The purpose of the municipal equalisation system is to create equal economic conditions for all municipalities and county councils in Sweden so as to enable them to provide all their inhabitants with the same standard of public service. The municipal cost equalisation system is intended to equalise structural cost differences between municipalities and between county councils. Differences which are the result of the service level chosen, efficiency, service quality and levels of user fees are not equalised. One fundamental principle is that municipalities and county councils should not be able to influence grants and fees in the cost equalisation system.

The cost equalisation system for municipalities consists of ten different sub-models for various activities. One of these sub-models is the individual and family care (IFC) model. Statskontoret (The Swedish Agency for Public Management) has been commissioned by the government to review the IFC model. In line with the government's commission, Statskontoret will investigate whether there are alternatives to influenceable variables and, where possible, will propose that such variables be included in the IFC model instead

In all sub-models within the cost equalisation system, a "standard cost" in Swedish kronor per inhabitant is calculated. The standard cost reflects the theoretical calculated cost which the municipalities would have with average service levels, efficiency, service quality and levels of user fees, with consideration instead being given to various external structural conditions. As with the current IFC model, we propose a comprehensive model for calculating the standard cost of IFC.

In the proposed IFC model, four explanatory variables are proposed:

- the proportion of inhabitants who are unemployed and who do not receive benefits,
- the proportion of persons born in Sweden (aged 20 to 40) who have low levels of education,
- the square root of the size of the urban population in the municipality and
- child poverty (defined as the proportion of children living at an economic standard below 60 per cent of the median income, adjusted for the number of dependents.)

The proposed changes to the IFC model are as follows:

- The variable
*proportion of the municipality's population receiving means-tested financial support for a period of more than six months*should be removed from the IFC model. - The variable
*proportion of the municipal population living in a block of flats built 1965–1975*should be removed from the IFC model. - Child poverty, defined as the
*proportion of children living in an economic standard below 60 per cent of the median income*, is proposed for inclusion in the IFC model. The basis for the calculation of child poverty is adjusted for cross-border commuting. Income from means-tested financial support is not included when calculating the variable for child poverty.

In its analysis, Statskontoret has examined a large number of factors that are likely to be decisive for the structurally determined need for IFC activities in the municipalities. One aim has been to reduce the risk that municipalities may be able to affect the outcome of the municipal cost equalisation system. The variable "proportion of the municipality's population receiving financial support for a period of more than six months" accounts for the bulk of the municipalities' expenses for means-tested financial support, which leads to a sort of circular connection in the IFC model. The risk also exists that this variable can be influenced by municipal decisions, which in turn may then also affect what level of compensation will be received from the cost equalisation system. Our starting point was thus that the variable should be removed. Without the variable, the other explanatory variables have more mutually distinct, significant and powerful effects in the IFC model.

In our analysis, we have also come to the conclusion that the variable "proportion of the municipal population living in a block of flats built 1965-1975" should be removed from the IFC model, since it proved to be unstable and also underwent variations from year to year that are difficult to explain. By excluding this variable, the IFC model becomes more transparent and operations at Statistics Sweden (SCB) are facilitated.

Poverty can be expected to have a significant impact on the number of cases where there is a need of individual and family care. After examining various alternatives, we have arrived at the conclusion that the IFC model should include the variable "the proportion of children living at an economic standard below 60 per cent of the median income, adjusted for the number of dependents," instead of the two variables proposed for removal. The underlying data is adjusted for cross-border commuting by removing commuters from the income-earning population entirely, since there is no information in the SCB's income statistics on these commuters' income from salaries earned in other countries. With the purpose of avoiding potential effects on the variable from the municipalities' side, income from means-tested financial support is not included in disposable incomes when calculating the variable for child poverty.

In line with our proposal, the IFC model's statistically calculated coefficient of determination is reduced from 67 to 57 per cent. The reduced coefficient of determination is partly due to the fact that the variable "proportion of the municipality's population receiving financial support for a period of more than six months" has been removed; i.e., the variable that leads to a circular connection in the IFC model is no longer included.

Our assessment is that the advantages of the changes outweigh the disadvantages, since the changes reduce municipalities' opportunities to influence compensation received from the cost equalisation system. Furthermore, the administration of the IFC model is simplified at SCB.

The cost equalisation system is a system for redistributing economic resources among municipalities and among county councils. The IFC model is included as a sub-model for cost equalisation in the municipalities. Changes to the IFC model therefore affect municipalities' compensation from the cost equalisation system.

With the current IFC model, the funds redistributed between municipalities in 2014 amount to a total of SEK 4.2 billion. The proposed changes, calculated in the same year, would increase the annual redistribution to SEK 4.5 billion, i.e., they would lead to somewhat higher redistribution among the municipalities. Most municipalities will experience small changes in compensation but for some municipalities the differences will be greater.

The differences between the models in terms of the outcome for the municipalities depend on several factors. There is however a pattern in the characteristics for the municipalities undergoing the biggest changes in the outcome from our proposed IFC model. The proposed changes to the model will serve to improve the outcomes of cost equalisation in municipalities with a relatively high number of unemployed people not receiving benefits and a relatively high degree of child poverty. In municipalities where a relatively high proportion of the population receives means-tested financial support for a period of more than six months, the economic outcome tends to deteriorate with the proposed IFC model.

Both the current and the proposed IFC models are based on a regression equation with a number of explanatory variables. The strengths of the connections between the variables and IFC costs can be expected to change over time. According to Statskontoret's assessment, the regression equation should therefore be updated on a regular basis, the suggested interval being three years.

Poverty is currently calculated based on family incomes. The register data used by SCB for the calculation can partly change where household composition is concerned when the new apartment register is introduced. The proposed parameters of the IFC model, and therefore also the parameter (coefficient) for child poverty, have been estimated with current household concepts. The IFC model therefore needs to be updated in a few years' time, at which point it will be based on the new register that SCB will use. This can be fully achieved in by the equalisation year for 2018.

The proposed IFC model can be implemented in the municipal cost¬ equalisation system starting in the equalisation year of 2016.