ABSTRACT
The article analyses the 15 elections that took place between 2010 and 2019 in four South European countries – Greece, Italy, Portugal and Spain – through the lenses of the retrospective vote theory. The large within-country variation of economic conditions justifies the adoption of an original subnational perspective, while the explicit test of alternative economic quantities and horizons provides a more credible assessment of voters’ behaviours. Besides offering a taxonomy of local retrospective voting, the research found that citizens assessed the incumbents against regional unemployment levels and national growth dynamics, further benchmarking the local economic conditions against their past performances. These results give credit to the idea that the South European electorate shares similar references in assessing the economic competences of incumbent governments.
Acknowledgments
I would like to thank the editors and the anonymous referees of the journal for their comments and suggestions, which offered me the opportunity to improve the original version.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Supplementary material
Supplementary data for this article can be accessed at https://doi.org/10.1080/13608746.2023.2202468.
Notes
1. NUTS3-level areas correspond approximately to provinces in Spain and Italy, to groups of regional units in Greece, and to inter-municipal entities in Portugal (Eurostat: https://ec.europa.eu/eurostat/web/nuts/national-structures).
2. See Chiaramonte (Citation2014), Conti & Memoli (Citation2015), Orriols & Cordero (Citation2016), Teperoglou & Tsatsanis (Citation2014), and Tsakatika (Citation2016).
3. Using objective measurements is also the only feasible research strategy with subnational units of analysis, since reliable and representative survey data are not available at such level of granularity.
4. I did not consider as incumbents parties that ruled for less than one year, thus resolving the problem of repeated elections in Greece and Spain; in that event, the lagged value was referred to the last standard election. I also avoided comparing support for government parties that had disappeared or had been formed during the mandate. More details are provided in the supplementary material.
5. To better reflect that standard 1-year time horizon of the electorate, and in the absence of more precise quarter data, it is common to use quarterly weighted averages of annual indices. The analysis reported in this article followed that same good practice.
6. The fact that random effects at the national level almost disappear is due to the fact that cross-country variability is already mostly captured at the level of single elections. Completely cancelling that hierarchical level from the multilevel model does not substantially modify the results, but I decided to keep it following the methodological advice of Schmidt-Catran and Fairbrother (2016). The complete table is presented in the supplementary material.