Constraining institutions, committed employees: unpacking the procedural legitimacy of welfare institutions in Sweden and Denmark

Tax-financed welfare institution depends on public legitimacy to be sustained. While the level of welfare institutions legitimacy has been investigated extensively, the contents – the reasons people have for considering welfare institutions legitimate – remain under-researched. This paper addresses this research gap by investigating legitimacy evaluations of central welfare institutions (health care, primary education, and cash benefit systems) among Swedes and Danes. Specifically, the study investigates evaluations of procedural legitimacy in regard to institutional equity and quality and employee competence, commitment, and fairness.

The paper is based on 115 qualitative interviews investigating normative valuations and evaluations of welfare and redistribution, carried out in Denmark and Sweden between 2013 and 2014. The Scandinavian case is of particular interest since institutional legitimacy is high, despite a high level of taxation and redistribution.

Initially, the notion of institutional legitimacy is unpacked, identifying the role of procedural justice and legitimacy within this wider field. Secondly, research questions and methodology is presented. Thirdly, the study findings in regard to equity, equality, competence, commitment, and fairness are presented and, finally, conclusions are drawn.

Rothstein (1998) identifies three distinct types of legitimacy relevant to welfare states and welfare institutions. First, the systems of redistribution must be perceived to pursue a just goal. Secondly, the distribution of burdens associated with redistribution requires popular legitimacy. Thirdly, the procedures for identification of needs and distributing wealth and opportunities must be seen as just.

The goal legitimacy of both welfare states and welfare institutions has had ample treatment in survey research, demonstrating that goal legitimacy is generally highest for encompassing, universal institutions and institutions catering to those deemed highly deserving, compared to means tested and selective institutions (Larsen 2006; Rothstein 2008; van Oorschot 2000, 2005). The distribution of burdens and benefits has also been investigated in regard to the legitimacy, finding that in universal welfare states the legitimacy of taxation and redistribution is higher, compared to conservative and liberal welfare states (Esping-Andersen 1990; Svallfors 2004, 2006; van Oorschot 2005). In contrast, research on procedural legitimacy is quite scarce.

The importance of procedural legitimacy has been stressed by welfare researchers (Svallfors 2006; van Oorschot and Meuleman 2012), but empirical operationalization of procedural legitimacy is usually limited to issues of corruption, benefit fraud, and non-take-up. A few studies have been conducted on procedural legitimacy, but these focus on user experiences, thus giving little knowledge about more general evaluations involved in public perceptions (Kumlin 2004; Kumlin and Rothstein 2005).

However, research on institutional trust or confidence connects strongly to procedural legitimacy. Confidence in institutions reflects people's expectation that they and others will get a fair and equitable treatment, regardless of wealth, status, or connections (Rothstein and Uslaner 2005). Research carried out on this topic predominantly engages issues of fraud, corruption, power abuse, and government legitimacy (Edlund 2006; Kornai et al.2004; Mishler and Rose 1997, 2001; Rose-Ackerman 2001) and suggest that such major flaws in procedural justice heavily impact confidence in – and consequently legitimacy of – institutions. Unfortunately, we know little of public evaluations of procedural legitimacy outside the areas of corruption and fraud. This is, as pointed out by van Ryzin (2011), problematic since citizens appear to care as much or more about the procedure, as they do about outcomes.

According to Rawls (1973), perfect procedural justice requires very clear and simple goals, making it possible to implement perfect rules and procedures. Imperfect procedural justice obtains when unclear, complex goals make implementation imperfect, with consequent injustices. Pure procedural justice involves the establishing of procedures which follow just processes and outcomes defined as just with reference to the processes. Most of the social policy finds itself bridging the two latter definitions, seeking on the one hand to achieve specific policy goals in the best possible way, while on the other hand following the general rules and procedures of public administration which seeks to ensure just processes. This means that rules and procedures on the one hand must be in line with policy goals and draw on legitimacy from the procedural justice of democracy (Rawls 1973; Rothstein 1998), while also being flexible enough to adjust to the specific needs of individual citizens (Sen 2011).

The social policy deals with complex problems, limited knowledge of what works, and great variation across cases. Consequently, precise and just targeting require imprecise rules allowing field personnel a high level of discretion. Procedures have to incorporate the employee ability to adjust measures in each case or run the risk of overly bureaucratic, complex, and internally inconsistent rules (Rothstein 1998). Discretion involves the application of judgment within the bounds of specific standards and under a specific authority (Dworkin 1978). This means that discretion involves a certain degree of freedom while subject to criticism and explicit standards of authority. From this perspective, discretion under the auspices of, e.g. professional standards and legal authority may be conducive to procedural justice and legitimacy.

However, discretion may also be used to navigate the conflicting demands and goals produced by the complex rules, norms, and standards of the social policy, potentially creating unjust procedures through substandard service, limiting access, unequal service like triage or creaming, and normative/classificatory biases (Lipsky 1980). Conflicts may arise in the cross-section of professional standards and institutional demands for efficiency, forcing street-level bureaucrats to choose between being citizen agents and form a professional alliance with the citizens or state agents loyal to the institutional framework and the public authority (Maynard-Moody and Musheno 2000). The legitimacy outcome of either strategy is, however, not straightforward, since client-centered procedures tend to create increased public legitimacy even if they conflict with democratic governance, while institutional loyalty increases democratic accountability.

Surveying the literature, five components of procedural justice emerge as pivotal in securing institutional legitimacy. First, in regard to the formal, institutionalized part of procedures, citizens must have confidence that (1) institutional rules and procedures are equitable (Rawls 1999; Tyler 2003), (2) that institutional procedures achieve institutional goals consistently and with high quality (Barber 1983; Luhmann 1979; Parsons 1967; Rothstein 1998). Furthermore, in regard to the exercise of discretion, legitimacy requires trust in employees to be (3) committed representatives of the institution in question (Garfinkel 1963; Giddens 1990; Kumlin 2004), (4) of unquestionable competence embedded in legitimate professions (Parsons 1967), and (5) fair and incorruptible in carrying out their duties (Morris and Klesner 2010; Rothstein 1998; Uslaner 2008; Warren 2004).

These five procedural justice requirements for institutional legitimacy guide the empirical analysis in this paper (Table 1).

This study aims to understand evaluations rather explain them. Consequently, the sources of information people draw on in evaluations are beyond the scope of this study, unless these are explicitly made part of the evaluation. Rather, this study contributes to filling another gap in our knowledge of institutional legitimacy by investigating the substantial contents of procedural justice evaluations and the relationship between the five procedural justice requirements for institutional legitimacy.

While the aim is to develop a general, analytical contribution to the discussion, the Scandinavian case comprises a particular empirical setting where legitimacy seems to have been successfully obtained, potentially yielding important insight on the procedural prerequisites for achieving institutional legitimacy. Furthermore, the Scandinavian universal case has played a prominent role in the debate on the sources of institutional confidence and generalized trust (Rothstein 2001; Rothstein and Uslaner 2005).

The two countries represent core characteristics of the universal welfare regime: tax-based, encompassing, generous welfare institutions covering most types of needs and life stages based on residence or citizenship. The three institutional areas included in this study are mostly similar between countries: health care and education is a universal entitlement in both countries and social security is means tested. Institutionally, the main difference is that unemployment benefit in Denmark is left in the charge of labor unions through insurance funds, whereas these are nationalized in Sweden. In both countries, active labor market polices have been implemented to a very extensive level. Welfare reform trajectories differ between countries. Denmark has historically had a higher level of decentralization and involvement of market and civil society in welfare provision, whereas Sweden has largely opted for centralized, state institutions (Knudsen and Rothstein 1994). In the last three decades, Sweden has chosen market strategies both in regard to out-sourcing of health care and marketization of education – including the controversial introduction of for-profit primary schools – to a much higher degree than Denmark. Denmark has followed the market-path later and to a more limited extends. Welfare professions remain educationally specialized with strong professional claims on their individual areas of expertise and the associated jobs in both countries.

The interview study was carried out in 2013–2014 which investigated Swedish and Danish valuations and evaluations in regard to welfare and redistribution. This paper engages perceptions of public primary education, public health care, and unemployment benefit/social assistance representing three core welfare issues: health, education, and social services. Interviewees were asked to evaluate both the institutional procedures and the employees within each of these institutions – in that sequence.

In the analysis, the following research questions were addressed:

• What are the dominant evaluations of procedural justice in regard to institutional equity and quality and to employee competence, commitment, and fairness in welfare institutions in Sweden and Denmark?

• Which types of justifications are employed in these evaluations?

A total of 115 semi-structured interviews (54 in Denmark and 61 in Sweden) were carried out face to face with an average duration of 1½ hours. To secure comparability, the interview guide provided a structured set of general themes, a number of mandatory subthemes and suggested probing questions in regard to each mandatory theme. This resulted in an interview with a clear agenda presented by the interviewer, while also giving the participant both time and opportunity to make detours, elaborate, and explain reasons and deliberations.

To be eligible for the study, the participants had to be employed because the interviews also connected to work and taxation. Background variables were collected on participants to ensure that the study population was balanced in terms of gender and social background. Following Lamont's (1992) neighborhood methodology, participants were contacted on the basis of residency in working class and middle/upper-middle class neighborhoods in the capitals Stockholm and Copenhagen and the industrial cities Gothenburg and Aalborg. Potential participants were contacted by telephone (both mobile and landline).

The inductive analysis identified prevailing patterns of evaluations and justifications, taking as a point of departure the analytical categories institutional quality and equity and employee competence, commitment and fairness (Alvesson and Kärreman 2011; Fereday and Muir-Cochrane 2006). The analysis involved identification of different evaluations and justification characteristic to each of these five analytical categories. This analysis started from descriptive coding of evaluations assigning them to one or more of the analytical categories. These five clusters of evaluations were then analyzed to identify both the content of the evaluation and the justification offered, if any. The derived set of justifications and evaluations was then employed in the second round of coding of the entire dataset to ensure consistent coding (Bazeley 2013; Saldaña 2015). The final analysis was carried out using coding matrices (Miles and Huberman 1994) of both descriptive and analytical codes employing cross-code queries in analyzing the interviews (Chamberlain 2006). The analysis presented here is focused on general tendencies within the data, though differences between country cases are presented where relevant. The analysis and the patterns identified in this study are grounded in the entire dataset and quotes have been selected within relevant codes to best represent and communicate the patterns of meaning identified in the analysis.

The following presentation of the study findings is divided into five sections reflecting the five main conceptual categories relevant to evaluations of procedural justice: institutional equity, institutional quality, employee competence, employee commitment, and employee fairness. Within each section, general findings connected to each of the three welfare institutions are presented as well as country differences where relevant.

Most of the Swedish and Danish participants have a confidence in the quality of public primary schools and public health care, even if most also express some type of reservations or critique of system quality. Educational practices are mostly evaluated positively in regard to quality of teaching delivered by primary education. Furthermore, the universal, tax-financed solution are thought to achieve more and better learning, preparing the pupils for democratic participation through inclusion in primary school and equipping students with the knowledge and competences necessary for them to succeed in the market through easy access to high-quality secondary and tertiary education.

Skilled health care worker, Denmark, Age 57:

Similarly, the health care system is perceived to be able to treat illnesses effectively both at the level of general practitioners, specialists, and hospitals. Here, however, there seems to be little difference between the two country cases in regard to these positive evaluations.

Nonetheless, what justifies these positive evaluations of quality to most participants is ultimately the competence and commitment of the employees, rather than characteristics of the education and health care systems as such. This will be investigated further in the sections on employee competence and commitment. Most evaluations ascribing high quality to institutions point to institutional goals and in some cases the distribution of burdens and benefits, such as universalism. The quality of the institutional procedures, in contrast, is given a rather harsh critique by many participants when delving into the details of how quality is actually delivered.

Many participants evaluate different administrative and market reforms as potential threats to quality. Generally, the interviewees perceive most of the changes in administrative procedures and in organizational structure to serve new public management purposes such as control and efficiency, rather than high-quality service in line with institutional goals. Such changes are seen as fundamentally disruptive to the quality of welfare institutions. This critique is in some cases directed at a perceived reduction in funding of welfare institutions which is thought to be more or less directly linked to the ambitions of efficiency and accountability driving policy reforms.

Primary school teacher, Denmark, Age 34:

Architect, Denmark, age 38:

Primarily participants from both Denmark and Sweden consider the health care system and public primary schools to be under pressure from insufficient funding and administrative reforms trying to increase efficiency at the expense of quality in service delivery. In the Swedish case, however, quality in the primary education system is furthermore thought to be under pressure from the advent of for-profit primary schools creaming pupils from public schools and fueling cut-backs in public spending on education. Similar evaluations can be found in the Danish data, but less frequent and voiced less strongly. This may be because private schools are strictly non-profit in Denmark.

In comparison, confidence is low in the quality of the two cash benefit systems even at the general level of the goals pursued in both countries. While participants have confidence in the correct and timely pay out of benefits, confidence in employment services is very low.

Clerk, Denmark, age 30:

Both systems are thought incapable of helping people into employment because they lack connections to potential employers, they waste the time and resources of the unemployed through ineffective training, activation programs, and ritualized tests and requirements, cumbersome administration renders the systems inflexible and inefficient, and they lack resources. The critique of the systems inability to get people employed is more prominent in the Danish case, whereas critique of the system for being pointless and cumbersome is equally prevalent in both Sweden and Denmark. The failings of these cash benefit systems are in part ascribed to institutional consequences of active labor market policies, which move focus away from the goal of helping the benefit recipients solve social problems and regain their footing.

Most respondents also express confidence in the equity of the schools and health care systems at the level of goals and distribution of burdens and benefits – including the principles of equal and universal access on which the institutions are built. However, many participants evaluate institutional equity negatively at the practical level. Interviewees criticize health care for de facto inequity in access to and quality of services. System complexity, resource shortage, and public funding for private care is thought to keep the less educated and resourceful people from getting the appropriate level of service and make the system impenetrable to them. Essentially, procedural inequities in the health care system are thought to reproduce social inequalities despite the manifest institutional goals of equality and universality.

Communication consultant, Denmark, Age 41:

School system inequities are claimed to emerge when differences in socio-demographic composition of pupils between schools threaten equal access to high-quality education and hampering social mobility. The system should – but fails to – rectify this by directing additional resources to schools with special challenges and, consequently, inequities in the access to high-quality education obtain, it is argued. Furthermore, increased privatization is thought to increase such inequities because the privileged can move their children to private schools, leaving public primary schools struggling to help children of the underprivileged and immigrants. This is an evaluation more prominent in Sweden than in Denmark, where respondent in the latter country sees the success of private schools as a symptom of public school failure, rather than a cause.

Most participants consider the cash benefit systems to perform inequitably, reproducing socio-economic differences as differences in the quality of services offered to – and the harshness of requirements made of – the individual benefit recipient. Even more consider these systems fundamentally inequitable because they operate haphazardly and treat everyone more or less unfairly. Many see this as a result of a disassociation between the actual goals of these institutions (securing employment and income replacement) and the ambition of policy changes aiming at punitive and controlling goals instead (active labor market policies and continuous tests of the benefit recipients willingness to work).

Unskilled sports instructor, Sweden, Age 55:

Confidence in the equity of these systems is markedly lower than in the educational and health care systems, even at the general level.

Turning to employee components of procedural justice, the first issue is competence. Employees within public health care and public primary schools are trusted to be highly competent by most interview participants, in both countries. School teachers are trusted to be knowledgeable of the subjects they teach, skilled in imparting that knowledge on the pupils. Some draw on their own experiences or those of their social network in making these evaluations, but most refer to the level of training required to qualify as a primary school teacher, which justify trust in the competence of teachers to many interviewees. Similarly, the participants generally express a high level of trust in health care employees in regard to the technical skills, the ability to administer efficient treatment, and care for patients. Again, the trust in the health care professionals is justified by the interviewees reference to the level of training that they have received and, by some, the positive experiences.

High school teacher, Sweden, age 38:

Skilled health care worker, Denmark, age 57:

While some participants are worried about the quality of teaching in primary schools and the risk of medical malpractice in public health care, such shortcomings are not ascribed to lacking employee competence. Rather, they are seen as a result of the time and resource constraints imposed on the employees by reforms of administrative practices and policies increasingly institutionalizing control and documentation procedures, thus reducing the amount of time employees can spend on carrying out their main tasks to the best of their ability.

Skilled social worker, Sweden, age 59:

In regard to the cash benefit systems, the level of trust in employee competence is generally much lower than in employees within primary education and health care. A sizable group of participants in both countries consider the employees incompetent in regard to unemployment services – more than social work – because the respondents deem the case workers least competent in this area. Such evaluations are rarely explicitly justified, but draw to some extent on media coverage and hearsay.

Nurse, Stockholm, Age 58,

Nonetheless, participants in many cases attribute these problems to system failure rather than employee incompetence. The constraints and irrationalities institutionally imposed on the employees result in a stressful and resource-starved work environment. Trust in employee competence is high in regard to two universal welfare institutions and markedly lower in regard to the selective institution, even if evaluations of employees as outright incompetent are rare.

Employees within all three institutions are trusted to be highly committed to their job and the associated professional standards. However, this trust is justified somewhat differently in connection with each of the three welfare institutions. In regard to the cash benefit systems, the participants consider employees to be committed and strive to do the best they can under the given circumstances. More than any professional commitment, this evaluation is justified by a general expectation that employees do their job to the best of their ability, making motivation and commitment beyond question.

Owner and CEO of a construction company, Copenhagen, Age 50:

Trust in the commitment of professionals within public health care and public primary schools is based on different evaluations. Health care and primary school professionals, in particular nurses, doctors, and teachers, are thought by many to be both underpaid and undervalued. The participants emphasizes that people in these professions are highly committed to their work and describe it as a vocation, more than a salaried job. Furthermore, within education and health care, the integrity and commitment of employees are thought to ensure institutional goal attainment despite deteriorating institutional structures and ruinous reforms.

Clinic assistant in partner's dental clinic, Copenhagen, Age 46,

In contrast, cash benefit systems employees are thought disempowered in the face of institutional pressure. Their commitment is thought to be of little use because the systems are fundamentally flawed, pursuing goals which the interviewees consider illegitimate. The evaluations of employee commitment differ between the means tested and insurance-based cash benefit institutions and the universal institutions in two regards. First, trust in the commitment of employees within the cash benefit systems is justified with reference to a general work ethics, whereas trust in the commitment of educational and health workers is additionally thought justified by their vocational and professional drive. Secondly, the employees within the cash benefit systems are seen as failing due to the flawed institutional pressures they are submitted to, whereas health and education works are thought to counteract flawed institutional pressures due to sheer strength of commitment and professional ethics.

Evaluations of employee fairness are generally positive within all institutions, even if constrained by institutional structures and policy reforms. However, fairness is thought more prone to succumb to human frailty and error than employee competence and commitment. The fairness of the teachers and health workers is a somewhat ambivalent issue. Most participants consider these employees fair or striving to be fair, while some concede that they can be prejudiced and biased and may unintentionally treat people unequally. However, the major concern is that employees cannot treat everyone fairly, due to the complexities and constraints of institutional structures and administrative procedures. This unfairness is seen as an exception from a general rule of fairness among teachers and health care workers and is thought not to change the general positive evaluations of employee fairness.

Bartender, Sweden, Age 25:

Cash benefit system employees are thought to strive for fairness both in terms following legal requirements, giving people the right type of benefits, and seeking to help everyone regardless of socio-economic status. While some suggest that the employees do not always manage to be entirely fair, this is seen as exceptions caused by human imperfection and the irrational and counterproductive constraints imposed by institutional structures.

Unskilled sports instructor, Sweden, Age 55:

While evaluations of administrative constraints are similar in all three institutional areas, evaluations of the fairness of cash benefit systems employees are less positive. They are generally thought to have less power to employ professional and personal values and ambitions in their work in the face of institutional pressures, compared to health and education workers. Consequently, the effects of inefficient and irrational institutional structures are more forcefully implemented as unfair practices.

The dominant evaluations in both countries are that institutions are seen as fundamentally legitimate, but institutional procedures are generally evaluated more negatively than the elements linked to employees. The citizens seem to favor professional discretion over democratic governance and precise institutionalizations of rules and procedure, in part because they see professions and fieldworkers as better keepers of institutional goals than bureaucracies’ chains of command.

In regard to the universal institutions – public health care and public primary schools – Danes and Swedes expect high-quality services from people who are technically competent, committed to their work, and striving to be fair. This includes the expectation that they are making discretionary decisions based on professional standards in line with institutional goals. Here the interviewees seem to side with Rothstein and Dworkin that procedural justice requires room for fieldworkers to make discretionary decisions, rather than standardized procedures. Evaluations of institutional procedures, in contrast, suggest that equity and quality are in decline and threatening the goal legitimacy of institutions. Policies emphasizing control, documentation and efficiency are thought to encroach upon the quality and equity of these services undermining institutional legitimacy – the conflicting goals and procedures characteristic to Rawls notion of imperfect procedural justice and Rothstein characterization of overly rigorous rules and procedures within complex policy fields. In regard to procedural justice, the institutional legitimacy of these universal welfare institutions is justified by the perceived intersection of strong professional competences and norms with vocational drive and general human fairness, despite deteriorating institutional quality and equity.

In contrast, a sizable group of participants in both countries who consider the cash benefit system employees incompetent to ensure that institutional goals are achieved, despite being generally committed and fair. In particular, they are unable and incompetent to help benefit recipients into employment, whereas employees doing traditional social work are evaluated as more competent. The legitimacy of employees seems to be split between the changing institutional goals of the cash benefit systems: employee ability to make sound discretionary decisions is only acknowledged in regard to social work goals, but not the ‘work first’ goals increasingly characteristic of Danish and Swedish social policy. Nonetheless, a majority of the participants predominantly attribute the perceived failings of the systems to constraints and irrationalities institutionally imposed on employees, resulting in a stressful and resource-starved work environment. Here it is questionable whether procedural legitimacy does in fact exist, but to the degree that it does, it is dependent on trust in general human work commitment and fairness combined with the specific, but inadequate competences of social workers. In the absence of a clear professional standard of discretion, the good intentions of employees become pivotal in judgments of procedural justice: the interviewees perceive cash benefit employees to opt for the citizen agent strategy, rather than the state agent strategy. This supports Rothstein's (1998) contention that more complex problems require a high degree of flexibility and autonomy. The respondents characterize the cash benefit institutions as the most rigid and constraining. Nonetheless, participants appear to prefer discretionary flexibility over bureaucratic control in all areas.

The type of justifications used seems to be relatively consistent across the three institutional areas, even if the result of the evaluation differs. Evaluations of institutional equity and quality are justified by whether institutional procedures seem capable of achieving institutional goals. The introduction of performance management and documentation systems is perceived as invasive and disruptive to institutional procedures in regard to quality – moving focus and resources away from goal attainment. Similarly, the introduction of welfare markets and geographical variations in services is thought harmful to institutional justice because equal access to services of uniform quality is threatened, undermining principles of pure procedural justice. These justifications distinguish between the legitimate, substantive welfare goals of institutions and illegitimate, administrative goals which new public management and the performance movement aim to achieve. This supports van Ryzin (2011) suggestion that institutional procedures are as important to citizens as outcomes. The present study shows that people are concerned about procedures both in regard to ethical aspects of equity and in regard to the quality and equality of outcomes. Interviewees are concerned with both the connections between procedures and institutional goals and the more general moral justifications of procedural practices.

Comparing the two country cases, there appears to be a more explicit criticism of increased market involvement in the Swedish case than in the Danish, presumably reflecting the more pro-market reforms of the Swedish social policy in the last few decades.

Similarly, the evaluations of employee aspects of procedural justice are more or less the same across all three institutional domains. Only the perceived lack of competence among cash benefits system employees diverge from the common trend. Evaluations of employee competence in general refer to professions: doctors, nurses, and teachers who are professional groups are considered highly competent. This assumption of competence apparently requires little justification in terms of experience, but is either assumed to be self-evident or is justified with reference to their education or their commitment. Interestingly, the perceived employee imperviousness to the illegitimate influence from welfare institutions builds on educational and professional institutions. This supports traditional claims regarding the role and power of professional norm systems in securing confidence in institutions (Kuhlmann 2006; Parsons 1967; Pellegrino et al.1991). Trust in employee commitment seems equally important and two types of justification underpin trust in employee commitment. The first, used across all three institutional domains, posits that it is human nature to do your job to the best of your ability. While this connects well with the literature on work orientations in Scandinavia (Hult and Svallfors 2002; Svallfors 2006; Svallfors et al.2001), it more generally connects to issues of trust in people and the assumption of shared norms within national communities (Sztompka 1999; Uslaner 2002). In the second, used within universal welfare institutions, compound vocation with general notions of work commitment. In this context, vocation seems to indicate that these professions are somehow beyond the limits of self-interest and commodified work ethics. Trust in employee fairness is justified in much the same way as trust in their commitment: all people are assumed to pursue fairness in their work. This trust in fairness as human nature connects well both with the literature on the high level of generalized trust in Scandinavia (Delhey and Newton 2005) and nations as moral communities where expectations of norm-driven behavior are assumed (Sztompka 1999; Uslaner 2002). The trust in employees seems generally founded in the professional competence to make good discretionary decisions even under institutional pressures – the citizen agent strategy of the street-level bureaucracy literature – but also in fundamental human nature. Employees are not only making discretionary decisions as citizen agents but also as citizens themselves.

The main finding of this study is that the procedural legitimacy of welfare institutions hinges on trust in employees as professionals and as humans guarding institutional goals. Institutional procedures, in contrast, are thought to become increasingly misaligned with institutional goals due to administrative reforms focusing on performance and accountability. This has serious policy implications since efforts to assure performance and accountability seems to undermine the procedural legitimacy of institutions, while the instrumentalization of employee judgment and decreasing professional autonomy infringes on the solid base of trust in employee discretion which upholds procedural legitimacy. Swedish and Danish citizens seem to agree with Rothstein's contention that procedural justice in complex systems in the final analysis hinges on employee discretion rather than procedural rigor. Emphasizing work commitment and human nature, however, they do not agree that normative commitments outside professional judgment are as irrelevant as claimed by Rothstein. Moreover, the employee coping strategies suggested by the street-level bureaucracy literature do not surface in interview evaluations of these welfare institutions, suggesting that they play a minor role in public perceptions of – and discourse on – welfare institutions compared to the system failures and constraints.

No potential conflict of interest was reported by the author.

Morten Frederiksen obtained his PhD in Sociology from Copenhagen University in 2012. He is currently Associate Professor at the Department of Sociology and Social Work, Aalborg University where he teaches sociological theory and research methods. His research interests are values and the process of evaluation – in particular trust, solidarity, and the symbolic boundaries between state market and civil society. His research has been published in international journals such as Current Sociology, Time and Society, and Acta Sociologica.

Morten Frederiksenhttp://orcid.org/0000-0003-2635-3827