Resumé

The community equilibrium model of local fiscal policy-making perceives of political leaders as 'invisible hands' or as anonymous, reactive figures who automatically adapt to the preferences of the citizenry. The party model, in contrast, views political leaders as willful representatives of certain interests in society who will pursue policies that are in accordance with the preferences of the party and its followers. In a context which is clearly biased in favor of the latter, the two models are tested. Under these extreme circumstances the community equilibrium model comes out as the most fruitful with respect to prediction of the fiscal behavior of local political leaders.

Two Conflicting Theories of Local Policy-Making

Most research on local fiscal policy-making falls into two categories. According to one tradition policy-making can be thought of as the result of a reaction to external forces by a political system which is constantly striving towards a sort of equilibrium with its environment. A second line of thinking stresses the specific characteristics of the political system as a determinant of public policy. Within the latter theoretical perspective it is customary among political scientists to look at the distribution of power between the political parties.

Although both perspectives are marked by great variation in terms of methodology, theoretical richness and consistency, it is nevertheless possible to represent them by means of two simple core models called the community equilibrium model and the party model.

The Community Equilibrium Model

At the beginning of the 1960s a search for systematic relationships between public policies, usually fiscal policies, and environmental characteristics was started among economists and political scientists. Within a decade a totally new 'determinant' or 'policy-output' literature emerged. Although the findings varied and although conflicting evidence showed up on a number of issues the message from these efforts was clear: Public policies

are mainly the results of non-political forces, that is, they are to be explained
by characteristics of the social and economic environment of political
systems.1

Much of this literature rested on implicit assumptions about the mechanisms through which socio-economic features were transformed into public policies. What these studies seem to have in common are the notions of utility-maximization and equilibrium, either in the form of what has been called a 'household consumption behavior analogy'² or in the form postulated by the 'median voter theorem' (cf. Søndergaard 1982). According to the latter, the logic of voting will result in policies that are congruent with the preferences of the median voter. The preferences of the median voter are therefore equivalent with the point of equilibrium.

Utility-maximization models are usually based on very restrictive assumptions, rarely corresponding to real world phenomena.³ Nevertheless, they may approximate the actual mechanisms of decision-making sufficiently enough to allow for forceful predictions. This may be true especially with regard to the fiscal behavior of local governments (Søndergaard 1982, 6).

To sum up, the core of community equilibrium models is that public policies are the result of adjustments to the needs and preferences of the population, that these needs and preferences can be represented by the social and economic characteristics of the community, and that these 'environmental' characteristics are translated into policies independently of who holds the power in the political system.

The Party Model

The assumptions of the community equilibrium model allow for the treatment of the political system as a black box. In contrast, the party model gives the holders of power some autonomy to decide on policies independently of the local environment. This relative autonomy has two sources, both of which have to do with the incentive structure of the local political elite.

First, political leaders are assumed to represent certain social classes or categories, which may hold different or even opposing preferences and ideologies. Political parties are accordingly perceived as representatives of specific interest constellations in society, and political leaders will seek to maximize votes subject to the constraints imposed upon them by the core supporters of the party. This is quite in contrast to the median voter theorem or other variants of the utility maximization model which simply assume that politicians will maximize votes or seek a majority.

Second, the stronger the link between party and social class, the more likely it is that strong national parties will emerge. These parties will mirror the administrative structures of government with branches at the central, regional, and local level. The higher the degree of congruence in party structures at different levels of society, the more likely it is that the local

party branches are subject to control from, and will be loyal to, the central party organizations. This loyalty may have two sources. One obvious source is the ideological basis on which many parties are grounded, a basis which is likely to be shared by political leaders at all levels.

More important, perhaps, is the fact that a strong party system in a country may break the link between public (local) policy and electoral success. In such a system there are few chances for an independent political entrepreneur to be elected. Nomination from within the party is normally a prerequisite to a seat in the local council. Nomination comes from party recognition, which is usually the result of years of loyal party activity. The individual politician therefore often owes more to the national party and its platform than to his particular stand on issues and often he does not even depend on the local party for his success at the polls. Election results are more likely to be a function of the party at the national level. A landslide in a national election may often have important consequences in the next local election.

The importance of the success of the national party may lead to a heavy orientation on the part of the local political leader towards the central party for guidelines, and party program is likely to be a major frame of reference with respect to decisions on local policies.⁴ The relationship between parties at the national and local level, and the distribution of power among these parties at the two levels, are therefore likely to be a major determinant of local policy-making. Party congruence, that is a party holding power nationally as well as locally, is likely to lead to 'reinforcement' (Sharpe & Newton 1984, 18): National policies towards local government are likely to reflect the preferences of the local party, and the local party is likely to implement those policies because the success of the party at the national scene, as judged by the degree to which its policies are implemented, may have important electoral consequences in the next local election.

The basic perspective inherent in the party model therefore contrasts sharply with the assumptions of the community equilibrium model. The former perceives of political leaders as willful implementors of certain preferences, ideologies and policies which favor certain social classes. The latter treats the local politicians as anonymous characters who - like the invisible hand - see to it that public policies are in accordance with the preferences of the community.

The Problem and the Context

Studies of the effect of socio-economics and politics are numerous. So are the conclusions.⁵ There is no doubt that the great variation in conclusions about the importance of politics is due to differences in design, methodology, and techniques as much as it is grounded in substance. Therefore it

is necessary to specify with great precision the problem studied here and
the space and time context in which it is studied.

The two models of local policy-making will be tested in an analysis of changes in fiscal policies. More specifically the study intends to explain changes in aggregate expenditures in Danish local (municipal) government from 1982 to 1986.

Concerning the systemic context there are some structures which will tend to favor the party model (in the sense of its relative explanatory power as against the community equilibrium model). First, the local party system is a reflection of the national system. At the local level two parties dominate, the Agrarian or Liberal Party and the Social Democratic Party, both of which had their origin in clearly demarcated social classes before and around the turn of the century. They possess 77 per cent of the mayoral positions with 50 to the former and 27 to the latter. The mayors in the remaining municipalities belong to either the Conservative or the Radical Liberal party, or to non-partisan lists which carry about 15 per cent of the votes at local elections. So generally the party system is strong and the parties are relatively coherent with branches at all levels of government.

Second, although two parties tend to dominate, we find in most jurisdictions a multiparty system - the average number of parties or lists represented being 5.2. This is not in accordance with one of the assumptions of the community equilibrium model which - at least in the median voter version - requires that two parties (or candidates) are competing for voters (Rowley 1984, 105). It is therefore rather unlikely that the political parties will play the role of the invisible hand.

The specific time period investigated may also entail a certain bias in favor of the party model. It was a period with relatively clear and explicit policies towards local governments implemented by the bourgeois coalition government which took power in the fall of 1982. In the same period ideologically based conflicts about the proper role of government in society and about the possible remedies to stop the growth in the public sector divided the political parties.

Finally, 1982-86 was a period of increasing fiscal pressure brought upon local governments mainly because reductions in the general grants were a major element in the fiscal policies of the bourgeois government. The degree to which this pressure was being felt can be illustrated by the fact that municipal governments with constant tax rates would have had to cut their operating expenditures from 1982 to 1986 by almost 9 per cent (cf. Mouritzen 1987a, Table 1). One study has shown that partisan differences are more marked in times of increasing resource scarcity than in times of abundant resources (Mouritzen 1985a). In the latter case politicians are inclined to fight for their particular sectors or programs. It has also been shown that the level of conflict in the municipal council increases with increasing degrees of fiscal stress (Mouritzen 1985b).

Conversely, there may also exist certain biases in this study favoring the community equilibrium model. One important source of such bias is that the analysis focuses only on changes in aggregate spending. It has often been suggested that partisan differences have a tendency to manifest themselves when spending is disaggregated into specific functions while they tend to be hidden at higher levels of aggregation (Sharpe & Newton 1984, 12). On the whole, however, both with respect to the particular time period and setting investigated, most is in favor of the party model of local fiscal policy-making.

Does Policy Matter?

One of the assumptions of the party model is that very weak relationships exist between local fiscal policies and the choices made by the voter at local elections. Analyses have shown that fluctuations at local elections do in fact correspond strongly with fluctuations at the national elections, although the relationship is weaker than it used to be (Thomsen 1986).

The 1985 local election presents an ideal case for a study of this major assumption of the party model. One of the most sensational results of that election was that the People's Socialist Party (PSP) increased its share of the votes from 5.7 to 10.5 per cent. This success was undoubtedly the result of national electoral swings but could perhaps also be attributed to local policies during the previous election period. In fact, if there is a relation between local fiscal policy-making and election results it ought to be found at the 1985 election. The reason is two-fold.

First, the election period from 1981 to 1985 was marked by highly visible and politically conflictual reductions in the grants to local governments and-in many municipalities - subsequent reductions in services. Potentially two parties could gain from these trends: the Social Democrats and the PSP. The former, however, did hold power in many municipalities, and could be held responsible for cut-back policies there. In many municipalities there is a strong tradition for cooperation between the Social Democrats and the bourgeois parties, in particular perhaps with respect to budgetary issues. In contrast, the PSP usually had been in opposition on fiscal issues and was by far the largest of the left-wing parties with seats in 95, or one third, of all municipalities.

Second, the 1985 election is an exemplary case because it was the most
'locally based' election since the grand municipal reforms at the beginning
of the 1970s (Thomsen 1986, 294-297).

In order to test the assumption of the party model about 'nationally based'
local elections, a simple model with three core variables was estimated.
The result is depicted in Table 1.

It is evident from the Table that the dominating determinant of PSP success⁶ is the relative strength of that party after the 1981 election. The stronger the party came out of that election the smaller the relative increase in share of votes. This can be taken as an indication of a kind of upper threshold of party support. The closer to that threshold the more difficult it is for the party to increase its share of the vote, and the weaker the fluctuations due to fluctuations in support for the party at the national scene.

More important in the present context, however, are the two remaining variables. The degree of fiscal stress is positively related to PSP-success. This indicator is a composite measure which has nothing to do with local fiscal policies -at least not in the short run.⁷ It mainly reflects the seriousness of the reductions in grants from the central (read bourgeois) government and of the crises in the local private sector. These externally defined conditions of fiscal policy-making seem to be correlated with the success of the PSP: the more serious the fiscal crises the larger the increase in votes for the People's Socialist Party.

With respect to the growth in expenditures there is no significant effect on the success of the People's Socialist Party.⁸ Real growth in operating expenses from 1982 to 1986, which averaged 2 per cent, ranging from a low of —11 to a high of 25 per cent, had no effect at all on people's voting behavior, at least not the votes cast for the PSP.

The analysis does not totally exclude the possibility of local policy effects - cf. the 65 per cent of the variation not explained by the three core factors - but is consonant with the party model of local policy-making which predicted that electoral success is a function of national political swings rather than local policies. The result is so much more convincing considering the case which was potentially not in favor of the party model.

Does Politics Matter?

According to the party model, local fiscal policy-making would reflect the preferences of party members and supporters as well as the policies of the national party. In particular, the likelihood that central government policies towards local governments would be implemented is predicted to be dependent on 'congruence' or identity between the party in power at the central and local level.

The model has no predictive power without a knowledge of the preferences
of voters and politicians, and of the policies of central government.
These three elements fortunately coincide more or less.

In the fall of 1981, after the passage of the 1982 budget and after the 1981 local election, two surveys were conducted of representative samples of citizens and the political elite. Several questions were posed concerning spending preferences. For citizens as well as local political leaders there are significant differences between the left and right. Bourgeois political leaders and voters express spending preferences which are significantly more moderate that those expressed by politicians and voters from the socialist parties.⁹ Also there was, on the part of bourgeois political leaders, a tendency to perceive of the regulations by the central government (in 1981 a Social Democratic minority government) as a force pushing for higher expenditures and a tendency for them to want less government regulation (Mouritzen 1984).

The policies which have been followed quite consistently since 1982 by the bourgeois government are very much in accordance with the wishes expressed by bourgeois voters and local politicians in 1981. Via yearly agreements with the associations of local authorities and/or legislative action, mostly in the form of reductions in general grants, the government has sought to bring expenditure growth in local government to a standstill. Every year explicit goals have been established concerning expenditure growth and other fiscal parameters. Except for the first year (where the goal was to reduce local spending) the spending goal has always been zero growth.

The relatively high degree of congruence between voters' and politicians' preferences and central government's policies is indicative of a clear and consistent frame of reference for the bourgeois fiscal decision-maker - if the party model has any explanatory power. This frame of reference would lead to fiscal policies that are significantly different in bourgeois and socialist dominated municipalities - if the party model has any explanatory power. Specifically, we can expect municipalities dominated by the bourgeois parties to stick to the spending goals of the central government, while municipalities dominated by the socialist parties would - in accordance with the preferences of their voters - decide expansionary fiscal policies.

The party model does not survive a first test. Real growth in operating expenditures from 1982 to 1986 were 1.7 per cent in municipalities with a Social Democratic mayor and 2.1 per cent in municipalities with a bourgeois mayor. There is a significant negative relationship between strength of socialist parties and real growth in operating expenditures (r = — .12).¹⁰

The Community Equilibrium Model

The community equilibrium model treats local governments as single entities and fiscal policies as largely predetermined by external constraints. The method of the community equilibrium model takes 'the preferences of political leaders as given and focuses attention upon changes in public preferences and therefore the demands made upon resources available to political leaders'. Therefore, fiscal policies 'would be expected to vary with changes in the constraint values of the conditions and external resource constraints' (Thompson & Williams 1979, 125).

Real growth in operating expenditures from 1982 through 1986 did vary a great deal. The community equilibrium model would seek the reasons for this variation in changes in demand and resources available. More specifically the model would predict:

Hypothesis 1. The larger the (positive) 'gap' between demand and actual
services at the start of the period the larger the increase in spending.
Hypothesis 2. The larger the increase in demand during the period, the

larger the increase in spending.

Hypothesis 3. The larger the increase in available resources, the larger the

increase in spending.

Table 2 reports the results of a test of these hypotheses (cf. the appendix
for technical details).

The first hypothesis postulates that decision-makers will take into

sideration the existence and size of a possible gap between demand for and supply of public services. The measure 'Demand 1982' is based on a survey of citizens' spending preferences in the late fall of 1981 (referred to earlier). As the data are only available in a national sample it has been necessary to build a simple simulation model based on policy characteristics of the individual municipalities (described in the appendix). The positive coefficient assigned to 'Demand 1982' is supportive of the first hypothesis: The higher the demand for spending in the community in 1982, the larger the increase in public spending from 1982 to 1986.ⁿ

The second prediction is also supported. Change in expenditures corresponds to changes in expenditure needs. Conventional economic theory teaches that demand increases with rising income. Therefore taxable income may be interpreted as 'demand' indicator. The best predictor of changes in municipal spending is in fact changes in taxable income. Finally, expenditures increase with decreasing unemployment. This may seem a bit odd, but in some areas demand clearly does rise with decreasing unemployment, notably for day care institutions where there is a close link between the employment situation of a family and the need for day care.¹²

According to the third hypothesis spending would increase with increasing resources. The two variables, loss of grants (from 1982 through 1986) and liquid assets 1982, measure the resource constraint imposed upon a municipality. Liquid assets obviously indicate the amount of resources which can be used in times of fiscal crises without visible and immediate effect on the tax payers' real income. Both resource constraint indicators affect local spending in the expected direction: The larger the increase in grant (or in reality the smaller the decrease) and the better the financial situation in 1982 the larger the increase in operating expenditures.

The test in Table 2 is really a test of a hybrid model, as it contains elements from the community equilibrium model and the party model of fiscal policy-making. The coefficient attached to the remaining variable, 'Social Democratic Mayor', shows that political party does matter on the margin. When the external conditions imposed upon political leaders are held constant, the preferences and policies for which these leaders and their party stand do have an effect on their decisions. The effect is in the expected direction, i.e. Social Democratic municipalities do - ceteris paribus - increase expenditures faster than bourgeois municipalities.

Conclusions

The result of this study is very similar to other studies of local fiscal policymaking.
In contrast to most popular beliefs and classical political science,

politics does not matter very much. The preferences, policies and platforms of political parties and the wishes of party followers only have a marginal effect on public budgets. The community equilibrium models which on the other hand rest on highly doubtful assumptions have considerable explanatory strength in comparison with the party model. As most other studies have proven, these assumptions, which are perhaps best described as 'useful fictions', do lead to rather strong predictions of the behavior of local government.

What distinguishes this study from most other studies is the case being investigated. For a number of reasons it discriminates against the community equilibrium model, i.e. it was biased in favor of the party model of policy-making. It was shown that the context of fiscal policy-making was one marked by strong, national parties and that local budgetary policies did not have electoral consequences. Therefore the study lends considerable weight to the findings of mainstream budgetary literature which under less extreme circumstances also tends to discredit the importance of the political party.

The methodology applied in a simple and straightforward way rejected the party model of fiscal policy-making. The prediction which follows from the model and which is based on a relatively precise knowledge of the preference of voters, political leaders and of the policies of the bourgeois government, is simply not true. All the predictions made by the community equilibrium model were found to be true. The final hybrid model, however, can be taken as an indication that the two models are not necessarily competing. The community equilibrium perspective seems to be the most fruitful with respect to the explanation of fiscal behavior, but at the margins the party in power did have a significant effect on budgetary decisions. Political leaders listen to public preferences but are able to pursue their own interests and preferences up to the limits of the constraints imposed by the public. In other words, local politicians are more loyal to the local constituency than to the party.

REFERENCES

Clark, T. C. & Ferguson, L. C. 1983. City Money, Political Processes, Fiscal Strain, and
Retrenchment, N.Y.: Columbia University Press.

Downs, G. & Larkey, P. 1979. 'Theorizing about Public Expenditure Decision-Making: (as)
if Wishes were Horses', Policy Sciences 11, 143-156.

Fried, R. 1976. 'Comparative Urban Policy and Performance', pp. 305-379 in Greenstein, F.
L. & Polsby, N. eds., The Handbook of Political Science 6. Reading, Mass: Addison

Gramlich, E. 1977 'Intergovernmental Grants: A Review of the Empirical Literature', pp.
219-240 in Oates, Wallace E. ed., The Political Economy of Fiscal Federalism. Lexington,
Mass: Lexington Books.

Gyford J. 1980. 'Political Parties and Central-Local Relations', in Jones, G. W. ed., New
Approaches to the Study of Central-Local Relations. London: Gower.

Jacob, H. & Lipsky, Michael 1968. 'Outputs, Structure and Power: An Assessment of Changes
in the Study of State and Local Polities', Journal of Politics 30, 510-538.

Kristensen, O. P. 1982. 'Voter Attitudes and Public Spending: Is there a Relationship?',
European Journal of Political Research 10, 35-52.

Madsen, S. 1984. 'Serviceydelser paa folkeskoleomraadet', pp. 117-143, in C. J. Skovsgaard,
Kommunernes service under pres. Copenhagen: Jurist- og Økonomforbundets Forlag.

Mouritzen, P. E. 1983. 'Modelling Central-Local Relations. Six Models for the Study of the Interplay between Central Government and Local Levels of Government'. Report no. 26, Central Control, Local Communities and Local Politics. Institute of Public Finance and Policy, University of Odense.

Mouritzen, P. E. 1984. 'Kommunalpolitikerne og det kommunale selvstyre: Ideologi eller
pragmatisme', pp. 187-209, in Mouritzen, P. E. & Jensen, H. T. eds., Samspillet mellem
staten og kommunerne. Copenhagen: Jurist- og Økonomforbundets Forlag.

Mouritzen, P. E. 1985a. 'Local Resource Allocation: Partisan Politics or Sector Polities', pp.
3-19, in Research in Urban Policy 1.

Mouritzen, P. E. 1985b. 'Concepts and Consequences of Fiscal Strain', pp. 13-33, in Clark,
T. N., Hellstern, G. M. & Martinotti, G. eds., Urban Innovations as Response to Urban
Fiscal Strain. Berlin: Verlag Europåische Perspektiven.

Mouritzen, P. E. 1987a. 'Fiscal Policy-Making in Times of Resource Scarcity. The Danish
Case', in Clarke, S. ed., Urban Innovation and Autonomy: The Political Implications of
Policy Change. London: Sage (in print).

Mouritzen, P. E. 1987b. 'The Demanding Citizen. Driven by Policy, Self-interest or Ideology',
European Journal of Political Research (in print).

Peterson, P. E. 1981. City Limits. Chicago: The University of Chicago Press.

Rakoff, S. H. & Schaefer, G. F. 1970. 'Politics, Policy and Political Science: Theoretical
Alternatives', Politics and Society 1, 51-77.

Romer, T. & Rosenthai, H. 1979. 'The Elusive Median Voter', Journal of Public Economics
12, 143-170.

Rowley, C. K. 1984. 'The Relevance of the Median Voter Theorem', Journal of Institutional
and Theoretical Economics 140, 120-126.

Sharpe, L. J. & Newton, K. 1984. Does Politics Matter? The Determinants of Public Policy.
Oxford: Clarendon Press.

Skovsgaard, C. J. et al. 1977. Studier i dansk kommunalpolitik. Aarhus: Politica.

Søndergaard, J. 1982. 'Alternative Approaches to the Modelling of Local Government
Expenditure Behaviour'. Report no. 7, Central Control, Local Communities and Local
Politics. Institute of Economics, University of Aarhus.

Thompson, E. & Williams, R. 1979. 'A Horse Race Around a Moebius Strip: A Review and
a Test of Utility-Maximizing and Organizational-Process Models of Public Expenditure
Decisions', Policy Sciences 11, 119-142.

Thomsen, S. R. 1986. 'Kommunale valg 1970-85: Landspolitisering eller lokalpolitisering?',
pp. 288-298, in Elklit, J. & Tonsgaard, O. eds., Valg og vælgeradfærd. Studier i dansk
politik. Aarhus: Forlaget Politica.