Introduction
The literature on policy implementation has placed great emphasis on the degree of autonomy possessed by street-level bureaucrats and the impact of this autonomy on the implementation of public policies. For instance, Lipsky (1980) and his proponents have documented the development of various coping mechanisms whereby street-level bureaucrats modify their job conception to reduce discrepancies between the demands of citizens and their limited resources (Jessen and Tufte, Reference Jessen and Tufte2014; Van Berkel and Van Der Aa, Reference Van Berkel and Van Der Aa2012). Conversely, critiques of this view have argued that the autonomy of street-level bureaucrats has been limited by increased managerial control and surveillance (Clarke et al., Reference Clarke, Gewirtz and McLughlin2000; Evans, Reference Evans2010; Evans and Harris, Reference Evans and Harris2004; Harris and White, Reference Harris and White2009).
In this paper, we argue that despite being subjected to forms of managerial control and regulation, street-level bureaucrats (i.e. local managers and front-line workers) play a central role in public reforms. According to this view, street-level bureaucrats are not passive agents who are forced to comply with reform changes; in fact, they may attempt to modify perceived discrepancies in various ways. A central aim in this paper is to highlight such modifications. We do so through the concept of ‘institutional work’, i.e. ‘the purposive action of individuals and organisations aimed at creating, maintaining and disrupting institutions’ (Lawrence and Suddaby, Reference Lawrence, Suddaby, Clegg, Hardy, Lawrence and Nord2006: 215; Lawrence et al., Reference Lawrence, Suddaby and Leca2009). As most public sector reforms can be perceived as attempts at institutional transformation – i.e. the simultaneous deconstruction of an old institutional order and the building up of a new one (Christensen et al., Reference Christensen, Fimreite and Lægreid2014; Lounsbury, Reference Lounsbury2002) – the concept of institutional work highlights the role of agency in influencing and responding to such institutional processes.
In our analysis, drawing on empirical data from the organisational implementation of a retirement pension reform in the Norwegian Labor and Welfare Administration (NAV), we focus on the institutional work undertaken by street-level bureaucrats in responding to the reform and its organisational implementation. In so doing, our approach responds to calls for approaches that put street-level organisations at the forefront of the analysis and which apply insights from (institutional) organisational theory to grasp how these organisations make use of their autonomy (Brodkin, Reference Brodkin, Brodkin and Marston2013; Winter, Reference Winter, Peters and Pierre2012: 273). We thereby aim to add to and nuance discussions about the impact of street-level bureaucrats in two ways: first, by describing types of institutional work through which street-level bureaucrats modify public policies, and second, by demonstrating the applicability of an institutional work perspective to the study and understanding of front-line work behaviour and its implications for the services.
Perspectives on street-level autonomy
Since the seminal work of Lipsky (1980) and Pressman and Wildavsky (Reference Pressman and Wildavsky1984), there is general agreement among public policy scholars that the services provided at street level typically diverge from the intentions of policy-makers and that much of this deviation results from the discretion of street-level bureaucrats (Brodkin, Reference Brodkin, Brodkin and Marston2013; Hupe et al., Reference Hupe, Hill and Buffat2015; May and Winter, Reference May and Winter2009; Meyers and Vorsanger, Reference Meyers and Vorsanger2003; Riccucci, Reference Riccucci2005). Autonomy involves the degree of freedom from organisational authority; for instance, the ability of street-level bureaucrats to provide a different type of service delivery from what is imposed by higher levels. The concepts of autonomy and discretion overlap, but the concept of discretion often points to more specific decision-making processes in which street-level bureaucrats interpret and apply rules and regulations (Hupe et al., Reference Hupe, Hill and Buffat2015: 17).
For instance, Hupe (Reference Hupe2013) makes a distinction between ‘discretion-as-granted’ and ‘discretion-as-used’, where the latter points to the specific practices of street-level bureaucrats regarding their services to the citizens and the former provides the context or degree of freedom (Hupe et al., Reference Hupe, Hill and Buffat2015: 17–18). The image of the hole in the doughnut – an autonomous area of discretion surrounded by restriction (Dworkin, Reference Dworkin1977: 31) – is often used as a metaphor for these differences. Overall, the literature indicates that the use of discretion varies according to individual (e.g. job experience), organisational (e.g. type of front-line agency, available resources), and institutional (e.g. types of welfare regimes) factors (for an overview, see Brodkin, Reference Brodkin, Brodkin and Marston2013; Maynard-Moody and Portillo, Reference Maynard-Moody, Portillo, Durant and Edwards2010).
Lipsky's (1980) work on street-level bureaucrats suggests that the street-level deviance from policy intentions stems from the existence of ambiguous rules or the lack of administrative control; in other words, from the existence of a void with regards to the actual services to be filled by the street-level bureaucrats. More specifically, Lipsky has pointed to the development of coping mechanisms wherein street-level bureaucrats modify their job conception to reduce the discrepancy between what they are supposed to do and what they actually do.
Critics of applying Lipsky's work to contemporary service organisations have, however, argued that autonomy has all but disappeared due to increased management of the welfare services and, by extension, increased control and surveillance of front-line workers (Clarke et al., Reference Clarke, Gewirtz and McLughlin2000; De Bruijn, Reference De Bruijn2011; Harris and White, Reference Harris and White2009). Street-level bureaucrats have traditionally possessed considerable freedom with regards to the performance of their work (Engel, Reference Engel1970). A central argument is that managerial and technocratic systems have hindered development of workplace conditions that enable this kind of professional autonomy of front-line workers (Harris and McDonald, Reference Harris and McDonald2000; Jones, Reference Jones2001). Further, Brodkin (Reference Brodkin, Brodkin and Marston2013) has shown how, because of limited resources, street-level bureaucracies may organise themselves in ways that limit assistance to citizens.
Other studies have emphasised how street-level bureaucrats do, in fact, possess considerable autonomy within contemporary welfare service organisations (Hupe and Hill, Reference Hupe and Hill2007; Jessen and Tufte, Reference Jessen and Tufte2014). A common example of this is activation policies, which are said to involve a considerable degree of discretion through the tailoring of services to suit individual needs (Fletcher, Reference Fletcher2011; Van Berkel and Van Der Aa, Reference Van Berkel and Van Der Aa2012). Studies have also emphasised how front-line workers increasingly operate as counsellors and facilitators, rather than as administrators of formal regulations (Stjernø et al., Reference Stjernø, Jessen, Johannesen, Ohnstad, Rugkåsa and Ylvisaker2014).
In the context of these debates, we need to know more about the possibility for front-line workers to create space within which perceived discrepancies in policy implementation could be modified or mediated. This is particularly interesting because it opens a view of autonomy not as something given, and thus inherent in the work context of street-level bureaucrats, but rather something that may be enacted and re-enacted in their work practices (Evans, Reference Evans2010; Lehmann-Nielsen, Reference Lehmann-Nielsen2006; May and Winter, Reference May and Winter2009). On this basis, we now turn to an elaboration of the concept of institutional work.
An institutional work perspective
The concept of institutional work has gained increased interest among organisational and management scholars as a response to what has been perceived as an oversocialised view of agency in neo-institutional theory (Powell and DiMaggio, Reference Powell and DiMaggio1991). The approach is concerned with agency and, in particular, with the awareness, skill and reflexivity of individual and collective actors in pursuing their interests in institutional contexts (Lawrence et al., Reference Lawrence, Leca and Zilber2013; Lawrence and Suddaby, Reference Lawrence, Suddaby, Clegg, Hardy, Lawrence and Nord2006; Lawrence et al., Reference Lawrence, Suddaby and Leca2009: 4). The approach entails an understanding of institutions as manifested in more or less conscious actions and belief systems and is thus inherently interwoven with practice (Bourdieu, Reference Bourdieu1990).
The development of welfare services is inherently laden with tensions between the political and administrative designers of the change processes and the street-level bureaucrats who implement them (Evans, 2012). On the one hand, central authorities have considerable power as institutional change agents due to their hierarchical empowerment and their responsibility for implementing political decisions (Christensen and Lægreid, Reference Christensen and Lægreid2011; Pollitt and Bouckaert, Reference Pollitt and Bouckaert2011). On the other hand, street-level bureaucrats also possess power through their exercise of discretion in interaction with the citizens.
The very existence of such power relations suggests that the transformation of public policies can rarely be undertaken solely through administrative coercion. Rather, from an institutional perspective, policy changes are dependent on the local institutional work of street-level bureaucrats in ways that support and enact the change processes, rather than undermine them. In this way, there is a distinction between the institutional work of top management and that of street-level bureaucrats. Whereas the former involves the design and initiative of the change processes, the latter are responses to these impulses. It is precisely in these responses that street-level agency is activated.
Lawrence and Suddaby (Reference Lawrence, Suddaby, Clegg, Hardy, Lawrence and Nord2006; Lawrence et al., Reference Lawrence, Suddaby and Leca2009) differentiate between three broad categories of institutional work. First, work aimed at creating institutions involves political work in which actors reconstruct rules, rights and boundaries, configure belief systems, or alter categorizations within the meaning system. In the context of welfare reform, the policy process encompasses most of the political work involved in creating institutions. A considerable part of this institution-creation involves altering the beliefs and meaning-systems of street-level bureaucrats and manifesting these in institutionalised, taken-for-granted practices. This may, for instance, be accomplished by top management through the creation of rules, through provision of skills and knowledge, through auditing and monitoring, or by embedding and routinizing activities. It may also be accomplished by front-line managers and workers by reinforcing these mechanisms for institution creation.
Second, work aimed at maintaining institutions involves supporting, repairing or recreating social mechanisms that ensure compliance. Although institutions have been understood to be self-sustainable (Scott, Reference Scott2001), even powerful ones require maintenance to continue being relevant and effective. This is particularly evident in the work of enforcement agents (Lehmann-Nielsen, Reference Lehmann-Nielsen2006). Institutional maintenance may also involve protecting institutional resources or co-opting groups into compliance (Currie et al., Reference Currie, Lockett, Finn, Martin and Waring2012) or withholding established beliefs and practices despite efforts of institution creation. Importantly, institutional maintenance can be seen as deviance from policy reforms, but it may also function as a constructive and productive factor for the organisation, i.e. as a way to ‘grease the wheels’ of the change processes (Courpasson et al., Reference Courpasson, Dany and Clegg2012).
Finally, work aimed at disrupting institutions involves undermining the mechanisms that lead to compliance. This may involve efforts to disassociate practices, rules or technology from their moral foundation, as well as undermining the core assumptions and beliefs upon which institutions are built (Lawrence and Suddaby, Reference Lawrence, Suddaby, Clegg, Hardy, Lawrence and Nord2006: 235–237), something which may involve processes of legitimation and/or delegitimation (Phillips et al., Reference Phillips, Lawrence and Hardy2004). Still, it can be argued that disruption rarely exists separately in practice during processes of public reform. Hence, if organisational reforms are simultaneously perceived as deconstructing and constructing a new institutional order, combining the institutional work of disruption and creation seems necessary for top management. This is also because top management possesses the authority to disconnect rewards and sanctions from established practices and link them to new rules, routines and forms of performance.
Overall, the concept of institutional work provides an important addition to the literature on policy implementation and street-level bureaucracy because it provides an analytical framework for assessing different types of discretionary practices and how they relate to the administrative goals and intentions – both those that deviate from them and those that support and reinforce them. These discretionary practices may exist interchangeably.
Research design
Case: The restructuring of the Norwegian pension administration
Our case context is that of a major restructuring of the retirement pension services in Norway. Like many other countries in Europe, Norway has recently reformed its pension scheme in response to predictions of an ageing population and thus a lack of economic sustainability. A common denominator in these pension reforms is the strengthening of the economic incentives for longer work affiliation – in accordance with a ‘work first’ agenda – by providing increased freedom of choice regarding the time and degree of pension withdrawal. A result of these changes is that the discretionary process of choosing a desired pension outcome – and by extension the risks and responsibility for obtaining relevant information – has been removed from street-level bureaucrats and placed in the hands of citizens. The shift of responsibility has in turn generated increased ‘pension confusion’ among citizens (Webb et al., Reference Webb, Watson, Ring and Bryce2014) and thus a need not only for information about the various alternatives and new legal concepts, but also for financial as well as digital skills in the decision-making process (Breit and Salomon, Reference Breit and Salomon2015).
In Norway, the implementation of the pension reform, decided by Parliament in 2005 and executed in 2011, was assigned to the newly established Norwegian Labor and Welfare Administration (NAV). NAV was itself the result of a major organisational reform in 2006, which consisted of a merger of the former local National Insurance Administration with the Labor Market Administration and, at the local level, integration with municipal social services. In each of Norway's nearly 430 municipalities, at least one local NAV office was established, operating as a one-stop shop (Askim et al., Reference Askim, Fimreite, Moseley and Pedersen2011). The aim of the NAV reform was to provide more integrated services to citizens and to promote a policy of activation that would increase the level of labour market participation and avoid extended use of public income security, in line with a whole-of-government approach (Alm Andreassen and Aars, Reference Alm Andreassen and Aars2015; Christensen et al., Reference Christensen, Fimreite and Lægreid2014).
To implement the pension reform – i.e. to handle the expected increase of information-seeking and pension applications – a new organisational structure was established. This new structure includes a web site with self-service application and information opportunities, three specialised call centres to support citizens via telephone, email or online chat functions, and five case processing units (so-called ‘pension units’) competent to handle the new pension legislation. Citizens are to use two primary channels to obtain information – the web site and the call centres – and secondarily the NAV offices. Thus, NAV's pension services now constitute a complex system of different organisational units with different tasks and types of expertise (see Figure 1).
Overview of the organisation of NAVs services
The changes denote how the pension reform, which started as a legislative reform, also turned into an organisational reform. Overall, the new pension system in NAV has involved a structural separation of case-processing and information. The case-processing is handled at specialized back-office units through a standardised decision-making process. This is done in order to prevent variation in the case-processing, as pensions are welfare benefits regulated by law and thus dependent on individual rights. Indeed, similar case processing units (and call centres) have been established for all other national insurance benefits in NAV.
Information to citizens is a service, not a benefit, and this task is handled by service workers at the NAV offices and call centres. These service workers thereby constitute the street-level bureaucrats in our case, since they exercise discretion regarding the level of service given to citizens seeking pension advice, based on their judgment of the (information) needs of these citizens. They do this despite the formal interface, which specifies that information provided at the NAV offices is to be a minimum type – i.e. restricted to general information already provided on the website (a so-called ‘nav.no-level’) – whereas the call centres are to be the primary information source involving more in-depth, person-specific information.
Data collection
The article draws on archival data and interviews at various organisational units in NAV. The archival data consist of government policy documents related to the pension reform. This information was collected with the aim of explaining the institutional underpinnings of the change. We also collected archival data related more to NAV and its tasks in the reform, such as allocation letters and evaluation reports (e.g. the Labor and Welfare Administration's own evaluation of the Pension Program). In addition, we have utilised the extensive documentation of the NAV reform provided by a research-based evaluation which has followed the reform from its start-up, as well as prior studies of citizens’ perceptions of the services, their pension knowledge, and their use of the digital self-service system (Fossestøl et al., Reference Fossestøl, Breit, Andreassen and Klemsdal2015; Bergene and Drange, Reference Bergene and Drange2015; Breit and Salomon, Reference Breit and Salomon2015).
The interview data consist of individual and group interviews collected at the three different units involved in pension issues and undertaken in the spring and summer of 2012 (see Table 1). We conducted most of the interviews with employees and managers at front-line units in NAV, i.e. in five local NAV offices and at all three call centres across the country. The interviews were conducted with a semi-structured interview guide covering themes such as competence and ability to provide adequate service to citizens, user experiences and encounters with users, relationships with other units in the NAV, and the impact of the reform on service quality.
Overview of the interviews
We also conducted group interviews in two of the three pension units. Although these units primarily conduct case processing, they also have professional responsibilities and respond to the most complex user inquiries (via the call centres). Finally, we conducted interviews with representatives of the top management of the Labor and Welfare Administration. This was done to obtain insight into the strategic choices as well as the administrative challenges associated with implementation of the pension reform.
Data analysis
The data collection and analysis was based on a multi-level case study design (Flyvbjerg, Reference Flyvbjerg2006; Yin, Reference Yin2009). Although it is difficult to provide an exact overview of the analytical process, two features have been central. First, we looked at the transformation processes through which the NAV administration operationalised and implemented the reform. We drew primarily on interviews with top management and on policy documents, but also on our own prior experience of NAV. This data provided insight into the institutional work of the top management in NAV and the reasons behind it.
Second, we explored the front-line responses, i.e. the practices of the managers and employees when faced with the institutional pressures. We differentiated between actors from various units in NAV, as well as their roles within each of these units. In this process, we also found that the results of the user studies we had conducted generally supported the contentions of front-line workers regarding the quality of the services. This data provided insight into their institutional work.
We coded and systematised the interview data to capture various theoretical dimensions in the responses (Alvesson and Kärreman, Reference Alvesson and Kärreman2007; Locke et al., Reference Locke, Golden-Biddle and Feldman2008). In this process, we linked forms of institutional work to the position of workers and to their work practices in order to gain insight into the likelihood that their work would achieve the intended reform outcomes. In the subsequent two sections we elaborate on the key outcomes of the analysis regarding the institutional work of top management and front-line workers.
Institutional work by the top management: combining institutional disruption and creation
The organisational restructuring of the pension administration meant a breach with the organisation of the previous pension (national insurance) administration. The former administration had been based for a long time on local insurance offices in every municipality, with institutionalised expertise and a professional identity related to competent (face-to-face) guidance and case processing. The reorganisation implied constructing a new institutional order and it took place via the combined work of disrupting the established practices, divisions of labour and professional identity of former insurance workers, and the work of creating new forms of practice and unit presentations.
To disrupt the institutionalised professional identity of the former national insurance offices, top management emphasised a lack of efficiency as the main strategy for delegitimising established practice. This meant that it was inefficient for citizens and NAV alike that the former had to visit local offices instead of being self-served through the web and a specialised call centre. In other words, in light of the technological modernisation, established practice was outdated; excellent service no longer meant face-to-face encounters, but 24–7 availability of information accessed by citizens themselves:
It is efficient for the user and for the NAV to shift enquiries to the call centre or to the local offices. [. . .] We wanted to pry people away from physical meetings at the NAV offices so that we could handle the increased workload [caused by the increase in the number of citizens eligible for old-age pensions]. (Senior manager in the Labor and Welfare Administration)
Moreover, top management's disruptive work consisted of undermining the core assumption that local offices could be sufficiently competent to provide a high-quality pension service. The argument was that the reformed pension legislation was too complicated to be handled by local pension workers. For the NAV offices, this implied that service practices institutionalised over the decades would have to be unlearned. Former national insurance workers were no longer to be experts in insurance legislation. As a senior manager formulated it, ‘We want them [i.e. the local NAV office staff] to stop answering questions they are not supposed to know the answer to.’
Institutional creation, in turn, was implemented through three processes. First, by outsourcing pension-related work from local NAV offices to citizens through introduction of an on-line service system through which citizens can search for relevant information, monitor their pension earnings, simulate alternative pension withdrawals, and then apply for pensions. Outsourcing also entailed transfer of pension advice, pension expertise and former pension workers to specialised call centres and transfer of case processing to the pension units, and the introduction of an extensive interface describing in detail each new unit's work tasks and the division of labour between them.
Second, it was implemented by scripting and standardisation of the advisory tasks. This has been done by restricting the scope of advice provided by NAV offices and the call centre. The formal interface specifies the expected level of competence, the types of inquiries to be answered by the various organisational units, and how the inquiries should be handled.
Third, it was implemented by reskilling front-line service workers. Reskilling involved reallocation of pension expertise from the NAV offices to the pension units, and reduction of the responsibility of the NAV offices regarding pension services. Efforts were taken to ensure a standard level of expertise at the various units of the NAV. For local NAV offices and call centres, this meant providing only a limited service to citizens and not extending the scope of their advice beyond their defined pension knowledge.
Institutional work by street-level organisations: institutional creation and maintenance
The changes by the top management in NAV heavily impacted the autonomy of front-line workers. They responded through other forms of institutional work: institutional creation and institutional maintenance.
Institutional creation – development of new front-line practices and skills
A key form of institutional creation at the street level involved development of new local front-line practices and skills that would meet the new demands of the transformation processes. Importantly, these new practices and skills would underpin the transformation processes undertaken by top management – despite not always being closely coupled to the original intentions.
One example was evident in the newly established call centres. The type of institutional creation at this level was primarily enacted by the local managers, with the blessing of their superiors in the upper echelons of the administration. Specifically, the telephone conversation model, as opposed to the traditional provision of face-to-face services, led to increased emphasis on what was often described as ‘guidance competence.’ This competence entailed a shift away from professional pension expertise toward relational and technological competence (Hutchinson et al., Reference Hutchinson, Purcell and Kinnie2000):
We're looking for a combination [of telephone and pension expertise]. But most have telephone expertise. [. . .] We are very efficient and spend little time on conversations. (Call centre manager)
The new focus on guidance competence not only entailed a shift in the relationship with each user, as relatively more emphasis was placed on the conversation itself, but also a shift towards increased scripting and standardisation of the conversations. As suggested by the quote above, an increased focus was placed on being efficient in the conversations. A call-centre manager formulated it this way: ‘you have to feel comfortable with this [scripted] structure to be able to work here,’ suggesting that this service model was not cut out for everyone. We were told that a typical day could involve up to 20 long conversations. Hence, being able to cope with continuous shifts between various problems and life situations was itself a skill:
When someone calls, they are focused. You have to be too, because you are jumping in and out of people's lives. That's often a heavy burden. (Call centre service worker)
The emphasis on guidance competence and on minimising transaction time also led to the development of new service practices. According to our informants, the formal aim for advisors was to answer 70 per cent of the incoming calls within 30 seconds, with an average call length of five and a half minutes. A common strategy was to cut down on the ‘chit-chat’ parts of the conversations:
People calling about their pensions generally have a lot of time. Hence, the challenge [when approaching or surpassing the average time] is to reject, but reject politely. We are to spend enough time, but not too much time. (Call centre service worker)
Another strategy, developed by call centre workers, but not publicly discussed, was revealed by some of our informants who explained that they had heard about how others had overcome the time limits by hanging up and/or asking the caller to call again. We were also told that citizens who contacted a call centre could be asked – despite official disapproval – to go to the nearest NAV office. This practice was defended locally by arguing that matters were too difficult to explain over the telephone and citizens should therefore have an opportunity to speak to an advisor face to face.
Another example of local practice creation was evident at the pension units, which were to become highly efficient case processors. These units were isolated from direct contact with citizens and had only limited responsibility to enhance the competence of front-line service at NAV offices and call centres. However, for the managers and workers at the pension units, a lack of competence at NAV offices could cause problems, as they had to answer more incoming calls than intended. To reduce this workload, the case processors provided courses for NAV office employees:
If there are new [managers and employees] the focus is mainly on ‘work first,’ and they don't prioritize pensions. We [pension unit workers] thus need to travel around to the NAV offices and inform the staff, because their knowledge is deteriorating. The NAV offices can't maintain it [pension expertise] when they are not working on it on a daily basis. So that's why we have to contribute. (Case processor at pension unit)
Moreover, instead of avoiding citizen contact and focusing solely on case processing, case processors also provided information courses to citizens. Hence, pension units did not leave front-line work to the designated front line, as was the original intention of the top management, but instead created a front-line role of their own. In so doing, they used their autonomy to develop a new local institutional form aimed at remedying deficiencies connected with the transformation processes implemented by the central authorities.
Institutional maintenance – protecting status quo, and legitimizing higher service level
Maintaining institutional work was particularly evident in efforts to preserve existing competence structures at the NAV offices, despite the top management's efforts to restructure it.
The increased scripting of the services implied that the NAV offices were expected to answer only a very minimum number of inquiries, and instead direct people to the self-service solutions (web, email, online chat, or telephone). As a top manager formulated it (somewhat ironically): ‘We expect the front-line staff at the local NAV offices to point with their hands in two directions: one towards the computer and the other towards the telephone.’
Nevertheless, there were still former pension specialists from the National Insurance Administration (NIA) located at the local NAV offices. Although they wanted to provide a high level of service to citizens who visited the offices, they were restricted by the formal service interface.
Such tensions are illustrated in a discussion between a former NIA specialist, her supervisor and a fellow service worker during a focus group. The fellow service worker explains that the NAV office workers feel uncertain about the instructions in the formal interface and how far they can go in giving advice to citizens. She also says that most inquiries by citizens are directed to the former NIA employee. NAV offices are allowed to assist citizens in calculating their pension level. Moreover, they are to advise citizens to contact pension advisors outside NAV – such as private pension consultants – for answers to their questions. Hence the service worker argues that ‘they [the citizens] have to figure it out for themselves.’
The former NIA employee continues as follows: ‘For the most part, we have had the pension expertise at the NAV office all the time. I think users have received decent assistance with pension calculations, which is the kind of information we are required to provide. But I have taken matters further because I know this stuff so well.’ To this the supervisor responds that the former NIA employee has probably provided more information and guidance than the formal interface instructs the NAV office to do.
Rather than attempting to silence workers who argue for more pension expertise at the NAV offices than the interface prescribes, local managers instead undertook institutional maintenance work by trying to have former specialists at the offices to ensure what they felt was adequate local pension expertise. As several of them argued, inquiries from citizens who came to the offices were simply too demanding to be handled at the level of minimum pension competence:
Fortunately we have had employees with the relevant experience. All the pension experts were transferred to the pension units [as case officers]. We have NN, who knows a lot, and we have also had some others. . . (NAV office manager)
In addition to this managerial institutional work, front-line staff also engaged in maintenance practices. For instance, the former pension specialists could act as local experts, not only in their own office, but also in nearby offices. According to several of our interviewees, they provided teaching or support and were more broadly assigned the task of keeping updated on legislative changes, or taking over tasks from other service workers at local offices. Others reported how they used informal networks of former colleagues with specialist competence now working at the pension units, thus ensuring that they could talk to someone whom they knew.
Overall, this type of institutional maintenance work was initiated and supported by the local managers. Although deviating in part from the formal interface and, by extension, operating against the organisational change processes in the reform, this institutional work nevertheless supported the change processes by ensuring an ability to provide services they experienced as important to specific users.
Another central form of maintaining institutional work involved deviating from the scripted level of service. Such deviations were often the result of the changes in the level and type of guidance and, by extension, of ambiguity regarding interpretations of new work operations. For instance, at the NAV offices, the scripting process involved increased uncertainty regarding the extent of the services workers were to provide. In this sense, there were tensions between the formal scripts and the needs of the citizens coming to the NAV offices to ask for help. A local manager commented as follows:
It's challenging for us because we have been used to providing a certain level of service and now we are supposed to tone it down. [. . .] So finding that level, finding the appropriate level of service for the NAV office, that's something we have to learn.
For the most part, deviating practices were ways to ‘bend the rules’ to serve users – and thus they were constructive, rather than destructive (Ellis, Reference Ellis2011). Moreover, rather than contracting their level of service to cope with the demands (Lipsky, 1980) they expanded it. Such practices included efforts by managers and service workers at NAV offices to spend more time on users than formally required. For instance, a local office manager explained how he had recently spent more than an hour with an older woman who needed extensive guidance on her pension inquiry. This kind of service is clearly outside the formal requirements of the NAV offices.
Such ways of dealing with the tensions between the administrative demands and the user demands indicated the ability to combine formal administrative demands with a higher level of service. Yet, in practice, it also indicated the feeling that there was not really any other choice regarding these users. Accordingly, workers at the NAV offices emphasised the citizens’ need for face-to-face contact, and hence their own crucial role:
Many citizens refuse to make that phone call to the call centre. They want to look someone in the eye. We especially see that among older people; it's the physical contact; they must be allowed to speak with someone. (NAV office worker)
Some of the managers we spoke with explicitly allowed local staff to exercise additional discretion when faced with these dilemmas:
I tell my staff that they should not worry about how much time they spend on users. I tell my own manager that we follow the interface – and use our heads. [. . .] The users want to hear what's best for them and they come here with a truckload of questions. We are stuck – we can't refuse to help them. It's not easy to tell users that ‘it says this and that in our manuals. . .’ (NAV office manager)
Overall, the front-line service workers found it difficult to reject deserving citizens. Hence institutional work served to minimise the discrepancies between what they felt were inadequate services and, by extension, inadequate policy implementation and interpretation by the central authorities. It was undertaken for the most part by front-line staff. Most often they were not sanctioned, but instead supported by local managers.
Discussion
In this article, we have sought to nuance the debate about diminished or extended autonomy of street-level organisations by exploring the institutional work of front-line workers in responding to higher-level policy and administrative changes. In so doing, our central argument is that a missing piece in the discussions of autonomy involves the agency of front-line workers in modifying perceived discrepancies – even within contexts of high managerial and bureaucratic control. Accordingly, we have illustrated the variety of local practices emerging even in organisational contexts characterised by efforts to limit the local level of service and by low acceptance for local variation. Given the central tenet in the policy implementation literature that street-level actions deviate from higher level goals based on a variety of features, we argue that the concept of institutional work is useful because it shows how successful policy implementation implies not only institutional work by top management, but also institutional work by street-level organisations.
As we have illustrated, institutional work may take different forms (see Table 2). At the central level, policy implementation and organisational restructuring involve the combined institutional work of disruption and creation – i.e. delegitimising established practices and instigating new forms of performance. At the local level, institutional creation involves practices that are in line with the higher-level goals. In other words, the perceived discrepancies between the higher-level goals and the needs of citizens are sorted out by developing new skills and practices regarding the guidance of citizens within the formal interface of the central authorities – albeit not always publicly announced or officially approved.
Summary of findings: Institutional work in organizational reform
Institutional modification, in turn, points to practices that seek to protect the status quo, yet without deliberate resistance or undermining of the higher-level goals. In this sense, attempts are made to resolve the discrepancies by reverting to the traditional mode of organisation – in our case to the traditional role of the front-line workers regarding pension services. Although maintenance in our case involves the provision of additional service to the users, it generally opposes higher-level policy implementation and involves efforts to bypass the formal regulations and thus ‘bend the rules’ and reduce the discrepancies. Paradoxically, rather than undermining the reform processes through active resistance or manipulation, such deviation from scripts required by central authorities shows the development of local practices that contribute to the transformation processes.
More specifically, the various responses are associated with a variety of justifications (cf. Lehmann-Nielsen, Reference Lehmann-Nielsen2006). The creation of local institutions by managers at the call-centres is already grounded in institutionalised beliefs and in the aims of the central authorities regarding the type and level of service to be provided. Thus it does not need any significant justification. In contrast, institutional maintenance is illegitimate to some degree; it involves a breach with the formal division of labour and thus needs justification. A central type of justification by the NAV office workers and managers involves references to the needs and expectations of citizens. This is also a justification supported by the citizens themselves (Breit and Salomon, Reference Breit and Salomon2015). The workers at the pensions units justify their actions by referring to the need to reduce their own workload, as well as to the institutionalised standards regarding the quality of the pension expertise.
There is an interesting difference in the degrees of justification needed in the different units. The institutionalised practice and identity of the former National Insurance Administration provided the managers and workers in the pension units and the local NAV offices with ideals of high-level service to citizens that they could mobilise in their constructions of legitimacy. In contrast, most of the call centre workers were newly recruited (e.g. from other call centres in the region) and thus did not possess such institutionalised service values. Instead, they developed practices – sometimes at the cost of service to the citizens – that they could not publicly justify, but which were rooted in silent or tacit understandings of the need to find ways to comply with standards set by management.
Conclusion
Our findings show the existence of considerable discretion employed by front-line workers in the context of managerial control and regulation. However, rather than being relatively fixed, and based on relative autonomy from higher-level authority, discretionary space is actively created and used in various ways by the front-line workers to modify the discrepancies. Thus, our analysis provides organisational and institutional explanations of how street-level bureaucrats make use of their autonomy (Brodkin, Reference Brodkin, Brodkin and Marston2013; Winter, Reference Winter, Peters and Pierre2012).
We believe that the article makes two specific contributions to the literature on policy implementation and street-level bureaucracy. The first contribution is an explication of the types of institutional work involved in organisational reforms and, specifically, how they reveal the ways in which street-level bureaucrats use their discretion to impact public policies through the delivery and organisation of services. The second contribution is the attempt to bridge theories of policy implementation and street-level bureaucracy with that of institutional work. We argue that institutional work provides a useful heuristic for conceptualising not only types of modification of public policies, but also the agency involved in this modification.
We believe that the types of institutional work described here are relevant beyond the context of pension services and that they illustrate important mechanisms in play where top-level attempts to (re)organise front-line services meet the ‘realities’ facing street-level bureaucrats. A central tenet demonstrated in our analysis is that policy implementation of reforms aimed at institutionalizing new forms of practice is dependent not only on local institutional work, in compliance with the objectives and instructions of central authorities, but also on the specific institutional work of street-level organisations in particular. In this sense, the variety of institutional work outlined in this article adds to the literature stressing the need for focusing on the behaviour of street-level bureaucrats in explaining the variations and outcomes of policy implementation (Winter, Reference Winter, Peters and Pierre2012: 273–275)
Furthermore, the article demonstrates how street-level deviation from central expectations need not be negative or dysfunctional, which has been the dominant focus in the street-level bureaucracy theory. We show how deviation can also represent positive and beneficial institutional processes as they ‘grease the wheels’ of the reform initiatives. From this point of view, policy implementation is dependent not only on creative (and disruptive) institutional work, but also on maintaining institutional work. Maintaining institutional work plays a key function in mediating between the past and the present in institutional reform processes and thus between demands from top management and the needs and expectations of citizens.
We argue that front-line service workers will play an important, active role in securing adequate services for citizens, even (or perhaps especially) for the most vulnerable citizens, despite the prominence of managerial control regimes. This optimism is rooted not only in the role of front-line workers operating as an organisational unit balancing between internal demands for consistency and efficiency and external demands for personalised services and protection of the legal rights of citizens (Needham, Reference Needham2011; Toerien et al., Reference Toerien, Sainsbury, Drew and Irvine2013). It is also rooted more broadly in their efforts to uphold trust and legitimacy in the welfare system, despite the introduction of new forms of interaction with citizens.
