Privatizing the central core of social work. Exploring the use of agency social workers in the Swedish social services

ABSTRACT

The use of staffing agencies and agency social workers in statutory social work has not been given much attention from scholars, despite the extensive implications it may have for the organizations, employees and clients. By analysing the result of structured telephone interviews with representatives from 102 Swedish municipalities, this article brings some knowledge to this scarcely researched field. The aim is to map the extent of use of agency social workers, to analyse the reasons behind the use, and to identify any potential variation between municipalities and branches of social work. The findings show that the use of agency social workers is common, widespread and surprisingly unrelated to socioeconomic, geographic and demographic characteristics of the municipalities. The most common reasons for hiring agency workers appear to be related to achieving numerical stability of staffing levels, despite vacancies and recruitment difficulties. Although there were few detected differences between municipalities, there were great differences between branches of social work. It was far more common for child welfare units to contract agency social workers compared to social assistance and substance abuse units. Considering the vulnerability of the children that come into contact with the social services and the importance of continuity in these cases, this is a cause for concern. In addition, practically all hired agency social workers were contracted to investigate and assess the clients’ needs, i.e. partaking in the exercise of public authority, which adds legal and professional complexity.

KEYWORDS:

• Social services
• staffing agencies
• agency social workers
• marketization
• privatization

Introduction

In a number of current studies and reports, the Swedish social services has been described as a system under pressure. Working conditions have been depicted as poor, turnover high and recruitment difficult (Tham 2016; Astvik, Welander, and Isaksson 2017). Reports have shown that the majority of Swedish municipalities have difficulties recruiting both experienced and newly qualified staff for their units (SKL 2016) and as a result, it appears as if many of them have come to rely on staffing agencies, i.e. for-profit companies hiring out agency social workers on short-term contracts, for solving the problem with understaffing. Other countries have seen a similar development in the social services; in for example the UK, the need for agency social workers, i.e. social workers that are employed by staffing agencies but rented out to user organizations, is reported to have increased in later years due to turnover and vacancy levels (Jones 2018).

The tendency to employ staffing agencies in the social services can be understood against the background of an increasing marketization of the public sector. From the late 70s and onwards, public sectors in most of the western world have undergone reforms in line with the movement often referred to as New Public Management (NPM) (Hood 1995; Pollitt and Brouckaert 2011). One of the central tenets of NPM is that the public sector has become ineffective and unaffordable, and that it would benefit from introducing market mechanisms and managerial ideas from the private sector. Consequently, NPM is often associated with a preference for contracting out and an increased focus on cost-effectiveness (a.a.). Just like other parts of the public sector, the social services have been subjected to changes in line with the ideals of NPM. Several studies have focused on effects of such changes, for example, changed working conditions for social workers and managers (Kirkpatrick et al. 2005; Shanks 2016; Astvik, Welander, and Isaksson 2017), and the increased outsourcing of service provision (Jones 2018; Storbjörk and Stenius 2019; Lundström, Sallnäs, and Shanks 2020). Studies regarding the latter issue have shown that the social services for several decades have outsourced a considerable amount of the service provision to private companies. While this type of outsourcing – i.e. services taking place outside of the social services – has gained interest among researchers and others, the outsourcing of tasks performed within the social services (i.e. statutory social work such as investigations and assessments of clients’ needs) has largely gone under the radar. This is surprising, considering that there appears to have been cases of outsourcing such tasks to agency social workers at least since the beginning of this millennium. This trend may be considered as a consequence of the same ideal that has caused the marketization of other parts of the social services, namely to increase the cost-effectiveness and decrease in-effectivity by introducing market solutions (Madell and Hollander 2003). However, unlike the outsourcing of service provision, we lack research on the outsourcing of tasks traditionally performed within the social services. We do not know the extent of the use of agency social workers in the social services, or what type of work that they are hired to do. We also largely lack knowledge regarding the consequences of the use of agency workers. British studies from some years ago have discussed the practice as a way to maintain service during periods of vacant positions and high workloads, but also identified risks related to e.g. working conditions and quality of work (Hoque and Kirkpatrick 2008; Carey 2011; Cornes et al. 2013). It is likely that some of these conclusions are valid also for the Swedish social services.

This article is a part of a larger project intending to fill the knowledge gap regarding the use of agency social workers and the consequences it may have on organizations, permanently employed social workers, the clients of the social services and the agency social workers themselves. In this article, we aim to explore the extent of the use, the potential variation between municipalities and branches of social work, and the reasons behind the use. This knowledge lays the foundation for the understanding of an under-researched phenomenon that appears to have increased during the last decade, and that is likely to be of great importance for all affected parties. We will discuss organizational motives and conditions related to the use of agency social workers, and the results will also be discussed in the light of the legal complexity surrounding this issue. In future publications within the project, this initial mapping will be followed by qualitative studies that explore the consequences of the use of agency social workers. Before presenting the findings, we will briefly outline the current knowledgebase regarding the use of agency social workers as well as the legal aspects of such use.

Reasons for and consequences of the use of staffing agencies

The use of agency social workers in statutory social work has received surprisingly little attention, not least in Sweden. Internationally, the issue has received some interest, although it is still a scarcely investigated field. Prior research comes from the UK, and was published in the beginning of the last decade. In this research, both benefits and drawbacks for relevant parties have been discussed.

For social work organizations, hiring agency social workers can be a way to maintain service during periods of vacant positions and high workloads, and it may be economically beneficial in some circumstances as it has the potential to limit the costs of recruitment and training (Hoque and Kirkpatrick 2008). In research focusing on the use of agency workers in general, two different types of organizational motives have been discussed; motives related to numerical stability and motives related to numerical flexibility of staffing levels. Whereas the former cover motives such as replacing absent staff, the latter involves being able to adjust the number of workers to match actual demand (and hence being able to quickly cut down on staff if needed). It has been shown that stability motives are more frequent in public sector organizations than in private sector ditto (Håkansson and Isidorsson 2007).

For the employed, staffing agencies may increase flexibility and expand the labour market. For social workers who possess public sector positions, the use of agency staff may decrease workloads and boost morale during times of understaffing (Hoque and Kirkpatrick 2008). In terms of negative consequences, it has been indicated that a long-term reliance on agency staff may be costly. It may also lead to an unequal distribution of tasks. As a result of agency workers’ lack of organizational-specific knowledge, the permanently employed may end up with the most demanding tasks. However, the opposite tendency has also been noted, i.e. agency social workers feeling compelled to take on the most demanding cases out of fear of being laid off if they do not (Cornes et al. 2013). As for possible consequences for the quality of work, these issues have been less studied. However, concerns in terms of lack of continuity and difficulties with relations with clients have been raised (Carey 2011). Research within adjacent fields (health care) indicates that the use of agency workers, in this case locum doctors, enable organizations to maintain appropriate staffing levels and allows flexibility in terms of staff, but also that it gives rise to concerns in terms of continuity, safety, teamwork and cost (Ferguson and Walshe 2019). Other studies indicate that locum doctors and/or agency nurses may contribute with new professional knowledge to the organization (Ruiner et al. 2017; Jansson and Engström 2017).

From the perspective of agency workers, there is a risk of being viewed as peripheral, and getting less access to in-house training, etc. (Carey 2011; Cornes et al. 2013). The general research on temporary agency staff (in all fields of work) has highlighted the triangular relationship between the agency worker, staffing agency and user organization. Agency staff have employment relations (determining contracts and wages) with the staffing agency and management relations (concerning every-day work) with the user organization. In addition, the user organization and the agency have a business relation and need to negotiate the price and quality (Håkansson and Isidorsson 2007, 2012). These circumstances may affect aspects such as the working conditions and potential peripherality of agency workers.

Within research that has analysed the driving forces behind an individual’s move into agency work in different fields of work, two explanatory models can be identified. One model is concerned with pull-factors of the staffing agencies, i.e. appealing factors that draw professionals towards work in agencies. This model views agency workers as ‘free agents’, i.e. professionals who opt out of permanent positions to enjoy the flexibility and beneficial pay rates within the staffing industry. The other model focuses on push-factors of the public sector, i.e. repelling factors that drive professionals away from permanent positions. This model depicts agency workers as personnel escaping from deteriorating working conditions (Kunda, Barley, and Evans 2002). In the empirical research focusing on the social services, support for both (or a combination) of these models has been found (Kirkpatrick and Hoque 2006; De Ruyter et al. 2008; Cornes et al. 2013). With regard to possible push-factors, there is considerable support in research for increasingly adverse working conditions in the social services, in Sweden as well as elsewhere (Mor Barak, Nissly, and Levin 2001; Lloyd, King, and Chenoweth 2002; Kim and Kao 2014; Tham 2016; Astvik, Welander, and Isaksson 2017). Connections between poor working conditions and social workers leaving their jobs have been established (Lloyd, King, and Chenoweth 2002), and literature has underlined the importance of these factors for social workers opting for agency work (De Ruyter et al. 2008).

To sum up; apart from a few British studies from the beginning of the last decade, there is very little research regarding the use of agency social workers in the social services, including basic research mapping the extent of use. In Sweden, such research is absent. Other sources (e.g. media reports and reports from trade unions) indicate that using staffing agencies is common among Swedish municipalities and that it may be a costly practice (SSR 2014). Research is needed to establish the extent and variation of use, as well as and the consequences for all parties.

Regulations

Since 1993, it has been legal to hire out agency workers for profit-making purposes in Sweden, both to public and private organizations. Unlike in some other countries, staffing agencies do not need a licence to sell their services in Sweden, and there are no restrictions regarding length of contracts etc. (Håkansson and Isidorsson 2016). Work in the staffing industry is regulated by collective agreements and agency workers are protected by the Employment Protection Act (Lag om anställningsskydd, 1982) in the same way as other employees on the Swedish labour market. Among other things, the act states that the typical employment contract should be permanent (with some exceptions), and that an employee that has been temporarily employed for two years should be offered a permanent position. Hence, it is reasonably common for agency workers to have a permanent position in the staffing agency (Håkansson and Isidorsson 2007).

The jurisdiction regarding the use of agency workers in the social services is complex, and the legal sources handling this issue are few. However, there are restrictions that are related to the exercise of public authority. Although it is possible that the agency social workers are hired to do work in the social services that is unrelated to the exercise of public authority (e.g. work in the municipally owned provider units), several sources indicate that this is normally not the case. Instead, agency workers appear to be engaged in investigating and assessing the needs of clients (Hollander 2005; SSR 2014). If a function of the social services involves exercise of public authority, the possibilities of handing it over to a private actor is limited. Therefore, the question of which functions within the social services that should be viewed as exercise of public authority becomes important. Whether or not an investigation of a client’s situation (undertaken by professionals) is such a function is a question that has engaged the Parliamentary ombudsman (JO). JO concluded in a statement (2001/02) that it is not only the decision-making that should be viewed as exercise of public authority, but also the preparatory work, i.e. the investigation. According to this statement, the possibilities for the municipalities to contract external staff to perform investigations are limited. This state of affairs limits the possibilities for the municipalities to contract external staff to perform investigations. However, despite these limitations, municipalities are allowed to contract agency workers as long as they contract individual workers (in such a way that they can be considered a part of the organization), and as long as the decision-making is not delegated to the agency worker. In an organization in which it is very difficult to monitor the daily work of the professionals, and where discretion is deemed necessary (Brodkin 2008), this arrangement raises questions regarding the balance between the professional discretion of agency social workers and the degree of insight that the decision-maker can have in these cases.

Just as there are legal constraints regarding the use of staffing agencies, there is also jurisdiction covering the procurement of such services. In fact, most countries have specific legislation regulating public procurement and, in the EU, all countries must comply with the EU procurement directives. In Sweden, most public procurement is done in accordance with the Public Procurement Act (LOU). It is common for municipalities to arrange procurement through framework agreements, which means agreements between one or more municipalities and one or more suppliers, with the purpose to determine the terms for an award of contract at a later date (i.e. when the service is needed). As seen in the results section, most purchases of agency social workers were done through such agreements. However, a relatively large proportion were not. The legal aspects, both those relevant for discretion and decision-making of the agency worker/organization and those relevant for purchasing services from staffing agencies, need to be considered if/when agency social workers are hired to perform work within the social services.

Methods

Sweden’s 290 municipalities have extensive autonomy regarding the provision of social services, meaning that the organization of the social services can differ between the municipalities. Nevertheless, child welfare, substance abuse treatment for adults and social assistance (i.e. a means-tested cash benefit which in Sweden is administered by trained social workers) are often seen as the three core branches of Swedish public social work, and are often referred to as the personal social services (PSS). These three branches of practice are often organized into separate units. It is also quite common that these units are divided into subunits based on function, e.g. specialized units for investigation and treatment (and sometimes intake) (Lundgren et al. 2009). Together with social work with disabled persons and the elderly, the PSS generally make up the social services in the Swedish municipalities. It is the use of agency social workers in the Swedish PSS that is the focus of this study.

The majority of municipalities in Sweden are small; around 80% of all municipalities have less than 50 000 inhabitants. In the selection of municipalities for this study, a stratified sample based on population was drawn in order to avoid excluding larger municipalities. The sample consisted of 40 municipalities with more than 50 000 inhabitants, and 62 municipalities with less than 50 000 inhabitants (in total 102 municipalities). In the larger municipalities such as Malmö and Stockholm, a reasonably representative district (in terms of socioeconomic characteristics and demography) was chosen. The selected municipalities are located in all parts of Sweden, and a comparison between the included municipalities and all Swedish municipalities classified into groups by the Swedish Association of Local Authorities and Regions¹ indicates that the selection resulted in a sample of municipalities that is reasonably representative in terms of municipality type (see Table 1). Overall, the municipalities that have been selected for this study can be viewed as fairly representative; there are no reasons to believe that the use of agency social workers differs significantly between these and other municipalities.

Table 1. Comparison between selected municipalities (n = 102) and all Swedish municipalities (n = 290).

Municipality type	Proportion of all municipalities (n = 290)	Proportion of selected municipalities (n = 102)
Larger city and municipalities close to larger city	37	41
Metropolitan cities and areas	16	20
Smaller municipalities and rural municipalities	42	39

Data were collected through structured telephone interviews conducted January 2019 – December 2019. The reasons for using telephone interviews rather than, e.g. a survey was partly to ensure an appropriate response rate, considering the general downward trend in the response rate to survey studies (Cook et al. 2009), and partly to increase the chance of directing our questions to the right person, i.e. an employee with knowledge of the municipality’s use of agency workers. The position of such employees differed between municipalities. In most municipalities, the interviews were held with the managers of PSS, who generally have the overall budgetary responsibility. However, in some municipalities, the organizational structures of PSS were such that the interviews had to be held with lower line managers who had personnel and budgetary responsibility for their departments. In these cases, it took more than one telephone interview to gain information from all the departments within PSS. In some municipalities, the PSS manager also referred the interviewer to a lower level of managers who had been responsible for hiring staff in their departments. Some of the managers preferred to answer the interview questions by email after the initial contact. The interview questions focused on the use of staffing agencies, the most important tasks of the hired social workers, the reasons for hiring, which agencies that had been used, etc.

Descriptive statistics were used to display the overall use of social workers in the participating municipalities (Table 2). As there is very little knowledge regarding the use of agency social workers and variation between municipalities, we explored bivariate associations between having hired agency social workers any or both of years 2017 and 2018 as well as having hired agency social workers during the month preceding the interview 2019, and socioeconomic/demographic variables (number of inhabitants, geographic location of the municipality, proportion of unemployed inhabitants, proportion of inhabitants that received social assistance, proportion of inhabitants with low income, proportion of inhabitants with low education, proportion of child poverty, proportion of inhabitants born in other countries or tax capacity for the relevant years) using a simple logistic regression. As socioeconomic factors have been found to correlate with adverse working conditions in the social services (Tham 2018), we wanted to explore the possibility of them also being associated with the use of agency social workers since these phenomena theoretically could be related. These analyses resulted in very few statistically significant associations, but indicated that the use of agency social workers was widespread. Need for hiring a large number of agency social workers could potentially also be related to socioeconomic/demographic variables, for the same reasons as having hired social workers at all. Therefore, we also wanted to explore such associations. Since the data did not allow linear regression (the dependent variable was not normally distributed, and the data was based on estimates of the interviewees), simple logistic regression was used to investigate possible bivariate associations between number of social workers hired in the municipalities (categorized as ten or more), and the socioeconomic/demographic variables. This analysis resulted in some statistically significant bivariate associations between having hired ten or more agency workers and percent of people born in other countries. A multiple logistic regression was thereafter used in order to control for potentially confounding variables, e.g. variables that have been found to be associated with adverse working condition in the social services (cf. Tham 2018) (Table 3).

Table 2. Use of agency workers 2017–2018.

	Percent
Proportion of municipalities reporting use of agency social workers (n = 102)
2017	64
2018	63
2017 and/or 2018	74
Number of hired agency social workers 2017 and 2018 (n = 74)
1–5	55
6-10	26
10-	19
Use in different units 2017 and/or 2018 (in municipalities that reported use, n = 75)
Child Welfare	95
Social assistance	27
Substance abuse	25
Other	15
Tasks of the agency social workers 2017 and/or 2018 (in municipalities that reported use, n = 75)
Performing investigations/assessments	100
Treatment/provision of care	0
Management positions	12
Most important reasons for hiring 2017 and/or 2018 (in municipalities that reported use, n = 75)
To fill vacancies	52
Recruitment difficulties	44
High workloads	41
Other reasons	13
Procurement procedure 2017 and/or 2018 (in municipalities that reported use, n = 74)
Framework agreement	65
Direct award of contract	35

Table 3. Associations between municipal-level socio-demographic characteristics and having hired more than ten agency social workers 2017–2018. N = 74.

	Multiple regression OR (95% CI)
Percent of inhabitants born in other countries	1.19 (1.08–1.32)**
Percent of inhabitants that received social assistance	1.49 (0.47–1.61)
Percent of inhabitants with low income	0.82 (0.59–1.42)
Percent of inhabitants with low education	0.91 (0.71–1.16)
Percent of unemployed inhabitants	0.87 (0.47–1.61)
Number of inhabitants	1.00 (1.00–1.00)

CI = confidence interval. OR = odds ratio, * = p < 0.05, ** = p < 0.01. Bold numbers indicate statistical significance.

Results

Use of staffing agencies

The result from the interviews confirms the general picture of agency social workers as a common phenomenon in the Swedish social services (see Table 2). Of the municipalities included in the study, around three quarters reported that they had hired agency workers during 2017 and/or 2018 (n = 75), the majority of these reported hiring agency workers during both years (n = 54). The length of the contracts varied considerably, from less than a month to the entire investigation period (24 months). The majority of municipalities that had hired agency workers had done so for a period of six month of more (n = 51), and the total number of agency social workers varied between one and 50 (median was five). Furthermore, more than one-third (37%) of all municipalities reported use of staffing agencies during the month preceding the interview (January–December 2019). The number of agency workers in the respective municipalities during this month varied between one and fifteen.

Almost all municipalities that reported use of agency social workers stated that they had hired them to fill positions in child welfare units (n = 71). However, around one-fourth of the municipalities reported that they had hired agency workers for positions in social assistance units and substance abuse units, respectively (n = 20 and n = 19). In all of the included municipalities, the agency social workers had been used to perform investigations, i.e. the preparatory work preceding a decision. Respondents from nine municipalities stated that they also had hired agency social workers to fill management positions. The most common reasons for using agency social workers were, according to the respondents, vacant positions, recruitment difficulties and/or high workloads. In a few municipalities, other reasons were described; e.g. need for agency workers in situations with conflicts of interests, need of social workers with certain competence etc.

Around one-fourth of the municipalities (n = 27) had not hired agency social workers during the investigated period. A bit more than half (n = 15) of these municipalities had according to the interviewees been able to cope with permanently employed staff or had not experienced difficulties with recruiting new permanent staff. However, respondents in the rest of the municipalities (n = 12) stated that refraining from hiring agency workers was an active strategy rather than as an absence of need. These respondents explained that they had experienced that the costs of agency social workers did not match the quality, and that hiring agency social workers negatively affected the working conditions in the unit. However, it is of course possible that the staffing situation in these 12 municipalities was more manageable than in the municipalities that reported use of agency social workers.

The market

Of the municipalities that had hired agency social workers during 2017 and 2018, the majority (n = 48) had framework agreements in place. However, several (n = 16) of these 48 municipalities reported that they had purchased services through direct award of contracts when they had not been able to find suitable services through their framework agreement, or when they had been displeased with the services that they had received from the procured companies. One-fourth of the municipalities (n = 26) lacked framework agreements altogether and had solely purchased services through direct award of contracts.

In Sweden, there appear to be a rather large number of staffing agencies to choose between, and although the exact number of such companies is unknown, 45 different companies were named in the interviews. Many municipalities had used a range of different staffing agencies; some respondents mentioned as many as ten different companies. There is no research regarding the characteristics or development of the ‘staffing agency market’ for social work, but judging from the Parliamentary ombudsman’s statement (2001/02) and websites of staffing agencies, it appears to have kicked off in a smaller scale around 2000. Data from annual reports from the company with today’s largest market share (which also was the most commonly mentioned company in this study, used by 28 municipalities) indicate a notable increase of the demand for agency social workers around 2015, when Sweden saw an increase in the number of refugees. According to the same annual reports, the demand continued to increase until 2017. After that, the reports instead describe a decreased demand (https://www.dedicare.se/om-oss/finansiellt-och-press/finansiella-rapporter/). These data should however be interpreted with caution, but lacking more reliable information, they provide an indication of how the market has evolved.

Variation between municipalities

Analysis of the data from the interviewees shows few differences between the municipalities that had and those that had not hired agency social workers during the investigated period (2017, 2018 and 2017 and/or 2018).² In fact, almost none of the demographic and socioeconomic characteristics that were analysed were associated with having hired agency social workers. Hence, instead of observing differences between, for example, disadvantaged and more affluent municipalities, we found that most types of municipalities at some point during the investigated years had hired agency social workers. The only exception was that municipalities with a higher proportion of immigrants 2018 appeared to a somewhat greater extent have hired agency social workers during the same year (OR = 1.07, p = 0.03). No associations between having hired agency social workers and factors related to the internal functioning/cost of the local authority (political governance and costs for the personal social services and its respective units) could be found. For 2019, much of the statistics regarding socioeconomic and demographic characteristics was unavailable. However, the information that could be retrieved (number of inhabitants, proportion of unemployed inhabitants, proportion of inhabitants born in other countries, tax capacity and geographic location) did not show association with having hired agency social workers during the month preceding the interview³ (January–December 2019).

Turning attention instead to potential differences between municipalities in terms of number of hired agency social workers during 2017–2018 (n = 74),⁴ we found some bivariate associations between having hired more than ten agency social workers and socioeconomic characteristics of the municipalities. Municipalities with a higher proportion of inhabitants born in other countries (aged 18–64, average of 2017 and 2018) appeared to a greater extent have hired more than ten agency social workers during the period of 2017–2018 (OR = 1.11, p = 0.003, not in table). In a model controlling for other socio-economic characteristics, the association between having a higher proportion of inhabitants born in other countries and having hired more than ten agency social workers was still valid (Table 3). None of the other measures made a significant contribution to the model.

Discussion

This study set out to explore the use of agency social workers in Sweden, to analyse potential variation between municipalities and branches of social work, and to explore reasons behind the use. The findings depict the use of agency social workers in Swedish municipalities as a widespread and common phenomenon. It appears to be largely unrelated to socioeconomic or demographic characteristics of the municipalities, although the proportion of inhabitants born in other countries appears to be associated with the number of hired agency social workers. The agency social workers are exclusively hired to perform investigations, and the use of agency social workers varies significantly between the different branches of social work. These results will be discussed below, as well as the reasons behind the use of agency social workers.

Recruitment difficulties, vacancies and high workloads

Looking at the most common reasons for hiring agency social workers as described by the interviewees in this study, it appears as if numerical stability of staffing levels is the main goal, not flexibility (Håkansson and Isidorsson 2007). The use is also not described as a way to increase effectivity or quality of the professional work. Instead, the reasons for hiring agency workers in the social services – vacancies, recruitment difficulties and high workloads – appear to be related to efforts of making the organization function in times of staffing difficulties. The use of agency workers may be interconnected and related to strenuous working conditions for social workers that has been reported in most of the western world (e.g. Mor Barak, Nissly, and Levin 2001; Lloyd, King, and Chenoweth 2002; Kim and Kao 2014; Blomberg et al. 2015; Tham 2018). Potentially, there is a risk for a downward spiral; recruitment difficulties contribute to vacancies, vacancies may mean high workloads for remaining staff, which at worst could lead to increased sick leave and turnover. Hiring agency workers could be a (temporary) way out of this spiral, as it relieves the permanent staff of some of their workload.

These findings are in line with those in the international research, where the use of agency workers in the social services has been discussed as a way for organizations to provide service despite vacancies and high workloads (Hoque and Kirkpatrick 2008). When services must be maintained irrespective of vacancies (which is the case in public social work), agency workers may be viewed as a solution. They can start working almost immediately and, at least judging from the information provided on the staffing agencies’ websites, they are experienced. This means that, unlike with newly employed staff, little or no introduction should be needed. Instead, agency social workers should immediately be able to start relieving the permanent staff of some of their workload. However, statements in the interviews in this study indicate that the quality of services provided by agency workers does not always live up to expectations, and that municipalities at times have sidestepped framework agreements as a result of experiencing poor quality from the procured staffing agencies. Some municipalities have contracted as many as ten different staffing agencies during the investigated years. These results indicate that there may be a considerable variation in quality between different agencies and/or agency social workers, and that the stability that the organization strives for is not always achieved with the agency workers.

Although not highlighted by the employees, it may be economically beneficial for municipalities to hire agency social workers in the short term, if formal and informal introductory training can be avoided (Strolin, McCarthy, and Caringi 2007). It is also possible for the municipality to achieve flexibility of staffing levels; being able to end the contract should the service no longer be needed. However, long-term reliance on agency staff may be expensive since the costs for such services exceed the cost for salary for permanent staff. Should agency social workers need significant introduction and support during their time in the municipality, and/or deliver substandard investigations, the price may be high – both economically and in terms of human distress. In addition, hiring agency social workers outside framework agreements and without following the proper procurement procedures may both be expensive and illegal.

Widespread use

There have been indications that working conditions for social workers are worse in areas that are disadvantaged compared to those that are more affluent (Tham 2018). Considering this, and e.g. the geographical differences of municipalities, it is a bit surprising that the use of agency social workers appears to be largely unrelated to socioeconomic, geographic and demographic characteristics of the municipalities. The reason for the lack of differences (both in terms use/no use, and in terms of extent of use) between municipalities with differing characteristics may simply be that the need for staffing backup is widespread and common. Given that recruitment difficulties are an important reason behind the widespread use of agency social workers, it may be that municipalities in general have failed to be viewed as attractive employers that are able to offer decent working conditions for social workers, irrespective of their socioeconomic and demographic status. In a social work labour market – which has become more extensive in Sweden due to an increase of private companies offering other types of social work employment – municipal/public social work may have trouble to compete and social workers may opt for other types of employment, even agency social work (Socialstyrelsen 2016; Shanks et al. 2018). Hence, both push and pull factors may be at work here and it is an empirical question which of those is the most important.

One significant exception to the general lack of association between socioeconomic characteristics and use of agency social workers was found – municipalities with a larger proportion of inhabitants born in other countries appeared to a greater extent have hired ten or more agency workers. This is not surprising. In Sweden, when a refugee/immigrant is provided a resident permit, it is the municipalities’ responsibility to accommodate them and to offer them services. The immigration to Sweden increased considerably during the years preceding 2017 (and peaked 2016), which caused additional strain on the municipalities. This situation may have caused an increased need for social workers in the municipalities that received a large proportion of refugees.

Performing child welfare investigations

Although few significant differences between municipalities were found, there were great differences between branches of social work. As discussed above, it was far more common for child welfare units to use agency social workers compared to other units. Considering that the situation in the Swedish child welfare has been described as particularly challenging, with high turnover, an increasing proportion of newly qualified social workers (Tham 2018) and difficulties to recruit new staff (SKL 2016), this comes as no surprise. The situation may however be described as suboptimal, seeing that turnover and workforce instability have been connected to failure to protect children (Healy, Meagher, and Cullin 2009), and negative outcomes for children in contact with the child welfare services (Williams and Glisson 2013). Relying heavily on agency social workers may have a negative impact on the workforce stability in the social services since the contracts of agency workers are temporary and usually short.

Furthermore, it is mainly for investigating and assessing the clients’ (chiefly children) situations that agency social workers are hired. As described above, the legal aspects surrounding the use of agency social workers for this purpose are complex. According to the legal advice, the municipalities are in principle not allowed to delegate the exercise of public authority to other legal entities or individuals. Nevertheless, agency workers are according to the same sources allowed to exercise public authority – e.g. performing investigations – under some circumstances, and provided that they do not make formal decisions based on their investigations. Under the prevailing circumstances, if this advice is followed, the representatives of the municipalities (often in the form of line managers) must make the decisions based on preparatory work (investigations) into which they have little insight. This set of circumstances means that the agency social worker is responsible for the often intrusive and delicate task of investigating the clients’ needs (which may involve gathering sensitive information from other authorities, etc.), but not for the decision that is made on basis of the result of the investigation. Responsible for the decision is instead a potentially uninformed manager.

Some municipalities stated that agency social workers were hired to fill management positions. This is an interesting finding. Although it has been reported that municipalities have experienced problems with recruiting managers (Vision 2019), we have not been able to find any (Swedish or international) research regarding ‘agency social work managers’. Managers in the Swedish social services are often decision-makers (making decisions delegated by the politicians). Some of the decisions they make cannot be further delegated to lower levels in the organizations. How the delegation of decisions is arranged when the manager is an agency worker (to whom the decision-making cannot be delegated) is a question that needs to be investigated further.

Limitations

This study builds on data from 102 of Sweden’s 290 municipalities. While attempts have been made to ensure representativeness, there is a slight overweight of large municipalities in this study. It is however unlikely that this has affected the results significantly. Somewhat more worrying is perhaps that the information regarding agency workers come from interviews with managers who may have provided us with estimates rather than exact information. This is likely to particularly have affected the information about the number of hired agency social workers and for this reason, the analysis of this data has been adapted accordingly (e.g. has been categorized into broad categories). The multiple logistic regression using these estimates (Table 3) is also based on relatively few observations, why this result is somewhat tentative. Finally, the scope of this study was limited to map the use of agency social workers in the Swedish social services, which means that there is no basis for drawing conclusions regarding its consequences. This is an issue that needs to be explored in future research.

Conclusions and implications

Overall, this study has established that hiring agency social workers is commonplace in Sweden, particularly within child welfare. The use of agency social workers appears to be viewed mainly as a way of maintaining numerical stability in the organization, e.g. filling vacancies and compensating for recruitment difficulties. The motives behind the use as described by the respondents are not related to flexibility in terms of being able to quickly cut down on staff, nor to increase effectivity or quality of the work performed in the organization (this does not necessarily mean that the use of agency social workers cannot contribute towards such development during their assignments). When hired, agency workers are almost exclusively used to investigate and assess clients’ needs, which makes the use legally complicated. Unlike with the provision of social services (e.g. treatment), there are legal constraints aiming to prevent the public sector from using private companies in matters that involve the exercise of public authority. Despite this, staffing agencies appear to have found a way into this very central core of the social services. Previous research has shown that the use of agency social workers can be helpful for strengthening and relieving an organization in times of vacancies and high workloads. However, it may negatively affect the stability that is required to build competence in the workforce. Also, the fact that the agency social workers most often are used in child welfare is somewhat worrying, given the sensitive nature of this field and the documented importance of continuity.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

Forskningsrådet för Hälsa, Arbetsliv och Välfärd [2018-00840].

Notes

1. Swedish municipalities are classified into three major groups by the Swedish Association of Local Authorities and Regions. The classifications are based on structural parameters such as population, commuting patterns and economic structure.

2. Dependent variables: Hired agency social workers during 2017 yes/no (1/0), hired agency social workers during 2018 yes/no (1/0), Hired agency social workers during 2017 and/or 2018 yes/no (1/0).

3. Dependent variable: Hired agency social workers during the month preceding the interview, yes/no (1/0).

4. Dependent variable: Hired more than 10 agency social workers 2017–2018, yes/no (1/0).