Regional Policy Trajectories in the Spanish Education System: Different Uses of Relative Autonomy

: Federal and highly decentralized political systems open different spaces to interpret, adapt


Regional Policy Trajectories in the Spanish Education System: Different Uses of Relative Autonomy
Education policy moves across different scales and in different directions in the context of globalization. While global mechanisms of influence in national education systems have been widely analyzed and discussed (Dale, 1999), less is known about policy influences at and between lower scales of governance. Federal and highly decentralized political regimes open different spaces to interpret, adapt, and enact external policy prerogatives and global educational reform ideas within specific country contexts (Savage & Lingard, 2018;Wallner et al., 2020). The multiscalar policymaking process has its own particularities in federal countries, as regional governmentsunderstood here as the governments of subnational units-may have education agendas either aligned or opposed to the national agenda. Regional policy trajectories can be understood as subordinated to a hierarchical distribution of policy responsibilities (including coercive relationships between federal and state actors) or as the product of policy learning and emulation dynamics between regions. However, if we pay further attention to their political particularities, regions can also be seen as spaces of struggle and singularization. From this perspective, regional governments are political subjects with their own agendas and interests that relate to national and supra-national scales through their own power games.
The case of Spain is highly relevant for exploring these dynamics. The process of education policy decentralization was initiated with the restoration of democracy by the end of the 1970s. For the last 50 years, the Spanish State has tried to balance the recognition of historical regions (i.e., Catalonia, Basque Country, Galicia) with a wider process of decentralization in main policy domains, including education, within the whole territory. This process, which has not advanced without tensions, is key to understanding the evolution of education policy in Spain. In the last 2 decades, Spanish regions have not been immune to global education reform pressures and the influence of global policy ideas that gear around the promotion of private provision, school choice mechanisms, and New Public Management (NPM) policies in education (e.g., school autonomy and performancebased accountability). However, the consolidation of the decentralization project, together with the fact that regional governments have aimed to imprint their own political profiles, have resulted in different translations of global reform ideas and, in some cases, a marked policy divergence. To show this, this article pays particular attention to recent changes in the educational governance (in areas such as school choice, accountability, and school autonomy) of two Spanish regions, Madrid and Catalonia, as they have gone through differentiated processes of exogenous and endogenous privatization reform.
The article is structured as follows. First, it presents key concepts of global education reform and its re-contextualization. Here, we focus on global trends regarding endogenous and exogenous forms of educational privatization and on the political singularities of these policies' recontextualization in highly decentralized educational systems. Second, the process of decentralization of education in Spain, and the current distribution of education responsibilities between central and regional governments, are summarized. The specificities of the Spanish quasifederal architecture are presented to understand the distribution of education policy responsibilities and the power relations between the central State and the regions. Third, the article reviews the historical conditions of the public-private divide in the Spanish education system and provides data to illustrate the processes of both exogenous and endogenous privatization experienced in this country's education. Fourth, we analyze recent educational reforms in Madrid and Catalonia and show how these reforms have resulted in remarkably different policy trajectories and policy outcomes. Finally, the article concludes by comparing the education reform trajectories in the two regions and by reflecting on ideas for future research on education policies in multi-level education systems.

Global Trends in Educational Reform: The Case of Exogenous and Endogenous Education Privatization
In recent decades, numerous education policy initiatives have focused on strengthening the presence of private sector actors and ideas in education. Countries from the Global North and the Global South have gone through reforms that promote the involvement of private actors in educational provision and give more centrality to market mechanisms in education. Nonetheless, far from a concrete policy agenda, education privatization crystallizes in a wide range of policy options. Ball and Youdell (2008) distinguish between two broad types of educational privatization. The first type is exogenous privatization, which involves the participation of private actors in the education sector, with the educational provision of schooling by non-state actors being its most common materialization. In the last 30 years, both the number of private providers and the level of private enrollment in primary and secondary education have increased significantly in almost all regions of the world (United Nations Educational, Scientific and Cultural Organization [UNESCO], 2021), and countries have engaged with exogenous forms of privatization through diverse policies (e.g., vouchers, charter schools, contracted schools, Low Fee Private Schools (LFPSs; Zancajo et al., 2021).
The second type of education privatization identified by Ball and Youdell (2008) is endogenous privatization, which can be defined as importing private-sector ideas, logics, and practices into the public sector. In recent decades, numerous countries have adopted NPM principles in the process of reforming public education (Gunter et al, 2016). The main assumption behind these reforms is that the private sector's management styles and techniques can improve the public sector's efficiency and responsiveness. As in the case of exogenous privatization, endogenous privatization has advanced through a diversity of policies, with school autonomy, market-driven competition between schools, and performance-based accountability policies at the cornerstone of the NPM agenda in the educational sector (Gunter et al., 2016).
Albeit the different focus and nature of exogenous and endogenous privatization, these two forms tend to be interconnected. Exogenous privatization frequently leads to the increase of market dynamics in education systems and tends to be accompanied by the adoption of school autonomy and performance-based accountability policies, which apply to both the private and public schooling sectors. These policies contribute to making public schools emulate private sector behaviors and management styles, as well as to become more performance-and competition-oriented. Thus, both exogenous and endogenous privatization have become increasingly entangled in the global education agenda. Policy programs that have gained international status, such as public-private partnerships (PPPs) and NPM in education-which to a great extent has promoted school autonomy with accountability policies (Verger et al., 2019)-have contributed to the reconfiguration of educational governance worldwide.

Scalar Tensions and Educational Reforms in Decentralized Systems
Despite their global reach and status, exogenous and endogenous privatization reforms have not followed homogeneous patterns. As pointed out by prominent scholars on policy movement, the global widespread of policies or the emergence of widely-shared reform paradigms does not mean a simple emulation and replication of policy solutions, but a transformative process in which policies "mutate and morph during their journeys" (Peck & Theodore, 2010, p. 170). Contextual mediations play a key role in understanding these divergences. Political architectures, administrative traditions, and other national institutions provide different conditions of possibility for education reform. Federalism, in its different forms, is also a key factor to understand policy divergence both between and within nation-states (see Savage & O'Connor, 2015, for a comparison between Australia and the United States on curriculum reform agendas). Federal (or quasi-federal) institutional frameworks are outstanding mediating factors shaping how global influences are interpreted and translated differently in national and subnational education systems; but these frameworks also illustrate how political scales structure and are structured by given power relations and social dynamics, which in turn modulate global policies and reform agendas. Recent studies have addressed the external influences of global education policy agendas, considering the mediating role of federal architectures in their divergent effects within national settings. For instance, Lingard and Lewis (2017) analyzed how the global influence of the Programme for International Student Assessment (PISA) generated distinct effects because of diverging political structures due to dissimilar educational federalisms in Australia and the United States.
Comparative analyses in the context of federal or highly decentralized education systems allow for controlling many institutional variables and better isolating the effects of other types of drivers (Carnoy, 2015). Indeed, in decentralized policy regimes, the different research units share a similar political architecture, administrative culture, and usually main socioeconomic characteristics-for instance, in our particular case, we are comparing two Spanish regions with a similar level of economic development. By exercising control over these institutional and economic aspects, we can pay better attention to the role of other factors, such as political ones, in triggering differential policy outcomes. Overall, quasi-federal settings appear to be a unique scenario to study the uneven diffusion of policy ideas, solutions, and institutional templates and how these patterns generate different layered cultural and policy complexities (see Wimmer, 2021).
Nonetheless, conducting research in federal systems also adds some complexity. It means considering that multiscale interactions intensify in educational reform, since additional scales of governance-which are mutually embedded by definition-intervene in adopting and enacting new policy instruments (Robertson, 2012). This implies looking at how scalar policy-making practices are produced by different political logics and to what extent certain political actors aim to use (or even produce) scalar tensions to their own advantage (Papanastasiou, 2017). Therefore, we should understand the concept of scale as not a mere analytical category or vertical level to look at but as a constituent part of the policy process. Indeed, the production and development of social spaces where policy actors intervene, and the policy process takes place, are at the very core of policy. Hence, the practice of scalecraft, understood as the mobilization and construction of scales, should be considered a key hegemonic strategy of policymaking (Papanastasiou, 2019), particularly relevant in (quasi-)federal systems. Scale-jumping is another matter of interest in decentralized education systems, where policy actors might privilege policy activism in some scales over others to achieve their goals. Therefore, for some policy actors, particular scales can be instrumentally or axiologically more adequate to exercise power and make policy in a particular way (Jessop, 2009), often contrasting with other actors in different scales and policy spaces.
Accordingly, quasi-federal systems are privileged policy environments to study the selective use of scales and their inner tensions to advance policy agendas. Competition dynamics can be exacerbated in decentralized contexts, especially in those where political and cultural identities penetrate multi-scalar interaction. This is indeed the case in Spain, where identity politics and territorial conflict are behind a long history of political contention, contestation, and construction of policy boundaries (Gallego et al., 2017). These types of tensions permeate and give a new meaning to the politics and economics of educational reform. Thus, despite previous research on federal education policy seeing federalism as a political organization that is prone to policy transfer, borrowing, and lending (Kerber & Eckardt, 2007), or even to new forms of more effective central steering (see Savage, 2016), in other countries, it is rather conducive to dynamics of political differentiation, competition, and decoupling between different administrative units (Swenden et al., 2006). Our analysis aims to contribute to this debate by exploring how education privatization reforms of a different nature (namely, endogenous and exogenous) have penetrated Spain with a particular focus on historical, political, and administrative dimensions.

The Historical State Inhibition in Education in Spain
Historically, the Spanish education system has been characterized by the inhibition of the State in the educational field. With few exceptions, the public authorities delegated the task of schooling to the Catholic Church and devoted few budgetary and regulatory efforts to administering the education system. Before the restoration of democracy in the late 1970s, this double system of public and private schools was also the main basis for the inequality of educational opportunities and results. Social and economic reproduction was closely related to participation in one sector or the other. Private schools (mostly religious) provided good quality education for the middle and high classes who could afford school fees. Students from low and middle-low classes used to attend State schools, where they were subject to a cultural transmission model based on ideological control rather than instrumental knowledge (Lerena, 1986). The first democratic government after the Franco regime inherited an education system characterized by great shortages (in 1976, there were more than one million children without a compulsory school place), institutionalized inequality of educational opportunities, old-fashioned pedagogic methods, deprofessionalized teachers with low salaries and poor training, and a bureaucratic and highly centralized educational administration. Thus, the challenge to transform the educational system was formidable.
The 1978 Constitution established the minimum necessary consensus for developing a democratic education system and set up the main goals of educational policy. The references to education in the Constitution illustrate the difficult negotiation process between left-wing and conservative political parties. Both groups had to renounce, at least, some of their principles to achieve an agreement. Thus, if left-wing parties accepted a significant presence of publicly financed private education, parents' right to choose religious education for their children, and a significant level of parental school choice, the conservatives had to accept some public control over the subsidized private sector: the non-compulsory character of religion in the curriculum, teachers' academic freedom, and the participation of the educational community in school decision-making. This ambivalence has been permanently present in the evolution of the Spanish education system, with opposite interpretations of the constitutional mandate between left-wing and right-wing parties and regular changes in the framework legislation every time power changed hands (Bonal, 2000).
The second permanent tension in the Spanish education system has been between centralization and decentralization forces. From the restoration of democracy, the historical regions in Spain (Catalonia, Basque Country, and Galicia) disputed the capacity to build their own education system with the central government. Economic, cultural, and linguistic factors formed the basis of the aspirations of these regions. Other Spanish regions reacted to the demands of the historical regions by requesting equal treatment and conditions to develop their political autonomy. Today, Spain has become a notably decentralized education system, with 17 regions (autonomous communities) holding a significant role in developing their own educational policies, especially with regard to the administration, management, and governance of the educational system (Calero & Bonal, 1999).

The Decentralization of the Spanish Education System
Spain is considered by some authors as one of the countries in Europe with the highest degree of decentralization in education at the regional level (Borrett et al., 2021). The central State retains a notably high legislative power as the ordering authority of educational policy and guarantor of the right to education, while the autonomous communities (or the regions, as we call them in this paper) are attributed the power to manage and organize the educational system (Puelles, 2002). Despite some authors arguing that the model of political decentralization in Spain can be considered as a federation in all but name (Romero Caro, 2022), other authors consider that the Spanish State enjoys a high level of decentralization but a much lower level of what is seen as federalism (Requejo, 2017). The absence of constituent units, the capacity of the central State to maintain its hegemony by setting state-level organic laws, the lack of influence of regional governments on the structure of the judiciary, and the practical absence of any form of fiscal federalism detach Spain from what should be understood as federations and federalism. The central power of the Spanish State reduces the legislative or executive capacity of Spanish regions in education, as compared to countries such as Germany or the United States (Sáenz Royo, 2021). However, the political decentralization of the Spanish State goes beyond a simple transfer of administrative responsibility. The high level of financial capacity and responsibility of Spanish regions (which manage more than 80% of the overall education budget) and the political prerogatives in most educational policy domains mean that Spanish political decentralization has some characteristics of federal systems. Nevertheless, the model is far from "pure" federalism as the regulatory powers of Spanish regions cannot contradict the basic legislation of the Spanish State.
The education decentralization process began in 1978 but did not end until the year 2000 (see Table 1). Powers were transferred at different speeds to regions by virtue of the diverse routes of access to autonomy provided for in the Constitution. The so-called historical regions, which had already enjoyed some autonomy before the dictatorship, could acquire educational policy responsibility through a "fast track." For the other regions, different mechanisms were devised that, although at different speeds, led to the same final scenario of political capacity. Thus, today Spain comprises autonomous communities that have enjoyed full powers for more than 40 years and others that have only accessed them after the year 2000. The experience and the possibilities of having built their own educational model were, therefore, aspects notably more developed in the former than in the latter (Bonal et al., 2005). The State powers in education, as defined in the Spanish Constitution, have been deployed in different organic laws. The education responsibilities of the Spanish State can be divided into two main functions: (1) those that correspond to it as the ultimate guarantor of the right to education, and (2) those assigned to it as the highest authority in the educational system. The State must ensure equality in the exercise of the right to education and, as such, develop mechanisms to compensate for inequalities. Likewise, it is responsible for establishing the legislative framework for the educational system, designing "the educational levels, as well as the modalities, stages, cycles and specialties of teaching," and setting "the minimum content to structure the national basic curriculum" (Puelles, 2002, p. 163) to be developed by the Spanish regions.
Spanish regions are responsible for the administration of the educational system. The Spanish Constitution gives them the capacity to develop all the norms set by the State and develop other legal and executive measures, including the capacity to create or suppress schools, make decisions on education funding, define teachers' contracts and careers, contract out services, or decide on a significant proportion of the curriculum, to name a few. Table 2 synthesizes the responsibilities of both the central government and the regions.
This division of responsibilities applies almost equally to all Spanish regions, although historical regions with their own language can define a higher proportion of the school curriculum. In all the other domains, all the Spanish regions have the same formal capacity to develop their own policy strategies. This homogeneous and rigid normative framework leaves to the will of the regions the possibility to develop their own policies. The use of this capacity has depended on two main factors. On the one hand, their "desire to be" (Embid, 1999, p. 52); that is, their political determination to build their own education system. On the other, the political and economic conditions of each territory have also guided the different possibilities within each region. Aspects such as the budgetary capacity, the higher or lesser presence of private education, and the political orientation of regional governments explain the different orientations that educational policies can take.

Central Government Autonomous Communities
-Regulation of the basic conditions to guarantee the equality of all citizens to realize the right to education

Exogenous and Endogenous Education Privatization in the Spanish Education System
Spain is one of the Organisation for Economic Co-Operation and Development (OECD) countries with higher enrollment rates in the private sector, particularly in private subsidized schools. Although private education has a long-standing tradition in Spain, the PPP model was institutionalized at the end of the seventies and the beginning of the eighties by the first democratic governments after the military dictatorship. The negotiation of the 1978 Constitution led to the school pact (Olmedo, 2008) between left-wing and right-wing parties that recognized the need to guarantee the principle of equality but also the freedom of instruction (Bonal, 1998). Based on this legal ambiguity, the first democratic governments (1977)(1978)(1979)(1980)(1981)(1982) ensured the continuity of the public funding of private schools, mainly managed by the Catholic Church, and pro-school choice policies (Bonal, 2000). However, it was not until 1985 that, in the context of increasing educational demand and budgetary restrictions, the social-democratic government elected in 1982 passed the Right to Education Law (Ley Orgánica del Derecho a la Educación [LODE]), which established the main principles of current regulation of public subsidies for private schools. This law also included the requirement of private subsidized schools to follow the national curriculum, not charge fees to families (though, in fact, private schools kept charging families for different concepts) and be nonprofit and follow the same rules as public schools in school admission policies (Calero & Bonal, 1999).
Despite there being a common framework regulating public subsidies for private schools since the eighties, the expansion and prevalence of private subsidized schools in primary and secondary education vary significantly among the different autonomous communities (Figure 1). Whilst in some regions, such as the Basque Country, Madrid, or Navarre, private subsidized schools represent a significant proportion of the total enrollment in primary and secondary education; in other regions, they only account for around 20%. These differences can be explained by factors such as the predominant ideology and policy preferences of regional governments, the level of economic development, demographic and urban trends, and the historical role of private education in the different regions.
In Spain, endogenous privatization trends are much more recent than in other OECD countries. In 2013, the conservative party, in office between 2011 and 2018, passed a new Education Reform Act (ERA), which involved a significant step forward regarding both exogenous and endogenous privatization policies, inspired by the tenets of NPM, at the State level. 2 Among other measures, this law established the obligation of the State to subsidize private schools if there is demand for it, opened the possibility of publishing the results of schools in large-scale assessments, and promoted a managerial approach to school autonomy through the professionalization of school leadership (Parcerisa, 2016;Verger et al., 2016).
Although exogenous and endogenous forms of privatization have been increasingly acknowledged in the Spanish regulatory context during the second decade of the 21st century, these policies have not penetrated all autonomous communities with the same force. Indeed, the existing level of administrative decentralization of education explains that the development of these policies differs substantially among Spanish regions (Olmedo, 2013). Even the regions that have been more active in managerial education reform have had varying levels of success in that venture.
To some extent, this explains that the levels of school autonomy and the presence of accountability policies in schools are significantly different among Spanish regions. According to principals' responses in PISA-2015 (Table 3), school autonomy, both curricular and managerial, is substantially higher in regions such as Madrid, Galicia, or Catalonia. Similar differences can be observed regarding the role and use of external performance assessments. The percentage of school principals who affirm using external assessment for comparative purposes is much higher in Madrid, the Balearic Islands, and Catalonia than in other regions. As we show next, the comparative analysis of contemporary educational reform in Madrid and Catalonia contributes to understanding these differences better within the Spanish context.

Primary education Secondary education
Source: Spanish Ministry of Education statistics. Source: Own elaboration based on PISA-2015 data.
a Based on the PISA school principal questionnaire, this index measures the capacity of the school to determine the curriculum (e.g., course content, courses offered) and student assessment policies. The index is standardized to having an OEDC mean of 0 and a standard deviation of 1. Higher values indicate a high level of school responsibility (OECD, 2017). b This index measures the responsibility of school staff in allocating resources such as hiring/firing teachers, determining salaries, and formulating and deciding budget allocations. The index is standardized to having an OEDC mean of 0 and a standard deviation of 1. Higher values indicate a high level of school responsibility (OECD, 2017). c This index is calculated based on the percentage of tasks related to the curriculum and resources of the school (i.e., principals, teachers, or school governing board). The index takes values from 0 to 1, which higher values indicating higher levels of autonomy (OECD, 2016).

Madrid: A Market Competition Approach to Education Policy
The region of Madrid is one of the most urbanized and developed regions in Spain, with the highest regional gross domestic product (GDP; Instituto Nacional de Estadística, 2019a). It is also the second region in terms of households' income level (Instituto Nacional de Estadística, 2019b) and the most densely populated. Madrid is an outstanding economic hub as well as the core of the political and administrative power in Spain. In contrast to other international capitals, especially in Western Europe, the capital of Spain is politically more conservative than the mean of the country due to historical, economic, and political reasons (Muñoz, 2021). Indeed, the conservative party has been in office uninterruptedly in the regional government of Madrid since 1995.
The region received educational responsibilities via the constitutional "slow path" in 1999, and thereafter, the education policy-making in the region has always been ruled by a conservative approach with important liberal inspirations. Interestingly, the region has developed a particular policy architecture with the implementation of minor norms, rules, and decrees, through which the conservative government has consolidated a market-oriented education system. The education policy approach in Madrid is openly conceived by the regional government as a counterbalance to the federal education reforms, which, according to the conservative perspective, are seen as an obstacle to excellence and merit due to their equity orientation. Recently, Madrid's regional government passed its first own regional educational law, the so-called Master Law of Educational Freedom, which represents a prolongation of the conservative reform agenda of recent decades and is framed as an opposed model to the national regulatory approach (Prieto, 2022). Accordingly, the regional reform is aimed "to guarantee the pillars of the regional education, which are being threatened by the implementation of the (last ERA) [adopted by the left-leaning Spanish central government in 2021]" (Consejería de Educación, 2022). To better understand what such "pillars of regional education" are, we review hereunder different dimensions of education policy in Madrid.
A first element that makes Madrid a particular case in terms of educational provision is the high presence of private education. Due to historical dynamics, the role of private religious schools in Madrid is salient. In this case, it is noteworthy that beyond historical precedents-configured by path dependency dynamics-the regional government adopted an active policy initiative to expand and fund traditional private schools, but also an emerging private sector with a more modern, business-oriented, and secular orientation. Currently, in Madrid, 54.6% of the students in primary and secondary education are enrolled in public schools, 29.5% attend private subsidized schools, while the private independent sector represents 15.9% of total enrollment (Consejería de Educación, Universidades, Ciencia y Portavocía, 2021). Overall, what makes Madrid a distinctive case compared to other Spanish regions is a notable percentage of private-subsidized schools and a relatively high proportion of students who attend fully private schools, which in other Spanish regions are almost marginal.
Similar dynamics can be identified with indicators of educational funding. Madrid is the Spanish region with the lowest levels of public funding per pupil in public schools, which contrasts with the fact that Madrid has the highest level of private household spending on education in Spain. More specifically, the public expenditure per pupil in public schools in Madrid is €4,727, the lowest in the country, significantly below the Spanish average (€5,779) and representing half of the expenditure per pupil in the Basque Country, the region with the highest investment (€9,298; Ministerio de Educación y Formación Profesional, 2020). In contrast, in Madrid, the average household expenditure on education is €812, almost duplicating the national mean (€437; Instituto Nacional de Estadística, 2020). In parallel, the private subsidized sector in Madrid experienced a significant increase in the public funding they receive during the last 2 decades (Ministerio de Educación y Formación Profesional, 2020b).
The advancement of exogenous privatization in Madrid has also been accompanied by education policies promoting school choice, autonomy, and accountability. An important wave of school reforms started at the beginning of the 2000s with the introduction of different programs of educational specialization, including sports, technology, and English. Among diverse initiatives, the Spanish-English Bilingual Program (BP) stands out for its extent and impact on the Madrilenian education system. The BP started in the academic year 2004-2005, aiming at improving English skills with the instruction of part of the school curriculum in English. Primary schools could ask to join the program adopting two different modalities, according to the percentage of the curriculum instructed in English. The program was initially implemented in public schools and quickly generated great interest among families and schools. In 2008, the publicly funded private schools adopted an analogous program with the support of the regional educational authorities, and the program was escalated to lower and upper secondary education in the following years (Consejería de Educación e Investigación, 2018). Currently, half of the students enrolled in primary public schools are part of the BP, while in the case of private publicly funded schools, this percentage rises to 59.7% (Consejería de Educación, Universidades, Ciencia y Portavocía, 2021). However, this reform model appears to be very limited in promoting school autonomy as schools simply joined educational programs predefined by educational authorities. In fact, some authors have stressed that this model provided schools with an "illusory freedom while increasing the control of the education system" (Prieto & Villamor, 2018, p. 23). It can thus be argued that the goals behind this policy approach were more related to the promotion of educational supply diversification rather than the enhancement of genuine professional and pedagogical autonomy. Paradoxically, such diversification neither took place as most schools joined the BP, interpreting it as a sign of quality or a market advantage in a context of competition. As already highlighted, almost half of the schools in Madrid have joined the BP; however, the schools adopting the program are not equally distributed within the educational system. Consequently, the BP operated as a mediator of school segmentation, contributing to worsening segregation dynamics in an already highly polarized education system (Gortazar & Taberner, 2020).
Together with the adoption of these programs of curricular specialization, the regional administration implemented a standardized test for accountability purposes. The test aimed to improve schools' results and enhance transparency by means of informing and orienting parental school choice. With this purpose, the educational authorities developed a digital school browser compiling diverse school information, including standardized test results. The regional government selectively used some OECD recommendations to legitimate and justify the dissemination of the school results. Although the test results were initially disseminated with public rankings, the lack of social and political consensus, and the discontinuity of political leadership prevented the retention of this policy. In this context, intermediate officials gained more centrality within the Regional Department of Education, who opted for pragmatic options, desisting from publishing the results of the test in order to align the test design and uses with the requirements established by the last federal reform . Nonetheless, although the test results are not formally disseminated to promote school choice anymore, some schools still consider testing as a source of performative pressure in local spaces of school competition (Pagès, 2021). In fact, as can be observed in Table 3, Madrid is one of the Spanish regions where schools more frequently use achievement data from the standardized test for different purposes, including the comparison of the school results with other schools (60%) and with the district or the national average (68%).
Finally, in parallel to such policy reforms, the regional government of Madrid promoted different norms and regulations aiming at expanding parental school choice. The most relevant change took place with the implementation of a school choice decree passed in 2013, which defined a single school-choice area at the city level in the whole region, contravening the federal regulation, generating a recent response from the central government, and requiring the compliance of basic state-level regulations on student admission policies (Sanmartín, 2022).Taken together, these different policy initiatives have contributed to increasing external pressures on schools and dynamics of school competition (Prieto & Villamor, 2012). Indeed, the regional education department boasts itself of being "the Spanish region with greater school choice" and stresses that, according to PISA, "85% of the Madrilenian schools compete with two or more schools in their environment" (Comunidad de Madrid, 2022). 3 In short, despite its apparent non-structural nature, the education reform process in Madrid has been continuous and intense in recent decades. The sum of different decrees and budgetary decisions, and the recent passing of the regional education law, promoted a market approach in education, the expansion of school choice, private and private subsidized schools, and performancebased accountability, converting Madrid into the Spanish region with the most marketized educational system.

Catalonia: Educational Reform à la NPM
Catalonia is one of the most industrialized and economically dynamic regions in Spain, with a GDP per capita much higher than the Spanish regional average. With its own language and longstanding political institutions, it is one of the so-called historical regions of Spain. Accordingly, it received educational responsibilities much earlier than other regions, in 1981.
The 1980s and the 1990s were decades of political continuity in Catalonia. The regional government was in the hands of the Catalan nationalistic conservative party uninterruptedly for 23 years. The most emblematic educational policy of this period consisted of promoting social cohesion in education through the adoption of Catalan as the main language of instruction. Of special relevance was the law on linguistic normalization approved in 1983, which, among other measures, established Catalan as the language of instruction to avoid the linguistic and social separation of students with different home languages. While the implementation of this linguistic policy has not been socially conflictive, there has been a gradual politicization of the school language (Garvía & Miley, 2013), which has become an arena of ideological struggle between Spanish and Catalan parties. Beyond language issues, this was also a period in which the government had a very liberal approach to PPPs and, accordingly, even elite schools benefited from public funding.
The change of the millennium inaugurated an era of power shift and political turmoil in Catalonia. In 2003, a broad progressive coalition (including social democrats, greens, and left-wing nationalists) took over and was in power for two parliamentary terms. In its second term, this coalition passed the first Catalan Education Act (Lei d'Educació de Catalunya, LEC) in 2009. The process for the approval of this act benefited from previous multi-stakeholder debates that had been in place since the beginning of the 2000s in Catalonia and which ended with the approval of a socalled National Agreement on Education in 2006. This agreement-which was signed by numerous teachers' unions, families' associations, pedagogical organizations, and private sector foundationsplaced a lot of emphasis on the importance of promoting school autonomy and an evaluation culture as a way to modernize the Catalan educational system and make it more equitable (Wilkins et al., in press).
The LEC expanded on the ideas of autonomy and evaluation, giving them a new managerial perspective. The Catalan education ministry at that time openly embraced the main principles of NPM and put much emphasis on promoting results-oriented management and on placing public regulations at the necessary minimum (Maragall, 2009). The central articles of the LEC would focus on strengthening the governance of the educational system through the promotion of schools as more autonomous managerial units, the professionalization of school leadership, and the strengthening of the evaluation system. Although the LEC got important input from domestic debates and stakeholders, it was also inspired by OECD ideas on school governance. In fact, it benefited from the direct technical advice of OECD staff during the reform process. That was the era in which PISA results had the biggest media impact, and accordingly, many of the decisions taken had the improvement of PISA results in Catalonia as a core benchmark (Verger & Curran, 2014).
Despite the relationship between the Spanish and Catalan institutions being often conflictual, the LEC was never conceived as antagonistic to the policy approach that predominates in the Spanish legal context. The Catalan education law, in fact, acknowledges the benefits of the structural reforms adopted at the Spanish level since the 1990s-in terms of the expansion and democratization of education and the promotion of constructivist pedagogies-and it takes it from there (Verger & Curran, 2014). The LEC pays so much attention to school governance and evaluation elements to complement and develop the dispositions included in the Spanish legal framework, not to offer a policy alternative.
With the LEC, the census-based standardized test acquired centrality in educational evaluation and school improvement policies. The main goal of the test was to measure the basic skills of students in order to improve and orient instruction and policy (Departament d'Ensenyament, 2009). The test has become, de facto, an instrument to measure schools' performance but without generalized consequences for teachers and schools. The results have never been published by public administrations, although they have been filtered to the media eventually. Also, some schools, especially private subsidized schools, post their results on their websites as a marketing strategy-a practice that, however, the government discourages. The LEC foresees the creation of an independent evaluation agency that would be able to evaluate a broad range of dimensions of the educational system, many of which had not been systematically evaluated yet, including teachers' performance (Bonal & Verger, 2013;Collet, 2017).
Thirty-four percent of students in basic education are enrolled in publicly subsidized private schools-a percentage even bigger in urban areas. (For instance, in Barcelona, more than 50% of the enrollment is in private subsidized schools.) The position of the LEC regarding the double provision network (public and private) is to favor the equivalence of the two networks and to advance toward a more genuine type of PPP (Zancajo et al., 2022). In a way, private schools that are part of the PPP are expected to be more inclusive and contribute to the equity goals defined by the government. For its part, public schools are expected to become more independent and adopt managerial techniques and logics from the private sector. Overall, in contrast to what we have seen in the case of Madrid, recent Catalan education reforms have focused more on endogenous than on exogenous privatization.
Although the LEC was not to everyone's taste, 4 it was not apparently biased toward a specific political ideology. Some qualify the Catalan act as a hinge law because it accommodates concerns and preferences from different educational ideologies. The act is so broad and diverse that it allows governments to develop part of the articles of the law while ignoring others. This was exactly what the government that took over in 2010 did. That year, the Catalan conservative party regained power after winning the election by a majority. This government discouraged the adoption of some of the LEC initiatives, including the creation of the independent Education Evaluation Agency or the distribution of disadvantaged students across the public and private subsidized school networks. It also advanced those policies that required less budgetary effort and were more ideologically sound-such as strengthening the figure of schools' principals by giving them more power in the selection of the teaching staff (Generalitat de Catalunya, 2014). This government applied severe budget cuts in education at the dawn of the global financial crisis. The main policy goal in that period was the achievement of educational success. For this purpose, the government promoted common core curricular standards in knowledge areas such as mathematics and literacy. The Inspection Services acquired new areas of responsibility and became more present, especially in underperforming schools (usually public schools serving socially disadvantaged populations). They adopted new school-evaluation programs aimed at improving underperforming primary schools, promoting an evaluation culture within schools through both the external and self-evaluation of schools' effectiveness, and attempting to introduce merit-based policies for teachers (Departament d'Ensenyament, 2014;Verger et al., 2020). The conservative government also tried to give more school choice opportunities and benefited families to choose for their children the school they attended-a policy clearly discriminating against the immigrant population (Bonal & Verger, 2013)-although these measures had a short run.
More recently, public education policy in Catalonia has evolved from focusing on school governance and NPM toward an approach that places more emphasis on discussing pedagogy, instructional improvement, and educational innovation. Innovative practices in the Catalan education system have been strongly promoted through a bottom-up process-specifically, by a civil society campaign backed by philanthropic and international organizations in the context of the financial crisis of the early 2010s (Torrent & Feu, 2020). The campaign-which made continuous references to UNESCO and OECD reports and programs to justify its goals and raison d'etre-was very successful in terms of school reach and media impact. The objectives of the campaign were quickly absorbed, and to some extent co-opted, by the regional authorities as a flagship educational policy in 2019. To understand this shift, it is important to mention that the education department of the region has been in the hands of a left-wing party, the left-nationalists, since 2018. 5 To date, "educational innovation" has become a catch-all policy that allows the Catalan government to intervene in schools and promote school improvement dynamics through a broadly engaging programmatic idea-even units of the education department have been named after it, and the "school improvement plans" are now called "innovation plans." Nonetheless, the most recent emphasis on promoting educational innovation as a public policy does not necessarily imply a rupture with the NPM approach inaugurated with the LEC.
Instead, the emphasis on innovation has worked as a pedagogic device to recontextualize core LEC policies (such as school-based management, management by objectives, assessment data use to inform instructional improvement, supervised teacher autonomy, and distributed leadership) in a reform program that is more normatively acceptable and professionally appealing to some factions of the teaching community (Verger et al., in press). Finally, the current government has also adopted new measures in an attempt to regulate the private school offer from an equity perspective-such as obliging publicly-subsidized private schools to enroll a bigger percentage of vulnerable students.

Conclusions and Discussion
In highly decentralized educational systems, different actors, scales, and administrative units take part in policy adoption and re-contextualization processes. The actors and the interests formed at the regional level, far from totally autonomous, are embedded within national and supra-national policy networks and legal frameworks. In this paper, we have analyzed how exogenous and endogenous privatization processes have materialized in Spain and, in particular, in two of its most populated regions. We have analyzed general trends in Spanish education policy and concurrent elements of tension between education policy and politics. Two cleavages emerge as the main elements of dispute. First, the dynamics of centralization and decentralization of educational policy responsibility have generated tensions between different scales of governance. Second, the ideological orientation of education reform with contrasting approaches between progressive and conservative positions. The way these two cleavages combine and how the related tensions are managed are key to understanding varieties of policy trajectories at the regional level.
Although the Spanish government keeps important political control over educational regulation in core areas, the regional scale has gained importance as a space for policy differentiation. The unclear frontiers regarding the division of powers between the central and the regional governments set by the Spanish Constitution have favored the emergence of particular forms of educational decentralization and habitual tensions between the most active regional governments and the Spanish State. As our analysis shows, Madrid and Catalonia have actively recontextualized supra-national and national educational policy frameworks with markedly different emphases. The two regions actively adapted policies of school choice, autonomy, and accountability to consolidate particular vernacular policy models with different approaches (Maroy et al., 2017). By doing so, they have been able to give different policy emphases to the governance of their education system in relation to the central-State approach and that of other Spanish regions. They have done so for both ideological and territorial motivations-with Madrid putting more emphasis on exogenous forms of privatization and Catalonia on endogenous privatization.
In Madrid, the differentiation mechanism is driven by a markedly ideological rationale, building a market-oriented and "meritocratic" education model in opposition to the state-level education reform approach, which is perceived as more equity-and public service-oriented. Although Madrid acquired the educational responsibilities relatively late, it has been eager to promote a singular policy approach in which pro-exogenous privatization and school choice policies are salient, despite coexisting with forms of endogenous privatization too. Education policy in Madrid is a political instrument through which the conservative government aims to challenge the Spanish social-democratic government, as well as what it sees as an education policy approach that is biased in favor of equity at the expense of excellence. In Catalonia, the differentiation mechanism is driven by a combination of political preferences and a discursive singularization of the Catalan education system, even though it is highly compatible with the policy approach that predominates at the Spanish State level. In the Catalan case, a more ambivalent and less ideologically charged process took place, gradually introducing forms of endogenous privatization, first with explicit NPM inspirations and, more recently, with more subtle strategies to enhance reform pressure, particularly using educational innovation as a discursive tool to trigger school-level processes of educational change and improvement.
Based on the analysis conducted, we suggest that the regional scale is not just another layer in a policy transmission chain subordinated to global and national education policy prerogatives, but rather an assembly with its own agency. In the Spanish case, this regional agency has been characterized by a certain regulatory ambiguity and by the willingness of specific regional governments to construct their own education model and, in doing so, reinforce the regional level as a privileged space of policy-making. Indeed, when political will is in place, the regional space shapes policy-making determinedly. In highly decentralized States, intermediate political and administrative power units might be interested in adopting their own reform approaches to join a global policy discussion, but also to internally develop and legitimize their own political agendas and gain political traction with their constituencies.
Apparently, the federal scale is less active in pushing governance and administrative educational reforms than the regional scale. However, both the Spanish and the regional governments' power overlap in numerous policy domains, often in contradictory ways. As a general rule, the federal space grounds the basis for education policy making, delineating a general framework where education policy takes place. Within this administrative and political frame set by the Spanish State, the politics of policy interpretation and translation happen at the regional scale.

Xavier Bonal
Universitat Autònoma de Barcelona xavier.bonal@uab.cat Xavier Bonal is a professor of sociology at the Autonomous University of Barcelona [Universitat Autònoma de Barcelona, or UAB] and director of the research group Globalisation, Education and Social Policies (GEPS). He has worked as a consultant for international organizations such as the