Courts, federalism, and judicial behavior

In federal and quasi-federal democracies, CCs are often expected to function as institutional referees in conflicts between levels of government. From Madison’s vision in Federalist No. 51 to modern comparative constitutional theory, courts have been described as counter-majoritarian actors tasked with upholding constitutional principles against both popular majorities and centralizing tendencies. In this view, CCs help protect minority or subnational interests and maintain the federal balance by resolving jurisdictional disputes and reviewing legislation that may infringe on subnational autonomy (Halberstam Reference Halberstam, Whittington, Kelemen and Caldeira1998; Watts Reference Watts1999; Kössler Reference Kössler, Requejo and Sanjaume-Calvet2022). However, whether CCs actually fulfil this function in practice remains an open empirical and theoretical question. Courts may just as easily reinforce central dominance, depending on the political, institutional, and legal context in which they operate (Aroney and Kincaid Reference Aroney and Kincaid2017; Popelier Reference Popelier2017).

Scholars of judicial behavior typically distinguish three major explanatory models: legalist, attitudinal, and strategic (Segal and Spaeth Reference Segal and Spaeth2002; Epstein, Landes, and Posner Reference Epstein, Landes and Posner2013). The legalist model treats judges as neutral interpreters of constitutional texts, relying on precedent and formal legal reasoning. While this model retains some explanatory power, especially in low-salience or technical cases (Hettinger, Lindquist, and Martinek Reference Hettinger, Lindquist and Martinek2006), it often fails to account for variation in politically charged or jurisdictional disputes, such as those involving center–periphery conflict.

By contrast, the attitudinal model emphasises the role of judges’ ideological preferences. In federal systems, this model predicts that CCs will exhibit centralist or devolutionist tendencies depending on the ideological alignment of their members with national or regional elites. High courts often include judges who are selected by or socially embedded within the federal political class, especially in systems where appointments are controlled by national institutions (Segal and Spaeth Reference Segal and Spaeth2002; Aroney and Kincaid Reference Aroney and Kincaid2017; Epstein, Landes, and Posner Reference Epstein, Landes and Posner2013). As a result, courts may disproportionately reflect the preferences of the political center or dominant federal parties.

The strategic model views judges as goal-oriented actors who pursue their preferences while navigating institutional constraints and political pressures. This framework suggests that courts may align with powerful actors, such as executives or governing coalitions, when doing so preserves their institutional legitimacy or avoids political backlash (Shapiro Reference Shapiro1981; Epstein and Knight Reference Epstein and Knight2000; Carruba Reference Carruba2009; Staton Reference Staton2010). Strategic considerations such as collegiality, dissent aversion, or reputational concerns may also guide behavior, particularly in high-stakes federal disputes where consensus is desirable or politically safer (Epstein, Landes, and Posner Reference Epstein, Landes and Posner2011).

In federal systems, these theoretical models help explain why CCs often exhibit a centralist bias (Bzdrea Reference Bzdrea1993; Aroney and Kincaid Reference Aroney and Kincaid2017; Filippetti, Rondinella, and Tuzi Reference Filippetti, Rondinella and Tuzi2025). This tendency is especially visible in multinational or asymmetrical federations, where national political majorities may seek to constrain regional autonomy. In such contexts, courts may contribute to this centralization, whether due to ideological affinity, institutional design, or strategic calculation.

Popelier (Reference Popelier2017) outlines four structural factors that help explain the centralist or non-centralist stance of courts: the extent of subnational representation at the federal level; judicial appointment procedures; the structure of the party system; and the degree of constitutional asymmetry. Later work focusing on Belgium also points to case salience, legal merit, and deliberation procedures as drivers of variation (Popelier and Bielen Reference Popelier and Bielen2019). In Italy, empirical analyses show a marked trend toward recentralization after 2011, with decisions by the Italian Constitutional Court increasingly favouring federal over regional claims (Filippetti, Rondinella, and Tuzi Reference Filippetti, Rondinella and Tuzi2025). Similarly, Dalla Pellegrina and Garoupa (Reference Dalla Pellegrina and Garoupa2013) found that political congruence between the Prime Minister and the Judge-Rapporteur significantly increased the probability of a centralist ruling.

In the Belgian case, Dalla Pellegrina et al. (Reference Dalla Pellegrina, De Mot, Faure and Garoupa2017) found that, contrary to traditional views of legal neutrality, political and ideological variables did influence judicial outcomes, though less strongly than in presidential systems like the U.S. These studies support the broader view that CCs cannot be presumed neutral in federal disputes; instead, their decisions often reflect embedded political and institutional dynamics.

Turning to the Spanish case, the empirical literature presents mixed findings regarding the political behavior of the SCC. On the one hand, several studies suggest that political alignment plays a significant role in shaping judicial outcomes. For instance, Garoupa, Gómez-Pomar, and Grembi (Reference Garoupa, Gómez-Pomar and Grembi2013) and Garoupa, Gili, and Gómez-Pomar (Reference Garoupa, Gili and Gómez-Pomar2021b) show that partisan and personal loyalties influence decisions, especially during periods of intense political polarization. Similarly, Dalla Pellegrina, Garoupa, and Gili (Reference Dalla Pellegrina, Garoupa and Gili2020), using ideal point estimation, identify bi-dimensional ideological preferences among judges in a Court-like institution in Catalonia. Muñoz and Rodilla-Lázaro (Reference Muñoz and Rodilla-Lázaro2024) further contribute to this line of research by demonstrating that party alignments are key to understanding which regional laws are challenged before the SCC.

On the other hand, a series of studies focusing on conflicts of competence and amparo appeals (López-Laborda, Rodrigo, and Sanz-Arceaga Reference López-Laborda, Rodrigo and Sanz-Arceaga2018, Reference López-Laborda, Rodrigo and Sanz-Arceaga2019, Reference López-Laborda, Rodrigo and Sanz-Arceaga2024) find limited evidence of systematic political bias in SCC rulings. These authors conclude that legal and procedural variables, rather than ideological congruence, explain the Court’s decisions in these domains. This divergence in findings may reflect differences in both the types of cases analysed and the unit of observation. While many studies treat entire Court decisions as homogenous outcomes, our approach disaggregates each decision into individual rulings and focuses specifically on legislation challenged by institutional actors, a subset of cases where political dynamics and intergovernmental tensions are most likely to manifest.

Building on this scholarship, our contribution systematically tests whether political alignment, court ideology, and parliamentary dynamics influence outcomes in center–periphery constitutional litigation. We do so by using a novel dataset of 1,888 rulings issued by the SCC between 1981 and 2023, coded by type of legislation, litigant, and political context. This granular design allows us to examine whether and how the SCC behaves as a politically responsive institution in disputes that directly involve the balance of territorial power in Spain.

To support our theoretical expectations, we draw on insights from recent literature on territorial politics and party systems. In the Spanish case, the two dominant state-wide parties (the center-left PSOE and the center-right PP) have historically converged in defending national cohesion, mainly during episodes of secessionist mobilisation (Maravall Reference Maravall2008; Amat Reference Amat2012; Brown-Swan and Cetrà Reference Brown-Swan and Cetrà2020; Amat and Rodon Reference Amat and Rodon2021). Ideologically, however, conservative parties tend to favour centralised authority, while culturally liberal and left-wing parties are more inclined toward decentralization and recognition of regional autonomy (Alonso, Cabeza, and Gómez Reference Alonso, Cabeza and Gómez2013; Toubeau and Wagner Reference Toubeau and Wagner2016). These ideological divides are reflected not only in legislative preferences but also in judicial appointments and patterns of constitutional litigation. Moreover, a growing body of work has shown that minority nationalist parties can influence the territorial agenda when they are pivotal to forming or sustaining national governments (Meguid Reference Meguid2008; Field Reference Field2016; Ponce-Rodríguez et al. Reference Ponce-Rodríguez, Hankla, Martínez-Vázquez and Heredia-Ortiz2018). These dynamics shape both the strategic incentives of litigants and the broader political context in which the Court operates.

These literatures suggest that judicial decision-making in federal and quasi-federal systems, notably those marked by asymmetry and intergovernmental tensions, is likely to reflect patterns of political alignment, ideological commitments, and coalition dynamics. These are the factors we explicitly incorporate into our theoretical model and hypotheses, which we introduce in the next section.

The case of Spain

Spain’s territorial model, forged during the democratic transition, gave rise to the Estado de las Autonomías, an asymmetrical and evolving system of decentralization. Although the 1978 Constitution does not define Spain as a federation, it granted substantial autonomy to the regions (Autonomous Communities, ACs), notably through a fast-track route for Catalonia, the Basque Country, and Galicia (Article 151 SC), and a more gradual process for the remaining regions (Article 143 SC). The decentralization process was shaped by negotiated transfers of competencies and institutionalised through the Pactos Autonómicos of 1981 and 1992 (Bossacoma Busquets and Sanjaume-Calvet Reference Bossacoma Busquets, Sanjaume-Calvet, Popelier and Sahadžić2019). Some scholars describe this model as ‘a federation in all but name’ (Watts Reference Watts1999), while others underscore its incomplete nature (Grau-Creus Reference Grau-Creus and Wachendorfer-Schmidt2000) and the imbalance between high levels of self-rule and limited shared-rule mechanisms (Elazar Reference Elazar1987).

Within this framework, the SCC was established as the key institution for constitutional interpretation, under the 1979 Organic Law (LOTC). Based on the Kelsenian model of concentrated constitutional review, the Court comprises 12 justices appointed for non-renewable nine-year terms. Appointments are made by the Congreso de los Diputados, the Senado (partially based on regional parliaments’ proposals), the government, and the General Council of the Judiciary, all requiring qualified majorities. While these procedures nominally uphold meritocratic standards, the process is ultimately controlled by central institutions, with only symbolic regional input. Compared to other European constitutional courts, the SCC handles a heavy caseload, which positions it not only as a legal authority but also as a key actor in mediating broader political and social conflicts (Vallbé Reference Vallbé and Pócza2024).

Despite its intended role as a neutral judicial body, the SCC’s structure renders it susceptible to politicisation. The selection of justices by political institutions, the absence of regional vetoes in the appointment process, and the Court’s involvement in high-stakes political disputes have all drawn criticism. While the Court has regularly included magistrates from Catalonia and the Basque Country, this has not translated into balanced territorial representation. Its jurisprudence has occasionally been perceived as favouring recentralization, deepening tensions in an already delicate territorial settlement. In regions with strong demands for self-government, the SCC’s legitimacy as a neutral arbiter has often been contested in the political debate.

The SCC’s functioning has been the subject of considerable empirical scrutiny. Garoupa, Gili, and Gómez-Pomar (Reference Garoupa, Gili and Gómez-Pomar2021a) found that judicial decisions were significantly shaped by judges’ personal loyalty to the party leaders who appointed them, mostly under the PSOE government of Prime Minister Zapatero. Similarly, Dalla Pellegrina, Garoupa, and Gili (Reference Dalla Pellegrina, Garoupa and Gili2020) identified the influence of ideological orientations, especially along the progressive-conservative axis, on rulings, particularly in the context of the Catalan territorial conflict. Rodilla-Lázaro (Reference Rodilla-Lázaro2025) further demonstrated that short-term partisan incentives influence which cases are brought before the SCC, while Muñoz and Rodilla-Lázaro (Reference Muñoz and Rodilla-Lázaro2024) found that the territorial dimension shaped state-wide parties’ behavior initiating judicial review processes, with conservative central governments championing the challenge of regional laws before the Court. Nonetheless, debate persists regarding the Court’s behavior in center-periphery disputes. While several studies point to a centralist bias rooted in political dynamics (Harguindéguy, Sola-Rodríguez, and Cruz-Díaz Reference Harguindéguy, Sola-Rodríguez and Cruz-Díaz2020; Sanjaume-Calvet and Paneque Reference Sanjaume-Calvet and Paneque2023; Vallbé Reference Vallbé and Pócza2024), other work finds limited or no empirical evidence of systematic favouritism towards the central government (López-Laborda, Rodrigo, and Sanz-Arceaga Reference López-Laborda, Rodrigo and Sanz-Arceaga2018, Reference López-Laborda, Rodrigo and Sanz-Arceaga2019).

Comparative evidence from other European constitutional courts further contextualises the Spanish case. In their analysis of the Belgian Constitutional Court, Dalla Pellegrina et al. (Reference Dalla Pellegrina, De Mot, Faure and Garoupa2017) conclude that ‘our empirical results seem to be inconsistent with the view, generally held by legal scholars, that ideology or political variables play no role in explaining behavior at the Belgian Constitutional Court’ (p. 345). That said, they caution that Belgian judges appear less politically motivated than their U.S. counterparts. In Italy, Filippetti, Rondinella, and Tuzi (Reference Filippetti, Rondinella and Tuzi2025) highlight a marked trend toward recentralization in CC rulings after 2011, driven largely by the central government. Earlier research by Dalla Pellegrina and Garoupa (Reference Dalla Pellegrina and Garoupa2013) also found that political alignment between the case rapporteur and the Prime Minister significantly increased the likelihood of the Court supporting centralist positions.

A strategic model of constitutional review in a decentralized polity

The institutional structure, political composition, and contested legitimacy of the SCC described above raise important questions about how it behaves in practice when adjudicating center–periphery disputes. To address these questions, we develop a strategic framework that models the behavior of CCs in decentralized systems. Our goal is to explain how political alignments and institutional incentives shape judicial decision-making in cases involving national and regional legislation. This framework builds on the patterns identified in the Spanish case and draws from broader insights in the comparative literature.

Constitutional courts in federal and quasi-federal systems are often portrayed as neutral arbiters in conflicts between levels of government. However, their behavior is shaped by political alignments, institutional arrangements, and broader strategic considerations. In the Spanish context, where a strong central state coexists with asymmetrically empowered regional governments, the CC operates in a political environment shaped by intergovernmental tension and dominated by two state-wide parties. Its decisions, while legally framed, are made within a set of predictable political constraints and incentives.

We consider three types of actors: the central government, regional governments, and the CC. The central government typically seeks to assert the primacy of national legislation and maintain constitutional cohesion. Regional governments vary in their political orientation: some are led by state-wide parties, while others, usually in Catalonia and the Basque Country, are governed by region-based or minority nationalist parties that advocate for greater autonomy. The CC, although formally independent, is appointed through procedures dominated by national institutions, which means that it is often composed of members whose preferences reflect those of the central political system.

We assume that the Court behaves as a strategic actor, sensitive to both political alignment and institutional incentives. Its decisions, to strike down or uphold challenged legislation, are shaped by the political configuration of the case, the ideological profile of the Court, and the broader dynamics of parliamentary coalitions. Based on this logic, we derive four expectations.

First (H1), we expect a general centralist bias: the Court is more likely to annul legislation enacted by regional parliaments than by the national legislature. This expectation is consistent with studies showing that federal high courts often reinforce central authority (Aroney and Kincaid Reference Aroney and Kincaid2017; Popelier Reference Popelier2017; Woelk Reference Woelk2020), particularly when subnational units lack strong representation in judicial appointments (Popelier and Bielen Reference Popelier and Bielen2019).

Second (H2), when the Court is politically aligned with the central government, meaning that both are controlled by the same party or ideological bloc, the likelihood of the Court supporting the federal government increases. Political congruence strengthens the Court’s role as a majority-enforcing institution, a dynamic consistent with theories of judicial deference under alignment (Vanberg Reference Vanberg2005; Garoupa, Gómez-Pomar, and Grembi Reference Garoupa, Gómez-Pomar and Grembi2013; Ginsburg Reference Ginsburg2019). This expectation stands in contrast to the findings of López-Laborda et al. (Reference López-Laborda, Rodrigo and Sanz-Arceaga2018, Reference López-Laborda, Rodrigo and Sanz-Arceaga2019), who argue that the CC’s decisions are largely insulated from political alignment or ideological bias. Our approach differs in both theoretical framing and empirical design. Theoretically, we conceive the Court not merely as an attitudinal body, but as an actor embedded in a wider strategic institutional setting. Empirically, we use a more granular unit of analysis (individual rulings within decisions) that allows us to capture variation often masked by aggregated decision-level coding. Most crucially, our focus is on judicial review of legislation (that is, the Court’s role as a negative legislator), whereas López-Laborda et al. base their analyses primarily on conflicts of competence between the central and regional governments. These disputes often center on executive actions or jurisdictional procedures rather than the constitutionality of legislative content. As a result, our analysis better captures the high-stakes confrontations in which political alignment is most likely to influence judicial outcomes. In this light, we expect political congruence to systematically shape the likelihood of the Court striking down regional laws or upholding national legislation.

Third (H3), the ideological composition of the Court shapes its decisions. A conservative-majority Court is more likely to adopt a centralist stance when evaluating legislation originating from region-based parties. Prior work suggests that conservative parties in Spain and elsewhere tend to support recentralization and uniformity both in their political discourse (Maravall Reference Maravall2008; Alonso, Cabeza, and Gómez Reference Alonso, Cabeza and Gómez2013; Toubeau and Wagner Reference Toubeau and Wagner2016) and in challenging regional legislation before the SCC (Muñoz and Rodilla-Lázaro Reference Muñoz and Rodilla-Lázaro2024), while region-based or progressive parties advocate for decentralization.

Fourth (H4), the Court adjusts its stance when the central government relies on parliamentary support from region-based parties. In such scenarios, the political costs of conflict with minority partners may encourage the Court to moderate its centralist bias. The literature on party strategies in multilevel systems highlights how governing coalitions that include minority nationalist parties tend to constrain centralization efforts (Field Reference Field2016; Basile Reference Basile2019), and how courts may strategically respond to avoid institutional backlash (Popelier and Bielen Reference Popelier and Bielen2019; Krehbiel Reference Krehbiel2021).

These expectations guide our empirical analysis of the Court’s rulings. In Section 5, ‘Case, data, and methods’, we test them using multilevel logistic regression models, treating the Court’s decision to annul a provision as a function of key political and institutional variables. This modelling approach accounts for the hierarchical structure of our data (rulings nested within years) and allows us to estimate the probability of annulment based on legislation type, political alignment, court ideology, and parliamentary dynamics. By doing so, we interpret judicial decisions not merely as legal determinations, but as politically embedded outcomes shaped by the broader institutional and strategic context of Spain’s Estado de las Autonomías.

Case, data, and methods

Dataset, variables, and units of analysis

We created a novel dataset by coding all rulings issued by the SCC between 1981 and 2023 in cases involving national or regional legislation. We depart from prior studies that treat the decision as the unit of analysis and instead follow Pócza (Reference Pócza2019) in treating rulings (the Court’s determinations on individual challenged provisions) as our unit of analysis. This avoids the distortion caused by treating internally inconsistent decisions as homogeneous outcomes.

Each ruling is coded with:

• the actor initiating the case,

• the type and origin of the legislation challenged,

• the political alignment between the Court and the relevant regional government(s),

• the ideological profile of the Court,

• whether the national government depended on regional nationalist party support in parliament, and

• the binary outcome: whether the Court annulled the provision (1) or upheld it (0).

To code the partisan identity of regional governments, we adopted a consistent rule centered on the party of the regional president and the relative dominance within coalition arrangements. When governments were formed by a single state-wide party or by coalitions clearly led by one (typically the PSOE or PP), we coded the regional government as being led by that state-wide party. Examples include the PSOE–IU coalition in Andalusia (2012–2015), the PSOE–regionalist coalition under Francesc Antich in the Balearic Islands, and the Tripartit government in Catalonia (2003–2006), all coded as PSOE-led. Conversely, when the head of government belonged to a regional party and this party either governed alone or was the dominant force in a coalition, even if supported by or in coalition with a state-wide party, we coded the government as Regionalist or Nationalist depending on the party’s ideology. For instance, the PAR–PP coalition in Aragón (1991–1993) is coded as Regionalist (PAR led the government), whereas the PP–PAR government of 1995–1999 is coded as state-wide (PP-led). Similarly, minority governments led by nationalist parties (e.g., in the Basque Country or Catalonia) but supported in parliament by state-wide parties were coded as Nationalist. This is the case, for instance, for the CiU government in Catalonia receiving parliamentary support from PP (1996-2000), or the instances when PSOE supported from parliament the PNV governments in the Basque Country. We classified the following parties as Nationalist: PNV, EA, Coalición Canaria, CiU, ERC, Junts per Catalunya, and Geroa Bai; and as Regionalist: PAR, Foro Asturias, Unión del Pueblo Navarro, Partido Regionalista de Cantabria, and Unión para el Progreso de Cantabria.

We do not classify other existing regionalist or nationalist parties in Spain (e.g., BNG, EH Bildu, Compromís, Partit Socialista de Mallorca) because, despite having been part of regional coalitions, they did not lead any government during the period covered by our data. While we recognise that this coding strategy may introduce some noise by grouping together, for instance, single-party governments and coalition governments involving regionalist or nationalist actors, it preserves internal consistency and emphasises the leadership and ideological orientation of the executive. Importantly, our empirical results remain robust despite this potential noise, reinforcing the strength of the findings. All our data and coding decisions will be made available in a public repository to ensure full transparency and reproducibility.

Table 1 displays the distribution of rulings by type of legislation and reporting actor. National legislation accounts for approximately two-thirds of rulings, with the remaining third concerning regional laws. Roughly one-third of cases are initiated by regional governments and another third by ordinary courts. The remainder comes primarily from the central government and national legislature. The Ombudsperson and regional legislatures are marginal actors.

Table 1. Frequencies of types of legislation and reporting actors

* n (%).

Figure 1 shows the evolution of litigation over time. Litigation involving national laws rose sharply in the late 1980s and again after 2010, the latter due to the Court’s backlog following the long deliberation over Catalonia’s Statute. By contrast, challenges to regional laws have remained relatively stable, peaking during moments of institutional reform in the 1990s, 2000s, and 2010s.

Figure 1. Evolution of constitutional litigation in Spain, by type of legislation.

Table 2 reports summary statistics for key variables. Unconstitutionality was declared in 43% of rulings. Regional laws were involved in 36% of cases. Conservative majorities dominated the Court in 34% of rulings. Court-government alignment occurred in 60% of cases, while Court-regional government alignment was present in 20%. In 39% of rulings, the national government depended on parliamentary support from region-based parties. Variables such as ‘Court–Regional government congruence’ and ‘Regionalist party in region’ can only be coded for rulings in which a specific Autonomous Community is clearly involved. This restricts the number of applicable cases to 754. This measurement strategy allows us to directly test the strategic logic laid out in the theoretical section, and to assess whether the SCC functions as a counter-majoritarian force or, conversely, reinforces central government dominance in legislative disputes.

Table 2. Descriptive statistics of the main variables

While our analysis is primarily quantitative and does not include a doctrinal examination of the legal content of each ruling, the consistent patterns observed across more than four decades suggest that structural and political dynamics, rather than legal complexity alone, account for the differential treatment of national and regional legislation. Our approach is thus designed to identify systematic tendencies in judicial behavior, not to assess the legal merits of individual rulings.

To do so, we estimate multilevel logistic regression models with random intercepts by year, reflecting the fact that constitutional litigation is shaped by institutional cycles, shifts in Court composition, and broader political dynamics. We do not include region as a grouping variable, for both theoretical and empirical reasons. First, our unit of analysis is the ruling, not the region, and many decisions either involve national legislation or are initiated by non-regional actors (e.g., ordinary courts), making the attribution of a unique region problematic. Second, the relevant territorial variation is already captured through substantive covariates, such as the type of legislation, the identity of the reporting actor, and the political characteristics of the regional government. Third, the representation of regions is highly unbalanced, with Catalonia accounting for 26.7 percent of decisions involving a regional government, while others fall well below 10% (see Table A1 in the Appendix below), which would create small, unstable clusters if modelled as random effects.

We also chose not to include decision ID as a grouping variable. Although some decisions contain multiple rulings, our explanatory variables are constant within each decision, and most decisions contain only a single ruling. As such, including decision-level random effects would add little explanatory power, while unnecessarily increasing model complexity and standard errors. Our modelling strategy, therefore, accounts for temporal heterogeneity while avoiding overfitting or redundant clustering. Robustness checks deal with the overrepresentation of Catalonia and will be commented on in the empirical section.

Empirical results

The majoritarian bias of the constitutional court

The first model (see Table A2 in the Appendix below) includes two covariates: the type of legislation (regional or national) and the type of case (constitutional complaint issued by ordinary courts, conflict of competence, or a regular a posteriori review). The results show that the probability of the Court declaring unconstitutionality is significantly higher when dealing with legislation produced by regional legislatures, controlling for the case type. This effect remains robust across all subsequent specifications with additional controls (see Table A2), providing strong support for our first hypothesis: that the Court is more likely to align with the federal government than with regional governments (H1). The marginal effects presented in Figure 2 illustrate a stark contrast in the likelihood of the Court declaring laws unconstitutional depending on the source of legislation. When ruling on national legislation, the Court’s predicted probability of annulment is just 0.23, while this jumps to 0.55 for regional legislation. This difference of 32 percentage points signals a strong and systematic majoritarian-centralist bias in the Court’s decisions, supporting H1. Furthermore, the model reveals notable differences in the likelihood of declaring unconstitutionality between reviews and conflicts of competence, compared to complaints brought by ordinary judges. These results hold consistently over time, as confirmed by year-specific predicted probabilities (see Figure A1 in the Appendix).

Figure 2. Predicted probabilities of the Court declaring unconstitutionality for regional and national legislation.

We expected that cases brought to the Court by the central government would have a higher probability of being supported by the Court compared to those brought by regional actors. The second and third columns of Table A2 (Appendix below) test this by including a variable indicating the actor that brings the case to the Court. Since the type of case is entirely dependent on the actor bringing it (e.g., ordinary courts can only file constitutional complaints), we exclude the type of case from the model. The results provide two key findings that support our initial hypothesis. First, as shown in the second column of Table A2, when cases are brought by actors other than the central government (the reference category), the chances of the Court declaring unconstitutionality are consistently lower. Further disaggregation by reporting actor in Figure 3 adds nuance: the central government enjoys the most favourable treatment, with a 0.68 probability of success in its appeals. In contrast, regional governments face only a 0.31 probability, less than half that of the central government. This is a striking finding given that regional governments are frequent litigants. The ordinary courts, interestingly, receive even less favourable outcomes when they appeal national legislation, indicating the Court’s additional deference to the legislature vis-à-vis the judiciary. These marginal effects emphasise that the Court not only defends the central government against regional challengers but also acts conservatively toward judicial review initiated from below the executive branch.

Figure 3. Predicted probabilities of the Court declaring unconstitutionality depending on who reports.

Our second hypothesis (that political congruence between the Court and the federal government will increase the likelihood of the Court upholding challenged national legislation and annulling appealed regional legislation) tests the impact of a Majority Front, or Court-Government political alignment. The results are presented in columns 4 and 5 of Table A2.

The marginal effects from models including Court-government congruence reveal that when the Court and the federal government are politically aligned, the Court is less likely to strike down legislation. This lower judicial assertiveness in scenarios of national-level congruence suggests a strategic behavior of institutional alignment: the Court appears more deferential to the legislature when both are controlled by the same party family.

Importantly, this effect is not conditional on the identity of the reporting actor, and the interaction terms are not significant. That is, the reduction in judicial activism under congruence applies across the board: whether the case is brought by a regional government, a national institution, or an ordinary court, the Court’s likelihood of declaring unconstitutionality falls when it is aligned with the central government. This gives support to the hypothesis that institutional alignment shapes judicial behavior at the systemic level.

In contrast, the last column of Table A2 in the Appendix below shows that there is no significant effect in a scenario where political alignment exists between the Court and a regional government that brings the case. In Section 1, ‘Introduction’, we raised concerns about whether the leverage held by territorialised minorities in the Court could offset the Court’s inherent bias toward federal interests, stemming from the appointment mechanisms of its members. The findings suggest that the Court’s bias toward the federal government typically outweighs any incentives to support regional governments, even when both the Court and the regional government are controlled by the same party.

The role of political ideologies and political majorities

To test our third hypothesis (H3) (the Court is less likely to oppose regional legislation when it is dominated by pro-decentralization parties than when it is dominated by federation-wide parties that oppose decentralization), we first need to assess whether there are differences between the two major Spanish parties in their stance on decentralization.

The alliance between leftist and minority nationalist parties was a defining feature of Spain’s constitutional moment and has remained a recurring element, to varying degrees, in the country’s territorial dynamics during its democratic period (Colino Reference Colino, Muro and Lago2020; Moreno Reference Moreno, Burgess and Gagnon2010). Moreover, evidence suggests that Spanish voters generally perceive the center-left PSOE, and even more so the leftist Podemos, as more supportive of decentralizing powers to the regions. In contrast, the center-right Partido Popular has traditionally advocated for the (re)centralization of political authority in Spain (Maravall Reference Maravall2008; Alonso, Cabeza, and Gómez Reference Alonso, Cabeza and Gómez2013). This ideological divide is significant in the ongoing debates over regional fiscal flows among others (León Reference León2007; Espasa and Bosch Reference Espasa, Bosch, Bosch, Espasa and Solé-Ollé2010). Consequently, we hypothesised that the Court’s majoritarian bias would be especially pronounced when it operates under a conservative majority.

The results of these models (see Table A3 in the Appendix below) provide strong evidence supporting our H3. The first model includes the main covariates without interaction terms, showing that a conservative-majority Court is significantly more likely to strike down legislation, controlling for both the type of legislation and the actor bringing the case. The second model incorporates an interaction between the Court’s ideological composition and the reporting actor, revealing that a conservative Court further decreases the likelihood of success for regional actors challenging national legislation. As shown in Figure 4, the marginal effects suggest that Court ideology matters significantly. Under a conservative-majority Court, the probability that a case brought by a regional government results in annulment rises sharply. Conversely, the probability of the Court ruling against central government positions falls.

Figure 4. Predicted probabilities of the Court declaring unconstitutionality by reporting actor and court ideology.

The Court’s majoritarian bias is not only institutional but also ideological: when the Court is dominated by conservative justices, it becomes more assertive in defending centralised authority. The marginal effect of a conservative Court is especially pronounced when cases involve legislation from regions governed by regionalist or nationalist parties. These findings corroborate the hypothesis (H3) that conservative majorities exacerbate centralist judicial tendencies.

The remaining results further illustrate the impact of a conservative-majority Court on judicial outcomes. When the Court is conservative, the probability of it declaring regional legislation unconstitutional rises significantly (third column). This likelihood increases even further when the reported region is governed by a region-based party (fourth column) and reaches its peak when the regional government is controlled by a minority nationalist party (fifth column). These findings underscore the heightened majoritarian-centralist bias of the Court under conservative dominance, particularly in cases involving nationalist-led regional administrations.

The results for our fourth hypothesis (H4) (that the Court’s opposition to regional legislation will be tempered when the federal government relies on parliamentary support from regional nationalist parties) are both surprising and illuminating. The first column of Table A4 (in the Appendix below) reveals that reliance on minority nationalist party support in the legislature is associated with a higher likelihood of the Court striking down regional legislation, contrary to expectations. However, introducing an interaction between minority support and the type of legislation (second column) aligns with our hypothesis. The interaction shows that the Court strategically calibrates its decisions in response to political dynamics: when the national government depends on minority nationalist parties, the Court is less likely to strike down regional legislation and more likely to invalidate national legislation. These findings reinforce the view of the Court as a highly strategic actor, balancing power dynamics by countering the central government’s leverage when it is bolstered by minority nationalist parliamentary support.

The interaction plot in Figure 5 shows how the Court’s behavior shifts in response to legislative coalitions. When the central government does not depend on minority nationalist support, the probability of the Court annulling regional legislation remains high. But when minority nationalist parties support the government in Parliament, this probability drops substantially, while the probability of invalidating national legislation rises.

Figure 5. Predicted probabilities of the Court declaring unconstitutionality depending on the type of legislation and on the support of nationalist parties to the central government.

This reversal of expected majoritarian outcomes gives support to H4: the Court adapts its rulings to reflect changes in political equilibrium when the central government’s position depends on regionally based parties. This strategic adjustment suggests that the Court is sensitive not only to institutional alignment but also to parliamentary interdependence.

Robustness checks

Catalonia accounts for 26.7% of all CC decisions involving regional governments in our dataset, raising the question of whether our main results are disproportionately driven by this single case. To assess this possibility, we perform two complementary robustness checks.

First, we re-estimate all models after excluding all rulings involving Catalonia. The results, reported in Tables A5, A6 and A7 in the Appendix below, remain broadly consistent with the full models. The baseline effects, including those linked to the political alignment between the Court and the federal government and the influence of conservative court majorities, are preserved. Although the effect of regional legislation loses statistical significance in some specifications, its direction remains substantively similar. Importantly, the hypothesised interaction between nationalist parliamentary support and reduced annulment of regional laws is still evident (see Table A7).

Second, we conduct a more targeted test, excluding only Catalonia between 2012 and 2021, a period of especially intense conflict between the regional and central governments. As shown in Tables A8, A9, and A10, A11, and A12 in the Appendix below, these results align even more closely with the full models than the complete exclusion of Catalonia. This suggests that our findings are not driven by the most politically charged episodes of Catalan politics but instead reflect broader institutional and partisan dynamics embedded in Spain’s constitutional framework.

While the Basque Country is the second most represented region in our data (13% of decisions involving regions), its influence is substantially lower than that of Catalonia. However, we also replicated our analysis excluding all cases involving the Basque Country, and the results (see Appendix below, Table A11) remain virtually identical to those obtained when excluding Catalonia, confirming that our findings are not driven by any single region. Other regions, such as the Canary Islands and Andalusia, contribute roughly 8% each. Given this distribution, we judge it unnecessary to repeat the same exclusion tests for each individual region. Moreover, our modelling approach (using multilevel logistic regressions with year-level random effects) already offers a conservative structure that helps absorb potential regional clustering effects without resorting to arbitrary case deletion. As such, the robustness checks reinforce our conclusion that the CC’s patterns of behavior are not driven by exceptional cases but reflect consistent strategic and institutional dynamics across time and space.

Finally, given the four-decade span of our data, we also examine whether the probability of annulment varies systematically over time. Using year-level random intercepts from the multilevel model, we compute predicted probabilities of unconstitutionality for each year. As shown in Figure A1 in the Appendix, the trends remain remarkably stable throughout the period 1981–2023, with no visible structural shifts or temporal breakpoints. This suggests that the Court’s behavior, although politically conditioned, does not exhibit major temporal discontinuities in its likelihood of annulling legislation.