Resumé

The two conservative parties in Norway, the old Conservative Party and the younger Progressive increased their electoral support from 23 percent in 1975 to 34 percent in the election of 1987. The electoral mobilization by these conservative parties was mainly based on an ideologically inspired rhetoric of reduced government spending and corresponding tax reductions. With the benefit of hindsight, it may be argued that these were policies to which the two parties did not live up. Whatever their political strength in the local arena, they were unable to reduce local government income taxes. But when analysing other fiscal strategies available to local governments we do find differences related to political strength, even if the findings are not always as expected. Municipalities dominated by the conservatives are run on the principle of families paying the actual costs of having children in public day-care institutions. But so are socialist-dominated municipalities. Conservative and socialist municipalities to subsidize fees for home-help services for the elderly. The main fiscal source of the conservatives is fees and charges on technical services used by every household and paid according to costs. The watershed between socialist and conservative parties appears as we analyse the use of property tax - a tax used much more frequently by socialist than by conservative and centrist parties.

Introduction

The local elections in 1979 introduced a new trend in electoral politics in Norway: support for the Conservative Party reached its zenith. Four and eight years later the Progressive Party, with its call for tax reductions and cutbacks in public spending, contributed towards a new conservative electoral victory, and the second conservative mobilization. Both these electoral surges led to major changes in the party composition of local governments.

Studies of the impact on public policy of political versus economic and social variables have a long tradition, inspired as they are by the question: Does politics really matter? (Dawson & Robinson 1963; Boaden & Alf ord 1969; Fry & Winters 1970; Castles & Borg 1974; Hansen & Kjellberg 1976; Castles & McKinlay 1979; Sharp 1981; Hansen 1981; Castles 1982; Boyne 1984). Few circumstances are more conducive to this inquiry than the

politics of local government in Norway between the elections of 1979 and 1987, since the political demands of the conservative winners of all three elections are so easy to identify. The Conservative Party has stubbornly maintained its opposition against tax increases and demanded public spendingto cut in areas where other parties have opted for expansion. Even more forcefully the Progressive Party has emphasized the need for tax reductions, making this issue into its very raison-d'étre. Since most studies of party impact have been concerned with spending or other performance variables, this gives us ample opportunity to turn to the fiscal side of politics.

During the period to be studied, the Conservative Party has on several occasions been part of national coalition governments, but perpetually dependent on centrist parties opposing cuts in spending as well as large tax reductions. Within the same period, the Progressive Party was excluded from national government influence by the same centrist parties. But, in the local arena, coalitions between the Conservative and the Progressive Parties, have become quite frequent, with the Oslo city government as the most prominent example. The question is: Does it matter? Or more specifically, does the strength of the conservative parties influence fiscal strategies of the local government? Does it induce local governments to increase fees and prices on the services provided by government in order to promote policies of tax reduction?

Local governments in Norway acquire their major revenues from a combination of direct income taxes, and government transfers. Taxation is by far the most important fiscal source for the larger municipalities, while transfers provide the income basis for the smaller. In 1990, an average of 47 percent of local government revenues was accumulated through income taxes. Transfers from national government accounted for 34 percent (NOS, 1992: C 32). Neither of those two income sources is controlled by local government since the maximum tax level is decided by parliament and transfer levels are set by national government. The tax rate is set to a maximum of 13.5 percent of taxable income, and for the last thirty years all municipalities have used the maximum tax rate.

Nevertheless, local government can choose between additional income options, and these opportunities are the focus of our study. The options are of two kinds: (1) property taxes, and (2) fees for various services provided by local government. Even if Sørensen and Underdal argue that property taxes and fees are of insignificant importance for the economy of the municipalities (Sørensen & Underdal 1993), these sources accounted for more than 15 percent of the total revenue in 1990 (Statistical Yearbook 1994, Table 447). It is readily apparent that the importance of these income sources has been growing rapidly since only 7 percent of the total revenue came from property taxes and fees in 1980 (5t.me1d.26:1983-84). And even

though taxes and fees provide a very small proportion of local government revenues, to the local politicians, fees and property taxes are the marginal income sources which reduce budget deficits. But most notably, for the local political debate, increases or decreases in property taxes and fees on public services are of symbolic significance, indicators of party cleavages and for the taxpayers, the level of property taxes and fees to be paid obviously has a political message attached to it. As a result, we have four dependent variables of interest:

1. The amount of the fees paid by parents for having children in daycare

2. The amount paid by elderly or disabled persons for home-help
services.

3. The fees paid by every household for technical services, such as
water supply, common sewer and garbage services.

4. The use of property tax.

The sovereignty of local government to decide on fees and property taxes gives us ample opportunity to investigate whether communities dominated by the conservatives differ from other communities in their approach to fiscal policy.

The Conservative Mobilization: Two Different Surges

Even though the Conservative Party expanded its proportion of the votes in the middle of the 19705, it was not until the local election in 1979 that the party experienced a new breakthrough, increasing its support from 23.3 percent in 1975 to 29.2 percent in 1979. In terms of electoral strength, the Conservative Party had established itself as the main contender of the Labour Party on the national level: it had become the majority party in several urban areas, and it was by far the largest of the non-socialist parties. However, this conservative mobilization had no impact on the support for the Progressive Party (0.8 percent and 1.9 percent in 1975 and 1979 respectively).

This changed at the next local election in 1983. The total support for the two conservative parties remained at exactly the same level, 31.1 percent in 1979 and 31.2 percent in 1983. But the Progressive Party was now approaching the size of the centrist parties, between 6 and 7 percent of the votes, and gained as much as the Conservative Party lost. This provided the political basis for the first conservative alliances at the local level.

The situation improved even more in favour of the Progressive Party at
the election in 1987. The Conservative Party continued to lose ground (23.3

percent), but the Progressive Party secured the best local election result ever for the two conservative parties combined by gaining 10.4 percent of the votes, and thus became the third largest party in the nation. By now the stage was set for new alliances with even more political influence for the Progressives. This development encourages the intriguing question of whether there are differences in the policies pursued during the first conservative surge, when the Conservative Party alone increased its support, and the second wave of conservative mobilization, when conservativepolicies on support from the representatives of the Progressive Party.

Expected Differences Related to Party Strength

According to conservative political theory and rhetoric, we expect the conservatives to avoid taxes on property. Both parties have fought, at the national as well as the local level, against taxes on private homes. However, we expect both parties to be in favour of increasing the fees and charges for services provided by local government. The logic behind increased service charges is the simple idea that prices should be set at an amount where the user pays for the larger part of the actual cost of services supplied. This is part and parcel of the conservative ideology of market-stimulated service production designed to provide the fiscal foundation for tax cuts and to prevent non-users from paying for other people's consumption of services. However, such policies contradict the socialist ideology of needbasedservices equal access regardless of personal income. It is by no means unusual to subsidize various services and set fees at a price lower than the actual costs. The socialists have basically approached general

taxation as the very foundation of the policy of solidarity to which they claim allegiance: those most in need of different welfare state services, such as parents with children in need of day care or pre-school care, or the elderly in need of home-help services, may not be among the most affluent of the citizens in the municipality. Through taxation everybody shares those burdens.

This implies more or less hidden transfers and income redistribution. Though not entirely unacceptable to the conservative parties, this policy has not been warmly embraced by them. The Progressive Party did, however, emphasize the needs of the elderly throughout the most recent electoral campaign. If this is not sheer rhetoric, we should expect the conservatives to be more reluctant to impose fees on home help- services for older people. This leaves duties on child care (basically used by the younger generation) and fees on utility services (used by every household) as their remaining tax targets.

As for the property taxes, they have traditionally served as a major issue
of contention between the parties of the left and the parties of the right in
Norwegian local government.

Our basic questions are the following:

• Does conservative strength in local government stimulate fiscal policies
tax reduction and increased fees and charges in accordance
with their respective programmes?

• Did the relationship between conservative strength and fiscal strategy become more pronounced as their majorities increased during the 1980s and as the Progressive Party became more of a substantial coalition companion?

• Do conservatives differ from other parties in the sense that they alone
are implementing these policies, or are other parties or coalitions
following the same general fiscal strategies?

• Do conservative local governments differ from others with respect to the fiscal instruments they use, or does the difference between conservative parties and centrist/socialist parties boil down to a matter of the level of taxation?

All our estimates are in fixed 1990 currency values.

Non-Political Explanations

Even if there are correlations between conservative strength, fiscal instruments
level of taxation, these relationships might be spurious and
explained by factors other than strength of parties and party politics.

Both conservative parties have their strongholds in urban areas. There -

fore control for population size of each municipality and density of populationis Urbanization and population size may also represent large-scale advantages which reduce the need for extensive fees and increased taxes.

On the other hand, there is a long rural tradition advocating strong and autonomous local authority for the explicit purpose of preventing the national government from dipping into the "purse" of local government. A maximum local government and a minimum of local spending and taxes have been the political strategy of rural municipalities since well back into the last century. These rural and mostly small municipalities have also been the strongholds of the centrist parties: the Agrarians, Liberals and the Christian People's Party.

The fiscal affluence of the municipality influences the local government's ability to subsidize services and avoid heavy tax burdens on its population. And again, the strongholds of conservative parties have been the richer municipalities.

Measuring municipality affluence in Norway is not entirely unproblematic. we talking about economic affluence before or after transfers? In the 1980s the poorest municipalities prior to transfers did in fact turn into the most affluent ones after transfers, at least in terms of local budget size per capita. With the introduction of the new income system in 1986, this tendency became even more pronounced in the latter part of the 1980s. We therefore introduce two concepts and measurements of municipality affluence: (1) pre-transfer as measured by the average income of the municipality's citizens; and (2) post-transfer as measured by the total disposable local government budget per inhabitant.

There are major variations among Norwegian municipalities regarding the kind of services given priority. The political ambition to keep services at certain levels varies, and with varying volumes of service the fiscal needs of the municipalities change. A control for the volume of services or activity level regarding various services of different municipalities is therefore appropriate.

Finally, some have argued that the number of potential "clients" - be they children or the elderly - might influence the fiscal strategies of the municipalities. The larger the number of clients as a proportion of population the larger the tax base in terms of fees. The temptation to increase municipality revenues by using service fees is therefore largest where the number of people using these services is greatest. However, increased fees may induce service users to vote against the fee-increasing parties. Regarding fees on day care, we also control for the proportion of private day-care institutions in each municipality.

Model Specification

The regression equations are specified as presented in Tables 1 and 2 above.

Note on Analytical Strategies

Our analysis is based on data from every Norwegian municipality. Our

principal analytical technique is ordinary least squares (OLS) regression, since every variable is measured at the interval level. However, some basic problems concerning regression analysis have to be taken into consideration.

Linear relationships?

We have grouped the parties into three categories:

(1) The Conservative Party and Progressive Party are amalgamated as
the prime group of interest for our investigation.

(2) The Left Socialist and Labour Parties are considered parties of the
opposition.

These two groups are inserted into the regressions equations.

(3) Which leaves the Liberal Party, the Agrarian Party and Christian People's Party, together with independents as the centrist party group, omitted in the regression and used as our statistical reference or comparison category.

Some would probably argue that linear regression is not appropriate since the support of the party groups varies from 0 to 100 percent (See Appendix 1). This argument rests on the assumption that parties are in the position to pursue their own goals and implement their own policies, if and only if their electoral strength approaches the 50 percent-mark; and as a result we are faced with non-linear relationships and even with the possibility of relationships plagued by interactions with other independent variables. Without ruling out these possibilities, we believe that the specific structure of Norwegian municipalities as political institutions provides us with a good case for retaining the linear model.

An electoral majority and even a parliamentary majority within the municipal council will not result in the dominance of the executive cabinet (formannsskapet) of the municipality by one single party, since this decisive body in Norwegian local government is elected by proportional representation. possibility of a one-party dominance in local government is therefore negligible; the opposition will always be represented within the executive. In 1975 the two conservative parties had an absolute majority, 50 percent or more of the votes, in only two municipalities. This number increased to five in 1979 and to six in 1983 as part of the first wave of conservative mobilization. The last election on which we focus -the election of 1987 - resulted in a conservative majority in a total of eight municipalities. We consider these numbers too small for a meaningful analysis of the impact of majority circumstances.

By and large there is therefore ample reason to expect the political

influence of parties to increase quite linearly with their electoral and
parliamentary strength. In faet most of the time as many as four parties
are represented in the executive cabinet.

For the analysis of day-care fees, only those municipalities that have such public facilities will be included. All municipalities have home-help services for the elderly and utility services, and they are free to introduce property taxes.

Multicollinearity

The simplest and preferred strategy using regression techniques would be to include every independent variable in one single regression equation. But then we run into the problem of multicollinearity. Primarily, size and degree of urbanization correlate with average citizen income to such an extent that it is inadvisable to use both variables in the same regression, considering the fact that the number of observations is only around 430. The same complication exists between budget affluence in general and the activity level or volume of services provided. Our strategy is therefore to analyse the impact of party strength under two different circumstances: First, we control for the size and urbanization characteristics of the municipality with the activity level or volume of services. Secondly, we control for the economic affluence of the municipality, measured by average pre-tax income and the total per capita size of the local government budget.

Some would assume that multicollinearity would exist between socialist and conservative strength, with rather high negative correlations. However, both groups of parties compete heavily with the centrist parties, leaving the inter-correlations between socialist and conservative strength at a rather moderate level; respectively 1975 (r = -0.06), 1979 (r = -0.28), 1983 (r = - 0.32) and 1987 (r = -0.30).

"Outliers"

Some municipalities will be obvious "outliers" due to size. Oslo in particular is a deviant case with close to half a million inhabitants, and governed by parliamentary principles; a one party majority in the municipality assembly produces a one-party executive cabinet. Oslo is therefore excluded from the analysis. Other municipalities are classified as the municipal "nobility" due to extraordinarily high levels of income. These are small municipalities with large industrial factories or hydroelectric plants, which account for large revenues, not generated by local political decision-making processes. These municipalities are excluded along with municipalities which have been amalgamated during the period analysed.

Time-lags

An analysis of the impact of party strength presumes a "time-lag" between the election and the time when effects on policy are felt. We have used the maximum time-lag of four years between election and measurement of fiscal policies, giving party strength the greatest possibility of making an impact. However, due to constraints of data availability some adjustments have had to be made. The election of 1987 should have been matched by data for 1991, but the most recent data available to the authors pertained to 1990.

The relationship between elections and fiscal years is as follows:

The budget for 1979 was made in the autumn of 1978 and is the second last budget for which the municipal council of 1975 is responsible, and this budget is neither influenced by the upcoming election campaign nor by the election outcome in 1979.

The Expansion of Day-Care Facilities

Ever since the adoption of the Child-Care Act of 1975, Parliament and various governments - socialist as well as non-socialist - have given priority to the expansion of child-care institutions. The national government established earmarked grant to stimulate expansion of child-care institutions, and the 1980s was marked by a considerable expansion. In 1975 there were 884 institutions, in 1979 2295 and in 1990 the number of child-care institutions had reached 4649. (These and figures elsewhere are retrieved from our data source: The Municipality Data Bank of the Norwegian Social Science Data Services).

In 1975 a total of 30,479 children were enrolled in day-care institutions. This constituted 7 percent of all children under the age of seven. In 1979 the number had increased to 78,189, covering 21 percent of all children under seven, and during the 1980s the coverage rate increased continuously and by 1990 a total of 139,350 or 36 percent of the children were enrolled (NOS 1990 nr. B-975).

Child-care institutions include both public and private or, rather, semipublicinstitutions. latter are child-care institutions that are privately owned and run, but dependent on national government subsidies that are administered and appropriated by the local government. The private

institutions are run by private firms for their employees, voluntary organizations,congregations, and are most often run on a non-profit basis. Of a total of 4649 institutions in 1990, 2737 were public, and 1912 were private, which results in a privatization rate at about 41 percent. Ten years earlier - in 1979 - 1370 were public, and 1184 were private of a total of 2554, corresponding to a privatization rate of 53 percent. The private institutions went through a phase of expansion in the 19705, while the public institutions account for the lion's share of the growth in the 1980s.

In this article we focus on day-care facilities owned and managed by the municipalities, where the size of the fees is a matter of political decision. Public day-care institutions are financed from three different sources: earmarked transfers from the national government, subsidies from the local governments and fees from the parents of the children. According to the municipal accounts, total expenditures for care institutions were 4.4 billion Norwegian crowns (NOK) in 1990, corresponding to NOK 11,800 per child under seven years of age, varying across municipalities within a range from NOK 185 to NOK 42,600 per child under seven years. The last category consists of small municipalities offering a day-care place for every child. Nine years earlier - in 1979 - total expenditure was only 0.67 billion, or NOK 2200 per child under seven, in the currency value of 1990.

Our dependent variable is the fee paid by parents for each child enrolled in a public child-care institution in different municipalities. Fees are either paid for full or half-time attendance. We assume that the proportion of parents using the two options is not dependent on political strength, and rather randomly distributed across municipalities.

The average fee paid by parents increased from NOK 3023 annually in 1979 to NOK 7001 in 1990, but the size of the fees varied substantially from one municipality to another. In 1979 the fees varied from a minimum of NOK 0 to a maximum of NOK 8837 per child enrolled; in 1990 the fees varied from NOK 0 to NOK 15,790. There is a clear increase in the variation (St.dev.) from 1979 to 1990, demonstrating that the fees paid by parents became less standardized during the 1980s.

Fees for Child Care: Politics Matter, But Not as Expected

When analysing the fees paid by parents for children in day care, we start by considering the impact of party strength, controlling for the percentage of the population living in urban areas, the size of the municipality, the percentage of private day-care institutions, the amount of money each local government spends on care institutions and the proportion of children under the age of seven in each municipality (Table 3).

The most decisive impact is of a "push and ambition" variety: when the
volume of day-care services expands, the fees also tend to increase. An

expansion of the activity by NOK 1000 per child under the age of seven increased the fees by NOK 61 in 1979, hit an all-time high of NOK 246 in 1983, but only in order to drop to NOK 161 in 1987 and to NOK 66 in 1990. This is a strong indication that the most ambitious local governments

regarding day-care activity apply fees as a part of the necessary fiscal foundation for the volume of expansion. The volume of spending is clearly related to political strength. As expected, socialist municipalities spend more (r = 0.21), while there is a negative relationship to conservative strength (r = -0.13).

This does not necessarily imply that the conservative-dominated municipalities a lower coverage of children in day-care facilities than do other municipalities. It is mainly a by-product of the fact that conservativedominated have a more privatized day-care system than socialist- as well as centrist-dominated municipalities - and in order to evaluate the coverage of day-care services we have to take both public and private day-care services into account. The strong growth in the provision of public day-care facilities that took place in the 1980s did, however, result in greater differences in day-care coverage between conservativedominated on the one hand and socialist- or centrist-dominated on the other. While we could not see any differences between the three groups of municipalities in 1980, there is a fairly strong negative relationship between conservative strength and day-care coverage in 1990 (r = - 0.33) as opposed to the positive relationship between centrist strength and day-care coverage (r = 0.15). Considering socialist strength, there is no relationship at all.

However, the difference in coverage rate is founded on different political approaches; while the centrist and conservative local governments rely on markets and private institutions, socialists rely on government institutions. This tendency of privatization in conservative municipalities becomes even stronger at the end than in the beginning of the decade; in 1990 the correlation between the number of private institutions and socialist strength was r= -0.16, conservative strength r = 0.28, and for the centrists no relationship.

Table 3 shows that as the private/public ratio increases, the familypaid in public day-care institutions decreases. This indicates that in municipalities with a high number of private day-care institutions, the care system tends to become dualistic; private institutions for the well-situated families who are able to pay, and public and inexpensive institutions for the families who are economically not so well situated.

Urban municipalities, where the conservative have their strongholds, are also disposed towards increasing fees, while size of population tends to decrease fees other things being equal - even though this effect is not significant until 1990. Smaller municipalities, which account for the majority of the municipalities, provide the battleground for the struggle between socialists and centrist parties.

The impact of party strength is obvious, but - as we expected - it does
not discriminate much between socialist and conservative parties. On the

contrary, socialist party strength is about as likely as conservative party strength to result in increasing fees for day-care facilities, and at very much the same rate. When the conservative mobilization - which was to make its impact felt until the fiscal year of 1983 - started in 1979, it was accompanied by an increase in family-paid fees in the range of NOK 20-40for percentage point gained by the conservative or, for that matter, the socialist group. The fees charged by the local governments in our reference category - municipalities dominated by centrist and non-partisan forces - are clearly lower than those imposed by socialist- or conservativedominatedlocal The socialists by no means seem to follow a fiscal strategy of subsidizing parents in need of day care. As in municipalitiesof strength: parents have to pay. The two kinds of communities seem to apply basically similar payment policies.

The proportion of children under the age of seven also seems to have a substantial impact; as the proportion increases by one percentage point, the fees increase by some NOK 200. This finding should be interpreted with caution since the variation is small. In 1990 this variable fluctuates between a minimum of 6 and a maximum of 14 percent. However, there is another story to be told concerning the number of children. The conservatives their strongholds in communities that are urban and with a large proportion of children under seven years of age (r = 0.25) - a type of community where the socialists do extremely poorly (r = -0.51). This suggests that the socialists run expensive day-care facilities, producing a high public activity level in municipalities with a rather moderate proportion of children. The conservatives also run expensive public day-care facilities, but at a moderate or low activity level where the proportion of children is at its peak. But as previously revealed, they have privatization of day-care facilities as part and parcel of their coverage policy.

The Impact of Affluence

When analysing the impact of local affluence, it is well worth keeping at least two arguments in mind: (1) In general, the expansion of day-care facilities seems to be dependent on an extensive use of fees, which should increase fees as activities expand. On the other hand, (2) municipalities dependent on transfers for their fiscal affluence must spend their money to display their need for next year's transfers, and may therefore be inclined to build and subsidize. Both tendencies are substantiated by our data (Table 4). In 1979 the transfer-dependent municipalities, which as mentioned have the largest budgets per capita, had the lowest fees. But from 1983 to 1990, during the most extensive expansion period of public child-care institutions, the larger the budgets the higher the fees. This indicates that fees rank among the general fiscal sources used for expanding activity levels, and

more so as local government could not exploit earmarked grants, but
choose among various political priorities on the basis of block grants
transferred from the national level.

In the municipalities with the highest per capita incomes there is probably little need for subsidizing day care, which they also avoid doing, even though they might seem more likely than others to be able to afford it. This relationship is stable and strong. Municipalities with above average incomes have larger fees than those with less affluent citizens. As the average income increases by NOK 1000, the fees grow by approximately NOK 50 or by 5 percent of the income increase.

The impact of party strength continues to induce increasing fees as long
as municipalities are dominated by either socialists or conservatives. And

again it is worth noting how similar the impact of conservative and socialist
strength is.

Ideology; the Family and the State

Our results so far indicate no major differences between socialist and conservative policy regarding the use of family-paid fees in public child care. There are, however, reasons to add some nuances to this general picture. Different norms and ideologies pertaining to the family, the state and the relationship between those two institutions could introduce consequential in the motives and logic behind similar policies.

The family concept of the conservatives is deeply rooted in traditional family values including opposition to promiscuity, divorces and abortions; and it is based on a pronounced and obvious patriarchal division of labour within the household (Morgan 1985). The socialist perception of the family is the micro-model of the ideal socialist society with its emphasis on values like equality, solidarity and egalitarianism. While the conservatives oppose state intervention toward the family, believing that bringing up children remains a private family matter, the socialists support wider intervention by the state in order to assure social control and equal opportunities for all children, and equal competitiveness for both genders in the labour market.

The priority of the socialists is readily apparent given this ideological heritage: to expand the volume of day-care facilities rather than holding costs at a low level. And they have indeed been increasing family-paid fees in order to fulfil the goal of expansion.

The modern conservative voter, normally urban, well-educated with an above average income, and living in a household where both parents have an occupation, needs and prefers day-care institutions. But to the conservative politician the monetarist principle of reduced public expenditures to be priority number one. These conflicting goals are reconciled in two different ways: first by stimulating day-care facilities as private enterprises; secondly, by relying on the principle of "actual costs paid by the consumer" if institutions are run publicly.

The consequence seems to be twin systems of day-care facilities;

1. Expensive ones where the conservative rely on, or take over public
institutions.

2. Less expensive public institutions competing with a dominant or
expanding private sector. But even here prices increase with the
strength of the conservatives.

Our political control and reference category, consisting of municipalities
dominated by Christians and Agrarians (in some cases non-partisan politicians)consists
advocates of the most traditional family ideology, in

which the family and the "mother" are the important care-taking institutions.These have most strongly argued for universal and direct economic support to individual families instead of subsidies to day-care institutions. The parents themselves can decide whether to use public day care or to leave the child care to the mother. This ideology could produce a low-issue priority: the lowest frequencies of public day-care institutions, and also the most expensive ones when they do exist. The paradox is that it is in these municipalities that we find both the best coverage and the least expensive public institutions.

In 1990 there were 139 municipalities dominated by the centrist parties in the sense that the centrist share of the vote had accounted for 40 percent or more over the last two elections. In these communities the average rate of children enrolled in day-care institutions is 33.1 percent at an average fee of NOK 5878. For municipalities with socialist strength above the 40 percent level in the two most recent local elections (N = 239) the comparable are 30.7 percent and NOK 7345. In conservative-dominated communities (N = 22) the corresponding figures are 20.9 percent and NOK 8318.

The explanation is probably less of an ideological than of a structural variety. Centrist parties have their strongholds in rural, and most of the time small municipalities with few children in actual numbers. This means that basic investment is relatively inexpensive, and that there are few children to subsidize or parents to tax, and that as soon as institutions are built, they cover a large proportion of the children.

Home-Help Services: A Substitute for Old-Age Homes

The aim of home-help services is to assist with the daily domestic functions of elderly and permanently disabled persons, people who are temporarily ill or for other reasons in need of care. These services serve as an alternative to residence in old-age homes, and are a less expensive option. Home-help services also include "housewife's substitutes", and various other services such as meal distribution and cleaning. These services exist in every municipality, the local government has to decide whether to charge for them, and, if so, to what extent.

While the number of homes receiving assistance from housewife's substitutes decreased from 45,800 in 1979 to 20,000 in 1990, the number of homes receiving home-help services has steadily increased. In 1979 and 1990 the number of homes receiving help was 98,400 and 148,400 (SSB 1995).

According to the municipal accounts, the gross expenditure for social

home-help services went up from NOK 1.7 billion in 1979 to 2.24 billion in 1990. The fees charged for home-help services accounted for about 4.5 percent of total expenditures in 1979 and increased to 8.8 percent in 1990. Again there is evidence of the increasing fiscal importance of the fees for local government finance.

The average yearly fee from homes that received home-help services which was NOK 683 in 1979 had gradually increased to NOK 1332 in 1990. However, major variations are evident. A few municipalities manage to run these services without charging any fees at all, and some of them generate considerable incomes from these fees, the most expensive municipalities NOK 3692 per recipient.

Conservative Rhetoric Matters

We have seen that the socialists and the conservatives are about equally unwilling to subsidize the younger generation through lower fees for daycare So what about the older generation? Are they subsidized or do they also have to pay? We would obviously expect the socialists to be less inclined to charge elderly people for the help they receive in their homes. Even charges varying with the recipient's ability to pay would be outside their ideology of the universal welfare state. Throughout the entire post-war era, the older generation has served as a major target group for the Labour Party which has consistently emphasized that illness, injuries, and handicaps should not be allowed to affect the economic situation of the old.

The conservatives, on the other hand, could either stick to their preferences
"users pay actual costs", or to their election campaign rhetoric
of welfare state priority for the elderly.

It is readily apparent from Table 5 that the latter alternative corresponds to the position consistently adopted by the conservatives in the local arena throughout the entire period, even though the coefficients are significant only at the bivariate level. With every percentage point which the conservatives in 1987 and 1990, the fees paid by the elderly for home assistance are reduced by about ten Norwegian crowns or by some seven crowns if statistical controls are introduced. At the bivariate level, the figures are virtually the same for the socialists, but they remain unaffected by statistical control. An increase in the socialist electoral strength by one percentage point reduces the charges by around NOK 10.

Socialist- and conservative-dominated communities would thus seem to be governed according to the same logic. Fees for day care are increased and the charges imposed on the elderly are reduced. In contrast, the centrist parties and the non-partisan local governments administered the least

expensive day-care institutions, but ran the most expensive services for
older people.

Expensive home-help services are related to the proportion of people aged 67 years (pension age) and older. Compared with the proportion of children, the figures for the elderly display much more by way of variation, ranging as they do from a minimum of 6 per cent to a maximum of 30 percent of the inhabitants. As the proportion of the elderly increases, so do the charges on home-help services, even if there are variations over time. In 1979 and 1987 a one percent increase in the proportion of the elderly increased the home-help charges by about NOK 20. The corresponding for 1983 and 1990 is NOK 30.

However, again there are interesting underlying relationships. Conservative strength is strongly negatively correlated with the age of the population (r = -0.53) as opposed to Socialist Party strength, which basically yields a zero correlation. This indicates that the conservatives protect the elderly against increasing fees in municipalities where the elderly are relatively few, while the socialists protect the older generation also when the proportion of the elderly grows. With their strongholds in municipalities an ageing population, the centrist parties are particularly vulnerable as the pressure from the proportion of elderly increases. In this case the size and urbanity of the municipality is of little importance. To the extent that there are effects, bivariate or controlled, they only account for nickels and dimes.

If we change the perspective from the structural properties of the municipalities economic affluence, the similarities between conservative and socialist governments become even more striking. In 1979, when the impact on priorities was still a reflection of the distribution of votes in the 1975 election and the Conservative Party had not yet been through the first mobilization wave, only a socialist impact is revealed. However, the first mobilization surge of 1979 gives an impact on priorities on the fiscal policies in 1983 and induces the conservative-dominated municipalities to reduce fees for the elderly. The same is evident for 1987, as well as in 1990, when the Progressive Party enlarged its share of the votes. For the last three elections and the following budget years there are hardly any differences between the socialist and the conservative camps: priority is given to older people and their services are subsidized to a much larger extent than we found for the younger generation.

On examining the impact of the affluence variables, we note that the average income for each municipality and per capita budget size continues to reveal intriguing patterns (Table 6). Neither affluence measure stimulated services for children's day care, while both encourage reduced fees on home-help services. As the local budget increases by NOK 1000 per capita, the fees for home-help services are reduced by NOK 20 to 30. Even though this does not seem to be a major subsidy, it could obviously make a big difference between living in the very rich compared to the poor municipalities regarding money available for personal spending among those in need of home-help services.

The municipalities with high average incomes behave in the same way. They charge their inhabitants for child care, but subsidize services for the elderly. Again a very stable pattern is established. As the average income increases by NOK 1000, the fees for services are reduced by approximately five Norwegian crowns. Again, this might not seem much. But it is well worth keeping in mind that the average per capita income of the municipalitiesvaries around NOK 25,000 in the least affluent communities

to well above NOK 100,000 in the most affluent municipalities and that a
gap of NOK 1000 might make a noticeable difference for the individual
pensioner.

As was the case for day-care policy, the similarities between the conservatives the socialists with respect to home assistance for the elderly are perhaps by-products of different ideological outlooks and motives. For the socialists this strategy is likely to be seen as a general welfare policy: a widespread public contribution combined with a low or no charge. The credo behind such a policy is public contribution according to need, not support according to the individual user's economic condition - supporting the principle of social rights.

The conservatives are more likely to argue from a liberalistic point of view; the elderly have through a long life of work earned their right to public support. The centrists, for their part, are more likely to put forward the responsibility of the family and to encourage the voluntary sector, especially the religious organizations.

Fees on Utility Services

Fees paid for utility services provided by local governments are obligatory for every house-owner in Norway. Utility services include water supply, common sewer services, and garbage removal, among the most important services. In 1979 the average charge for these services was NOK 137 per capita. In 1983 this average had increased to NOK 294 per capita, in 1987 to NOK 488, and in 1990 to NOK 648 per capita.

The variation among the municipalities is substantial. The fees on utility services vary from a minimum of NOK 281 per capita to a maximum of NOK 2317 per capita in 1990; while nine years earlier some municipalities provided these services free of charge, and the most expensive municipalities only one-quarter of the amount charged by the most expensive 1990. Again this demonstrates the increasing significance of fees and charges as a monetary resource for the municipalities. The variation in technical fees has also increased quite substantially over the years surveyed.

Technical Fees: The Conservative Money Source

Contrary to fees for day-care and home-help services, which are fees imposed on people in certain situations, technical fees are of a universal kind. Everybody needs water, access to a sewer system and garbage removal. With a few exceptions, there is no differentiation in the size of the fees related to the amount of the services used; just being a houseowner the category and the size of the charges. We expected the conservatives to be most inclined to pursue fees on these services at a level where the users have to pay most of the actual costs.

Table 7 indicates that our assumptions are correct and that the impact of electoral victories widened the conservatives' ability to use these fees as their fiscal strategy. In 1979 a one percent increase in the support for the conservatives increased the technical fees by less than one Norwegian crown. But there is evidence of a steady increase, and by the end of the 1980s the impact of a one percent increase was almost nine Norwegian crowns. There is also a positive association between the strength of the socialist parties and technical fees, but it is of a much smaller magnitude, close to four Norwegian crowns as the strength of the socialist parties increases by one percentage point at the end of the decade.

Among the non-political variables, the structural properties of the municipalities, both size and urbanization, are important. As the proportion of the population living in urban areas increased by one percent, technical fees increased by around two Norwegian crowns in the early years of the decade, and around five crowns at the end. As the size of the population increased by 1000 inhabitants, the technical fees went up from about one Norwegian crown in 1979 to two in 1990.

There is, as revealed in Table 7, a strong positive correlation between utility fees and local government spending on technical services - and this effect persists as we introduce controls for the political and urbanization variables. During the entire period the large urban municipalities, with conservative dominance, come closest to applying technical fees at a level where the users have to pay the actual costs.

With control for affluence, the effect of party strength continues to reveal its impact (Table 8). A much larger increase in fees is related to conservative strength than to the strength of the socialists, and the difference is intensified throughout the decade, even though conservative and socialist municipalities more inclined than centrist municipalities to increase technical fees.

As we already know, the conservatives have their strongholds in municipalitieswith
incomes, and when controlling for conservative strength
we find a direct relationship between average income and amount of the

fees paid: people in richer communities pay more. As the average income increases by NOK 1000, the fees for technical services increase by slightly more than two Norwegian crowns during the 1980s and by almost three and a half crowns in 1990.

Local budget affluence has no significant impact on the size of technical fees in either of the years examined. It is mainly the affluent conservative municipalities that use technical fees as an important part of their fiscal foundation.

Property Taxes: The Cleavage Line?

So far, we have not been able to detect any major differences in fiscal policies between socialist- and conservative-dominated municipalities. They seem to use the same options to improve their budgets. It is basically a matter of higher or lower fees on day care, home-help services and technical services. The major difference is that conservative-dominated municipalitiesapply

palitiesapplytechnical fees more extensively than socialist-dominated communities. With respect to these policies, there is no cleavage between conservatives and socialists. The demarcation line runs between the socialist- and the conservative-dominated municipalities on the one hand and the centrist- and non-partisan-dominated communities on the other.

Does this pattern change when we evaluate the impact of political
strength on property taxation? We expected avoidance of tax increases to
be the first political priority of the conservatives.

Unfortunately, data on local government property taxes are not available prior to 1987. Therefore, we will only be able to analyse the party impact on property taxes for two elections: 1983 and 1987, with the impact measured for the fiscal years of 1987 and 1990 respectively.

First, it is appropriate to say a few words about property taxes and the decision to implement this income potential. When deciding on fees for day care, home-help services or technical services, local government may adjust fee sizes to such an extent that real costs are covered. In this way they can actually determine the quantity of local government revenue. This does not apply to property taxes. It is essentially a question of deciding whether to introduce them or not. If the municipality assembly decides to do so, the taxation level is regulated by the national government in such a way that private property will be taxed by 0.002 percent during the first year; during the next four years it will automatically increase to 0.007 percent - a level above which it will never be raised. After four years the quantity of money generated by the tax is exclusively a function of the amount and value of each municipality's private property, and as such the amount of money raised has nothing to do with political decision-making. For our analysis this means that we will introduce a dichotomous variable where the municipalities are divided into a group having introduced property and another one that has not. Of course, we expect increased support for socialists to stimulate implementation of property taxes while conservative-dominated municipalities will avoid this tax. This means that we are working with a dichotomous variable which precludes OLS regression, which is no longer appropriate, since the probability estimates can either drop below oor exceed 1. We shall instead apply Probit regression where the estimated probability impact for each variable, and the cumulative of the variables are confined by the probability limits of 1 and 0. The major disadvantage with Probit regression is that the coefficients do not have an intuitive interpretation and that the probability estimates of one unit change in an individual independent variable will vary with the value assigned to other independent variables. In our analysis the estimate of probability changes is based on a one unit change from the mean value for each individual independent variable, while holding the remaining independent variables constant at their mean values.

Our analysis utilizes data from two years. In 1987, 167 out of the 426 municipalities on which the analysis is based had implemented property tax. In 1990 this number had increased to 189; 23 additional municipalities had introduced property tax, while one had abandoned this source of income.

By assigning a mean value to all variables in Tables 9 and 10, we obtain an initial probability of implementing property taxes at 39 and 44 percent in 1987 and 1990 respectively. It is the change in probability from this average that we are measuring when the value of each independent variable varies by one unit change.

By controlling for structural properties such as population size and urbanization, we can assess the impact of political strength as revealed in Table 9. As expected, the probability of using property tax increases with socialist strength and decreases with conservative strength, but not by very much. An increase of one percentage point in socialist support above the socialist average support of slightly more than 43 percent in both elections, raises the probability of property tax by less than one percent. An increase of one percentage point above the conservative average of 21 percent and 22 percent in 1987 and 1990 respectively increases the probability of not

having property tax by 1.1 percent for both years. We need rather large deviations from the means to reach major differences in the chances that property tax is or will be introduced. A support for socialists close to 60 percent increases the likelihood of having property tax by around 10 percent. We need support above the 30 percent level for the conservatives to reduce the probability of property tax by 10 percent. However, this situation is not uncommon. In 1983 as many as 69 of our 426 municipalities had socialist majorities of 60 percent or more. This figure dropped to 53 percent in the election of 1987. The two conservative parties had a combined strength of 30 per cent or more in 88 and 108 municipalities in 1983 and 1987 respectively. Under these circumstances we arrive at a probability difference of around 20 percent for having introduced property tax.

With the other variables we also notice certain impacts. As the size of the population increases by 1000 inhabitants above the mean (around 10,000 inhabitants), a probability change of almost one percent for property taxation is seen in both years. In 1990 urbanization also had an independent impact.

Controlling for fiscal and private affluence reveals more of the story (Table 10). The impact of socialist strength hardly changes at all, indicating that, whatever the circumstances, the socialists are more willing to use property tax compared to all other parties, but by now the negative impact of conservative strength fades to nothing. Both measurements of affluence influence the probability of property taxation, but in opposite ways. As the average pre-tax income increases by NOK 10,000 above the average, the probability of having property taxes implemented is reduced by almost 4 percent in 1987 and above 2 percent in 1990. However, an increase in the per capita local government budget by NOK 1000 raises the probability of property taxes by about the same amount.

This indicates that if local governments generate substantial revenues from direct income tax, they are reluctant to introduce property taxes and most of the time this is exactly where the two conservative parties reign. Municipalities with per capita large budgets, basically financed through national transfers tend to supplement their income by property taxes. These are the municipalities where the socialists and centrist parties compete for political dominance.

Finally, during the period from 1987 to 1990 only one municipality decided to end property taxation, which means that some of the municipalities the two conservative parties gained in support in both 1987 and 1990 were municipalities which already had introduced property taxes. This explains part of the weak and non-existing association between conservative and the probability of property taxation, but it also demonstrates that even conservative politicians are reluctant to surrender fiscal resources.

Conclusion

This article has examined the fiscal policies of Norwegian municipalities, asking the key question for local democracy: does party strength really matter? The question is raised against the background of the political mobilization by the conservatives during the local elections of 1987 and 1990.

Similar questions have been asked by several Norwegian scholars during the past 25 years of local government studies; and the answers have been somewhat ambiguous. Nordbø (1968) and Viken (1974) demonstrate that party strength during the 1960s and early 1970s had a significant impact on local government spending on social welfare, even though the effects of demographic and economic factors were even more salient. Yet, Pettersen (1984) indicates that the importance of party politics was declining during the 19705, especially regarding the impact on local expenditures, and hardly recognizable at all by the beginning of the 1980s. The conflict between parties, and the party political impact, seems to have vanished - a conclusion in line with that of research findings from a variety of countries.

However, as Hansen & Kjellberg (1980) advise: it is necessary to distinguishbetween
policy areas according to the autonomy of local
government. Only when the autonomy is extensive, where local political

authorities are free to choose between various goals and means, can we expect the strength of parties in local government to have any political bearing. Local autonomy provides the rationale for the fiscal strategies that were singled out for analysis in this article.

Some would argue that, at least at the local level, the conservative rhetoric and proposed policies of tax reduction have been a complete failure, since they have never been able to increase fees and charges to such an extent that they have been able to cut the direct income taxes. This is common knowledge, but the conservatives have been free to pursue such a policy. In the process, we have discovered differences between municipalities with distinct party political profiles which are less directly related to political rhetoric and controversy.

Having examined the political impact of local fiscal strategies of the 1980s, we can draw but one conclusion: parties do matter, but not always as expected. Conservative strength in local government does produce fiscal policies in accordance with conservative programmes and rhetoric - prices are defined so as to ensure that the consumer pays a larger share of the actual cost of the services. Local governments controlled by the socialists pursue similar policies, which leaves the centrist and non-party administrations those that most heavily subsidize services provided by local government.

There are, however, differences between various welfare services; the elderly are protected against this fiscal strategy by both conservatives and socialists, and are given priority in the domain of welfare policies - a tendency which became more pronounced as the Progressive Party expanded its political base. Fees for children in day centres - basically used by the young generation - and fees on technical services used by every household constitute the preferred fiscal instruments of the conservatives and socialists alike. As for the latter, the conservatives are more likely than the socialists to impose tax burdens on the households.

For the most part, conservatives and socialists do not differ much with respect to fiscal policy. There is, however, reason to believe that their motives are different. While the conservatives pursue fiscal policies according the "user pays the cost" principle, the socialists promote fees as a strategy for expanding the volume of welfare services. This seems to be especially true for day-care services.

The cleavage between the conservative and socialist parties is brought out by their approach to property taxes. The socialist are inclined to implement such taxes as their strength grows whatever the structural and fiscal circumstances of the municipality. The conservatives shun them, but controlling for the affluence (average pre-tax income) of the municipality, the impact fades away. The implication is that the conservatives have their strongholds in affluent communities, where property taxes are redundant.

The zero correlation between conservative strength and the probability of utilizing property taxes obtained when controlling for affluence also serves as an indication that property taxes, if introduced, are hard to abandon even if the conservatives increase their share of the votes.

We have finally seen evidence of an increasing reliance on fees, charges and property taxes in the wake of the replacement in 1986 of earmarked transfers by block grants from the national government. This fiscal conservatism to be linked to a new sense of responsibility for local finances among local politicians which was not encouraged when the services rendered were heavily subsidized by the national government, as was the case prior to 1986.

There are those who suggest that people will end up choosing where to live according to the quality and price of the services provided by local government (Tiebout 1956). Even though we have found differences between municipalities with respect to prices and service volume related to party strength, it should be emphasized that these differences are not large enough to recommend that people move according to their preferred party colour of local government. Moving expenses will definitely outweigh the economic advantages.

ACKNOWLEDGEMENT

Data for this article have been provided by the Norwegian Social Science Data Services.

REFERENCES

Boaden N. & Alford, R. R. 1969. "Sources of Diversity in English Local Government",
Public Administration 47, 203-223.

Boyne, G. A. 1984. "Review Article: Theory, Methodology and Results in Political Science -
The Case of Output Studies", British Journal of Political Studies 15, 473-515.

Castles. F. G. ed., 1982. The Impact of Parties. London: Sage.

Castles, F. & Borg, S. G. 1981. "The Influence of the Political Right on Public Income
Maintenance Expenditure and Equality", Political Studies XXIX, 604-621.

Castles, F. G. & McKinlay, R. D. 1979a. "Public Welfare Provision, Scandinavia, and the
Sheer Futility of the Sociological Approach to Politics", British Journal of Politics 9, 157—
171.

Castles, F. & McKinlay, R. D. 1979b. "Does Politics Matter: An Analysis of the Public
Welfare Commitment in Advanced Democratic States", European Journal of Political
Research 7, 169-186.

Dawson, R. E. & Robinson, J. A. 1963. "Inter-Party Competition, Economic Variables and
Welfare Policies in the American States", Journal of Politics 25, 265-289.

Fry, B. R. & Winters, R. F. 1970. "The Politics of Redistribution", American Political Science
Review 64, 508-522.

Hansen, T. 1981. "Transforming Need into Expenditure Decisions", In Newton, K., Urban
Political Economy. London: Frances Pinter Ltd.

Hansen, T. & Kjellberg, F. 1976. "Municipal Expenditures in Norway: Autonomy and Constrains
Local Government Activity", Policy and Politics 4, 25-50.

Hansen, T. & Kjellberg, F. 1980: "Kommunale utgifter i Norge: Autonomi og sentral kontroll

i lokalforvaltningen", in Kjellberg, F., Den kommunale virksomhet. Oslo: Universitetsforlaget.

Kjellberg, F. 1985. "Local Government and the Welfare State: Reorganization in Scandinavia",
Dente, B. & Kjellberg, F., The Dynamics of Institutional Change. Local Government
in Western Democracies. London: Sage.

Morgan, D. H. J. 1985. The Family, Politics and Social Theory. London: Routledge & Kegan
Paul.

Norbø, E. 1968: Kommunenes ytelser i sosialsektoren. Oslo: Hovedoppgave. Institutt for
Statsvitenskap. Universitetet i Oslo.

NOS 1982: B-298: Statistisk årbok 1982. Oslo: SSB.

NOS Municipal Council Elections. Oslo: SSB. 1975, 1979, 1983, 1987.

NOU 1974:53. Mål og retningslinger i lokalforvaltningen. Oslo: Kommunaldepartementet.

Pettersen, B. H. 1984. Innsatsutviklingen i den kommunale sosialsektoren. Oslo: Hovedoppgave.
for statsvitenskap. Universitetet i Oslo.

Sharp, L. J. 1981. "Does Politics Matter?", in Newton, K., Urban Political Economy. London:
Frances Pinter Ltd.

SSB 1975-1994: Structural Data from Municipal Accounts. Oslo: Central Bureau of Statistics
of Norway.

SSB 1994: Statistical Yearbook 1994. Oslo: Central Bureau of Statistics of Norway.
SSB 1995: Historical Statistics 1994. Oslo. Central Bureau of Statistics of Norway.

Sørensen, R. J. & Underdal, A. 1993. "Coping with Poverty: The Impact of Fiscal Austerity
on the Local Budgetary Process in Norway", Scandinavian Political Studies 16, 49-71.

Tiebout, C. M. 1956. "A Pure Theory of Local Expenditures", Journal of Political Economy
LXIV, 416-424.

Viken, A, 1974. Kommunalpolitikk og sosialomsorg. Oslo: Hovedoppgave. Institutt for
Statsvitenskap. Universitetet i Oslo.