Decentralisation and regional cabinet size: the Spanish case (1979–2015)

Abstract

This article explores under what conditions regional governments tend to have larger or smaller cabinets. The main contention is that cross-regional variation in cabinet size is partly explained by the dynamics set up by the multilevel system of government, mainly territorial decentralisation, multilevel government (in)congruence or the existence of nationally distinct regions. The hypotheses are tested with a new and original dataset built upon the Spanish case (1979–2015). Findings show that regions with more welfare state policies, especially when the region’s economic capacity is high, and nationally distinct regions tend to have bigger executives. In contrast, decentralisation in the form of basic state functions and government incongruence do not have a significant effect. Results have important implications for our understanding of sub-national territorial institutions and their interaction with decentralisation dynamics.

Keywords:

• Decentralisation
• cabinet size
• multilevel system
• ministers
• portfolios

Acknowledgements

We thank Laura Chaqués and Albert Falcó-Gimeno for sharing data with us. We also thank the two anonymous referees for their insightful comments. Both authors contributed equally to this work.

Notes

1. The dataset and the replication materials are available at https://sites.google.com/site/pauvallprat/data.

2. Deviations from full proportionality have been the focus of recent research, see Bäck et al. (2011) or Falcó-Gimeno and Indriðason (2013).

3. This was the case of Navarre (1979), the Basque Country and Catalonia (1980), Galicia (1981) and Andalusia (1982). The case of Andalusia is special, since the region was finally included in the fast-track path after a period of elite and popular mobilisation in favour of self-government.

4. As compared to other decentralised states, Spanish ACs are characterised by a moderate level of regional self-government (behind federal states) and a very low level of shared government. See RAI dataset (Marks et al. 2008).

5. El País’ yearbooks collect the composition of each regional cabinet on 1 January and, therefore, regional cabinets that lasted less than a natural year could not be covered. However, more than one reshuffle per year is not usual. We follow Falcó-Gimeno (2014) and data was double-checked using the database of the Observatorio de Gobiernos de Coalición (http://www.ub.edu/OGC/index_es.htm).

6. Ministers without a portfolio are highly uncommon and are only identified in the early years of regional governments.

7. Although the ministers at the Spanish regional level are named as consejeros (councillors), we stick to ‘ministers’.

8. It is important to focus on reshuffles as these new cabinets can imply changes in cabinet size and by not including them we would be losing relevant units of data. We follow a similar approach to Verzichelli (2008), who showed that reshuffles did not affect cabinet size in a different way than post-electoral cabinets. In our dataset, half of the cases (43.4%) correspond to reshuffled cabinets. Reshuffles are fairly distributed across regions (35–50% of each region’s observations correspond to reshuffles), which does not lead to some regions having more influence than others because of the number of observations by region (robustness checks weighting for the number of cases across regions provide the same results).

9. Regional budgets were gathered from a variety of sources, including original budget laws, official regional gazettes or regional newspapers. We converted budget figures to euros and adjusted them to inflation. In Spain, economic decentralisation is very low although there is an exception for the Basque Country and Navarra, which are allowed to collect their own taxes and give a portion to the central government in order to pay for functions such as Defence or Foreign Affairs. In the rest of the ACs, all taxes are levied and collected by or for the central government.

10. Data comes from the Spanish Ministry of Territorial Policy and is publicly available at the Ministry for Finances and Public Administration (http://www.seap.minhap.gob.es, accessed 28 October 2014). Chaqués and Palau (2011) also include another category of transfers, relative to environmental protection and territorial management planning. Due to the nature of the transfers, we have considered them as welfare state policies. Disentangling both, however, provides the same results. Finally, Falcó-Gimeno (2014) did not include all regions in his analysis. We complemented the dataset following the same coding approach. Table B in the online appendix shows the concrete coding for each topic.

11. Alternatively, we also computed an additional measure of congruence, which tackles the ideological (dis)similarity across levels of government. Further details and results, which are robust, can be found in the online appendix.

12. Sources: Political Opinion Barometer 746, April 2014. Catalan Statistics Institute (CEO, http://ceo.gencat.cat/). Post-election Survey of the Basque Elections (Ref. 2964), October 2012. Centre for Sociological Research (CIS, http://www.cis.es). Post-election Survey of the Galician Elections (Ref. 2958), September 2012. Post-election Survey of the Navarre Elections (Ref. 3097), June 2015. Centre for Sociological Research (CIS, http://www.cis.es).

13. Operationalisation detailed in the online appendix.

14. One might think that decentralisation and congruence interact: with more powers, regions are more likely to act as they deem necessary and the central capacity to influence regional governments’ decisions is lower. This interaction, however, is not statistically significant.

15. Operationalisation detailed in the online appendix.