Placement Advisors as Innovators. How Professionals Use Enhanced Discretion in Germany’s Public Employment Services

Abstract Employees of the public employment services (PES) are street-level bureaucrats who shape activation policy on the ground. This paper examines how PES staff use enhanced discretion in an innovation project carried out by the German Federal Employment Agency. Applying a bottom-up perspective, we reconstruct PES employees’ logic of action and the dilemmas they face in improving counselling and placement services. According to our findings, placement staff use enhanced discretion to promote more individualised support and an adequate matching of jobseekers and employers. The use of discretion is framed by organisational norms and reward mechanisms and by the current labour market situation. Our analyses are based on qualitative interviews and group discussions with placement staff.


Changes in job placement services as a result of activation policies
Activation policies focus on promoting employability and labour market participation, emphasising the jobseekers' active involvement (Newman, ; Serrano Pascual, ). Although the forms of activation differ from country to country, all the countries exhibit a growing importance of commodification: in other words, an increasing relevance of labour market participation for social security and integration (Dingeldey, a). In this way, labour market risks are individualised more strongly instead of being cushioned by society as a whole. Commodification can be enabling in character or may emphasise the obligation to work (workfare) (Brodkin, ). One element of the workfare-oriented strategy is prioritising timely transitions into employment, including atypical forms of employment, such as agency work or marginal part-time employment.
Atypical forms of employment bear greater social risks. Especially the lowskilled, agency workers and people with marginal employment contracts (so-called "mini-jobs") have little upward mobilitywith regard to both leaving the low-wage sector and improving their employment status in more stable employment forms (Eichhorst et al., : ff.). Empirical impact analyses and debates on activation policy also characterise the discussions in the Journal of Social Policy (e.g. Dengler, ; Hohmeyer and Lietzmann, ; Orton et al., ; Jordan, ; Eriksen, ).
The public employment services (PES), in particular the counselling and placement of jobseekers undertaken there, constitute a core element of delivering social policy. Research focusing on the frontline delivery of activation policies has argued that the way in which placement staff apply standardised rules to individual cases shapes the welfare state in practice (e.g. van Berkel et al., ). Discretion is a key concept in this research tradition, referring to Lipsky's classical study of street-level bureaucracies (Lipsky, ).
There is an ongoing debate surrounding the curtailment of frontline discretion in the course of New Public Management reforms, procedural standardisation and the use of IT systems (Evans and Harris, ; Jessen and Tufte, ). However, many scholars stress that discretion is an inevitable feature of frontline work in a human service organisation. Although standardisation and IT systems have certainly changed the nature of frontline work, new dimensions of discretion have emerged with activation policies, such as making judgements and evaluations of individual clients' behaviour (van Berkel et al., : ).
Drawing on the literature on frontline discretion in activation policies, our paper examines how employment agency staff use an expanded scope of discretionary power within the framework of an innovation project. What are their goals and what dilemmas do they face? Investigating these questions can provide insights into the mechanisms that affect the use of discretion and thus the shape of activation policy in practice. Qualitative data that were gathered to evaluate an innovation project conducted by the Federal Employment Agency (Bundesagentur für Arbeit (BA)) form the empirical basis of the paper. Based on the street-level bureaucracies approach we show that the PES staff use their discretion to promote individuality in their relationships with unemployed people even though this approach requires more effort. Focusing on the perspective of PES employees and reconstructing their logic of action, we find that it is crucial for this practice to be legitimised in the organisation. We conclude that PES staff with greater discretionary powers, enabling organisational norms and a favourable labour market situation are conducive to delivering tailor-made services and placement in stable and durable jobs. This study focuses on the German unemployment insurance scheme. The unemployment insurance constitutes the "upper" tier of Germany's two-tier system of unemployment protection. Benefits are granted on the basis of previous contributions, and a proportional income replacement is intended to secure status protection for a certain period of time. Activation requirements are less stringent than those in the means-tested basic income support scheme (popularly known as "Hartz IV"). Nonetheless, policy reforms since the mid-s have weakened the principle of status protection and stressed the conditionality of benefits (Bothfeld and Rosenthal, ). With the aim of decreasing reservation wages and accelerating re-employment, a narrower definition of "suitable employment" and more demanding regulations concerning occupational mobility, availability and job search requirements have been introduced (ibid: ).

Placement advisors as street-level bureaucrats
As part of the public administration, the BA is responsible for implementing the unemployment insurance scheme and related employment services. In this context, Germany's PES staff deliver services associated with job placement and career guidance in some  employment agencies. The placement advisors at the employment agencies have an important role to play in performing the BA tasks that are defined by law. The PES employees take the provisions of social security legislation that are standardised by law and apply them to individual cases, thereby translating standardised rules into practical action. Due to the framework conditions of their work, placement advisors can be defined as street-level bureaucrats. Street-level bureaucrats (Lipsky,  []) are characterised by possessing considerable discretionary powers in their day-to-day work. By using this discretion in sometimes very different ways, street-level bureaucrats influence the concrete implementation of abstract policy specifications.
In social policy research, Lipsky's approach has proved helpful for examining the delivery of activation policy at the street level of the PES (for an overview see Brodkin and Marston, ; van Berkel et al., ). The basic conditions of the employee action and dilemmas in job placement in an activation-oriented welfare state constitute one main topic of the analysis. Reference is made, for example, to the fact that frontline staff are confronted by multiple role expectations. They have to reconcile a bureaucratic and a professional rationale: in their role as administrative employees they are responsible for enforcing rules, also making decisions, for example, about benefit sanctions. At the same time, in their role as professional advisors, they work together with jobseekers to develop strategies aimed at labour market integration (Sainsbury, : ). In doing this, they combine "carrot and stick" elements of activation policy (Tabin and Perriard, ; van Berkel and van der Aa, ).

    
Organisational demands are also a source of further dilemmas in job placement work. In many countries a trend towards standardisation is being observed, which is reflected in detailed procedural regulations valid within the organisations (Ponnert and Svensson, ). This trend is reinforced by the use of IT procedures to assist job placement. While standardisation can increase transparency and enhance the external legitimacy of the organisation, it can simultaneously mean a curtailment of the employees' professional discretion, and therefore a form of de-professionalisation, as well as creating new uncertainties. This leads to placement staff being required to find a balance between standardised central demands and requirements on the one hand and the call for solutions that are tailored to meet jobseekers' individual needs on the other hand.

Activation-oriented job placement
In Germany's PES, as in other European countries, structures have increasingly emerged since the s which reflect a managerial interpretation of administrative action as is formulated by the concept of New Public Management (Hood, ; Jantz and Jann, ). Considerable changes were implemented in particular during the course of the so-called "Hartz reforms" in  and  (Fleckenstein, ). The job placement services at the employment agencies were reorganised with the aim of modernisation (often IT-assisted) (Dingeldey, b; Eichhorst et al., : ff.). Although these reforms were intended to make services more user-friendly, the new processes were criticised as being too standardised and lacking a service orientation (e.g. Hielscher and Ochs, ; Sondermann et al., ).
Another core element of the reform process was the implementation of management by objectives. In the PES context, management-by-objectives systems are based on labour-market policy targets that are operationalised using various target indicators, such as the number of vacancies filled or the average duration of benefit receipt. In the management system of the BA, performance objectives are set out in annual target agreements drawn up between the BA headquarters, the regional directorates and the employment agencies. In order to boost their performance, the target achievement of agencies in districts with similar labour market conditions is compared.
The monitoring of performance plays an important role in outcome-oriented management by objectives. Every agency and branch office has at its disposal an instrument known as the management information system, which makes it possible to compare actual values with the target values continuously. The introduction of the management-by-objectives system resulted in extensive changes in the work methods at the employment agencies. Studies described the new management system as "omnipresent" (Ochs and iso, : ). The managers' day-to-day business was found to be characterised by dealing with figures (ibid.: Chap. .), and the       placement advisors' daily work routine was also seen to be characterised to a considerable degree by quantifying accounting practices (Sowa and Staples, ).
This increased importance of systems to evaluate performance and outcomes creates certain operational rationales and requirements in placement work (van Berkel and Knies, ). In the context of the US, for example, Evelyn Brodkin () describes PES employees as rationally calculating actors who keep their action in line with formal guidelines and other central criteria of the organisational environment. She argues that the street-level workers orientate their discretion towards the logic of management-by-objectives systems and therefore that performance and outcome monitoring systems in the sense of New Public Management may tempt PES staff to concentrate the resources on jobseekers who are closer to the labour market and easier to place in employment (creaming). Creaming becomes particularly evident in means-tested basic income support schemes (see e.g. Hohmeyer and Lietzmann,  for Germany; Jordan,  for UK).
The findings presented above prove that it is rewarding in both theoretical and empirical terms to investigate the discretion of PES staff when delivering public employment services. While other studies look at the discretion exercised by street-level bureaucrats, particularly in terms of consistent rule settings, this study focuses on a setting which permits enhanced discretion in the course of an innovation project. Due to the fact that placement staff were actively involved in initiating improvements, the project design allows us to observe what changes placement staff undertake when they are granted enhanced discretionary authority. Moreover, we can reconstruct which factors foster or hinder the use of value discretion in particular when implementing innovation ideas. Conclusions can be drawn concerning the targeting of employment services by PES employees and dilemmas associated with the use of discretion in counselling and placement services. The study thus shows how advisors redefine "carrot and stick" elements of activation policy when they change aspects of job placement.

Evaluation of an innovation project
The analysis is based on qualitative data gathered during the evaluation of an innovation project. During the two-year project period ( to ) three German employment agencies were permitted to organise their placement processes differently from the way stipulated in the internal rules of the BA. The project was initiated by the BA headquarters; its development was also strongly promoted by staff representatives. The management presented the project as a reaction to staff representatives' criticism of too detailed procedural rules and a high degree of standardisation in placement services.

    
The three agencies chosen to participate in the innovation project were requested to actively involve their employees in initiating changes in the placement process. The agencies were granted enhanced discretion at the level of the organisation. They were free to conduct their placement work in a different manner to that stipulated in the regulations issued by the BA. Legal provisions regulating aspects such as benefit duration, sanctions and the use of labour market policy instruments, however, had to be respected. Moreover, performance monitoring systems remained in place and performance targets were set as usual during the two-year project period.

Data and methods
In the course of the scientific supervision of the innovation project,  interviews were conducted with skilled personnel and managers at the participating agencies, at the regional directorates responsible for those agencies and at the BA headquarters. The main focus of the survey was the agency level ( interviews). We also held seven group discussions with four to six placement advisors at the three agencies. In addition, we conducted  participant observations of projectrelated workshops (e.g. creative workshops, themed workshops), project meetings and counselling interviews involving advisors and jobseekers.
The interviews and group discussions were semi-structured to allow the respondents to decide what was relevant for them and to permit narrative responses. At the same time, the interview guide ensured that important topics were covered, such as experiences made in placement and counselling services as well as assessments of discretion.
The interviews were conducted at the respondents' places of work and generally lasted between  and  minutes. The group discussions lasted between  and  minutes. The majority of the interviews and group discussions were recorded and fully transcribed with the participants' consent. For the analysis, we first combined inductive and deductive procedures to develop thematic categories which were then taken as the basis for encoding the material using the software MAXQDA (Mayring, ). Categories included, for example, project implementation, job placement and advisory services, rules and regulations, as well as organisational context. Based on this initial inspection of the material, we identified text sequences that provided information about interpretations of discretion and its use. On the basis of a comparative analysis, the relevant text passages were condensed and typical argumentation patterns were then identified.

How placement staff use discretion
In Lipsky's analysis of the Dilemmas of the Individual in Public Services, discretion is used as an encompassing term referring to street-level bureaucrats' decisions and actions "determining the nature, amount, and quality of benefits and       sanctions provided by their agencies" (: ). In a more fine-grained perspective, different elements of discretion can be identified (Taylor and Kelly, : ff.): task, rule and value discretion. PES advisors have relatively broad autonomy in their daily tasks to decide which placement services they provide to whom and when. This discretion arises from the complexity of the tasks and the resulting lack of complete managerial control (task discretion). Despite the existence of legal, fiscal and organisational constraints, placement advisors determine the type and frequency of client support through their interpretation of the regulations (rule discretion). Furthermore, based on their professional knowledge and professional codes of conduct, advisors make choices about the objectives of activation and the practical application of social policy (value discretion).

The scope of discretion during the innovation process
As can be expected due to the above-mentioned forms of discretion, our data reveal discretionary and decision-making authority, which also exists independently of the innovation project. In interview sequences referring to discretion, rule discretion is a prominent theme. PES employees illustrate, for example, how they interpret rules regarding so-called placement proposals. These formal placement proposals comprise job-vacancy details that are sent to jobseekers with the aim of initiating an application. When unemployed individuals receive information about a job vacancy from the employment agency as a formal placement proposal, they are expected to apply for the job. Non-compliance leads to benefit sanctions. However, placement advisors explain that they accept credible reasons for jobseekers not applying for suggested jobs if the jobseekers can prove that they have applied for other jobs more in line with their own personal inclinations instead. In a similar way, PES advisors describe a leeway for interpreting rules when assessing reasons given by clients for missing appointments. One advisor puts this in a nutshell when explaining "we have directives but have to fill them with life" (,   ).
Besides this discretion that exists in general for street-level bureaucrats, however, advisors in the innovation project were given a say in decisions regarding whether organisational directives were suited to their work and should be continued or suspended. To this end, they submitted proposals via informal notes or IT systems. Especially for major changes or more complex topics, concepts were developed in workshops. There, the PES advisors discussed what would make their work easier as well as how employment services could be more closely aligned with the interests of their addressees. For this purpose employers and unemployed people were invited to the workshops or were asked to take part in surveys accompanying the change processes. Some of the workshops were observed by the researchers. There, professional demands and dilemmas regarding workload, the compatibility of the implemented changes with the software currently in use or the performance system were addressed. During the      conceptualisation of the changes, the PES staff made choices about the design of employment services and the practical application of social policy. Although the decisions concerning the changes to be introduced were made by the management, the placement advisors set new professional standards. The major benefit of the innovation project was the extended value discretion.
The various changes implemented in the innovation project can be distinguished by the value discretion they entail and whether they are aimed at jobseekers or employers (see table ). In Germany, job placement work is divided into jobseeker-oriented and employer-oriented tasks, which are performed by different advisors. Employer-oriented placement staff collect information about job vacancies which are reported to them by employers or are published publicly. Jobseeker-oriented placement advisors are responsible for advising jobseekers and for finding suitable vacancies for them. During the innovation project, advisors reduced documentation tasks and adopted direct, advisory and less bureaucratic forms of communication and counselling. These are presented below.

Using discretion for innovations in jobseeker-oriented services
The PES staff used their enhanced discretion to adopt less bureaucratic forms of communication with their clients. They simplified forms, information and processes. One example that many interview partners mentioned in relation to de-bureaucratisation is a greater individual leeway in deciding about the durations and intervals of job-counselling interviews. One PES advisor explains: Now it's possible to make the interview shorter or longer for various reasons or to delay the interview or to conduct it by phone, and this is welcomed because it is personalised. Unlike in The quotation is an example of enhanced value discretion in placement work. Before the project, an internal rule of the BA set out that counselling interviews had to be conducted in person, with a defined length and at defined intervals. This rule was suspended by the employment agencies participating in the innovation project. Furthermore, a process indicator that had previously been recorded and monitored by managers with IT assistancewith the aim of ensuring that advisors spent at least  percent of their working time on counselling interviewswas discontinued. For the PES employee cited above, it is not only the organisational rule that is important ("it's now possible") but also the fact that his or her action is esteemed ("this is welcomed"). This quote, together with the fact that not all advisors deviated from the previous specifications on interview duration and intervals, also reveals a dilemma between the conflicting aims of time saving and the representation of the work in the performance system. On the one hand it can free up time if appointments do not need to be documented and appointments are made on a case-by-case basis. On the other hand, diverging from ritualised processes entails risks because this could diminish the work performance. Some advisors continued to follow the previous guidelines, such as entering counselling interviews in the calendar in order to document them "for oneself and of course for the manager" (, )as one PES employee reports. This suggests that the use of discretion is also linked with the credible legitimacy of new procedures and the actual loss of importance of previous rules. In this way, decisions made at the value level also affect the rule and task levels. In other words, advisors developed changes in the job placement services by using their value discretion during the innovation projectsuch as abolishing the above-mentioned rules for scheduling counselling interviews. Whether the advisors really invite jobseekers for interviews on a more individual basis or whether they maintain the old rules within the framework of their rule discretion depends on whether the innovations in the performance system are rewarded. They also organise their daily tasks in their work with jobseekers in accordance with this.
Furthermore, the PES staff adopted direct and less formal forms of communication with their clients. For example, so-called job cafés were developed in order to enable meetings on a more equal footing. There, advisors meet a few jobseekers in a café to give them information on the job market situation. Advisors may also decide to invite others: for example, external experts (to present findings on job application strategies), or employers (who provide information about their vacancies). In addition, advisors present current job vacancies at the meetings and ask the jobseekers whether they want to apply for them.

    
Another important change in jobseeker-oriented services was the partial discontinuation of so-called personal action plans. Personal action plans are written agreements between the employment agency and jobseekers that record jobseekers' obligations and the support services provided by the employment agencies. If jobseekers fail to demonstrate that they have made the efforts set out in the personal action plan, this will result in benefit sanctions.
Before the innovation project, personal action plans were concluded with each jobseeker in the first counselling interview. All employment agencies involved in the project decided to grant placement advisors enhanced discretion regarding whether or not to draw up personal action plans with jobseekers in the first months. The following extract from a discussion between advisors shows their opinions regarding this change: Advisor : Well I find that it's more personalised. Without saying "I have to do this in a certain way" you can simply try to do the best you can for the person in front of you. And in my opinion, that's the whole purpose of an employment agency. In other words the service idea is simply more important. And what they call employer brandingin other words, our reputationhas automatically improved too. ( : : : ) Advisor : ( : : : ) I also think we can treat the people in front of us more personally. We don't have the rules we used to have, if this is the situation then you have to do this, instead we can really look to see what suits the individual person. ( : : : ) I've been working here for ten years now and it was never like this before, and I really find that it's an extremely positive development for us. (, -) The advisors associate not drawing up a personal action plan with greater open-mindedness and service orientation. It is important to "do the best you can", to "treat the people more personally" and to follow the principle of what the client needs. The advisors' statements do not simply refer to the service for the unemployed, but already convey a new type of "employer branding" in order to define their work as personalised and themselves as "nice people". According to the activation-policy reforms, job placement was to be modernised and service delivery should take into account the perspective of the client, often referred to as the "customer" (Dingeldey, b; Eichhorst et al., ), However, terms such as "customers" and "service" are discussed quite critically, as they disguise power processes and do not depict existing asymmetries (Hasenfeld, ; Ludwig-Mayerhofer et al., ). Although there is a fundamental power asymmetry between the unemployed and the placement advisors, the PES staff interviewed here see their enhanced discretion as following a service logic rather than an administrative logic. Earlier studies already showed that the advisors' communication style is influenced by their professional self-perception and role and that service-oriented advisors exist (Boockmann et al., ; Hielscher and Ochs, ). In the analysed innovation project, not only did the individual       advisors dispense with personal action plans, but this was also actively promoted by the agencies.
Job placement without sanctionable targets in personal action plans reduces disciplinary elements for clients. The dilemma of the service orientation lies between equal treatment and creaming. The possibilities for reducing pressure can be associated with the advisor's assessment of the particular client. It emerges from other statements that discretion is used above all for jobseekers who are regarded as motivated and active. In contrast, it is important to "have things more firmly under control" (, ) when dealing with people whose motivation is considered to be low. Discretion in favour of the jobseeker then tends to be used less, and during the counselling interviews the advisors refer to the externally defined limits of discretion: for instance, "Look, this is what the legislator says. This is where my discretion ends" (, ). Consequently, regulations are a desired method for communicating more strictly and more authoritatively with certain jobseekers and for legitimising boundaries. Presumed or actual restrictions in discretionary leeway then serve the purpose of control, whereas exploiting discretion to the full stands for a service orientation.

Using discretion for innovations in employer-oriented services
Concerning the employers, one change was to involve them in the placement process. Instead of employers simply providing information about vacancies on paper, personal meetings with employers were increasingly usedespecially in the case of small and medium-sized enterprises. Moreover, brief job interviews between jobseekers and employers were initiated: either with jobseekers being accompanied to employers or the interviews taking place in the context of information events at the agencies. Employers were then asked to provide information about the vacancies and obtained direct feedback from jobseekers. One aspect of this involvement was to create transparency regarding working conditions and applicant profiles prior to the potential job interview. The analysis of the interviews reveals that placement advisors also face a dilemma here: the enhanced transparency is opposed to less effort without personal meetings at the employer's premises and without the event organisation for the advisors.
Rule and task discretion were also used to link the knowledge of the advisors responsible for jobseekers with that of those responsible for employers. The division of tasks for jobseeker-and employer-oriented advisors creates specific knowledge on both sides. Their knowledge exchange occurs mainly via an IT-based matching procedure or formalised exchange formats and individual contact between colleagues in specific cases. Placement advisors aimed to create a more tailor-made service for employers by means of new forms of knowledge transfer between jobseeker-and employer-oriented advisors. The agencies tested forms of team cooperation or made advisors responsible for both sides of the labour market to improve information about employers, vacancies and jobseekers. Further      training and an intensive exchange were necessary in order to link the two work areas that were previously divided and monitored separately in the performance system. The knowledge exchange was intended to compensate for limitations in the electronic matching procedure for less standardised occupational profiles. This knowledge about jobseekers' and employers' needs was not only used to supply work-ready jobseekers. One agency implemented the new guideline to use staff resources in particular to find employment for people who were further from the labour market or were not entirely suitable for a particular vacancy. Those jobseekers were personally recommended to employers by the advisors. The advisors therefore faced the dilemma of balancing the actual aim of better service quality with that of increasing the number of placements.

Using discretion to promote stable employment through jobseeker-and employer-oriented services
Throughout our data on employer-and jobseeker-oriented services, a higher weighting of stable employment becomes evident. One PES advisor summarises this tendency as follows: Well, the client is the focus of our attention, and our meta goal, we want to get people into employment and to bring employers and jobseekers together for sustainable integration ( : : : ) These are the goals we pursue and we simply have more possibilities via [the innovation project] (, ).
In the BA logic, "sustainable" integration, or "stable employment" is an employment relationship lasting at least six months. This is the operationalisation used in the BA performance monitoring system. Using this aggregated management accounting indicator, a statistical analysis showed that the agencies participating in the innovation project achieved equal or better values in promoting stable employment than comparable synthetic control agencies (Freier et al., : f.)  . According to our analysis, from the perspective of PES employees, the term "placement in stable employment" refers to the aim of achieving a labour market integration that is as stable as possible (minimum six months) by providing a consultation service that focuses intensively on matching applicants and vacancies. In the innovation project, the goal of placement in stable employment is also the basis for many of the more service-oriented offers and the knowledge exchange between the two sides of the labour market described above. Advisors obtain more cooperation from jobseekers if the personal action plan is dropped and employers are involved in the placement process. The statements reported earlier about categorising jobseekers according to their willingness (creaming), however, also indicate that prioritising placement in stable jobs could depend on the advisor's estimation of the jobseekers.       However, placement in stable employment entails a dilemma associated with balancing increased cooperation and reduced standardised efforts for services oriented towards short-term employment. PES staff mention the problem that aiming for lasting employment is generally more time-consuming, as the formal placement proposals are of a higher quality, being more precise, but fewer of them are made. At the same time, the formal placement proposals serve to quantify placement activities in the performance monitoring system. A successfully filled vacancy is only registered if a formal placement proposal was issued beforehand. According to PES employees, concentrating on high-quality formal placement proposals sometimes led to poorer results in their own target achievement compared with other teams. When the focus is on performance indicators, there is a risk that the individual jobseeker's situational needs may be neglected in the effort to meet targets (Bender and Brandl, ; Brodkin, ; Sowa and Staples, ). Moreover, re-employment that cannot be depicted in the monitoring system may not receive any esteem. Our material suggests that the extent to which street-level bureaucrats achieve placement in stable employment is decisively influenced by certain indicators in performance monitoring. The researchers observed a turning point in the innovation project: when the integration rate became too low, managers urged advisors to promote timely integrations again by increasing the number of formal placement proposals. The previously shared norm of a high fit and stable placement was thus rejected. The following statement by a PES advisor illustrates such dilemmas: "Because when I always have to meet my targets, then I have to set my priorities in a certain way for certain demands ( : : : ) and that gives me less leeway, less time, less effect again and more rigid specifications in my mind" (, ).
Another example of dilemmas with regard to placement in stable employment is that high demand in firms offering temporary employment and seasonal work can lead to high labour market integration rates if people who were placed in short-term temporary jobs reappear in the jobseeker registers and are placed in work (and counted) again. A study of PES advisors' attitudes reaches the conclusion that placement advisors who focus on finding stable employment or on stabilising the client's personal situation exhibit poorer target achievements in terms of labour market integration than those whose main aim is a rapid transition into employment (Hofmann et al., : ).
Despite these dilemmas, our respondents orientate their placement decisions towards the quality of a placement in long-term employment, even if a short-term job may have a stronger impact on the highly relevant performance indicator of the integration rate. In addition, the statements made by our respondents suggest that the positive labour market trend and the resulting demand for skilled labour facilitated increased cooperation with employers and placements in more long-term employment. The project period ( to      ) was characterised by a favourable labour market with falling unemployment and dynamic employment growth. Firms were increasingly willing to make compromises with regard to the qualifications, work experience and additional skills required in the low-wage sector and to conclude employment contracts for longer periods (Gürtzgen et al., ). Accordingly, the demand for skilled labour appears to have levelled out the dilemma of balancing the conflicting aims of achieving measurable outcomes in terms of timely labour market integration and finding lasting employment for groups that are in demand.
With their focus on long-term labour market integration via more personalised advice and matching jobseekers with vacancies, the PES advisors in the study not only promote stable employment, but to a certain extent they also balance out the individualisation of the responsibility for unemployment that is associated with activation policy, the pressure to take up work and the demands on benefit claimants to make concessions (Dingeldey, a; Kettner and Rebien, ; Lessenich, ). The placement advisors in the study choose an employment-oriented balance, in contrast to those who promote pressure reduction for vulnerable people in social activation programmes (Freier, ).

Conditions for innovations to benefit jobseekers, employers and advisors
Against the background of the PES advisors' interests impacting on job placement, of what advisors are expected to deliver according to the legislation, of the organisation, the jobseekers and structural basic conditions, the PES staff fill directives "with life" by applying abstract social-policy and organisational rules to individual cases. They thus act as street-level bureaucrats (Lipsky, ).
Based on an innovation project that permitted enhanced discretion at the organisational and street level, we showed that frontline workers used their discretion to achieve personalised services and placements in lasting employment. In particular, they use value discretion to define changes within the job placement services. Our respondents did not see themselves simply as employees of a public authority, but also identified with the concept of tailoring their service to their clients' needs and reducing pressure, as is seen in the example of dropping personal action plans.
Due to the exploratory nature of our study, the results cannot be generalised. However, the findings are in line with other exploratory studies suggesting that institutionalised discretion fosters a focus on individualised, tailor-made patterns of service provision (Rice, ; van Berkel et al., ). These patterns may include a broader portfolio of service provision as well as more holistic forms of employment services for vulnerable clients (Rice, ). In the job café described earlier, after a presentation, it can be easier for the jobseekers to discuss       current job vacancies with the advisor and decide which are most suitable for an application. The individualised service, however, also focuses on control and does not correspond to an ethical and non-discriminatory conception of social policy (regarding conceptions of the individual subject, see Yeatman et al., ).
In contrast to existing studies, we examined not only how placement advisors use discretion, but also what factors influence their use of discretion. By developing innovations, advisors have been able to expand their value discretion. However, whether this value discretion was implemented in their everyday tasks depended on several aspects: illustrating the turning point with regard to the emphasis on stable employment, we show that the use of discretion depends on targeting mechanisms within the organisation. Taking individual counselling time as an example, we argue that value discretion is more likely to be used if the action is esteemed, especially in the eyes of the manager, and corresponds with organisational norms. In light of the fact that, in the context of New Public Management, established standardisations such as contacts to jobseekers are regarded as a reduction of discretion (Jantz and Jann, : f.), advisors reclaimed their discretion in the innovation project.
The thesis presented is that the interplay between PES advisors having greater discretionary powers, changing organisational norms and a tight labour market is beneficial to individualised services supporting placement in lasting and stable employment. We showed that changes were accepted more easily when they were recognised and valued within the organisation. The performance management system proved to be an important mechanism for creating and promoting esteem in the organisation. Endowed with greater discretion, the employees interpret "service orientation" as personalised and long-term job placement and orientate their behaviour towards this organisational norm as long as the performance indicator "integration rate" guarantees a good outcome in the monitoring system. The favourable labour market situation reduces the dilemma regarding the conflicting aims of service orientation and the implementation of activation policy. Dilemmas are also seen, for example, in the conflicting aims of timely labour market integration and placement in long-term employment. The presented qualitative study was able to reveal mechanisms, but is limited in its ability to determine whether PES staff promote stable and lasting employment beyond the innovation project. Future research could take this as a starting point to analyse whether in a positive economic climate PES advisors systematically offer stable and lasting employment.
What is remarkable in the study at hand is that advisors opt for placement in long-term jobs and justify this professionally even though the outcomes are less visible in the monitoring system. This leads to the conclusion that PES staff are able to interpret organisational rules in the sense of stable jobs in organisations where placement in lasting employment is legitimised. Nonetheless, a conflict remains between professional self-perception and management via      performance indicators, and steps aimed at re-employment that are only measurable to a limited extent (Bender and Brandl, ; Brodkin, ; Sondermann et al., ; Sowa and Staples, ).
Beyond the German case, our results suggest that the use of discretion does not depend solely on advisors' personality traits (Boockmann et al., ; Lagerström, ) or social-structural characteristics of the unemployed (Ludwig-Mayerhofer et al., ) or their motivation (Hasenfeld, : ). Structures of New Public Management not only shape advisors' work on a formal level via bureaucratic procedures, standardisation and managerial control (Diefenbach, : f.; Jessen and Tufte, ), but also influence professional action on an emotional level via esteem. This brings the employment service closer to Eriksen's view that street-level interactions should be based on respectful behaviour and professional decisions instead of rules (Eriksen, ). From this, we deduce that it is not sufficient to expand advisors' discretion in order to promote placement in stable employment. Discretion in itself is neither good nor bad, but the interplay between discretion, organisational norms and the labour market situation can promote individualised services and lasting employment.