Fiscal Federalism-Decentralisation and the size of government

Master's Thesis, 2015

57 Pages


Table of Contents

1. Introduction

2. Definitions
2.1 Definition of Fiscal Federalism
2.2 Definition of Fiscal Decentralisation and Fiscal Centralisation

3. Fiscal Federalism in Austria
3.1 Decentralisation in Austria compared to other OECD countries

4. Theoretical Perspectives on Fiscal Federalism
4.1 First Generation Theory of Fiscal Federalism
4.2 Public-Choice Perspective of Fiscal Federalism
4.3 Second Generation Theory of Fiscal Federalism
4.3.1 Centralization and Decentralisation in the Second Generation Model
4.3.2 Potential dangers of Fiscal Decentralisation
4.3.3 Risk- Sharing and Interjurisdictional Insurance
4.4 Conclusion

5. Fiscal decentralisation and size of the government
5.1 Introduction
5.2 The Leviathan hypothesis
5.2.1 Empirical studies on the Leviathan hypothesis The control variables The original test The inseparability of taxing and spending decision Evidence from Swiss state and local governments References on panel data of 18 OECD countries References on panel data of 29 OECD countries
5.2.2 Summary and Conclusion
5.2.3 Critical assessment on the empirical tests Conclusion
5.3 Tax competition
5.3.1 Introduction
5.3.2 Definition of tax competition
5.3.3 Tax competition and output of local public goods Summary
5.4 Local benefits with federation-wide costs
5.4.1 Introduction
5.4.2 The model Summary
5.4.3 A simplification of the model Summary
5.4.4 A model of tax sharing and tax illusion Summary
5.4.5 Conclusion

6. Conclusion

7. Literature
7.1 Articles
7.2 Internet sources

8. Equations

9. Figures

10. Tables

1. Introduction

A large number of countries are trying to improve their ability to serve their inhabitants more efficiently and more effectively. To accomplish this ambition, a reawakening of interest in the practices and in the principles and of fiscal federalism is mandatory. Questions arise such as:

How many taxes are necessary to provide an efficient amount of local public goods? How should the taxes be allocated most reasonable between the different levels of governments?

Which level of government should have how much freedom of choice concerning tax revenues and tax expenditures, or differently, which degree of decentralisation is most constructive?

Generally, two conflicting possibilities to provide an efficient level of local public goods are existing: The first one proposes a high degree of centralisation of the tax system which would lead, if the government is totally benevolent, to an efficient output of local public goods. An argument against centralisation is that a centralised system cannot serve the different needs and preferences of the inhabitants of unequal regions. The second perception states that an efficient level of local public can be provided if the system is decentralised. A possible disadvantage of decentralisation is the appearance of tax competition which may lead to an inefficient low level of local public goods. The discrepancy between these two conflictive systems is going to be discussed in this thesis.

Furthermore, politico-economic mechanisms, which are supposed to explain the correlation between degree of centralisation and size of government are introduced which leads to the research question of this thesis:

Does a decentralised tax system lead to a smaller size of government compared to a centralised tax system?

Which politico-economic mechanisms are responsible for this correlation?

In relation to these questions, different approaches are discussed, underlined by theoretical and empirical models.

The first one, a public choice approach, states that central governments operate like monopolists, or leviathans, extracting as much tax revenues from the citizens as possible. The aim is not to maximise social welfare, but to increase their control over the resources of the economy. This approach was formulated by Brennan’s and Buchnan’s Leviathan hypothesis (1980), stating, subject to the above mentioned conditions, that a higher degree of decentralisation leads to a smaller size of government. The hypothesis was tested frequently since 1985, some of the models and their results are presented in this study.

This approach is followed by a chapter about tax competition which may appear in a decentralised system. In the literature tax competition is treated as a way to tame the Leviathan. But tax competition may also not lead to an efficient output of local public goods. The danger of a decentralised system underlying tax competition is an underprovision of local public goods. The chapter collectively treats an extract of existing literature and models, starting with the “standard tax competition model” by George R. Zodrow and Peter Mieszkowski (1986).

A second approach states that in a centralised system, which produces local benefits with federation-wide costs, an inefficient high level of taxes and provision of local public goods are the consequence. Politicians from sub-central governments can influence political actions e.g. by paying campaign contributions to provide more public goods for their own jurisdiction. Through this behaviour, called capturing, an inefficient high level of taxes is levied by the central government. A model by Persson and Tabellini (1994) explains this phenomenon in detail. A simpler version of the model was developed by Weingast et al. (1981).

An additional approach is presented in this study, which is supposed to model the Austrian situation, called "a model of tax sharing and tax illusion". In this approach the taxes are collected by the central government but the level of taxes is effecively decided by the representatives of the provinces. The model includes a behavioural component, which represents the idea that tax burdens are considered too little by the provinces as they do not have any visible responsibility imposing them.

The structure of the thesis is as follows: In chapter two, some important definitions are formulated. Chapter three describes fiscal federalism in Austria, its degree of decentralisation and the comparison to other OECD countries. Chapter four provides a short overview of various models and elements of the First and the Second Generation Theory of fiscal federalism. Chapter five switches back to the research question, starting with the Leviathan hypothesis, treating five empirical models in detail, followed by a critical assessment of the empirical tests. Afterwards, an overview of the topic of “tax competition” is provided, giving some definitions and a short summary of existing studies and models. Then the above mentioned second approach follows, explained by a model of Persson and Tabellini (1994) and a simplification of the model by Weingast et al. (1981). A third approach tries to model the main mechanism in the Austrian system. Chapter six concludes the results.

2. Definitions

2.1 Definition of Fiscal Federalism

According to Pfaundler (1931) fiscal federalism can be described as the whole complex of arrangements and facts that treats the financial relations between the main state and its integrated political subdivisions and, in addition, the relations between the political subdivisions.

Basic requirement for fiscal federalism is that some of the political subdivisions obtain the resources needed for various (public) goods through taxes.

The relations which are formed through fiscal federalism arise in groups which are arranged in subordinated order (main state, member states, communities), which is also called vertical fiscal federalism and/or in groups which are arranged in horizontal direction (horizontal fiscal federalism).

2.2 Definition of Fiscal Decentralisation and Fiscal Centralisation

Following Golem (2009), fiscal decentralisation in its purest form, implicates that sub-national governmental units have the autonomy over financing and provision of public goods and services. In such a system, the sub-national governments control their own budget without any intervention from the central government.

As reported by Porcelli (2009) p.1 the term “ can be defined as a two-dimensional policy institution that involves either decentralisation of a tax instrument, when local governments have the power to raise taxes, or decentralisation of expenditures when local governments bear the responsibility for implementing expenditure functions. ”

Marlow (1988) p.262 defines and measures the extent of decentralization in his paper “Fiscal decentralisation and government size” as the “ ratio of state and local government expenditures - to - total government expenditures. ”

Most authors agree on the opinion that fiscal decentralisation increases social welfare as the provision of public goods and services can be tailored more precisely to the needs of the respective citizens.

Fiscal centralisation assumes that financing and provision of public goods is exclusively dictated by the national governments [Feidler and Staal (2008)].

3. Fiscal Federalism in Austria

Austria is a federal state with four sectors: the federal government, nine provincial governments, many local governments and social security institutions which are not part of the system of fiscal federalism.

The financial constitutional law from 1948 regulates the intergovernmental fiscal relations and sets the principles for the fiscal equalization law. The fiscal equalization law treats the rules of tax sharing and intergovernmental transfers in detail and it determines the cost bearing between federal government, provincial governments and local governments. The fiscal equalization law is valid for four years after which it is replaced by new regulations.1

As Schneider and Dreer (2013) write, about 95% of the enforcement of the tax laws and of the levy of the taxes are carried out by the federal government, which characterizes the Austrian system as a centralised one.

Following Brunner (2015) in detail it can be described as follows:

In Austria, the distribution of tasks between federal government, provinces and communities which have to be financed by the taxes depends on their responsibility. For example, the federal government is responsible for the social security system, provinces are responsible for nature conservation and communities take care of their infrastructure.

All of these mentioned tasks are financed through taxes, whose allocation can be summarized as follows:

In Austria, there is a big difference between the regulation of revenues and expenditures of taxes. Tax revenue is mainly regulated in a centralised way which means that it is accomplished by the federal government. The most important taxes collected by the federal government are for example the corporation tax, sales tax and the income tax.

Provincial and local governments hardly have any responsibility for levying taxes. The taxes levied by the federal government are split between itself and provincial and local governments which decide how to spend them. The amount of taxes a province receives mainly depends on the number of inhabitants.

This process, the sharing of the joint federal taxes between the federal government, all the provinces and all the communities, is called vertical fiscal equalization.

By contrast, horizontal fiscal equalization governs the allocation of financial resources between governments which are on the same level.

The distribution of the taxes to each community is calculated in the following way:

In a first step, the share of taxes for the communities is split by province following the graded population key (abgestufter Bev ö lkerungsschl ü ssel), which means that the amount of taxes received for all communities in a province is higher the more population of this province is concentrated in bigger communities. From this amount, the province keeps a maximum of 7.6% for purposes such as financing province budgets (Landesumlage) and 12.7% for funding allocations (Bedarfszuweisungen). Next, communities which could not reach the average revenues of all communities in Austria in the previous year, receive 30% of the difference. Following the RIS (Rechtsinformationssystem), the remaining taxes are split up between the communities according to the graded population key which is calculated since 2015 as follows:

- In communities with less or equal 10,000 inhabitants, the real number of residents is multiplied by 1 41/67;
- In communities with 10,001 to 20,000 inhabitants, the real number of residents is multiplied by 1 2/3;
- In communities with 20,001 to 50,000 inhabitants, and in cities with own statute with a maximum of 50,000 inhabitants the real number is multiplied by 2;
- In communities with more than 50,000 inhabitants and in the city Vienna, the real number is multiplied by 2 1/3.

Depending on this "fictive" number of inhabitants, the taxes are distributed to each community.

The traditional reason for this is the "Brechtsches Gesetz" which states that denser settlement requires more regulations and then higher expenditures per head.2 Moreover bigger communities provide public goods which may also be used by the inhabitants living in the neighbouring smaller communities which may also explain higher expenditures per head in bigger communities..

Adesamer and Höferl (2004) summarize the results of this complex system of fiscal equalization in Austria which leads to the following distribution: the federal government levies about 95% of all taxes. About 60% of the tax revenues through the federal government is left to the federal government itself, the remaining 40% are split between provinces and communities.

3.1 Decentralisation in Austria compared to other OECD countries

Following Blöchliger (2013), the degree of decentralisation varies strongly across the OECD countries. The following Figure 1 shows that the sub-central expenditure share in Canada in 2011 was on average 66%, while the smallest sub-central spending share in Ireland was only 11%. Austria is approximately in the middle with about 31%.

Also sub-central tax revenue shares varie strongly across the countries. Canada's share in 2011 was about 50%, while the revenue share of Ireland was only about 2%. Sub-central tax revenues in Austria are very low with a percentage of about 6%.

Generally, spending is more decentralised than revenues.

According to the figure, the most decentralised OECD country is Canada, as it has the highest sub-central expenditure share and the highest sub-central revenue share. The most centralised country is Ireland with one of the lowest expenditure shares and the lowest revenue share. As already mentioned, Austria's degree of revenue decentralisation is very low, while it's degree of expenditure decentralisation is about five times higher.

Figure 1: Decentralisation across OECD countries in 2011

illustration not visible in this excerpt

Source: Bl ö chliger (2013) p.5

Considering the change of the degree of decentralisation from 1995 to 2011 in all OECD countries, it can be found that it has been very small over the last years. The following Figure 2 shows that the sub-central revenue and expenditure share were growing, if only slightly, the last years.

Figure 2: Change of decentralisation in the OECD, annual average, 1995 to 2011

illustration not visible in this excerpt

Source: Bl ö chliger (2013) p.5

4. Theoretical Perspectives on Fiscal Federalism

The following chapter three follows closely Oates (2005). It gives a short overview of various models and elements of the First and the Second Generation Theory of fiscal federalism.

4.1 First Generation Theory of Fiscal Federalism

This part summarizes the earlier view of fiscal federalism which prevailed in the 1950's and 1960's and explains the basic vision of the early normative theory of fiscal federalism.

One of the main elements in the First Generation Theory (FGT) is the implicit view of the working of the public sector which implies that if market failure occurs, the government would interact and correct the failures by introducing appropriate policy measures. Public economists try to diagnose the source of the failure, work out a solution for the problem and leave it to public officials to establish the remedy. The assumption is that government agencies act in public interest to maximize social welfare because of electoral pressure.

In this context the Arrow-Musgrave-Samuelson (AMS) perspective states that the aim of each level of the government is to maximize social welfare in the respective electorate. In particular, decentralised finance opens up the possibility to provide so called "local public goods", which meet the demand of the residents in the respective area more accurately, compared to a central government which provides only a uniform level of public output.

Oates (1972) formalized this proposition as "Decentralisation theorem": Two key assumptions are made:

- Each level of government is benevolent;
- With centralisation the per person levels of public good provision are uniform across jurisdictions.

The Decentralisation Theorem states that the level of government which should interact in the cases mentioned above depends on the size of regional or public good spillovers as well as on differences in preferences for public goods between the areas.

More precisely, decentralised provision is always more or at least as efficient as a centralised provision if:

- the preferences of the residents differ between the administrative units but not inside a single administrative unit;
- no externalities are existing.

In this model the exclusion of interregional externalities is problematic, and there is the argument that a regionally differing outputniveau could be allocated under central allocation as well. This argument was envisioned by the FGT as if governments provide an efficient output of public goods at different levels for goods whose "spatial patterns of benefits were encompassed by the geographical scope of their jurisdictions" [Oates (2005) p. 351], there are a number of local public goods which cannot be totally controlled in their scope, so it was recognized that this system called "perfect mapping" could hardly exist for every local public good.

The reason is that some local public goods can produce spillover effects: Imagine a road which everyone can use, if only to drive through, or a clean river from which residents of other jurisdictions benefit as well.

By a simple application of the traditional Pigouvian theory of subsidies the FGT dealt with the issue of allowing decentralised provision an approach to a centralised and uniform level of public outputs and still having welfare gains: The central government should provide subsidies (e.g. grants) to decentralised governments to internalize the benefits. The amount of local public goods would then be extended by recipient governments to the point on which marginal social benefits for the whole society equal marginal costs.

Another key element of the FGT is taxation in a federal system, known as the "tax assignment problem". It tries to answer the question of what form of taxation are best at different government levels. In this theory decentralised, especially local, levels of governments, should place a primary focus on benefit taxes which are for example property taxes. The central, or higher levels of government have a bigger scope for the use of other forms of taxes as for example progressive income taxes just as a part of a much broader program - the redistribution of income. Or in other words: Local governments should focus on benefit taxes like property taxes rather on non-benefit taxes on mobile factors, such as labour, as these could lead to distortions in the location and levels of economic activity.

In order to provide assistance to poorer jurisdictions and to correct distorted migration patterns, the FGT addresses the issue of equalizing lump-sum grants from the central to the local governments.

Another model which occupies a decisive role in the literature of local public finance is the Tiebout Model by Charles Tiebout (1956). As Oates (2006) writes, Tiebout's intent is to show that there exists a solution for the free-rider problem by so-called "local public goods" and simultaneously to challenge the claim by Paul Samuelson that decentralisation could not result in an efficient provision of public goods.

According to Tiebout, communities can be seen as a participant in competitive market for mobile citizens, or, to express it differently, tax payers. The driving force of this competition is the potential migration of individuals to the respective communities. On this basis, they have an incentive to provide the optimal amount of public goods in order to "convince" the individuals to choose the optimal community according to their desired living conditions. This distribution of individuals leads to a Pareto-efficient equilibrium in which a lot of different communities provide different levels of public goods. The problem of this behaviour is that it may lead to tax competition and an inefficiently low level of public goods- chapter 5.3 treats this aspect in detail

To summarize, in the 1950's and 1960's, a vision of fiscal federalism was produced in which the central government is the main responsibility for macroeconomic stabilization while decentralised governments had the task to provide an efficient level of local public goods.

4.2 Public-Choice Perspective of Fiscal Federalism

According to Porcelli (2009) the most important contribution to the public choice perspective of fiscal federalism is the so called "Leviathan hypothesis" by Brennan and Buchanan (1980). According to this hypothesis, central governments do not maximise social welfare. They operate like monopolists, or leviathans, with the aim to increase their control over the resources of the economy. Under this assumption, fiscal decentralisation acts as a mechanism for constraining the inefficient behaviour of the government. Empirical studies on the hypothesis are presented in chapter 5.2.

4.3 Second Generation Theory of Fiscal Federalism

The more recent literature on fiscal federalism reaches across many different disciplines with contributions of political scientists, economists, and many others.

Trying to describe the "Second Generation Theory" in a systematic way, two basic sources are decisive:

1. The literature on public choice and political economy with focus on the behaviour of political agents and political processes
2. The literature on problems of information, for example asymmetric information to mention a keyword.

Out of these two sources, a new literature on fiscal federalism emerges, in which the topic of (de)centralization appears in a different light

4.3.1 Centralization and Decentralisation in the Second Generation Model

Different approaches have been found in order to describe the topic of centralisation and decentralisation in the Second Generation Model.

One example is a paper by Seabright (1996). He uses the "incomplete contract" approach to analyze fiscal federalism. Elections are viewed as incomplete contracts in which some important information is unverifiable. In this setting "it may matter very much who has the power to action, because we should presume that agents will take actions according to their interests." [Seabright (1996) p. 64-65]. The benefit of centralisation is that it allows a greater coordination of policies while decentralisation promotes accountability. Seabright defines accountability as the increased probability that the welfare of a given region can determine the re-election of the government.

From that follows that the choice between centralisation and decentralisation depends on the "the relative importance of interjurisdictional spillover effects versus the losses from reduced accountability under central control." [Oates (2005) p.358] Even if the tastes are homogeneous, decentralisation still may be preferred as long as local control increases.

Another approach which is the basis of important papers in the SGT is the so called "Principal Agent Model". To summarize shortly, this model can be seen as a special form of a game with principals and agents, who have asymmetric information of which they can make use.

The Second Generation theory (SGT) literature has made use of this model in different ways. One example is that the central government acts as principal. The principal tries to structure intergovernmental fiscal relations in such a way that the agents, which are the regional or local governments, behave in ways that assist the aims of central officials. Asymmetric information takes place as the centre only has imperfect information over the fiscal activities of decentralised public agents.

4.3.2 Potential dangers of Fiscal Decentralisation

Although it is mentioned that with fiscal decentralisation, local governments are closer to their constituencies and are therefore able to meet the demand of the citizens more exactly, there is also some potential danger.

Fiscal decentralisation gives the opportunity to exploit the system by shifting the burdens of local programs on the whole nation. The literature on tax exporting in the FGT recognized this issue in following way: "... exposed the tendencies for decentralized finance to be overly expansive in cases where the incidence of local taxation could be shifted onto residents of other jurisdictions." [Oates (2005) p. 360]. Chapter 5.4.3 treats a theoretical model which is based on these mentioned assumptions.

In this context, the more recent literature provides a richer treatment of fiscal breakdowns which can result through exploiting the system. Kornai (1979) makes a distinction between soft and hard budget constraints. He describes the difference as follows:

"A budget constraint is hard if it is asserted with iron discipline: the firm can spend only as much money as it has. It has to cover its expenses from its incomes from sales. It is entitled to take out credit, but the bank is prepared to grant credit only under "conservative" and "orthodox" conditions. This can be, therefore, only an advance for subsequent proceeds from sales. The budget constraint is soft if the above-mentioned principles do not get asserted consistently." [Kornai (1979) p. 806]

According to this defintion, hard budget constraints are effecitve as it can be invested only the amount of money which is existing. Soft budget constraints are not effective as money can be borrowed without any risk for the person or institution getting the money.

For the purpose of fiscal (de)centralization the term refers to regional or local governments which can rely on a central- or a higher level of government to rescue them from fiscal problems.

This problem leads to the following questions: The first is why central governments cannot commit themselves credibly to avoid fiscal bailouts? What is the motivation for what in the end turns out to be destructive behaviour?

The motivation can be explained in the following way:

If the central government does not "rescue" the local government, there will be electoral consequences for central authorities as voters put blame on the center and the welfare of the locality will decrease because of low levels of public outputs. In other words: central governments commit fiscal bailouts as otherwise they may harm themselves.

Another question is what kind of economic and political instutions are able to minimize these incentives?

One option is to minimize the role of local governments by greater centralisation. Due to this there will be less scope for locals to seek fiscal help.

Features which can contribute to hard budget constraints for local governments are for example efficient credit markets in the context of a mature banking system as it disciplines the finance of local governments. If a local government performs fiscally poor the consequences are reduced access to credit and also higher interest rates. This component can encourage responsible local fiscal descision-making.

Institutions which are essential to hard budget constraints are a local effective system of taxation to ensure the financing of local programs, and a functioning system of intergovernmental grants in a way that its not subject of manipulation to provide fiscal bailout.

Especially in developing and transitional nations, such institutions do not always exist. The literature provides some kind of measures which can help to harden local budget constraints:

"(1) Constitutionally or legislatively imposed balanced-budget constraints that efficiently make it unlawful for local governments to run deficits on current account spending;
(2) Limitations on debt issues that constrain borrowing to the finance of capital projects with careful definitions of what capital spending encompasses.
(3) Well designed public bankrupcy laws that specify clearly how fiscal crisis will be handled."

[Oates (2005) p.363]

It has to be mentioned that there does not exist a general recipe, the above statements are more guidelines, as the exact form of institutions in a country depend on the historical traditions and on the specific political cultural and economic institutions in the country itself.

4.3.3 Risk- Sharing and Interjurisdictional Insurance

Another intensively discussed topic in the literature of the SGT is the role of the central government in representing an insurance function in the intergovernmental fiscal system.

A system of intergovernmental assistance, which is sensitive to imperfectly correlated stochastic shocks which reduce income, increase costs and overall reduce social welfare, can exist. The function of this system is to provide assistance to jurisdictions, which experience negative shocks which lead to one of the mentioned outcomes.

A fundamental problem is a standard moral hazard problem. It appears in an insurance model where the central government employs intergovernmental grants in order to provide risk sharing and to internalize spill over effects from outputs of local public goods. With the existence of a central insurance, the local jurisdictions are induced to underprovide outputs or programs which have the task to increase the value of the local economy's capacity to adjust to exogenous shocks.

Persson and Tabellini (1996) analyzed the trade-off between risk sharing and moral hazard. In their framework, the extent of the moral hazard problem depends on the nature of the federal fiscal constitution. They found that the moral hazard problem can be addressed better in a system where the central government provides insurance directly to individuals than a system where insurance is provided to governments- a system of confederation.


1 cf.: July 2015 6

2 cf.: 7

Excerpt out of 57 pages


Fiscal Federalism-Decentralisation and the size of government
Catalog Number
ISBN (eBook)
ISBN (Book)
File size
1416 KB
fiscal, federalism-decentralisation
Quote paper
Ina Walcherberger (Author), 2015, Fiscal Federalism-Decentralisation and the size of government, Munich, GRIN Verlag,


  • No comments yet.
Read the ebook
Title: Fiscal Federalism-Decentralisation and the size of government

Upload papers

Your term paper / thesis:

- Publication as eBook and book
- High royalties for the sales
- Completely free - with ISBN
- It only takes five minutes
- Every paper finds readers

Publish now - it's free