as well as Indridi Indridason, Robert Rohrschneider, Daniele Caramani, Gunnar Helgi Kristinsson, Thorunn Palina Jonsdottir, Thorbergur Thorsson, Stephanie Cox, my co-students in the CDSS programme at Mannheim University and many more.

Notes

1. Wessels and Giebler label them as ‘party delegates’.

2. Wessels and Giebler group trustees and delegates together and contrast them with partisans.

3. CSES module 3 covers Australia 2007, Ireland 2007, Netherlands 2006, Germany 2009, Iceland 2009, Portugal 2009 and Switzerland 2007.

4. In the EES study, voters were asked about what party they voted for in the previous election. In both countries, Greece and Belgium, the previous election was the 2007 election also covered in the CCS data.

5. This applies to the candidate data from Portugal 2009, Germany 2009 and Ireland 2007.

6. This applies to the candidate data from Greece 2007, Iceland 2009 and the Netherlands 2006.

7. This applies to the candidate data from Australia 2007, Belgium 2007 and Switzerland 2007.

8. The formula for policy congruence using the means is: LRDP_m = |V_m – P_m|. LRDP_m is the absolute left–right distance between the mean placement of party by its candidates and the mean self-placement of party voters, V_m is the mean left–right placement of party voters and P_m is the mean left–right placement of the party as placed by its candidates. The scales for left–right are on an 11-point scale, ranging from 0 (left) to 10 (right).

9. The interpolated median takes account of the number of respondents that choose each layer on the congruence measure, in our case the left–right position. The interpolated median is computed as follows: IM = M + (ng – nl) / (2ne), IM is the interpolated median, M is the standard median, nl is the number of responses less than M, ng the number of responses that are greater than M and ne is the number of responses equal to M. The distance measure is calculated using the same formula as computing the distance between the means (see note 8); replacing the mean with the interpolated median.

10. Andeweg uses a non-cumulative function, while Golder and Stramski use a cumulative function. See discussion on this in Andeweg (2011: 43).

11. The one exception is that the coefficient for the proportion of partisans is non-significant when policy congruence is measured as the absolute difference between the interpolated medians; in all other models it is significant and positive.

12. There is a semantic difference between questions 1 and 2, when asking voters’ opinions. The response category to the first question refers to the MP’s own party voters, and the second to constituency voters. However both are contrasted with either the party position or the MP’s own opinion. I assume that the responses reflect a difference between a bottom-up process as the role of delegate assumes, and a top-down process inherent in the role of partisans and trustees. Based on that I take those together who choose voters in question 1 and 2 as delegates and contrast them with partisans and trustees.

13. Weighing the data by country or by party results in an increase in partisans, a slight increase in delegates and a drop in trustees. The difference between the un-weighted and weighted replies, and the harmonisation between the replies when weighted by country or by party, could indicate that there is a country difference in representatives’ replies to the questions about the role of the MPs.

14. Information on the size of the parties in parliament is from the macro-data codebooks in the CSES study for all elections except the Greek and Belgian elections in 2007. Information on the size of Belgian and Greek parties is retrieved from Patrikios and Karyotis (2008) for Greece and from Pilet and van Haute (2008) for Belgium.

15. Various sources for party age are used, such as party websites, Wikipedia.org and macro-data codebooks from the CCS project. National experts have, in all cases, been asked to confirm the coding for party age. National experts are: Rachel Gibson and Ian McAllister (Australia), Lieven de Winter and Pierre Baudewyns (Belgium), Hermann Schmitt (Germany), Theordore Chadjipadelis (Greece), Eva Heida Onnudottir and Olafur Th. Hardarson (Iceland), Michael Marsh (Ireland), Dorien van Rheenen (Netherlands), André Freire (Portugal) and Georg Lutz (Switzerland).

16. Another possibility would be to use the number of years since the party was founded. However, there can be problems in pin-pointing the exact year for when the parties were founded, as well as mergers and splits of parties before they first enter the parliament. Whilst these problems can also occur when using the year the party first enters the parliament, I believe this is a more reliable measure. Most importantly, it should be found that party age is related to policy congruence; older parties should have higher policy congruence with their voters compared to younger parties.

17. Information on the number of days parties have been represented in government is from the Parliament and Government Composition Database (2011).

18. The absolute difference between the mean party voter and the mean country voter.

19. Formula used is: Number of effective parliamentary parties = 1/∑((proportion of seats in the parliament)²).

20. Polarisation of the party system is calculated using the following formula from Dalton (2008): SQRT {∑(party vote share_i)^∗([party L–R score_i – party system average L–R score]/5²}, where i represents individual parties and L–R stands for left–right. For the parties’ L–R means, I use the candidates’ placement of their own party from the CCS data.