Massification, unification, marketisation, internationalisation: a socio-political history of higher education in Sweden 1945–2020

ABSTRACT Since the mid-twentieth century, higher education has been subject to spectacular growth. While the expansion of higher education is undoubtedly a general trend, its actual characteristics in terms of its specific conditions and driving forces vary by context. In this article, our aim is to develop such a socio-political historical narrative of Swedish higher education since 1945. We focus on how the Swedish case evolved in four salient aspects: (I) the transformations of the welfare state and the economy broadly defined; (II) the policies directly involved in redesigning the higher education system; (III) the scale, composition and structural relations between the parts of the higher education system; and (IV) the way different social groups (vis-à-vis gender and class) have used the system. We show that higher education policy has shifted its focus from massification and unification to marketisation and internationalisation. We notice that the conditions for the expansions differ substantially during the first three decades after 1945 compared to the 1990s in political, economic and demographic terms. While the system has opened up for increased participation overall, the way students are sorted into different institutions and fields of study by gender and class reveals a remarkably rigid social structure.


Introduction
Since the mid-twentieth century, higher education has been subject to spectacular growth worldwide (Schofer and Meyer 2005;Charle and Verger 2012). Initially a prospect for the few, higher education has mutated into a societal sector of strategic interest for a large variety of stakeholders. While the expansion of higher education is undoubtedly a general trend, its actual characteristics in terms of its specific conditions and driving forces vary by context. In higher education systems such as the Swedish one, where the level of state coordination and integration has been high (Clark 1983), political action and reforms have highly impacted development, especially in terms of increasing scale, composition and the structural relations between component parts. In this way, examining the relationship between higher education and the state is critical in order to understand how such expansion has unfolded since 1945 (cf. Stevens and Gebre-Medhin 2016) and how higher education has changed as it has expanded. Another pivotal category of actors that define the characteristics of a higher education system are students, particularly how they utilise the system. To move beyond research that merely illustrate general models of diffusion, it is necessary to situate higher education in its immediate context, such as its relations to the state and the economy broadly defined as well as its users. By doing so, socio-political historical narratives can be developed for precise cross-national comparisons.
In this article, our aim is to develop such a socio-political historical narrative of Swedish higher education since 1945. To do so, we focus on how the Swedish case evolved in four salient aspects: (I) the transformations of the welfare state and the economy broadly defined; (II) the policies directly involved in redesigning the higher education system; (III) the scale, composition and structural relations between the parts of the higher education system; and (IV) the way different social groups (vis-àvis gender and class) have used the system. By comparing these aspects, we synthesise the history of higher education expansion in Sweden based on an otherwise rather fragmented scholarship, that nevertheless is crucial to understand the development of each aspect.
Our synthetic approach aligns with previous attempts to combine a historical account of a specific country's educational system with explanations of its development, whether economic (Bowles and Gintis 1976), cultural (Collins 1979) or multifaceted with attention given to the relative autonomy of educational fields (Bourdieu 1989). A complete socio-political modern history of higher education in Sweden is beyond the scope of this paper; however, we offer a preliminary contribution to such a narrative.
A useful tool for synthesising these four aspects is periodisation. The most famous periodisation of higher education system expansionthe transition from elite to mass to universal access (Trow 1972)has the virtue of accounting for the scale of the system and the consequences of its growth, but is too general for our ambition to understand the specificities of the Swedish case. Through a nuancing of the inner logic of each of the four aspects and their historical development, we are able to highlight converging trends and developments that seem to be incongruous with each other. In short, we can observe how the actions of different categories of actors are relatively autonomous vis-à-vis each other. For example, the social structure of higher education remained remarkably stable even when the system expanded enormously and radically shifted shape and function over time.
We begin by briefly describing the development of the omnipresent Swedish welfare state, the road to becoming one of the most equal countries in the world and the recent trends of rising economic inequalities. We then focus on higher education policy and the changing roles of higher education, the funding schemes for institutions and students as well as access to higher education. Following this, we characterise the Swedish higher education system in terms of its overall size, types of institutions and fields of study. Finally, we take the social uses of the higher education system into account, adding another layer to the relations between the types of institutions and fields of study. We conclude with a discussion of our main results.
From a social democratic regime in a well-organised state to rising inequalities and shifting powers Sweden developed dramatically over the twentieth century, from being a poor, largely agricultural-based economy in the northern periphery of Europe to generating one of the highest gross domestic products per capita and becoming well integrated in the global economy. Like industrial nations in general, the development of economic inequalities in Sweden followed a U-shaped curve during the twentieth century: huge inequalities at the beginning of the century, economic equalisation in the middle of the century and rising inequalities again from the 1980s (Piketty 2013). What made Sweden stand out from this general pattern was that it became exceptionally equal in terms of economic means and social insurance.
We divide modern and contemporary Swedish history from 1945 to 2020 into two different phases. First, a long economic boom took place between 1945 and 1973 which, politically, was a period entirely encompassed by the Social democrats' hegemony. The expansion of the welfare state over this period epitomised the social democratic welfare state in the sense of Esping-Andersen's typology (1990) and included policies aimed at economic redistribution and a universal social insurance system. Social and economic democracy were the prime objectives of Social Democratic policy, and their educational policy had a similar aim of removing financial obstacles so that everyone could continue beyond the level of primary schooling, first to secondary and then to higher education (Rothstein 1996).
One way of achieving this goal was by establishing a unified system of education, starting with the creation of a nine-year comprehensive elementary school system in 1962 that replaced a previously organisationally diverse elementary school system . A similar reform was launched for secondary schools in 1964, creating one integrated gymnasium with both preparatory programmes and technical education previously provided by technical institutes. Finally, in 1977, a unified higher education system was created.
The second period began because of the oil crisis of 1973 and the subsequent economic downturn, which exacted a sharp break with what had been long-lasting economic growth. From then on, the economy became more oscillatory, steady economic growth paired with setbacks in the early 1990s, 2001, 2008 and 2019. Even though the overall direction of the economy up until today has been growth, the consequences of the economic crises have been severe when considering unemployment rates. The crisis of the 1990s elevated unemployment levels from 2 per cent to 12 per cent in just a few years. Since then, unemployment rates have never fallen below 6 per cent, and spiked immediately in reaction to economic crises. Furthermore, the economy has also been restructured with a stronger focus on the service economy and a continual dismantling of the industrial sector. Paired with an increasing degree of digitalisation and automation of the economy, Sweden is often highlighted as a pioneering country with high levels of internet access, computer capacity and digital literacy, producing fertile soil for IT entrepreneurship that boasts companies such as Spotify, King and Klarna (Mauno Pettersson and Wisterberg 2018).
If Sweden ca. 1950-1970 epitomised the social democratic welfare regime in the sense of Esping-Andersen (1990), there are several reasons for questioning whether the same category could be used for its development since the 1980s. Politically, the hegemony of the Social Democrats was broken in 1976, after which political power has bounced between right-wing (1976-1982; 1991-1994; 2006-2014) and left-wing (1982-1991; 1994-2006; 2014-) governments. Political orientation overall has also shifted more towards the right and towards more neoliberal politicsa turn that the Social Democrats have participated inand right-wing governments have only further bolstered this shift. This has meant the introduction of market solutions for and the privatisation of the public sector, the deregulation of markets, the cutting and suspension of taxes as well as the engendering of larger social and economic differences. Deregulations of financial markets in the 1980s and the burst in housing markets altered the relationship between capital and income as well as the relationship between ownership and leasing. Since the 1980s, economic inequalities in Sweden have increased at a rate that is among the highest of the OECD countries. Nevertheless, it is important to acknowledge that Sweden was globally one of the most economically equal countries in 1980 and still is fairly equal from an international perspective (Therborn 2018;Bengtsson 2020).
A similar drastic political shift also occurred in Swedish educational policy. From an ideal of equal and equitable schooling, policy has shifted towards market-oriented buzzwords such as 'freedom of choice', 'efficiency' and 'transparency'. In the early 1990s, privatisation and market solutions were introduced in compulsory and secondary education. Since then, the number of private actors and the shares of students in both for-profit and non-profit independent schools has risen steadily (Forsberg 2015). Most strikingly in international comparison is that this expansion is publicly funded and is combined with the possibility for commercially oriented private providers to reap substantial profits from their activities. The tides have turned since the 1970s.

Higher education in political context
If we consider the most striking novelties and lasting legacies of the higher education policies that drove or accompanied the post-1945 expansion in Sweden, we can identify four eras: massification (1955-1977), unification (1977-1993), marketisation (1993-2006), and internationalisation (2007-). Each of these eras, delimited by years of major reforms, created distinct conditions for the higher education system to evolve. For example, the era of massification fuelled expansion by creating an automatic funding scheme for faculties (academic divisions) with open admissions. The subsequent era of unification brought restricted intake throughout the entire system, a nationwide centralisation of administration, and an incorporation of all post-secondary (and some secondary) education into the definition of higher education. Decentralisation of administration, however, including decisions on intake restrictions, was a hallmark of the marketisation era. Finally, the era of internationalisation saw the revision of the degree structure for entry into international student markets. 1955-1977: attempts to solve a looming crisis

Massification
In the decades following World War II, expanding higher education was deemed necessary for two main reasons: the rising number of eligible entrants due to an expanding secondary school and increasing demand from the labour market for professionals with academic degrees because of both the general technical developments of the mid-twentieth century and the expansion of public administration (SOU 1959:45). At the time, there was an open admissions policy to the faculties of theology, law and most areas of philosophy for anyone with studentexamen (higher certificate examination), a preparatory degree for university studies. In 1953, university admission was extended to groups without studentexamen, such as folkskollärare (elementary school teachers) and läroverksingenjörer (technical college engineers).
The first commission to explicitly target the problem of expansion was appointed in 1955. It was established to meet the rising demand for higher education that was already apparent in the early 1950s and that was expected to rise with increasingly larger cohorts of university age. The main issue was how to facilitate expansion, not whether higher education should expand. Two main solutions were proposed that would have significant implications for the extraordinary expansion and transformation of the university in the 1960s. First, the financing of higher education in the faculty of philosophy, which had mostly open admissions, was continuously tied to the enrolment of students. This principle, called universitetsautomatiken and modelled after a similar solution for the gymnasium, thus granted the universities financial conditions for growth (Andrén 2013). Second, a new position was introduced, universitetslektorer (university lecturer), with teaching as its main task, which improved the opportunity for institutions to expand their permanent teaching staff to meet new demands. Faculties with open admissions could now expand at pace with the demand from students.
Further actions were taken in the 1960s to facilitate expansion, such as a new system for student financial aid in 1965. Between 1939 and 1964, state-funded financial aid had been conditional on the aptitude of the student, as measured by studentexamen performance, their grades at their higher education institution and the economic means of the student, their spouse and their parents; aid consisted of a mixture of in-kind (1939) and rent-free loans (1918). When the new system was introduced in 1965, with grants and loans, it was still conditional on aptitude measured by their post-secondary academic performance and financial means, but only the means of the student itself. Further facilitating expansion, a new university was established in northern Sweden in 1965 as well as branches for the old universities in new geographical locations in 1967 (Askling 2012, 43-49).
Unification (1977)(1978)(1979)(1980)(1981)(1982)(1983)(1984)(1985)(1986)(1987)(1988)(1989)(1990)(1991)(1992): labour market orientation at the height of rational planning The 1977 reform turned out to be a pivotal moment in the modern history of Swedish higher education. It was the result of a long series of commissions, most of which were appointed in the 1960s (UKAS in 1966, PUKAS in 1968and U68 in 1968, and was the final piece in the social democratic project of creating a unified educational system at all levels. The universitetsautomatiken had led to very high cost for the state, and the watchwords for the commission of the Swedish Higher Education Authority (UKAS) in 1966 were 'efficiency' and 'throughput'. The solutions to these problems took the form of fixed study programmes at the faculty of philosophy in 1969but with some room for local variation -(Tengner 2006, 10-11), a new funding scheme of block grants and the introduction of a numerus clausus for all higher education in 1979. When UKAS initially presented the proposal of fixed study programmes in 1968, it met resistance, particularly from students. Later that year, a new commission, PUKAS, presented a revised proposal that was less rigid in terms of the fixed study programmes and allowed more freedom of choice than UKAS. However, less than ten years later in 1977, fixed study programmes were introduced throughout the entire higher education curriculum. Disciplines were to some extent replaced by these programmes, and the committees tasked with planning the curriculum were populated with faculty as well as external representatives. This shift has been described as a vocationalisation of higher education (Nybom 2012). To facilitate a closer link to the labour market, six regional boards were established to plan the educational offer at an intermediary level between the national agencies and the individual higher education institutions.
Financially, a new budget model was introduced with block grants in place of the previous automatic enumeration of resources based on investigated needs. This budgetary shift was accompanied by a clear demarcation between resources earmarked for education and for research respectively. Different fields of study were given different price tags and intake was restricted, leading to the numerus clausus in 1979. All of this was accompanied by a productivity deduction of 2%, based on the general increase in productivity (Gustavsson, forthcoming).
One of the most important and lasting contributions of the U68 commission and subsequent government bills was probably the delimitation and definition of higher education. Up until the 1977 reform, one spoke of universities and högskolor (university colleges). U68 suggested that some parts of the later years in gymnasium (upper secondary school) could be regarded as part of higher education (SOU 1973:2, 9-10). Thus, in 1977, a central admissions agency took over all admissions for the sector, and admissions criteria were to some extent harmonised according to the various forms of education now included in higher education. The educational system had now been administratively unified at all levels.

Marketisation (1993-):free-for-all competition
Towards the 1990s, the role of higher education was increasingly becoming a means for a variety of objectives, such as reducing social inequality and unemployment among young people and increasing democracy. After having been adapted to labour market demands in the 1970s, higher education institutions became 'development or innovation centers' in national or even regional economic policy (Nybom 2012). Previously, they were tasked with providing regional employers with labour, but now they should take on an active role in the regional economy. Higher education was increasingly treated as a commodity to be traded on a national market, and the 1993 reform of higher education laid a foundation for this development.
In 1989, the Social Democratic government appointed a new Higher Education Commission, whose final report was a foundation for the 1993 reform (SOU 1992:1). In between, in 1991, a right-wing government had come into power, but this political shift should not be overestimated. The market orientation and decentralisation that the reform contained had already been laid out by the Social Democratic government; the right-wing government only emphasised it. The 'freedom for quality' government bill (1992/93:1) that proposed the reform was the product of both a previously observed decline in the quality of education and research and a fear of its continuation. From a strategic standpoint, enhanced quality was to be combined with an increased efficiency in resource management as well as increased mobility and innovation. To reach these targets, a double strategy was developed: higher education institutions were to be given enhanced autonomy with regard to the organisation of studies, the educational offer, student admissions, professor appointments, etc. and such increased autonomy and institutional diversity were to be combined with incentives, evaluations and competition. Market logic should determine the relation between the supply of education and the demand from students; the shape and orientation of the educational offer and the resource allocation should be determined by student choice (Gov. Bill 1992/93:1: 21-22). Moreover, this reform also introduced the possibility of turning existing state institutions into private institutions, run by foundations. The general direction of the reform could be described as creating market-like conditions in higher education (Bauer et al. 1999: 85-88;cf. Clark 1983).
Financially, the block grant funding scheme of 1977 was still intact in 1993, but a significant novelty was the quality measures. From 1993, allocation of resources became based on student performance, and institutions could not receive any grants for failing students. The official idea behind this proposal was to provide incentives for increasing quality, but practically this budgetary scheme seemed to create an incentive to simply ensure that all students passregardless of the quality of their learning and performance. This was later adjusted so that financing was related to both enrolment and student performance.
It also has to be underscored that the increased autonomy of the higher education institutions obtained in 1993 was reinforced further in 2011 with the so called Autonomy reform (SOU 2007:98;SOU 2008:104;Gov. Bill 2009/10:149;Ahlbäck Öberg et al. 2016). Organisation structure and the professional categories were deregulated and it became possible for each institution to decide on its own internal organisation and its professional categories.

Internationalisation (2007-): reshaping the system according to international models
Despite the structural changes to higher education as a result of the 1993 reform, just 14 years later the entire system was transformed at its foundation. Such a radical change often only comes about as a result of severe crisis or strong external force; the latter was the case here, adoption of the international Bologna model.
Internationalisation of higher education has been a political ambition and priority since at least the early 1970s when a special commission was appointed to propose reforms and initiatives that would strengthen the internationalisation of higher education. While some internationally oriented educational programmes were established, the direct effects were limited, and additional efforts would not be made until the late 1980s and the early 1990s. Through introducing the opportunity to obtain loans for studies abroad in 1989, joining the European Union in 1995, and becoming part of the Erasmus programme in 1992, studies abroad both as freemover and exchange students were considerably enhanced for Swedes. In fact, this meant an internationalisation of the educational offer. Still, the Swedish system of higher education remained highly Swedish in most of its dimensions (Börjesson 2005).
This, however, would drastically change with the adoption of the Bologna model in 2007 (DS 2004:2;Gov. Bill 2004/05:162). Sweden had been part of the Bologna process since its initiation in 1999 and was able to restructure its higher educational system in less than a decade. While not shortening its four-year doctoral programme to the prescribed three years, Swedish higher education did adopt a three-tier system with a 3 + 2+4 structure. This stood in sharp contrast to the former two-tier system, where basic education ranged from two years to five-and-a-half-year programmes. As a consequence, shorter programmes became prolonged by a year, and longer programmes often divided into two programmes, a basic programme followed by an advanced programme, the latter of which became the master's level. Undoubtedly, master's programmes have become the prime vehicle for the internationalisation of higher education catering to an international student marketstressing that internationalisation should not be seen as a rupture of the market paradigm introduced in 1993, but rather a deepening of it (Rider 2014). This was further emphasised in 2011 with the introduction of tuition fees for international students from so-called third counties, i.e. outside of the EU/EES (Bryntesson and Börjesson 2019).

Expansion and reshaping of the institutional landscape
Overall number of students Swedish higher education has followed the global trend in expansion of the number of students. The number of registered students has increased from 14,000 in 1945 to 360,000 in 2019a 2,500 per cent increase. This expansion has not been successive, and it is possible to identify three phases of expansion paired with phases of slow increase, stagnation or even contraction in between and afterwards.
From 1945 to 1959, the first expansive phase took place with a slow but steady increase from 14,000 students to 33,000 students, representing a more than doubling in 15 years and a yearly average growth rate of 6 per cent. This phase was followed by an exceptional expansion phase in the 1960s, raising the numbers from 37,000 to 120,000 in a decade, a tripling of the number of students and an average annual increase rate of 12 per cent.
The subsequent period from 1970 to 1988, phase three, marked a rupture of the expansion, and even initially saw a decrease in student numbers, declining from 120,000 to 114,000, with a low of 109,000 in 1974. The jump in figures from 1976 to 1977, rising from 114,000 to 148,000 was a result of the 1977 higher education reform's changing definition of higher education. The trend of non-expansion from the previous years continued after the 1977 reform up to roughly 1988, when the numbers hovered between 155,000 and 160,000 students.
Starting in 1989, the numbers started climbing again and the rise lasted until 2003 (forming a fourth phase and the second phase of rapid expansion). The expansion was not as extraordinary as in the 1960s, but was still important: rising from 160,000 in 1988 to 340,000 in 2003, more than a doubling in 15 years and an annual growth rate of 5 per cent.
The fifth period started around 2004, when the previously continuous upward trend was broken and still continues today. It is defined by a wave-like increasing pattern with peaks in 2003 (340,000) and in 2010 (365,000) with a low in 2007 (319,000).
Four sets of factors are essential in order to understand this development: demographic, aspirational, economic, and political factors. First, the demographics explain certain parts of the pattern. The expansion during the latter half of the 1950s and the first half of the 1960s coincided with a rise in the number of twenty-year-olds by more than 60 percent from 1952 to 1965 (cf Ohlsson 1986). Additionally, the phase of stagnation during the 1970s and 1980s was paralleled with a demographic plateau, and the period after 2004 also shifted in accordance with the demographic trend. However, the major expansion from 1989 to 2003 in fact occurred while the number of twenty-yearolds decreased by 20 per cent in little more than a decade, from 1986 to 1998.
In this way, demographics do not fully account for the expansion, stagnation and retraction; we also need to take into account the demand for higher education studies. This can be understood as containing two parts: the number of eligible students and the willingness among those eligible to apply for and enrol in higher education. After World War II, upper secondary education started to expand, continued to do so throughout the twentieth century and reached a point in the late 1980s where almost everyone started upper secondary education. In addition, aspiration levels increased gradually, and this factor came to play a major role in the expansion during the 1990s.
It is very clear that enrolment patterns are shaped by economic development, but two different and opposing logics can be identified. During a first phase, from 1945 through the 1980s, the pattern is that a positive economic development correlated with the expansion of higher education. During the 1970s, economic recession and stagnation occurred. However, from the 1980s and onwards, the logic shifts; higher education thereafter expanded in times of economic crisis. This is true for the 1990s, the Dot-com burst in 2001, the financial crisis in 2008 and now the COVID-19 pandemic. In times of economic prosperity, there is less expansion and even stagnation (1980s) and retraction (2010)(2011)(2012)(2013)(2014)(2015)(2016). It appears that when the system has reached a certain level and is approaching universal access, the focus shifts from a general expansion to a compensatory expansion.
In sum, given the developments in Swedish higher education from 1945 to 1970, it is obvious that in the Swedish case the massive expansion that occurred over this period was driven by a very strong political will to expand the system. At the same time, these efforts also implied that the politicians lost the possibility to actually control the level and direction of the expansion. The higher education politics of the 1960s and 1970s was dictated by a will to take a firmer grip on expansion, which, to some extent, became even more precious as the economic foundations for the politics weakened with the crisis of the 1970s.
Further, the expansion during the 1990s was highly driven by political rationales, but of different sorts. Higher education became a crucial means for society to combat high unemployment rates among the younger generation but also was a way to restructure the competence of the labour force and render it more compatible with the demands of the new economy. Since the government had the possibility to decide the overall size of the system by setting goals for each higher education institution (cf. Jackson 2019), it was relatively easy to steer investments towards regions severely hit by the economic crisis and high unemployment rates as well as towards fields of study that were considered strategic for the country. It is thus not enough to study the development of the system as a whole, but more detailed analyses of the expansion in terms of types of higher education institutions and fields of study are called for.

Unification under one system amplifying institutional differences
When attempting to solve the problem of how to expand higher education capacity for an increasing number of students, the discussions and solutions concerned two issues. One was whether to expand existing institutions or to establish new ones, and the other issue was whether to continue with the Humboldtian legacy and combine education and research or create an institutional division of labour. Although the paths taken often contained elements of each of these model solutions, we can clearly see how one solution dominated in each epoch.
At the beginning of the era of massification, ca. 1945, the notion of higher learning encompassed universities (Uppsala, Lund, Karolinska), university colleges (Göteborg [university in 1954], Stockholm [1960]) and specialised institutions (polytechnics, etc.). Between establishing new or expanding existing institutions, the course of action was to expand the existing institutions. However, this did not meet rising demand, and in the 1960s, a new university in the northern city of Umeå was established in 1965 along with regional branches for the other four universities in 1967 (Göteborg Uni-versity→Karlstad branch, Stockholm→Linköping, Lund→Växjö, and Uppsala→Örebro). The regional branches were all about undergraduate education, with no direct funding for research and with universitetslektor as the most senior faculty (Agevall and Olofsson 2013). However, the idea of creating branches rather than independent university colleges was to keep with the Humboldtian legacy of combining teaching and research through their organisational connection to their main universities. The era of massification saw a ten-fold increase in the number of students at universities (and their branches) in 1945-1975, while the specialised institutions saw a five-fold increase.
As higher education was administratively unified in 1977, institutional differences within the system crystallised and later amplified. While existing institutions continued to expand, the main solution was to establish new institutions: the undergraduate university colleges. One of the university branches was upgraded to a university (Linköping), and the others were upgraded to university colleges. In addition, ten new university colleges were established, often on the basis of institutions already in existence, such as training colleges for elementary school teachers, nursing colleges and schools of social work. Just like the branches of the previous era, the new university colleges were not allowed to have direct funding for research. In a sense, the era of administrative unification marks a clear break with the Humboldtian legacy, and an institutional division of labour between research universities and undergraduate university colleges was cemented.
The university colleges have, ever since their creation, strived to obtain direct research funds, either by advocating for policy change or by obtaining university status. The era of marketisation would see both of these strategies realised. The bill on research policy for the years 1997-1999 awarded university colleges direct funds for research in order to strengthen the ties between education and research (Gov. Bill 1996/97:5). Six colleges were awarded university status between 1995 and 2018, with an additional one announced for 2022. These developments suggest a slight return to the Humboldtian legacy. Since the marketisation reform in 1993, only three new institutions have been established (Södertörn, Malmö, and Gotland). Thus, for the most part, increased capacity to mitigate the monumental expansion that began in 1989 was created by expanding existing institutions. This time, it was the university colleges and specialised institutions that expanded the most by tripling their number of students, while universities merely doubled in size.

Shifting balances between fields of study
With expansion as a consistent feature of Swedish higher education since 1945, most fields of study have experienced excellent conditions for growth. Any periodic decline, if at all, has only been relative to other fields. The most transformative period for fields of study was the era of massification. In 1945, the social sciences made up about 10% of students in the academic division they shared with the humanities (SOU 1943:17). As a consequence of this division having open admissions, both the social sciences and the humanities displayed faster growth than the overall expansion of the universities. The social sciences even rose to become the numerically dominant area of study (Dalberg, Börjesson, and Broady 2019).
Since the era of unification, expansion has remained relatively consistent within each broad field of study (Börjesson, Bertilsson, and Dalberg 2014). The social sciences have been the largest area of study, accounting for around 30% of all students in higher education. The humanities have been in relative decline, as have health and (teacher) education, while engineering and natural sciences have increased their share of the student population. These latter fields have since the 1950s been subject to various and massive recruitment campaigns, implicitly or explicitly proclaiming that the welfare of the nation rests on an adequate supply of engineers and natural scientists (Lövheim 2016).
The stability on the level of larger areas of study, however, hides significant evolution on the level of disciplines. Within the social sciences, business studies has become completely dominant, while computer science and computing have become the most numerous fields of study within the natural sciences and engineering, respectively. With recruitment plummeting in the years following the burst of the Dot-com bubble in early 2000, the close association between labour market trends and the fields of computing highlights a characteristic of the marketisation era: student demand, not centralised planning, should determine enrolment and allocation of resources.

Differentiated uses of higher education
Reducing social inequalities through participation in higher education has been a consistent feature of educational policy since 1945. Widening access to include underrepresented social groups and geographic regions was the main objective of policies targeting social inequalities from the 1950s to the 1970s. In the 1980s and 1990s, the focus shifted to gender inequality (Schenk 2005). There has indeed been an overall increase in participation that has been transformative in terms of new groups increasing their share. The rise of women since 1977 as the undergraduate majority is a case in point. However, when we consider the relative chances of participation conditional on social class, or the horizontal stratification between institutions and fields of study in terms gender, social class and academic achievement, we observe remarkable stability. These significant structural continuities are often ignored, especially in the case of Sweden and the other Nordic countries, which are frequently perceived as egalitarian islands in an unequal world. This is also why we emphasise the inertia of inequality by proposing that social stratification should be identified as almost a constant in an otherwise changing system of higher education.

Increased participation, maintained inequality
Higher education expansion has undeniably led to a general increase in participation implying a transformation of the whole system in terms of Trow's classification from an elite system educating up to 15 per cent of a birth cohort, which lasted up to the end of the 1950s, to a mass system established during the 1960s with levels reaching up to 25 per cent of a birth cohort, expanding further during the 1990s, when the levels passed 40 per cent, towards finally the stage of universal expansion in the 2000s (Melldahl 2015). Less evident, however, has been the claim that previously overrepresented groups have maintained their advantage relative to other groups across cohorts, even as higher education has expanded. As the theory of maximally maintained inequality predicts, equalisation of participation usually happens when the rate of participation from advantaged groups reaches saturation (Raftery and Hout 1993). Equalisation in participation did occur in Sweden and was most pronounced for birth cohorts 1925-1950, but then stabilised for cohorts born 1950-1970(Jonsson and Erikson 2007. Evidence from a recent study suggests that inequalities are being maintained, albeit at lower levels than before the massification (Thomsen et al. 2017).
Relating the trend of inequalities to the overall expansion of the system and policy context, we note that the decrease in inequality for birth cohorts expected to enter higher education in 1943-1968 took place when the expansion of higher education was most rapid (1950s-1960s). This also coincided with the transition from meanstested to universal student aid in 1965. As the expansion halted in the 1970s, the competition for the positions increased, and the decrease in social inequalities stalled. Once the system expanded again in the 1990s, there seemed to be decreasing differences again, but not at the level of altering the fundamental social differences in the system. However, when it comes to gender, the story differs.

The rise of women
A general trend in higher education systems over the last quarter of the twentieth century was the reversal of gender inequalities (Vincent-Lancrin 2008). Sweden was perhaps an early indicator of this trend, with the gender balance tipping in the favour of women in 1977just over a century after women as a group were allowed to even be admitted to (parts of) the university in 1873. This shift was due to the redefinition of higher education (discussed above) that year whereby former non-university, and predominantly female, degree programmes were awarded higher education status; male-dominated specialised institutions such as the polytechnics (KTH, Chalmers) were already considered parts of the higher education system prior to 1977.
Although the institutional upgrade through redefinition provides a definitive date for the gender reversal in Swedish higher education, it is only a minor aspect in the overall context of gender in higher education. As indicated by Figure 1, women expanded their participation more rapidly than men over the entire period, meaning expansion was thus primarily driven by women. This also implies a shifting balance in programmes where both genders compete. For elite education, most programmes were still male dominated directly after the 1977 reform, but today most elite programmes are female dominatedengineering programmes being the exception (Lidegran et al. 2014). This indicates that besides the radical shift in the gender balance from the 1977 reform, there are no distinct breaks in the trend of increasing shares of women, making a periodisation plausible.

Classed institutions, gendered fields
Patterns of social stratification in higher educationsometimes referred to as horizontal stratificationhave proved remarkably rigid since at least the late 1970s (Broady and Palme 1989;. That is, during the same timeframe as equalisation stalled along the vertical dimension of socially stratified participation in higher education, sons and daughters of different social origins sorted into different institutions and fields of study in strikingly patterned ways. The most distinct features of this social structure may be summarised as classed institutions and gendered fields: institutions being differentiated according to the over-and underrepresentation of students from different social classes and fields of study being differentiated by gender.
The social differentiation of institutions distinguishes the oldest and largest universities along with the specialised institutions from the younger and smaller universities and university colleges; students originating from the upper-middle and upper class are overrepresented at traditional universities and specialised institutions while working class students are overrepresented at university colleges. In other words, the oldest, largest and most resource-rich institutions manage to recruit a disproportionate amount of the students with the largest amount of social and academic assets.
Fields of study are highly segregated by gender. Men are overrepresented in engineering, technology and natural science; women are overrepresented in healthcare, education and languages; and many social sciences are fairly gender balanced. These patterns are common around the world (Charles and Bradley 2009;Barone 2011). What the Swedish case shows, however, is the interaction of class and gender with institutions and fields of study. Gender segregation is most pronounced among the sons and daughters of the working class: the sons take shorter engineering programmes or computing while the daughters study to work in social services or kindergarten teaching, typically at more geographically peripheral university colleges. At the opposite end of the social hierarchy, sons and daughters from the upper-middle and upper class with high grades meet as equals at longer professional programmes in medicine, law or economics at the Karolinska Institute, Uppsala University, Lund University or the Stockholm School of Economics.

Discussion
The Swedish higher education system provides an interesting case for examining the higher education system's relation to the state and the economy; its relationship to politics; its own structure, composition and size as well as its uses by social groups, men and women, older and younger. We have in this article tried to sketch these as four separate histories with their own periodisation and dynamics. In the following, we will synthesise these histories to each other and draw some general conclusions. We will also discuss the Swedish case in an international context in order to highlight its particularities and communalities.
For the higher education system, we have defined and labelled four larger reform periods: massification (1955-1977), unification (1977-1993), marketisation (1993-) and internationalisation (2007-). Massification laid the ground for the enormous expansion that occurred throughout the 1960s. Unification enlarged, unified, bureaucratised and centralised the system and took control of the expansion rate by introducing numerus clausus in the whole system. The marketisation period meant a sharp break with the former and shifted towards decentralisation, deregulation and marketisation, and to some extent also privatisation, which all implied the introduction of new public management into higher education; a paradigm that is still in place. The fourth transformed the Swedish higher education system from being a nationally designed system to an internationally modelled one in line with the Bologna process. It did not change the steering and financing of the system, but rather altered the composition of course content, programmes and degree structure.
How can we understand these reforms in terms of continuation, stability and path dependency on the one hand, with change, disruption and fracture on the other? Two interpretations are plausible. One interpretation puts a sharp dividing line at 1977. Prior to 1977, the Swedish higher education system could be described as a dual system with power at the bottom in the hands of the professors and the departments as well as at the central level of the ministry and the government. With the 1977 reform, more power was concentrated at intermediary and bureaucratic levels with the introduction of regional boards and the central steering of the educational offer. In such a perspective, the 1993 reform decentralised the power further to the level of the higher education institutions and their leadership, with neither the professorship nor the central political bodies receiving more mandate than the other. This line of development was then accentuated with the Autonomy reform in 2011 which distributed even more power to the higher education institutions, and decreased the autonomy and power at both the base and the top levels.
A second interpretation puts the focus on the type of steering that is built into the system. Then the dividing line can be drawn at 1993 with the shift from political to market-oriented steering. Market solutions where introduced to accommodate the educational offer, which should be oriented towards demand. This is still in place, and this market orientation has increased with internationalisation, opening up the Swedish system to a larger European and global market of higher education.
Our suggestion is to not prioritise any interpretation over the other, but to regard them as complementary and adding to our understanding of the modern history of higher education. In fact, on a policy level the two dimensions are related in the rhetoric: internationalisation and competition on a global market are argued to create an increasing need for more autonomous higher education institutions that can act as strategic actors. However, the empirical evidences for this link are sparse.
Furthermore, the size of the system, here measured in the number of students, needs to be added to the picture. Swedish higher education, as many other higher education systems, has expanded drastically from 1945. The conditions for the expansions differ radically between different phases (see above). The first two phases coincided with the long economic boom after World War II, while the third was triggered by the economic crisis of the early 1990s and its historically high levels of unemployment. We also identified a shift in logics where until around the end of the 1970s, expansion was related to a positive economic trend, whereas afterwards, expansion became more closely tied to economic downturns and crisis.
The transition from an elite system towards the stage of universal expansion in the 2000s highlights that expansion of the system implies its opening up towards new groups, such as women; the middle, the lower-middle and the working classes; and immigrantsincreasing and spreading the use of higher education in society. Furthermore, such an expansion and transition of the system has been conditioned by a strong political will to expand along with an input of massive economic resources, augmenting demand for higher education in the population as well as in the labour market. For the system, this has meant an increase in the number of institutions, but especially a growth of the number of students per institution, a geographical spread of the educational offer, and a more diversified provision of education, by types, content and orientation.
Our conclusion here is that as the system has expanded, many strict boundaries have been blurred and even dissolved, thus resulting in a more complex and opaque system. From a differentiated system organised around types of institutions, the expansion of the size of the individual institutions has increased the complexity within each institution.
It is also striking that while the higher education system has expanded enormously and also radically shifted shape and function over time, the social structure remains remarkably stable. It is true that women have increased their participation much more than men have, but the basic gendered preferences for fields of study persist. Further, the share of working-class offspring has increased, but the social differentiation according to types of institutions and programmes has not been altered. This implies a strong preserving character of the system, linking it to an increased importance for mediating social reproduction.
Finally, positioning the socio-political history of Swedish higher education in an international context, we put forth Sweden as an example of a country that over time has become less national and more internationally oriented with the adoption of the Bologna model, the increased focus on institutional autonomy and the larger diffusion of market solutions. This is, of course, not unique for Sweden, but certain traits may be. Even if it is doubtful whether we can continue to label Sweden a social democratic welfare state, we would argue that the heritage of the social democratic welfare state model clearly overshadows its recent development. We still see a large focus on widening participation in higher education, less direct cuts in the state budget for higher education as well as rather substantial investments due to the ongoing economic crisis. It will be interesting to follow the current development and see if the recent focus on big governance will increase the political steering of higher education, or if the substantial critique of new public management will lead to more professional autonomy or finally if the trend towards more institutional autonomy will be intensified.