FACTORS OF CENTRIFUGAL POLITICAL TENDENCIES IN SPAIN: A QUALITATIVE COMPARATIVE ANALYSIS

The purpose of this article is to analyze the centrifugal tendencies in 17 Autonomous Communities of Spain by demonstrating that certain differences in the configurations of factors have lead to varying degrees of the loss of autonomy. The main conclusion is that for the configurations leading to less autonomy, there is a low level of economic development, a low volume of EU funding, and an absence of local languages as a factor of identity. The cases of Andalusia, Asturias, and Galicia are of particular importance in these configurations. For the configurations that lead to greater autonomy, two patterns are identified. The first group (Catalonia and the Basque Country) demonstrates a link between centrifugal tendencies and three conditions: the existence of regionalist parties in the parliaments of Autonomous Communities (AC), a high level of economic development and a high level of identification with a regional language. The second group of cases (Valencia and the Canary Islands) demonstrates the added importance of funding under the EU regional policy programs for the period from 2014 to 2020. The scientific novelty of the article lies in its analysis of the dynamics of centrifugal tendencies in Spain via crisp-set Qualitative Comparative Analysis (csQCA).


Introduction
The factors that influence the dynamics of centrifugal political tendencies in Spain, has received a great deal of attention.For example, several authors claim that the existing territorial model of the "state of autonomies" was laid down in the Constitution of Spain in 1978, and the socio-economic discrepancies in the development of regions, namely, the gap between the "developed periphery" and developing regions (Shashkova, 2018;Busygina, 2001;Aláez Corral, Bastida Freijedo, 2019).In this article, we define the notion of centrifugal tendencies as a course of development that leads to a gradual depletion of the powers of the political centre in the interests of regional politicians, but does not allow their recognition as actors separate from the state (López-Basaguren, San-Epifanio, 2019, p. 212).
In recent discussions, EU regional policy has been viewed as a factor influencing Spanish centrifugal trends and contributing to centrifugal rather than to centripetal forces.The regional programs of the EU are one of the many drivers pushing regions away from the centre.This factor has played a role in virtually all centre-regions conflicts (the less developed and more developed parts of a country are each feeling exploited by the other).It does not specifically aim at alleviating grievances within member states, but it can be used as a tool by central governments to counter secessionist moods.However, it is not the only one: ethnic, linguistic, religious and/or cultural differences, play a role as much as historical memories, such as past independence.These are indirect effects that require careful study to identify causal effects because there can be many reasons for the ebb and flow of such moods.The potential explanatory factors of centrifugal trends in Spain, identified by scholars, are surveyed in the section "Research criterion".Spain is not the only such case in the EU political landscape; in many European countries the conflict between the centre and its regions is on the political agenda.In the UK, the dynamics of the centre-region conflict play a special role in the development of the country.The peculiarity of Spain is that it is a Western European country that peacefully transitioned to democracy after autocratic rule (Prokhorenko, 2015).
Research on this topic has still left several important questions unanswered.There have been a limited number of studies investigating the factors and reasons that lead to decentralization calls.In this regard, the case of Spain allows us to identify the multidimensional aspect of the adaptation of regional political structures to the institutional framework of the EU.
The study reveals which factors contribute to centrifugal rather than to centripetal forces, based on the analysis of the dynamics of centrifugal tendencies in Spain using crisp-set Qualitative Comparative Analysis (csQCA).Therefore, the exploration of this issue presents a worthwhile research avenue.
The research is divided into two phases.First, csQCA is employed to identify all the possible reasons which might have an impact on calls for autonomy and find them in the 17 Autonomous Communities (AC) in Spain.Secondly, we obtain a combination of factors to see what cases are more viable and whether the EU regional policy makes a difference.We also highlight the main results obtained using csQCA in greater detail and outline promising areas of future research in the field.

Methods
Many studies have explained the growth of centrifugal trends in Spain.However, the influence of the EU on these dynamics is not adequately addressed in the literature, especially testing these theories using csQCA.
QCA was introduced by Charles Ragin to analyse complex causality, i.e. the interplay between conditions within the case.The factors which determine the outcome are called conditions (not independent variables as in quantitative approach).In contrast to mainstream quantitative methods, where the influence of separate independent variables on the dependent variable is measured, QCA is a case-oriented approach capable of dealing with asymmetric 1 and complex conjunctural causality 2 (Schneider, Wagemann, 2012).In order to measure the significance level of independent variables, QCA uses set theory concepts such as sufficiency (all cases with condition X also have conditioned outcome Y) and necessity (all cases with outcome Y also have condition X, i.e. without X, Y cannot occur).It allows the researcher to determine correlations and casual relationships between the conditions and the outcome, i.e. so-called deterministic relationships (Schneider, Wagemann, 2012).
QCA combines quantitative and qualitative approaches.It qualitatively addresses the heterogeneity and complexity of cases, in terms of causality, and it quantitatively defines common patterns for all the cases using Boolean algebraic logic.That is why QCA is the best fit to analyse a population of 10-100 cases.This sample size is the most problematic for traditional qualitative (too many cases for detailed study) and quantitative (too few cases for statistical interference) methods. 1 Asymmetric causality-the recipe for the occurrence of the outcome is not simply the opposite of the recipe of the non-occurrence of the outcome.The explanation of the non-occurrence of the outcome cannot be derived from the explanation of the occurrence of the outcome.The non-occurrence of the outcome needs to be assessed separately.For example, the gender quotas in the Parliament lead to the increased political participation of women but the non-existence of such quotas does not necessarily lead to the decrease of women political. 2Complex conjunctural causality-different combinations of conditions can lead to the same outcome.For example, the increase of women political participation can be conditioned by the introduction of gender quotas and high economic level, or the same outcome can be conditioned by the low level of corruption and high level of democracy.
The result of QCA is the configuration (deterministic combinations of conditions in which the researcher is interested in) of conditions leading to the occurrence or non-occurrence of the outcome.But such configurations should not be interpreted as a final solution.They are measured in relation to the concrete case as the result of the logical minimisation process 1 .Then, based on theoretical knowledge, the final interpretation of the result can be concluded.
QCA is the most appropriate method for this research.First, we study possible reasons for the increase of centrifugal tendencies in Spain at the level of the autonomous communities.There are 17 autonomies in Spain: QCA is a reasonable choice of method for sample size of 12 or more cases (Fainshmidt, Witt, Aguilera, Verbeke, 2020).Secondly, we are interested in looking for reasons that can increase support for autonomy or result in their absence.Since there is no single reason that leads to this outcome, it is necessary to study the conjunctural causal relationships in the complex, and QCA allows us to test these.Thirdly, the use of QCA is a new approach to the study of causality in the domestic policies of Spain.

Research criterion
This section focuses on the five main conditions used in this research: Regional Policy of the EU (or Cohesion policy of the EU); Regional Authority Index (RAI); the level of economic development; regionalist parties, and regional language as a condition of identity.

Regional Policy of the EU
The EU is widely recognised as the most developed form of supranational integration in the modern world.Several studies have shown that participation in integration structures allows regions to gain access to wider free trade areas without appealing to the centre (Alesina, Spolaore, 2003).Large states provide economic benefits to their regions, including low social spending per capita for the regional budgets, and access to trade zones without customs restrictions.However, supranational economic integration can lead to a decrease in a region's necessity to be the part of the state by removing interstate trade barriers as the supranational association is the guarantor of the region's access to the free trade area.(Brancati, 2014: 70) Accordingly, participation in the EU could result in more requests for increased regional powers.The EU's regional policy aims to ensure the harmonised economic development of EU regions, through investments from the European structural and investment funds (ESIF).The Maastricht Treaty has strengthened the role of regions at the European level, by creating a Committee of Regions.This institution consists of local and regional representatives who are capable of influencing the decision-making process in the supranational institutions of the EU (Debus, Gross, 2018.Regional politicians can also influence the process of decision-making at the first stage in the European Commission.
To operationalise the regional policy influence on centrifugal tendencies in Spain, the gross volume of ESIF injections in each Spanish AC is considered according to the financial plan of 2014-2020.The preference for this over the dependency index 1 or per capita ESIF injections in the region is because the overall investment volume is more important.Business and political elites have the easiest access to, and management of, ESIF money, not the Spanish civil society.
Consequently, a greater amount of investment creates a demand for more independence in managing these funds.As Downs (2000: 69) states, "The EU, as resource provider, certainly encourages regional aggressiveness, yet there is dramatic cross-national variation in the form, as well as the results of efforts, to secure funds and other benefits of EU membership".

Regional Authority Index
The territorial structure of the state and the institutional relationship between the centre and regions can be viewed as the most important prerequisite for the dynamics of centrifugal tendencies (Bache, Jones, 2000).Some researchers agree that there is a strong interdependence between economic and political decentralization (Bache, Jones, 2000).An important component of accelerating the process of internal devolution is the presence of pre-established territorial communities.The regions that receive additional powers quickly (historical autonomies in Spain) have an opportunity to participate more effectively in the decision-making process and request more autonomous rights (Bache, Jones, 2000).
In order to assess the impact of the regional powers on centrifugal tendencies, we use data showing the real rights of autonomies.The Regional Authority Index (RAI) project measures the political independence of a region, i.e. the ability to make decisions on their own and to participate in making decisions regarding their status (Panov, 2020: 105).Within RAI, there is a "self-rule" and "shared rule" dimension: the powers that the regional authorities possess (institutional authority, political powers, tax autonomy, level of representation); and the powers that regional authorities have on a national scale (tax control, the possibility for constitutional reforms, control of the executive branch) (Hooghe et al., 2010).The combination of these indicators is the final index, according to which it is possible to operationalise this factor in a QCA model.

Level of economic development
Developed countries are more likely to adopt decentralization policies than developing ones (Schakel, 2010).A high level of per capita income leads to demands for greater autonomy because the benefits of decentralization are more tangible than in less developed countries (Bahl, 1999).Milanovic and Sambanis (2011) show a correlation between the demand for higher sovereignty and regional income; richer regions can demand more sovereign rights to control linguistic, educational, and/or tax policies.However, these opportunities have consequences, including reduced economic integration with the rest of the state, costs for business, and slowing economic growth.In this regard, the size of the region matters, since a larger economy results in lower costs per capita.Small but wealthy regions also tend to demand more sovereignty, especially if they are budget-donor regions (Milanovich, Sambanis, 2011: 8-9).The problem of "internal colonialism" arises when state policy creates economic inequality by developing some regions at the expense of others.(Hechter, 1975).
In this study, we use GDP per capita as an index of the level of economic development in different regions.This approach is widely used as an indicator of the economic well-being of a region (Debus and Gross, 2017; Brancati, 2016; Crescenzi and Giua, 2018).Within the EU, GDP per capita is the main parameter according to which the regions receive investment from the ESIF.
The calibration of the Spanish autonomies is carried out based on the median, which makes it possible to objectively identify the aggregate of regional economic development.

Regionalist parties
Regionalist parties can be considered the political manifestation of the existence and relevance of the centre-periphery cleavage in a given political system (Massetti, Schakel, 2015: 866).The key issue for such parties is the promotion of the territorial issue and a greater degree of self-government.Colino (2020: 72) suggests focusing attention on the activities of regionalist parties and social movements as engines of demand for more autonomy.
In this article, we define regionalist parties by the criteria proposed by Massetti and Shakel (2015).First, regionalist parties are self-contained political organizations that contest elections.
Second, these parties are present only in a particular sub-territory of the state, and/or are candidates for contestation in regional elections but not national ones.Third, the political activities of such parties are limited to the territory of the region, which results in the protection of interests for only their region.Finally, the main task of regionalist parties is to achieve, protect, and/or strengthen territorial self-government.According to Martínez-Tapia (2016), regional elites are interested in intensifying and maintaining the latent stage of the centre-region conflict to achieve political goals.
We refer to the results of the regional elections in Spain in 2019 to determine the presence/absence of regionalist parties in sub-national bodies (Parliaments, Cortes, Assemblies, General Junta).The presence in sub-territory bodies means opportunities for regionalist parties to influence the decision-making process.The election results were obtained based on data from the Spanish socio-political newspaper, 'El País'.

Regional language as a condition for identity
The language of an autonomous community, or a regional language is emphasized, by some authors as a condition for the identification of citizens with their region (Álvarez Pereira, Portos, Vourdas, 2017: 3).The rise of secessionism in such highly decentralised regions as Catalonia, can be explained from cultural and linguistic points of view.Koval (2012: 71) highlights that the Catalan language in Catalonia can be considered as an identity foundation.Regional language was embedded in the Statute of Autonomy as mandatory, and prevailed over Castilian ones.The regional language in other autonomous communities does not play an important role in connection with regionalist sentiments.For example, autonomous communities such as Valencia and the Balearic Islands, demonstrate resistance to any separatist sentiments and there is moderate support for greater autonomy.A distinctive minority language should be evaluated only in connection with other factors that determine support for greater decentralization (Álvarez Pereira, Portos, Vourdas, 2017: 3).
As an indicator of the regional language level as a matter of identity, we use the results of surveys conducted in each AC by the Center for Sociological Research in 2012 1 .In particular, the respondents were asked whether they consider the language of their AC as a condition for the identification of citizens with this region.Calibration of the results was carried out on the basis of the median.
Regionalist parties were chosen only if they have seats in parliament, and therefore have an opportunity to influence the political decision-making process.Language as a condition of identity was taken from a large-scale survey in 17 ACs by the Center for Sociological Research conducted in 2012.The ESIF amount was obtained by combining the distributions of EU funds.
In this case, the regions of Spain that have received funds from the European Regional Development Fund, the European Social Fund, and the European Agricultural Fund for Rural Development.GDP per capita at the beginning of 2014 was used because it is necessary to reveal the allocation of funds from the EU and not their direct expenditure.In other words, the GDP is 1 The most recent study where the regional languages as a factor of identity in Spain are measured.
taken "clean" without injections from the ESIF, which allows us to consider this indicator as a separate condition from the ESIF.Finally, support for autonomy was calculated by combining the following two responses from the Center for Sociological Research polls conducted in 2019: "I believe that Spain should represent a state in which the autonomous communities have more rights than they currently have" and "I believe that Spain should represent a state recognising the autonomies' right to independence".

Sources: Compiled by the authors based on data from El País, CIS Estudio N2956, CIS Estudio N3247, EUROSTAT, Hooghe et al. (2010, p. 359), DG Regio
The results of the calibration of conditions and the outcome using csQCA are presented in Table 2. Regionalist Parties (RP) equal to 1 if such parties are represented in the regional parliaments, and 0 otherwise.RAI equals to 1 if the result exceeds the average number (13.5) for the ACs.For the other conditions, the calibration was carried out via the median for two reasons.
Second, if we study the level of autonomy in the autonomous communities carefully, we will notice that the average level is higher than the median.This result can be explained because support for more autonomy in Catalonia and the Basque Country is significantly higher than in all other regions.These rates increase the average for Spain as a whole; therefore, we use the median in order to neutralise this effect.

Sources: compiled by the authors
In the process of logical minimisation, a condition or configuration of conditions can be considered sufficient if it meets the consistency criterion not less than 0.75 (Schneider, Wagemann, 2012).The consistency criterion of csQCA is the proportion of cases in the truth table that lead to a given outcome.Thirteen significant configurations were consistent.The analysis was performed in a fuzzy-set QCA program.The program generates solutions of three types: a complex solutionthe interpretation of the configuration with logical remainders that are possible combinations of conditions, which, however, do not occur in the existing database; a parsimonious solution-all configurations with potential combinations are considered as positive; an intermediate solutiononly some configurations with possible combinations are considered (easy counterfactuals).In this study, we chose the intermediate solution as a compromise option 1 (Table 3). 1 Such an approach was used in Inguanzo (2018) to analyse the recognition of indigenous rights in South East Asia.Catalan coalition has the declared aim of not only expanding autonomy but also gaining sovereignty.This declaration of sovereignty is reinforced in the political discourse by the potential benefits from independent economic development.
In the cases of Valencia and Canarias, the role of EU funds is emphasized.Both autonomous communities cannot be included in the list of economic leaders of Spain (GDP per capita is below the national median) and they receive assistance from EU funds, which results in a certain financial dependence on the ESIF.According to QCA results, the existence of regionalist parties serves to strengthen the authorities' ability to manage the EU funding streams on their own.In this sense, the regional policy of the EU stimulates attempts to directly interact with the EU, bypassing the central authorities.
In configurations that lead to less autonomy, we can observe the following configuration: a low level of economic development, a low volume of EU funds and the absence of language as a factor of identity.The cases of Andalusia, Asturias and Galicia are of particular interest in this configuration since the lower level of economic development coupled with a high RAI result mean that there is a lower level of centrifugal tendencies.If we compare this group with the group including Catalonia, the Basque Country and Navarra (as the most "centrifugal" ones), the difference leading to other outcomes lies in both the level of economic development and the level of dependence on the ESIF (as a ratio of EU funds to GDP) 1 .
At first glance, the level of regional authority as a factor does not play a decisive role in our analysis.Most Spanish autonomous communities have a similar set of rights and are included in a single fiscal system; the only exceptions are the Basque Country and Navarra.The lack of differentiation of regional powers leads to the contradictory influence of RAI on centrifugal tendencies.However, the pre-established regional powers should be taken into account as a factor because the presence of quasi-federal rights creates additional channels of influence and expands the opportunities for lobbying at different levels of governance, including at the EU level.In this regard, there is a clear relationship between the opportunities created by the EU regional policy and sub-territory structures capable of using these opportunities effectively.

Conclusion
In this research, we trace the impact of EU regional policy on centrifugal tendencies in modern Spain via csQCA.This method determines that even though Catalonia and the Basque Country are traditionally distinguished by strong centrifugal sentiments, there is no direct influence of EU regional policy in these regions.Centrifugal tendencies are fueled by the EU regional programs in such regions as Valencia and Canarias.In Catalonia and the Basque Country, a direct relationship was found between the centrifugal tendencies and the existence of regionalist parties in the parliaments of ACs, a high level of economic development and a high level of regional language as a component of identity.
In future studies, we propose to focus on the role of EU regional policy in reinforcing the centrifugal tendencies in Valencia and the Canarias using other methods, for instance, processtracing, which can provide a deeper understanding of causal dynamics in causally similar cases. 1 Dependency index median calculated by Debus and Gross approach in Spain is 2,06.Andalucia -4,47; Asturias -3,27; Galicia -3,92; Catalonia -0,76; the Basque Country -0,51; Navarra -1,15.

Table 3 .
Minimal formulas for intermediate solutionsSources: compiled by the authors (Capital letter conditions indicate the presence of a condition for the outcome, small letter-their absence.Left column contains conditions that lead to centrifugal tendencies, right-to their absence)First, the results demonstrate a direct interdependence between the condition of RPs, the level of economic development, a high level of regional language, and demands for more autonomy.In both formulas, these conditions are necessary for a high level of autonomism.The examples of Catalonia and the Basque Country as the most economically developed regions are indicative here.There are strong RPs in the regional parliament in both autonomous communities.In the Basque Country, 28 of the 75 seats are held by the Basque Nationalist Party [Partido Nacionalista Vasco] and it has formed a ruling coalition with the local branch of the Socialist Party.In Catalonia, the ruling coalition is Together for Catalonia [Juntos por Cataluña] and Republican Left of Catalonia [Izquierda Republicana de Cataluña].Both coalitions have wide support among the population and can influence the national political agenda.In particular, the