The mitigating effects of political capacity and political alignment on COVID-19 mortality, an examination of Spanish Autonomous Communities

ABSTRACT

We analyze the first wave of COVID-19 fatalities in Spain at the subnational level to identify why some Autonomous Communities (ACs) were more successful at minimizing mortality levels. Our work adds to emerging literature on policy implications of pandemics and the role of governments in containing crises. Additionally, because national responses to crises are not homogeneous, we develop empirical measures of subnational political capacity. We investigate the relative capacity of governments in mobilizing the population to follow national health policy recommendations and extracting resources to implement those policies. Furthermore, we examine the effects of political alignment between the subnational and central governments in determining the country’s success containing the outbreak. Through a random-effects panel regression and a generalized additive model, we find that wealth and demography account for half of the variance in COVID-19 deaths across ACs, while including political capacity and alignment increases the variance explained above 70%.

RÉSUMÉ

Nous analysons la première vague de décès liés au COVID-19 en Espagne au niveau infranational afin d’identifier les raisons pour lesquelles certaines communautés autonomes (CA) ont mieux réussi que d’autres à minimiser les taux de mortalité. Notre recherche s’ajoute à la littérature émergente sur les implications politiques des pandémies et le rôle des gouvernements pour contenir les crises. En outre, les réponses nationales aux crises n’étant pas homogènes, nous élaborons des mesures empiriques de la capacité politique infranationale. Nous étudions la capacité relative des gouvernements à mobiliser la population pour suivre les recommandations de la politique nationale de santé, et à extraire des ressources pour la mise en œuvre de ces politiques. En outre, nous examinons les effets de l’alignement politique entre les gouvernements central et infranationaux dans la détermination du succès du pays à contenir la flambée épidémique. Par le biais d’une régression de panel à effets aléatoires et d’un modèle additif généralisé, nous constatons que la richesse et la démographie expliquent la moitié de la variance des décès liés au COVID-19 entre les CA, tandis que l’inclusion de la capacité et de l’alignement politiques augmente la variance expliquée au-delà de 70 %.

KEYWORDS:

• COVID-19
• political capacity
• subnational politics
• Spanish politics

Acknowledgements

We thank Nicholas Stowell from Claremont Graduate University for providing the RPE subnational data and Subnational level COVID-19 deaths, and Dr. Jacek Kugler for his guidance and feedback.

Disclosure statement

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

Notes

1 Data from Our World in Data. (Roser et al., 2020)

2 Data from Spain’s Ministry of Health (Centro de Coordinacion y Alertas de Emergencia Sanitaria, 2020a).

3 Mobility trends are compared to “the median value, for the corresponding day of the week, during the 5-week period Jan 3–Feb 6, 2020” as a baseline (Google LLC, 2020). Baseline days represent a normal value for that day of the week. As such, daily percentage changes from the corresponding day of the week baseline are provided.

4 The protected population equivalent is calculated by weighing the population covered by the National Health System by the relative per capita health expenditure for seven age brackets. See Grupo de Trabajo Sobre Gasto Sanitario (2007) for a detailed explanation.

5 Data from Spain’s Ministry of Health (Centro de Coordinacion y Alertas de Emergencia Sanitaria, 2020a).

6 The Spanish Congress is the most powerful of the chambers since, unlike the Senate, it can propose new laws, write, and debate legislature. Furthermore, in the most recent election PSOE’s majority in the Senate is clear—it holds 45% of the Senate’s seats.

7 We run random-effects linear models with first-order serial correlation (using the command XTREGAR in Stata) which transforms the data to remove the AR(1) component (see Stata manual for more information).

8 The GAM specification includes fixed effects controls for Autonomous Community.

9 Population density was not included given collinearity with GDP pc. When GDP pc is replaced with population density in the regression, its coefficient is insignificant. Results with population density are included in Table B in the Appendix.

10 The normalized mobility coefficients with lags 3, 4, and 5 are -1.03, -0.64, and -0.40 respectively (all statistically significant).

Additional information

Notes on contributors

Yuzhu Zeng

Yuzhu Zeng is a PhD Candidate in Political Science at Claremont Graduate University and TRC research fellow, focusing on International Relations and Comparative Politics. Her research interests include governments' political capacity, relations between the US and China, the impact of inequality in the developing world, and demographics changes within a subnational framework.

Marina Arbetman-Rabinowitz

Marina Arbetman-Rabinowitz PhD is a Research Associate at the Maldonado Institute at Claremont Graduate University and at La Sierra University. As a consultant, she worked for the US Government, World Bank, private banks, and financial institutions worldwide. Her research focus is on political and economic development, natural disasters, and conflict. Her expertise is in empirical modelling, data analysis, and measurement. She has done research on political capacity computation and evaluation at the international and subnational level for the last 35 years.