Increasing school autonomy in Western Europe: a comparative analysis of its causes and forms Open Access

European education systems have traditionally been regulated by the state. More recently though, the state's monopoly over education has softened and new governance forms have emerged, which generally entail a downward shift of authority to individual schools. Despite an increased pace of change both in western OECD countries and globally (Mons 2004; Jakobi and Teltemann 2009), increasing school autonomy remains under-researched in comparative educational and social science research. This article aims to advance the debate by pursuing a theoretical and empirical conceptualization of school autonomy, its causes, and forms.¹

Our analytical approach is divided into two steps. First, we explore the catalysts behind greater school autonomy in Western Europe by focusing on two forms of influence, which may fuel changes in educational governance: (1) internationalization processes and (2) shifts in socio-political values. Moving beyond previous research highlighting the spillover of welfare state and New Public Management (NPM) reforms to education (Klitgaard 2007, 2008; Verger and Curran 2014), we argue that the spread of post-materialism (Inglehart 1971, 1977) as well as transnational diffusion processes may be crucial variables in explaining educational reforms. In a second step, we examine context-specific variations in national reform trajectories. Have different school autonomy models emerged and, if so, what explains the varying pathways? We assume that partisan differences can be decisive. While both center-left and center-right parties have aimed to transform educational governance, we postulate that parties pursue different political rationales (Gingrich 2011) and different ‘school autonomy agendas’.

We first offer analytical insights into the concept of school autonomy and embed it in the discussion on the transformation of welfare states, before providing a short discussion of previous research. We then present our theoretical approach, which – like the research questions – is divided in two explanatory components. After briefly discussing the research design, we present our empirical results regarding school autonomy and its driving forces in Sweden, England, Italy and Germany (Hessen). To conclude, we critically reflect on our results and provide an outlook for future research.

The trend toward school autonomy can broadly be seen in the context of overarching shifts in the public sphere, which occurred in varying degrees in Western Europe starting in the early 1980s. Faced with low economic growth, decreased productivity, mass unemployment and ageing populations, western governments came to see the welfare state and public sector as a burden and source of socioeconomic problems (Pierson 1996; Starke 2006). Policies under the label of ‘retrenchment’ aimed to downsize the welfare state and recalibrate public services to meet public demands more effectively. These demands were frequently met with various policies aimed at promoting choice and competition (Klitgaard 2007, 2008).

These developments coincided with discourse on the decreasing capacity of governments to control and regulate various social spheres (Weiler 1990; Rose 2000) and the difficulties of central governments in coordinating multiple stakeholders and local interactions between public and private organizations (Pressman and Wildavsky 1984). This discourse provided legitimacy to new policies aimed at enabling the state to engage in ‘action at a distance’. Increasingly viewed as partners, animators and facilitators of independent agents and powers, central governments promoted policies based on ‘localization’, that is, governing through self-steering forces of trust and obligation (Rose 2000: 223–4).

These trends ran parallel to the rise of NPM (Pollitt and Bouckaert 2011). Seen as a new ‘post-bureaucratic’ policy paradigm (Hood and Peters 2004), NPM was ‘an approach in public administration that employs knowledge and experiences acquired in business management and other disciplines to improve efficiency, effectiveness, and general performance of public services’ (Vigoda 2003: 813). Against the background of economic and fiscal austerity, NPM and welfare state reforms fed into each other and, in some cases, spilled over into education (Klitgaard 2007). Hence, numerous authors have viewed school autonomy in this larger context (Cole and John 2001; Verger and Curran 2014) and highlighted how the state liberates educational providers from central regulations, while simultaneously demanding greater accountability. To this end, the state withdraws from the inner workings of schools and instead focuses on educational ‘outputs’, that is, the skills and qualifications acquired by pupils.

A multi-faceted phenomenon, school autonomy essentially entails the shift of authority to individual schools. This may affect the administration of funds and material resources, personnel management, learning organization, as well as class organization, planning and structural matters (OECD 1995, 1998; Bottani 2000; Rürup 2007). School autonomy policies are often closely related to or go hand in hand with the phenomenon of educational decentralization (Weiler 1990; Hanson 1998; Mons 2004), which has the effect that both concepts are referred to interchangeably or not sufficiently differentiated in the literature. While school autonomy policies chiefly pertain to the increase of self-governance capacities of individual schools, the concept of decentralization generally refers to shifts in territorial competences from the central government to regional, state or local authorities (e.g. the German Länder, Swiss cantons, French départements). Depending on the context, an increase of school autonomy may run parallel to educational decentralization, leading to an increase of authority of both individual schools as well as enhanced decision-making leverage for local authorities (e.g. Sweden, see below). However, school autonomy may still remain low in territorially decentralized systems. This is often the case in federal systems, like Germany or Switzerland, which traditionally represent a certain degree of territorial decentralization, even though their component parts (Länder, cantons) have historically exercised a similar form of top-down governance dynamics over ‘low autonomy schools’ to centralized systems (Weiler 1990: 434). Thus, for the sake of clarity, we opt for the concept of school autonomy for our analysis, as decentralization may be confounded with shifts in territorial competences, which do not result in actual decision-making autonomy for schools (for the French case, see Dobbins 2014).

Against this background, Table 1 exemplifies five dimensions of educational decision-making power identified by Rürup (2007),² which may be transferred both to individual schools or other territorial entities, resulting in diverse context-specific constellations of how these dimensions are re-organized in autonomous school settings and which actors are empowered (see below).

School autonomy is currently a widely discussed topic in educational research. Proponents point to purportedly more efficient decision-making and resource utilization, the diversification of educational offers, local innovations and the stronger democratic character of education (Chubb and Moe 1988; Weiler 1990; Hanson 1998; Karlsen 2000). These factors are in turn expected to improve educational outputs and quality. Opponents fear the duplication of administrative structures, the unwarranted influence of interest groups, personnel recruitment problems and the aggravation of social inequalities (see Meier et al. 2000; Mons 2004).

Numerous scholars have also examined school autonomy policies as well as the often closely related phenomenon of territorial decentralization in education and identify different institutional configurations of education systems (Mons 2004; Altrichter and Maag-Merki 2010; Altrichter and Rürup 2010). One important analytical contribution is Meyer, who elaborates on the multidimensionality of decentralized school governance by pointing to situation-specific and functional variations, while emphasizing that varying forms of interaction of state and local steering mechanisms and new forms of stakeholder governance among school management, local authorities, teachers and parents have developed (Meyer 2009: 468 et seq.).³ Hence, both school autonomy and decentralization add new social dynamics to school governance to the extent that they open new channels of influence to a broader array of local, regional and internal schools stakeholders over these dimensions.

Despite these conceptual advancements, Mons (2004: 47) criticizes the lack of solid theoretical foundations regarding school autonomy policies. More recently, researchers have contributed to overcoming this research deficit by exploring the role of international organizations such as the OECD⁴ in educational governance (Martens et al. 2007, 2010). One important German-language study is Rürup (2007), who analyzes the impact of various political and socioeconomic features of the German Länder on school autonomy policies. Several authors also deal with the impact of increased school autonomy on educational performance (Chubb and Moe 1990; Wößmann 2007). For example, Walberg et al. (2000) determined that students perform better at schools, which have autonomy over personnel decisions.⁵ An analysis of PISA⁶ data conducted by Schlicht-Schmälzle et al. (2011) revealed, however, that educational deregulation – in this case school autonomy and privatization – leads to a trade-off between efficiency and educational equality: while increased school autonomy and privatization result in improved performance of most students, socially privileged groups benefit more, which in turn increases educational inequality.

Despite significant progress, important questions remain unanswered. In view of the above-discussed literature, it is an open question to what extent school autonomy coincides with the broader introduction of NPM or whether other causal mechanisms are at play. Research has also yet to comparatively explore variations in the forms and causes of increased school autonomy. In other words, why have most Western European OECD countries substantially enhanced the autonomy of educational providers? And, importantly, have different school autonomy models emerged and, if so, what circumstances explain diverging pathways?

As noted above, we by no means reject the widely discussed assumption that enhanced school autonomy is the result of spillover effects of broader societal and economic changes, welfare state reforms and NPM. However, we argue that such overarching processes may only account for part of the picture and insufficiently explain different school autonomy pathways.

To advance the state of research, we pursue a dual theoretical approach. First, we draw on research dealing with international policy convergence (Knill 2005). The hypothesis is frequently put forward that national policies are becoming more similar due to increasing international institutional interlinkages. Scholars have pointed to the influence of international organizations and transnational education policy platforms. In particular the OECD has established itself as an important transnational actor with its PISA study. PISA revealed that high school autonomy generally correlates with better school performance (Schlicht-Schmälzle et al. 2011). Along these lines, school autonomy has been promoted as ‘best practice’ by the OECD, in particular since the first PISA study in 2001 (OECD 2011b) and research has shown that national decision-makers have increasingly aligned their education policies with top PISA performers (Martens et al. 2007, 2010).

Thus we assume that governments will align themselves with OECD reform proposals and educational policy forerunners when reforming their educational systems.

However, there is a risk of overestimating international stimuli, in particular in education, which has always been closely attached to the nation-state. Therefore, we alternatively explore a previously neglected potential education reform catalyst – the shift toward post-materialism. Post-materialist theory deals with a cultural shift in the late twentieth century reflected in transformed value systems of modern societies (Inglehart 1997). It assumes that classical materialist values are being replaced by non-utilitarian values such as self-realization, quality of life and authentic political participation.

Inglehart (1997) sees the increasing level of education of the post-war generation as an important cause for this transformation. A higher level and broader offer of education have produced critical citizens, who not only are more interested in new means of self-realization and political participation, but also increasingly critically assess the performance of the state and administrative hierarchies (Inglehart 1997). This results in demands for the democratization of governmental activity and a stronger will to shape political processes.

The downward shift of decision-making authority is one way of meeting demands for more individualized public services. For example, the desire for greater school choice and more influence over everyday school activities could strongly affect the direction of educational reforms. As noted above, demands for more individualized governmental services (Inglehart 1997) in welfare states may be at odds with highly centralized school systems: ‘A uniform and centrally planned public school system that also restricts parental choice is at risk of losing legitimacy and public support’ (Klitgaard 2009: 221).

However, one might also postulate that increasing post-materialism could bring about the exact opposite, namely the rejection of school autonomy policies. As noted above, school autonomy also may be seen as part of overarching changes in the socioeconomic sphere and, in particular, efforts to introduce more market-oriented and utilitarian approaches in the provision of public goods. This is currently reflected in the diffusion of NPM and output-oriented approaches in school governance and – in particular in view of the PISA study (Martens et al. 2010) – a stronger focus on measurable skills applicable in economic contexts. This has coincided with the expansion of standardized testing at the school and national level and greater inter-school competition. These trends in turn may be perceived as undermining classical humanist (i.e. non-utilitarian) values in education such as individual creativity, personal development, self-fulfillment and overall well-being, not least due to the additional stress imposed on students and teachers by the myriad of new output-oriented steering mechanisms. Thus, autonomy policies linked with a utilitarian understanding of education and school management may be seen as undermining classic post-materialist values.

As hinted in Hypothesis 3 and noted above, we are also interested in variations in school autonomy constellations. Along these lines, Altrichter and Rürup (2010) distinguish two governance models, which are not mutually exclusive – the participation model and competition model. The competition model is chiefly concerned with promoting quality, innovation, and enabling schools to strategically position themselves as service providers, which in turn frequently entails greater school choice. The participation model primarily aims to ‘democratize’ school policy-making by incorporating and empowering local stakeholders.⁷

Both bring about shifts of authority between numerous stakeholders, most notably the principle/headmaster, the collective of teachers and/or teachers unions as well as outside stakeholders connected with schools (e.g. parents and local government). Although highly context-specific, the competition model generally strengthens the entrepreneurial capacity of the headmaster/principal. The participation model generally enables new input channels for parents and local government through new participatory bodies. As a result, school autonomy and the territorial decentralization of policy to the local level may go hand in hand. The extent to which teachers organizations assert themselves in governance processes is also highly context-specific and dependent on their organizational capacity and constraints (Dobbins 2014).

Crucial for our analysis is the distinction between and justification for the competition-oriented and participative models, which we trace back to partisan politics. We argue that measures promoted by center-left and center-right parties differ regarding their legitimization and form (Ansell and Gingrich 2003). Specifically, center-right parties pursue a stronger market- and quality-oriented argument in educational governance reforms so that the outcome will resemble the competition model. We argue, by contrast, that center-left parties promote a ‘democratizing’ school autonomy agenda, which highlights citizen participation in the provision of educational services, that is, the participation model.

We pursue a most different cases design (Przeworski and Teune 1970), which enables us to explore cases demonstrating the same phenomenon under different contexts. We peg these different contexts to varying welfare state traditions, as education can be viewed as a central part of welfare states (Jensen 2011). Thus, we select one country from four welfare state traditions, which have enhanced school autonomy. The Swedish welfare state tradition (Esping-Andersen 1990) is manifested in its education model characterized by high public spending and a high share of the population with advanced education. Education plays a crucial role in social integration, as reflected in the comprehensive (i.e. non-stratified) secondary education system (Busemeyer and Nikolai 2010). Generally classified as a ‘conservative welfare state’ (Esping-Andersen 1990), Germany also has low private education spending and high share of the population with at least upper secondary education. However, it remains committed to a stratified secondary school system in which pupils are placed on different academic or vocational tracks early (Busemeyer and Nikolai 2010). As a liberal welfare regime (Esping-Andersen 1990), Great Britain's education system is characterized by moderate public spending and high private spending, while education is seen as a form of ‘preventive’ social insurance against labor market risks (Busemeyer and Nikolai 2010). Italy, by contrast, represents the ‘Mediterranean model’ characterized by both low spending and low shares of the population with upper secondary or tertiary degrees (Busemeyer and Nikolai 2010). Due the federalist structure of German education policy, we limit our analysis to the state (Land) Hessen. Our justification is that – unlike CSU-dominated Bavaria or SPD-dominated North-Rhine Westphalia – Hessen has a long tradition of being governed by both center-left and center-right coalitions. This partisan variation also applies to the other three cases and helps us to isolate the impact of partisan constellations.

The data regarding the dependent variable, school autonomy, are drawn from the PISA studies, OECD policy papers, the EU (Eurydice), and a PISA-based data set of school autonomy policies compiled by Schlicht-Schmälzle et al. (2011). We define ‘post-materialist value shifts’ and ‘international policy diffusion’ as our independent variables. To trace the shift toward post-materialism before the respective reforms, we evaluate national and international representative surveys such as the European Value Study (EVS)/World Value Survey (WVS)⁸ and secondary literature. We also draw on country-specific studies from independent research institutes to gather additional data on shifts in values. We measure international diffusion by examining the influence of the OECD on national education reforms based on primary and secondary literature.

To explore variations in school autonomy policies (dependent variable), we focus on country-specific reforms in secondary education.⁹ The data were acquired from governmental documents and secondary literature. As for ‘partisan differences’, we explore government documents, party programs and secondary literature to discern the respective governing parties’ positions on and legitimization for educational reform. The timeframe of analysis is from 1980 to 2010. This enables us to assess whether the growing impact of the OECD since the first PISA study (2001) has intensified the autonomy trend. To preliminarily assess the validity of Hypothesis 3 (i.e. post-materialist backlash to school autonomy), we also briefly touch on a few very recent educational reforms in the examined countries.

Besides the increasing percentage of decisions taken at school level we also consider school autonomy policies in a broader sense by referring to local or school governance reforms as well as parental choice.

To what extent were these developments driven by international policy diffusion and/or the spread of post-materialism (Hypotheses 1 and 2)? While our analysis of the impact of international education policy actors such as the OECD primarily draws on primary and secondary sources, we derive our data on the post-materialist shift from the EVS/WVS which best reflects classic post-materialist values (e.g. emphasis on individuals, future aims of the country). We then analyze Inglehart's 4-item post-materialism index (1971) from the EVS/WVS (share of post-materialists), which is comprised of two questions concerning future materialist or post-materialist priorities from the survey:

If the two post-materialist items are chosen the score is 3, if one post-materialist is chosen the score is 2, otherwise it is 1. In Figure 5, we develop the post-materialist index for our four countries by focusing on the increasing share of post-materialists.

What were the driving forces behind enhanced school autonomy in England? Altogether, a moderate shift toward post-materialism can be discerned in Great Britain starting in the 1980s.¹¹ This is reflected in the relatively strong emphasis on individualism, a greater desire for post-materialist political objectives, and an increased share of post-materialists (see Figures 2–5). During the reforms citizens placed increasingly little trust in the education system. This was apparent in a national survey, in which the inefficiency and weak performance of the school system were decried almost immediately before the governance reforms were introduced (Brook et al. 1992: A-27).¹² Thus, at a time when values were shifting toward the post-materialist paradigm, large segments of the population began to question the education system. This fell in line with a more critical assessment of the performance of the state and public management.

The skepticism regarding the efficiency and performance of the education system was also shared by the Thatcher government and triggered public debates (Pring 1995). Altogether, the public's critique of the education system strongly impacted the government's reform course:

In detail, the reforms aimed at creating a secondary education system based on choice and diversity (Mattei 2012: 253). The Education Reform Act 1988 introduced local school management as a form of school autonomy by taking away financial and personnel responsibilities from local education authorities (LEAs) (Schmid et al. 2007: 95), strengthening head teachers as well as parents, teachers and community members through school governing bodies (Schmid et al. 2007: 254). In addition, the publication of school performance data through league tables and a new funding system based on ‘money followed the pupil’ (Schmid et al. 2007: 253) established a quasi-market for secondary education.

To what extent were the reforms influenced by international diffusion dynamics? The evidence shows that not only overarching public sector reforms under Thatcher, but also shifting socio-political values and an increasingly skeptical view of public services fueled the education reforms before the OECD emerged as a significant actor. There is also little evidence that Britain's lackluster PISA performance further contributed to the reform process, as Britain already was largely aligned with OECD-promoted policies (e.g. school autonomy, accountability and multilateral quality assurance). Instead, implementation deficits during the introduction of school autonomy were largely made responsible for the sub-par performance (Knodel et al. 2010: 185).

Changing socio-political values in Sweden appear to have had an even more manifest impact on school autonomy reform than in England. The analysis of the EVS/WVS for Sweden reflects that a shift toward post-materialism largely coincided with education reform. The surveys revealed that a growing share of the population called for greater individualism, more direct means of participation in reforms and held views that were largely in line with post-materialist principles (Figures 2–5).

Yet, how exactly did the shift in values impact the actions of the Swedish government? In the late 1970s, the Swedish welfare state was criticized for its purportedly excessive bureaucracy (Klitgaard 2007: 228). Furthermore, it was increasingly difficult for voters to differentiate between the Social-Democratic Workers Party (SAP) and the state. After an electoral defeat, the SAP returned to government in 1982 and announced comprehensive public sector reforms as a reaction to its previous legitimacy crisis (Klitgaard 2007: 228). The Swedish government mandated a study on power and democracy, which provided in-depth insights on the quality of Swedish democracy and the distribution of resources (Petersson 1988: 144 et seq.). The study not only confirmed a greater desire for individualization and choice in governmental services, but also showed that it is difficult for parents to influence the educational offer and choose schools (Petersson et al.1989: 262). This was symptomatic of the perceived lacking legitimacy of governmental activity in the public sector, as ‘[ … ] welfare state activities did not accord precisely enough with citizen preferences’ (Gustafsson 1989: 179 et seq.). This prompted the Social-Democrats to put secondary school reform high on their agenda (Klitgaard 2007: 183).

Thus, it appears that the government explicitly reacted to the shift in political values by introducing far-reaching educational reforms. These entailed mutually reinforcing policies of both territorial decentralization and enhanced school autonomy, whereby the role of the state was reduced to the formulation of broader objectives, the allocation of financial resources and quality assurance. The state enhanced the decision-making autonomy of individual schools, enabling them to administer their own personnel affairs (Helgøy and Homme 2006: 148), develop their own profiles, and engage with their local environment (Klitgaard 2007: 456). Local communities, in turn, were given block grant funding for local schools. This gave them greater leeway to focus on local needs, while the introduction of a voucher system expanded freedom of choice between public and private schools.

What impact did international policy diffusion have on school autonomy in Sweden? Similar to England, Sweden had already restructured its education system in line with OECD-promoted principles before it became a strong reform catalyst (Martens et al. 2007). Thus, the Swedish reform efforts appear to have been largely driven by the shift in values and the government's reactions. However, the recent notable deterioration of the Swedish PISA results has generated new reform pressures and provided argumentative ammunition to the critics of school autonomy. For example, a recent governmental inquiry (SOU 2014) traced the decline in performance and increasing mistrust in the school system to the purported excessive transfer of authority to schools, which left teachers and management staff unprepared for their new tasks and aggravated social inequalities. This has increased calls for a renationalization of primary and secondary education (see below). Thus, the OECD's PISA results may actually have recentralizing effects on Swedish education, which would run counter to the perception of the OECD as a promoter of school autonomy.

Similar to Sweden, a post-materialist trend can also be observed in Germany. The desire for greater individualism increased the most significantly in Germany, while the desire for more active political participation has also grown over time (see Figure 2–4). Figure 5 shows that Germany has the largest share of ‘post-materialists’ among the four countries between 1980 and 1990 just before the reforms were initiated.¹³ A national survey also revealed that relatively negative view of the efficiency and performance of the education system in the 1990s (Allensbacher Berichte 2002/2005), while a survey from the Institute for School Development (Dortmund) showed that (future) parent generations wished for greater means for participation in schools (Kanders 1996).

Before the reforms kicked in, Germany had a territorially decentralized education system, which granted nearly all decision-making authority to the Länder, while the autonomy of individual schools remained low. Regarding the case of Hessen, the government made targeted efforts to react to the shift in values – long before public management and welfare state reforms began to spillover into education (Altrichter and Maag-Merki 2010). This was particularly apparent in the far-reaching reforms of the Social-Democratic-Green coalition in the early 1990s (Hepp 2003: 29), which focused on internal school developments. As stated in their coalition agreement, ‘Schools should change to cater to the interests of children and parents. Enhancing schools’ means for shaping their own affairs and increasing the participative rights of parents, local school authorities and pupils are important prerequisites for doing so’ (Koalitionsvereinbarungen 1991: 13). Furthermore, parents were granted greater freedom of choice over their children's secondary education pathways (Weiß 1993: 311). According to Education Minister Holzapfel, rapid societal changes made it necessary to rethink school policy and emphasize school quality. This required changes in internal school governance, which elevated school development to a collective project of the school community and the external school environment (Hepp 2003: 34). These ambitions led to a law in 1992 regarding the internal restructuring of schools in Hessen, which enabled greater participation rights and freedom of choice for parents regarding educational pathways.

As shown above, Hessen – along with several other Länder (Rürup 2007)¹⁴ – initiated major school autonomy reforms before the German ‘PISA shock’ (Martens and Niemann 2013). A particularly important role was played by Education Minister Holzapfel, who essentially acted as a ‘reform entrepreneur’ in designing the new school policy (Hepp 2003: 30). Although it cannot be ruled out that the recourse to international developments may have indirectly impacted the reforms in Hessen, it is imprudent to pinpoint transnational policy diffusion as the sole driving force. However, after the publication of comparative PISA results for the German Länder – in which Hessen's performance was only average – additional legitimacy pressures emerged for the German education system (Pütz 2008) and school autonomy remained on the political agenda under the subsequent Christian-Democratic and Liberal coalition (CDU-FDP) (Rudloff 2008: 344). The educational debate was also increasingly framed by the buzzword ‘quality assurance’ (Rudloff 2008: 341). The below-average PISA performance prompted Hessen to take action with several laws focusing on quality assurance by means of institutional empowerment (1999, 2002 and 2004). These measures were further flanked by German-wide efforts initiated by the Standing Committee of Ministers of Culture and Education (KMK) to enhance quality assurance and set educational standards in exchange for greater school autonomy. It is therefore safe to say that some degree of transnational diffusion, most notably from the OECD, fueled a paradigmatic change from input-oriented toward output-oriented governance aimed at enhancing the capacities of individual schools (Martens and Niemann 2013).

Did the spread of post-materialist values or international policy diffusion also drive educational reform in Italy? Figures 2–5 reflect a post-materialist shift among Italians before the major educational reforms in the late 1990s. Although no data dealing specifically with the impact of the shift in values on education exist, the government's new education policy approach lends evidence to our hypothesis that post-materialism was a catalyst for school autonomy. Contrary to the previous top-down policy-making style, the government actively promoted citizen participation and relied on public opinion in its reform thrust. Specifically, it conducted a public survey on the desired direction of educational reforms before their introduction to parliament (OECD 1998: 11):

Education Minister Luigi Berlinguer received over 1000 responses and triggered a broad debate on the weaknesses of the Italian education system (OECD 1998). Thus, the public response shaped the character of the reform bill, which entailed greater civic participation rights in shaping the affairs of increasingly autonomous schools (OECD 1998).

What role did international diffusion play? The Italian education system was extensively reformed in the late 1990s (OECD 1998) – precisely at a time when the OECD was increasingly evaluating national education systems. Specifically, Italian education policy-makers mandated to the OECD to conduct such an evaluation in early 1998, whose results impacted later reforms concerning educational quality and defined school autonomy as a lever for educational innovation (Vitteritti 2009: 15). Thus, the OECD, in particular, appears to have provided a further reform impetus. The below-average Italian PISA performance also played into a perception that the education system is inept in preparing pupils for the knowledge economy. However, we could not find any evidence that PISA and transnational diffusion gave a further impetus to school autonomy policies. Instead, more recent educational reforms passed in 2010 under Gelmini are driven by austerity aimed primarily to downsize the teaching corps and educational administration, while introducing several curricular modernization measures (e.g. more math and natural sciences).

Looking back, our empirical analyses lend most evidence to Hypothesis 2. The shift toward post-materialist values temporally coincided in all four cases more or less with the enhancement of school autonomy. As our national surveys revealed, the change in values led to a more critical assessment of state services such as education and demands for ‘custom-made’ services, stronger participative rights and less state authority. The governments’ actions reflect this shift and can be viewed as an explanatory factor for governance reforms (for Hypothesis 3, see below), resulting in increased school autonomy and in the Italian and Swedish cases a territorial decentralization that strengthened local authority as well.

We found little substantial evidence that school autonomy policies were significantly enhanced by international diffusion (Hypothesis 1) – neither in the ‘pre’- or ‘post-PISA’ phase. The case studies show that only the Italian reform approach was considerably influenced by the OECD, while the reforms in Hessen were intensified after the ‘PISA shock’.

Do partisan variations impact the form and justification of school autonomy (Hypothesis 4)? Table 2 provides a preliminary overview of our findings.

Our results confirm partisan differences in the justification of reform measures. For example, conservative governments under Thatcher and Major introduced three comprehensive educational reforms to foster choice and diversity (Department for Education 1992; Whitty 2008: 167), while a funding system was established based on the principle ‘money followed the pupil’ (Mattei 2012: 253). Financial resources were pegged to the number of pupils in schools, which then distribute the LEAs as global budgets (West et al. 2010: 457). Thus, the reforms resulted in greater freedom of choice for parents, who became increasingly involved in governing boards, while the new funding system and the publication of league tables¹⁵ boosted the quasi-market character of education. Parents and pupils were subsequently increasingly seen as consumers and schools as producers of educational services.

In contrast, the Blair Labour government emphasized social inclusion by mitigating some of the inequitable effects of the Conservative quasi-market reforms (West and Pennell 2012: 206, 218). Reforms focused on centralized standards and (further) reducing the power of local government (Bache 2003: 313), thus strengthening school autonomy. Moreover, the emphasis on partnership within the community (e.g. Education Action Zones, Excellence in cities) responded to the political imperative for New Labour to be seen as promoting ‘joined-up government’ (Cole and John 2001: 113). The administration sought to turn competitive market structures to their advantage by improving educational outputs and reducing social inequalities. Thus, while the reforms were chiefly driven by a social equality agenda, they reflected a stronger participatory element than the Conservative reforms.

‘Decentralization’ (in a territorial sense) and ‘school autonomy’ were the buzzwords behind the center-left Italian government's education reforms (Vitteritti 2009: 12). Under the center-right Berlusconi government though, the notion of autonomy was overshadowed by the concepts of entrepreneurial schools and ‘market-making’ (Vitteritti 2009: 14), leading to the strengthening of school management. Yet in both cases, it is difficult to pinpoint clear-cut partisan differences, as the reforms only hinted at notions such as market-orientation, accountability, efficiency, and democratization. It can, however, be plausibly assumed that the center-left government pursued a democratization agenda, as the 1997 reform increased both school autonomy and decision-making powers of regions and municipalities. However, in line with the Italian tradition of lacking local community participation in schools (Barzanò 2011), there were no efforts to broaden societal participation.

The Swedish Social-Democrats’ 1984 party program already expressed the idea of remodeling welfare state policies on the notion of freedom of choice (SAP 1984: 10 et seq.). An additional argument was to strengthen ‘consumers’ (Anell and Carlsson 1985: 24) by enabling parents to choose alternatives to public schools for which the costs could be largely covered by vouchers (Klitgaard 2007: 183). This Social-Democratic government also simultaneously sought to democratize education by strengthening local authorities and making schools more autonomous. This was supposed to enable them to react better to individual preferences than a centrally governed system. Increasing parent choice in turn compelled individual schools to take more entrepreneurial approaches in marketing schools to parents and creating their own profiles. Thus, it is safe to say that the Social-Democrats simultaneously pursued both a leftist and rightist autonomy agenda, which linked the democratization of governance (e.g. parent's rights and strengthening municipalities) with market-making policies (school choice, vouchers and quality monitoring). The subsequent center-right coalition (1991–1994) then instilled stronger market logic into the ongoing reforms by enabling so-called ‘independent schools’ to compete with public schools (Klitgaard 2008).

The underlying philosophy behind the 1992 and 1997 reforms in Hessen was to promote the individual development of pupils and schools (Holzapfel 1992). Proponents argued that schools can only be successful when governed as a common project of all involved stakeholders (Holzapfel 1992). This understanding of school development was particularly well reflected in the so-called School Conference (Schulkonferenz). This body consisting of teaching staff, parents and pupils was created by the Social-Democratic-Green coalition to promote democratic responsibility (Avenarius et al. 2003: 46). By contrast, the policies of the Center-Right-Liberal coalition (CDU-FDP) after 1999 were more geared toward quality assurance and further developing the personal responsibility of individual schools (Hepp 2003: 69). This resulted in the introduction of management methods and weakening of the School Conference (CDU-Hessen/FDP-Landesverband Hessen 1999).

Altogether, it is apparent that center-right parties have generally promoted school autonomy reforms based on competition, accountability and efficiency, while center-left parties tend to justify school autonomy as a means of enhancing democracy and political participation. However, these results must be put into perspective, as there were only traces of the democratic legitimacy argument in Italy and England. Moreover, the Swedish Social-Democrats pursued a dual strategy, which catered to both rightist demands for greater choice and competition and leftist demands for educational ‘democratization’.

What became of our counter Hypothesis 3 that increasing post-materialism could also lead to an aversion to school autonomy (when linked with management-oriented and utilitarian approaches to education)? Unlike Hessen, which has largely clung to the democratic-participative model (Hessisches Schulgesetz 2012), developments in Sweden, Italy and England enable us to preliminary assess this hypothesis. In Sweden, there has indeed been a recent public and political backlash against school autonomy due, in particular, to the deteriorating PISA results (Syll 2011). However, we found no evidence (yet) of ‘post-materialist’ or ‘anti-utilitarian’ motives, rather a wide-spread perception that school autonomy has resulted in an unequal distribution of resources among schools and thus undermined educational equality (Kornhall 2012). Critique of school autonomy has centered on the purported lacking educational governance capacities of schools and local authorities (Eklundth 2014), and not – as presumed in Hypothesis 3 – the downgrading of humanist educational values.

Public criticism in England is directed against the purportedly test-obsessed and unequal education system. The use of benchmarking is more widespread in England than in any other OECD state (OECD 2011a). It is argued, that ‘high-stakes’ tests can produce perverse incentives for teachers and pupils (OECD 2011a). Furthermore, school autonomy and accountability are criticized for not having generated greater equity outcomes for marginalized students (Lingard and Sellar 2012). Observers also point out that performance indicators used to secure accountability do not match the wider goals of education and do not work toward meeting the broader needs of society. Competition between autonomous schools based on league tables is thus seen to promote ‘teaching to the test’, which can undermine classical humanist and post-materialist values (West 2010; Boyle and Charles 2011).

In Italy we also found recent evidence of a public backlash brewing against Premier Renzi's very current school reform (buona scuola), which further institutionalizes the idea of entrepreneurial schools (scuola azienda) and vesting school managers with considerable steering autonomy. Critique is currently being voiced regarding the further deterioration of the historically weak participation of society and local stakeholders in education and the purported ‘hierarchization’ and ‘over-meritocratization’ of school policy (Magni 2015; Mauro 2015). While the evidence still remains inconclusive and should be scrutinized in future research, these preliminary findings may shed light on an interesting phenomenon: post-materialist attitudes not only largely were a catalyst in the spread of school autonomy policies, but also may contribute to their dismantling (when linked with a highly utilitarian approach to education).

We pursued two primary objectives in this paper. Focusing on international policy diffusion and shifts in socio-political values, we aimed to identify the mechanisms leading to greater school autonomy in four countries. Despite difficulties in empirically grasping post-materialism, we showed that changing socio-political values gave a significant reform impetus. This was reflected both in the relative correlation between the spread of post-materialist values and the introduction of school autonomy as well as the ‘post-materialist arguments’ (i.e. mistrust in public services and authentic political participation) in the governments’ reform justifications. By contrast, there was only scattered evidence that international organizations such as the OECD significantly contributed to the school autonomy trend. This quite surprising finding indicates that internationalization effects on national education policies may be somewhat overrated (Martens et al. 2010 for PISA and Bologna; Dobbins and Knill 2014 for the Bologna Process) and that changes in values and preferences of domestic decision-makers or the population in general should be given more consideration (see Martens et al. 2014; Busemeyer 2013).

Moreover, we showed that center-right and center-left parties generally promote a different understanding of school autonomy. Our empirical case studies underscored Gingrich's argument (2011) that conservative and social democratic parties pursue different rationales, which – despite the common autonomy trend – can lead to different forms of educational governance. Altogether, we are convinced that our analytical framework and empirical results provide a modest, but important advancement in the literature on educational reforms. Nevertheless, we see numerous possibilities for further research. First, our limitation to four countries and to EVS/WVS trend studies restricts the generalizability of our findings. Second, as noted above, researchers may consider subjecting Hypothesis 3 (post-materialist pushback against utilitarian, managerial approaches) to more robust comparative data over a broader timeframe. Third, additional explanatory variables could be incorporated. For example, the influence of education interest groups such as teachers unions on school autonomy remains relatively unexplored. In this regard, researchers may consider giving greater attention to the preferences of teachers unions as key educational stakeholders and how they inject their interests into the different school autonomy models.

No potential conflict of interest was reported by the authors.

Claudia Christ is a doctoral candidate at the University of Tübingen in the transdisciplinary doctoral programme ‘International comparative research on education and education policy in welfare states’. Her research interests are contemporary education reforms, welfare state politics and European integration.

Michael Dobbins studied political science and Slavic studies at the Universities of Konstanz, Warsaw and Rutgers. He completed his doctoral dissertation in Konstanz, which dealt with comparative higher education policies in post-communist Central and Eastern Europe. In 2013 he was appointed junior professor for policy analysis with the specialization education policy at the Goethe University of Frankfurt. His main areas of research are higher education policy, secondary education policy, policy-making in the EU, policy-making in the USA, transformation processes in Central and Eastern Europe as well as the Caucasus countries.