Have the Welfare Professions Lost Autonomy? A Comparative Study of Doctors and Teachers

Abstract This article investigates the impact of NPM reforms on two prominent welfare state professions; medical doctors and teachers. The case study context is Sweden, where the impact of NPM led to a series of reforms in health care and education after 1990. The focus in the paper is on professional autonomy, which can be seen as a core trait in professional work. The findings in the article point to both medical doctors and teachers having lost professional autonomy as a result of NPM-reforms, particularly with regards to the dimensions of work organization and evaluation. Autonomy in individual decision-making seems to have been least affected, even if there are indications of this being infringed upon as well. Despite these broad similarities, the loss of autonomy is found to be more pronounced in the case of teachers. One reason behind this difference, which manifests itself in teachers having become subject to higher level of direct administrative control by school managers while at the same time experiencing a more distinct loss of autonomy to evaluate their work, appears to be that medical doctors have been more successful in establishing themselves as experts in relation to new public audit agencies.


Introduction
The history of the welfare state is intertwined with the development of professions.Medical doctors, nurses, teachers, psychologists and social workers are all examples of professions whose work has made the development and delivery of high-quality welfare services possible.The relationship between the state and the welfare professions has sometimes been described as an implicit contract, where the state grants the professions a high level of discretion and self-organization in return for which the professions provide their expertise in service of the public interest (Mintzberg, ; Cousins, ).Many have argued, however, that something changed in the symbiotic relationship between the welfare state and the welfare professions after  when the New Public Management (NPM) movement gained traction.Within this movement, the authority and independence of professions were often portrayed as harmful, leading to waste of public resources, inefficiency, and neglect of the true interests of welfare users (Clarke and Newman, ; Exworthy and Halford, ; Flynn, ).Consequently, reforms inspired by the NPM doctrine, such as more assertive administrative management, the introduction of economic incentives to increase cost awareness, strengthening of the rights of users, and new forms of performance measurement often seemed to imply a diminishing of the powers of professionals (Pollitt and Bouckaert, ; Kirkpatrick and Ackroyd, ; Freidson, ).Over the last decades, a wide range of studies in different countries and policy areas have found that the authority of professions has in fact been reduced as a result of NPM reforms (Dent, ; Brandsen and Honingh, ; Healy, ; Heffernan, ; Kuhlmann, ; Kirkpatrick et al., ; Broadbent and Laughlin, ; Flynn, ; Maroy, ; Gunter et al., ; Blomgren, ).A commonly identified pattern is that professions have been subject to stricter managerial oversight and stronger cost pressures, while their work is more often evaluated by outside actors such as audit agencies.While such developments may not be altogether negative, they carry the risk that the specific knowledge, experience, and ethics that have been associated with professional authority becomes devalued (Ahlbäck Öberg et al., ; Connell et al., ; Freidson, ).
The view that NPM reforms represent a threat to, even attack on, professions within the welfare state is widespread.Nonetheless, there are also those who claim that this concern is over-stated and that professional groups in practice often adapt successfully to the organizational changes associated with NPM (Evetts, , ; Levay and Waks, ; Noordegraaf, ).Others have pointed to a variety in how professional groups have been affected by NPM reforms, with some professions appearing more resilient than others, and some even gaining status and prestige in their wake (Moberg et al., ; Jarl et al., ; Ackroyd et al., ).Several researchers have argued that in order to better understand the impact of NPM reforms on professions, more studies are needed which compare developments across professional fields.This would make it easier to identify what role is played by the specific organization and knowledge basis of professions when they try to adapt to new managerial logics (Ackroyd et al., ; Kirkpatrick et al., ; Ferlie et al., ).
In this article, we respond to the call for comparative studies by investigating the impact of NPM reforms on two prominent welfare state professions; medical doctors and teachers.Education and health care also constitute relevant cases for comparison in that they are both sectors were NPM have had a pervasive influence on policy-making and organization during the last decades, both nationally and at the international level.The case study context is Sweden, where       the impact of NPM led to a string of reforms in health care and education after .Starting off with a public, bureaucratic, welfare system of 'social democratic' type (Esping-Andersen, ), NPM in Sweden led to a break-up of the de facto public monopolies in health care and education and paved the way for more pluralist systems where competition and user choice have come to play a significant role.Reforms started with decentralized budget responsibilities and purchaser-provider splits in the s, moving on to quasi-market competition and various forms of user choice.In the  and s, a "third wave" of NPM reforms have centred on quality measurement and new forms of audit.The fact that NPM reforms have followed largely similar trajectories in the health care and education, which are both services provided through local authorities, makes the Swedish context particularly suitable for comparing the two sectors.
The research questions addressed in the article are: to what extent doctors and teachers in Sweden can be said to have lost professional autonomy as a result of NPM reforms; and whether there are any differences in how such effects have manifested in their respective professional domains.The focus in the paper is on professional autonomy, rather than other forms of professional power.This is motivated by the fact that professional autonomy, or the ability of self-management and autonomous decision-making, has typically been seen as a core trait in professional work (Abbott, ).Three dimensions of professional autonomy are identified in the article: autonomy in work organization, autonomy in individual decision-making, and autonomy in work evaluation.The findings in the article point to both medical doctors and teachers having lost professional autonomy as a result of NPM-reforms, particularly with regards to the dimensions of work organization and evaluation.Autonomy in individual decision-making seems to have been least affected, even if there are indications of this being infringed upon as well.Despite these broad similarities, the loss of autonomy is found to be more pronounced in the case of teachers.An important reason behind this difference, which manifests itself in teachers having become subject to higher level of direct administrative control by school managers while at the same time experiencing a more distinct loss of autonomy to evaluate their work, appears to be that medical doctors have been more successful in establishing themselves as experts in relation to new public audit agencies.Another finding is that the introduction of user choice appears to have had stronger de-professionalizing effects in the educational sector in Sweden, where increasingly competitive school markets have led to parents gaining a stronger voice.
The paper is organized into five sections.In the first, the theoretical concepts of professionalism and professional autonomy are reviewed, identifying three dimensions of autonomy.The second section presents the research methodology.The third and fourth sections turn to the empirical case studies, first the medical profession and thereafter the teaching profession.The last section summarizes and discusses the findings.

Professional power and autonomy
A profession is usually understood as an occupational group that displays certain characteristics such as a distinct status, or position, in relation to other occupational groupsspecifically, a strong identity and a relatively high amount of independence in organizing their work (Brante, ; Torstendahl and Burrage, ).Professions are characterized by requiring relatively long, university-based education, which implies that the knowledge professionals possess has an abstract rather than practical base.They typically claim monopoly jurisdiction over their area of expertise, which means that others cannot do the same work without explicitly belonging to the profession through a formal certification or licensing.Strong professional organizations are often central in upholding such controls, which also limit the number of people that are educated into a certain profession in a society.In addition, professionals are usually able to maintain a high level of discretion in their work, meaning that the professional has the mandate to use his or her own judgement in deciding how work tasks should be executed (Abbott, ).Archetypical professions are medical doctors and lawyers: "classical" professions that emerged in the wake of early scientific discoveries and industrialization in the th century.In the th century, the rise of the welfare state created new demands for professional knowledge.Examples of professions that developed during this period are teachers, nurses, psychologists, and social workers (Brante, ).
Professions are not only distinct in and of themselves; they also shape the organizations in which they work.A professional organization can be described as an organization where work is discretionary rather than standardized; leadership is exercised through collegiality rather than hierarchy; and the quality of the work is evaluated by the professionals themselves, or their peers, rather than by administrative superiors (Kirkpatrick and Ackroyd, ; Freidson, ).According to Freidson, the professional organization logic can be distinguished from two other ideal types of organizational logics, the bureaucratic and the market-based.A bureaucratic organization is based on values such as legal order, administrative hierarchy, standardization of work routines, and rationalistic pursuit of specified goals.A market-based organization is oriented towards satisfying consumers, or users, while at the same time containing costs, thereby generating profits to owners.In practice, most organizations have elements of all three organizational logics, even if their relative importance differs (Friedson, , ).
The broad impact of NPM after  has led to a strengthening of both the bureaucratic and market-based organizational logics in many welfare sectors.NPM is commonly understood as a collection of ideas about how public organizations can be made more efficient by introducing steering logics inspired by the market, such as decentralization, cost-effectiveness, competition, performance measurement and clear lines of managerial authority (Laegreid, ; Hood, ).NPM has tended to enforce bureaucratic elements through its emphasis on administrative authority and leadership (Kirkpatrick et al., ).Another example of bureaucratic steering logic strengthened through NPM is its orientation towards measuring performance and the demand that all activities within an organization be documented for audit purposes (Bringselius, ; Power, ).Market-based steering was introduced in welfare sectors in many countries in the s with the creation of quasi markets or user choice systems.In such systems, self-governing organizational units (for instance, schools, nursing homes, or health clinics) are forced to compete for public funding through contracts with public agencies or on the basis of user choice (Le Grand, ; Le Grand and Bartlett, ).
While it has often been argued that NPM has led to a decline in professional power, or de-professionalization, the exact meaning of this term has remained somewhat elusive.While most agree that professional groups in many areas have been "pushed back" or lost authority as a result of NPM reforms, there is less agreement on how this has manifested and how pronounced the tendency is.Many studies have described reforms that have strengthened managerial authority over professions, introduced more detailed regulation of their work practices through protocols and guidelines, and paved the way for new forms of external quality evaluation (Hall and McGinity, ; Scott, ; Exworthy and Halford, ; Dent, ; Broadbent and Laughlin, ).Examples of professional groups having lost authority as a result of NPM reforms also include the introduction of price competition, forcing upon professionals a logic of cost-efficiency when making decisions, and the strengthening of consumer power, resulting in professionals having to try to attract new users though marketing events (McDonald et al., ; Kirkpatrick et al., ).Other studies have described how professionals, rather than succumbing to NPM-related organizational changes, have opposed them or found ways to re-assert their power in the new organizational landscapes in which they have found themselves (Bezes et al., ; Thomas and Hewitt, ; Levay and Waks, ).Some, like Evetts, note that professions have responded to NPM by developing new skills, such as those related to management and performance evaluation (Evetts, ).This, Evetts argues, can be understood as a new form of professionalism rather than a decline in professional authority.Nordegraaf, similarly, talks of the tendency of welfare professions to develop into "hybrids"where some groups enter into managerial positions, thereby fragmenting professional identities (Noordegraaf, , ).The hybridization tendency can thus be seen as a form of professional adaption to new managerialist logics threatening traditional forms of professional control.The      ? question remains open however if all welfare professions are equally capable of developing into hybrids and whether such developments strengthen or undermine their position in relation to other occupational groups (Croft et al., ; Bode et al., ; Noordegraaf and Van der Meulen, ).
Even if there are many traits associated with professionalism, it has often been argued that autonomy is at the heart of professional power (Johnson, ; Abbott, ; Friedson, ).Professional autonomy refers broadly to the ability of a professional group to organize its work as it sees fit without interference from others, and can exist at several different levels, such as the societal level, the work-place level, and the individual level.Another distinction has been between external and internal autonomy, where the first, external autonomy refers to control over the organization of workfor instance, where and when work tasks are carried out, by whom, and in what way.Internal autonomy 'refers to the ability of individual professionals to make independent decisions, or judgement calls, in the course of their daily work, a characteristic also referred to as self-governance or discretion (Ballou, ).Discretion has been argued to be essential not least in the area of welfare services, which often involves decisions regarding individual needs or the allocation of resources between individuals and groups.To confer the right to make such decisions to the professionsfor instance, in areas like health care or social workreflects the high level of trust that policy makers and citizens traditionally have had in the professions' judgement and ethics.A third dimension of professional autonomy is discussed by Ahlbäck Öberg et al. (), who highlight the right of professions to decide for themselves, on the basis of shared professional knowledge and norms, what constitutes 'a job well done'.This autonomy in evaluation, or interpretation of what constitutes quality within a professional field, can be said to be another essential component of professional autonomy both in the work place and on a societal level.In the empirical case study below, the three dimensions of professional autonomy (autonomy over work organization, individual autonomy, and autonomy over evaluation) are used as a framework structuring the examination of the impact of NPM reforms on the medical and teaching professions in Sweden.

Methodological approach
The main method used in this article is comparative case studies.The empirical study investigates and compares the effects of NPM reforms on professional autonomy in health care and education, targeting the main professions in these areas: medical doctors and teachers.A comparative case study is suitable when a predominantly qualitative research method is used to try to describe relations between causal factors and outcomes and when it is important to contextualize the analysis of how interactions between causes and effects are connected (Bennett and Elman, ).The case study method is also appropriate in this case as it makes it possible to draw on a variety of sources to try to understand how different types of NPMs reforms at various points in time came to influence professional work (Kaarbo and Beasley, ).The main empirical source used in the study is previous research, complemented with sources such as public laws, public committee reports and investigations from professional associations.Following George and Bennett, the case studies are organized as structured, focussed comparisons (George and Bennett, ), which means that they are structured in similar ways so as to enable systematic comparison, while certain aspects of the cases focussed on, in this case, professional autonomy.
The comparative case study method uses case selection to try to isolate causal effects.The fact that both cases are set in the same national context (Sweden) makes it possible to hold constant several factors that typically vary when professions in welfare services are compared across countries, such as national legislative frameworks or basic type of welfare system.Health care and education are two very different public services, but in Sweden, their organization share several similarities.First, they are both public services, which means that the overwhelming majority of both doctors and school teachers in Sweden are public employees.In both areas, services are tax-financed and provided to all citizens by local, self-governing bodies; in the case of health care,  regions, and, in the case of education (primary, secondary, and upper secondary),  municipalities.Secondly, NPM reforms in the two sectors have followed a similar trajectory, with early reforms in the late s being based on the management by objectives (MBO) doctrine; a second wave being initiated in the early s promoting marketization through competition and choice; and a third wave emerging after  which introduced quality management.This implies that the Swedish case context can be seen as relatively favourable for a comparative study of the effects of NPM on professions in health care and education, as it reduces some of the differences that often exist between these sectors, such as with financing, employment, and reform content.A difference between them however is that the medical profession is what is sometimes called a "classical" profession, established already in the  th and  th centuries, while teachers became professionalized during the  th century (Brante, ).With the development of the welfare state both professions became integrated within public institutions but maintained independent associations and strong professional identities.Even so, the medical profession as a classical profession is usually regarded as having higher status and prestige then more recent welfare professions such as teachers, social workers and nurses (Suddaby and Muzio, ; Brante, ).This difference can thus be considered a factor that varies between the cases, making it a possible explanation for differing outcomes with regards to professional autonomy.A first wave of NPM reforms were introduced in Swedish health care at the beginning of the s.They brought organizational changes such as devolution of budgetary and administrative power to clinics, purchaser-provider splits, management by objectives, and performance-related payment systems (Anell and Svarvar, ).In some cases, like the Stockholm region, so-called internal markets were introduced whereby hospital clinics and health centers were financially reimbursed based on their performance.Another part of this first wave of reforms was a new focus on management and leadership at the clinical level.A new regulation on clinical management in  strengthened the powers of clinical managers while at the same time removing the previous rule that only doctors could assume this position.The second wave of NPM reforms in healthcare was more focused on competition, privatization, and patient choice.After a legal change in , private providers were allowed to establish freely in the formerly almost all-public system (Saltman and von Otter, ).Making user choice a key part of their political reform agenda, a right-wing government that took office the same year launched the so-called General practitioner reform (), which made it easier for individuals to choose freely among competing GPs.This reform gave patients the right to choose a GP or primary care center all over the country, while at the same time improving conditions for private care providers.During the s, the steady rise in private care providers within the publicly financed and regulated system continued.In the mid-, about  percent of all primary care providers were private (Isaksson et al., ).
A third trend in Swedish health care that can be linked to NPM reform is new ways of enhancing and evaluating medical quality.Quality management in the health sector has been introduced both at the national level by the National Board of Health and Welfare (NBHW), and regional level through new regional quality evaluation systems.Quality management has also been used to steer the physicians in their clinical work through the introduction of guidelines for diagnosis and treatment choices.Another reform has been collection and publication of clinical outcome data through the so-called medical quality registries (Fredriksson et al., ; Levay and Waks, ).The clinical guidelines are issued by the NBHW, but medical experts are consulted when they are developed (Fredriksson et al., ).The quality registers are managed by the Swedish Association of Local Authorities and Regions (SALAR), together with medical specialist associations.
Taken together, the NPM reforms introduced in Swedish health care have affected the working conditions for the medical profession in numerous ways.Due to the increased number of private care providers, doctors have been given more opportunities to choose private employers, which can be seen as a way to increase their professional status.At the same time, there has been a       strengthening of the administrative management at the clinical level and introduction of new cost pressures in the form of financial reimbursement systems.The role of patients as consumers of care has been strengthened through choice models, which tie the choices of patients to financial flows within the system: for instance, in the form of listing in the primary care sector.Last but not least, the quality management reforms have led to new demands on doctors to document their work and report clinical results, leading to a substantial increase in administrative work (Funck, ; Gadolin and Andersson, ).
Regarding the impact of the NPM reforms on doctors' professional autonomy, the findings are mixed and partly contradictory.Starting with autonomy over work organization, there appears to be a clear negative impact, not least through doctors being forced to spend a large amount of their time documenting their work and taking on new administrative tasks.Several surveys of doctors in Sweden have shown that they experienced a loss of influence over working conditions following managerial reforms in the s (Brante et al., ; Härenstam and Bejerot, ; Bejerot and Hasselbladh, ; Bejerot et al., ; Forsberg et al., , ).On the other hand, doctors who took only new administrative positions, such as clinical managers, reported a higher level of influence over their work (Petterson, ).The introduction of performance-based financing systems appears to have had a particularly strong impact in that it forced doctors to report all treatment decisions, leading to a general sense of bureaucratization (Charpentier and Samuelson, ; Kastberg and Siverbo, ).Autonomy over the organization of work was also challenged by new forms of clinical leadership.A survey of doctors showed that the number of leadership positions held by them was significantly reduced between  and  (Bejerot et al., ).Autonomy over the organization of work appears to have been less affected by the patient choice reforms.Even though these reforms were meant to lead to competition for patients, studies suggest that doctors were relatively unaffected by them (Winblad and Andersson, ; Winblad, ).
The individual autonomy of doctors has clearly been affected by the NPM reforms as well but it is hard to determine how much.Performance-based payment systems have created new incentives to contain costs and prioritize differently between patient needs.Several studies have found that doctors felt that they had become more aware of financial considerations in everyday clinical work after the introduction of various economic incentives in the s (Lampou, ; Forsberg et al., , ).Results from a newer interview study on GPs confirm that payment systems rewarding many patient visits make doctors feel that they have to prioritize shorter visits by healthier patients and that this leads to less available time for patients with more complex and timeconsuming health problems (Vengberg et al., , Swedish Government Accounting Office, ).In this sense, the individual autonomy of the medical      ? profession can be said to have been reduced by the new financial renumeration schemes.
Another impact on individual autonomy comes from the clinical guidelines, which contain detailed instructions to doctors on medical assessments and prioritizations, and thereby undermine the discretion of professionals in their clinical work (Funck, ; Hasselbladh and Bejerot, ).It should be noted, however, that even though the guidelines are formally issued by the NBHW, they are developed through cooperation with medical specialists.There are also indications that most doctors consider the guidelines to be useful tools in their clinical practice, seeing them as a support rather than as a constraint (Winblad and Andersson, ).When it comes to the rights of patients to choose care providers, it is worth noting, finally, that this organizational feature, which was introduced with the stated purpose of "empowering" patients in relation to professionals, does not appear to have much impact on clinical decision-making.Several studies show that doctors do not feel pressure to compete for patients, or try to increase their satisfaction, as they perceive that there are always new patients coming if some leave (Winblad, ; Vengberg et al., ).
Regarding the third type of professional autonomy, evaluation of work, the picture is mixed as well.As shown above, new forms of quality evaluation and control, such as inspections, patient surveys, quality registers, and comparisons between care providers on the basis of performance data, have been introduced.These new control measures have enabled public agencies both at the regional and state levels to monitor and evaluate the profession's performance and compliance with national standards in a way not seen before in Swedish healthcare.Several studies have found that these new forms of performance measurement indeed undermine the profession's authority to decide what is considered 'good medical quality', as this evaluation has been transferred at least in part to other actors (Bejerot and Hasselbladh, ; Jespersen and Wrede, ; Hasselbladh and Bejerot, ).At the same time, it can also be argued that the medical profession has been able to preserve some of this authority.Most important in this respect is that doctors are deeply involved in formulating the clinical guidelines in their respective areas of specialization, and when medical quality indicators are developed (Fredriksson et al., ).This implies that the profession has in fact retained considerable influence over defining what is 'best practice' or qualitative outcomes in their field.Levay and Waks conclude, with reference to the medical profession in Sweden, that 'the professionals submitted themselves to the gaze and judgement of others, but in the end, they retained considerable control over judgement criteria' (, p. ).It should be noted it is mainly an 'elite' group of highly specialized doctors who are involved in the quality management systems, not all doctors.This may be interpreted as a tendency towards fragmentation, or hybridization, of the profession (Noordegraaf, ).Finally, it does not appear that the right of patients to choose care       providers has led to any real shift in authority towards the users to evaluate the quality of professional work.Available data indicate that relatively few patients in Sweden make use of their right to actively choose providers, and that if they do, they do not use any type of performance data to guide their choice, not even on accessibility or waiting times (Hoffstedt et al., ).Summarizing the impact of NPM reforms on the autonomy of the medical profession, it seems that there is a negative effect, particularly with regards to work organization and individual autonomy, but that this is more limited than could be expected, given the extensive NPM reforms in this area.It is also noticeable that the medical profession has become actively involved in the new forms of performance measurement and steering associated with NPM, thereby extending their professional domain.
NPM in Swedish education: what effects on the professional autonomy of teachers?
NPM has had wide impact within the Swedish education sector as well.Management by objectives (MBO) was introduced in  when the responsibility over education was decentralized to the municipal level after previously being divided between the state and local governments.The reform led to elected representatives and civil servants at the municipal level gaining a stronger direct authority over the schools, both with regard to formulating goals and evaluating their fulfilment.At the same time, the schools were made into selfgoverning units with budgetary and personnel responsibilities, which led to a stronger managerial role on the part of the principals (Persson et al., ).The second and perhaps best known NPM reform in Swedish education was the School Choice reform introduced in .The reform was radical, in effect transforming the Swedish public education system from a public bureaucracy to a quasi-market, where private organizations were given the right to compete on equal financial conditions with public schools.Public resources are allocated to schools on the basis of the choices of students.The reform led to a competitive education market, particularly in urban areas, and a rise in privately operated schools (known in Sweden as 'independent' schools).In , % of the students in primary and secondary education, and % of the students in upper secondary education, attended privately operated schools, in most cases owned by for-profit firms.The third wave of NPM reform in the Swedish education sector consisted of an array of regulatory initiatives from national authorities making up a new quality management regime.This regime included new directives to municipalities to monitor and qualitatively evaluate the schools, more national achievement tests, new requirements of documentation of student performance, new and more transparent gradating systems, and the development of new audit and inspection systems (Bergh, ; Rönnberg, ).In , a new      ? agency for inspecting and collecting performance data from all schools was established, the Schools Inspectorate.The regular publication of student achievement statistics by the National Agency of Education added to the media interest in school rankings and can in this way be seen as contributing to the transformation of the public school system into a competitive and increasingly stratified school market.
Taken together, the NPM reforms introduced within the Swedish education sector have clearly affected the working conditions for teachers.There has been a general strengthening of administrative authority within the school sector, reflecting the new managerial powers of both municipalities and principals.Parallel to this development has been the introduction of market dynamics in the form of privatization, competitive pressures, and an increased focus on the satisfaction of students and their parents (Parding and Berg-Jansson, ; Stenlås, ).It should be noted that there have also been a few reforms during the period with the explicit purpose of strengthening and professionalizing the role of teachers, such as the introduction of a teacher license in  and the First Teacher reform in , which introduced a new form of career advancement for teachers (Bergh and Englund, ; Lilja, ).
The implications of the NPM reforms for the professional autonomy of teachers in Sweden are complex and not easily summarized.Starting with autonomy over work organization, it seems that this has been challenged in several ways during the period.One of the most important factors in this respect has been the transformation of the role of school principals from part of the collegiality to administrative managers.In Sweden, this process was prompted not only by the MBO reforms introduced in , but also by formal removal in  of the requirement that principals should have a teaching exam, and changes in the education program for school principals that made it mandatory and strengthened its orientation towards business management and leadership.Jarl et al. summarise the changes in the principals' role in Sweden by concluding that they have turned into a profession of their own, distinct from the teachers ().The changes in their work organization reported by teachers in Sweden are many and point generally towards a loss of professional power over how they spend their working time.Examples of such changes include: the introduction of individual, performance-related wage-setting; demands that all student achievements are documented in writing and communicated regularly to students; demands that teachers participate in school marketing events; the introduction of regular "mentorship" meetings with students to review their performance; demands that the teachers participate frequently in organizational activities at the workplace (staff meetings, committees, etc.); requirements that they work in "teams" when planning their teaching; less freedom to work from home; demands that teachers are accessible to students and their parents through email, phone and meetings; and, finally, demands that they participate regularly       in different quality evaluation programs (Frostenson, ; Stenlås, ; Bergh, , ; Parding, ).Taken together, these changes indicate that managers have gained influence in schools and that teachers today have considerably less freedom to organize their work independently than they did prior to the reforms.
The individual autonomy of teachers in their professional work seems to have been less affected by the NPM reforms.Several interview studies reveal that teachers generally feel that the freedom to organize activities in their classrooms remains more or less intact.This means that they can decide independently on pedagogical methods, reading materials, and interact with the students as they see fit (Frostenson, ; Ringarp, ).Frostenson () notes that school leaders rarely seem to interfere with, or even discuss, the teachers' pedagogical work, but concern themselves foremost with administrative tasks such as scheduling, planning, and staff meetings.As explained by one teacher: 'There is freedom in the pedagogical work.The school leadership is not very interested in pedagogy.Never had a discussion about that with them.My principal is totally uninterested.Other issues are important.That's where I am extremely managed.' (cited in Frostenson, p. .own translation).
At the same time, there are also indications that the teachers' discretion is being more infringed upon by pressures relating to grade-setting.A report from the National Union of Teachers in  showed that % of the teachers claimed to have been exposed to what they saw as non-legitimate pressure from either principals, students, or parents to set higher grades than warranted by the students' performance (NUT, ).As noted by Parding and Berg-Jansson, the competitive school market has created economic incentives to satisfy students and boast the school's achievements, as a loss of students will have direct negative ramifications for the school's budget (Parding and Berg-Jansson, ).
When it comes to the third form of autonomy, the right to evaluate professional work, it seems clear that this, too, has been undermined.Particularly the shift to quality management as the main steering doctrine for the Swedish school system in the s has had far-reaching implications for how and by whom the work of teachers is evaluated.This power has shifted from the profession itself towards public authorities like the municipalities and the National Agency of Education (NAE) at the domestic level, and organizations like the OECD and EU at the international level (Sellar and Lingard, ; Martens and Wolf, ).Moreover, it appears that the teachers have also lost autonomy in defining how their work is evaluated, as this is increasingly done through quantitative measurements of student achievements and grades, evaluation criteria with which many teachers disagree (Bergh, , ; Lundström, ).Lundström argues that the quality management doctrine has shifted the interpretation of quality in education from a broad and multi-dimensional understanding, where values such as the spread of democratic values and      ? socialization of students were given prominence, to a preoccupation with quantifiable forms of achievement measurement.This shift, Lundström notes, is not driven by the profession but by political actors at the national and international level, and can be seen in light of a global drive for economic competitiveness ().Similarly, Bergh argues that the quality movement in education represents a challenge to the autonomy of the teaching profession in Sweden because it implies a view of teachers as valuable only to the extent that they can deliver student achievement: '[f]rom the early s onwards : : : there is a linguistic change as arguments about teaching and teachers become more instrumental in character' (Bergh, , p.).
With the introduction of competitive markets for schools in , the power to evaluate the work of teachers can also be said to have shifted to the consumers of education, e.g. the students and parents.This shift in authority was a central motive behind the school choice reform, which sought to transform the users of public education from passive recipients to 'customers', whose free choice of school would contribute to a quality improvement within the system as a whole (Government proposition, /: ).While there is little evidence of a general quality improvement in Swedish schools following the reform, it has contributed to increased social and ethnic stratification within the system (Forsberg, ; Böhlmark et al., ).This, in turn, has led to a growing importance of school branding and the fact that factors like location, student body composition and rumour have become more important when parents and students chose between schools.The implication for teachers is that they work in a marketized environment, where activities such as marketing and customer satisfaction have become more important (Lundström, a; Fredriksson, ), but also that their work environments have become increasingly differentiated.In sum, it can be concluded that the NPM reforms introduced in the education sector in Sweden after  have had a largely negative impact on the professional autonomy of teachers.They have experienced reduced autonomy to organize their work and a significantly reduced autonomy in evaluating its quality.The dimension where their autonomy has been best preserved appears to be the individual, as self-governance and discretion still characterizes the teachers' work in the classroom.

Concluding discussion
The findings in this paper show that there has been a reduction in professional autonomy of both medical doctors and teachers in Sweden following as a result of NPM reforms.Both professions have experienced a strengthening of managerial authority over their work organizationfor instance, in that they have less control over how they spend their time and how the work is divided.Most notably, they have been forced to undertake more administrative work as a result of       increased demands on documentation, performance reporting, and quality evaluation.Additionally, both doctors and teachers have experienced pressures which infringe on their individual autonomy.In the case of doctors, such pressures have been foremost related to cost containment as new financial reimbursement systems have created stronger incentives to ration services and reduce the time spent with patients.The medical profession has also experienced reduced individual autonomy due to the introduction of clinical guidelines.For teachers, the reduction in individual autonomy has been related foremost to grading, where students, parents, and principals have come to exert more pressure in this regard, but also to new national curricula detailing what areas they should cover in different fields.When it comes to the last autonomy dimension, autonomy to evaluate the quality of one's work, both the medical and teaching profession have experienced how this power has shifted towards external actors such as public agencies and service users.These findings of reduced professional autonomy are broadly consistent with previous research on the effect of NPM reforms on doctors and teachers (Noordegraaf,  Even if the findings in the paper show many similarities in how autonomy has been reduced in the medical and teaching professions, there are also differences.First and foremost, the findings indicate that the teaching profession has experienced more loss of professional autonomy than the medical profession.The managerial authority of school leaders and municipal bureaucracies appears to have been strengthened more, resulting in a more manifest loss of autonomy for teachersfor instance, with regard to how and where they spend their work time, and how they plan their teaching.The teachers' loss of autonomy over their work environment is also enhanced by the fact that principals are no longer seen as peers but rather representatives of local governments or, in the case of independent schools, private employers.It is also evident that teachers have been more affected by the introduction of user choice models, which in Sweden transformed a bureaucratically organized public school system to competitive school markets; a development which has forced teachers to pay more attention to the expectations of students and parents.Finally, teachers have suffered a more marked loss of autonomy in evaluating their work, as this power has shifted almost completely to external actors, both at the national and international levels.While doctors have been able to extend their professional domain to become consultants, or experts, when the state sought to develop new systems for quality evaluation in health care, this does not appear to have happened in the same way in the case of teachers.This finding indicates that not all professions may be as apt at developing new forms of organizational professionalism or become 'hybrids' (Evetts, ; Noordegraaf, ).It is also striking that school managers have begun to develop into a profession of their own,      ? whilst it is still common for doctors to combine managerial roles with clinical practice.The differences in impact on professional autonomy between the medical and teaching professions are summarized in Table .
The findings in the paper have several implications for how the NPM on professional autonomy should be understood.First, the comparative case studies make it possible to identify common mechanisms through which NPM reforms impact medical doctors and teachers in Swedenfor instance, in the form of stronger work place management and new requirements for documentation and performance evaluation.It is notable how similar the developments are: a fact that can be seen as testimony of the relative coherence of NPM reforms in Sweden despite the long time period under which reforms have taken place.Second, the comparison also makes it possible to identify differences in how NPM reforms have manifested themselves in these two professional domains.Two examples which become manifest in the Swedish case show how user choice has been more detrimental to professional autonomy in the school sector than in health care and how doctors appear more successful than teachers in developing organizational professionalism.To provide a full explanation for the apparent higher capacity to preserve professional autonomy on the part of doctors is beyond the scope of the paper, but a few possible factors can be identified.First, as a classical profession doctors can be expected to have a higher level of authority in relation to other occupational groups.They have also been historically more successful in "closing" their profession to all who do not have formal medical training, which remains an elitist and highly selective education.For these reasons, doctors retain a unique position as leading experts in the field of health care.Another factor behind the more pronounced loss of autonomy of teachers as compared to doctors might be the more technical nature of medical knowledge, which makes it easier for the medical profession to close their field to other occupational groups (Currie and Suhomlinova, ; Turner, ).
A third contribution of the article to the research field is, arguably, the disaggregation of the concept of professional autonomy, which makes possible a deeper understanding of the ways in which the professional autonomy of doctors and teachers has been affected by NPM reforms.By distinguishing between autonomy over work organization, individual autonomy (or discretion), and autonomy over work evaluation, the paper has clarified that it is foremost in the first and third domains that professional autonomy has been reduced.Both doctors and teachers retain the right to make discretionary decisions in their work and, in that sense, they have preserved a significant part of their professional authority in their interactions with users (cf.Ahlbäck Öberg et al., ).
The implications of the main finding in the paper, that both the medical and teaching professions in Sweden have lost professional autonomy as a result of NPM reforms, can be understood in different ways.To the critics of professional privilege and authority, it may be welcome news that public and private managers now can assert more control over how professions carry out their tasks and evaluate what the results are.Likewise, the shift in authority from the professions to the users of public welfare services can be seen as beneficial both for individual freedom and democratic participation (Saltman and von Otter, ; Rothstein, ).Furthermore, it is not the case that all reforms introduced in welfare services under the umbrella of NPM have been seen as negative by professions themselves.There is ample evidence, for instance, of professions welcoming more systematic quality evaluation, which has been possible through the development of new and improved quality indicators (Heiwe et al., ; Shortell et al., ).Another example is improved documentation and standardization of professional routines, which help maintain professional quality standards across different localities.In this light, it seems important to nuance what sometimes has been portrayed as an inherent conflict between NPM and the welfare professions.At the same time, there might also be negative effects for the quality of services associated with reducing professional autonomy.As demonstrated by Freidson, when organizations shift from a professional logic to logics associated with bureaucracy or the market, there are other forms of power that reassert themselves.A more bureaucratic steering logic implies that welfare services become more standardized and that professions are obliged to follow protocols rather than their professional judgement.This undermines flexibility and the aptitude for situational adjustment, both important values in human services.A stronger market-based logicfor instance, in the form of professionals needing to satisfy user demandsis likely to result in some users, being more apt at articulating such demands, gaining an advantage over others.Last but not least, if the welfare professions were deprived of too much of their autonomy, they might lose some of their motivation, ethical guidance, and sense of duty.Such traits, which have been of central importance when professions have earned the trust of citizens to make decisions concerning deeply personal matters in their lives, might be hard to recreate if lost.
     ? NPM in Swedish health care: what effects on the professional autonomy of doctors?

TABLE  .
Assessed impact of NPM reforms on professional autonomy