**1. Introduction**

Neoliberalism has provided the ideological framework behind the transformation of the Western liberal state since the 1980s. Apart from deregulation and privatization its dominant feature has been a subjection of the public sector to market logic or market-imitating principles. The wide range of reforms initiated since the 1980s to re-shape the public sector in this new image has been conveniently subsumed under the label 'new public management', or *NPM*, [1–4]. While, like neoliberalism, hardly a well-defined term it has been guided by a re-conceptualization of the relationship between public sector and citizens in terms of producer and consumers, the conviction that market principles provide more e fficient forms of governing and control, signaling the introduction of the 'three M's': Markets, measures, and managers [5,6]. As the view of the state, and its role, underwent a transformation in Western (liberal) political imaginary, so did its role in international development policy. Neoliberal reform has been central to international (Western) development policy since the 1980s and became the underpinning rationality for promoting 'good governance' [7–11]. While this export idea has come under pressure from an emerging actually existing alternative with China's growing role in development, the conception has remained that it provides the bedrock of a well-functioning and

transparent state and it is crucial to democratic development.<sup>1</sup> Indeed, it provides the core for much of what remains of liberal development theory and practice.

The conception that economic and political liberalization go hand in hand, and the assumption that economic liberalization supports democracy has a long tradition in liberal theory and is deep-rooted in democratization theory [12–16]. Although the conception has been challenged, and although part of the Marxian-inspired left have always claimed that market liberalism is incompatible with true democracy, the actual inversion of the idea, i.e., that liberal notions and conceptions of democracy are damaged or destroyed by liberal development itself, has not been fully explored from a liberal conception.<sup>2</sup> In this article, we aim to pursue such a course, arguing that the neoliberal tradition and new public managemen<sup>t</sup> reforms of the public sector effectively erode core (liberal) democratic traditions and undermine good governance, in the sense of transparency and the rule of law. We do so by first outlining key characteristics of neoliberalism and new public management, in the sense, we conceive of it, and then move on to empirically illustrate how reforms in Sweden systematically challenge the transparency of the public sector and the rule of law in a core liberal state, essentially through the emergence of an illiberal practice in the form of what we call 'shadow management'. We depart from a theoretical review of how NPM transforms the public sector and from empirical observations drawn from the Swedish higher education sector. The empirical observations are derived from case studies by the government-independent watchdog Academic Rights Watch (ARW), which monitors academic freedom in Sweden. ARW has documented over a hundred cases in detail in a database available online (www.academicrightswatch.se).<sup>3</sup> This documentation, which mainly consists of public records, is aimed at violations against academic freedom, including freedom of speech. However, in this material, we also observed the phenomenon we call 'shadow management' (see Section 3 for a definition). While we were critical about major aspects of new public managemen<sup>t</sup> (managerialism in the public sector) especially in higher education already from the theoretical point of view, deeming its principles in conflict with core values in the public sector, with academic freedom and in governing the university, it was the observations in the empirical cases that led us to induce our hypothesis that NPM generates the emergence of shadow management. We hypothesize: (i) That this is a general trend, of which we have only documented in detail some examples and in one sector (higher education), and (ii) that the emergence of this practice can be tested against the 'controlling institutions', which provide checks and balances for the public sector in Sweden. While every single case of shadow managemen<sup>t</sup> practice is a problem with regard to transparency, the rule of law and the constitution, and a pattern of increase is particularly worrying, it is only through the lack of response by the controlling institutions that we really ge<sup>t</sup> systematic erosion in the core of the liberal state. Corruption and nepotism may be incidents in the best of governing structures, but as long as there are checks and balances, functioning control institutions and people are held accountable, the public trust in the institutions can be safeguarded. Hence, addressing the effect of NPM on the controlling institutions is an important part of this study. In the Swedish case, it is of special interest to demonstrate how the institutions of the Parliamentary Ombudsman and governmental Chancellor of Justice work, since they have been international export products for securing democratic functioning. A crucial finding concerns how those legal institutions introduced to safeguard the constitution, the rule of law, and democracy fail to fulfill their purpose and instead operate to undermine and strategically reinterpret NPM values. While there have been

<sup>1</sup> This connection has been regularly repeated in numerous international documents on development, for example, by the World Bank.

<sup>2</sup> The idea has also been questioned within democratization theory but remains a die-hard assumption within liberal development policy and especially US Foreign policy (for a questioning and nuances see for example [17,18].

<sup>3</sup> The documentation is too large to address in detail here, whence we have selected a few cases which we consider to the representative of the overall picture. The detailed documentation is available online at www.academicrightswatch.se. While this material is in Swedish, some cases are summarized in English and available at www.academicrightswatch.com and there is also a summary report in English of about two dozen cases from 2013: 'Academic Freedom 2013' (available from the aforementioned website).

indications that NPM leads to more complaints to ombudsmen and a decline in the percentage of complaints that are actually investigated (Hood and Dixson, 2015, Ch. 6) [4], to our knowledge, the dilemmas and complex choices facing ombudsmen and other officials under neoliberalism have not been fully appreciated. For the positioning of our research in relation to other work on and criticism of the institution of ombudsmen, we refer to Section 7 below.

#### **2. Neoliberalism and Its Misconceptions**

Neoliberalismhas been widely misconstrued as a mere extension of liberalismembracing laissez-faire, various civil and individual rights and constraints upon governmen<sup>t</sup> intervention. A common misdiagnosis and misconception of neoliberalism is that it reduces governmen<sup>t</sup> and, by liberating the economy from governmen<sup>t</sup> intervention, dispels distortions and corruption associated with political meddling in the market. In fact, in many cases, it does the opposite. The trend that mass surveillance of citizens has reached a previously unseen historical peak in the neoliberal state can only partially be attributed to technological development. A strong intrusive state is clearly at the very foundation of neoliberalism and recent explorations into its theory and practice have begun to provide more clarity into its nature [19–23]. Neoliberalism constitutes a break from classical liberalism in several senses. A crucial difference is that neoliberalism views the 'free' or 'self-regulated' market as the best model for all social relations. The foundation for this is the neoliberal understanding of knowledge and the conception that the market is superior to any human or collective of humans in processing information [24–27]. The market is a superior information processor, and as such, all interference in it will be illegitimate [28]. Neoliberalism typically also rejects the idea of the market as something natural. Instead, it considers the market as an artificial construct, something that has to be introduced through a political decision and by legislation (a foundational theoretical text is [24]).

Michael Sandel has suggested a distinction between market economy and market society [29]. In a market economy the market is an important sphere that can support the development of a society, provide economic growth (that may be distributed in various ways), channel and make use of entrepreneurship, skills, and creativity, but must be kept within limits and have democratic political regulation, since an untamed market tends to generate social effects that may be unwanted, such as deep inequality and economic and social injustice. Hence, the market is an important instrument but neither a goal in itself nor an overarching principle for society as a whole. Some areas and values will typically be safeguarded from the market and economic profit. In a market society, by contrast, the market becomes the model for all social relations and the market premises are introduced at all levels of society, including the public sector, in education, culture, etc.

Market society, by definition, transforms all values to market, and ultimately economic, values. In doing so, it, again by definition, corrupts other values. It is a value-transforming machine. In a market society, the market will trump the very foundation of democracy and human choice is reduced to consumer choice.

Neoliberalismtransforms democracy in a number of dimensions, but most concretely, it affects democratic governance and culture through the philosophy underpinning the array of reforms that have become known under the label 'new public management' [4,6,30]. While the label new public managemen<sup>t</sup> (NPM) is a loose umbrella term encompassing an array of different reforms in different sectors and countries that have been implemented with different organizational models in mind (with different labels such as 'lean' or 'command and control', or 'systems view'), it is underpinned by the idea that the public sector needs to introduce forms of steering and control that exist in the private market, which, it is assumed, provides for more efficiency and thereby better service. The basic idea is simple. The public sector provides services in much the same manner as private companies do. These services are paid for and utilized by the citizens, in a similar way as with customers to a company. Through competition, privatization, opening up for private actors to implement services, separate budgets and result accounting for different units within agencies, institutions, municipalities, etc., and by goal setting and control of the goals through measurement, the public sector is assumed to

become more cost-e fficient and flexible as well as more e ffective in meeting citizen needs. The citizens, in this context, are thought of as 'customers' or 'users' (the actual customer may be a municipal agency, etc., providing service to the 'users'). Hence, the language itself is a reflection of this philosophy, or ideology, which constitutes a break from the classical public administration model. The classical model was based on rule-governance (the rules are the foundation, and each service or agency receives an estimated lump budget sum), and the assumption that the public sector had a unique, special role in serving the citizens and thereby also democracy (the 'public'). Constituting the very architecture of the democratic state, it should be governed by its own principles and rationality that are di fferent from those of the market [31–33].

An immediate consequence of NPM is that all activity must be possible to evaluate in numerical and economic terms. Another consequence is that professional judgment (for example, by physicians or teachers) will be subordinated to the evaluations by the administrative control, thereby in a sense, leading to both a form of de-professionalization and a build-up of bureaucracy. Arguably, a further effect is that there is more room for direct political and ideological goal setting and setting of specific targets, within a variety of areas that concern either the activity in question or is more general and ideological (i.e., a specific goal for a school or hospital, or a more general goal like promoting gender equality or an 'environmentally certified' institution). The reason the space for direct political and ideological goal setting is wider, is because the space for the actual profession is diminished. Hence, actual specific professional influence has to stand back for political directions, which may not be directly related to the activity in question but to broader ideological motives).<sup>4</sup>

However, the most problematic aspect from a democratic perspective is that public agencies and institutions suddenly start to behave as if they were private companies, focusing on their 'brand name' (or reputation) and issuing policies against the backdrop of this new rationality. As we will see, this also creates conflicting goals, forcing authorities and governmen<sup>t</sup> agencies to choose between following the law or the governmen<sup>t</sup> directives.

#### **3. The Emergence of Shadow Management**

In the Swedish case, constitutional law, such as freedom of speech and freedom to disclose information (a right for public servants to whistle-blowing), and the laws regulating the function and operation of the public sector (ultimately resting on the constitution), were developed in relation to the building of democratic institutions. They are interwoven with the public sector as an instrument for democracy. This is one of the foundations for how the public sector shall be governed. The other foundation for the governing of the public sector is the governmen<sup>t</sup> directives and instructions (Regleringsbrev), which are now guided by the demands of managemen<sup>t</sup> philosophy, with managerial demands for market-imitating behavior, result-orientation, competition, and measurability, with inbuilt frequent auditing, evaluation and control.<sup>5</sup> The latter (managerial philosophy) contains goals and demands that are sometimes in tension with the former (the constitution). To take an obvious example,

<sup>4</sup> To provide a concrete example, which is also empirically confirmed in informal discussions with several regional politicians in Sweden: *Profession-oriented goals* relate to the actual activity, such as health care within the health sector, whereas *political-ideological goals* would relate to other goals set for that sector, for example related to identity politics (such as promoting diversity, gender equality) or sustainable development, etc. This also explains why NPM has an attraction within the green and left parties, although it comes from neoliberal premises (the Green Party in Sweden has for example been a strong advocate of NPM, since it enables them to better pursue identity political goals within the public administration, universities, etc.). In the Swedish case, virtually all public service and public agency institutions now have a range of goals related to identity politics. Often, an agency or a region may have several dozen goals that reach well outside the actual activity. Often, these goals are unnecessary since the issue is anyway covered by law (such as non-discrimination, etc.). With too many goals, the risk of goal-conflicts increase and the focus is diluted.

<sup>5</sup> A clear expression of how the public sector is now guided by managerial demands is found in the views of the Swedish Agency for Government Employers (Arbetsgivarverket [34]), which is the governmental authority responsible for negotiating with the trade unions on employment issues on behalf of governmen<sup>t</sup> agencies. While several documents show the same managerial-oriented trend, a good single source is its strategy for state employment policy "*Strategi för den statliga arbetsgivarpolitiken"* (2013). See also the analysis of its policy in [35].

freedom of speech and freedom to disclose information is not unproblematically combined with a managerial concern for protecting the organization's brand name.

NPM may, due to its pre-programmed tension with constitutional requirements, seem a highly implausible philosophy to begin with. Nevertheless, this type of reform started to be implemented during the second half of the 1980s by the Social Democratic government, led by then PM Ingvar Carlsson, and supported by Finance Minister Kjell Olof Feldt and chief economist Klas Eklund. In 1988, the Parliament brought a decision about a general transition in the form of governing the public sector from 'rule-oriented' towards 'goal-' or 'result'- oriented, and the following three years this policy change spread to the regional and municipal levels [36]. The economic crisis that hit Sweden in the 1990s propelled this development even further, regardless which party constellation was in power, and by 1995 a study by Christopher Hood estimated that Sweden, along with UK and Ireland, was among the most diligent of countries in transitioning to NPM [2].

As we shall argue, the conflict between the core constitutional guarantees (of openness, transparency and full disclosure of information, freedom of speech, freedom of disclosure of information)<sup>6</sup> and managerially-influenced governmen<sup>t</sup> goals and instructions, is, in practice, solved by the introduction of shadow management, whereby sensitive errands and managerial decisions that may be unconstitutional or otherwise illegal are dealt with in internal memos and secret consultancy reports. The consequence is a duality in the public sector, which damages public trust in institutions and erodes their democratic legitimacy.<sup>7</sup>

According to Swedish law, public agencies must fulfill a high degree of transparency, for example, in recruitment and sta ff issues, as enshrined in constitutional and administrative law. Freedom of speech must not be restricted except in ways that are specified in the law. In the private sector, by contrast, the legal requirements on transparency are not as demanding. There, freedom of speech may be restricted by the company, which can also hire sta ff in much the way it chooses to. This condition has led to criticism that employees in the private sector do not enjoy the same protection of freedom of speech and freedom to disclose information as public servants.

What happens in cases where the new kind of governance, required by the government, is incompatible with constitutional and administrative law? Which one has to give way? One possibility is that the authorities and public agencies rebel against governmen<sup>t</sup> directives and decide to follow the law. However, public agency leaders are under a legal duty to obey governmental directives or face the risk of being dismissed. Another possibility is to ignore the law and follow the governmental directives. This, however, risks leading to exposure in media, which tends to pick up on legal irregularities. In the worst case, it could activate a much feared 'media drive' against the public agency leadership, whereby the governmen<sup>t</sup> sooner or later would have to step in and show decisiveness, typically by replacing the leadership of the agency in question. In practice, therefore, this is also not really an option.

In order to manage this dilemma, public agencies (authorities) and their jurists have discovered a 'third way'. Towards the public, they must appear to follow the law and constitution, but inwards and hidden from public view, a quite di fferent practice has developed, namely a shadow practice, which operates much like in the private sector. By appearing to obey the law, the media and public

<sup>6</sup> Enshrined in the two constitutional laws 'Regeringsformen' (the Instrument of Government) and 'Yttrandefrihetsgrundlagen' (The Freedom of Expression Act). The other two (of four) constitutional laws, not relevant in the present context, are 'Successionsordningen' (The Order of Succession), regulating the Royal House, and 'Tryckfrihetsförordningen' (Freedom of Press Act).

<sup>7</sup> For the connection between the functioning (and non-corruption) of institutions and trust in them, as well as the connection between trust in institutions and social trust is well-researched, in a Swedish context see for example: [37–39].

scrutiny can be kept at bay, while the governmental directives to apply managerial (NPM) practices are met in the background.<sup>8</sup>

By shadow management, we mean administrative practices for parallel treatment of issues sensitive with respect to law and regulations, and which addresses these issues without transparency, i.e., in the 'shadow'. Thus, 'shadow management' denotes managemen<sup>t</sup> or public administration with a parallel structure or parallel set of practices invisible to an outsider, where all errands and cases that do not fit the relevant legal requirements are dealt with, in direct or possible violation of the law, especially constitutional and administrative law.<sup>9</sup> By not registering such cases and just dealing with them internally, they are di fficult to find for an external reviewer and are, therefore, rarely subject to independent scrutiny. This leads to cultivation of informal practices, networks, and routines that are not recorded or regulated. The public sector ceases to be transparent or conform to the rule of law. For instance, fairly regular cases in which there is a minimal risk of irregularity or illegality are archived and presented to the public on display, as a law-abiding façade. Cases which are in potential violation of constitutional or administrative law are no longer recorded or archived, but addressed and dealt with in internal memos, secret consultancy reports, or labeled 'working material', which does not legally require archiving. This is particularly the case with regard to basic civil and administrative rights, which are not prioritized or part of the architecture in the NPM system.<sup>10</sup> For example, decisions that may interfere with the basic rights of individuals are, at best, stored locally in the departments rather than in the central university registry, which makes it practically impossible to make them subject to scrutiny. An outsider would not even know what to look for in the first place, since there is no trace of the documents in question in the central records or archives. In administrative terms, these documents 'do not exist' or they exist only for those in the administration who already know where to look.<sup>11</sup> To give a few illustrations of these, we will briefly summarize five cases from the documentation we draw upon.

#### **4. Cases of Shadow Management in the Higher Education Sector**

In 2006, a student at a major university undertook to study, from an organizational-theory perspective, a failed reorganization process, which shocked the Business School in which he studied and led to deep personal conflicts and deterioration of quality (e.g., loss of international accreditation).<sup>12</sup> The student concluded in his Master thesis that the Vice Chancellor of the university was ultimately responsible for the failures. As a consequence, the university refused to register the student's grade ("pass with distinction"). The grade was registered only after the student had filed a complaint to the regional Administrative Appeal Court. At the end of 2013, the student's Master thesis was still missing in the public electronic database on the department's website, in violation of the department's publication practice.<sup>13</sup>

<sup>8</sup> This would also be the 'rational' way to react to the dilemma introduced by NPM in the public sector. NPM creates a new rationality which creates shadow management, a new structure of rationality which, as shown below, affects even the Ombudsman under the parliament.

<sup>9</sup> The fact that an organization can have a 'shadow' hidden from and in complete or partial conflict with the o fficial image of the organization has been observed by several authors. According to [40], p. 387, for example, "Organization Shadow is understood as facts which organizations wish to deny about themselves, due to the threat posed to self-image and self-understanding and, more generally, the need to be viewed in a favourable light by others". In the few works in which the term 'shadow management' has been used before (e.g., [41,42], it signifies a situation in which someone other than those formally in charge are actually managing an organization. As should be clear from how we define the term, this is not the meaning intended here.

<sup>10</sup> Although such rights are not logically incompatible with NPM, there is, in practice, grea<sup>t</sup> tension between the former and the latter, since they have different goals (the traditional goal of the 'Weberian'-model of public administration was to serve the public, equal treatment etc., whereas the goal of the private sector, and imitations of it, is to maximize profit/minimize costs). 11

Finding this material is usually dependent on whistle-blowers, as in many of the cases, the ARW documentation. 12

This case is detailed under the date 9 Dec 2013 on the ARW website documentation: http://academicrightswatch.se/?p=798. 13 It was immediately made public following a complaint by ARW to the new Vice Chancellor, which is detailed under the date26Jan2014inthewebsitedocumentation: http://academicrightswatch.se/?p=845.

A similar case involving a senior researcher played out at a Swedish university of agriculture.<sup>14</sup> The professor had, for many years, published articles critical of Swedish forestry policy in the university's public electronic database. He also criticized studies made by some of his colleagues for legitimizing governmen<sup>t</sup> policy in the face of what he took to be evidence to the contrary. This criticism culminated in 2012 in an article in which a scientific investigation attributed to the university was attacked for relying on "insu fficient scientific methodology", meaning that the article became a potential threat to the university "brand". The professor archived the article in the database as usual. However, this time, someone with su fficient formal authority interfered with and deleted all his latest publications in the public database, which was thus cleansed of material that the university leadership might find disturbing.

Another particularly instructive example of shadow managemen<sup>t</sup> is the handling of a memo issued by one of Sweden's largest universities prohibiting researchers in a department from expressing "excessive criticism" among colleagues, following a conflict that was interpreted as a threat to the psycho-social workplace environment.<sup>15</sup> The prohibition, in e ffect, gives the manager level greater power and control over sta ff in perfect accordance with managerial or neoliberal principles. After all, the prohibition only concerns excessive criticism among colleagues. The regulation means, in practice, that serious criticism can only be given top-down and not between employees at the same level. Such an arbitrary restriction, issued by a public authority, runs the risk of infringing on the freedom of speech as regulated in Article 10 of the European Convention of Human Rights (ECHR). The vagueness of the prohibition makes it open to arbitrary interpretation, and therefore, likely to be legally unacceptable according to ECHR case law. For the point we wish to make, it su ffices to concur that the prohibition, at least potentially, infringes on freedom of speech, and therefore, is an important document that should be registered in the public archives for critical assessment. However, the memo was never o fficially registered or archived at the university in question.

The case with the memo is instructive also for another reason since it documents how internal tensions arise within the leadership of a public institution about what should be registered and archived and what not, which, on our analysis, is what can be expected to happen when NPM clashes with legal constraints. On request, the Head of Administration decided that the potentially troublesome memo should be registered in the central archives, but the decision was not executed at the lower levels in the hierarchy, possibly due to the interference of another high-level o fficial.

A variation on this problem is the practice of registering dubious documents in the o fficial archives but then, in violation of the Swedish constitution, denying access to them when they are requested by a citizen. This happened to a professor at a university college in central Sweden who, following a conflict with the university leadership that ultimately led to his dismissal, wanted to know what was stored in his files.<sup>16</sup> This case, too, concerned freedom of speech, since the employee had originally irritated the university leadership by informing students about the existence of diverging views regarding the suitability of moving an educational program to a new location. When some concerned students protested against the treatment of the professor, the Vice Chancellor proposed at a meeting that their critical note be "accidentally" left behind in her o ffice for the purpose of avoiding registration in the central archive.<sup>17</sup>

Our fifth and final example concerns public agency communication policy, which is a related area where neoliberal policy tends to clash with fundamental law. Communication policies usually stipulate what employees may or may not say, internally or externally. The phenomenon can only

<sup>14</sup> This case is detailed on 7 February 2015 in the ARW website documentation: http://academicrightswatch.se/?p=1376.

<sup>15</sup> This case is detailed on 27 December 2012 in the ARW website documentation: http://academicrightswatch.se/?p=194.

<sup>16</sup> This case is detailed on 4 April 2014 in the ARW website documentation: http://academicrightswatch.se/?p=914.

<sup>17</sup> This is the account given by the students. An administrative o fficer later explained away the event as a "misunderstanding" since, as she explained, leaving the document in the o ffice would still mean that the document was o fficially received, and therefore, should, by Swedish law, be archived in the Registry. The Vice Chancellor herself has, to the best of our knowledge, not commented on the event.

be understood against the backdrop of the pseudo-market conditions created by neoliberal policy, whereby publically funded universities are required to compete with each other for students and resources. Thus, the ultimate purpose of a communication policy is to protect the brand name of the university and thus to ensure that the university remains competitive on this pseudo market. Needless to say, there is little legal room for the university to act in this regard since freedom of speech is protected in the constitution. Thus, communication policies tend either to directly violate the constitution or to create uncertainty as to what can be said or not. This phenomenon has now become widespread among state, as well as municipal, authorities. To the extent that these rules are stored only internally, e.g., in the intranet of a public agency, or elsewhere where outsiders cannot access them, or cannot be expected to search for them, they are part of a practice of shadow management. Generally, they illustrate the duality and uncertainty that is generated by introducing NPM ideology in the public sphere.

In 2013, ARW began to review internal communication policies issued at Swedish universities. The starting point was the policy of a university college in mid-Sweden whose communication policy contained a number of peculiar formulations (also from a linguistic point of view), including exhortations to "communicate internally before externally". All employees were also expected to "stand behind and represent the decisions taken". Again, rules like these run the risk of infringing on free speech, which is a constitutionally protected right.

However, the set of practices we call shadow managemen<sup>t</sup> could not emerge and consolidate in the universities or public sector without the intermarriage with the controlling judicial institutions that either ignore or help legitimize them.

#### **5. A New Dual Role of the Ombudsman?**

Sweden is noted for having introduced the function of an ombudsman as part of the public system of checks and balances.<sup>18</sup> Many other countries now have ombudsmen, which handle various complaints from citizens. The governmental ombudsman in Sweden, the Chancellor of Justice, can even settle damage claims against the state. When state policy is in line with the law, the ombudsmen can perform an important role in a democracy by considering complaints against the state from the general public and public employees in a quick, free, and non-bureaucratic way, without the complaint having to go through the courts.

However, when state policy is in tension with the law, as is the case when neoliberal governmental policy is introduced in the public sector, the ombudsmen can be expected to reinforce neoliberal policy rather than constitutional or administrative law.<sup>19</sup> The reason is that the head ombudsmen in Sweden are appointed by either the parliament or the government. Here, we will focus on the Chancellor of Justice, the governmental ombudsman. If the Head of the Chancellor of Justice, who bears the same name as her institution and needs to sign every non-trivial decision, does not settle complaints in a way that is seen by the governmen<sup>t</sup> as "adequate", he or she will run the risk of being replaced by someone who is seen as "more competent" or "flexible" in her decision-making.<sup>20</sup> This is not a big concern in the normal state of things when policy and law cohere, but, again, it becomes one if state policy is in conflict with the part of the legal system overseen by the ombudsman because then "adequate" may, from a governmental perspective, very well mean "in accordance with state policy" rather than "in accordance with the law". In this abnormal situation, the ombudsmen in practice take on a whole new role, namely that of defending state policy against the law.

<sup>18</sup> See Jägerskiöld (1961) [43] for the history of the Swedish (parliamentary) ombudsman.

<sup>19</sup> See also Hood, C (2015) [4]: Ch. 5.

<sup>20</sup> Cf. Bull (2000) [44], p. 342: "Another worry about these 'lesser' ombudsmen is that they are not as independent as the Parliament's ombudsman. They are all agencies under the government. This means that the governmen<sup>t</sup> may choose to replace any one of them as it sees fit and that the governmen<sup>t</sup> controls their budgets. This close connection between the ombudsmen and the governmen<sup>t</sup> does not promote the status of the ombudsmen." This is true not only of the "lesser ombudsmen", such as the discrimination ombudsman, but also of the Chancellor of Justice.

The obvious strategy for resolving this conflict, consciously or subconsciously, is, in e ffect, for the ombudsmen to reinterpret the law so that state policy conforms to it, which amounts to proclaiming through various pseudo-legal discursive strategies that the conflict between state policy and law was only apparent. In general, one would expect conflicting decisions and generally fragmentary decision-making from the ombudsman in such cases since the situation will be interpreted di fferently by di fferent employees at the ombudsman, whereby some lower o fficials will attempt to abide by the law while others, in particular the head ombudsman, will focus on implementing the directives from above.<sup>21</sup> In particular, it will be increasingly di fficult to retain consistency over di fferent decisions and over time. One would also expect that doubtful decisions defending neoliberal policy should not be made directly available for public scrutiny, e.g., on the ombudsman's homepage. Rather, these decisions will be made in the background and the relevant documents silently conferred to the internal archives.

A potentially detrimental consequence is that the decisions by the Chancellor of Justice tend to be seen as precedents for the courts, which, in turn, makes it possible for the governmen<sup>t</sup> to influence, via the Chancellor, the rulings in the courts in a particularly expedient fashion. Following the Chancellor's decisions, laws will then, in practice, be reinterpreted in courts in ways that systematically favour governmental neoliberalism. Similar observations apply to the Parliamentary Ombudsman, although this institution is not organized directly under the government, but as the name indicates, under the parliament. The latter is a detail of little practical consequence since, in a parliamentary democracy such as Sweden's, the governmen<sup>t</sup> normally is in the majority in the parliament.

A further way in which the new emerging practices are legitimized is through shallow investigation. In Sweden, the Chancellor of Justice is supposed to regularly check the recording and archiving practice of public authorities. However, in order to expose the universities and colleges, months of excavation in the archives would be required, coupled with an investigation and questioning of witnesses. It is highly unlikely that the Chancellor of Justice, a relatively understa ffed agency considering its mission to oversee hundreds of governmen<sup>t</sup> agencies, could meet this task. Following the necessarily shallow investigations of the Chancellor of Justice, it will appear as if all is operating according to regulations, and the practice will be legitimized.

#### **6. The Case of the Swedish Ombudsman**

Following our analysis, one should expect that the ombudsmen would not take action but instead produce pseudo-legal justifications for neoliberally inspired decisions that are in conflict with legal prescriptions. Moreover, we would expect there to be inconsistencies and tensions within and across decisions made by the same ombudsman. Finally, it is likely that all problematic decisions, which in effect, conflict with legal requirements, should be missing from the public database on the ombudsman homepage containing "decision which the Chancellor of Justice judges to be of general interest".<sup>22</sup> These expectations were, in many cases, confirmed. Below are some salient examples.

The case of the memo with the neoliberal speech prohibition against "excessive criticism" among colleagues was reported by two a ffected researchers to the Chancellor of Justice, who, as expected, went to grea<sup>t</sup> lengths to defend it, against all odds, as lawful. A central premise in the Chancellor's argumentation was the claim that a prohibition against excessive criticism is not a prohibition against speech with particular content, but only against speech of a certain form. However, the falsity of this premise should be obvious. Moreover, Article 10 of the ECHR explicitly forbids not only restrictions

<sup>21</sup> In Sweden, an ombudsman, or other higher o fficial, who is perceived by the governmen<sup>t</sup> to have served the state particularly well can expect to be awarded, under pompous circumstances, a gold medal by the Swedish king. The current Chancellor of Justice received the most prestigious medal in 2016. Arguably, the royal medals serve, among other things, as an incentive for high o fficials to remain loyal to state policy.

<sup>22</sup> Quoted from the Chancellor of Justice homepage (http://jk.se/Beslut.aspx as of November 22, 2015), our translation from Swedish.

of speech with particular content, but explicitly also "formalities" that have a chilling e ffect on free speech (and which cannot be defended as necessary in a democratic society).<sup>23</sup> Thus, a case could be made that the Chancellor's claim that the prohibition in question only concerns form and not content was not only inaccurate but also irrelevant.

At any rate, the Chancellor acted, in confirmation of our expectations, as a defender of neoliberal speech regulation in the public sector rather than as a guardian of the law. Her decision has conferred pseudo-legal legitimacy upon the prohibition in question. Other public authorities can now use the same prohibition to strengthen their managerial rule in violation of the constitution, with scrupulous reference to the Chancellor's authoritative evaluation. The case illustrates the expedience with which a neoliberally oriented governmen<sup>t</sup> can introduce managerial principles with apparent legal justification in the public sphere through the system of the ombudsmen. Without the ombudsmen, the governmen<sup>t</sup> would have to influence the courts to achieve the same authoritative e ffect, which is far more di fficult and time-consuming and, above all, politically risky.

The case of the professor who was denied access to his own files was dismissed by both the Parliamentary Ombudsmen and the Chancellor of Justice, in both cases without motivation. The case of the troublesome communication policy at the university college was considered by the Chancellor of Justice, who found several passages problematic from a constitutional standpoint, including those already mentioned. The Chancellor also issued an important general recommendation in her statement: That, namely, communication policies may only regulate o fficial speech that is made on behalf of the public authority in question, in our case, a university, and not speech that employees make as citizens. Surprisingly (but then again perhaps not), she resolved nevertheless not to issue "criticism" in the particular case, leaving the actual applicability and seriousness of her recommendation in some doubt.

Two further cases may, in conclusion, serve to illustrate the expected conflict in the Chancellor's decision-making over time. Both cases involve attempts by universities to sanction academic whistleblowing and yet, although the cases are strikingly similar, the Chancellor accepted one complaint and rejected the other. A detailed analysis showed that logically contradictory principles for freedom of speech were appealed to in the two cases. In the first case, from 2013, it was ruled that fundamental legislation was not violated with reference to the neoliberal principle that civil servants need to be loyal to the employer not only in implementing decisions taken at higher levels but also in refraining from public criticism of those decisions.<sup>24</sup> In the second case, from 2015, it was ruled instead that fundamental legislation had, in fact, been violated, with reference to the conflicting principle that while public o fficials have to be loyal in the implementation of decisions, this loyalty does not extend to their speech behavior.<sup>25</sup> The latter principle coincides with the standard legal interpretation of the relevant parts of the Swedish constitution. There is no reference in the latter ruling to the earlier one.

Finally, our hypothesis was that only the Chancellor decisions that are in accordance with the constitution and other relevant laws should appear in the public electronic database on the Chancellor's homepage. We expected decisions that try to confer pseudo-legal legitimacy on principles of neoliberal governance rather to be part of the Chancellor's practice of shadow management. Thus, we expected to find the last decision from 2015 in the database but not the earlier 2013 decision. We also expected not to find the ruling on the problematic memo in the database. As for the communication policy, we lacked a clear prediction. On the one hand, the Chancellor criticized what was, in e ffect, an attempt

<sup>23</sup> Article 10 of the ECHR (our emphasis): "1. Everyone has the right to freedom of expression. This right shall include freedom to hold opinions and to receive and impart information and ideas without interference by public authority and regardless of frontiers. This Article shall not prevent States from requiring the licensing of broadcasting, television or cinema enterprises. 2. The exercise of these freedoms, since it carries with it duties and responsibilities, may be subject to such formalities, conditions, restrictions, or penalties as are prescribed by law and are necessary in a democratic society, in the interests of national security, territorial integrity or public safety, for the prevention of disorder or crime, for the protection of health or morals, for the protection of the reputation or rights of others, for preventing the disclosure of information received in confidence, or for maintaining the authority and impartiality of the judiciary."

<sup>24</sup> This case with the Chancellor of Justice is filed under the registration number: Dnr 5416-13-40.

<sup>25</sup> This case with the Chancellor of Justice is filed under the registration number: Dnr 8322-14-30.

by the university college to blur the distinction between the public and the private, in the service of neoliberal policy, and she also, in a surprising move, issued general guidelines that could block such future attempts. This would count in favor of open public access. On the other hand, the Chancellor did not formally criticize the university college or demand that the policy be revised. This tension in the decision, which could be interpreted as a minor concession to neoliberalism, would count in favor of no open access.

As a matter of fact, all our predictions turned out to be true. As for the communication policy decision, the Chancellor chose, after all, not to make it openly accessible on the homepage. While the ruling affects speech policies at some 300 public agencies, it was thus "judged not to be of general interest".

#### **7. Discussion and Conclusions**

We have argued that the neoliberal tradition and new public managemen<sup>t</sup> reforms of the public sector effectively erode the core (liberal) democratic values of the rule of law and transparency. The tension between public law and managerially-influenced governmental policy is, in practice, resolved by introducing what we referred to as "shadow management" in public administration, whereby managerial decisions that clash with constitutional and administrative law are dealt with in internal memos or consultancy reports and hidden from public view. The consequence is the rise of a double standard in the public sector, which potentially reduces public trust in institutions undermining their democratic legitimacy. While our case study came from the Swedish higher education sector, we conjecture that the phenomena identified are general in nature. However, further research from other sectors and countries would be needed to establish this general claim.

The institution of the ombudsman, while playing an important role of supervising state agencies in the normal state of things, has been criticized from a number of perspectives in the literature [44]. One concern is that the role of the main ombudsmen in Sweden is not only to supervise the administrative agencies but also to supervise the courts. This can lead to peculiar situations where the ombudsman is prosecuting someone before a court, and the same court is under the ombudsman's supervision regarding how the case is handled, potentially threatening the independence of the court. Internationally, however, ombudsmen rarely play the role of overseeing the courts.

The second type of criticism is, in a sense, the opposite of the first. The point, which we touched upon above, is that the Ombudsmen do not have enough power in their hands to actually play a regulating role. Recent statistics indicate that the Parliamentary Ombudsman dismisses 50 percent of the complaints without any investigation. Of the actually investigated cases, less than 10 percent result in some form of "criticism". A public official who is criticized can, in principle, carry on "business as usual" without there being any form of severe consequence. Only a few cases per year actually lead to prosecution.

Finally, the institution of the ombudsman, as it is interpreted in Sweden, has been criticized for political dependence on either the parliament or the government. It has, therefore, been suggested that at least the Parliamentary Ombudsman should be converted to a public Ombudsman free of political dependence [44].

The critique we have raised against the Swedish ombudsmen connects to, but also importantly deepens and extends, the last type of objection. We argued that when governmental neoliberal policy clashes with legal requirements, the likely effect is that the ombudsmen, including the governmental Chancellor of Justice, become double-edged swords. Rather than monitoring law-abidingness in the public sector, the ombudsmen take on the new deceptive role of providing pseudo-legal justification for neoliberal reform, making neoliberalism and ombudsmen a particularly dangerous mix from a democratic and legal perspective. We supported our contentions by a case study of Swedish higher education, drawing on documentation and public records collected by an NGO, and hypothesized that the mechanisms we highlight are general in nature, leaving the assessment of the wider validity of the hypothesis for future research. While it is not the purpose of this article to provide policy recommendations, we would conclude that the market principles inherent in NPM are highly unsuited to many parts of the public sector, including higher education.

**Author Contributions:** Both authors have made substantial contributions to the article in all stages, including conceptualization, method, empirical analysis and interpretation. J.S. has been somewhat more involved in conceptual-theoretical framing and design while E.J.O. has been somewhat more involved in legal document analysis. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors wish to thank Magnus Zetterholm for continuing valuable discussions on the subject.

**Conflicts of Interest:** The authors declare no conflict of interest.
