After German reunification: the implementation of a two-tier school model in Berlin and Saxony

ABSTRACT In the last two decades, many German states have enforced a partial integration of school types and have transformed their school systems into a two-tier model. The traditional tripartite school model, for which Germany has long been known, is thus no longer a characteristic of their school systems. This article analyses the determinants of the structural changes in Berlin and Saxony by applying concepts of historical institutionalism. In both states, the reform of the schools’ structures can be described as path dependent, but also a specific constellation of interests, strategies and actors led to a partial erosion of the respective traditional structures.


Introduction
In comparative social stratification research, Germany is known for its stratified school system, which involves early academic selection and the tracking of students into hierarchically structured and spatially segregated school types. 1 While many nations converged towards comprehensive schooling models after the Second World War, 2 Germany retained its traditional multi-tiered school system. 3 After four years of joint schooling, 4 students are referred to distinct educational tracks, each associated with a different curriculum and certificate (the Hauptschule, Realschule and Gymnasium). Germany is also known for its differentiated system of special schools (Sonderschule) for students with special educational needs. 5 Some German states established the This article adopts a historical-institutionalist perspective, examining the sources of stability and the drivers of change in secondary school systems in two German states. For the empirical reconstruction of the event sequences and causal narratives, the study carried out a qualitative content analysis of parliamentary documents, official statistics, secondary literature, semi-structured interviews and newspaper reports. The material was interpreted with the help of a coding frame, which was developed based on the theoretical concept and evaluated over the course of the analysis. 9 The following section outlines the theoretical approach, introducing the key concept of path dependence and mechanisms of change in school structures, and the methodological approach. The article then provides an in-depth account of policy developments in Saxony and Berlin. The final section summarises the findings and provides an outlook for further research.

Theoretical approach and methodology
In comparative political research, the study of education systems and education politics has historically been neglected. 10 Especially in the field of welfare research, education was seen as 'special' and was often excluded from comparative studies of welfare politics. 11 Nevertheless, this has changed in the most recent years. A growing literature in political science has discovered education politics as a field of investigationespecially for the fields of vocational education or higher education. 12 The field of primary and secondary education politics is still mostly being ignored. The few studies with an explicit political science perspective concentrate on the role of teacher unions 13 or the political origins of primary education systems in the nineteenth century 14but political scientists rarely study the stability or change of school systems. 15 With the analysis of the reform processes in two German states, this article will demonstrate how fruitful the historical-institutionalist perspective can be for investigating the stability and change of school systemsnot only for the German context, but also for the comparative and international education literature. 9 The coding frame can be found in Nikolai, Schulpolitik im Wandel. Theoretical approach: path dependency and institutional change The structural reforms implemented in many German states after reunification and in the 2000s deviate from the historical developmental path of the tripartite model, whose origins go back far beyond the foundation of the FRG and the GDR in 1949. 16 For many decades, this developmental path proved remarkably resistant to profound change. 17 The concept of path dependency 18 stresses the lasting stability of institutions, describing it as a historical legacy caused by trend-setting decisions at critical junctures. Institutional configurations have a lasting effect over long periods of time because actors tend to hold on to institutions. Due to investments (e.g. in school buildings, administrative routines, teacher education for different school forms at universities), institutional developments become deterministic in terms of positive feedback processes. Thus, a chosen path is reinforced as time passes, resulting in so-called 'lock-in effects', and deviations from a certain path become increasingly unlikely. Accumulated commitments and investments in the selected path make it difficult to effect any profound change. 19 On the one hand, path-dependent analysis allows for the identification of the restrictive general conditions that determine the extent to which political decision-makers have room to implement reforms. On the other hand, the concept allows for an identification of the drivers of change that may alter these general conditions and may thus effect a (more or less) pronounced destabilisation of the established institutional order. The structural continuity of the stratified school system with its early selection has to be understood as such a path-dependent development. The origins of its forces of inertia can be traced back to the nineteenth century. 20 The status quo of the rigorous tracking system after the Second World War was supported by teachers, school-trackspecific interest groups, 21 political parties 22 and parents.
However, institutional change and the dissemination of new knowledge are also possible beyond the paths chosen at one point. For instance, new findings or critical junctures (brought on by wars, crises or major scientific discoveries, for example) may necessitate new institutional arrangements or cause institutions that were thought to be stable to collapse. The formation of the FRG and the GDR, the division of Germany and German reunification must be regarded as such critical junctures.
Institutions are characterised by a high degree of stability as long as their specific mechanisms of reproduction are left to operate undisturbed. If these mechanisms are suspended or begin to erode, however, there is room for change that reform-oriented actors may take advantage of. 23 It is the merit of Benjamin Edelstein to adapt the concept of reproduction mechanisms to the field of school systems and to the stability of the German school structures. 24 Functionalist mechanisms of institutional reproduction may erode if an institution is no longer capable of fulfilling its function in the overall system, or if the functional needs of the overall system have changed. Utilitarian mechanisms of reproduction may deteriorate if competitive comparisons show an existing institution to be inferior to other alternatives, or if information about performance deficits initiates learning processes among the relevant decision-makers. Legitimisationbased mechanisms of reproduction may erode if societal values and/or the subjective convictions of relevant decision-makers begin to change, for instance due to the diffusion of ideas or alternative paradigms that delegitimise established institutional forms and practices. Power-based mechanisms of reproduction may be suspended if the balance of power in a given policy area has changed. This may occur, for example, if dominant actors begin to lose influence and are no longer able to enforce their interest in preserving an institution against other actors, or if actors redefine their interests and subsequently 'switch sides'. 25 The article will especially focus on the role of actors and their changing ideas, motives and strategies. A change of attitudes among stakeholders may not only be caused by exogenous events but also by endogenous factors such as changing values or ongoing learning processes. 26 Accordingly, stakeholders pass on and justify their knowledge through political discourses (including parliamentary debates). Stakeholders are not only carriers of knowledge: they also transform that knowledge. In that sense, knowledge is not static; it changes and can be changed. Old certainties may be forgotten, but new societal knowledge may emerge as well. New knowledge is more likely to become established if exogenous events upset our existing knowledge, thereby creating an open space for challenging existing certainties and injecting new knowledge into the discourse.
Actors such as teacher unions or parents have the potential to act as powerful forces for stability. Plans like school form integrations may influence the vested interests 27 of actors to protect school institutions. For example, teacher unions may fear changes in salaries, social prestige or the method of instruction. In addition, parents are also an important group of the electorate and may be afraid that a school form integration will restrict the educational chances of their own children. 28 With the rising share of parents who have acquired their certificate for a higher education entrance qualification, the proponents of the Gymnasium as an educational clientele may gain power in 23  educational discourses. Political parties tend to refrain from large structural school reforms, as these may jeopardise their re-election. 29 However, it is possible that actors who are convinced by the need for structural change suspend power-based mechanisms. Change actors may alter the institutional protection by mobilising resources, by developing convincing strategies and argumentations, and by creating new alliances with other actors. The concept of change agents, which is stressed by James Mahoney and Kathleen Thelen as an important factor for driving institutional changes, is similar to the concept of institutional entrepreneurs or policy brokers. 30

Methodological approach
This study employed different empirical sources: (a) political documents, (b) media coverage and (c) expert interviews with education-policy-makers.
As political documents, the following document types were analysed: party manifestos, coalition agreements, protocols of parliamentary debates, statements of interest groups and ministerial documents (such as school development plans) and governmental statistics. With these documents, it was possible to trace the political process and to capture the contextual factors. Political documents were also a source for analysing the interests and ideational foundations embedded in the different texts. An important source for the analyses was parliamentary debates, which 'are especially wellsuited for studying the transformation of public justification'. 31 After all, even though decisions in Germany's parliamentary system are mostly made in parliamentary committees rather than in parliament, parties do use parliamentary debates as a public forum to justify their decisions to the electorate. 32 Their goals, motivations and arguments are presented intentionally and carefullyin other words, these are not 'shortterm interpretations' but 'official party statements'. 33 The analysis of media coverage (together with the document analyses) included different articles from daily newspapers in Saxony and Berlin. 34 By analysing the media coverage, it was possible to identify the relevant actors and to contrast the findings with the document analysis and expert interviews.
At the core stands an analysis based on expert interviews with education-policymakers (10 for Berlin, 10 for Saxony). The results of the interviews have been anonymised. The expert interviews included interviews with ministers of education, members of the educational administration, members of parties in the parliaments (mainly 29 Nikolai and Rothe, 'Konvergenz in der Schulpolitik?'; Nikolai and Rothe, 'Entscheidungen in der Schulpolitik'. 30  Christian and Social Democrats) and representatives of interest organisations (e.g. teacher unions). The interview partners were in office in Berlin or in Saxony in the years 1990 and 1991 (in Berlin also in the 2000s) and they were surviving witnesses of the school reform processes at this time. The interviews considered problem-focused and episodic elements using a semi-structured guideline. The interviews were conducted between 2011 and 2016. All the empirical sources (documents, media coverage and expert interviews) were analysed using qualitative content analysis based on Schreier 35 and the typology of mechanisms of change presented in the theory section.

The institutional evolution of school structures in Berlin and Saxony
This article selected two cases representing different trajectories of change. In Saxony, sweeping structural reforms were implemented over the course of only a few months as part of German reunification, whereas structural reform in Berlin's school system was driven by long-term incremental processes.
The common point of departure for both states was the short period immediately after the Second World War, which can be regarded as a critical juncture as defined by the concept of path dependence. In the western occupation zone of the United States, the United Kingdom and France, which later formed the states of the FRG, the tripartite school structure continued. 36 In contrast to the western powers, the Soviet occupation force did in fact implement a comprehensive structural reform in 1946, thereby turning away from the traditional three-tier model. 37 The fundamental institutional decisions of the post-war period are essential for understanding subsequent developments, because from then on, reform-oriented actors defined their interests within the established institutional logic and adapted their reform strategies to that logic.
Saxony: continuity in times of changethe short road to a two-tier model The analysis of the reform in Saxony will start with the developments of the GDR school system and its impact on educational inequality as these have framed the ideas, strategies and motives of the East German education-policy-makers who were in office after reunification.

Developments in the GDR school system
In the GDR, a socialist-style unitary school system was introduced with a 10-year comprehensive school, the Polytechnische Oberschule (POS) or polytechnical upper school. The secondary school type leading to the Abitur after POS completion was preserved under the title of Erweiterte Oberschule (EOS) or expanded upper school ( Figure 1). 38 The GDR also introduced the apprenticeship Berufsausbildung mit Abitur, which enabled students to acquire a general higher education entrance qualification while simultaneously completing a vocational training programme. Its design as a unitary school system notwithstanding, the GDR school system was highly differentiated. Access to the EOS was limited; besides academic achievement and societal demand, political attitudes and party involvement were the most important criteria for admission. 39 Therefore, only 13% of an age cohort graduated from the EOS with an Abitur. 40 Additional measures such as Russian foreign-language classes (so-called 'r-classes'), Abitur preparatory classes at the POS and EOS (until 1982), and special classes and schools for gifted children (from 1963 onwards) provided privileged access to the EOS for a small group of students, especially the children of the 'socialist intelligentsia'. 41 Due to the prevailing ideological indoctrination and the high degree of social and political selectivity, the GDR school system experienced a massive loss of legitimacy over the years. By the time of German reunification the GDR was discredited, as numerous statements by the different groups of the citizens' movements, newly founded teacher unions and articles in the media show. The GDR school system enjoyed hardly any societal support, mainly due to the social and political selectivity for admission to the EOS and its ideological indoctrination. As a result, major parts of the population did not think preserving the existing school system was a desirable option. 42 When the stakeholders in the GDR school system finally lost their political power with the fall of the Wall, they also lost their ability to enforce their interest in preserving that system. 43 The above-mentioned developments in the GDR school system provided the starting point for the fundamental changes in Saxony's school system. According to the Unification Treaty of 31 August 1990, the new states were supposed to refer to the 1964/1971 Hamburg Agreement for guidance when redesigning their school systems. The Hamburg Agreement was a consensus between the Western German states on unified forms for organising secondary schools such as the Hauptschule, Realschule, Gymnasium, and later also the comprehensive schools. In September 1990, however, the negotiators in the Joint Education Commission of the GDR and the FRG agreed to establish a 'common and comparable basic structure' rather than a school structure as in West Germany. 44 This agreement can be considered a 'fundamental choice of direction', 45 loosening the institutional regulations to the extent that the introduction of a new school type became possible in the new states.
Like Saxony-Anhalt and Thuringia, Saxony decided to establish a two-tier school system (see the School Act of 20 June 1991) (see Figure 1). 46 Beginning with the 1992/1993 school year, in addition to the Gymnasium, a Mittelschule (middle school) was introduced, combining the Hauptschule and Realschule tracks into one school type. However, it was not clear from the outset that a two-tier model would eventually be implemented. Early on in the decision-making process, a tripartite model that was based on the West German tradition was on the table as well. 47 So how did it happen that the idea of a two-tier model prevailed in the months between the foundation of the Saxony state legislature in October 1990 and the passing of the School Act in June 1991? submitted. It suggested that the following secondary school types should be established in addition to a four-year primary school: the Werkrealschule as form of a Hauptschule, Realschule and Gymnasium, as well as the option of creating comprehensive schools (Gesamtschule) as a supplementary school type. Due to the parliamentary majority, one could assume that the Christian Democrats would enforce this proposal for a tripartite school system. 48 However, the next few months showed that the proposal was not supported by a broad consensus in the population and that reform-oriented actors in the government and in the ministry administration followed an alternative approach for the school structure in Saxony. The proposal met with massive criticism by the public. The Ministry of Education and Culture received about 1200 letters rejecting the envisaged Werkrealschule as a 'Hauptschule in disguise'. 49 Parents and teachers saw it as a 'step backwards compared with the universal polytechnical upper school, which awarded the middle school leaving certificate after grade 10'. 50 Many Saxons did not want their new government to go too far in dismantling the structures of the GDR school system. 51 Against the backdrop of their experiences with selective admissions in the GDR school system, the envisaged tripartite system, which was known for its social selectivity, did not enjoy unrestricted legitimacy either. To be sure, the introduction of the Gymnasium was welcomed by the population because, compared with the EOS with its restricted access policies, the Gymnasium promised broader and, most important, strictly performance-based access. The Werkrealschule, by contrast, was met with scepticism because parents feared that, unlike the Mittlere Reife, education qualifications awarded by a Werkrealschule would not make graduates eligible to transfer to a Gymnasium later on. This is why there was strong opposition to establishing the Werkrealschule as an independent school type. A two-tier school system, by comparison, promised to provide more opportunities for upward educational mobility. 52 In many expert interviews the two-tier model was also seen as an independent solution which would not cope with the school structure of the Western German states and which was understood as a form of 'home-grown' solution. The following statements emphasise this: It was in the air. When we looked at the school system it was clear for us, that we don't want to introduce the Hauptschule. . .. We didn't want to repeat the mistakes of the West German Länder. 53 They planned to impose the tripartite school structure. However, we said no, we would like to create something different and something that will fit to our situation in Saxony. 54 48 Interview SN5, interview with a representative from the education ministry in Saxony, May 5, 2012. 49  From the beginning of the reform processes in Saxony, the introduction of the Gymnasium was not questionedeven by the Social Democrats in opposition in parliament. In contrast to the limited chances in the GDR of receiving the higher education entrance qualification, the Gymnasium was perceived as a school form that would offer significant opportunities and that was seen as part of the western culture. Moreover, many parents or grandparents had their own experiences with the Gymnasium before this school form was abolished in the GDR: With the Gymnasium, we had the association that more young people can now get the higher education entrance qualification. The Gymnasium was not a restriction; it was an opening for us. . .. With the Gymnasium, we were able to participate in the western culture. 55 Concerning legitimisation-based mechanisms, the government proposal failed to convince teachers and parents that the introduction of the Hauptschule as a single school form was appropriate. Many statements of parents and teachers in letters to the ministry or in newspapers demanded distinctive school forms, but not so segregated as within a tripartite school system.

Influence of change agents
The Werkrealschule failed to garner widespread support, not only in the population at large but also among Saxony's Christian Democrats. While the group supporting the Minister of Education and Culture (i.e. Stefanie Rehm) lobbied for the adoption of a tripartite school system in the winter of 1990/1991, 56 another group within the CDU argued in favour of taking an independent route in Saxon school policy. 57  State Secretary Nowak had a reputation as an experienced school policy-maker and was considered to be the driving force in the school structure reform. 62 Along with a small group of staff members at the Ministry of Education and Culture, he drafted a proposal for a new school bill. By launching an intensive PR campaign, SPD member 55 Interview SN3, own translation. 56  Nowak sought to forge an alliance with like-minded actors in the CDU cabinet to win approval for the Mittelschule and thus a two-tier model. The teacher union landscape, which consisted of the regional associations of the Trade Union for Education and Science (GEW), the Philological Association (PhV) and the Saxon Teacher Union (SLV), was divided into two camps. The GEW supported the implementation of comprehensive schools, whereas the PhV and the SLV advocated a tripartite school system. However, the camp of the tripartite advocates was divided again into the SLV and PhV and this made it difficult for the conservative teacher unions to enforce a tripartite school system, especially as the SLV was ready to accept comprehensive schools alongside the Gymnasium, the Realschule and the Hauptschule. However, for the reformers in the CDU the introduction of comprehensive schools was unacceptable and a two-tier model was an alternative to this. At the end of the decision process, the SLV and the PhV accepted the two-tier model, as this model still promised the introduction of the Gymnasium. 63 The role of the school building infrastructure and the declining role of the Hauptschule for vocational education The supporters of the two-tier solution increasingly found allies in the Saxon state legislature as well, not at least because the introduction of the Werkrealschule was considered to be inefficient from an economic point of view. Given the existing building infrastructure, the creation of a tripartite school system would have required enormous investments. 64 The building infrastructure of the GDR school system was designed for a comprehensive school system. The schools, especially in rural areas, were small with a maximum of two classes per grade. The introduction of a two-tier model was a better fit to the existing infrastructure of school buildings than a tripartite model. Furthermore, a two-tier school system was in line with the people's interest in preserving the full range of secondary options close to students' homes. Finally, a two-tier model was a better match for the existing professional qualifications of the teaching staff. 65 Teachers of the GDR school system were trained to be teachers at the POS or the EOS. Compared with these school forms, being a teacher at a Hauptschule was less attractive to the majority of teachers as regards salary and prestige. 66 From parents' letters to the Ministry of Education and Culture, newspaper reports and discussion forums, it became clear that parents were unlikely to choose the Werkrealschule for their children. 67 The Hauptschule's ongoing decline in the West German states and its devaluation as a 'school for leftovers' was anticipated in Saxony as well. 68 In many states of West Germany, the Hauptschule had already lost a number of students in the 1980s as students from the Hauptschule had increasing difficulties in finding a training place in the vocational sector (see also the developments in West Berlin). The Hauptschule was no longer the entry gate to vocational education. Parents were increasingly afraid to send their children to the Hauptschule and preferred to 63  choose another secondary school form. In West Germany, the share of students in the seventh grade at the Hauptschule decreased from 79. 3% in 1952/1953 to 33.9% in 1987/ 1988. 69 Doubts were raised about whether the Saxon business community would accept the Werkrealschule leaving certificate as a sufficient qualification for admission to vocational training programmes. Wolfgang Nowak, as the change agent in Saxony, knew the difficult situation of the Hauptschule in his home state of North Rhine-Westphalia. He used the 'bad' image of the Hauptschule to convince reform-oriented actors in the Saxon government and the education ministry to choose a two-tier model.
In summary, the proponents of a two-tier school system were able to count on the broad political and public acceptance of their reform model. The decision to introduce a two-tier model in Saxony did not represent a break with the past because, despite all of the changes, the new school structure also offered a great deal of continuity with the GDR school system. It was the broad wish of parents, teachers and education-policymakers in the government and the education ministry to implement a form of differentiation with a strong Gymnasium. However, compared with the traditional tripartite model, the Saxon model was seen as less differentiated than the tripartite school model. The two-tier model was a compromise between the comprehensive school structure of the GDR and the West German tripartite school system. The decision not to establish the Hauptschule as a distinct school type reflected the ongoing prevalence of societal ideas concerning a minimum level of general education that should not fall below the middle school leaving certificate awarded after 10 years of schooling. Furthermore, the selected model could be more easily linked to the existing infrastructure of the GDR school system. Finally, the Hauptschule was not seen as an appropriate school because of its declining significance for vocational education.
Berlin: incremental changethe long road to a two-tier school model Since the Second World War, the school system in Berlin has experienced several structural changes. 70 In the first instance, the article will describe the structural developments in West and East Berlin before reunification, whereby the focus is on West Berlin as there are more structural changes than in East Berlin. Then, the transformation of the school system in the reunified Berlin in 1991 and in the 2010s is analysed.

Developments in West and East Berlin before 1990
When school operations were resumed after the end of the Second World War in May 1945, the Berlin authorities initially continued the pre-war tripartite structure (Volksschule, Mittelschule, Oberschule) (Figure 2). 71 Berlin was not assigned to a single occupying force, and its four sectors were directly subordinate to the Allied Control Council made up of the United States, the Soviet Union, the United Kingdom and France. In the first parliamentary elections in October 1946, the Social Democratic Party (SPD) won an absolute majority in all four sectors. Also represented in parliament were the Christian Democratic Union (CDU), the Socialist Unity Party of Germany (SED), and the Liberal-Democratic Party of Germany (LDP, later FDP). Hence, the left-wing camp, consisting of the SPD and SED, had a clear parliamentary majority and could enforce in November 1947 the implementation of a 12-year comprehensive Einheitsschule for all children. It consisted of a six-year first cycle without tracks and a second cycle with tracks. Starting in eighth grade, students were divided into a practical branch ending after ninth grade and an academic branch ending after 12th grade (Figure 2). 72 After the division of Berlin in 1948 and of Germany in 1949, the two halves of Berlin followed different paths in their school structures. East Berlin, as part of the GDR, adjusted its school system to the comprehensive school structure of the GDR (see Figure 1). 73 Before reunification, the school structure in East Berlin was not modified.
In West Berlin, the Social Democrats lost their majority in the 1951 elections and formed an all-party coalition with the CDU and the FDP. 74 In this all-party coalition, the SPD had no other choice than to accept the conversion of the comprehensive school model into a tripartite model in the school year 1952/1953 (see Figure 2). 75 Nevertheless, the SPD could enforce that the primary school lasted six years instead of four years as in the other Western German states.
With the social-democrat-liberal coalition in the 1960s, the SPD made a renewed push for structural changes. Following several school pilot projects that had emerged by the late 1960s, in the 1972/1973 school year Berlin was the first federal state in the FRG to introduce the comprehensive school as an additional school type and transformed its tripartite system into a four-tier school system (Figure 3).
Since the 1950s, the enrolments have experienced long-term erosion. The number of students at the Hauptschule was constantly decreasing, from more than 50% (1952), to 20% (1980) and to less than 10% by the end of the 1980s (Figure 4). The Realschule also experienced a decline in enrolment. In contrast to the Hauptschule and the Realschule, the comprehensive school and the Gymnasium experienced increasing demand. By the end of the 1980s, the Hauptschule was perceived as a 'school for leftovers' and was rarely chosen by parents as a secondary school for their children. 76 The structural reforms carried out in West Berlin may be characterised as path dependent until the reunification of Germany. Although West Berlin established an integrative school form by introducing the comprehensive school, it was only a fourth additional school form so that the model of a vertically structured school system still continued. Despite its domination in West Berlin, the Social Democrats could not and did not want (due to their desire for re-election) to enforce their interest in changing the school structure against a conservative block (Christian Democrats, German Philological Association) with its strong electoral base for mobilisation. When it became evident in the 1980s that the Hauptschule had lost its function for vocational education and was less and less chosen by parents and their children as a school after primary school, the SPD was not in government. From 1981, West Berlin was governed by a coalition of the Christian Democrats and the Liberals. As the SPD won the elections in 1989 and formed a coalition with the Alternative Liste, a regional association of the Green Party, there was the chance to transform the school structure in West Berlin. The Berlin Wall came down only a few months later, in November 1989.

Developments in the unified Berlin after 1990
With German reunification in 1990 and the reunification of East and West Berlin to the one-city state of Berlin, it became necessary to unify the different school structures. As in Saxony, for the education-policy-makers in East Berlin preserving the existing school system of the GDR was not a desirable option. As in Saxony, many education-policymakers in East Berlin 77 criticised the fact that access to the EOS was limited to students from the 'socialist intelligentsia' and the ideological indoctrination of the curriculum (especially for the subjects of military education and civic education). In West Berlin, supporters of an integration of school types, particularly in the ruling coalition of the SPD and the Greens, called for a two-tier model for East and West Berlin consisting of the Gymnasium and the comprehensive school. 78 In the absence of clear political support for structural reforms, reinforced by the electoral defeat of the red-green coalition in the first citywide elections in 1990, the comprehensive school structure of East Berlin was replaced by the school system of West Berlin with its four-tier model (see Figure 3). 79 The electionwinning CDU enforced this transfer of school structures during the government negotiations with the SPD. 80 The SPD feared a 'culture war' as happened in the 1950s and accepted the traditional tripartite school structure including the comprehensive schools: We [as the SPD in Berlin] favoured the comprehensive schools, but it was too dangerous for us to weaken the Gymnasium. This would result in a culture war in Berlin. In addition, the parents, who favour the Gymnasium, are assertive in elections. . . Therefore, in the end we were not interested to change the school structure in Berlin. 81 The reunification can be characterised as a window of opportunity. In West Berlin, the Hauptschule had a poor image and therefore the implementation of a two-tier model was discussed. However, power mechanisms overlapped any considerations in finding an alternative solution for the unified Berlin. The election victory gave the CDU a powerful position to enforce its structural model and to transfer the tripartite model from West to East Berlin. Only 20 years later, after reunification, could the original idea of a two-tier school model be realised. As we shall see, changes in all four mechanisms of institutional reproduction made it possible to reform the school structure.
Increasing pressure in the 2000s years enabled a structural shift In the 2000s, enrolment continued to decline at the Hauptschule (see Figure 4). Only 6.3% of all students were enrolled at the Hauptschule in 2009. The long-term changes in school choice behaviour led to a problematic concentration of at-risk students at Berlin's Hautpschule, especially students with a migration background. Whereas enrolment at comprehensive schools and the Realschule stagnated and even slightly declined, the Gymnasium faced a huge influx of students. In 2009, the Gymnasium accounted for more than 46% of all students in grade 7. Due to rapidly declining enrolment, the preservation of the Hauptschule as an independent school type was made increasingly difficult and too costly from the perspective of utilitarian dimensions. Further, the Hauptschule had lost its function in the Berlin school system, as students from the Hauptschule had massive difficulties in finding a training place in vocational education. In 2009, only 29% of the students with a certificate from the Hauptschule could enter the dual vocational sector in Berlin. 82 The Hauptschule in Berlin and the four-tier school system also increasingly lacked legitimisation. Over the course of the German PISA shock that began in 2001, a heated debate arose regarding the performance and fairness deficits in the Berlin school system. In the first German state comparison of the PISA survey results, called PISA-E in 2002, 83 Berlin revealed a strong relationship between student competencies and their social background and a high proportion of at-risk students at the Hauptschule. The PISA-E test also demonstrated the existence of school-type-specific developmental milieux. Finally, the legitimacy of the Hauptschule was shaken by the debate over the Rütli School in 2006. Teachers at this Hauptschule lamented the violence at their school in a public letter. Subsequently, the end of the Hauptschule was debated in newspaper headlines in Berlin and nationwide. 84 The loss of legitimacy, the strong decline in enrolment and the lost function for vocational education made it impossible to maintain the Hauptschule as an independent school form. The reform pressure was high, but a reform in Berlin was only possible because education-policy-makers from the different political camps were willing to reform the school system. Education-policy-makers from the CDU as a strong opponent of school reforms in the past had to reconsider their views that the four-tier school model no longer represented a viable solution for Berlin. In both parties (i.e. SPD and CDU), an increasing number of voices announced the two-tier model as a political solution. 85 In this climate of willingness to reform, one change agent become important: With the appointment of Jürgen Zöllner in 2006 as Senator for Education and Research in the left-wing coalition of the SPD and the Linke, a major structural change occurred. In September 2008, Zöllner proposed to integrate the Hauptschule and the Realschule into one school form. This proposal had to be understood as a tactical manoeuvre because it initially left out the comprehensive schools. The supporters of the comprehensive schools in the SPD and the parents, as well as the principals, were vehemently against the reform proposal. However, numerous informational events and small informal discussion groups were able to mitigate the concerns of the comprehensive schools about being potential losers in any school form integration. It was always highlighted that the comprehensive schools would be the model and threshold for school reform. With this argument, it was possible to integrate the opponents from the SPD and the Linke. 86 With his moderation and his ability to involve many stakeholders, Zöllner must be characterised as a change agent. In January 2009, Zöllner proposed to integrate the Hauptschule, the Realschule and the comprehensive school into one single school form, the Integrierte Sekundarschule (integrated secondary school), which would offer all secondary school degrees including the Abitur (see Figure 3). In addition, Zöllner was able to garner broad and bipartisan support including the following supporters: the government coalition, the left-wing oriented Union of Education and Science (Gewerkschaft Erziehung und Wissenschaft), the parents' association (Landeselternausschuss), the association of head teachers of the Gymnasium (Vereinigung der Oberstudiendirektoren), and different industry associations (among them the Berlin Chamber of Commerce and Industry and the Berlin Chamber of Handicrafts). 87 As the Gymnasium was preserved as an independent school type, Zöllner's reform proposal met with broad approval. In January 2010, the parliament of Berlin decided to implement the school reform at the beginning of the 2010/2011 school year.
In summary, the school form integration in 2010/2011 was possible due to changes in all four dimensions of institutional reproduction. The declining enrolments of students at the Hauptschule, its lost function for vocational education and the problematic concentration of at-risk students left the Hauptschule a costly and ineffective school form. However, the reform was only possible due to the reform willingness of the education-policy-makers from different political camps and the capability of a change agent to convince and bring together different actors.

Summary and outlook
In recent years, many German states have implemented sweeping structural reforms of their school systems. This article has shown that concepts from historical institutionalism with a focus on the concept of path dependency and the role of actors could be used effectively in the analysis of persistence and change in educational institutions. In particular, Edelstein's concept for the mechanisms of institutional reproduction in the context of school politics is a fruitful instrument for analysing the stability and changes of school systems. On the one hand, the structural reforms carried out in Saxony and Berlin may both be characterised as path dependent because, in a two-tier model, students are still selected after grade four or grade six for different secondary school forms. The Gymnasium remains unaffected by the school reform integration and still exists as an independent school form in all German states, and so do the special schools. According to Paul Pierson, the reforms in Saxony and Berlin could be understood as 'bounded change', 88 institutional change that is hampered by path dependency. Nevertheless, also within the remaining institutional logic of tracking students, the two-tier model led to substantial change in the school systems in the German states. In Saxony, the introduction of a two-tier school structure was the result of political negotiations taking place in a decision-making situation that lasted for only a relatively short time. As part of a path-dependent sequence, actors could choose between a limited number of alternatives with regard to school structure, one of which eventually prevailed as a political compromise. By contrast, the school structure reforms implemented in Berlin represented the culmination of long-term incremental changes that had made the preservation of the Hauptschule as an independent school type increasingly difficult. As part of a path-dependent sequence, the switch to a two-pillar model finally emerged as the only politically feasible option.
In both case studies, the integration of school forms was possible because institutional reproduction was suspended or eroded in four dimensions: Whereas in Saxony education-policy-makers perceived and discussed the declining role of the Hauptschule in the Western German states in 1991, Berlin experienced a tremendous descending role of the Hauptschule in the 2000s. In both states education-policy-makers doubted that the Hauptschule could play a functional role for parents and vocational education. Concerning utilitarian mechanisms, the Hauptschule was also perceived as a costly school form. In Saxony, the school building infrastructure of the former comprehensive system was not suitable for the implementation of a tripartite school structure. Introduction of the Hauptschule would have meant enormous investments in school buildings. Concerning legitimisation mechanisms, the quality of the Hauptschule and its role as an appropriate school has been questioned by education-policy-makers in both states. Nevertheless, the erosion of reproduction mechanisms in the functional, utilitarian and legitimisation dimensions could only lead to a structural reform, as long as power-based mechanisms also changed. Furthermore, reform-minded actors and specific actor strategies of individual entrepreneurs were identified as significant factors. In both states, reform-oriented actors put aside their ideological beliefs and as change agents could form a broad reform coalition and could convince regarding the necessity of reforms. The example of Berlin showed that in 1991 a change agent who could bring together different political camps was absent and only by the end of the 2000s, with Zöllner as a new education senator, could such a change agent transform the power-based mechanisms.
Although this paper has focused on the cases of Berlin and Saxony, the analyses can easily be extended to comparative studies on school reforms, e.g. across other German states or reform processes in other countries. To that effect, the present analysis offers a wide range of options for further analyses, such as focusing on path dependency and the drivers of change.