The frontline delivery of activation: workers’ preferences and their antecedents

ABSTRACT

This article intends to contribute to the increasing body of academic literature on the frontline delivery of activation policies. In line with the broader literature on frontline work, this literature focuses attention on frontline delivery practices and preferences and how these can be explained. Activation work is an interesting case in this context, because activation workers’ jobs designs vary considerably and workers have a diversity of educational backgrounds (including but not limited to social work) in the many countries that have implemented activation policies. This article looks at workers’ preferences rather than actual delivery practices. It analyses how job design and educational background are related to workers’ preferences concerning servicing clients, managing activation workers and spending time on various activation-related activities. Although the findings of the research that are presented in the article show that educational background and job design play only a modest role as antecedents of workers’ preferences, future research into this issue remains interesting. For it may help to increase our insights into the optimal match between what activation aims to achieve and how the provision of activation is organized on the one hand, and workers’ skills/qualifications and job design on the other hand.

SAMENVATTING

Dit artikel draagt bij aan het toenemende academische onderzoek naar de wijze waarop activeringsbeleid in praktijk wordt uitgevoerd. Dit onderzoek richt de aandacht op handelingspraktijken en voorkeuren van uitvoerders (in Nederland ‘klantmanagers’ genoemd) en op de vraag hoe die kunnen worden verklaard. Activering is een interessante casus omdat taken en verantwoordelijkheden van uitvoerders heel divers zijn, en omdat uitvoerders in de landen waar activeringsbeleid is ingevoerd zeer diverse opleidingsachtergronden hebben. Onder de uitvoerders bevinden zich sociaal werkers, maar zij domineren dit werkveld zeker niet. Dit artikel richt zich op de voorkeuren (en niet zozeer de praktijken) van uitvoerders. Het analyseert hoe de taakomschrijving en de opleidingachtergrond van uitvoerders samenhangen met hun voorkeuren ten aanzien van de dienstverlening aan klanten, de wijze waarop uitvoerders gemanaged willen worden en de tijd die ze wensen te besteden aan diverse activiteiten in het kader van activeringsdienstverlening. Hoewel we tot de conclusie komen dat taakomschrijving en opleidingsachtergrond slechts een bescheiden rol spelen als antecedenten van de voorkeuren van uitvoerders, is dit type onderzoek wel degelijk van belang. Want het kan bijdragen aan ons inzicht in de optimale match tussen wat activering beoogt en hoe activeringsdienstverlening is georganiseerd aan de ene kant, en de kennis/vaardigheden en taakomschrijving van uitvoerders aan de andere kant.

KEYWORDS:

• Social work
• social policy
• activation
• frontline work
• discretion

SLEUTELWOORDEN:

• sociaal werk
• sociaal beleid
• activering
• klantmanagement
• discretionaire ruimte

Introduction

With the increasing emphasis in social policies worldwide on the activation of unemployed people, research into the frontline delivery of these policies has gradually received more attention (Brodkin & Marston, 2013; Tabin & Perriard, 2016; Van Berkel & Van der Aa, 2012). The last decade has witnessed an increasing number of academic publications devoted to this issue and, not surprisingly, these publications echo many of the debates scholars of policy implementation and frontline agency have been engaged in for long (cf. Hupe, Hill, & Buffat, 2015). First of all, there is the debate about how policies and services are being delivered and clients being served at the frontlines, and about how frontline workers in activation (from now on referred to as activation workers) use their discretion in doing so. Various studies showed that workers may use discretion in rather diverse ways, resulting in diversity in treating and servicing clients. This points to the risk that frontline delivery becomes a rather individualized and potentially even arbitrary process (Eikenaar, De Rijk, & Meershoek, 2015; Nothdurfter, 2016). This issue is directly related to the second debate in studies of the frontline delivery of activation: the management of frontline work and the room for professionalism. In line with the broader ‘professionals under pressure’ debate (Noordegraaf & Stijn, 2013), several authors conclude that the activation project, with its emphasis on labour-market participation, conditionality and sanctions, may challenge professional service provision (Kjørstad, 2005; McDonald & Chenoweth, 2009; Røysum, 2013). Some authors argue in favour of a (re-)professionalization of the frontline delivery of activation in order to promote transparency in the use of discretion (Eikenaar et al., 2015), strengthen the professional treatment model in delivering activation (Sainsbury, 2008) and encourage the provision of personalized services (Van Berkel, Van der Aa, & Van Gestel, 2010). Finally, and related to the former, there is the debate about how the roles of frontline workers should be perceived. Maynard-Moody and Musheno (2000) distinguish two narratives in studies of frontline workers: the worker as state agent (‘policy implementer’) and the worker as citizen agent. These narratives point at activation workers’ dilemmas in deciding ‘who to serve’ in activation (the government, the organization and/or the client) (Thornton & Marston, 2009) and the role conflicts (Tummers, Bekkers, Vink, & Musheno, 2015) they are confronted with in making these decisions.

This article aims to contribute to our knowledge of the frontline delivery of activation. More specifically, it aims to increase our insights into two types of antecedents of frontline delivery: workers’ educational background and workers’ job design. As we elaborate in more detail below, studying the role of educational background in the frontline delivery of activation is interesting given the large variety of educational backgrounds of workers involved in activating unemployed people. This issue has received some attention in the literature (see below), but mostly in the context of studies focusing on comparing workers with an educational background in social work with those without such background, or on comparing workers with various levels of social work training. This article will compare workers with various types of educational background, rather than focusing on social work only.

When analysing the role of educational background in frontline service delivery, we take the design of workers’ jobs (Hill, 2006) into account as well. Activation workers have a variety of job descriptions, which has implications for the tasks and client groups they are responsible for and the authority given to them (Jewell & Glaser, 2006). As we cannot rule out that workers with specific educational backgrounds are more likely to be found in some types of jobs rather than others (for example, social workers may be more likely to serve the most hard to employ groups and less likely to be servicing clients close to the labour market), our study combines the analyses of educational background and job design.

Whereas most available frontline activation studies focus on antecedents of workers’ practices, we focus on how workers would like their work to be, that is, on workers’ preferences, and the antecedents of these preferences. A focus on workers’ preferences rather than practices may provide a more pronounced insight into the role of educational background. For example, Scott (1997) looked at the role of organizational, worker and client characteristics in frontline practices and found that worker characteristics were least important, whereas organizational characteristics were most important. By focusing on workers’ preferences, we hope to reduce this dominant role of organizational characteristics while gaining a better picture of whether or not educational background matters. As will be elaborated below, the preferences involved in our study reflect the core issues in frontline delivery studies: the servicing and treatment of clients, the management of frontline work, and the ‘who to serve’ dilemma workers are confronted with.

These considerations result in the following research question that will be addressed in this article:

Taking job design into account, how is workers’ educational background related to their preferences regarding characteristics of their work?

The article is structured as follows. The next section looks at the diversity in the job design and educational background of activation workers. In the third section, we elaborate on why we may expect that activation workers’ job design and educational background are related to workers’ preferences. The fourth section introduces our research and the methods that were used. Then, the results of the study are presented and the concluding section discusses limitations of our research and reflects on its main findings.

Job design and educational background of activation workers

The literature review discussed in this and the next section is based on a literature review on frontline activation work we prepared for an edited volume on activation work in Europe (Van Berkel, Caswell, Kupka, & Larsen, in press). For this review, a literature search was carried out using ‘frontline work’, ‘activation’, ‘welfare-to-work’ and ‘social work’ as keywords. The literature search focused on articles published in English-language academic journals, specifically social policy, social work, public administration and public management journals. For this article, we focused on those publications resulting from this literature search that (1) provide insight into the job design and educational background of workers involved in delivering activation policies; and (2) provide direct or indirect evidence of the impact of job design or educational background on activation workers’ agency and delivery of activation policies.

The academic literature on activation work reveals that the job design of activation workers and their educational background are highly diverse. Workers’ job design varies along various dimensions. The first dimension of job design we look at concerns whether activation work is integrated with benefit administration and sanctioning (the ‘integrated model’) or whether activation work is separated from benefit administration and sanctioning (the ‘traditional model’) (Hill, 2006).¹ The academic literature discusses various pros and cons of these models. For example, the traditional model prevents a situation in which activation is neglected, among others because workers tend to prioritize benefit issues in the integrated model, specifically in contexts where their caseloads are high (Jewell & Glaser, 2006). The traditional model also separates the ‘good cop’ and ‘bad cop’ roles (service provision versus sanctioning), which may help workers to build rapport with their clients (Jewell & Glaser, 2006). Nevertheless, a study by Scrivener and Walter (2001) showed that the integrated model is more effective.

The second dimension of job design discussed here concerns the client groups for which workers provide activation services. Client groups of activation are very heterogeneous, which is often framed in terms of clients’ distance from the labour market. In many countries, this had led to a categorization of clients in terms of their labour-market distance which may also be reflected in the job design of workers: activation workers may be responsible for specific client groups or may service mixed groups of clients.

No systematic data are available comparing the type or level of professional training attained by activation workers in various countries. Nevertheless, the literature provides a picture according to which activation work is done by workers with very diverse educational backgrounds. The role of social workers receives particular attention in the literature, although it is not unusual to denote activation workers as ‘social workers’ in cases where these workers do not have an educational background in social work. Social work is sometimes used as a branch or as the umbrella of social professions rather than as a specific profession (‘social work’) (Van Ewijk, 2009). Despite this caveat, we think that we can conclude that the role of the social work profession in activation work differs considerably in countries: social workers have a stronger role in the delivery of activation in the Nordic countries than in countries such as the UK, Australia or the Netherlands. Traditionally, social workers used to be especially involved in providing social services to social assistance recipients in local welfare agencies. But the role of social work is not static. A stronger emphasis in social policies on activation and conditionality may result in a situation where trained social workers are considered a ‘liability’ in implementing these policy reforms. In Denmark, for example, social workers have been accused of frustrating the implementation of stricter activation policies as a consequence of which the proportion of social workers among activation workers seems to have decreased (cf. Larsen, 2013). Furthermore, reorganizations and mergers of public agencies responsible for benefit administration and activation services may result in a reshuffling of workers’ tasks and a redistribution of workers with specific professional backgrounds across these tasks (e.g. Røysum, 2013). Finally, the outsourcing and marketization of activation services may also have an impact on the role of social work in activation, for example, because contracted providers may recruit workers with other types of educational backgrounds. For the Australian case, McDonald and Marston (2008) argue that case managers working for Job Network, the Australian network of non-profit activation providers, are recruited from various backgrounds and that social workers only form a rather small group among them.

Job design, educational background and workers’ agency

Apart from Scrivener and Walter’s study (2001) that showed that job design is related to the outcomes of activation, we are not aware of any studies that investigated the way in which activation workers’ practices (and the outcomes they realize) are related to job design. Nevertheless, some studies provide indirect evidence that job design may be important. In terms of the models distinguished by Hill (2006), activation workers in the traditional model can focus on developing people-changing or professional technologies in servicing their clients, whereas workers in the integrated model need to combine people-changing technologies with more bureaucratic and rule-guided people-processing technologies. As Hasenfeld (2010) has argued, realizing this combination is a challenge for organizations, and we expect that it will provide a challenge for workers as well. We already pointed at possible tensions resulting from the combination of the ‘good cop’ and ‘bad cop’ roles in the integrated model. It could be argued that reconciling both roles is part of the ‘professional repertoire’ of activation workers, but at the same time, as has been argued by Sainsbury (2008), acting as a bureaucrat might go at the expense of acting as a professional. Another possible though not necessarily unresolvable dilemma confronting workers in the integrated model concerns that between the bureaucratic equal treatment ethos and the professional personalized services ethos.² For example, a bureaucratic approach of sanctions could result in a strongly rule-oriented way of enforcing sanctions whereas, from a professional personalized service perspective, a pedagogic way of dealing with sanctions might be more likely (cf. Marston, Larsen, & McDonald, 2005). A similar argument could be made concerning the way in which workers make use of individual action plans, a frequently used instrument in activation in European countries. A bureaucratic approach of these plans could focus on rights and obligations of clients, whereas a professional personalized service approach will more likely focus on the nature of interventions and services. Finally, returning to Hasenfeld’s concerns about the organizational challenge to combine professionalism and bureaucracy, it could be argued that activation workers in the ‘traditional’ model will be more averse towards bureaucratic and hierarchical styles of management than activation workers in the ‘integrated’ model: as they do not combine activation and benefit administration, they are more likely to perceive their job as requiring professional interventions rather than accurate rule following.

As far as workers’ client groups are concerned, we can again refer to indirect evidence that client groups matter in terms of workers’ agency and preferences. Compared to clients less remote from the labour market, clients most remote from the labour market are confronted with complex problems which often require diverse services and support, making it more important for workers to be able to provide personalized services (cf. Fletcher, 2011; Toerien, Sainsbury, Drew, & Irvine, 2013). Workers working for these clients may also need more time for their clients to provide the support clients need. In order to be able to do so and act in a responsive way, it could be argued that they prefer more discretionary room and are more averse against regulations that limit their abilities to provide personalized services.

Several studies pointed in a more general way at the potential impact of professional training on workers’ agency, for example, when it comes to providing personalized services (Toerien et al., 2013) or the treatment of vulnerable clients (Fletcher, 2011). A small number of studies looked in more detail at how activation workers’ educational background is related to frontline practices or to workers’ attitudes and perceptions that are expected to be related to these practices. Studies investigating the role of the nature of professional training usually compare workers with an educational background in social work with other groups of workers and find, among others, that social workers have more positive attitudes towards the unemployed (McDonald & Marston, 2008) and the deservingness of welfare recipients (Kallio & Kouvo, 2015), are more sensitive to clients’ needs and more able to provide personalized services (Jewell, 2007). Blomberg, Kroll, Kallio, and Erola (2013) compared groups of workers with various levels of social work training and arrived at similar conclusions: workers with lower levels of education are more strongly inclined to adopt individual rather than social blame explanations of causes of poverty. Some studies looked at how social workers and other groups of workers vary in their perceptions of management. For example, already in 1965, Scott published an article that showed that social workers more strongly than other groups of workers prefer professional rather than bureaucratic forms of supervision (cf. Liljegren, 2012). Jessen and Tufte (2014) write that social workers report stronger experiences of reduced discretion than other groups, which may reflect that policy and organizational reforms have a stronger impact on social workers or a stronger sensitivity of social workers to reductions of discretion.

Summarizing, there is some direct and indirect evidence that warrants the expectation that job design and educational background matter in terms of how workers (prefer to) activate their clients, how they want to be managed and how they cope with the ‘who to serve’ issue. We will now turn towards the design and results of our study.

Research and methods

Our study took place in the Netherlands, which is an interesting country for studying our research question, given the large variety of educational backgrounds and job design of activation workers in this country. The study focused on two groups of workers involved in activating social assistance recipients. The first group is workers in local welfare agencies: the municipal public agencies responsible for social assistance administration and activating social assistance recipients. The second group of workers works in so-called sheltered employment agencies. Traditionally, these agencies provide jobs for people with severe intellectual, mental and physical handicaps. Because of their experience in job-placements for people remote from the labour market, they are sometimes contracted by local welfare agencies to provide activation support for groups of social assistance recipients. As municipalities are free to contract these agencies, the role of sheltered employment agencies differs between municipalities in terms of whether or not they are contracted, and in terms of the size of groups of social assistance recipients they serve.

In the Netherlands, the activation of social assistance recipients is relatively highly decentralized: municipalities and their local welfare agencies have considerable room for policy-making and implementation. Apart from the use of contracted agencies, it also involves decisions concerning the nature of services and, relevant in the context of this article, decisions concerning workers’ job design and qualifications considered necessary when hiring frontline workers. Thus, considerable differences can be observed across Dutch local welfare agencies. As far as sheltered employment agencies are concerned, it should be mentioned that they have no role in benefit administration and imposing sanctions even though they are expected to inform local welfare agencies about sanction-worthy behaviour of clients.

In our study, workers working in 14 local welfare agencies participated. These agencies were not randomly selected from all Dutch local welfare agencies, so we cannot assume that they make a representative sample. When these agencies contracted sheltered employment agencies, the latter were also asked to participate in the research, resulting in the participation of six sheltered employment agencies. In both types of agencies, all frontline workers involved in activating social assistance recipients remote or very remote from the labour market (the core target groups of Dutch activation) were invited to participate in an online survey. The survey was developed on the basis of a literature review and a series of interviews with 31 frontline workers and 7 managers of 3 local welfare agencies and 3 sheltered employment agencies. One hundred ninety-six frontline workers participated in the survey: 163 local welfare agency workers and 33 sheltered employment agency workers. The overall response rate was 52% and hardly differed between both types of agencies. Data were collected in the second half of 2013.³

Job design, educational background

The two groups of antecedents were operationalized as follows. For job design, we looked at the client groups workers work for (in terms of labour-market distance) and at their tasks, distinguishing between the integrated model and the traditional model. Educational background was operationalized in two ways as well: the type of educational background of workers, distinguishing social work, social administration, Personnel & Labour (Human Resource Management)⁴ and other professional training; and the level of education distinguishing university education, higher professional education and lower levels of education.

Workers’ preferences

For operationalizing workers’ preferences, three categories of preferences were used. The first category refers to preferences concerning servicing and treating clients; the second refers to preferences concerning managing activation workers; and the third category refers to how workers prefer to distribute their working time over various types of activities. The first two categories were operationalized in terms of statements, phrased in the form of ‘Ideally, … .’ (For example: ‘Ideally, implementing rules correctly is more important than achieving results’). On a 5-point Likert scale, respondents could indicate to what degree they agreed with each statement. For preferences concerning the distribution of working time, respondents were asked what proportion of their working time they ideally want to spend on a given work activity.

Preferences in the first category, servicing and treating clients, contained statements about monitoring clients, the content of individual action plans, decisions on sanctioning, equal treatment versus personalized service provision, evidence-based decision-making, rule-oriented versus result-oriented decision-making and objectives of work. Preferences in the second category, managing frontline work, contained statements about autonomy of workers, the regulation of services, the prioritization of client contacts, the consultation of supervisors/peers in difficult situations and individual performance management. In the third category, the distribution of working time, the following work activities were distinguished: client contact, contact with colleagues/supervisors, external contacts (e.g. employers, external service providers) and administrative tasks. The categories of preferences reflect the core issues in frontline work studies discussed in the introduction. The ‘who to serve’ issue is reflected in how workers prefer to distribute their working time over various activities, and in several statements about priorities (for example, client contacts versus administrative tasks; correct rule implementation versus realizing results with clients).

Current work characteristics

We also looked at current work characteristics in order to compare current and preferred work characteristics. Current work characteristics were formulated using a parallel structure to workers’ preferences (for example: ‘In my current work, implementing rules correctly is more important than achieving results’), and were also measured using a 5-point Likert scale. In addition, workers were asked what proportion of their working time they currently spend on specific work activities.

Control variables

In the data analysis, the following control variables were used: type of agency workers work for (local welfare or sheltered employment agency), workers’ age and number of years of experience in activation work.

Analysis

For the analysis of antecedents of workers’ preferences, linear regression was used. Several dummy variables were created to analyse the antecedents and to compare groups. For example, dummy variables were created to compare workers working for people remote from the labour market with other workers, and social workers with workers with other educational backgrounds. In the regression analysis, the control variables, job design variables and educational background variables were entered in separate blocks. In the results section, we look at job design while controlling for control variables; and at educational background while controlling for control variables and job design.

In comparing current and preferred work characteristics, Pearson bivariate correlations were used.

Results

Descriptive data

The average age of the respondents was 44 at the time of the research, and respondents had a work experience in activating clients of 10 years on average. Table 1 presents data on job design and educational background. A majority of respondents are responsible for activating their clients solely, and the others combine activation and benefit administration. About one half of the respondents work for clients either remote or very remote from the labour market, the others work for a mixed client group. A large majority completed higher education: in most cases (75%, not in table), this refers to higher professional education. Respondents with a university degree are a rather small group. In terms of the nature of education, three types of education are most common (which is why our analysis focused on workers with these types of education), but a rather large group of 40% of respondents has completed another type of professional training.

Table 1. Job design and educational background.

	Percentage of respondents
Job design: nature of tasks
Activation only (traditional model)	69
Activation plus benefit administration (integrated model)	31
Job design: client group
Remote from the labour market	28
Very remote from the labour market	23
Mixed client group	49
Educational background: level of education
Higher education (university or higher professional education)	83
Other (i.e. lower) level	17
Educational background: nature of education
Social work	22
Social administration	15
Personnel and Labour	23
Other	40

Table 2 provides insight into workers’ preferences, distinguishing the three categories of preferences mentioned before.

Table 2. Workers’ preferences (means and standard deviation) and their correlation with current work characteristics.

Workers’ preferences	M	SD	Correlation current work characteristics
Servicing and treatment of clients (1 = not at all preferred, 5 = strongly preferred)
Monitoring clients focuses on compliance with obligations	2.87	0.851	.333**
Action plans focus on rights/obligations	2.31	0.851	.336**
In decisions on sanctioning clients the impact of sanctions on clients’ activation is decisive	3.54	0.931	.431**
Equal treatment is more important than personalized service provision	2.35	0.870	.302**
Evidence on the effectiveness of interventions plays an important role in decision-making	3.45	0.827	.261**
Implementing rules correctly is more important than achieving results	2.15	0.825	.444**
Objective: Implementing rules and regulations meticulously	2.04	1.55	n/a
Objective: providing clients the services they need	3.23	1.35	n/a
Objective: promoting clients’ sustainable participation	2.87	1.30	n/a
Objective: promoting clients’ benefit independence	3.37	1.49	n/a
Managing frontline work (1 = not at all preferred, 5 = strongly preferred)
Sufficient room for decision-making	4.14	0.680	.086
Services are less subject to regulation	3.32	0.901	.240**
Client contacts are prioritized, even when that goes at the expense of administrative tasks	3.80	0.784	.168*
In difficult situations I consult supervisor rather than peers	2.12	0.809	.286**
My supervisor and I enter into a performance agreement	3.40	0.929	.233**
Distribution of working time
Proportion of working time spent on client contact	45%	15.03	.473**
Proportion of working time spent on contacts with colleagues/supervisors	13%	6.04	.459**
Proportion of working time spent on external contacts	24%	7.90	.642**
Proportion of working time spent on administrative tasks	14%	9.27	.645**

*p < .05.

**p < .01.

When we look at workers’ preferences concerning servicing and treating clients, the mean scores (M) are not very extreme, although standard deviations (SD) are rather high which reveals diverse opinions on preferences. Respondents are focused somewhat stronger on realizing results in activation than on implementing rules (M = 2.15). They also tend towards prioritizing personalized service provision rather than equal treatment (M = 2.35), and to prefer evidence-based decision-making (M = 3.54). The importance workers attach to various objectives of their work also shows that rule implementation is not their primary concern (M = 2.04). Nevertheless, promoting clients’ benefit independence – the core objective of Dutch activation policies – is considered the most important objective (M = 3.37), although closely followed by providing clients the services they need (M = 3.23).

Diversity is also found when we look at how respondents prefer to be managed. With one exception: sufficient room for making decisions concerning activation is strongly preferred and the standard deviation is relatively low (M = 4.14, SD = .680). There is modest aversion among respondents concerning rules regulating their work (M = 3.32) and stronger aversion against hierarchical supervision (M = 2.12). Prioritizing clients rather than administrative tasks is preferred rather strongly. Respondents have a somewhat positive attitude towards performance management (M = 3.40), which is in line with the result-oriented culture pointed at before. In terms of how respondents want to spend their working time, they want to spend the largest proportion of working time with their clients (45%), although the standard deviation is quite large.

The final column of Table 2 shows how workers’ preferences are related to how they characterize their current work. All but one of the workers’ preferences is related to how they do their current work. In most cases, workers’ preferences are in line with how they currently do their work. There are only two exceptions. Firstly, the more they experience administrative tasks as hampering service provision, the more they prefer to prioritize client contacts over administrative tasks (the current work statement read as follows: ‘Administrative tasks hamper service provision to my clients’; 1 = strongly disagree, 5 = strongly agree). The administrative pressure is also reflected in data on how respondents spend their time and prefer to spend their time: currently, they spend 31% of their time on client contacts and 32% on administration, whereas they prefer to spend 45% of their time on client contact and only 14% on administration. Secondly, the more workers experience regulations as hampering service provision, the stronger they prefer services to be less subject to regulation (the current work statement read as follows: ‘Regulations hamper service provision to my clients’; 1 = strongly disagree, 5 = strongly agree). These findings indicate that workers are more satisfied with how they service clients than with the way in which they are managed. This is also reflected in how workers rate satisfaction with their work and with their organization: on average, these ratings are 8 and 7, respectively.

Analysis of antecedents

Job design

In trying to understand variance in workers’ preferences, we will now analyse the role of the two groups of antecedents, starting with job design. The data are summarized in Table 3, column job design.

Table 3. Workers’ preferences and their antecedents (standardized regression coefficients (β); only significant effects are displayed).

Workers’ preferences	Job design			Educational background
	Client group		Tasks	Type of education			Level of education
	Very remote	Remote	Activation only	Social work	Social admin	P&L	Low
Servicing/treatment of clients (1 = not at all preferred, 5 = strongly preferred)
Monitoring clients focuses on compliance with obligations
Action plans focus on rights/obligations					−.173*
In decisions on sanctioning clients the impact of sanctions on clients’ activation is decisive
Equal treatment is more important than personalized service provision				.187*
Evidence on the effectiveness of interventions plays an important role in decision-making
Implementing rules correctly is more important than achieving results			.281**	.191*		.199*	.158*
Objective: implementing rules and regulations meticulously			.305**
Objective: providing clients the services they need
Objective: promoting clients’ sustainable participation			−.217**
Objective: promoting clients’ benefit independence	.247**					−.269**
Managing frontline work (1 = not at all preferred, 5 = strongly preferred)
Sufficient room for decision-making			−.188*
Services are less subject to regulation				−.267**
Client contacts are prioritized, even when that goes at the expense of administrative tasks			−.214**
In difficult situations I consult supervisor rather than peers			−.153*
My supervisor and I enter into a performance agreement						−.178*
Distribution of working time
Proportion of working time spent on client contact			−.293**	−.166*
Proportion of working time spent on contacts with colleagues/supervisors
Proportion of working time spent on external contacts
Proportion of working time spent on administrative tasks			.272**

*p < .05.

**p < .01.

The client group that respondents work for hardly relates to their preferences, with only one exception: respondents not working with clients very remote from the labour market consider the objective of promoting clients’ benefit independence more preferable than workers working for very remote clients (β = .247, p < .01). Working in the integrated or traditional model is somewhat stronger related to workers’ preferences. Overall, the preferences of respondents working in the integrated model show a stronger match with the bureaucratic model of service provision: the administration of rules and regulations is considered more important (β = .281, p < .01), whereas having sufficient room for decision-making is less strongly preferred (β = −.188, p < .05). Furthermore, workers in the integrated model prefer to spend more time on administration and less on clients than their colleagues in the traditional model. They are also less inclined to prefer prioritizing client contact over administrative tasks (β = −.214, p < .01). Somewhat contradictory to this picture, they are less inclined to consult supervisors rather than peers in difficult cases.

Educational background

The column educational background presents data on the relation between workers’ educational background and their preferences. Level of education hardly matters, and we only find a moderate number of cases where type of education is related to workers’ preferences. Social workers’ preferences deviate most strongly from the other respondents: they are more averse towards working according to rules (β = −.267, p < .01), prefer to spend more time on contacts with clients, and more strongly perceive providing personalized services as more important than equal treatment (β = .187, p < .05). Respondents with a Personnel & Labour background share social workers’ orientation towards results rather than rules. In addition, they are more in favour of performance management (β = −.178, p < .05). Finally, they more strongly prefer promoting clients’ benefit independence (β = −.269, p < .01).

Conclusion and discussion

This article provides insight into the antecedents of preferences of workers involved in the activation of social assistance recipients. Moderate evidence was found of the role of job design and educational background as antecedents of workers’ preferences. The sample size in our study can be an issue here; different results might be found in a larger-scale investigation.

As far as the role of job design is concerned, workers in the integrated model have a somewhat more bureaucratic ethos, which, of course, is understandable given their responsibilities for benefit administration. At the same time, this also has an impact on how they provide activation: they put a stronger emphasis on rule implementation and administration, and prefer to spend less time on client contact. Whether this impact on the activation process also has an impact on activation outcomes is an issue interesting to look at in future research (cf. Hill, 2006). In addition, we find that workers’ client groups hardly relate to their preferences, the only exception being the activation objective workers consider most important, which is in line with what one would expect (for clients very remote from the labour market, realizing benefit independence is considered hard to accomplish). Other expectations are not supported. For example, workers working for very remote client groups do not prefer personalized service provision more strongly, nor do they prefer to spend more time with their clients than workers with other client groups. One explanation for this might be that workers have different views on how to treat and service clients very remote from the labour market: some might prefer that they be left alone, as they already have a hard time managing their lives; others might prefer to support clients’ social participation. Another explanation concerns a broader issue: to what degree do feasibility considerations play a role when workers are asked to express their preferences? In the Dutch case, workers are used to the fact that activation policies are primarily aimed at promoting labour-market participation and benefit independence; and that resources available for people very remote from the labour market have been declining for years. Thus, workers’ preferences may not merely reflect considerations of desirability but considerations of feasibility as well.

Although our evidence of the nature of workers’ educational background as an antecedent of workers’ preferences is modest, the findings reveal a pattern reflecting that workers’ educational backgrounds may correspond with various professional service profiles. In terms of the ‘frames of reference’ distinguished by Eikenaar et al. (2015), social administrators come closest to the procedural frame of reference, which emphasizes allegiance to rules. Social workers resemble the client-as-needy frame of reference, with a strong orientation on clients and less emphasis on rules. Workers with a Personnel & Labour professional training resemble the work-focused frame of reference. In terms of Maynard-Moody and Musheno’s (2000) narratives of state versus citizen agents, we find that the workers in our study combine both roles but in different ‘mixes’. Social workers’ preferences most strongly resemble the citizen agent role, although their preferences concerning objectives of activation or dealing with sanctions do not deviate from workers with other educational backgrounds. Of course, we cannot rule out that ‘self-selection’ plays a role here: social workers who decide to work in agencies responsible for activation may be social workers who support activation policies. Workers with a Personnel & Labour educational background are citizen agents in as far as they consider realizing results with their clients more important than rule implementation. At the same time, they are state agents in the sense that they most strongly emphasize the core policy objective of activation: promoting benefit independence. Finally, social administrators tend to perceive themselves most strongly as state agents. Although our findings allow us to draw tentative conclusions only, we think this is an interesting venue for future research.

As we mentioned before, our research focused on the activation process rather than its outcomes; and it would be interesting to investigate whether diversity in workers’ preferences results in differences in outcomes, in terms of quantitative outcomes as well as in terms of the nature of outcomes. Insight in process and outcomes will provide stronger ground to make recommendations concerning job design and educational background of workers in activation. We conclude this paper with some preliminary thoughts on this. Despite the conclusion of Hill (2006) concerning the effectiveness of the integrated model, we do have concerns about the more rule-oriented service ethos that the integrated model seems to trigger. Closer investigation of the two models in terms of activation outcomes is relevant, as both are still being used in the job design of activation workers. As far as workers’ educational background is concerned, this issue is potentially important in terms of hiring and recruiting activation workers. The variety of policy objectives related to activation (income provision; enforcing conditionality and sanctions; promoting employability, social inclusion and participation) and the variety of ways in which the provision of activation is organized throughout the world (ranging from a mainly administrative task to forms of professional service provision) help to explain the diversity of educational backgrounds that can be found in activation. Promoting the social inclusion of the most vulnerable groups of unemployed may require other skills, qualifications and professionals (social workers?) than promoting the labour-market participation of clients (Personnel & Labour professionals?). And hiring social administrators in a context where activation work is a mainly administrative task (irrespective of whether activation and benefit administration are integrated or separated) seems sensible whereas hiring social professionals such as social workers and Personnel & Labour professionals is more suitable in an organizational context that gives room for professional service provision processes. In other words, a debate on the preferable educational background and job design of activation workers cannot and should not be detached from a debate about the objectives and organization of activation.

Acknowledgements

The authors would like to thank the anonymous reviewers for their constructive and useful comments.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Dr Rik van Berkel is Associate Professor at the Utrecht University School of Governance. His research interests include welfare state transformations, the introduction of new models for the provision of social services, the frontline delivery of welfare-to-work and the engagement of employers in promoting the labour-market participation of vulnerable labour-market groups. He participated in several EU funded networks and research projects, has been co-editor of several academic books and published, among others, in Journal of Social Policy, Social Policy & Administration, Public Administration, and European Journal of Social Work.

Dr Eva Knies is Associate Professor of Management Studies and HRM at the Utrecht University School of Governance. Her present research interests include strategic Human Resource Management (HRM); public service performance; HRM in the public sector; the role of line managers in HRM implementation (people management); and sustainable employability. She published, among others, in Social Policy & Administration, Public Management Review, Public Administration and Personnel Review.

Additional information

Funding

The study reported in this article was funded by the Dutch foundation ‘Stichting Kennis Ontwikkeling HBO’ [reference number PRO-2-012].

Notes

1 Hill (2006) also looked at specialization of activation tasks, but this dimension of job design was not included in our study.

2 In the academic debates on social services and social work, the concept of ‘personalisation’ is highly contested. See, for example, Ferguson (2012), Lymbery (2012) and Owens, Mladenov, and Cribb (2016).

3 Apart from approval for doing this research from the management of the organisations involved in our study, we did not need approval from an ethical committee to carry out this research.

4 In the Dutch context, Personnel & Labour (or Human Resource Management) is a type of education offered by schools for Higher Professional Education and universities. In most cases, people with this type of education work in HRM or personnel departments of public or private organisations. Some of them, however, start a career as activation worker.