Conservation or disappearance? The public provider of home care services in a system of choice

ABSTRACT

This article aims to analyse the position of the public provider (i.e. share of recipients) of home-care services in Swedish municipalities with marketization: a system of choice. Following the literature, an assumption is that the public provider has difficulties in surviving the competition with private home-care providers. In addition, the relevancy of this assumption could differ between different municipal settings. To test this assumption, we use statistical analysis. The main result is that the public provider is a ‘strong player’ in most municipalities experiencing marketization. However, there exists a variation in this respect – the position varies between municipalities.

KEYWORDS:

• Home-care
• elderly care
• public provider
• system of choice
• privatization
• marketization
• welfare mix
• competition
• Sweden

Background and aim

This article deals with the position of the public provider in a system with marketized elderly care. In Sweden, elderly care, which is a municipal area of responsibility, consists mainly of nursing homes for the elderly and home-care services. Over time, there has been a shift towards fewer residents in nursing homes and a greater emphasis on home-care services (Szebehely and Trydegård 2018).¹ Within the field of home-care services, marketization is an apparent feature.² This is evident through the 2009 implementation of the Act on System of Choice, hereafter referred to as LOV (lagen om valfrihetssystem).³ This act gives Swedish municipalities the option to introduce so-called choice systems in home-care services. By October 2018, the system was adopted in home-care services by 158 of 290 Swedish municipalities (SKR 2018). The marketization discussed has resulted in a ‘welfare mix’ with different types of providers: public, for-profit and non-profit (Ascoli and Ranci 2002; Sivesind and Trætteberg 2017).⁴ The exact configuration of this welfare mix is for the elderly persons to decide (given the precondition that they have been granted elderly care by the municipality’s care manager). If they choose only for-profit providers, the ‘mix’ will, accordingly, be very lopsided.⁵

In this ‘welfare mix’, the public provider is an intriguing actor, but little attention has been paid to how the public provider develops in competition with other providers in a system of choice. In the literature on ‘marketization’ and the ‘welfare mix’, most attention has been devoted to for-profit and non-profit providers. It is this research-gap that is being addressed in the article. The public provider’s role in a ‘welfare mix’ has been commented in terms of two scenarios – its conservation or its disappearance (Ascoli and Ranci 2002). One scenario is that the public provider continues to be important because of the accumulation of valuable human resources such as experience and professionalism. The contrasting scenario is that the public provider disappears because of deliberate actions taken by politicians (Ascoli and Ranci 2002) or simply because of a disappointing result in competition with other providers.

Considering this last aspect, an assumption is that the public provider encounters difficulties surviving the competition and, thus, loses care recipients to private welfare providers. This assumption is derived from previous research on the performance of the public provider in delivering services. According to Hodgkinson et al. (2017), public ownership is commonly associated in the literature with less efficiency than private ownership.⁶ This argument about the poor performance of the public provider, most commonly associated with public choice (Tullock 1965; Downs 1967; Niskanen 1971), does also denounce the public provider as inflexible and impersonal (Dahl and Lindblom 1953). The assumed inflexibility of the public provider could have a negative impact on its ability to meet the needs of particular groups. The argument has also been put forth that public organizations, in comparison with non-profit ones, are less successful in targeting particular groups of the population (Salamon 1987).

This assumption about the position of the public provider is interesting to explore in the case of Sweden, a country with competitive elderly care and strong municipalities. Sweden’s case offers an excellent opportunity to advance and nuance the knowledge about the position of the public provider regarding competition. We believe that the position of the public provider varies among different municipal settings in Sweden. In fact, despite the same laws and regulations, it is difficult to believe that the effects of LOV on the public provider would be the same in all municipalities in Sweden that have implemented the system. Accordingly, we have developed several hypotheses that take the municipal context into consideration. We will use the number of care recipients in total, population density, the history of LOV (number of years in operation) and electoral support for left-wing political parties, respectively, to explore the position of the public provider in Swedish municipalities.

To summarize, the aim of this article is to describe and analyse the position of the public provider of home-care services in those Swedish municipalities that have adopted a system of choice. We define the position of the public provider as its share of home-care recipients in the municipality.⁷ The body of research data for this article comprises official statistics from Statistics Sweden (SCB) and the National Board of Health and Welfare (Socialstyrelsen). By adopting a quantitative research design, we address the following questions: (1) What is the position of the public provider, and is there variation among the municipalities in this respect? (2) If so, how can this variation be explained? Does the municipal context have an effect or play a role?

The overall structure of the article comprises six sections, including this introduction. In the next section, we introduce the Swedish case further. We first describe central–local government relations and the decision-making bodies of the municipalities and, thereafter, the system of choice. In the third section, the theoretical part of the article is presented and we introduce our hypotheses. The fourth section discusses the research method as well as the data collected and applied to this study. The fifth section presents the findings of the empirical investigation. In the final and concluding section, we discuss the findings and their implications for future research.

Marketization and choice in Swedish elderly care

Introducing the case of Sweden

Sweden, with a population of 10 million, is described as a ‘decentralized unitary state’ (Lidström 2011). It has 290 municipalities, and local self-government is strong and is protected in the constitution. A central part of this protection includes the right for the municipalities to levy taxes from their citizens to finance different types of services (Lidström 2011). Central-government grants constitute only about 20% of the municipalities’ revenues (Erlingsson and Wänström 2015).

Local self-government is also strong in regard to statutory social services such as care for older persons, childcare and care for those with disabilities, as well as other mandatory tasks such as schools and infrastructure (SKR 2020). This robust local self-government challenges the notion of one single welfare state in the country. Instead, it has been suggested that it is more relevant to describe the Swedish welfare state in terms of a multitude of welfare municipalities (Trydegård and Thorslund 2001).

The main decision-making body in the municipalities is the directly elected municipal council, which is elected every four years by residents of the municipality. There are preparatory bodies consisting of elected politicians, i.e. municipal committees, which discuss each issue in depth before a final decision is made. These committees are organized according to the different policy fields of importance to the municipality. For instance, matters that concern caring for older people are considered by the committee of social services or similar (Gustafsson 1999). Mandatory care for the elderly population is an important task for municipalities, accounting for 19% of the municipalities’ expenditures and 3% of GDP (Socialstyrelsen 2019a).

Marketization

The Swedish municipalities have the right to decide about marketization of care for their older population. They also have the power to decide about the type of model to use for the procurement of such care. The most established of these models derives from the Public Procurement Act (lagen om offentliga upphandlingar, LOU) and can be summarized in the phrase ‘winner-takes-all’. In this model, the contract is awarded to the actor with the most attractive bid in terms of the quality descriptions or cost of services – or both (Segaard and Saglie 2017). In Sweden, this model has been applied, for instance, in the procurement of nursing home services for elderly persons (Feltenius 2017; Broms, Dahlström, and Nistotskaya 2020).

By contrast, a system of choice such as LOV accommodates more than one ‘winner’. In fact, all welfare providers that meet the specified criteria in the tender document (förfrågningsunderlag) stipulated by each municipality are given permission to offer services (Government Bill 2008/09:29). Service providers who gain permission then become an option that an individual can choose from in the system (Jordahl 2013; Feltenius and Wide 2015, 2019; Moberg, Blomqvist, and Winblad 2016; Wide and Feltenius 2016). However, in the case of home care services, the elderly must first have been granted this particular service, a decision made formally by a responsible care manager of the municipality. Moreover, this decision includes the content of care provided (the particular input) and the number of care hours granted (Wolmesjö 2014).

This procedure reflects the fact that the municipality pays for the larger portion of the home-care services regardless of who the provider is (Erlandsson et al. 2013). Hence, approved providers receive a fixed sum from the municipality based, for instance, on the number of hours of service provided. The elderly person’s fee for the services, which is regulated by and paid directly to the municipality, covers only a small portion of the total sum. The national average for this is, according to the latest survey, about 6% (Socialstyrelsen 2014). The remaining portion of the cost is financed by the municipalities through local taxes and central government grants.

Providers of home-care services

The practice of LOV has been most evident in home-care services (Feltenius and Wide 2015, 2019), which has resulted in different types of providers that include for-profit, non-profit and public. It is notably in the context of LOV that the size of different types of providers depends on the choice made by the elderly individual. The inclusion of a provider in a system of LOV does not imply any guarantee in terms of a certain number of care recipients. The same applies to the public provider. Similarly, providers are not allowed to reject a request from an individual recipient who wants their services (Government Bill 2008/09:29).

In 2018, there were about 516 unique private providers in all in the 158 municipalities with LOV. Most private providers are small with a local base. The two largest private providers are the for-profit companies Attendo Care (in 24 municipalities) and Team Olivia (in 11 municipalities) (SKR 2019a). Non-profit providers are only occasionally evident in Swedish municipalities with LOV. This has been explained by the attempt in the postwar period to establish universal eldercare services in the Swedish municipalities, with no room for other providers except for the public one. When marketization was introduced, the non-profit sector was slow to adapt to a new role as welfare provider (Blomqvist and Winblad 2019).

The public provider in a ‘welfare mix’

The ‘welfare mix’

The presence of different types of welfare providers in a system of choice and marketization has been described in terms of a ‘welfare mix’ (Bode 2006; Theobald 2012; Johansson, Arvidson, and Johansson 2015; Sivesind and Saglie 2017). For example, in their book ‘Dilemmas of the Welfare Mix’, Ugo Ascoli and Costanzo Ranci analyse the impact of privatizations on welfare systems. According to Ascoli and Ranci, the new structure arising is mixed considering the existence of both different types of providers and different forms of regulation and coordination between actors (Ascoli and Ranci 2002).

In this welfare mix, two providers have received special attention: non-profit and for-profit. The non-profit providers have received attention, for instance, through comparative research on the role of the non-profit sector in welfare states (Salamon and Anheier 1996; Salamon et al. 1997). Attitudes of local political parties towards non-profit providers in the welfare area have also been examined (Johansson, Arvidson, and Johansson 2015) as well as why countries as Sweden have a smaller proportion of non-profit providers than other countries (Blomqvist and Winblad 2019). Another example compares non-profit providers with other types of providers (Sivesind 2017; Trætteberg 2017; Trætteberg and Fladmoe 2020).

Attention has also been given to for-profit providers, for instance comparisons of the biggest chains in long-term care in countries such as Canada, Norway, Sweden and the US (Harrington et al. 2017). Studies has also been performed on the growth of for-profit providers in the sector of residential care for children and youths in Sweden (Meagher et al. 2016). Another example investigates contract relations between public authorities and for-profit providers, along with questions on accountability in relation to the contracts (Isaksson, Blomqvist, and Winblad 2018; Blomqvist and Winblad 2020). A further example is research examining if for-profit welfare providers also have become influential actors in the political process (Pieper 2018).

However, the public provider has received significantly less attention in the literature. One reason might be that the public provider has been ‘taken for granted’: it has always been present and will continue to be so. Ascoli and Ranci mainly pay attention to the role performed by non-profit providers, but they are commenting on the role played by the public provider. In their previously mentioned book ‘Dilemmas of the Welfare Mix’, two scenarios have been developed regarding the public provider: ‘conservation’ and ‘disappearance’ (Ascoli and Ranci 2002).

The explanation to the state of ‘conservation’ is the perceived waste to discard valuable human resources, such as experience and professionalism, built up for many years in the organization of the public provider. Interestingly, Ascoli and Ranci believe that those resources could be highly valuable in an expected cooperation between public, for-profit and non-profit providers. The opposite scenario, ‘disappearance’, is explained by deliberate actions among politicians in favour of far-reaching privatization (Ascoli and Ranci 2002). However, we also believe that there could be another reason for ‘disappearance’: a disappointing result in the competition with other providers due to poor performance.

Public organizations

In their classic ‘Politics, Economics and Welfare’, Robert Dahl and Charles Lindblom discussed the pros and cons of public administration. Among the disadvantages mentioned were ‘red tape’, that is, an excessive use of rules and procedures that result in rigidity in the operations of public administration. Other disadvantages mentioned were ‘inflexibility’, ‘passing the buck’, ‘failure in economizing’, ‘excessive centralization’ and ‘a culture of impersonality’. For instance, impersonality is evident in relation to both employees within public organizations and the clients that they are hired to serve. Another example is inflexibility, i.e. an excessive rigidity resulting in an incapacity for making changes (Dahl and Lindblom 1953). These noted disadvantages of public administration are still being discussed, 70 years after the original publication of Dahl and Lindblom’s study. In fact, there is a widely held belief in the literature that public organizations operate less efficiently and less effectively than private ones (Hodgkinson et al. 2017; Rainey and Chun 2005).⁸

This notion of the public provider has been made referring to different aspects of publicness, such as ownership, control and funding (Andrews, Boyne, and Walker 2011). ‘Ownership’ builds on arguments from the economic theory of property rights. According to this theory, owners and shareholders in private organizations have stronger motives for controlling the behaviour of managers. By contrast, individual voters are not expected to have the same strategic motives for controlling the performance of public organizations, with negative consequences in terms of efficiency as a result (Andrews, Boyne, and Walker 2011).

Another dimension of publicness is ‘control’. In public sector organizations, there are several expressions of control, such as audits, inspections and performance reports. Further, these methods can be performed by principals at different levels – central, regional and local. If the total of all these controls is too large, a negative impact on the efficiency and effectiveness of public service delivery may result. That is, instead of delivering the actual services they are expected to deliver, public organizations may be too busy reporting different aspects of their operations. Too many controls may also cause harm for public organizations in the way that they can impose conflicting demands (Andrews, Boyne, and Walker 2011).

Finally, the category of ‘funding’ considers Niskanen’s public choice theories of bureaucracy (Niskanen 1971). Niskanen argued that it makes a difference whether citizens pay directly for a service. If the citizen pays the organization directly, then the organization delivering the service is believed to be more attuned and sensitive to feedback and any complaints articulated by the citizen. This implies that public organizations, which foremost obtain their resources indirectly through taxation, are considered unresponsive to the people who receive the services (Andrews, Boyne, and Walker 2011).

The public welfare provider

This rather gloomy view of the performance of public organizations is also evident in the literature on marketization and different types of welfare providers. It reflects the fact that the public provider constitutes a segment of the wider public administration. An important point of departure when discussing the public provider is that it belongs to the political steering chain of representative democracy (Christensen, Lægreid, and Røvik 2020). The public provider is the ‘agent’ in this chain, while the elected politicians operate as the ‘principal’ that issues the instructions. In this system, accountability is vital. The ‘principal’ is accountable to the citizens and, in turn, must be able to hold the ‘agent’ accountable for its actions (Warren 2014). One way of ensuring this is the practice of the Weberian system for organizing the bureaucracy, with hierarchies of command and control (Pollitt and Bouckaert 2011).

In this context, the public provider and its services are governed by written plans adopted by municipal politicians. Considering this top-down steering, it is assumed that politicians make their decisions with the ‘best interests of the municipality’ and ‘the average citizen’ in mind, rather than particular groups (Trætteberg 2015). As a result, the content of care offered by the public provider ultimately tends to be designed according to the ‘one size fits all’ principle, which leaves little room for ‘tailor-made’ solutions that meet the specific individual’s needs for care (Salamon 1987; Trætteberg and Sivesind 2015).

The inflexibility of the public provider to offer such individualized, tailor-made solutions has prompted the conclusion that such solutions are, for the most part, provided by other actors in the ‘welfare mix’. Here, it is assumed that non-profit actors have a greater likelihood of providing tailor-made services (Salamon 1987). This is possible since non-profit providers, as argued in the literature, often apply other strategies in their operations than public alternatives do (Salamon and Abramson 1982; Salamon 1987; Osborne 1998, 2010; Mariani and Cavenago 2013).

Besides the inflexibility in offering tailor-made solutions, the public provider has also been criticized of being impersonal. In previous research, one factor identified as important for the individual’s choice between different providers, was the size of the provider. Smaller private providers, with a limited staff, have a competitive advantage since they can offer a service that becomes more personal (Svensson and Edebalk 2006; Vadelius 2015). Staff continuity has also been pointed out as one of the most important quality aspects of home-care services (Edebalk, Samuelsson, and Ingvad 1995; Moberg, Blomqvist, and Winblad 2016; Socialstyrelsen 2019c). The public provider, on the other hand, represents a larger staff and organization, increasing the probability that the care recipients receive a more impersonal service and meet many different employees. One example is the LOV-municipality Kungsbacka in western Sweden. In Kungsbacka, elderly (in early 2020) who had chosen the public provider, met in average 22 different employees within a period of two weeks, while the figure for the private providers were 11–13 different employees (Norra Halland 2020).

In summary, although the public provider has attracted less attention than for-profit and non-profit providers, the attention has often been critical. First, the public provider has been described as impersonal in the sense that due to its large organization, the care recipients meet too many employees. Second, it is described as organized hierarchically and steered by elected politicians, with little input from the care-giving organizations. This is considered to allow little – if any – room for differentiation of public welfare services. In sum, it leads to the assumption that the public provider has a difficult time trying to survive in the competition with other types of providers. Accordingly, our first hypothesis is that the public provider (in municipalities with LOV) has a lower share of care recipients than other welfare providers. However, we also assume that the position of the public provider differs between municipalities, which will be elaborated below.

Variation among municipalities

An important variable for consideration in this respect is population structure and geography. Previous research about marketization in general has identified the existence of a geographic pattern: private providers are more successful in central cities than in peripheral rural areas (Hartman 2011). Within nursing homes for the elderly, an uneven geographic distribution in marketization has been shown. The privatization of nursing homes in Sweden is more common in areas with a higher population density (Winblad and Isaksson 2013). It is likely that municipalities with fewer potential customers and a lower population density are less attractive to private providers of home-care services because of a limited possibility of economy of scale and profitability (Stolt and Winblad 2009). Instead, the public provider is expected to carry a heavier assignment and responsibility in municipalities with one or both of these characteristics. Our second hypothesis is that the position of the public provider (in municipalities with LOV) is dependent on the number of care recipients in total. Our third hypothesis is that the position of the public provider (in municipalities with LOV) is dependent on the population density.

Moreover, the tradition of LOV, and the length of time it has been in use in the municipality, is likely to be relevant as well. The system of LOV and the entry of private providers represent a major shift in home-care services in Sweden. LOV was first introduced in 2009, and the number of municipalities with LOV has increased over time, which will be shown later. People are often described as ‘creatures of habit’, and adapting to new practices can take time. Hence, in a newly introduced system of LOV, the public provider might have an advantage since the ‘new’ providers are not yet known. As the awareness of LOV grows, the share for the public provider might decrease. We also believe that the establishment of private providers increases over time, for example, as new companies are formed. Our fourth hypothesis is that the position of the public provider (in municipalities with LOV) is dependent on the history of LOV in the municipality.

In the case of marketization, ideology concerning the privatization of care for elderly persons also matters clearly in several respects. Liberal and conservative parties are more in favour of privatization than left-wing parties are (Guo and Willner 2017). Previous research has indicated a relationship – although weak – between individuals’ ideological orientations and their opinions about welfare services. In the case of Sweden, citizens supporting left-wing parties tend to be less satisfied with social services when the public share of welfare services decreases. This reflects a more sceptical attitude towards marketization and, hence, a higher propensity that citizens choose the public provider for ideological reasons (Hardell, Johansson Sevä, and Öun 2020). In short, in municipalities with a political orientation towards the left, the public provider would, indeed, survive the competition with private providers. In municipalities with a political orientation towards the centre-right, the private providers are expected to be more successful. We choose ‘voter support’ as a variable instead of the parties’ seats in the municipal council since it is, ultimately, the citizens who choose between the private or public providers. Another important reason is that cross-bloc local government occurs frequently in Sweden (in one third of the municipalities in 2014–2018, see SKR 2019c) which makes a possible effect of the parties’ seats in the municipal council or the municipal board difficult to interpret. Our fifth hypothesis is that the position of the public provider (in municipalities with LOV) is dependent on the electoral support for left-wing parties.

Research method and data

To be able to discuss and evaluate the accuracy of our hypotheses, we conducted an empirical study of Swedish municipalities. The case of Sweden, with a system of choice, provides an excellent opportunity to advance and nuance the knowledge about the development of the public provider in competition with other providers. We used statistical analysis to explore the position of the public provider of home-care services in Swedish municipalities. By ‘position’, we refer here to the public provider’s share of all home-care recipients in the municipality. We calculated this variable from data on the number of individuals accessing home-care from different types of providers, i.e. public and private, on municipal level. We limit the study to elderly care and accordingly recipients who are 65 years and older. The data has been gathered by the National Board of Health and Welfare (Socialstyrelsen 2019b).

We investigated those Swedish municipalities that had operated according to the Act on System of Choice (LOV) in home-care services in 2018 (158 of 290 municipalities), which was the most up-to-date statistics. Data is not available for all municipalities since twelve municipalities have not reported the required statistics to the National Board of Health and Welfare. Accordingly, 146 out of the 158 municipalities with LOV are included in the analysis.

In the empirical analysis, we first set the context by showing the number of municipalities with LOV as well as the public provider’s share of all home-care recipients in the municipalities. Second, the relationships between the position of the public provider and the independent variables (as stated in the hypotheses above) are investigated by bivariate correlation analysis. In the third step, the effects of the independent variables on the position of the public provider are tested by multivariate regression analysis.

A list of the municipalities with LOV and the year they adopted LOV have been provided from the Swedish Association of Local Authorities and Regions (SKR 2018, 2019b). Data regarding the population structure and election results of the municipalities have been gathered from Statistics Sweden (SCB 2020). Since all data investigated constitute official statistics, we consider them to be reliable and of high quality. The variables and measures of spread are presented in Table 1. In the forthcoming analysis, we have log-transformed the number of care recipients and population density to prevent extreme values from potentially distorting the results.

Table 1. Included variables in the empirical analysis. Minimum value, maximum value, mean value and standard deviation. N = 146

Variables, municipal level	Minimum	Maximum	Mean value	Standard deviation
Number of home-care recipients in total (all providers), 2018	106	18,073	1,119	1,882
Share of home-care recipients with the public provider (as a percentage of home-care recipients in total), 2018	0	100	84.2	20.6
Population density (inhabitants per square kilometre), 2018	0.80	5,819	256	766
Years with LOV (up to 2018; adoption of LOV 2018 = 0)	0	6	5	1.53
Votes on the Social Democrats (as a percentage of all votes), municipal election 2014	6.9	46.7	30.9	7.2
Votes on the ‘Alliance’ (as a percentage of all votes), municipal election 2014	21.9	82.6	42.9	8.6

Sources: Data from SCB (2020), SKR (2019b) and Socialstyrelsen (2019b).

Findings

The position of the public provider in Swedish home-care services

In Figure 1, the number of municipalities with LOV (all services) during 2010–2019 is shown. Statistics by type of service is not available over time. However, most municipalities that have decided to adopt LOV have done so within home-care services.⁹ There are a few municipalities (fewer than five per year) that have implemented LOV only within other services. For example, in 2018 there were 160 municipalities with LOV, of which only two municipalities (Avesta and Lomma) had not adopted LOV within home-care services but only in other social services (SKR 2018; Valfrihetswebben 2021).

Figure 1. Number of municipalities with LOV 2010–2019

Note: The data includes home-care services as well as other services, but few municipalities with LOV has not implemented it in home-care services. Total number of municipalities in Sweden: 290.Source: Data from SKR (2019a).

Figure 1 illustrates a clear trend demonstrating that the number of municipalities with LOV has plateaued in the past few years. However, every year, several municipalities decide to abolish LOV, whereas others choose to adopt it. There are two main reasons that municipalities abandon LOV within home-care services: either few care recipients have chosen private providers, or no private providers have been established in the municipality. In these cases, the cost of managing a system of choice is considered greater than the benefits it provides (SKR 2019a).

By deciding on LOV within home-care services, the municipality attempts to create a market of different providers. In 2018, a total of 75% of the care recipients in Swedish municipalities with LOV still had their home-care services delivered by the public provider (SKR 2019a; Socialstyrelsen 2019b). Accordingly, public home-care is dominant in the country as a whole. Despite this, the picture is more diverse: Sweden can rather be described as a heterogeneous landscape of marketized home-care services. Some municipalities with LOV host only private providers, while in others, there are no private providers at all (SKR 2019a).

The data in Table 2 highlight that, in a group of 9 municipalities, 0–50% of the care recipients have the public provider. In a group of 31 other municipalities, the percentage is 50–80%. In a large group of 106 municipalities, 81–100% of the care recipients have the public provider option.

Table 2. Share of care recipients with the public provider in home-care services, 2018

Percentage of care recipients with the public provider
	0–20%	21–40%	41–50%	51–60%	61–70%	71–80%	81–85%	86–90%	91–95%	96–99%	100%	Total
Number of municipalities	5	3	1	5	11	15	15	14	17	27	33	146

Note: Number of municipalities with LOV within home-care services in each category, 2018. Total: 146 municipalities because of missing data. Total number of municipalities with LOV within home-care services: 158. Total number of municipalities in Sweden: 290.

Source: Data from Socialstyrelsen (2019b).

But why do we see this variation between the municipalities? One explanation might be that some municipalities are simply more lucrative for private providers than others. Another might be that it takes time for the private providers to have an impact, and different municipalities have introduced LOV in different years. Finally, a third explanation is that there is a higher demand by care recipients for home-care services offered by the public provider in some municipalities than in others. In summary, we suggest that the municipal context matters.

Municipal structure and the public provider of home care

Densely populated urban municipalities with large potential client bases are generally more enticing to private providers than are sparsely populated areas with limited opportunities for efficiency and economies of scale (see Stolt and Winblad 2009; Hartman 2011). One example of a municipality with LOV in the first category is the Swedish capital of Stockholm, with 962,154 inhabitants and a population density of 5,140 inhabitants per square kilometre. An example of a municipality with LOV in the second category is Storuman, a small municipality in northern Sweden with 5,912 inhabitants and a population density of <1 inhabitant per square kilometre (SCB 2020). In total, there are 18,073 home-care recipients in Stockholm, and 185 home-care recipients in Storuman (Socialstyrelsen 2019b).

In Figure 2, we show the relationship between the percentage of home-care recipients with the public provider and the number of home-care recipients in total and, in Figure 3, the relationship between the percentage of care recipients with the public provider and the population density of the municipalities. From the figures, it can be seen that there is a negative relationship between the variables. The fewer recipients in total and the lower the population density, the more recipients have the public home-care provider. We have further analysed our data with bivariate correlation analysis, in which we log-transformed the number of care recipients and population density to prevent extreme values from potentially distorting the results. There is a significant negative correlation between care recipients with the public provider and the total number of care recipients (Pearson’s r = −.356***; p = .000; N = 146), as well as between care recipients with the public provider and population density (Pearson’s r = −.464***; p = .000; N = 146).

Figure 2. Correlation between care recipients in total and care recipients with the public provider

Note: To facilitate interpretation, we excluded municipalities with the largest number of care recipients (extreme values) from the figure.Sources: Data from Socialstyrelsen (2019b).

Figure 3. Correlation between population density and care recipients with the public provider

Note: To facilitate interpretation, we excluded municipalities with the highest population density (extreme values) from the figure. Sources: Data from SCB (2020) and Socialstyrelsen (2019b).

As Figures 2 and 3 illustrate, there are deviant municipalities. This might depend on structural characteristics other than the population structure. For example, there are several suburbs of Stockholm with a small population size. In these cases, the closeness to other suburbs and good opportunities to coordinate the work are probably more critical to private providers than the mere number of care recipients to compete for (Stolt and Winblad 2009). However, this might also depend on other explanatory variables. Thus, we need to go further and investigate other types of contextual characteristics of the municipalities as well.

Years with LOV and the public provider of home care

The tradition of LOV in the municipalities is expected to affect both the care recipients’ inclination to choose a provider other than the public one, as well as the formation and establishment of private companies in the municipalities. This means that the public provider should be subject to more pressure over time. A bivariate correlation analysis shows a significant but weak negative correlation between care recipients with the public provider and the number of years with LOV in the municipality (Pearson’s r = −.268***; p = .001; N = 146). It suggests that the public provider is larger in the municipalities that have recently adopted LOV. Yet it is important to note that these ‘latecomers’ can be characterized by other common factors, such as political ideology and socioeconomic structure.

Political ideology and the public provider of home care

Based on previous research, we also assume that the ideological preferences of politicians and citizens might matter for the position of the public provider (see Stolt and Winblad 2009; Elinder and Jordahl 2013; Guo and Willner 2017; Hardell, Johansson Sevä, and Öun 2020). More specifically, the public provider is expected to have a stronger position in municipalities with a political orientation towards the left and vice versa. Thus, we investigated the role of political ideology as well. We analysed the relationship between the percentage of care recipients with the public provider and the percentage of votes for the Social Democrats (the dominant left-wing party) and for the ‘Alliance’ (a coalition of four liberal and conservative parties), respectively, in the 2014 municipal election, i.e. the election to the term of office (2014–2018) which the data on the dependent variable refer to.

Bivariate correlation analyses show, firstly, a significant positive correlation between care recipients with the public provider and the percentage of votes for the Social Democrats (Pearson’s r = .351***; p = .000; N = 146), which is also made visible in Figure 4. Secondly, there is a significant negative correlation between care recipients with the public provider and the percentage of votes for the ‘Alliance’, i.e. the four liberal and conservative parties (Pearson’s r = −.294***; p = .000; N = 146).

Figure 4. Correlation between votes on the Social Democrats and care recipients with the public provider. Sources: Data from SCB (2020) and Socialstyrelsen (2019b)

Explaining the variation

In the last step of the analysis, we used multivariate regression analysis (OLS) to explain the variation between the municipalities. We constructed a basic contextual model to estimate the effect of population structure, years with LOV and political ideology on the position of the public provider. The results of the regression analysis are presented in Table 3.¹⁰

Table 3. Estimation of the models’ effect on percentage of care recipients with the public provider. Unstandardized regression coefficients and heteroskedasticity consistent standard errors

	Model A		Model B		Model C
	B	Std. Error	B	Std. Error	B	Std. Error
Care Recipients in Total	−4.460**	1.618	−3.898*	1.768	−4.417**	1.549
Population Density	−3.973***	1.208	−3.837***	1.217	−4.130***	1.188
Years with LOV	−2.639**	.874	−2.512*	.991	−2.704***	.815
Voter support for the Social Democrats	.115	.289	.574*	.287	.	.
Voter support for the ‘Alliance’	−.538*	.224	.	.	−.597**	.216
Constant	162.542***	18.275	120.435***	12.416	169.349***	13.790
R^2	.343		.317		.342
Adj. R^2	.320		.297		.324
N	146		146		146

*** p ≤ 0.001; ** 0.001 < p ≤ 0.01; * 0.01 < p ≤ 0.05

Note: Care recipients in total and population density have been log-transformed to prevent extreme values from potentially distorting the results.

Sources: Data from SCB (2020), SKR (2018, 2019b) and Socialstyrelsen (2019b).

In the first model (model A), we included all the independent variables. The analysis indicates that the public provider of home-care services is more limited in municipalities with many care recipients, a high population density, a longer tradition with LOV and strong voter support for the ‘Alliance’ parties. Interestingly, no significant effect of the voter support for the Social Democrats is detected. Therefore, we chose to exclude the voter support for the ‘Alliance’ (model B), which resulted in a significant effect for the Social Democrats but a somewhat reduced explanatory power. When we, instead, excluded the voter support for the Social Democrats (model C), the explanatory power increased. The political context does have an effect on the position of the public provider, but the voter support for the ‘Alliance’ (-) is clearly more important than voter support for the Social Democrats (+). The explanatory power of the models in Table 3, however, is 30–32%. Accordingly, a large proportion of the variation remains unexplained, which leaves the floor open for discussion on other types of explanatory factors. Prospective research questions will be discussed next along with a summary of the main results.

Discussion and conclusion

This article aimed to analyse the position of the public provider of home-care services in Swedish municipalities that have adopted a system of choice (LOV). We defined the position of the public provider as its share of home-care recipients in the municipality. The case of Sweden is interesting since municipal home-care services consist of a ‘welfare mix’ with different providers: public, for-profit and non-profit. The configuration of this mix is a matter for the elderly persons in need of care in the municipality to decide. Hence, there is no guarantee that a provider in a system of LOV will have a certain share of care recipients. The following research questions have been addressed: (1) What is the position of the public provider, and is there a variation among the municipalities in this respect? (2) If so, how can this variation be explained? Does the municipal context have an effect or play a role?

The empirical findings suggest that the public home-care provider continues to be a ‘strong player’ in the market in most Swedish municipalities that have adopted LOV. This finding contradicts our first hypothesis, that the public provider should have a lower share of care recipients than other welfare providers. However, our findings also illustrate a variation among the municipalities. Taken together, the results of the multivariate analysis show that the political, institutional and structural variables have an effect on the position of the public provider in Swedish municipalities. All the hypotheses concerning the variation, stated in the beginning of our article, were confirmed. The public provider has a stronger position in smaller municipalities with higher voter support for left-wing parties and a more limited experience of LOV (i.e. number of years in operation). According to the literature, the public provider could be expected to have a stronger position in sparsely populated areas since these areas are less attractive for the establishment of private providers. This theoretical proposition seems to be evident also in the case of home-care services. In addition, the proposition that ideology matters hold true. If the population in a municipality is politically oriented towards the left on the left–right scale, the prospects for private providers to attract elderly persons are less optimistic. Instead, there is stronger support for the view that the provider should be public, which also reflects the choice of provider that is being made.

The empirical results presented here contribute valuable insights to the debate on the role of the public provider in a ‘welfare mix’, evident in systems with marketized care for the elderly. In this debate, there are two scenarios on the role of the public provider: ‘conservation’ and ‘disappearance’ (Ascoli and Ranci 2002). The results presented here give limited support to the scenario of ‘disappearance’, which suggests that the public provider has difficulties surviving in the competition with other providers due to poor performance (in terms of inflexibility, impersonality and inefficiency). Even though we have not investigated performance per se, our research has shown that the public provider is successful in attracting elderly care recipients. Contrary to the theoretical assumption of the poor performance of the public provider, our results might, therefore, suggest that the public provider is performing well in the sense that it can offer a competitive service in relation to other type of providers.

Although our research findings overall support the scenario of ‘conservation’, we believe that this conclusion must be further investigated with a focus on development over time. One of the variables tested here, number of years with LOV, is of particular interest since it suggests that the position of the public provider might decrease over time. Our interpretation of this finding is that people, especially those who are older, might stay with providers who are familiar and have been in the municipality for a long time. The public provider certainly meets this criterion, considering the fact that, for a considerable amount of time, it has been the one and only provider of home-care services in Swedish municipalities. However, as years go by, other providers will become familiar too. It is, therefore, important to repeat this study and to follow up on the position of the public provider. This will enable us to draw robust conclusions over time. Thus, a prospective question might be: Is it just a matter of time before the public provider decreases significantly?

We also see a need to explore other factors that can explain the position of the public provider. It has been argued that the municipalities, in the case of marketization, could tilt the playing field in their favour (Blix and Jordahl 2021). In our case, there is a possibility that the municipality would prefer a design of the tender document more favourable to the public provider. For instance, the municipality is able to create different geographical areas within the system (Konkurrensverket 2013), which could promote the public provider. Besides geography, it has been claimed that the public provider does not, contrary to the intentions of the law, compete with private providers on the same terms financially.¹¹ In this study, we have not been able to take those types of factors into account, but this is important to consider in future studies.

We also believe that there are other explanatory factors that are difficult – or even impossible – to capture and analyse in a quantitative study. Among them are specific, local contextual factors such as the actors’ use of strategies, i.e. long-term planning with the aim of maximizing future benefits. Traditionally, research on organizational strategies in the public sector has been limited, but the research field has been growing in recent years (Andrews et al. 2009; Höglund et al. 2018). In this case, it would be of interest to explore whether municipalities use strategies to position the municipal home-care service provider in relation to the private providers. We suggest that, if the municipalities abstain from formulating strategies for the public home care provider, the result might be a smaller share of care recipients – and vice versa.

If strategies have been formulated, these could be either towards ‘convergence’ or ‘divergence’ in relation to private providers. By ‘convergence’, we mean that the public provider models itself after private providers by imitating their ideas, solutions and profiling. Accordingly, by ‘divergence’ we refer to the opposite scenario, namely that the public home-care provider attempts to distinguish itself from private providers. One way is to present eldercare recipients with the idea that the public home-care provider represents something completely different in terms of fundamental values. In a strategy of ‘divergence’, economic values (spending public funds economically) are important, but democratic values are prioritized. In these cases, we also believe that the public provider relies more on its historical reputation than on seeking to maximize the number of ‘customers’ by advertising and marketing.

The importance of strategies concerning the public home-care provider in a context of marketization is an interesting topic for future research. We believe that in-depth attention to the public provider is a significant element of research on marketization within eldercare. This could contribute to a better understanding of what happens with the public provider when the inflow of care recipients is not guaranteed as it was in the days before the system of choice. Some aspects of the public provider’s operations and activities might have improved, while others have not. Of particular concern here is whether the democratic values traditionally associated with a public provider have weakened.

Acknowledgments

The authors would like to thank participants at the EGPA Study Group IV (Belfast, 2019), the SWEPSA Annual Conference (Norrköping, 2019) and the NORKOM Research Conference (Åbo, 2019) as well as three anonymous reviewers for valuable comments and feedback.

Disclosure statement

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

Additional information

Funding

This work was financially supported by the Swedish Research Council for Health, Working Life and Welfare (FORTE) under grant dnr [2018-00423].

Notes on contributors

David Feltenius

David Feltenius is an Associate Professor at the Department of Political Science, Umeå University. In addition to studies on marketization and the welfare state, his research focuses on central-local government relations and territorial politics. His recent publications include a chapter on local state-society relations in Sweden (together with Anders Lidström) in the book Close Ties in European Local Governance as well as an article on civil society organizations in a marketized Swedish welfare state published in Journal of Civil Society. Feltenius is currently working in a research project on public welfare providers in a marketized welfare state.

Jessika Wide

Jessika Wide is a Senior Lecturer at the Department of Political Science, Umeå University. Her research interests deal with elderly care and marketization; user participation; local democracy; and women in politics. Her recent work appears in the Swedish Journal of Political Science, Journal of Civil Society and others. Wide is currently working in a research project on public welfare providers in a marketized welfare state.

Notes

1. By home-care services, we mean service as well as personal care in the individual care recipient’s own home. In 2019, 8% of the population aged over 65 and 22% of the population aged over 80 received home-care services (Socialstyrelsen 2020).

2. Marketization is also evident in the welfare states of other Nordic countries; see for instance Rostgaard (2006); Hansen (2010); Anttonen and Häikiö (2011); Meagher and Szebehely (2013).

3. The law is issued by the Swedish government in the following bill: Government Bill (2008/09:29).

4. Often the literature only makes the difference between ‘public’ and ‘private’ providers, thus not distinguishing between ‘non-profit’ and ‘for-profit’ providers. Se further: Trætteberg and Fladmoe (2020).

5. For an overview of the pros and cons of ‘choice’ in this context, see: Rostgaard (2006); Moberg, Blomqvist, and Winblad (2016).

6. For a literature overview, see for instance: Boyne (1998); Rainey and Bozeman (2000); Andrews, Boyne, and Walker (2011).

7. This variable is more suitable when comparing municipalities than the public provider’s share of performed hours, which varies depending on the number of recipients who require more comprehensive care.

8. However, this notion has been debated and it has been argued that there is a lack of enough empirical evidence (Boyne 1998; Pollitt 2003).

9. As mentioned, LOV is most common within home-care services (158 municipalities in 2018) but is also adopted within other social services, such as adult daily activities (32 municipalities), nursing homes for the elderly (21 municipalities), companion service (19 municipalities), relief service in the home (19 municipalities) and family counselling (15 municipalities) (SKR 2018).

10. A control for multicollinearity does not indicate any problems (collinearity statistics result in VIF <2; tolerance value >.045). However, since problems with heteroskedasticity occur, the models have been reestimated by using heteroskedasticity consistent standard errors, HC3.

11. This claim has been made by the Swedish Association of Private Care Providers who are referring to statistics produced by the Swedish Association of Local Authorities and Regions (SKR). See further: Vårdföretagarna (2016).