Political competition and representation in municipal councils

ABSTRACT

I study whether the numbers of high- and low-social status-individuals in French municipal councils are affected by the level of political competition. I argue that because the most-eligible positions within lists of candidates are obtained by high-social-status persons, political competition increases the number of high-status municipal councilors and is thus detrimental to the election of low-social-status candidates. Using the elections of 2008 and 2014, I provide suggestive evidence that the socioeconomic status of individuals is indeed correlated with their position within lists of candidates. Moreover, I show that the presence of an additional competing list during an election is associated with 0.2 additional high-social-status individuals within the municipal councils. I also observe that those additional municipal councilors are males, which is a result that might have been caused by the gender quota policy implemented in French local elections.

KEYWORDS:

• Municipal council
• socioeconomic background
• representation
• French municipal election
• political competition

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/03003930.2024.2350627

Notes

1. Carnes (2016) strikingly recalls that in the US, if millionaires were a political party, ‘it would have a supermajority in the Senate, a majority in the House, a five to four majority in the Supreme Court and a man in the White House’. For a historical discussion of political recruitment in legislative chambers in Europe, see also Best and Cotta (2000) and Best (2007).

2. Sweden might be one exception. Dal Bо́ et al. (2017) suggests that social classes have little role in Sweden’s local politics, where politicians seem to be selected based on competences.

3. Pitkin (1967) provides an extensive discussion of the concept of representation. She defines substantive representation as acting ‘in the interests of the represented, and in a manner responsive to them’ (Pitkin 1967, 209). She also defines descriptive representation as ‘the making present of something absent by resemblance or reflection, as in a mirror or in art’ (p.11). This is often understood as having one or several representatives who share the main characteristics of the represented. The concept of descriptive representation may actually be criticised; Pitkin and Hayat (2013) argues that thinking of the representation of an assembly as ‘a representative sample’ leads one to focus on its ‘composition rather than on its activity’.

4. Such gender quota policies have generated important discussions in recent years. (a) Are the policies effective? The recent literature on France or Europe during local elections includes Lassébie (2020) Bagues and Campa (2021), Rodrigues (2022). (b) Do these quotas affect the policies that are enacted? As far as I know, studies focusing on local governments have found mixed evidence (Andreoli, Manzoni, and Margotti 2021; Beach and Jones 2017; Casarico, Lattanzio, and Profeta 2022; Chattopadhyay and Duflo 2004; Ferreira and Gyourko 2014; Pande 2003).

5. Smaller municipalities follow an open-list system. The threshold above which the closed-list system is implemented was 3,500 inhabitants before 2013 and 1,000 inhabitants afterwards.

6. Starting in 2018, it is possible to add two more names as substitutes. However, this was not the case during the elections that I focus on.

7. The three largest French cities, namely, Paris, Marseille and Lyon, have elections organised at the district (arrondissement) level. Thus, they are excluded from the current analysis.

8. Moreover, lists that obtained more than 5% of the vote may be merged with other lists.

9. As mentioned in the introduction, it is not clear whether voters react to the socioeconomic status of candidates (Adams et al. 2021; Albaugh 2020; Carnes and Lupu 2016). It is also doubtful that voters know the socioeconomic backgrounds of all the corunners on a list. This is why I insist on the fact that what matters is what heads of list believe.

10. Although this will not be investigated below, it appears that previous considerations have depended on the fact that seats in municipal councils are allocated (at least partially) proportionally to the votes obtained by the list. In a ‘winner-takes-all’ electoral system, all the individuals from the winning list would be elected as municipal councilors, and political competition would therefore not affect the allocation of seats.

11. In addition to genuine mistakes, there might be errors due to accents in the names. Those accents may have been written in one file and forgotten in the other. I therefore removed accents in the names.

12. Descriptive statistics are available in the appendix.

13. The relative rank is $\frac{\mathit{Rank}}{\mathit{No}.\mathit{Seats}}\times 100$. In the above regression, the size of the lists is controlled for by means of fixed effects.

14. In a previous version of the text, I also analyse whether occupations within the high- and low-socioeconomic-status groups are positioned similarly on lists of candidates. It appears that there is a hierarchy of occupations within the two groups. I also investigate whether political labels of lists (left or right) or the level of political competition within cities affected those rankings. Those analyses are available upon request.

15. This would be coherent with articles that discuss the creation of lists, such as Achin and Paoletti (2002) or Lévêque and Cavallaro (2020).

16. Regarding precision, this outcome is likely due to the sample size and the difference in identification strategy. Localities with fewer than 3,500 inhabitants did not follow the list-based system prior to 2014. Consequently, they could not be used in the panel regression. The identification strategy also focuses on cities that experienced changes in the number of seats or in the number of competing lists. Many cities had the same numbers of competing lists between the two elections.

17. Both the average and median value of low-status persons is 5, while the median (average) difference between low- and high-status persons is −4 or −4.6.

18. In particular, I control for a proxy of the “median income’’ in cities. I am not able to find the same measure in both 2008 and 2014.

19. Adding seats is often discussed in the literature (Gerring, Jerzak, and Erzen 2023; Kjaer and Elklit 2014; Kjaer, Dittmar, and Carroll 2018). Increasing the number of seats in municipal councils and therefore, the number of corunners on lists could also affect political competition. This is the case if doing so impedes the creation of lists of candidates that could not gather enough corunners.

20. Once more, see Lassébie (2020) for a discussion.

21. I use a note from the INSEE (the National Institute of Statistics and Economic Studies; Roth et al. 2012) to attribute the skill level.