One or two arenas? The break-up between national and regional elections

ABSTRACT

This paper shows that the role of national electoral dynamics on regional elections is highly mediated by institutional and electoral constraints at the regional level. Using data on statewide parties’ electoral competition in regional and national elections in Spain and Italy, results show that the contamination of regional elections is lower in regions where decentralization has travelled further and where strong regionalist parties dominate electoral competition. The paper also shows that these two channels -more regional authority and more regionalist competition- shape the regional manifestos of statewide parties by increasing their pro-regional positions. These findings represent a contribution to a better understanding of the extent to which regional elections are a separate electoral arena from the national one.

Disclosure statement

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

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1 The most common measure in the literature has been a “dissimilarity index”, (Schakel 2013) which measures the proportion of the electorate who would have to change their vote in order for the regional election to exhibit the same results as in the previous federal election in that region (see Amat, Jurado, and León 2009 on the limitations of this index).

2 Other literature refers to this as electoral externalities (Rodden and Wibbels 2011).

3 Thorlakson (2009) shows that in layer cake models like Canada or US where subnational governments exercise have large powers over taxes and expenditures that are separate from the federal government, subnational branches become more autonomous. Conversely, in marble cake federal models like Germany or Austria, parties are more integrated and intergovernmental cooperation becomes smoother.

4 The formal distribution of powers that is enshrined in party statutes does not quite capture the actual balance of power between national and subnational party elites: changes in formal internal organization may only happen well after a modification of regional-national power relations has been in place.

5 We are focusing here on electoral results. For an analysis of the differential impact of national elections on individual-level behavior in regional elections across regions, see Liñeira (2016).

6 The models do not include a lag dependent variable (LDV) in order to avoid a potential Nickell bias in our estimates. The results are robust, however, to the inclusion of a lagged dependent variable.

7 In the Online Appendix we show results using the more general Regional Authority Index, which remain the same. We also display in the Appendix analyses using contemporaneous values that yield insignificant estimators, confirming that changes in the two hypothesized channels precede changes in the level of contamination.

8 The analyses show the results for PP and PSOE. In Navarre, PP contests in some elections as part of a coalition with UPN, which at the same time can be labelled as a regionalist party. We do not include the results for this party in this region.

9 Notwithstanding our argument, Müller and Bernauer (2018) analyze parties’s regional deviations from national-level position in Switzerland and account for other ideological, temporal, vote- and canto-specific factors that can explain variation in the positions.

10 We include as a covariates the vote share of the party at the regional level in national elections, the subnational left-right position of the regional manifesto -as Massetti and Schakel (2015) have shown that left wing positions in Spain they can correlate with party’s regionalism-, and the saliency of the center-periphery dimension in the regional manifesto, as well as the regional dummies.

11 In alternative analyses not reported here, we have replicated the model specification including an interaction between both dependent variables. However, we do not find interactive effects. Both variables seem to have independent and unconditional effects.

12 In these two regions, regionalism has been also above 20%. In Cantabria regionalism has been high since the mid-nineties mostly due to PRC (Partido Regionalista de Cantabria)

Additional information

Funding

This work was supported by Ministerio de Economía y Competitividad: [grant number CSO2017-82881-R]; Juan March Institute, Center for Advanced Studies in the Social Sciences (Funding Programme 2).