Introduction

A fundamental dilemma for the Scandinavian welfare state is that policies – supposedly generous, universalistic, egalitarian and uniform – are formulated on the national level but implemented on the local level (Esping-Andersen 1990, Hanssen 1997, Kautto, Fritzell, Hvinden, Kvist & Uusitalo 2001, Kröger 1997a). But how can it be a uniform policy, if there is a local influence?

This tension has been the focus of research in different social welfare areas. Investigating of public old-age care in Sweden, Trydegård and Thorslund find that in spite of national goals of equality, universality and equity a “considerable regional inequality exists in Sweden in the care and services provided for the oldest in the community … [I]t might be more appropriate to talk about a multitude of different ‘welfare municipalities’ … rather than one single welfare state” (Trydegård & Thorslund 2001:183). In his study on child day-care provision, Kröger states that “the traditional principle of Scandinavian local self-government, requiring local autonomy, and the goal of a welfare state providing a uniformly equal service to all is … a contradiction in terms”. A centralist model stands against a local autonomy model (Kröger 1997b:485). O'Higgins argues that “territorial justice can only be ensured by continually intervening to restrict or direct choices which local authorities may make” (O'Higgins 1987:11). Taking a look outside Scandinavia, scholars emphasize the necessity of a “spatial strategy for equality” since the geographical variation, for example in the British welfare service, seems haphazard and difficult to understand (Powell & Boyne 2001). Local government finance in England often goes contrary to expectations. National control is a prerequisite for equal treatment; “who says equity, says centralization” (Boyne, Powell & Ashworth 2001:33). Daley, Johansson, Malmberg and Sundström (2006), however, give a brighter picture of the social welfare state. In their study of home and community-based services for the elderly and people with disabilities in Sweden they argue that equal treatment, in the sense of uniformity and measured as equal coverage rates, does not necessarily mean that the support is given in an equitable fashion in response to need.

This article aims at widening the research on the problem of spatial equity to a welfare policy area so far not investigated: social services in kind to people with severe congenital disabilities or other severe lifelong disabilities that have occurred before the age of 65. The case studied is that of Sweden. Support to disabled people varies to a great extent between the Swedish municipalities. Some give 10 times as much support per 10,000 inhabitants than others (Lewin & Westin 2005). Mass media depict municipalities with a more restrictive policy for people with disabilities as “the municipalities of shame” (Expressen 2003).

In the 1990s a new law, LSS (SFS 1993:387, Act concerning support and service for persons with certain functional impairments), was introduced in order to give disabled people a stronger position in the welfare system and with an explicit purpose to overcome the local variation in the support. To quote the commission that prepared the act:

There are many municipalities and their qualifications, such as competence and resources, as well as ambitions for disabled people vary to a large extent. If you regard the rights of disabled people as associated with citizenship, that is as citizen rights, it is hard to find any arguments why the conditions should depend on where you live. (SOU 1991:46, pp. 113–114, our translation)

The commission did not want people with disabilities to be left to the “arbitrariness of the local political majority” (p. 114, our translation). In contrast to international legislation in this policy area (see, for example, the report on European disability legislation, Council of Europe 2003) LSS has a strong mandatory character. Support is given as an unconditional right to people with the most severe disabilities attained before the age of 65. Available municipal resources are not allowed to influence the decision-making, as under the Social Services Act, which is a more general, goal-oriented framework law, leaving ample room for the municipalities to interpret and realise the goals in accordance with local ambitions and resources. The latter offers a similar kind of personal support as LSS but has a lower ambition: “decent” living conditions, as compared to “good” living conditions. Achieving good living conditions demands a higher level of ambition and less room for discretion and, in practice, typically more far-reaching and expensive social support measures of a higher quality. With Kröger (1997b) and Boyne et al. (2001) it can thus be argued that LSS is an exponent of a centralist perspective and a tool for the political endeavour to reach the spatial equity that has not seemed attainable through the Social Services Act. The unique legal construction of LSS makes the implementation especially worthy of note.

The purpose of this article is to analyse the local variation in the provision of LSS support and to explore whether the variation is associated with a variation in needs or rather with political and structural conditions in the municipalities. Two more specific questions are raised: do different local ambitions, as noted by the commission, still characterize Swedish disability politics? Is there a spatial inequity in the implementation of the disability support?

The study has an explanatory purpose as well as a normative one. In the next section the theoretical framework is elaborated. The hypotheses are then tested in bivariate and multivariate regression analyses and the results are discussed. Finally, we take a stand in the matter of whether the municipalities have succeeded in realising the political intention of spatial equity.

Theoretical Framework

Municipal variation in the provision of LSS support can be expressed in several ways. In his discussion of the welfare state, Gøsta Esping-Andersen points out that research on welfare output should not narrowly focus on welfare costs. Instead he stresses the importance of the rules and standards of welfare programmes. He uses the term de-commodification, i.e. “the degree to which individuals, or families, can uphold a socially acceptable standard of living independently of market participation”. The potential for de-commodification is dependent on how accessible the welfare programmes are and what level of support they permit (Esping-Andersen 1990, chapters 1 and 2, cit. p. 37). Since the expressed goal of LSS is to make it possible for disabled people “to live as others do” (section 5), we regard social support as a vital aspect of de-commodification.

Different measures of expenditure are a common way to assess welfare output. However, we do not want to equate costs with needs fulfilment, which is sometimes done in welfare research (see e.g. Hörstedt, Prütz, Wells, Edebalk & Lindgren 1996). Our solution is to use various measures of the provision of LSS support, related to the size of the municipal population. As can be seen in Table 1, the municipal variation is considerable (see Methods section below). Some municipalities only have 13 LSS persons per 10,000 inhabitants; others have 10 times as many. A similar pattern is seen for the other variables.

Table 1. Descriptive statistics for variables measuring the provision of LSS support.

Variable Min Max Mean Standard deviation N
LSS persons* 13 131 57 17 289
LSS measures* 20 289 99 34 289
LSS costs** 589 5856 2501 852 289
*Measured per 10,000 inhabitants.
**Measured in SEK per inhabitant.
Sources: Data on LSS costs, LSS persons and LSS measures from the National Board of Health and Welfare (Socialstyrelsen 2003a, b); Data on population 2002 from Statistics Sweden (SCB).

As Hanssen (1997) points out there is a longstanding research tradition looking at the variation in the policies of local governments. So-called output studies commonly seek explanations to local welfare policies in social needs, local government resources and local political preferences. Similarly, provision of support is usually seen as governmental reactions to demands for welfare. In accordance with Tarschys (1975), Hanssen recommends an interactive model with different explanatory factors on three levels: socio-economical, socio-cultural and institutional. We too recognize this “complicated mix of factors in interaction” (Hanssen 1997:113) in our attempt to explain the municipal variation in provision of LSS support. Our model contains four concepts: needs, propensity to apply, political ambitions and resources. The first two phenomena concern the demand side and the latter two the supply side. Each phenomenon is associated with a number of hypotheses. Many of the explanations proposed here frequently occur in welfare literature, though some are more specific to the case of municipal disability care.

The theoretical framework is presented in Figure 1.

Figure 1. 

Theoretical framework explaining the municipal variation in the provision of LSS support.

Needs

The most intuitive answer to why disability support varies among municipalities is that there is a variation in the actual need for support. It is a fundamental problem in this kind of research that we have no objective measure of needs. Public statistics cannot give information on the number of potential rights-bearers, that is, disabled people who have difficulties in their daily life making them eligible for LSS support.

If all needs were met, the number of persons with LSS support and the number of LSS measures would together mirror the “true” prevalence of severely disabled people and their needs; the coverage rate would be 100%. In research and evaluation of disability reforms, an administrative definition is commonly used that presupposes such a relationship; provision of support serves as a measure of prevalence of needs (Barron, Michailakis & Söder 2000). This may be true in an ideal process, where the extent of needs equals that of given support. In reality, certain groups, such as those with mental disabilities, seem to have difficulties in getting LSS support, either because they do not apply for support, or because their applications are turned down. Thus, experienced needs are most probably more extensive than the statistics indicate (Qvarlander, Flyckt & Brusén 2004).

Instead of using provision of support as a measure of needs, social researchers turn to different indicators of the level of need. The age of the population, for example, is a conventional indicator when care of the elderly is studied, as it has been proved to be a predictor of home-help use (see, e.g. Trydegård & Thorslund 2001). When it comes to estimating the “true” prevalence of disabled people and their needs, a corresponding indicator is lacking. Still, even though we are also restricted to using available statistical data on provision of support, our ambition is to find indicators that enable us to separate the legitimate variation from the illegitimate, that is, a variation in support that is not based on needs.

There is no comprehensive data that reveals any particular differences in the municipal environment (e.g. toxic emissions) that could influence the prevalence of disability. Instead we have to content ourselves with a search for variations between municipalities in the number of potential rights-bearers that could reflect other municipal characteristics, such as the former location of institutions for disabled people, volume of earlier county council care and settlement patterns.

Residential institutions

LSS persons do not constitute a homogenous group. Three target groups have the right to LSS support: target groups 1 and 2 concern people with intellectual disabilities, while target group 3 concerns people with severe physical and mental disabilities. Several intellectually disabled people have a previous life history of care in residential institutions. Disabled people in target group 3, as well as young intellectually disabled people of today, do not have these experiences. The institutions were unevenly spread across the municipalities, and it is likely that this is reflected in the number of LSS persons today.

The closure of residential institutions for intellectually disabled people was decreed in legislation and was almost completed at the end of the 20th century. Municipalities where such institutions used to be situated, have on average a higher number of LSS persons (Socialstyrelsen 2006) and our hypothesis is thus that the volume of care is higher where there were more residential beds in a municipality. This would be a consequence of intellectually disabled people choosing to stay in these municipalities when the institutions closed down, and should thus only influence the number of older intellectually disabled living in the municipalities.

Mental hospitals

Mentally disabled people make up around one fifth of target group 3 on the national level (Socialstyrelsen 2003a). The closure of mental hospitals has not been regulated in the same way as institutions for people with intellectual disabilities. However, it is reasonable to believe that there may be a higher prevalence of mentally disabled people in municipalities where there used to be mental hospitals.

County council care

In the early 1990s, the municipalities took over the responsibility for many intellectually disabled people, as well as the actual services from the county councils, a regional body that also manages health care. The volume of care varied substantially between the county councils and we believe that these differences are reflected in the present municipal variation. Since these kinds of services cannot be directly pinpointed to specific municipalities within the county, we have to compare municipalities in one county with those belonging to another. This measure is obviously very blunt, but the underlying assumption is that although the different municipalities in the county may have taken over a larger or smaller share of the county council services, compared to municipalities in counties with relatively less disability care, municipalities in counties with a large volume of support generally received responsibility for a larger number of intellectually disabled people.

Population density

Population settlement patterns may also serve as indicators of the prevalence of needs. Sweden is characterised by a concentration of the population in the southern parts of the country, while the north is sparsely populated. Sparse population may be a result of depopulation; these variables are strongly correlated. The National Board of Health and Welfare has found that municipalities with a net depopulation in 1999–2004 have more LSS persons, a tendency that is attributed to an assumption of a lesser mobility among disabled people (Socialstyrelsen 2006). Apart from depopulation, there could be other explanations for why sparsely populated municipalities display a higher volume of LSS care. Researchers point to endogamy and consanguinity as potential causes for genetic diseases (Bittles & Egerbladh 2005; Peltonen 1996). If those diseases also have disabling consequences, a greater need for support may follow. In sum, it can be hypothesised that disabled people make up a higher proportion of the population in sparsely populated municipalities.

Propensity to apply

Demand is usually, but misleadingly, used as a synonym to need (see e.g. Lien & Pettersen 2004, Trydegård & Thorslund 2001). We want to distinguish another concept – propensity to apply for LSS support – since a variation in the propensity to apply can be influenced by other factors than need; and those factors can be more or less legitimate. As is the case for need, there are no public records on the number of applications.

Before LSS, many intellectually disabled people had support under a specific act concerning that group. It had a similar mandatory character as LSS. Those persons automatically qualified for LSS support and no new application was needed. New potential rights-bearers of the first and second target groups, as well as those persons with physical or mental disabilities that may qualify for the third target group, are expected to apply for support themselves or through a guardian (Lewin 1998).

There are both facilitating and obstructing circumstances, related to the individual or to the environment, that influence the propensity to apply for LSS support. For instance, he or she may choose to apply for support under the Social Services Act, or find it difficult on the whole to establish contact with the social services due to long travel distances. Sometimes, the circumstances in the social context are decisive: ageing parents that can no longer provide the necessary support. Some hypotheses thus concentrate on individual characteristics, while others concern environmental factors.

Human capital

The task of the welfare state is not only to “de-commodify” the individual from the market, to use Esping-Andersen's word. He or she should also be independent from family and relatives, and the state should fill the blank when this traditional support system has been mobilised into the labour market.

The general shift from utilitarianism to theories of rights as the ideology of the welfare state has also had an impact on legislation for disabled people. In the language of LSS the disabled person has the right to a good life, regardless of his or her close network. The individuals have the right to be independent, and the welfare state a responsibility to lift the care burden from the closest persons. Still, disabled people, like children and old people, are often cared for by relatives and other close persons.

General population characteristics, such as age, health, employment and education, may be hypothesised to influence the possibilities or propensity to provide care and thus also affect the application for public support even if the disabled person's individual needs have not changed. When family members get older, they may no longer be able to take care of their disabled relatives. This may also apply in the case of ill-health. Low educational level and unemployment may also have effect on caring ability. Influenced by Gary Becker and James Coleman, we use the concept of human capital as a heading for these characteristics (Becker 1964, Coleman 1988). This concept is also used in welfare research (Hyggen 2006). The above mentioned variables correlate and are brought together in a human capital index.

Non-Nordic immigrants

Research shows that disabled people born in other countries are cared for by a person from their own household to a greater extent than disabled people born in Sweden (Hörstedt et al. 1996, Szebehely, Fritzell & Lundberg 2001). Lack of information may also be a more pressing problem for this group and hinder an application. Consequently, data may show a variation in culture rather than in needs.

LASS

If a person receives other forms of support, he or she might not experience a need for LSS support. Personal assistance in the person's private home, financed through national insurance (LASS), may be given instead of residential services under LSS.

Social Services Act

As mentioned earlier, a similar but less far-reaching support than LSS can be given under the Social Services Act. Such support may also substitute a LSS measure.

Care allowance

Parents may be compensated by the national insurance for the extra care of a disabled child and therefore not feel the need for LSS support.

Local culture

It has earlier been shown that municipalities in the northern counties display a higher number of intellectually disabled LSS persons than the rest of the country (SOU 2002:103). It can be questioned whether these findings are due to a greater prevalence of genetic disabling diseases, a higher concentration of disabled people due to depopulation, or if there is a greater propensity to apply for support. A similar pattern has also been observed with regard to sickness insurance and sickness benefits, and researchers have found different regional attitudes. In the northern counties, many applicants accept the possibility of receiving benefits even if it is against the rules, for instance in problematic situations such as being harassed at work. Socially accepted norms – also held by social insurance officers – rather than actual differences in health, determine the outcome (Olsson 2006). Our hypothesis is that there is also a local culture with such liberal attitudes and norms when it comes to applying for LSS support.

Land area

Municipal geography may have an impact. In municipalities covering large land areas, services may be remote, making it difficult to apply for support even if there is a need.

Political ambitions

Needs and propensity to apply are one side of the coin. The other is the ambitions of local politicians and bureaucrats to satisfy the applicants, and the resources they can mobilise for this purpose. A simple but not uncontroversial assumption is that politicians and bureaucrats actually try to respond to the needs of citizens in accordance with the principles of representative democracy (Sharp & Maynard-Moody 1991). However, they have a choice to be more or less ambitious and choose an active or a passive strategy. Implementing reforms in an active way means that forceful policy guidelines are issued and generous resources are allocated, resulting in accessible and high quality services. Alternatively politicians may choose a more passive strategy, acting as little as possible in order to formally obey the law (Lewin 1998).

The Swedish welfare state has been built during a long political struggle between a pushing left and a resistant right concerning state intervention in the economy (Lewin 1967). In spite of this fundamental conflict, the disability reform of the 1990s was introduced in total unanimity by Parliament. However, the tension between the national and local level – the very research problem of this article – was present right from the beginning in the legislation process: The Swedish Association of Local Authorities strongly opposed the reform, their argument being that a mandatory rights legislation would inevitably curtail the self-governance of the local authorities. However, the federation was overruled (Lewin 1998).

Even though Parliament sets the national goals for disability care as for other policy areas, the municipalities, together with the county councils, are the main implementers of welfare politics: they allocate resources, determine the organisation of the bureaucracy, and issue policy guidelines. The role of the social worker is to assess the needs of the applicant and establish his or her beneficiary status in accordance with the national and local policy rules. For practical reasons, the politicians typically delegate the right to make decisions in individual cases to the social workers. This delegation may be withdrawn, and power returned to the political arena (Gustafsson 1999).

Walter Lipsky (1980) stresses the political role of bureaucrats, as do other researchers after him (see e.g. Keiser 1999). Even though politicians have the formal responsibility, it is the officials at different levels of the municipal bureaucracy that in practice implement disability politics. However, it is beyond the scope of the study to investigate the interplay between local politicians and local bureaucrats; we have not considered existing public data sufficient for that purpose.

Left-wing government

The effect of political parties on the size of welfare spending is perhaps the most interesting explanatory variable and also an ardently debated issue. Left-wing parties are assumed to spend more due to their stronger willingness to adjust perceived market failures by means of public interventions. The empirical evidence for this claim has however been mixed (see e.g. Castles & McKinlay 1979, Hanssen, Pettersen & Sandvin 2001, Imbeau, Pétry & Lamari 2001, Kittel & Obinger 2003, Lien & Pettersen 2004, Trydegård & Thorslund 2001). It takes time for an ideology to make its mark (Blais, Blake & Dion 1993). The analysis of partisanship will therefore also include the effect of length of incumbency and majority status.

Women in the council

The composition of the local governments may also be of importance for the level of ambition. Some citizen groups may be more “welfare friendly”. It has been hypothesised that women's preferences and priorities differ from those of men as a consequence of their traditional responsibility for the major part of both informal and professional care. These experiences might make them more concerned about local social policies (Lien & Pettersen 2004). Women are thus hypothesised to have a greater ambition to meet the needs of disabled people.

Separate organisation

Municipal disability care may be a separate organisation within the municipality, or organised together with care for the elderly or with the so-called individual and family service. The latter used to be the case before LSS. The National Board of Health and Welfare has found that municipalities with a separate disability organisation on average provide more LSS measures (Socialstyrelsen 2005). We know little about the reasons for the municipalities introducing a separate organisation. A separate organisation might mean a greater possibility for the politicians to control the provision of LSS support and contain costs, or quite the opposite – to increase the competence and thereby probably the volume of support. Whatever the political ambition, it can be hypothesised that a separate organisation makes the activity more visible and thus easier to ask for.

Disability organisations

Disability organisations are among the most active pressure groups when it comes to contacting local politicians (Bäck 2000) and they actively promoted LSS reform. The municipalities have a legal duty to cooperate with the disability organisations. However, the extent of cooperation varies when it comes to how involved these organisations are in the municipal decision making. It can be hypothesised that municipalities with more formal cooperation with disability organisations, due to pressure from these organisations, are positively associated with the volume of care.

Resources

It is reasonable to assume that disability care varies with the economic resources of the municipalities. Municipalities with large resources can afford to spend more on disability services.

But politicians are not only constrained by the economy, they are also in a position to master conditions by efforts such as economising, altering priorities or changing the tax rate. We would like to distinguish economic variables from the political ones just analysed as “political ambitions” (see e.g. Sharp & Maynard-Moody 1991). Thus, local government revenues or tax rate measures are excluded. This distinction is not always upheld by welfare researchers (Hanssen et al. 2001).

Compensated tax base

We regard taxable income per inhabitant, i.e. the tax base per inhabitant (Hörstedt et al. 1996), as a good indicator of potentially available resources. In Sweden, an inter-municipal income compensation transfer system levels out the differences in the tax base and is therefore also taken into account (Barrilleaux, Holbrook & Langer 2002).

Economic result

We also consider the earlier economising of the municipalities in terms of their economic result (Hanssen et al. 2001).

Long-term debt

Another way to measure earlier economising of the municipalities is to study the debt situation (Hanssen et al. 2001).

Population

If we assume that the number of LSS persons is roughly proportional to the population, it can be hypothesised that municipalities with large populations, and thereby more LSS persons, may benefit from economics of scale (Kalseth, Rattsø & Sørensen 1993). The costs of giving support to one more person would then be lower for a large municipality than for a small one. Correspondingly, a small municipality may have a relatively small tax base even if the average taxable income is large and may thus have difficulties in providing support to many needy persons.

Methods

A quantitative cross-municipal study has been undertaken to investigate the implementation of LSS in all 289 Swedish municipalities. We focus on the output of LSS care in 2002, which was the final year in the political term of office 1998–2002. Data from public statistics have been compiled and edited within the project. The data producers are mainly government authorities with responsibility to assure data quality: Statistics Sweden (SCB), National Board of Health and Welfare (NBHW), Swedish Social Insurance Agency and Swedish Disability Ombudsman. In some cases databases compiled by other researchers have been used. See Tables 1 and 2.

Table 2. Definitions of independent variables.

Variable Definition Sources
Demand
Needs
Residential institutions Number of residents in institutions for intellectually disabled in 1978 per 10,000 inhabitants in 2002. NBHW (Socialstyrelsen 1978); SCB
Mental hospitals Beds in mental hospitals in 1968 per 10,000 inhabitants in 2002. NBHW (Socialstyrelsen 1982, appendix 3); SCB
County council care Total number of beds in institutions, daily activities, and schooling per inhabitant in the counties in 1990. Landstingsförbundet 1992; SCB
Population density Average population density 1995–2002, 100 people per square kilometre. SCB
Propensity to apply
Human capital Factor score index consisting of the variables: percentage of population older than 64 years of age; ill health rate; percentage of population with a post upper secondary school education; unemployment rate. All measured in 2002. SCB; Swedish Social Insurance Agency
Non-Nordic immigrants Percentage of the population in 2002 born outside the Nordic countries. SCB
LASS Number of persons per 10,000 inhabitants with personal assistance according to LASS in 2002. Data on excel file from Swedish Social Insurance Agency; SCB
Social Services Act Number or persons per 10,000 inhabitants below the age of 65 with home-help or special housing according to the Social Services Act in 2002. NBHW (Socialstyrelsen 2003c)
Care allowance Number of children per 10,000 inhabitants whose parents received care allowance in 2002. Swedish Social Insurance Agency; SCB
Local culture The municipality belongs to one of the counties Norrbotten, Västerbotten, Västernorrland or Jämtland (dummy). SCB
Land area Municipality area (100 square kilometres) in 2002. SCB
Supply
Political ambitions
Left-wing government A government consisting of at least either the Social democrats or the Left party and no non-socialist party formed after the 1998 election (dummy). Bäck 2003
Majority Left-wing government Left-wing government with at least 50% +1 of the seats on the Municipal council KFAKTA03; SCB 1995; Bäck 2003.
Long term left-wing government Left-wing governments formed after both the 1994 and the 1998 election. KFAKTA03; SCB 1995, 1999; Kommunaktuellt1995.
Stable left-wing government Left-wing governments with the same party composition formed after both the 1994 and the 1998 elections. KFAKTA03; SCB 1995, 1999; Kommunaktuellt 1995; Bäck 2003
Women in the council Percentage of women on the Municipal council after the 1998 election. KFAKTA03
Separate organisation Presence of separate administrative organisation for disability care in 2003. Data not available for earlier years. NBHW (Socialstyrelsen 2005)
Disability organisations Additive index been based on survey questions from the Disability Ombudsman: (1) Is there a disability council in your municipality? (2) Are the disability organisations heard early in the municipal decision making process? (3) Did the disability organisations take part in preparing the municipal disability policy plan? Swedish Disability Ombudsman 2001
Resources
Compensated tax base Sum of the tax base per inhabitant and the intermunicipal income compensation (1000 SEK) in 2001. SCB
Economic result Economic result per inhabitant (1000 SEK) in 1998. SCB
Long-term debt Long term debt per inhabitant (1000 SEK) in 1998. SCB
Population Population (1000 people) in 2002. SCB

Three variables make up the dependent variable volume of LSS support: number of persons receiving LSS measures (LSS persons), total number of LSS measures (LSS measures) (it is possible to receive more than one such measure) and municipal costsof providing LSS support (LSS costs). A factor analysis of the three variables only produces one factor, indicating that all three variables measure the same underlying concept, and they have therefore been joined in a factor score index called the Volume index (Lewin, Lewin, Bäck & Westin 2008).

The Volume index was used in bivariate and multivariate regression analyses to explore the association with the independent variables. The definitions of the independent variables are presented in Table 2.

Results and Discussion

As can be seen in Table 3, simple bivariate analyses confirm most of our hypotheses and also yield statistically significant results in most cases. There are some exceptions, though, which deserve comment.

Table 3. Bivariate effects on volume of LSS support.

  Bivariate effects
Variable B (Std. error) R2
Residential institutions 0.012* (0.001) 0.189
Mental hospitals −0.001 (0.00) 0.005
County council care 0.024* (0.007) 0.035
Population density −0.047* (0.014) 0.037
Human capital 0.290* (0.056) 0.084
Non-Nordic immigrants −0.046* (0.016) 0.028
LASS 0.041* (0.012) 0.040
Social Services Act 0.008 (0.006) 0.006
Care allowance 0.011* (0.006) 0.012
Local culture 0.908* (0.155) 0.107
Land area 0.005* (0.002) 0.015
Left-wing government 0.391* (0.116) 0.038
Majority left-wing government 0.430* (0.012) 0.042
Long term left-wing government 0.484* (0.116) 0.057
Stable left-wing government 0.592* (0.151) 0.051
Women on council 0.014 (0.011) 0.006
Separate organisation 0.180 (0.126) 0.007
Disability organisations 0.271 (0.240) 0.005
Compensated tax base −0.018* (0.005) 0.042
Economic result 0.028 (0.031) 0.003
Long-term debt −0.002 (0.007) 0.000
Population −0.001 (0.001) 0.003
*Significant at the 0.05 level (one sided test).

Support in accordance with the Social Services Act has no negative effect on volume of LSS care, but rather there exists a weak, though not statistically significant, positive relationship. Personal assistance and benefits from the national insurance for the extra care for a disabled child show a significantly positive association with volume of LSS support. Contrary to what might be expected (cf. Daley et al. 2006, Trydegård & Thorslund 2001), different forms of support do not serve as corresponding vessels. Rather, the positive relationships suggest that they complement each other, and that the same individual often receives support through different channels.

A Model for Variation in LSS Support

All variables yielding statistically significant effects on volume in the bivariate analyses are included in the multivariate analysis. The support measures supposedly serving as substitutes for LSS support (LASS, Social Services Act and care allowance), proved instead to act as complementary support, and are thus excluded from the multivariate analyses.

The variable concerning the takeover of care from the county councils ceases to be statistically significant in the multivariate analysis. A reasonable explanation is that this has to do with the variable local culture. Some of the counties with the most services for disabled people in 1990 are situated in the northern part of Sweden. However, the relationship is not perfect (Pearson's r=0.273), and in fact the effect of local culture is the stronger one. The non-Nordic inhabitants variable also proves insignificant, together with the compensated tax base. It should be noted that tax base is not the same as tax revenues. Just because a municipality has a population with a comparatively high income, it does not necessarily mean that it will collect more taxes. However, the correlation is quite high (r=0.775).

The variables that remain siginificant constitute the final multivariate model and can be seen in Table 4.

Table 4. Multivariate effects on volume of LSS support.

  B (Std. error) t Adj. R2
(Constant) −0.257 (0.067) –3.833  
Residential institution 0.011 (0.001) 8.245  
Population density −0.022 (0.013) –1.692  
Human capital 0.130 (0.061) 2.121  
Land area −0.006 (0.003) –2.183  
Local culture 0.881 (0.181) 4.869  
Stable left-wing government 0.352 (0.130) 2.698  
Model summary     0.333

A higher volume of LSS care is found in municipalities that previously used to house a residential institution for intellectually disabled people, have a low population density, a lower accumulation of human capital (older population, fewer highly educated, more unemployed and more sick people), a small land area, are situated in one of the four most northern counties, and have had a stable left-wing government during the past two terms of office.

The significant effect of residential institutions in the bivariate analysis strongly supports the notion that prevalence of needs affects the volume of provided LSS support. One more resident in 1978 per 10,000 inhabitants in 2002 is associated with 0.2 more LSS persons and 0.4 more LSS measures per 10,000 inhabitants, while the costs are nine SEK higher per inhabitant. The hypothesised mechanism is the tendency for intellectually disabled people to stay in the municipality where the institution was located. If so, the municipal variation will be reduced in time as former residents age and die. A great majority of new LSS persons have intellectual disabilities and usually already receive support as children or young adults. They have no institutional experience and the variation in need in the future will not depend on an earlier existence of residential institutions. A comparison over the years shows that the variation (the standard deviation as percentage of the mean) in the number of intellectually disabled has actually decreased between 1998 and 2002.

We have specifically looked at the effect of residential institutions on the three target groups that make up the variable LSS persons. It is interesting to note an unanticipated effect on disabled people with other disabilities than intellectual. Perhaps having an institution in the municipality has contributed to a more positive attitude towards disabled people along with better staffing, which has led to a higher municipal ambition to provide support. This would then be an effect on the supply side –“path dependence” –a concept used in research to grasp the impact of history and tradition on politics today (Thelen 1999). When care was transferred from the county councils to the municipalities in the early 1990s, staff at different levels also joined. It is reasonable to assume that the higher level bureaucrats in particular have been able to inspire politicians to pursue an active strategy and develop disability care in the municipalities.

Sparsely populated municipalities tend to have a greater LSS volume, which we believe is due to greater need. There may also, however, be a question of propensity to apply. In sparsely populated municipalities the employment situation may be difficult. Having a close person entitled to LSS support may mean a possibility for a family member to work as a personal assistant.

Municipalities with low human capital also have a greater LSS volume. Human capital is measured as an index, where changes are expressed in standard deviations. The variable “old age” is chosen to illustrate this determinant. One more person above the age of 64 per 10,000 inhabitants is associated with 1.2 more LSS persons and 2.0 more LSS measures per 10,000, while the LSS costs are 33.1 SEK higher per inhabitant. We cannot say whether individual needs are greater in these municipalities or if there is an inability of close persons to provide care. Further, even if LSS is strictly aimed at the individual's needs and his or her independence, some measures are specifically intended to make daily life easier for close persons. Need can thus be said also to encompass close persons, which means a broadened individual-need concept.

Research on sickness benefits (Olsson 2006) shows that liberal attitudes towards illegitimate use of this system among the population, staff management at working places, physicians and social workers at regional social insurance offices are more common in regions where many collect sickness benefits. These are also regions where human capital tends to be low. Maybe such attitudes, relating both to the propensity to apply for LSS support and to ambitions, also extend to disability care among the disabled people themselves and their close persons as well as social workers in the social services. The two former groups would then be more demanding and the latter group more resilient in their supply.

In the bivariate analysis, land area had a positive impact on LSS volume, mostly through its effect on LSS costs: a larger area implied higher costs. However, in the multivariate analysis, the effect is reversed and thus in line with our hypothesis that long distances may lessen the propensity to apply for support. The changed effect may be due to the fact that the region variable is included; many large area municipalities are situated in the northern part of Sweden where the LSS volume is larger (r=0.634).

Municipalities in northern counties tend to have a greater volume of LSS care. They support 14.4 more persons per 10,000 inhabitants, give 29.7 more measures per 10,000 inhabitants and spend 676.3 SEK more per inhabitant on LSS care than other municipalities. This cannot be explained with needs, measured as earlier location of institutions and a sparse population. Nor can it be explained with supply factors, such as the presence of a stable left-wing government and resources. As the effect is also independent of the degree of human capital, our conclusion is that the effect is due to a local culture. This result is in line with the above mentioned research on the sickness benefits system: liberal attitudes are more common in the northern part of Sweden (Olsson 2006). This tendency to depend on public support as well as a generosity from the social workers seems also to be the case in the disability area.

As mentioned, the effect of left-wing governments is contingent on the length of incumbency. We do not find any significant effect of left-wing government if we only study the party composition of the cabinet in power 1998–2002. This holds for both minority and majority left-wing governments in 1998–2002. However if we narrow it down to left-wing governments who have stayed in power for two consecutive terms of office (1994–2002), the presence of a left-wing government is positively and significantly associated with a larger volume of LSS care, thus indicating a higher ambition. Left-wing municipalities tend to support 9.2 more persons per 10,000 inhabitants, give 19.3 more measures per 10,000 inhabitants and spend 449.9 SEK more on LSS care per inhabitant than other municipalities. It takes time before a new regime can make its own mark.

Even if we do not claim to have covered all possible explanations for the variation in LSS care, we believe that we have identified some crucial contenders. However, public data are not comprehensive. Lack of data prevents us from studying the effect on the propensity to apply for LSS support by such factors as special schools and rehabilitation services as well as the effect on supply by organisational changes. Nor does our quantitative approach permit a deeper study of the explanatory variables. The role of real enthusiasts among politicians and bureaucrats acting as driving forces when implementing the disability policy might be studied with qualitative methods such as case studies. Case studies could also elucidate the extent to which disability organisations appear as active lobbyists and not only as passive recipients of information.

Need, propensity to apply, and in the long run, political ideology are thus represented in our final explanatory model of the variation in the provision of LSS support between Swedish municipalities (Figure 2). Money is not the cause of the variation. It should be noted, however, that the explained variance of our model (adj. R2) is 33.3%, and consequently there is still variation left to be explained.

Figure 2. 

An explanatory model for the effect of municipal characteristics on the provision of LSS support.

Conclusions

Our results support a need-responsiveness model of welfare support; social problems engender local government responses. Reality is obviously more complicated than those critics maintain who talk about “the arbitrariness of the local political majority” or “the municipalities of shame”. Municipalities with many severely disabled people provide much LSS support and have high LSS costs. It is natural that municipalities where there have been no residential institutions have fewer LSS persons. It is also beyond the control of the municipal politicians as to whether disabled people remain in sparsely populated municipalities. Of course, such variation is legitimate.

However, there are some results that could be a political concern. The lower propensity to apply for support in municipalities with large distances is one that may signal an accessibility problem. The effect of low human capital is another. Municipalities may take on a greater responsibility if there is a low capacity of close persons to provide care. If this is an effect of attitudes that are independent of need, the equal treatment norm may be threatened and we have a legitimacy problem. The same can also be said about effect of local culture.

Even if the LSS reform was introduced in total unanimity by the Swedish Parliament, The National Organisation of Local Governments strongly opposed the LSS reform, and the criticism has not ceased. In the municipalities, utopia meets reality and for the local politicians the idea of self-governance is perhaps the most important political value of all. Old people and severely disabled people seemingly have the same difficulties in their daily lives, but the former group are given support according to the more flexible and less costly Social Services Act, while severely disabled people have special rights through LSS. Hence, although national politicians have emphasized that only needs should be decisive for the welfare responses by the local politicians, other regional factors seemingly play a role. There is no doubt that there actually is a conflict between local self-government and national welfare politics and that the former has an advantage in the view of municipality politicians. The principle of spatial equity is violated. Although it should not, it still matters where you live.